PLACE N RETURN BOXtomnovoflibchockomnom you: record.
To AVOID FINES Mom on o: More duo duo.

DfiEE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU Is An Mmuflvo ActiorVEqnl Oppommy Institution
WM

 

i
E
g
E

F .

 

 

LIBRARY
Michigan State
University

 

 

 

This is to certify that the

thesis entitled

Experimental Observations of Composting

Dairy Manure Solids , '
presented by i '

Stephen Earl Ferns

has been accepted towards fulfillment
of the requirements for

 

 

Masters degreein Agric. Engr.

 

Dmt_flovember 25, 1987

0-7639
MS U is an Affirmative Action/Equal Opportunity Institution

 

EXPERIMENTAL OBSERVATIONS OF COMPOSTING
DAIRY MANURE SOLIDS

BY
Stephen Earl Ferns

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE
in
Agricultural Engineering

Department of Agricultural Engineering

1987

 

ABSTRACT

EXPERIMENTAL OBSERVATIONS OF COMPOSTING
DAIRY MANURE SOLIDS

BY

Stephen Earl Ferns

Dairy manure solids (DMS) were composted in a batch
windrow process of 3 weeks duration with weekly turnings.
Diffusion or natural convection provided heat and mass
transfer. Temperature, gas concentration (02 & C02),
ambient conditions, insitu physical properties and pile size
were monitored. A temperature and gas probe and a rotary
corer developed show promise for future insitu sampling in
composting systems of this type. Sensitivity analyses were
performed on the experimental methods and suggestions for
improvements were made.

Results show that the assumption of spatially constant
physical and thermal properties used in a previous model of
distributed heat and mass transfer in compost windrows is
not justified in all cases. Significant changes in physical
properties over time were not observed.

DMS compost behaved similarly to other compost
substrates. However, DMS composted at higher moisture
contents and lower maximum temperatures were observed than

with other compost substrates. Distinct color zones that

 

_. #_.~—_fi.._

correlated with time-temperature patterns were observed in
the windrow cross-sections. Analysis of time-temperature
patterns indicated areas where either significant thermal
death or regrowth of mastitis-causing coliform organisms
could occur. Several methods were discussed for managing

the DMS composting process in order to produce acceptable

Mam

Major Professor

a AKA

Major Professor

Department Chairman

dairy bedding material.

. _. _—«.__ __. _

To my wife Lori.

iv

ACKNOWLEDGMENTS

Many individuals have contributed to the success of my
MS studies and research at MSU. My Co-Major Professors, Dr.
John Gerrish and Dr. Ted Loudon provided direction,
assistance and support. Dr. Gerrish was an inspiration in
his ability to creatively solve research problems. Dr.
Loudon provided pactical research advice and was generous in
his sharing of knowledge on soil and water engineering.

Dr. Fred Bakker-Arkema served on my committee and provided
insight into heat and mass transfer processes and vigorously
tried to move me along in my program. Dr. Larry Segerlind
graciously discussed applying Finite Element modeling to DMS
composting and served on my committee. One of my biggest
disappointments was not being able to actually use this in
the thesis.

Dr. John Speicher and Dr. Roger Mellenberger provided
invaluable insights into practical dairy management and the
nature of the coliform mastitis problem.

Bob Gardner, a fellow MS student from the Animal
Science Department, offered invaluable moral and technical
support: there was at least one other person in the
university crazy enough to play in the DMS.

The Kellogg Biological Station dairy manager, Rob

Ashley and his staff cheerfully put up with many somewhat

V

unusual requests and created a supportive research
environment.

Several individuals very generously shared their
laboratory space and time. These people included Dr.
Mackenzie Davis, of the Environmental Engineering
Department; Dr. David Dilley and Dr. Steve Sargent of the
Horticultural Department; and Dr. James Tiedje and Peter

Groffman of the Crop and Soil Science Department.

vi

TABLE OF CONTENTS

Page

List Of TablesO O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O OXiii
List Of FigureSO O O O O O O O O O O O O O O O O O O O O O O O O O O O O‘O O O O O O O O OXVi
WER 1: IntrOductionO O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O 1

1.1: Criteria for Use of Dairy Manure Solids as a
Bedding MaterialOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

1.2: Effects of Mastitis on the Dairy Industry........

10201: PhYSical EffeCtSOOO O OOOO OOO O O O OO O O O O O O O O O O
1O 2 O 2: Economic EffeCtSo O O O O O O O O O O O O O O O O O O O O O O O O O
1.3: Micro-organisms Associated with Mastitis.........

1.3.1: General Organisms.........................
10302: COliform organismSOOOOOOOOOOOOOOOOOOOOOOOO

1.4: Treatment of Dairy Manure Solids for Bedding.....

1.4.2: Treatment by Composting...................

1.5 The Composting Process...........................

Definition................................
Parameters Affecting Composting...........
Spatial Variability.......................
Need for Distributed Heat and Mass

Transfer Composting Model.................
Modeling Composting Heat and Mass

Transfer Processes........................
: Recent Advances in Composting. ..... . ....

.1
2
2
3
3
3
3
5
1.4.1: General Treatment Methods.................5
5
6
6
7
7
7
8
9

CHAPTER 2: Objectives................................ll

CHAPTER 3: ‘Literature ReVieWOOOOOOOOOOOOOOOOOOOOOOOOOlz
3.1: Description of Composting.......................12

vii

3.5:

Analogous Systems...............................13

MicrObiOIOQYOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO14

30301
30302

GeneralOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO14
Types of Micro-organisms.................lS

l: Bacteria and Actinomycetes........15
2: FungiOOOOOOOOOOOOOOOOOOOOOOOOOOOOO16

Compost Microbial Ecology........ ...... ..l7

. : General...........................l7
. : Temperature.......................18
. : Moisture Content and Water
Activity..........................27
: pH................................29
: Redox Potential...................31
: Compost Microbial Ecology
Studies...........................31

BiOChemistrYOOOOOOOOOOOOOOOOOOOOOO OOOOOO O OOOOO OO35

3.4.1:
3.4.2:

StOiChiometrYOOOOOOOOOOOOO OOOOOOO OOOOOOOO35
compost StUdleSOOOOOOOOOOOOOOOOOOOOOOOOOO36

Organic Substrates................36
Nitrogen Transformations..........38
Effect of Carbon Source...........41
Effect of Inorganic Materials.....4l

tthH

mpirical StUdieSOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO43

UIUIU'IUIU‘
o. coo

4»wa“
oo oo o

MDWNH

Temperature..............................43
Oxygen and Carbon Dioxide................43
M01sture Content.........................43
Windrow Size.............................47
Windrow Zones............................48

Kinetic Modeling................................49

Microbial Growth.........................49
Microbial Death..........................51
Regrowth................... ..... .........55
Combined Equations.......................56

Physical Properties..... ............ ............58

U'IPU-DNH

Particle Size......... ........ ...........58
Particle Density.........................59
Bulk Density.............................59
Derived Quantities.......................59
Compressibility... ................. ......60

viii

3.7.6: Settlement Behavior......................60

3.8: Heat Transfer...................................61
3080l: DifoSionOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO61
3.8.1.1: Thermal Conductivity... ..... ......61

3080102: SpeCifiC HeatOOOOOOOOOOOOOOOOOOOOO66

3.8.1.3: Thermal Diffusivity...............67

3.8.2: Natural convectionOOOOOOOOOOOOOOOOOOOOOOO67

309: Mass TranSferOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO67

3.9.1: Windrow Void Space Transfer...... ...... ..68
3.9.1.1: DifoSionOOOOOOOOOOOOOOOOOOOOOOOOO68
3 O 9 O 1 O 2: Evaporation. O O O O O O O O O O O O O O O O O O O O O O 69
3.9.1.3: LiquidSOOOOOOOOOOOOOOOOOOOOOOOOOOO70

3.9.2: Transport in Liquid Films................71

3.10: Complete Heat and Mass Transfer Models.........73

CHAPTER 4: Experimental Methods......................75
4.1: Data Requirements...............................75
4.2: General Experimental Design and Methods.........75
4.3: Probe Design and Placement................ ..... .78
4.4: Temperature Measurement.........................83

4.4.1: Equipment and Method.....................83
4.4.2: Error analysis and calibration...........85

4.5: Gas Sampling and Analysis.......................87

: Equipment and Method.....................87
: Error AnalYSiSOOOOOOOOOOOOOOOOOOOOOOOOOOO90

4.5.2.1: Effect of Air Drawn into Sample
from Outside of Sampling Area.....91
4.5.2.2: Effect of Withdrawing Successive
Samples on Changes in Gas

Concentration....... .............. 94
4.5.2.3: Effect of Leaks into Gas
Sampling System........ ........... 96

4.5.2.4: Effect of Diffusion and/or Gas

Flow into Syringe and Gas

Storage Containers.......... ...... 97
4.5.2.5: Gas Chromatograph Calibration

ix

4.6:

4.10:
4.11:

CHAPTER 5:

5.1:
5.2:
5.3:

5.4:
5.5:
5.6:
5.7:
5.8:
5.9:

and ErrorSOOOOOOOOOOOOOOOOOOOOOOO104

Core Sampling Method...........................106

Review of Sampling Methods and Corer
DeSignOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO106
Comparison of Core Sampling Methods.....109
Description of Experimental Core

Sampling Method.........................112
Moisture Content and Volatile Solids...........115

Methods and Ca1cu1ations................115
Sensitivity Analysis....................118

4.7.1:
4.7.2:
31.11]!DenSitY.....o.............................122

calcu1ationSOOOOOOOOOOOOOOOOOOOOOOOOOOOO122
senSitiVity AnalYSiSOOOOOOOOOOOOOOOOOOOO122

4.8.1:
4.8.2:

Porosity and Free Air Space....................127

calCUIationSOOOOOOOOOOOOOOOOOOOOOOOOOOOO127
Sensitivity Analysis....................129

4.9.1:
4.9.2:

Additional Physical Properties................135

Windrow Size ChangeSOOOOOOOOOOOOOOOOOOOOOOOOOO138

Experimental Results..... ......139

OOOOOOOOOOOOOOOOO139

General Observations..

Windrow Size and ShapeOOOOOOOOOOOOOOOOOOOOOOOOO139

TemperatureOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO146
Temperature Variability.................146

Temperature Patterns....................149
Time-Temperature Relationships..........157

UIU'IU'I

.3.1:
.3.2:
.3.3:

Gas concentrationSOOOOOOOOOOOOOOOOOOOOOOOOOOOOO157

Moisture Content................... ............ 173

VOIatile SOlidSOOOOOOOO OOOO17'7

Wet Bulk Density ....... ........

POtOSitYooooo-ooo

Free Air Space......... ..... . .................. 191

 

5.10: Additional Moisture and Physical Properties...197

5.11: Windrow COlor ChangeSOOOOOOOOOOOOOOOOOOOOOOOOO197

CHAPTER 6: DiSCUSSiOnOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO212

6.1: Experimental Methods...........................212

6.1.1: Probe Design............................212
6.1.2: Temperature.............................212
6.1.3: Gas Concentration.......................213
6.1.4: Gravimetric Moisture Content............215
6.1.5: Volatile Solids.........................215
6.1.6: Dairy Manure Solids Sampling and Bulk
Density.................................216
6.1.7: Porosity and Free Air Space.............216
6.1.8: Dairy Manure Solids P1acement...........217
6.1.9: Windrow Size Measurements...............217
6.2: Experimental Results..... ....... ...............217
6.2.1: Temperature.............................217
6.2.2: Gas Concentration.......................218
6.2.3: Volatile Solids.........................223
6.2.4: Moisture Content........................225
6.2.5: Bulk Density............................226
6.2.6: Additional Derived Physical Parameters..227
6.2.7: Thermal Properties......................229
6.2.8: Windrow Size Changes....................232

6.3: Reaction and Transport Characteristics.........232
60301: Type Of ReactionOOOOOOOOOOOOOOOOOOOOOOOO232
6.3.2: Unreacted Core Model....................233
6.3.3: Pore Space Heat and Mass Transfer.......235

6.4: Pathogen surVivalOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO237

6.5: Application of Composting to Dairy Manure
Solids for Production of Bedding Material......249

6.6: Preliminary Assessment of Heat and Mass
Transfer Models for Composting Dairy

Manure Solids....................... ..... . ..... 250
CHAPTER 7: Conclusions...... ......... .. ........ .....252
CHAPTER 8: Directions for Future Research ....... ....255
8.1: Experimental Methods...... ...... . ........ . ..... 255

xi

8.2: Windrow

8.3: Compost

APPENDICES
Appendix
Appendix

Appendix

Appendix
Appendix
Appendix
Appendix

Appendix

Compost Processes........... ........... 257

Heat and Mass Transfer Modeling........258

Temperature Teasurement Error
Calculations.......................260
Sample Thermocouple Water Bath
Calibration Data........ ..... ......264
Effect of Sequential Gas Samples

and Location on Carbon Dioxide

and Oxygen Concentration...........269

' Gas Concentration Correction

Factor Calculations................270
Volumetric Gas Standards Error
Calculations.......................283
Preliminary Bulk Sampler
Experiments........................286
Derivation of Error Propagation
Equations..........................293
Windrow size, temperature, gas
concentration, and physical

property data and sample location
availablity........................296

ReferenceSOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO297

xii

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table
Table
Table

Table

Table

Table

3.1:

LIST OF TABLES

Growth temperatures for micro-

organismSOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOlS

910 values for garbage composts.......19

Decimal reduction times for selected
micro-organisms...........................23

2 values for selected micro-organisms.....23

Approximate minimum levels of water
activity permitting growth of micro-
organisms at temperatures near optimal....29

The limits of pH allowing initiation
of growth by selected micro-organisms.....30

Confidence levels that all material

will obtain a temperature equal to or
greater than a particular temperature

for a desired number of days..............55

Experimental values of diffusion
coefficients in gases at one

atmosphereOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO68

Diffusion coefficients at infinite
dilution in water at 25°C.............73

Sampling methods and schedule.............76

Dairy characteristics.....................77
Temperature measurement errors. ....... ....86

Water bath calibration of 32
thermocouples................... .......... 87

Potential sources of error in gas
sampling, storage and analysis...... ...... 90

Internal volume of sampling tubes and
diffusion chamber...... .. ...... 95

xiii

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

4.7:

4.8:

4.9:

6.2:

6.3:

Gas concentration standards used in gas
Chromatograph calibration................105

Comparison of moisture contents
expressed on a wet and dry solids
baSiSOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOll7

Comparison of volatile solids content
expressed on a dry and ash solids
baSiSOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO118

Sensitivity analysis for moisture
content expressed on a dry

Selids baSiSOOOOOOOOOOOOOOOOOOOOOOOOOOOO121

Sensitivity analysis for volatile
solids expressed on a dry ash and
salids baSiSOOOOOOOOOOOOOOOOOOOOOOOOOOOO121

Values of constants A and B from dairy
manure solids specific gravity
measurementSOOOOOOOOOOOOOOOOOOOOOOOOOOOO128

Partial derivatives of porosity
calCUIationOOOOOOOOOOOOOOOOOOOOOOOOOOOOO130

Error levels of contributing factors to
porosity and free air space
varianceOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO13]-

Partial derivatives of free air space
calcu1ationOOOOOOOOOOOOOOOOOOOOOOOOOOOOO134

Summary of windrow weight and size.......l41

Time-temperature data for monitoring
locations in Windrow l...................158

Time-temperature data for monitoring
locations in Windrow 2...................159

Windrow color changes....................209
Windrow penetration depths...............211

Effect of windrow physical property
variability on the calculated thermal
properties in Windrow lB.................23l

Length of time spent in selected
temperature ranges by location and

windrow........ 239

Temperature penetration time as a

xiv

————

 

Table

Table

Table

Table

Table

Table

Table

Table

Table
Table
Table
Table

Table

Table

6.4:

A.1:

A.2

B.l

C.1:

D.4:
D.5:
D.6:
E.1:

F.l:

F.2:

function of clump diameter and thermal
properties...............................246

Effect of initial decimal reduction,
fraction of lethal temperature and

thermal death coefficent on the number

of windrow turns required....... ......... 247

Limits of error for thermocouple wire
(Reference Junction at 0°C...........261

Summary of sources of error in
thermocouple measurements................263

Water bath thermocouple calibration
--Run lOOOOOOOOOOOOOOOOOOOOOOOOOOOOO OOOOO 264

Effect of sequential gas sample and
location on carbon dioxide and
oxygen concentrations....................269

Gas Chromatograph trace length as a
function of withdrawal number............272

Oxygen concentration as a function of
withdrawal number........................273

Effect of initial pressure, sample

container volume, and sample size

on sample container and syringe

internal pressure after sampling.........273
Low error intermediate calculations......275
Final gas correction calculations........277
Gas concentration calculations...........280
Concentration and syringe use data.......284
Effect of sampler type and sample

depth on calculated bulk density,

porosity and free air space......... ..... 286

Statistical analysis of bulk density
and free air space data .................. 292

K V

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

LIST OF FIGURES

Hypothetical bacterial growth curve
(From Olson and Nottingham, 1980)........l4

Effect of temperature on the

generation time of a typical mesophile
Escherichia coli) (Adapted from Olson

and Nottingham, 1980)....................19

Generalized types of survivor curves
observed in studies of heat

inactivation of microorganisms (From

Hang, 1980).coco-coo00.000000000000000.0.25

Observed oxygen consumption rates for
various composting mixtures and reactor
types as a function of temperature.

Each curve represents the best fit of
observed data (From Haug, l980)..........44

Effect of moisture content on oxygen
consumption rates of various composting
substrates (From Haug, 1980).............46

Section through a mushroom compost heap
illustrating the typical differences in
temperature and aeration (From Lambert,
1941)OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO48

"Apparent” thermal conductivity of

various materials at different

densities. A, kapock; B, baggasse;

C, cork board; D, slag wool;

E, mineral wool (From Pratt, 1969).......63

Effective conductivity of porous

material due to latent heat transfered

by movement of water vapor (From Pratt
1969)....................................66

Conceptual illustration of mass
transport of major components during
composting (From Haug, 1980).............72

Probe for sampling temperatures and

xvi

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

4.11:
4.12:

gas showing alternating wooden and

copper sections. Probes had one to

four perforated copper diffusion

chambers depending on location...........82

Typical temperature and gas sampling
locationSOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO83

Orthogonal projection of test windrow
showing temperature and oxygen probe
placement. Windrow height and base
dimensions are typical...................84

Gas sampling apparatus. Valve is in
position to allow sample to be drawn
from diffusion chamber in pile...........88

Cross-section of probe insertion hole
showin annular void space created by
probe 1nsertion..........................92

Sampling points for sequential gas
samplingOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO95

Sequential sample gas concentration,
Location lOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO96

Effect of cummulative gas withdrawal on
Oxygen concentrationOOOOOOOOOOOOOOOOOOOOO99

Effect of sample pressure (calculated)
on oxygen concentration..................99

Percent influx as a function of
internal pressure......................100

Rotary core sampler....................113

Core sample locations. Cylindrical

in shape, the center of the sample was
located at approximately the same

point in the y-z plane as the
temperature/gas sampling point.........114

Effect of length errors on total
variance as a percent of total
VOIUmEOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO124‘

Effect of changes in radius variance
on total variance as a percent of
total v01umeOOOOOOOOOOOOOOOOOOOOOOOOOOO124

Effect of two different radius errors
on total variance as a percent of bulk

xvii

Figure

Figure

Figure

Figure

Figure

Figure

Figure
Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

density measurements...................126

Relative contribution of various

sources of error to total variance

as a percent of bulk density
measurements...........................126

Relative contribution of radius and

height errors to total variance as a
percent of bulk density percent
measurement............................127

Total porosity variance for 3 levels
Of factor varianceOOOOOOOOOOOOOOOOOOOOO132

Contributions to total porosity
variance due to high factor variance...133

Total free air space variance for 3
levels of factor variance..............136

Contributions to total free air space
variance due to high factor variance...136

Ambient weather conditions..............140
Cross-sections of Windrows 1A and 2A....141

Initial elevations for front half of
Windrow 18 (plan view)..................143

Elevation changes in front half
Windrow 18 after one week...............l44

Cross section of Windrow lB elevation
ChangeSOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO14S

Cross section of windrow showing

distances from a monitoring point to
windrow surface along the probe (Dp)

and the nearest surface point (Ds)......145

Measurements of temperature
variability in Windrow 2................147

Temperature profiles for Locations 1,

2’ and 3' Windrow1....OOOOOOOOOOOOOOOOOlSO

Temperature profiles for Locations 4,
5' ands, Windrowloo-00000000000000....150

Temperature profiles for Locations 8,
9, 10 and Slab, Windrow 1..............151

xviii

 

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

5.11:

5.12:

5.13:

5.14:

5.15:

5.16:

5.17:

5.18:

5.19:

5.20:

5.21:

5.22:

5.23:

5.24:

5.25:

5.26:

for Windrow
at noon......152

profiles
on Day 1

Temperature (°C)
1A cross-section

for Windrow
at noon......153

profiles
on Day 3

Temperature (°C)
1A cross-section

for Windrow
at noon......154

profiles
on Day 5

Temperature (°C)
1A cross-section

Final temperature (°C) profiles for
Windrow 1A cross-section on Day 7......155

Final temperature (°C) profile for
Windrow 2COOOOOOOOOOOOOOOOOOOOOOOOOOOOO156

Time spent in each temperature class
by location in Windrow 1A. Total
time for windrow was 6.96 days.........160

Time spent in each temperature class
by location in Windrow 13. Total
t1me for windrow was 6.89 days.........161

Time spent in each temperature class
by location in Windrow 1C. Total
time for windrow was 6.81 days.........162

Time spent in each temperature class
by location in Windrow 2A. Total
t1me for windrow was 5.80 days.........l63

Time spent in each temperature class
by location in Windrow 2B. Total
time for windrow was 6.84 days.........164

Time spent in each temperature class
by location in Windrow 2C. Total
t1me for windrow was 6.88 days.........165

Gas concentrations at Location 10
showing suspect high oxygen levels.....166

Difference between original and
corrected oxygen concentration for
selected dataOOOOOOOOOOOOOOOOOOOOOOOOOO167

Oxygen concentration data for
Windrow lOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO168

Oxygen concentration data for

Windrow 2000OOOOOOOOOOOOOOOOOOOOOOOOOOO169

Carbon dioxide concentration data
for Windrow 1OOOOOOOOOOOOOOOOOOOOOOOOOO170

xix

 

 

Figure
Figure

Figure
Figure
Figure

Figure
Figure

Figure
Figure

Figure
Figure
Figure
Figure
Figure

Figure

Figure

5.27:

5.28:

5.29:

5.30:

5.31:

5.32:

5.33:

5.34:

5.35:

5.36:

5.37:

5.38:

5.39:

5.40:

5.41:
5.42:

Carbon dioxide concentration data
for Windrow 2OOOOOOOOOOOOOOOOOOOOOOOOOO171

Gas concentrations by location,
Windrow 13' Day 3000OOOOOOOOOOOOOOOOOOO172

Gas concentrations over time, Location
1. Lines indicate general trends in
data, not intermediate values..........172

Gas concentrations over time, Location
3. Lines indicate general trends in
data, not intermediate values..........l74

Gas concentrations over time, Location
10. Lines indicate general trends in
data, not intermediate values..........174

Gravimetric moisture content (% wb)
by location for Windrow 1..............175

Gravimetric moisture content (% wb)
by location for Windrow 2..............l76

Average gravimetric moisture content
(% wb) by location and turning for
Windrows l and 2.......................177

Average volatile solids content (% db)
by location and turning for Windrows
land 20.000000.00000000000000000000000178

Volatile solids content (% db) by
location for Windrow 1.................l79

Volatile solids content (% db) by
location for Windrow 2.................180

Wet bulk density (kg/m3) by location
forWindrow1OOOOOOOOOOOOOOOOOOOOOOOOOO183

Wet bulk density (kg/m3) by location

for Windrow 2000OOOOOOOOOOOOOOOOOOOOOOO184

Average wet bulk density (kg/m3)

by location and turning for
w1ndrow51and 2000OOOOOOOOOOOOOOOOOOOOlBS
Wet bulk density changes over time.....187

Wet bulk density changes over time by
locationOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO187

 

Figure

Figure
Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure
Figure

Figure

Figure

5.43:

5.44:
5.45:
5.46:

5.47:
5.48:

5.49:

5.50:

5.51:

5.52:
5.53:

5.54:

5.55:

5.56:

5.57:

5.58:

5.59:
5.60:
5.61:

5.62:

Wet bulk density changes over time
greater than 11 % coefficent of
variationOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO188

Porosity by location for Windrow 1.....189
Porosity by location for Windrow 2.....190

Average porosity by location and
turning for Windrows l and 2...........191

Porosity changes over time.............l92

Porosity changes over time by
locationOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO0192

Free air space by location for
Windrow1000OOOOOOOOOOOOOOOOOOOOOOOOOOO193

Free air space by location for
Windrow 2000OOOOOOOOOOOOOOOOOOOOOOOOOOO194

Average free air space by location
and turning for Windrows l and 2.......l95
Free air space changes over time.......l96

Free air space changes over time
by locationOOOOOOOOOOOOOOOOOOOOOOOOOOOO196

Free air space changes over time
greater than 7 % coefficent of
variationOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO197

Volumetric moisture content by
location for Windrow l.................198

Volumetric moisture content by
location for Windrow 2000OOOOOOOOOOOOOO199

Fractional solids content by
location for Windrow l.................200

Fractional solids content by

location for Windrow 2.................201
Void ratio by location for Windrow 1...202
Void ratio by location for Windrow 2...203

Degree of saturation by location for

Windrow 1..... .204

Degree of saturation by location for

xxi

Figure 5.63

Figure

Figure

Figure 6.2

Figure 6.3:

Figure 6.4

Figure

Windrow 2000OOOOOOOOOOOOOOOOOOOOOOOOOOO205

Windrow 1 cross-section showing color
zoneSOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO206

Windrow 2 cross-section showing color
zoneSOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO207

Summary of assessment of time-
temperature effects on microbial
surVival and growt'hOOOOOOOOOOOOOOOOOOOOO240

Windrow 1B temperature and 210918
2, 3

reduction at Locations ...... 42

Windrow lB temperature and 1091
reduction at Locations 4, 5, an 6......242

Windrow lB temperature and 1091
reduction at Locations 8, 9, an 10.....243

Decimal reductions in Windrow 18 by
locationOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO244

xxii

 

CHAPTER 1

INTRODUCTION

In the last decade the use of separated dairy manure
solids as a bedding material for dairy cattle has received
increasing attention and use. Depending on herd size and
the availability and cost of other bedding materials, the
separated dairy manure solids can provide an economic
alternative (Keys et a1., 1977; Allen et al., 1979). The
direct economic benefits in dairies using liquid flush
manure handling systems can be augmented by savings from
decreased labor in the manure flushing system, the
construction of smaller waste storage lagoons and less
frequent removal of manure solids (Hermanson, 1985;
Lindemann, 1985). Some studies have shown that cows prefer
separated manure solids to other types of bedding (Keys et

a1, 1977).

1.1: Criteria for Use 95 Dairy Manure Solids as a Bedding
MateriaI

In order to be a satisfactory bedding material, the
dairy manure solids must meet several criteria. Moisture
content affects general cow cleanliness and the ease of
handling the bedding material. More importantly, high
levels of certain micro-organisms in bedding material have

been associated with mastitis in dairy cows (Bramely and

 

Neave, 1975).

1.2: Effects of Mastitis on the Dairy Industry

 

 

1.2.1: Physical Effects

 

Mastitis is an infection of the mammary gland. In
dairy cows it is characterized by swelling, heat, redness,
pain and disturbed function. Clinical signs of the disease
are a result of the cow's defense mechanism attempting to
destroy invading micro-organisms, return milk production to
normal and repair the damaged gland (Jarrett, 1981).
Mastitis can begin with injury to the teat or the entry of
pathogenic bacteria into the teat (Bramley, 1974; Carroll,
1977). The entry of pathogenic bacteria can be affected by
the milking routine, disease prevention procedures, faulty
milking machines and the sanitary quality of the cow's
environment, ie. the bedding and water supply.

The severity of mastitis attacks can range from
chronic, low-level infections which affect the quality and
quantity of the milk produced, to peracute in which all
signs of inflammation are present, together with fever,
depression, shivering, loss of appetite and rapid weight
loss (Jain, 1979). Peracute mastitis can lead to the cows'
death within hours; cows that survive may not recover their
milk production and must be culled.

Mastitis severity is related to the type of bacteria
and the resistance of the host (Jarrett, 1981). The
pathogenicity of coliform organism serotypes, for instance,

varies widely (Carroll et al., 1973). Dairy cows are

 

3
particularly susceptible to coliform mastitis just prior to
parturition and in the early stages of lactation (Bramely,
1974; Eberhart & Buckelew, 1972). 3

1.2.2: Economic Effects

 

Economic losses from bovine mastitis can be serious.
‘Estimates of the United States' annual losses due to
mastitis during 1976 were $1,294 million (Blosser, 1979).
Reduced production and discarded milk accounted for $897
(69.3 %) and $142 (11.0 %) million, respectively. Increased
replacement costs due to death and early culling caused a
$103 (8.0 %) million estimated loss.

1.3: Micro-organisms Associated with Mastitis

 

1.3.1: General Organisms

 

A wide variety of micro—organisms has been associated
with mastitis. Streptococcal and staphylococcal organisms
account for many cases of chronic, low level mastitis that
reduce the quantity and quality of milk production (Nat.
Masitis Coun., 1978). Coliform organisms have been
implicated in outbreaks of peracute mastitis and the
subsequent death of cows (Nat. Mastitis Coun.,ll978).
Bovine mastitis produced by Cornebacteria (Packer, 1977a),
Psuedomonas aeruginosa (Packer, 1977b), mycoplasma (Jasper,
1977) and yeast-like fungi (Farnsworth, 1977) have also been
reported.

1.3.2: Coliform Organisms

 

 

Mastitis caused by coliform organisms is of particular

concern when dairy manure solids are used as bedding.

4

"Coliform" loosely refers to lactose-fermenting organisms of
the family Enterobactericeae, including the genera
Escherichia, Klebsiella, and Enterobacter (Eberhart, 1977).
Coliform organisms are prevalant in the dairy environment
(Brander, 1973). Levels of coliform organisms greater than
1x106 have been associated with outbreaks of mastitis
(Bramely, 1974; Bramely and Neave, 1975). This level of
coliform organisms occurs in many types of fresh and used
bedding in dairy environments (Bramely and Neave, 1975;
Neave and Oliver, 1962; Jasper et al., 1975, Carrol and
Jasper, 1979; Bishop et al., 1980; Rendos et al., 1975).
High-producing cows tend to spend longer times in contact
with bedding materials than low producers (Francis, 1981);
this leads to high temperatures and increased coliform
growth in bedding used by high producers (Francis, 1981).

The existance of high coliform organism levels does not
not necessarily lead to coliform mastitis, however.
Coliform mastitis is considered a disease of glands not
affected by other pathogens (Jain, 1979). Colonies of
Staphylococcus aureus and Streptococcus agalactia adhere
much more easily to the teat than do coliform organisms
.(Frost, 1977). In the absence of specific therapy, colonies
of S. aureus and S. agalactia grow into the cow's teats and
produce low levels of chronic mastitis (Jain, 1979).
Coliform infections are difficult to superimpose upon other
infections. 500,000 leukocytes/m1 of foremilk from any

irritant, will usually prevent the establishment of

5

experimental innoculums of coliforms (Schalm et al., 1971).

Significant progress has been made in the development
of programs to reduce Streptococcus agalactiae and
Staphylococcus aureus (Carroll, 1977). These programs for
teat dipping and therapy of non—lactating cows are not
effective against coliform organisms, however (Carroll,
1977; Eberhart, 1977). The removal of the staphylococcal
and streptococcal infections opens the way for infection by
coliform organisms.

1.4: Treatment of Dairy Manure Solids for Bedding

 

 

1.4.1: General Treatment Methods

 

 

One means of reducing the incidence of coliform
mastitis is to treat the bedding material so that it will
not support high levels of the organisms. Early work with
paraformaldehyde indicated that its application would
initially reduce levels of coliform organisms (Bramely and
Neave, 1975). Recovery of coliform numbers was swift,
however, and the economics of frequent chemical applications
was judged to be prohibitive (Bramely and Neave, 1975). The '
use of lime additions to raise pH levels has been used with
some success (Smith, 1985).

1.4.2: Treatment by Composting

 

 

The application of composting to reduce coliform
numbers and produce a relatively inert bedding material from
separated dairy manure solids has also been attempted
(Carroll and Jasper, 1979; Bishop et al., 1980; Allen et

al., 1980). Composting has been shown to reduce the numbers

6
of mastitis-causing pathogens below the critical level of
1x106 (Carroll and Jasper, 1978; Bishop et al., 1980; Allen
et al., 1980). While regrowth of mastitis causing organsism
has been noticed in compost used as bedding (Bishop et al.,
1981), similar levels of regrowth have also been shown in
other bedding materials (Rendos et al., 1975). Furthermore,
the compost used in these experiments was not subjected to
an intensely managed compost process such as a force
aeration system.

Coliform control in composting dairy manure solids can
be achieved in two manners. Temperatures inside compost
piles of refuse and sewage sludge reach 80°C (Wiley, 1957;
Shultze, 1962; Willson et al., 1980). Dairy manure solid
compost temperatures up to 55°C have been reported (Allen et
al., 1980). Bramely and Neave (1975) found that
maintaining a temperature of 50°C in sawdust bedding samples
killed the organisms, while 30 and 40°C temperatures led to
increased numbers. Secondly, if the pile is correctly
managed, the substrate should be reduced, thus inhibiting
regrowth.

1.5: The Composting Process

 

1.5.1: Definition

 

 

Composting is a microbial self-heating process in which
heat given off during aerobic respiration, in conjunction
with reduced local oxygen concentration, depleted substrate,
or extreme moisture level, tends to limit microbial growth

(Finstein et al., 1980). Heat transfer and mass transfer of

7
oxygen, water and carbon dioxide is governed by diffusion,
natural convection, forced convection or by a combination of
processes.
1.5.2: Parameters Affecting Composting

Research on composting has focused on empirical
descriptions of the overall composting process and the
impact of selected parameters on the desired objectives
(Snell, 1957; Schulze, 1962; Jeris and Regan, l973a,b,c).
Temperature, oxygen level, moisture content, free air space
and substrate properties, such as the amount of volatile
solids present and the carbon/nitrogen ratio, have been
found to be important parameters.

1.5.3: Spatial Variability

 

The distribution of temperature and oxygen in a compost
pile has been found to be non-uniform (Lambert, 1941).
Furthermore, Carrol and Jasper (1978) have observed that the
distribution of mastitis-causing organisms within a
composting pile varies with time and temperature.

1.5.4: Need for Distributed Heat and Mass Transfer
Composting Modelf

 

Because of the spatial variability and the difficulty
in obtaining time-temperature relationships, it is difficult
to predict the overall level of mastitis-causing organisms
in a compost pile based on grab samples. This is
particularly true when different parts of a compost pile are
mixed together and used as bedding material. For example,
if 90 % of the compost has coliform levels of 1x103 per gram

of bedding, and only 10 % has levels of 1x107, the

I.

thoroughly mixed pile will still have dangerous levels of
1x106 per gram of bedding. Extensive experimentation is
necessary to determine the correct pile size and shape to
achieve uniform reductions in mastitis-causing organisms. A
model that predicts spatial and temporal changes in
temperature would show where in the pile treated bedding
material could be obtained.

1.5.5: Modeling Composting Heat and Mass Transfer Processes

 

 

Little work has been done to model composting.
Comprehensive simulations of continuous-feed completely-
mixed and batch models of the composting process have been
developed by Haug (1980). Finger (1975) developed a
distributed heat-and mass-transfer model for windrow
composting that is based on constant property thermal heat
conduction and the diffusion of oxygen into the compost
pile.

Modeling spatially-distributed heat-and-mass—
transfer processes in a compost pile can be improved.

Haug's models, while sound in theory, are not distributed
models. Finger's model, while a distributed model, has four
potentially serious shortcomings: (1) it does not take into
account the reduced respiration that occurs at high
temperatures; (2) it ignores the role that water could play
in the heat-and-mass-transfer process, (3) it is based on
constant physical and thermal properties, and, (4) it can be

applied only to rectangular geometries.

9

1.5.6: Recent Advances in Composting

 

 

 

Recent research has indicated the nature of the
shortcomings of Finger's model. McKinley and Vestal (1984)
found that the maximal respiration for sewage sludge
composting occurs in the range of 25°C to 45°C. Other
studies (Finstein et al., 1980) have noted that the
temperature optimum for composting is in the range of 50°C
to 55°C. Moisture content has been found to affect the
thermal and physical properties of compost. Bonhoff, et al.
(1984) found that for separated dairy manure, the specific
heat, thermal conductivity, thermal diffusivity, and free
air space are functions of the percent moisture content.

The thermal diffusivity is also a function of temperature
and bulk density. Mears et a1. (1975) found that the
specific heat and thermal conductivity of swine manure
compost is a function of moisture content. Oxygen diffusion
rates in ground garbage compost material are linearly
related to the free inter-connected air space and inversely
affected by increases in moisture content (Shell, 1955).
Water can play a large role in heat and mass transfer within
a compost pile. Finstein et a1. (1983) estimated that when
vaporization is used as a strategy in forced aeration
systems for temperature control, almost 88 % of the heat
removal is through vaporization, while only 10 % is through
dry air convection and 2 % by conduction. Bulk density
increases and pile volume decreases with time in long—term

undisturbed swine manure compost windrows (Mears et al.,

10
1975).

Reliance on experience with other types of composting
may not lead to an accurate understanding of the problems
related to dairy manure solids composting, however.

Creation of adequate free air space (air filled porosity) is
an important concern with sewage sludge composting because
of its small particle size (Haug, 1980). Mote and Griffis
(1980) found that moisture management is of crucial
importance with composting cotton gin trash. Composts made
from different carbon sources exhibited very different
temperature profiles when carbon/nitrogen ratio and bulk
density were held constant (Mote and Griffis, 1980).

Since the few studies of composting dairy manure solids
(Carrol and Jasper, 1978; Bishop, et al., 1980; Allen et
al., 1980) do not focus on the actual composting process, it
would be desirable to observe the composting of this

substrate as part of an overall modeling effort.

CHAPTER 2

OBJECTIVES

The objectives of this research are (l) to conduct
experimental studies of a partially enclosed windrow dairy
manure solids (DMS) composting systems aimed at assessing
the importance of variables affecting composting heat and
mass transfer; (2) to assess the accuracy of the data
collection effort; and (3) to make suggestions for future

modeling of heat and mass transfer in DMS composting.

ll

CHAPTER 3

LITERATURE REVIEW

3.1: Description pf Composting

 

 

 

Composting is a member of a class of processes that
exhibit self-heating behavior (Cooney and Emerson, 1964).
There is general agreement about the sequence of events
leading to self-heating and, potentially, the spontaneous

combustion of materials.

...If the material is reasonably fresh, the initial
warming may be ascribed in part to the enzymatic
activities of the still living plant material itself.
However, it is generally conceded that the major;
buildup of heat results from the metabolism of the
mixed saprophytic microbial flora that promptly
develops. The temperature rises quickly,...often in a
matter of hours if the decomposing material is finely
divided and rich in readily available nutrient (such as
fresh grass, chopped guayule, or manure), and soon
passes the optimum for mesophilic forms. Here the
thermophiles become of critical importance; they
multiply rapidly and raise the temperature to the peak
that can be reached by microbial activity--about 70°C
or slightly higher. Subsequent heating results then
from the autocatalytic chemical processes which can
begin to operate at this temperature. The processes
cause further heating, marked chemical changes in the
heating mass, and ultimate ignition if the necessary
conditions are maintained (Cooney and Emerson, 1964).

Composting differs from this general course of events
in two manners. Since most composting processes involve
dead or waste materials, the inital temperature rise is

primarily caused by mesophilic saprophytic microorgansims

12

13

instead of the respiration of live plant cells. Secondly,
temperatures do not generally reach the autocatalytic stage.
Instead, temperatures are maintained at levels close to the
maximum temperature for thermophilic growth until either the
microbial population is so debilitated that it cannot
support the high temperatures (Finstein, 1980), moisture
content is reduced to inhibitory levels, or the substrate is
exhausted (Finstein and Morris, 1975). At this point the
temperatures begin to fall.
3.2: Analogous Systems

Insight into the underlying principles of dairy manure
solid composting can be gained from a number of sources.
Work in the late 19th and early 20th century led to a basic
understanding of the self-heating and spontaneous combustion
of a variety of agricultural products (Cohn, 1889; Miehe,
1905, 1930, Waksman et al.,l939). Some of the products
investigated include hay, manure, barley, cornmeal, cracked
corn, oats, cotton, hemp and hops (Cooney and Emerson,
1964). Self-heating caused large losses from fires and
overheating and was estimated to be over $20 million per
year in the 1920's (Browne, 1929). Losses due to
overheating and spontaneous combustion have also been noted
in the peat (Finnish Peatland Society, 1982) and forest
products (Hajny, 1966) industries.

On the positive side, microbial thermogenesis has been
used to process a number of important products including

cacao (Chatt, 1953), tobacco (Garner, 1946), horticultural

l4
container media (Hoitink and Poole, 1980) and composts for
mushroom production (Haiser and Sinden, 1953)

The application of composting to waste processing and
treatment has also been extensively studied. Substrates
examined include municipal wastes (Jeris and Reagan, 1973),
leaves (Strom et. al., 1980), sewage sludge (Haug, 1979),
septage (Lombardo, 1977), fruit and vegetable cannery wastes
(Rose et al., 1965), cotton gin trash (Griffis and Mote,
1978), grain dust (Chang et al., 1980), and horse (Pizer,
1950) swine (Martin et al., 1974), poultry litter (Bell and
P05, 1971), and dairy manure (Hummel and Willson, 1975).

3.3: Microbiology

 

 

3.3.1: General

 

The growth of a micro-organism in a favorable
environment procedes as shown in Figure 3.1. There is an
initial period of adjustment to the environment, called the
lag phase. Growth begins and accelerates into a period of

rapid and constant exponential growth, called the

 

LOGARITHM OF NUMBERS
OF BACTERIA

l

0 4 8 12 15 20 24 28 32 36 40

 

UNITS OF TIME

Figure 3.1: Hypothetical bacterial growth curve
(From Olson and Nottingham, 1980)

15
exponential or logarithmic growth phase. This growth is
eventually slowed and then stops due to the accumulation of
toxic materials or depletion of nutrients. At this point,
the stationary phase, cell division and death are in
approximate balance. The final phase occurs when cell death
becomes greater than new cell production (Olson and
Nottingham, 1980).

The temperature ranges for growth can be used to
distinguish between three major groups of micro-organisms.
These groups are shown in the Table 3.1. Within each group
there are organisms that can survive into the temperature

ranges of the adjacent groups (Olson and Nottingham, 1980).

Table 3.1: Growth temperatures for micro-organisms.

 

 

Temperature
(°C)

Group Minimum Optimum Maximum
Psycrophiles -5 to,+5 12 to 15 15 to 20
Mesophiles 5 to 15 30 to 45 35 to 47
Thermophiles 40 to 45 55 to 75 60 to 90

 

3.3.2: Types 93 Micro-organisms

 

A wide variety of micro-organisms have been identified
in composts. Bacteria (including actinomycetes) and fungi
are the predominant organisms found.
3.3.2.1: Bacteria and Actinomycetes

Bacteria have a number of morphological forms and are

typically unicellular, but multicellular associations are

16
common. These associations can be true multicellular states
or accumulations due to cell division. Actinomycetes have
an elaborate multicellular structure that arise from a
single reproductive cell. Extensive chains and branches
develop into mycelia which structurally resemble molds.
Actinomycete filaments are one tenth to one fourth the
diameter of fungi; actinomycetes have a procaryotic
structure compared with eucaryotic fungi, as well as
biochemical and sexual differences (Haug, 1980).

The most common mode of bacterial reproduction is by
binary fission, although sexual reproduction or budding can
also occur in some cases. When unfavorable environmental
conditions are encountered, some bacteria are capable of
forming dormant cells which are more resistant to harsh
environmental conditions. These forms include endospores,
cysts, and exospores. Endospores are the most stable of the
three forms (Haug, 1980).
3.3.2.2: §u_gi

Fungi are eucaryotic and heterotrOphic spore-bearing
organisms that lack chlorophyll (Atlas and Bartha, 1981).
Slime molds and true fungi make up the two broad divisions
of fungi. The true fungi can be further divided into molds
and yeasts. Both aerobic and anaerobic metabolism is
observed among yeasts, while molds are aerobic (Haug, 1980).
Fungi can tolerate low moisture conditions, have a larger
optimum pH range and often have a lower nitrogen requirement

than bacteria (Haug,l980). Bacteria and fungi rely on

 

17
similar organic substrates in composting systems and they
are often in direct competition.
3.3.3: Compost Microbial Ecology

3.3.3.1: General

 

Environmental conditions determine which organisms can
be present on a sustained basis in a given ecosystem. From
an ecological standpoint the interaction between a
biological system and the abiotic environment is best
described by Oden's Combined Law, which states

...the presence and success of an organism or a group
of organisms depends upon a complex set of conditions.

Any condition which approaches or exceeds the limits of

tolerance is said to be a limiting condition or a

limiting factor (Atlas and Bartha, 1981).

Atlas and Bartha (1981) discuss the wide variety of
environmental factors that can affect micro-organism growth:
nutrients, temperature, water activity, pH, redox potential,
pressure, and inhibitors. Tolerance ranges for any one
parameter are interactive with other parameters. Nutrients
or inhibitors must not only be present in the micro-
organism's environment, but they must also be in a form that
is available to the micro-organism. The microhabitat of a
micro-organism exerts the most direct effect on microbial
growth and survival. The ability of micro-organisms to
alter their environment can have a positive, negative or

neutral effect on what microbial populations will be

present.

18

3.3.3.2: Temperature

 

The metabolic activity of an organism is affected by
temperature. Higher, but non-lethal temperatures increase
metabolism. The change in enzyme activity caused by a 10°C
rise in temperature is called the Q10 value (Atlas and
Bartha, 1981). Experimentally determined Q10 values for
forced aeration garbage compost are shown in Table 3.2
(Moore, 1958; Wiley and Pierce, 1955).

Temperature levels affect the length of time for a new
generation of micro-organisms to be produced, or the
generation time. Figure 3.2 shows a curve for a typical
meSOphile (E. coli). Initially, as temperatures increase,
growth also increases and generation times decrease. In the
optimal growth range, generation times are short and fairly
constant. Generation times go to infinity at a temperature
only slightly greater than the optimal growth maximum (Olson
and Nottingham, 1980).

Despite the general increases in biochemical activity
with temperature, thermophiles have longer generation times
than mesophiles and have lower cell yields in proportion to
the amount of substrate utilized. This occurs because
thermophillic organisms spend a great deal of time and
energy repairing heat damage (Atlas and Bartha, 1981).

Micro-organisms make a number of adaptations to high
temperature conditions (Atlas and Bartha, 1981). These
adaptations include increases in the proportions of

saturated lipids in membranes, synthesis of heat resistant

19

Table 3.2: Q10 values for garbage composts.

 

 

 

 

 

Temperature Q10
Range
(?C) Moore (1957) Wiley and Pierce (1955)
30 to 40 2.6 1.70
40 to 50 1.80 1.65
50 to 60 1.77 1.60
_~1soo~
‘23:
E
51000-
g
g E 66/!
L”;
5 mm»
5

 

 

l - ,
10 20 30 ‘30 50
TEMPERATURE (’C)

Figure 3.2: Effect of temperature on the
generation time of a typical meSOphile
(Escherichia coli) (Adapted from Olson
and Nottingham, 1980)

20

enzymes, and the inactivation of different enzyme systems at
different temperatures. In addition, thermophilic organisms
exhibit amino-acid and growth factor requirements at high
temperatures that are not apparent at optimal temperatures.
The maximum heat tolerence of vegetative thermophiles is
correlated with the heat stability of their ribosomes.

When temperatures exceed the optimum for each organsim,
injury or death can occur. Mild time/temperature exposures
above the optimum temperature can cause stress which leads
to injury. While an injured cell may remain viable, it may
be unable to reproduce until the injury is healed (Olson and
Nottingham, 1980; Busta, 1978).

The type of organism can affect survival rates. Spores
of bacteria are resistant to high temperatures. Survival
for minutes at 120°C or for hours at 100°C is not uncommon.
Vegetative cells are killed after brief exposures to 70 to
80°C whether they are spore-formers or not (Olson and
Nottingham, 1980).

A number of possible mechanisms have been proposed to
explain the lethal influences of temperature on vegetative
bacteria, including (1) coagulation of protein; (2)
inactivation of enzymes; (3) disruption of cellular lipids;
(4) damage to the genetic apparatus; (5) breakdown of RNA
(Allwood and Russell, 1970).

Adams (1978) reviewed the heat injury of bacterial
spores. Spore injury manifestation can fall into four major

classes:

21

.....(1) requirements by survivors for non-nutrient
germination stimulants, (2) modified optimum
temperatures for the enumeration of survivors, (3)
increased sensitivity of survivors to inhibitors, and

(4) altered nutritional requirements by survivors.

Others, such as changes in the influence of pH, or Eh,

and the choice of the recovery medium, have not been

studied sufficiently to place them in specific classes

(Adams, 1978).

Spores must germinate and complete outgrowth before
vegetative cell growth can begin (Adams, 1978). During this
period, the resistance of a spore to environmental stresses
caused by heat, radiation, chemicals and extreme pH
decreases. Spore germination is a multistage process that
is mediated by a variety of agents (Adams, 1978). Injury
that is expressed during outgrowth can involve a number of
spore structures and metabolic activities.

Actinomycete temperature optima are not well defined
(Finstein and Morris, 1975). Some researchers report
temperature maxima in the range of 55°C, while others have
found actinomycetes at temperatures.up to 75°C.

Data on temperature ranges for thermophilic fungi have
been presented (Cooney and Emerson, 1964). The minimum
growth temperature is between 20 and 30°C; the maximum
temperatures are 55 to 60°C. The relationship between
temperature and growth of 6 fungi isolated from municipal
waste compost was studied by Kane and Mullins (1973).

Growth of fungal colony diameters was greatest in a range of
35 to 50°C. Between 55 and 60°C the growth dropped to zero.

A study of the temperature and pH optima for 21 species of

both thermophilic and thermotolerant fungi was conducted by

22

Rosenburg (1975). His results agreed well with those of
Kane and Mullins. No correlation was found between
temperature optimum and pH optimum among members of the
group tested.

An organism's resistance to these higher temperatures
is expressed as the thermal death time. The thermal death
time is the amount of time at a particular temperature that
is required to kill a given number of organisms (Atlas and
Bartha, 1981). By plotting the numbers of micro—organisms
that survive a given temperature after a certain time on a
semi-logaritmic plot, the decimal reduction value (Dr)' or
time to give a ten-fold reduction, can be calculated (Atlas
and Bartha, 1981). Some common decimal reduction times are
given in Table 3.3.

The decimal reduction time for a given organism will
change with temperature. By plotting the Dr values for each
temperature on semilog paper, a straight line will generally
be formed if death is first-order. The slope of this line,
2, is the temperature interval required for a log reduction
in the Dr value (Atlas and Bartha, 1981). Values of z for
selected microorganisms are given in Table 3.4. Values of z
are not constant over the entire temperature range because
the value becomes infinite at some point (Hansen and
Riemann, 1963). Differences in the 2 values between species
occur. High heat resistance of an organism and high 2
values may be correlated (Hansen and Riemann, 1963).

Despite the simplicity of logarithmic survivor curves,

Table 3.3:

23

 

 

organisms.
Organism Time Temperature
(min.) (°C)
Escherichia coli 20-30 57a
Pseudomonas aeruginosa 2 55b
Staphlococcus aureus 19 60a

 

a Source: Atlas and Bartha (1981)
b Source: Olson and Nottingham (1980)

 

 

Table 3.4: 2 values for selected micro-organisms.
Organism 2
(°C)
Yeasts 3 - 5
Most non-sporing bacteria 4 - 6

 

Source: Hansen and Riemann (1963).

Decimal reduction times for selected micro-

24
curves of other shapes are frequently observed (Figure 3.3).
Curve A is the standard logarithmic curve. Curve B, with
its lag in deaths, is frequently observed with clumped cells
(Hansen and Riemann, 1963). Curve C is of particular
concern because it seems to indicate that there is some
temperature where the lethal effect of temperature is
reduced (Haug, 1980).

Two theories have been advanced to explain the
deviations from the logarithmic survivor curves. The
”multiple target" theory was used by Moats (1971) to explain
for thermal injury and death in bacteria. This theory could
account for the lag period in curve C but not curve B.

Wei and Chang (1975) advanced a theory that random
collisions between disinfectant molecules and micro-
organisms were the cause of microbial death. The
probability for this collision was modeled with a Poisson
distribution. Clumping of varying numbers of micro-
organisms lead to a multi-Poisson distribution model. The
result was that organism clumping was the major reason for
the different survival curves. The lag period of curve C
was due to a high proportion of the population being present
in clumps. Curve B was explained by a large proportion of
non-clumping cells combined with a few large clumps.

The heat resistance of spores and vegetative bacterial
cells increases with an increase in their growth temperature
(Hansen and Riemann, 1963; Olson and Nottingham, 1980).

Slow growing cells exhibit greater heat resistance than fast

25

 

LOG POPULATION

 

 

 

TIME

Figure 3.3: Generalized types of survior curves
observed in studies of heat
inactivation of microorganisms (From
Haug, 1980)

26
growing cells (Hansen and Riemann, 1963). If large numbers
of cells are initially present, longer times will be
required to reduce the number of survivors to a given level
(Olson and Nottingham, 1980).

Heated organisms may show an increased lag period
before beginning their exponential growth phase (Lembke,
1937). This occurs even if the heat treatment did not kill
large numbers of organisms (Kaufmann et al., 1959). Injured
micro-organisms lose resistance to selective chemical agents
including salts and antibiotics. Repair of sublethal damage
to microorganims occurs more readily in simple than more
complex media (Olson and Nottingham, 1980).

Bacteria, including E. coli, exhibit the lowest heat
resistance during the exponential growth phase. The initial
lag phase and the stationary phase have higher heat
resistance (Hansen and Riemann, 1963; Olson and Nottingham,
1980).

Bacterial levels can continue to decline even after
heat treatment stops and optimal growth conditions are
restored (Jackson and Woodbine, 1962).

The heat resistance of microbial cells increases with
decreasing humidity. Dry air at 140 to 150°C has less
killing effect than wet steam at 100°C (Hansen and Riemann,
1963). Heat resistance of vegetative cells and spores can
be increased in substances with relatively reduced water
activity (Christian, 1980). This must be tested for each

microbe and substrate combination, however (Olson and

27
Nottingham, 1980).
3.3.3.3: Moisture Content and Water Activity
Water must be available for micro-organisms to grow. A
useful measurement of the availability is water activity

(a Water activity can be changed by either reducing the

w).
amount of water or by adding solutes (Christian, 1980).

Water activity of a substance is the ratio of the water
vapor pressure of the substance (p) to that of pure water
(p0) at the same temperature:

aw = p / po (3.1)

If a solution becomes more concentrated, the vapor pressure
decreases and the aw falls from the pure water value of l.
aw values can be related to the equilibrium relative
humidity (ERH) and osmotic pressure.
Ideally, water activity can be related to solute

concentration using Raoult's law in the form

aw = p / po = n2 / (n1 + n2) (3.2)

where n1 = moles of solute
n2 = moles of solvent

Non—ideal behavior can result, however, either from
reductions due to interaction between solutes or increases
because of dissociation. Experimentally determined molal
osmotic coefficients are used to compensate for this
nonideal behavior (Christian, 1980). aw values can be
calculated from

loge a = -vm ¢ / 55.51 (3.3)

V

where v = number of ions generated by each molecule
of solute (v = l for nonelectrolytes)

28

molal concentration of solute
molal osmotic coefficent

m

¢

Sources of information on the molalites for various solutes

at a range of aw values are available (Scott, 1957).

Despite the predictive value of the preceding equation
for relatively simple solutions, for complex substrates it
is more accurate to directly measure the relationship
between water activity and moisture content. A number of
factors that can influence vapor pressure and thus water
activity include adsorption of water molecules onto surfaces
and capillary forces, as well as solution concentration
effects (Christian, 1980). Hysteresis is also present
between drying and wetting moisture isotherms. At high aw,
micro-organism growth may be higher on the desorption
isotherm than on the adsorption isotherm at the same aw
(Acott and Labuza, 1975). The effect of water activity on
the growth of many species is largely independent of the
solute that controls aw (Scott, 1957).

Micro-organism growth is most rapid at a levels

w
ranging from 0.995 to 0.980 (Christian, 1980; Atlas and
Bartha, 1981). At lower aw, growth rates and the stationary
population decreases and the lag phase length increases.
With sufficently low aw, the lag phase becomes infinite
(Scott, 1957).

Values of the limiting aw for several selected
organisms are shown in Table 3.5. Gram-negative bacteria,

including Pseudomonas spp. and the Enterobacteriaceae only

grow well above 0.96 and 0.93 aw, respectively. Gram-

29

positive non-spore-forming bacteria such as Staphylococcus
aureus have a lower limit of 0.86 aw. Spore-forming
bacteria can grow at aw levels of 0.94 to 0.89, with the
Table 3.5: Approximate minimum levels of water activity

permiting growth of micro—organisms at
temperatures near optimal.

 

 

Organism aw
Molds
Aspergillus fumigatus - 0.82
Bacteria
Bacillus cereus 0.95
B. stearothermophilus 0.93
Enterobacter aerogenes 0.94
Escherichia coli 0.95
Psuedomonas fluorescens 0.97
Salmonella sp. 0.95
Staphylococcus aureus 0.86

 

Source: Troller and Christian (1978).

common limit being 0.90 to 0.91 aw. Fungi grow at much
lower aw than bacteria; fungal growth at high aw levels is
much slower than bacteria (Christian, 1980).
3.3.3.4: pH

The pH of a substance is an important factor in
determining the survival and growth of bacteria in that
substance. pH effects are difficult to separate from those
of other effects such as the concentration of undissociated
weak acids that are affected by pH (Corlett and Brown,
1980). The relative rate of proton leakage into the cell

vs. the proton-rejecting capacity of the cell determines if

30

an environment is inhibitory due to the activity of weak
acids (Freese et al., 1973).

pH growth limits can vary among micro-organisms but
most have optimum growth near pH 7 with ranges between pH 5
and 8. Minimum and maximum pH values for several selected
micro-organisms are shown in Table 3.6.

Table 3.6: The limits of pH allowing initiation of growth
by selected micro-organisms.

 

Organism Minimum Maximum
PH PH

 

Gram-negative bacteria
Escherichia coli 4 4
Klebsiella pneumoniae

(aerogenes) 4.4
Pseudomonas aeruginosa 5.6
Salmonella spp. 4.0

Gram-positive bacteria
Bacillus cereus 4.
B. stearothermophilus 5.
Staphylococcus aureus 4.
Streptococcus pyogenes 6.

 

Source: Corlett and Brown (1980).

The heat resistance of bacteria is decreased by acid or
alkaline conditions. The heat resistance generally has a
rather narrow pH range outside of which resistance falls off
quite rapidly. The optimum heat resistance typically occurs
at pH between 6 and 8 (Hansen and Riemann, 1963; Corlett and

Brown, 1980).

31

3.3.3.5: Redox Potential

 

Enzymatic reactions of micro-organisms are often
oxidation-reduction reactions. Whether an organism can
carry these redox reactions out depends on the redox state
of the environment (Atlas and Bartha, 1981). Micro-organism
such as strict aerobes or anaerobes which have only one
terminal metabolic system are limited to a relatively narrow
range of environmental redox potentials (Atlas and Bartha,
1981; Brown and Emberger, 1980). Facultive anaerobes have
alternate systems which can be switched by either the
environmental redox potential or oxygen concentration (Brown
and Emberger, 1980).

Rapid decreases in redox potential have been associated
with the early logarithmic phase of micro-organism growth
(Hewitt, 1950), spore germination, and regrowth (Douglas et
al., 1973).
3.3.3.6: Compost Microbial Ecology Studies

 

Finstein and Morris (1975) reviewed the microbiology of
composting as it applied to solid waste. The density and
succession of various micro-organisms as a function of
temperature during composting has been studied by several
authors (Waksman et al., 1939; Chang and Hudson, 1967;
Stanek, 1972).

Bacteria are present in all stages of composting
(Waksman et al., 1939; Chang and Hudson, 1967; Stanek, 1972;
Finstein and Morris, 1975). Bacterial diversity decreases

as temperature increases to inhibitory temperatures in the

32
thermophilic zone.

Actinomycete populations were slower to colonize fresh
substrates than either bacteria or fungi (Lacey, 1973; Chang
and Hudson, 1967). This was attributed to the ability of
actinomycetes to degrade more complex substrates including
cellulose, hemicellulose, chitin, and perhaps lignin (Lacey,
1973). Actinomycetes have been found to be more visible on
dry vs. wet particles (Schultze, 1962), restricted to within
6 inches of the compost surface (Erickson, 1952), and to not
grow well in poorly aerated masses of compost (Anon., 1953).
Explicit studies of Actinomycete sucession during the
temperature ascent are lacking (Finstein and Morris, 1975).

Different zones of temperature were noticed in the
piles and fungi were active in the cool, dry exterior
regions of the piles as compared with the hot interior
(Chang and Hudson, 1967). High temperatures, acidity and
anaerobic conditions were felt to limit fungal growth to the
exterior of the compost pile (Kane and Mullins, 1973). Low
pH values in stable manure composts were associated with
slow decomposition rates and anaerobic conditions (Lambert
and Davis, 1934).

Initially, large mesophilic and smaller thermophilic
fungal populations were present (Chang and Hudson, 1967).
Both populations fell, with the thermophilic fungal
population disappearing after 5 days, and the mesophilic
population disappearing after 5 to 8 days. A thermophilic

population began to appear by day 8 and recovered to levels

 

33
of 1x107 by approximately day 16. The mesophilic fungal
population reappeared 20 to 34 days into the compost period
and only recovered to levels of 1x103 to 1x105.

The ratio of thermophilic to mesophilic fungi rose
during composting, and high counts of thermophiles remained
in areas that had cooled (Chang and Hudson, 1967).

After studying the biochemical changes that occurred
during the experiment and the ability of fungi to utilize
cellulose as a carbon source, Chang (1967) concluded that
the ability to use complex carbon sources and the ability to
thrive at high temperatures are the two important
characteristics of successful colonizers of composts.

Waksman et a1. (1939) studied substrate and microbial
changes in stable manure held at four different constant
temperatures for 47 days. The micro-organisms involved in
the decomposition of the manure were found to have the
following characteristics:

At 75°C, the animal population and the fungi were
completely repressed. Actinomycetes appeared only
seldom, at the surface of the compost. Only certain
types of bacteria were active, belonging largely to the
spore-forming, hemicellulose-decomposing types...

At 65°C, the bacteria and actinomycetes were
chiefly concerned in the decomposition process. Fungi
appeared only seldom, and animal forms were absent.
The first two groups were represented by a number of
characteristic thermophilic groups. After a certain
period, the bacteria were gradually reduced and the
actinomycetes became the predominant organisms. The
thermophilic actinomycetes are limited to very few
species, but comprised several genera.

At 50°C, certain thermophilic fungi were very
active, in addition to the bacteria and actinomycetes.

This selective population, in which fungi and
actinomycetes played the predominant role, was

34

responsible for the most rapid decomposition of the
manure...The actinomycetes were similar to those
developing at 65°C.

Lower temperatures, as typified by 28°C, gave
rise to a highly heterogeneous population. Bacteria,
fungi, actinomycetes, protozoa, and nematodes were well
represented by a great variety of forms. A few days
elapsed before certain active types became established,
a fact which accounts for the delay in the rapidity of
the decomposition process at this temperature.

Stanek (1972) summarized the knowledge about microbial
succession of mushroom compost as follows:

1. Number of micro-organisms growing at 25°C decreases
during the fermentation process in compost; number of
thermophilic and thermotolerant micro-organisms
increases.

2. At first number of bacteria and then of
actinomycetes and thermotolerant fungi increases. The
dominant types of micro-organisms are:

(a) at the start of the fermentation process:
mesophilic and thermotolerant spore-yielding and
non-spore-yielding bacteria and quickl growing
fungi (Phycomycetes--genus Mucor, etc.).

(b) 'during the period of the peak—fermentation
process in the pile and during pasteurization:
thermophilic actinomycetes (Streptomyces,
Thermonospora spp.) and non-sporulating bacteria
(Pseudomonas spp.),

(c) at the end of pasteurization: thermotolerent
fungi (Humicola spp.etc.).

(3) The compost piles are colonized in various layers
(Lambert and Davis, 1934) by various micro-organisms.
This phenomenon is particularly striking before the
first turning of piles: on the surface there appear
bacteria and fungi (especially Phycomycetes) growing at
a temperature of 25°C; in the layer where intensive
aerobic processes occur grow thermophilic bacteria and
actinomycetes; the centre of the pile is mostly
colonized by anaerobic spore-forming bacteria
(Clostridium spp. etc.). After turning the pile
differences decrease by the homogenization of material.

(4) Various groups of micro-organisms take part in the
decomposition of various substances. Some thermophilic
actinomycetes and thermotolerant fungi decompose

35

cellulose, some of the thermotolerant bacteria,
actinomycetes and fungi decompose pectin etc.

(5) The occurrence of individual kinds and groups of
micro-organisms and their activity depend on the
momentary conditions (humidity, aeration, etc.) and
expecially on the quantity of accessible nutrients in
the compost (C:N ratio, etc.: After adding glycides the
number of bacteria had increased and the amount of the
ammonium had decreased; by changing sources of nitrogen
the ability of thermophilic actinomycetes and
thermotolerant fungi to decompose cellulose also
changed.

3.4: Biochemistry

 

 

3.4.1: Stoichiometry

 

Haug (1980) presented information on the general
chemical composition of a variety of organic materials.
Assuming an average compositon of sludge organics of
C10H1903N, Haug determined the stoichiometric oxygen
requirement.

C10H1903N + 12.5 02 + 10 C02 + 8 H20 + NH3
Elevated temperatures and pH > 7.0 would lead to the
volatilization of ammonia and Haug felt that nitrification
oxygen demands would not normally need to be considered.
Based on this assumption, about 2.0 g 02 would be needed per
gram of organic matter oxidized.

Finger (1975) took a slightly different approach using
the following overall equation for compost reactions.

(C6H1205)n + w 02 + d NH3 + CaHbOCNd + Y coz + 2 H20

where a = 2.82n W = 3.03 n moles OZ/mole subst.
b = 4.69n Y = 3.18 n moles COZ/mole subst.
c = 1.40n Z = 4.31 n moles HZO/mole subst.
d = 0.436n

Finger based his stoichiometric equation on the assumptions

that the substrate could be represented as a carbohydrate

36

macromolecule, that the elemental composition of the microbe
is similar to yeast, and that the conversion efficiency of
the substrate into cell material is 0.4 by weight.

The respiratory quotient (RQ) is the ratio of carbon
dioxide produced to the oxygen utilized by microbes. The
theoretical RQ for the complete oxidation of carbohydrates,
protein and fats was given as 1.0, 0.8 and 0.7, respectively
(Braithwaite, 1956). The RQ for forced aeration ground
garbage composting was determined to average about 0.9
(Moore, 1958). Braithwaite (1956) found that the RQ was 1.0
or greater for the first few days and progressively lower
for the remainder of the composting period, eventually
approaching 0.6.

3.4.2: Compost Studies

 

Changes in a number of the biochemical constituents of
compost or compost-like material have been studied by many
authors. Gregory et a1. (1963) carried out an extensive
investigation on microbial and biochemical changes in moldy
timothy and fescue grass hay. Gerrits et al. (1965) and
Muller (1965) studied changes in biochemical constituents
during the preparation of synthetic mushroom composts.
Higgins et a1. (1982) compared the organic composition of
aerObic, anaerobic, and compost-stabilized sludges.

3.4.2.1: Organic Substrates

 

Chang (1967) analyzed changes in hemicellulose,
cellulose, lignin, diastase and ethanol soluble fractions,

total N and ammonium and nitrate N in wheat straw composts.

37

The straw lost over half of its dry weight after 60 days of
composting; almost all of the loss occured in the first 34
days. The first 5 days had the greatest rate of loss with
an average loss of 2.66 % per day as opposed to an average
of only 1.3 % per day for the next 30 days. The ethanol
soluble fraction, which contained sugars, glucosides, and
essential oils showed a slight increase. Starch and
glycogen were hydrolysed by diastase. This remained
essentially constant over the composting period. Losses of
hemicellulose and cellulose acounted for most of the total
dry weight lost. Hemicellulose decreased from 35.6 to 16.9
percent. Cellulose changed from 45.3 to 13.2 percent. The
rate of hemicellulose decomposition was uniform through the
composting but cellulose decompositon rates varied. High
cellulose decomposition rates were associated with
thermophilic temperatures. Lignin did not change
appreciably from approximately 10 percent.

Waksman et al. (1939) studied the influence of
temperature upon the microbiological population and
decomposition processes in horse manure composts. Manure
was heated to four constant temperature levels for 47 days.
Samples were withdrawn on the 9th, 19th, 33rd and 47th days
for analysis of microbial populations, moisture, ash,
ammonia, hemicellulose, cellulose, lignin, water soluble
organic matter and water insoluble protein.

Total decomposition at 50°C was always greater than

that at 28°C. The 75°C samples by far had the smallest

38

total decomposition. 65°C total decomposition was the
highest at day 9 but by the 47th day it was approximately
the same as the 28°C sample. The rate of cellulose
decomposition was initially steeper than that of
hemicellulose; the decomposition rates and levels became
very similar by the end of the experiment. The effect of
temperature on changes in cellulose content paralleled the
order of the total decomposition rates. Almost no cellulose
degradation occured at 75°C, however. .Hemicellulose
degradation was highest at 50°C and lowest at 75°C.
Degradation of hemicellulose at 65°C was greater than that
at 28°C. Increases in lignin contents were in the order of
28 > 50 > 65 >> 75°C. Protein content increases were
ordered as 50 >> 65 = 28 >> 75°C. The results of Gerrits et
a1. (1965) largely confirm Waksman et a1. (1939) results.
3.4.2.2: Nitrogen Transformations

Initial accumulation of ammonia was inversely related
to the rate of decomposition in fresh horse manure (Waksman
et al., 1939). Whenever decomposition was delayed in the
initial composting period, large nitrogen losses occurred.
After 19 days the only traces of ammonia were found at 28°C
and 50°C. Ammonia was present at 75°C for the entire
experimental period, despite active volatilization due to
high temperatures and alkaline conditions. Nitrate nitrogen
began to appear in the 28°C and 50°C samples in 33 and 61
days, respectively. The 65°C samples had large accumulations

of ammonia by the end of the experiment. Over the 61 day

39

experimental period, nitrate formation was greatest at 50°C
followed by 28°C. Ammonia formation was greatest at 65 and
75°C.

Burrows (1951b) analyzed the changes in nitrogen,
phosphorus, carbon/nitrogen ratios, and pH during composting
and mushroom cropping. In a companion paper (Burrows
(1951a), he described his methods of analysis and conducted
an analysis of errors due to sampling and variance in
measurements of total ash, acid insoluble ash, moisture,
total nitrogen and calcium in the compost.

Burrows (1951b) found an increasing loss of nitrogen
with nitrogen level without an increase in organic matter
loss. Carbon/nitrogen ratios therefore narrowed during
composting. Changes in pH occurring during the same time
period indicated that ammonia volatilization was probably
the cause of the nitrogen losses.

Temperature was found to have a marked independent
effect on nitrogen losses (Burrows, 1951b). Nitrogen losses
plotted against the mean maximum temperature between turns
indicate that stack losses are quite high at 70°C but almost
negligible at 60°C.

The retention of several nitrogen sources in a compost
pile was evaluated (Burrows, 1951b). Differences between
sources was highly significant with urea showing the
greatest losses.

Burrows speculated on the results of this experiment:

...Previous work, confirmed by the present

40
investigation, shows that a higher nitrogen content
tends to increase the rate of formation of ammonia but
a higher carbon-nitrogen ratio enable the ammonia
formed to be reassimilated rapidly by organisms growing
on the abundance of carbohydrate. The temperature
effect may operate in two ways: first, by affecting
the rate of bacterial transformation of proteins into
ammonia. According to Waksman the rate appears to
increase with rising temperature until an optimum in
the region of 65°C is attained, when the rate
decreases, no doubt owing to the adverse effect of the
higher temperature upon the bacterial population.
Secondly, a rise in temperature will cause a more rapid
removal of ammonia, by physical agencies, from the
neighbourhood of the growing bacteria in the compost
heap, thus increasing nitrogen losses as the
temperature rises.

Low nitrogen composts had very high initial
immobilization of ammonia (Beckwith and Parsons, 1980). Two
thirds of the ammonia and half of the nitrate was
mineralized and immobilized, respectively, in low nitrogen
composts. All inorganic nitrogen was immobilized by day 10.
The maximum incorporation of nitrogen into the organic
nitrogen fraction occurred by the 20th day. Ammino acids
showed similar patterns to the organic nitrogen values and
were essentially constant after day 10. Amino sugars first
appeared on day 10; maximum levels occurred on day 20 and 30
for the low and high nitrogen composts.

Recovery of added fertilizer was highest in the low
nitrogen (94 %) as opposed to the high nitrogen (88 %)
compost (Beckwith and Parsons, 1980). In the initial stages
of decomposition, there was a large increase in biomass. As
activity declined, dead cells and cell contents provided a

carbon source.

Suzuki and Kumada (1977) concluded that nitrogen

41
transformation during the rotting process of rice straw
compost involved ammonification followed by the simultaneous
occurrence of nitrification and denitrification and finally
nitrification. A spatial pattern of nitrogen transformation
was also evident: the outermost layer underwent
nitrification, the interior ammonification, and the
intermediate layer had both processes occurring.

3.4.2.3: Effect pf Carbon Source

 

The availability of the carbon source to micro-
organisms in the composting process is as important as the
nitrogen source. Mote and Griffis (1980) composted three
different mixtures of materials with different carbon
sources. Bulk densities and amounts of nitrogen, water, and
carbon were held the same. Each compost exhibited different
combinations of three composting rates.

Gerrits et a1. (1965) found that the addition of
amendments to horse manure compost that had high levels of
readily available carbon caused rapid increases in compost
temperatures. Amendments with the same nitrogen levels but

less readily available carbon showed smaller increases.

 

3.4.2.4: Effect 93 Inorganic Materials

Bretzloff and Fluegel (1962) studied the changes in
several inorganic chemical constituents in a mushroom
compost pile made of manure, corn cobs, hay and other
supplements over a 30 day period. Moisture content,
conductivity, pH, total nitrogen, Kjeldahl nitrogen,

phosphorus, potassium, calcium, magnesium, sodium and

42
several ash fractions were determined from 9 locations along
a compost windrow. Temperature and oxygen contents were
measured at the same locations. The pile was turned 5 times
during the 30 day composting period.

Moisture content started at 52 % wb and increased to 72
% wb in 10 days, after which it leveled off for the
remainder of the composting period (Bretzloff and Fluegel,
1962). Total ash content showed an initial jump from 20 to
25 % after the first 8 days and a linear increase to 35 % by
the end of the 30th day. Acid-insoluble ash increased
linearly from 10 % to 18 % over the composting period. pH
showed an initial increase from 7.0 to 7.6 and an
approximately linear decrease to 7.0 by the end of the
composting period. The nitrogen, total phosphorus, calcium
and magnesium increased during composting. Potassium and
sodium contents were quite variable but showed small
increases.

Mineral amendments (bentonite, kaolinite and finely
ground calcined aluminium oxide and ferric oxide) had little
effect on the rate of decomposition in low nitrogen
synthetic composts (Beckwith and Parsons, 1980). Kaolinite
and bentonite produced a small accumulation of organic
material on high nitrogen synthetic composts.

The addition of ground gypsum improves aeration, water
movement and drainage in composting due to flocculating

action on the colloidal compost materials (Pizer, 1950).

43

3.5: Empirical Studies

 

3.5.1: Temperature

 

Bartholomew and Norman (1953) carried out a series of
experiments to study the influence of initial temperature on
the rate of heat evolution. Decomposing straw was tested
under adiabatic conditions at three separate initial
temperatures. Lower initial temperatures had prolonged
incubation periods. Maximum hourly temperature increments
occurred when the temperature had risen 4 to 8°C above the
starting temperature.

McKinley and Vestal (1984) found that the maximal
respiration for sewage sludge composting occurred in the
range of 35 C to 45°C. Other studies have noted the
temperature optimum for composting was in the range of 50 to
55°C (Finstein et al., 1980).

3.5.2: Oxygen and Carbon Dioxide

Oxygen and carbon dioxide concentrations were measured
in mushroom compost windrows (Lambert and Davis, 1934).
Haug (1980) summarized data on oxygen uptake as a function
of temperature (Figure 3.4).

3.5.3: Moisture Content

Vigorous heating of a variety of substrates up to
temperatures of 55 to 60°C when gravimetric moisture
contents are between 40 to 60‘% wb has been reported by
James et a1. (1928). Schultze (1961), Jeris and Regan
(1973b) and Snell (1957) studied the relationship between

oxygen uptake and the wet basis moisture content in

44

 
   
 
  
  
 

 

 

50
JERIS. et al
CONTINUOUS.
HIXED REFUSE.
pH - 8.0 to 8.3 SCHULTZ
1: CONTINUOUS.
3 GARBAGE AND SLUDGE
g SNELL ,- ‘
' BATCH / \
m I
> caouuo GARBAGE\/ \
e 10- /
3
I
N
0
% SD—
~ CLARK
,° BATCH. __
WILEY SYNTHETIC REFUSE
:: BATCH,
O MIXED GARBAGE AND
:- REF 5 JERIS. et a!
a U E CONTINUOUS.
2 mxao aims:
8
Z O
0 top
0 CHELL :mo son:
2 CSNTI‘IUOUS.
3 an SLuacE
>- 05—
x
O
u
0
(U
.—
( \ JERIS. er. JI \ JERIS. ct JI
0C 3ATCH. dATCH,
CCHPOSTED MIXED 2EFUSE MiHSPRINT
0.1 l 1 1 4 ' 1
10 20 30 40 50 60 70 ‘ 80

TEMPERATURE. “C

Figure 3.4: Observed oxygen consumption rates for
various composting mixtures and reactor
types as a funciton of temperature.
Each curve represents the best fit of
observed data (From Haug, 1980)

45
substrates of rewetted compost, refuse and ground garbage,
respectively. This was summarized by Haug (1980) as shown

in Figure 3.5. Below about 50 % wb moisture content

\0

negatively affected oxygen uptake to the point where at 20 o
wb almost no biological activity occurred (Glathe, 1960).
Between 50 and 70 % wb, oxygen uptake rates were at their
maximum; rates began to decrease at higher moisture
contents.

Haug (1980) indicates that low moisture contents can
lower the rates of reaction because the bacteria that cause
the composting action require an aqueous environment, and
mass transport limitations for soluble components may be
encountered under low—moisture conditions. Haug (1980) also
notes the difficulty in isolating effects of moisture alone
because of the relationships between moisture, bulk weight
and FAS.

Bartholomew and Norman (1953) studied the effect of
initial moisture content on the temperature rise of
decomposing straw under adiabatic conditions using forced
aeration. The experiments were started at 25°C and a range
of moisture contents (% db) from 75 to 275 % were examined.
In the mesophilic range, the time-temperature profiles were
similar; deviations were noticed in the thermophilic range.
Microbiological activity was highest at higher moisture
contents, as measured by heat production and decomposition.

Increases in average moisture content in a naturally

ventilated horse manure compost windrow lead to increases in

46

Aomma .momm EoLmv mmumuumnsm

mcmumOQEou moofium> mo moumu cofiuQEDmcoo

comaxo co ucmucoo muoummoe uo uommmm

Ewommd .hzmpzoo $55.02

"m.m whomfim

 

om on ok om on o4 on ON 0—
n d 1 4 1 d 4 w ‘ 1
4 owu
U 4.
C
4 o
D
D
n 4
4
4 4
4 ‘
4 o 4 .
4 4 D “323 9323 Adz... D
4
4 4 one: .235. 92 254 4
O
4 4 4 52:8 Steam £525... 0
4 :4 o

 

0—

ON

on

0?

.0m

00

ON

on

00

00..

BLVH 3>W1dfl NEUAXO I’IDI'IIXVI'I :JO 1I~133t13d

47

the carbon dioxide content at locations in the windrow
(Gerrits, 1972). Very wet composts using the same substrate
exhibited slower temperature rises and higher maximum
temperatures than drier mixtures. These trends were
attributed to delays in microbial activity and the larger
thermal mass present in the wetter compost, respectively.
Moisture contents of between 70 to 72 % wb were associated
with aerobic (5-20 % 02) and micro-aerophilic (0-5 % 02)
conditions in the largest part of the heap. Observations of
lower moisture content (68 %) composts indicated that they
were heavily aerated with cold air. Temperatures were
observed to be lower and decreased quickly after maximum
temperatures were reached. Gerrits (1972) concluded that
approximately 75 % of the heat generated through the
combustion of dry matter by micro-organisms is used for
evaporation.

Some substrates, such as cotton gin trash dry out very
quickly and need intensive water management if decomposition
is to proceed in a rapid manner (Mote and Griffis, 1980).

3.5.4: Windrow Size

 

Size reduction in a domestic refuse windrow from 10 ft
x 20 ft to 5 ft x 13 ft improved natural aeration at the
expense of increased heat loss (Horstmann and Ehgelhorn,
1969). Strom et al. (1980) recommended a leaf composting
strategy that utilized small piles in the fall to promote
aeration. Two of the small piles were combined into one

larger windrow in the winter to retain heat. Proportions of

48
the compost subject to acid, anaerobic and non-reactive
zones increased with the height and width of the windrow
(Lambert and Davis, 1934).

3.5.5: Windrow Zones

 

 

Lambert (1941) presented a detailed description of
different zones commonly encountered in mushroom compost
heaps (Figure 3.6). The patterns in physical conditions
were dependent on the size, shape and compactness of the

windrow and changed over time. Zone A is an outside layer

 

 

 

 

 

 

 

 

Figure 3.6: Section through a mushroom compost heap
illustrating the typical differences in
temperature and aeration (From Lambert,

1941).

that has temperatures varying from ambient temperatures to
43°C. This zone is well aerated. The sides can become very
dry due to excessive aeration and the top wet in wet or cool
weather. Zone B is moderately moist, well aerated and has
temperatures from 43 to 60°C. The third zone, C, extends

like a huge doughnut around the windrow 2 to 4 feet from the

49

sides and 3 feet from the top. The highest temperatures are
found in this zone: 60 to 82°C. The fourth zone (D) is
anaerobic and occupies the entire lower central part of the
windrow.

3.6: Kinetic Modeling

 

 

3.6.1: Microbial Growth

 

Whang and Meenaghan (1980) developed a kinetic model of
the composting process. They assumed that enzyme kinetic
concepts were applicable to the development of the
composting model, that an intermediate complex of micro-
organisms and substrate was formed under a quasi-equilibrium
state and that endogenous reaction is irreversible. An

equation similar to the Michaelis-Menten equation was

derived:
R = K2 (C) / (K1 + (C)) (3.4)
where R = consumption of substrate
Kl = (k_1 + k2) / kl
= su strate
XT = total micro-organism concentration
= X + CX*
X = free micro-organism
CX* = activated substrate-organism complex
kl, k-l' k2 = specific reaction rates

Values of K1 and K2 were determined from an
experimental compost reactor using humidified forced
aeration with daily turning in a batch mode.

Haug (1980) discussed the general sequence of events in
the catabolism of solid substrate by composting micro—
organisms. Some of the steps included:

1. Release of extracellular hydrolytic enzymes by the

50

cell and transport of the enzymes to the surface of the
substrate;

2. Hydrolysis of substrate molecules into lower
molecular weight, soluble fractions;

3. Diffusion transport of solubilized substrate
molecules to the cell;

4. Diffusion transport of substrate into the microbial
cell, floc or mycelia;

5. Bulk transport of oxygen (usually in air) through
voids between particles;

6. Transport of oxygen across the gas-liquid interface
and the unmixed regions which lie on either side of
such an interface;

7. Diffusion transport of oxygen through the liquid
region;

8. Diffusion transport of oxygen into the microbial
cell, floc or mycelia; and

9. Aerobic oxidation of the substrate by biochemical
reaction within the organism.

Haug (1980) presented a derivation of the substrate
consumption and microbial growth equations based on single
substrate limitations. A form of the Monod equation was
used. Microbial growth can be related to substrate use by

the equation:

dx / dt = Ym(-dS/dt) - kex (3.5)
where dX/dt = net growth rate of microbes, mass/volume-
time
Ym = growth yield coefficient, mass of
microbes/mass of substrate
ke = endogenous respiration coefficient,

l/time or mass of microbes respired/mass of
of microbes-time

X = concentration of microbes, mass/volume
maximum utilization coefficient, maximum
rate of substrate utilization at high
substrate concentration, mass
substrate/mass microbes-day

Under substrate limiting conditions the combined growth
equation can be derived:

dX/dt = Y (kmSX/(Ks + 5)) - k x (3.6)

m
01'

(dX/dt)/X

YmkmS/(KS + S) - ke (3.7)

where (dX/dt)/X = net specific growth rate

51

k maximum net specific growth rate

half velocity concentration, mass/volume

(DI/)5

concentration of rate limiting substrate,
mass/volume

The four kinetic coefficients, Ym, km, Ks and ke need
to be known for a specific substrate and microbe
combination. Haug (1980) presents data on a range of
biological processes.

Sinclair and Ryder (1975) compared the ability of two
different growth models to explain continuous culture data
under both carbon and oxygen limiting conditions. The
interacting substrate model postulates an interaction
between two substrates and is based on double enzyme
kinetics. Only one of two substrates would enter into the
model at any one time in the alternate substrate model.
Both Monod/Monod and Monod/Contois expressions worked well
in both models. The models were able to satisfactorily
explain experimental results.

3.6.2: Microbial Death

 

Haug (1980) developed a first-order decay model for
inactivation kinetics which is often refered to as "Chick’s
Law." The most common form is given as

dn / dt = -kdn (3.8)

viable cell population
thermal inactivation coefficient

where n
kc1

If kd is a constant, integration of the above equation

yields
nt = noexp(-kdt) (3.9)

where nO = initial cell population

52
nt = later population at time t
The temperature effect on kd is typically modeled by

the Arrhenius form

kd = C exp(-Ed/RTk) (3.10)
where C = constant
Ed = inactivation energy, kcal/mole
Tk = temperature, K

Inactivation energies for many spores and vegetative cells
range between 50 and 100 kcal/mol (Bailey and Ollis, 1977).
By taking the logarithim of the above equation, a plot of
the log of kd vs. l/Tk from survivor plot data can be used
to find C and Ed (Haug, 1980).

Since the temperature of the compost changes with time,
kd will not be constant. Combining Equations (3.8) and
(3.10) gives

dn / dt = -c n expi-Ed/RTk(t)]n (3.11)
where Tk is a function of time, Tk(t). This expression can
be used to evaluate the kill resulting from various time-
temperature profiles by separating the variables and
integrating from the intial to the final conditions:

1'.
1n(no/nf) = f c exp [-Ed/RTk(t)] dt (3.12)
t

o
where the subscript 0 indicates initial and f final
conditions. Haug (1980) noted that integrations for
hyperbolic, exponential and linear time-temperature
functions have been performed, in particular by Bailey and
Ollis (1977). In composting situations, the time-

temperature curves may not follow profiles for which

53

analytical solutions are available. Graphical (Haug, 1980)
or numerical procedures could be used to calculate microbial
death due to temperatures in those cases.

Large clumps of compost particles have been observed to
form during composting. Haug (1980) examined the effect of
this solids clumping on microbial destruction. He assumed a
spherical homogeneous ball of compost material heated from
the outside and thermal and physical properties based on
those for water and compost. Using these assumptions he
calculated the length of time that it would take for the
interior of clumps of various sizes to heat to 0.90 times
the ambient temperature. Heating times for particles of
between 1 and 10 cm radius were found to be negligible;
only when clumps were greater than 20 cm radius did heating
times become significant.

Nonuniform temperature distributions can affect the
overall thermal deactivation of pathogens. The effects of
these nonuniformities have been examined for windrow (Haug,
1980) and static pile systems (Haug, 1980; Burge et al.,
1978). For windrow systems, thermal inactivation can be
described as:
and nt = noifl + fhexp(-det)]N (3.13)

f1 + fh = 1 (3.14)

number of organisms surviving

where nt
number of organisms initially present

0 . . . .

fl fraction of composting material in the low-
temperature, sublethal zone

fh = fraction of composting material in the high-

temperature zone
At = time interval between pile turnings

54

thermal death coefficient

Rd
N

number of pile turnings

Two different fl/fh values and several det values were
evaluated. The fraction of surviving organisms was much
greater than that Calculated by examining a specific time-
temperature curve. Haug (1980) notes that this indicates
that the exposure of all portions of a windrow to lethal
temperatures is as important as achieving time-temperature
curves at any one location in order to increase the average
‘thermal inactivation.

Furthermore, the above analysis is based on the
assumption that all particles are randomly mixed during the
turning process. As Haug notes, this may be true if
commercial turning devices are used. If a front-end loader
is used, however, complete mixing will probably not occur.
The loader can be used, however, to place portions of the
windrow that need additional heat treatment in zones that
should receive higher temperatures. If only a small portion
of material on the outside of the pile comprising 1 % of the
total compost mass escapes mixing, the average thermal
inactivation would be limited to no more than 3 Dr (Haug,
1980).

Burge et al. (1978) measured the temperature
distributions in 15 static piles of sewage sludge and wood
chips using forced aeration and used this data to predict
the confidence levels for achieving a particular time-

temperature relationship. The number of days that different

55

Table 3.7: Confidence levels that all material will obtain
a temperature equal to or greater than a
particular temperature for a desired number of

 

 

days.
Confidence Levels (%)
Temperature 95 99 99.9
(°C) (Days) (Days) (Days)
>50 13.8 13.3 12.6
>55 10.6 10.1 9.4
>60 7.3 6.8 6.3
>65 4.3 3.9 3.4
>70 1.2 1.0 0.8

 

temperatures could be reached are shown in Table 3.7.

Burge et a1. (1978) also examined the mean and standard
deviation of temperature in the toe area of the 15 test
piles. Examining the data, Haug (1980) notes that

Significant reduction of bacteriophage f2 would be
predicted not only on the basis of the mean temperature
but also the temperature obtained by subtracting the
standard deviation from the mean.

The confidence levels shown in Table 3.7 apply only to the
15 test piles as they will vary with the substrate material,
bulking agent and manner of operation (Haug, 1980).

3.6.3: Regrowth

Coliform and fecal streptococcus bacteria have been
observed to regrow in sterile liquid sludge heated to 35°C
(Brandon, 1878). The most dramatic regrowth of fecal
streptococcus bacteria was observed in sterilized material
that was then recontaminated (Ward and Brandon, 1977).

Occaisional regrowth of coliform organisms was observed

during windrow composting of sewage sludge, particularly in

56

wet winter (California) conditions (Selna and Smith, 1976).

3.6.4: Combined Equations

 

Haug (1980) proposed a net rate coefficient equation
based on empirical expressions from experimental data to

describe the overall kinetics of sewage sludge composting:

kd = (Fl)(F2)(F02)kdm (3.15)
where Fl = moisture content correction
F2 = free air space correction
F02 = oxygen content correction
kdm = maximum rate coefficient determined by

substrate and temperature conditions

Haug assumed that the oxidation of biodegradable
volatile solids (BVS) is first-order with respect to BVS
quantity:

d(BVS) / dt = -kd(BVS) (3.16)

The distinction between BVS and nonbiodegradable
volatile solids is made to distiguish between materials that
are readily degradable and those that take longer to degrade
(Haug, 1980).

Haug assumed that the rate constant kd was a function
of temperature only. Based on studies of garbage composting
by Schultze (1962) that determined oxygen consumption, Haug
(1980) developed the following empirical equation:

. kd = 0.00632 (1.066)’1"’20 (3.17)
Where T is in degrees Celsius and kd (g BVS oxidized/g TVS-
day).

Heat inactivation due to increased temperatures was

accounted for in Haug's models using an expression developed

57
by Andrews and Kubhu (1973) to describe similar effects

during aerobic digestion of liquid wastes:

_ (T-TRl) - (T-TRZ)
kd - de1 [C1 C2 ] (3.18)
where kd 1 = rate constant at temperature TRl' day“l
Cl,82 = temperature coefficients

TRl'TRZ reference temperatures, C

Equations 3.17 and 3.18 were combined and reasonable
values were assumed to arrive at the effect of temperature
on kd (Haug,l980):

kdm = 0.0126 [1.066(T‘20) - 1.21‘T‘50)] (3.19)

The effect of moisture content on oxygen consumption
rates and rate of BVS oxidation was estimated with data from
Schultze (1962), Jeris and Regan (1973b) and Snell (1957).
An "S" shaped curve was fit to the data by two equations for
different moisture content ranges.

The effect of low free air space (FAS) on oxygen
transport was modeled by Haug (1980) using data from Jeris
and Regan (1973b) for refuse compost. Thirty percent was
the optimum FAS in terms of oxygen uptake rates. Two
separate equations covering the range of FAS values were fit
to the data. FAS effects were considered to be important
only for those composting systems that did not use bulking
agents. As long as some FAS was present, Haug (1980) felt
that it would have a negligible effect on reaction kinetics.
Knowledge of the bulk weight of the compost material was
necessary to calculate the FAS.

Oxygen content effects on compost reaction kinetics

were felt to be very complex by Haug. The effect of free

58
air space oxygen concentration was assumed to follow a
Monod-type expression:

F02 = VOLP02 / (VOLP02 + 1.0) (3.20)
where VOLP02 is the volume percent oxygen in the FAS. The
effect of oxygen concentrations above 5 % were minimized by
the assumption of a half-velocity constant of 1.0.

3.7: Physical Properties

 

 

3.7.1: Particle Size

 

 

Chang and Rible (1975) analyzed fresh, deposited and
composted livestock wastes to determine their particle size
distribution and to characterize each size fraction in terms
of its value as fertilzer, feed suplement and fuel.
Composted dairy wastes showed a higher concentration of
large fibrous materials than fresh waste. Moisture contents
of larger size fractions remained the same while smaller
sized particles showed large decreases. Crude fiber and
protein decreased by nearly 50 % in most size fractions.

Fat content decreased in all particle sizes. Ash contents
increased by 75 to 100 percent.

Mears et al. (1975) analyzed the particle size of a
number of different compost mixtures. They had problems
with measuring the particle size of material with greater
than 45 % wb. The method of drying affected the particle
size determinations for this material. Particle size
measurement of material that was initially drier than 45 % wb
was not affected by drying method. Particle size followed a

log normal distribution. Geometric mean diameter was

59
determined for each windrow on a weekly basis and decreased
over time. The rate of decrease varied between the
substrate mixtures and the reported work did not consist of
sufficent experimental windrows to draw conclusions about
any one substrate.

3.7.2: Particle Density

 

Chen (1982) reported mean values of beef cattle manure
dried solids density of 1524 kg/m3. The average mean
particle density of dairy manure solids obtained from three
separate farms was 1551 kg/m3 (Bohnhoff and Converse,
1986a). Regardless of the source, composted material had
mean particle densities that were significantly greater than
the mean particle density of freshly separated solids.

Bohnhoff et a1. (1984) and Bohnhoff and Converse
(1986a) present two equations for determining particle
density. The earlier equation is a totally empirical linear
regression on volatile solids. The later equation is
derived from the definition of mean dry particle density
with terms for the densities of volatile and fixed solids
obtained by nonlinear least squares regression.

3.7.3: Bulk Density

 

Mixtures of swine waste and refuse were initially
between 275 and 500 kg/m3 and increased to between 650 and
850 kg/m3 after 27 days of composting (Mears et al., 1975).

3.7.4: Derived quantities

 

 

When bulk density, particle density and gravimetric

moisture content are known a number of other quantities can

 

60
be calculated (Hillel, 1982). These include the porosity
( f ), void ratio ( e ), volumetric moisture content ( 8 ),
degree of saturation ( s ) and air filled porosity ( fa )
or free air space (FAS).
3.7.5: Compressibility

Based on a method developed for evaluating silage
materials, Mears et a1. (1975) determined the
compressibility of swine waste based compost materials.

They found that a relationship between bulk density, Oh, and
applied axial stress, a, adequately described the
compressive properties of the composted materials:

1n 0 = 1n 00 + C 1n 0 (3.21)
where 1n 00 and C are regression coefficients formed by the
statistical analysis of the data.
3.7.6: Settlement Behavior

Stentiford et a1. (1984) reported changes in 1.7 m tall
triangular windrow cross-sections for mixed refuse
composting by the static pile method. The maximum change in
the pile height came in the first 5 days with 60 percent of
the 36 cm total decrease.

Mears et a1. (1975) measured windrow volume reductions
in swine waste based windrow composts. Reduction in
windrows composed entirely of swine waste ranged from 8 to
16 % after 10 days and 23 to 45 % in 40 days. A mixture of
swine waste and straw had a reduction of 28 % after 50 days.

The time-settlement behavior of milled urban refuse

under saturated conditions has been studied by Chen et al.

61
(1977). A mathematical model for solid waste settlement was
developed by Zimmerman et al. (1977). It consists of two
simultaneous equations, one of which is non-linear. The
effects of finite strain, biological and chemical activity,
and the time variation of saturation have been included.

3.8: Heat Transfer

 

3.8.1: Diffusion

 

 

3.8.1.1: Thermal Conductivity

Thermal conductivity is a basic heat transfer property
of a material. Materials can be considered conductors (high
conductivity, high heat transfer) or insulators (low
conductivity, low heat transfer). The material in question
can consist of a single phase (ie. gas, liquid or solid) or
can be a combination of phases. Each phase can consist of a
mixture of materials, as well.

Heat conduction in gases, vapors and liquids depends
largely on the molecular transfer of the kinetic energy of
molecular movement (Karak and Yener, 1979). Temperature and
pressure have an important influence on the thermal
conductivity of these substances. Liquids with their closer
molecular spacing have much higher thermal conductivities
than gases.

Thermal conductivties of mixed state materials are
affected by the properties of each state and the way in
which they are combined. Based on the structure of the
solid state, materials can be classified as fibrous

(textiles, fiberglass wool, straw), granular (powders, coal,

62
grain) or cellular (cork, foam insulating boards). Factors
affecting thermal conductivity of such material include the
(l) conductivity of the component material, (2) pressure of
interstitial fluid, (3) temperature, (4) bulk density, (5)
particle size, (6) particle size distribution, (7) void
space and porosity, and (8) moisture content (Chun-Yung
Chen, 1969).

Simple models of thermal conductivity relate the
porosity and conductivity of the two phases as separate
resistances arranged in series and parallel (Pratt, 1969).
Other theoretical models are based on the porous material
having a continuous solid or continuous air phase (Hillel,
1981).

Pratt (1969) presented experimental data showing the
variation in effective thermal conductivity due to the
presence of different gases in a fibrous insulation. Helium
containing insulation had five time the thermal conductivity
of air while the thermal conductivity of insulation filled
with C02 was only 72 % of air. The thermal conductivity of
the solids in the matrix has a direct effect on the
effective thermal conductivity.

Temperature changes effect the apparent conductivity
through two mechanisms: the thermal conductivities of the
components and the contribution of radiation (Chen-Yung
Chen, 1969). Thermal conductivities of gases increase with
temperature. Some liquids, such as water, show the same

trend.
In addition to the purely conductive effects, actual

63

heat transfer in heterogenous materials can be by convection
or radiation as well as by conduction. Depending on the
method or scale of measurement, these transfer modes are
sometimes included in an apparent or effective thermal
conductivity. At typical composting temperatures, the
radiation effects are minute.

Increases of fibrous material bulk density increase
thermal conductivity. Pratt (1969) observed that many
researchers have shown that the thermal conductivity in low
bulk density fabrics with porosities of about 90 percent are
almost independent of the component fiber. Allcut (1951)
performed a series of measurements of apparent thermal
conductivity for various material and densities. A
characteristic hook shape is shown in all thermal
conductivity curves (Figure 3.7). The cause for this shape

is heat transmission by convection. Increasing densities

 

Jud,

:JLJ I" -‘

 

Thermal (cnduclmly
I04 I .‘J (m.| act)"

Hm 3....21

U'U LIIIH‘ f"

 

 

 

O 5’.) ICU E.) Cox) Li. J
C8PSIT‘,‘. h ;""3

Figure 3.7: "Apparent" thermal conductivity of
various materials at different
densities. A, kapock; B, baggasse;
C, cork board; D, slag wool;

E, mineral wool (From Pratt, 1969).

64

reduce the convection heat transfer contribution to the
apparent thermal conductivity by dividing the air layers
into smaller layers, thus increasing the medium's tortuosity
and increasing it's resistance to fluid flow.

After a certain density is reached, however, the
apparent thermal conductivity begins to increase. The exact
density varies between materials and is dependent on, among
other things, the ratio KS/Kg of the solid to the gas
thermal conductivity. Deissler and Boegli (1958) found that
the effective conductivity of a void was strongly influenced
by material arrangement for high KS/Kg. Chun-Yung Chen
(1969) visualized this as being due to the contact point
between the solid particles. Gas is a relative insulator in
high KS/Kg systems, so most heat flow takes place near the
contact points. For values of Ks/Kg of approximately 1000,
nearly all the heat transfer takes place near the point of
contact (Deissler and Boegli, 1958). Changes in bulk
density increase the number of contact points, leading to
greater contributions of conductive heat transfer to the
overall apparent thermal conductivity.

A bulk material contains two types of gas spaces:
intra- and inter-particle. If the gas has a lower thermal
conductivity than that of the solid, intra-partical gas
lowers the particle effective conductivity. The cells can
be large or small, closed or open. Closed cell material
voids are essentially air and vapor tight, while open cell

voids are interconnecting and permit free movement of air

65

and vapor through the material (Pratt, 1969).

The effect of inter-particle gas space on the effective
thermal conductivity depends on the characteristics of the
system. At very low bulk densities, with large porosities
and pore sizes, convective heat transfer is large and
important. As overall porosity and pore size decrease,
convective heat transfer decreases. This leads to a
decrease in the effective thermal conductivity. It is
important to note that decreases in pores sized can also
lower thermal conductivity by increasing the tortuosity and
resistance to convective heat transfer (Verschoor and
Greeber, 1952).

The influence of latent heat transfer by water vapor in
air-filled pores can be significant (Hillel, 1981; Pratt,
1969). Latent heat transfer effects can be taken into
account by the additon of an apparent conductivity due to
the evaporation, transport, and condensation of water vapor
to the thermal conductiviy of air (Hillel, 1981). This
value is strongly temperature dependent and rises rapidly
with increasing temperature (Figure 3.8). At about 55 to
60°C it is approximately the same value as the thermal
conductivity of liquid water (Pratt, 1969).

Mears et a1. (1975) calculated thermal conductivity for
composts consisting primarily of swine wastes from
previously determined values of thermal diffusivity, bulk

density and specific heat:

66

 

I I l I I
7 4‘“
$6-
7
E
g
- «- 2g._e..flk
"g 4‘. ‘\-A°+er p'va or A.
2‘
IE
I}
3
‘8
8 2'
2
:- ‘0

 

 

 

 

. 1 1 I L I
IO 20 30 4O 5O 60 TO 80
Ternpevotuve.°C

0

Figure 3.8: Effective conductivity of porous
material due to latent heat transfered

by movement of water vapor (From Pratt
1969).

K = 0.1163 a p C (3.22)

P
Statistical analysis of the data indicated that thermal
conductivity varied linearly with moisture content. Thermal
conductivity at any given moisture content increased
significantly as the compost matured.

Houkom et a1. (1974) and Chen (1983) measured the
thermal conductivity of beef cattle manure. Chen developed
three separate regression equations for different manures
that related thermal conductivity to bulk density, total
solids and porosity.

Bohnhoff and Converse (1986a) measured the thermal
conductivity of dairy manure solids at 2 different moisture
contents, three temperatures and three bulk densities. A

four parameter equation relating thermal conductivity to

temperature and volumetric moisture content was selected.

67

3.8.1.2: Specific Heat

 

Specific heat has been found to vary linearly with the
wet basis moisture content for swine waste (Mears et a1,
1975) beef cattle manure (Houkom et al., 1972; Chen ,1982)
and separated dairy manure solids (Bohnhoff and Converse,
1986a).

3.8.1.3: Thermal Diffusivity

 

The thermal diffusivity of a swine waste compost was
determined by Mears et a1. (1975) using the transient
method. Thermal diffusivity of dairy manure solids was
calculated by Bohnhoff and Converse (1986a).

3.8.2: Natural Convection

 

Principles of convection through porous media are
discussed by Bejan (1984). Haug (1980) developed a simple
natural convection model to assess the ability of this
process to supply oxygen at stoichiometric rates.

3221 Mass Transfer

The spatial scale in a composting system is associated
with different types of mass transfer processes. Gas and
vapor transfer in windrow voids could occur by
diffusion or natural convection. Liquid transfer in void
spaces could occur due to saturated or unsaturated flow.
Gas transfer from the free air space across the liquid film

to the substrate-microbe complex is by diffusion.

68

3.9.1: Windrow Void Space Transfer
3.9.1.1: Diffusion

Experimentally determined values of diffusion
coefficients in gases at one atmosphere are shown in Table
3.8. The binary diffusion of gases can be predicted for
given temperatures and pressures using the Chapman-Enskog
kinetic theory or empirical correlations (Cussler, 1984).
The Chapman-Enskog theory assumes nonpolar gases and this
excludes water and ammonia (Cussler, 1984). Diffusion of
gases vary with the 1.5 to 1.8 power of temperature

(Cussler, 1984).

Table 3.8: Experimental values of diffusion coefficients in
gases at one atmosphere.

 

 

Diffusion
Gas Pair Temperature Coeficient
(°K) (cm2 sec")
Air--C02 276.2 0.142
Air--02 273.0 0.1775
Air--H20 289.1 0.282
298.2 0.260
312.6 0.277
333.2 0.3050

 

Source: Cussler (1984).

Shell (1955) studied the diffusion rate of oxygen
through a ground garbage compost. The relative diffusion
rates were shown to be linearly related to the free air
space of the material. Increases in moisture content

decreased the relative oxygen diffusion rate. Increases in

69
bulk density decrease the relative diffusion rate.

Hillel (1981) discussed a number of experimental
efforts to relate the diffusion of gases in soils to the
diffusion rate in air and derived a general equation for
transient diffusion in the soil.

Equimolar counterdiffusion occurs when the same number
of moles of two gases in separate reservoirs diffuse towards
one another (Geankopolis, 1983). Finger (1975) assumed that
the equimolar counterdiffusion of oxygen and carbon dioxide
accounted for mass transfer in his distributed heat and mass
transfer model of a compost pile.

Diffusion in many systems involves more than one
component. In composting systems, observations of steaming
indicate that water should be considered in addition to
ongen and carbon dioxide. Multicomponent diffusion is
estimated by converting the problem to a binary problem,
solving it and then converting back to a multicomponent
solution (Cussler, 1984).

Diffusion of gases in porous solids and capillaries is
discussed by Geankopolis (1983) under Fickian, Knudsen and
transition regimes. Isothermal diffusivity of water vapor
with respect to soil volumetric moisture content was derived
by Jackson (1964).

3.9.1.2: Evaporation

 

 

The drying rate of biological products with initial
moisture contents above 70 to 75 % wb can be constant if

external drying parameters such as air velocity, air

7O
temperature and air humidity are constant (Brooker et al.,
1981). Constant rate drying will be observed under constant
external conditions when the internal resistance to moisture
transport is much less than the external resistance to water
vapor removal from the surface of the product (Brooker et
al., 1981).

Brooker et a1. (1981) presents an expression for the
constant rate moisture loss for biological products. The
surface area and either the heat or mass transfer
coefficents must be known for the constant drying rate to be
calculated.

Bohnhoff and Converse (1986b) developed desorption
isotherms for water desorption equilibria at five
temperature levels. Three isotherm models were fit to the
data; the relationship between moisture content, relative
humidity and temperature was best described by the four
parameter Chen-Clayton equation.

Equations for estimating the isosteric and integral
heats of desorption over a temperature range of 0 to 70°C
were also developed (Bohnhoff and Converse, 1986b).
3.9.1.3: Liquids

Hillel (1981) discusses the concept of hydraulic
diffusivity to describe the convective transport of water in
soil pores. Hydraulic diffusivity is the ratio of hydraulic
conductivity to the specific water capacity. All three
terms can be written as functions of the volumetric moisture

content. An equation for the simultaneous transfer of both

71
liquid and vapor can be derived (Hillel, 1981).

3.9.2: Transport in Liquid Films

 

 

Haug (1980) has presented a conceptual illustration of
mass transport and reaction in compost particle water films.
This is shown in Figure 3.9. Consumption of oxygen by
microbes causes a concentration gradient and oxygen diffuses
from the free air space into the substrate-water-microbe
matrix. Aerobic microbial metabolism consUmes the substrate
and oxygen. Microbial mass is synthesized and carbon
dioxide, heat, water, and ammonia are produced. The
metabolic end products are at elevated concentrations in the
liquid phase and will diffuse toward the airspace.

Diffusion coefficients in a gas are about 105 times
greater than in a liquid. Due to higher liquid vs. gas
concentrations, however, the flux in a gas is not that much
greater, being only about 100 times faster. (Geankopolis,
1983). The slowness of diffusion in liquids often limits
the overall rate of processes that occur in liquids
(Cussler, 1984). Some experimentally determined diffusion
coefficients are presented in Table 3.9.

Cussler (1984) presents a number of methods for
calculating diffusion in liquids and compares them for
diffusion of oxygen in water at 25°C.

Diffusivities in liquids are often dependent on the
concentration of the diffusing components. Oxygen diffusion
in bacterial slime layers can be significantly lower than in

water, sometimes as low as 0.04 x 105 (Bailey and Ollis,

72

H- H20 In Iced

Suburuo

 

  

Ralouohon

 

Endoconouu
Baton-lion

On flow In In.
an lane.

02

 

NM‘HCOJ

NO;

0.. PthI <-— ——- Subllnu. WIIor, Microbe Mull:

Figure 3.9: Conceptual illustration of mass
transport of major components during
composting (From Haug, 1980).

73

Table 3.9: Diffusion coefficients at infinite dilution in
water at 25°C.

 

 

Solute D (-10’5 cmZ/sec)
Carbon dioxide 1.92

Oxygen 2.10

Ammonia 1.64

 

Source: Cussler, 1984.

1977).

Haug (1980) used a simplified model of gas transfer in
a saturated matrix of solid substrate and microbes to
estimate the effect of particle size on oxygen flux and the
time required to satisfy the Stoichiometric oxygen
requirement. He concluded that diffusion can match the
oxygen consumption rate if the particle size is sufficiently
small. Particles thicker than 1.0 cm would have large
diffusional resistances that could dominate the process
kinetics. Particles of about 0.10 cm would be small enough
for diffusion to meet demand. Oxygen diffusion would no
longer exert control over the overall composting rate if
particles are less than 0.05 cm in diameter.

3.10: Complete Heat and Mass Transfer Models

 

The method of volume averaging was used to derive the
governing equations for heat and mass transport in a rigid
medium by Ryan et al. (1981). Excellent agreement was found
between theory and experimental work using a spatially

perodic model of a porous medium for conductive and

74

diffusive transport.

Latif and Lissik (1986) developed a non-distributed
respiration model for heat and gases released during grain
storage. The rate equation was a function of initial
temperature and moisture content, and grain damage.

Chau et a1. (1984) presented a numerical model for heat
and mass transfer in spherical products. The effects of
respiration, transpiration, conduction, convection and
evaporative cooling were included.

Haug (1980) developed comprehensive simulations of
continuous-feed completely mixed and batch models of the
composting process. Finger (1975) developed a distributed
heat and mass transfer model for windrow composting that was
based on thermal heat conduction and the diffusion of oxygen

into the compost pile.

CHAPTER 4

EXPERIMENTAL METHODS

4.1: Data Requirements

 

 

A windrow system without enhanced natural convection or
forced aeration was selected for study. This system was
selected because it closely resembled current dairy manure
solids (DMS) handling practices: simply piling and turning
the separated DMS before using it as a bedding material.

Based on a review of the literature, the data which
must be used to validate a time-dependent spatially-
distributed finite-element model capable of predicting the
time/temperature histories in a compost windrow include the
following: temperature, gas concentrations (02 &/or C02),
moisture content, volatile solids, bulk density, porosity,
air-filled porosity, and windrow size change. The last four
items were measured in core samples; the first two were
measured using a probe, the design of which is described
below. Table 4.1 gives a summary of the sampling methods

and schedule for each variable.

 

4.2: General Experimental Design and Methods
The project was located at the dairy facility of the
Kellogg Biological Station of Michigan State University.

The facility, constructed in 1985, is a research and

75

Table 4.1:

76

Sampling methods and schedule.

 

Variable

Method

Sampling Date
1 2 3 4 5 6 7

 

Temperature:
Windrow:

Slab:

Air:

Rel. Humidity:

Gas
Concentration:

Core Samples:

Windrow size &
Shape:

* Campbell Scientific
CR21X datalogger and AM32
multiplexer; 9 locations:
Windrow 1-3 reps. averaged,
Windrow 2-3 reps. recorded;
1 minute sampling interval,
averaged each 15 minutes.

* Same instrumentation;

1 location, 3 reps.
averaged; same sampling
interval.

* Same instrumentation;

1 location; same sampling
interval.

* Same instrumentation;
1 location; same sampling
interval.

* Probe sampling, evacuated
cylinder storage, analysis
for OZ and C02 with gas
Chromatograph; 10 locations
3 replications.

* Rotary corer; moisture
content, volatile solids,
coliform number, bulk
density, porosity, free air
space; 9 locations.

* Windrow height, 30 to 36

locations; windrow height at

probes, 12 locations; see
text for method details.

XXXXXXXXXXXXX

XXXXXXXXXXXXX

XXXXXXXXXXXXX

XXXXXXXXXXXXX

X X X X

X X X X

 

77
demonstration dairy farm that includes a 192 cow free stall
dairy and milking parlor and a variety of support buildings.
Four times each day, manure and spent bedding were flushed
from free-stall alleys into a settling tank. The tank
contents were agitated daily and the slurry was pumped over
an inclined separator. The manure solids that came from the
separator were then conveyed to a series of 4.5 m x 4.5 m
bins inside a naturally ventilated building. The separated
dairy manure solids (DMS) were allowed to accumulate for a
week. Table 4.2 summarizes the physical characteristics of

the dairy.

Table 4.2: Dairy characteristics.

 

Number of cows: 192 Type: Holstein-Friesian
Housed in free stalls Four stall alleys Four feed alleys

Waste system: Flushed 4 times per day with tip and gated
tanks
5 % slope
Well water and recycled lagoon water
Stationary screen separator
Conveyer
Dual lagoon
Bedding: 1/3 composted DMS--l/2 skid
steer bucket per week per stall
2/3 chopped straw

 

When an experimental windrow was constructed, a batch
of accumulated DMS was removed by a "skid steer" bucket
loader, placed in a silage mixer, agitated, weighed and
dumped on the ground. Samples for analysis of moisture

content, volatile solids and coliform count were taken from

78

the mixed windrow. The experimental windrow was constructed
across a bin using a skid steer to place the DMS. Great
care was taken to assure uniform placement. The skid steer
operator was directed where to put each bucket load. Each
bucket was placed so that the material fell from a uniform
height of 46 cm (18 inches) above the windrow surface. The
skid steer bucket was not used to shape the windrow: all
final shaping was done manually with a shovel. All windrows
had a trapezoidal prismatic shape.

The windrow was taken apart weekly, mixed in a silage
mixer, weighed and formed into a new windrow using the
methods described above. Exact uniformity of windrow size
and shape was not possible between turnings because of
losses due to composting and the difficulty of windrow
construction. Each batch of solids was treated for 3 weeks.
Two separate batches were monitored: one during the summer
and the other during the fall of 1986.

4.3: Probe Design and Placement

Methods of measuring temperature and gas concentration
with probes in natural convection/diffusion based composting
systems must meet several criteria. Accurate measurement
must be possible. Probe placement and operation should
disturb the windrow as little as possible. The probe should
be sufficently rugged to withstand rough handling as well as
the corrosive compost environment.

Accurate measurement of temperature levels and gas

concentration are covered more extensively in the

79

appropriate sections. In general, temperature and gas
samples must accurately reflect the actual temperature and
gas concentration at the sampling location. The thermal
conductivity of the probe, for instance, must be small
enough that heat from neighboring high temperature zones is
not conducted along the probe to lower temperature
locations, leading to an inaccurate, high temperature
measurement. Similarly, the gas sampling process should not
draw gas from areas other than the intended sample area.

The method of probe placement can affect physical
properties such as the bulk density and porosity in the area
around the probe. This can in turn affect the heat and mass
transfer which is being studied. If the probe is forced
into the windrow, the surrounding DMS are compacted. If a
hole for the probe is drilled into the windrow, the less
dense material may settle away from the probe, exposing the
probe to ambient conditions; this can be a serious problem
with long term windrows (Mote, 1986). Placement of probes
while the windrow is being made could lead to areas of the
windrow "downstream" from the probe having different bulk
densities than those on the "upstream" side.

The timing of probe placement must also be considered.
Probes can either be inserted into the windrow whenever a
measurement is made or they can be left in place for the
duration of the study. Measurement on demand reduces both
the number of probes that are needed, as well as their

complexity. Large homogeneous windrows are required,

80

however, to permit probe insertion at a new location at each
time without affecting the process. In-place probes allow
many measurements to be made over time in the same location.

Operation of the measurement system should not affect
the heat and mass transfer process in the windrow. This is
particularly true for gas sampling. Suction of large
volumes of gas draws outside air into the windrow. In
forced aeration or suction composting systems this is not
important. In natural convection or diffusion systems,
however, this could greatly change the heat and mass
transfer in the windrow, if only for a while.

There have been many attempts to develop probes for
sampling temperatures and gas concentrations. Finstein
(1980) placed a thermocouple in a slotted wooden dowel to
reduce resistance to the probe placement. While studying
forced aeration composting systems, Singley et a1. (1982)
used a probe that consisted of a perforated cylindrical
diffusion chamber on the end of a pipe to measure both gas
concentrations and temperatures at the same point in a
windrow. A thermocouple was placed in the diffusion
chamber. The probe was placed while the windrow was being
built; air samples were removed as needed for analysis.
Finger (1975) used a dissolved oxygen electrode mounted in a
diffusion chamber at the end of a metal pipe. The probe was
inserted into the compost windrow and the oxygen
concentration measured at equilibrium.

Minimization of windrow disturbance is of prime

81

importance in this study; probes are therefore left in
place. The probes are rigid, are inserted into the windrow
after the windrow is formed and then left in the windrow,
and permit both measurement of temperatures and sampling of
gasses in the windrow.

The probe used in this study is shown in Figure 4.1.. It
allows gas collection and measurement of temperature at
several points along its length. It consists of alternating
sections of slotted wooden dowels with perforated copper
tubing. The wooden sections reduce conduction along the
probe and are not structurally affected by the temperature
differences in the compost windrow environment, as is the
case with materials such as plastic tubing. The perforated
copper pipe acts as a diffusion and gas storage chamber.
Dowel and copper pipe outside diameters are matched to
reduce resistance to probe insertion and prevent oversized
holes from being formed. The tip of the probe is turned to
a point. Depending on the number of measurement locations
on a probe, between one and four slots are cut in each dowel
section. Thermocouple leads and a polyethylene tube are
laid in each of the slots; this pair continues down the
probe until it reached the location to be measured. The
tubing terminates inside the diffusion chamber while the
thermocouple is extended out one of the holes in the copper
tubing so that it senses the compost temperature outside the
probe.

Probes were constructed and placed so that 10 locations

82

.cowumoofi co ocwpcmdmp mumoEmco

commowwmp Lmaaou omumLOMLmQ Lao“

ob oco pm: monoLm .mcomuoom Loadou

0cm cocoo3 mcwumcumuam mcm3ocm mom
can mmuoumLmQEmu mamaQEmm Low mooLm "H.a muomwm

umamm .2262

 

 

 

 

 

 

 

 

 

Eu ma (HP) Eu ma . Eu ma )b) Eu ma III III III III.
who:
u>ooLu cm
uumz a ocmnae
A Uh.

 

 

 

330° oooz / 9.3:... cu

cucunw\m ouuoLOuuum

83
in a cross-section of the compost windrow could be
monitored. Triplicate sets of probes were constructed to
allow for three cross-sections (0.4 meters apart) to be
studied for a total of 30 points or 10 locations with 3
repetitions. Figure 4.2 is a cross sectional view of the

windrow showing typical probe placement. Location 7 was not

Sampling location

 

7,
l/
. 21 Z). I
3 ' l L0n1
"10L

 

1 3 6T $
Dimensions typical
'4--1.22m 1.22m 1.22m

Figure 4.2: Typical temperature and gas sampling
locations.

 

monitored because the windrows were not large enough.
Figure 4.3 shows an orthogonal projection of the windrow
and a collection system.

4.4: Temperature Measurement

 

4.4.1: Equipment and Method

Copper constantan thermocouples were connected to a
Campbell Scientific CR21X datalogger through a 32 channel
multiplexer. The datalogger provides reference temperatures
and the calibration curves for voltage—to-temperature

conversion. A different method of connecting the

 

 

 

A. Bulkrnatefialsanuflinglocafion D. Datahagger
B. Reiativehumiditysensor E. Multiplexer

C. Gas sampling/

Figure 4.3:

F. Gas sampling ports
thermocouple line

Orthogonal projection of test windrow
showing temperature and oxygen probe
placement. Windrow height and base
dimensions are typical“

85
thermocouples to the datalogger was used in each of the two
experimental windrows. In Windrow l, the thermocouples from
the corresponding locations on different probes were
connected in parallel and then run to the datalogger. This
allowed an average temperature for each of the 10 locations
to be obtained while using up fewer input ports on the
datalogger. In Windrow 2, the thermocouples from each probe
location were run directly to the datalogger so that the
temperatures at all 30 locations were separately monitored.
Temperature in the concrete slab under the compost, ambient
temperature and ambient relative humidity were also
monitored. The temperatures and relative humidities were
sampled every minute and averaged every 15 minutes.
4.4.2: Error Analysis and Calibration

Table 4.3 summarizes the sources of possible error with
the Campbell CR21X datalogger and AM32 multiplexer used with
copper-constantan thermocouples. According to the Campbell
literature (Campbell Scientific, 1985) the maximum error is
:1.5°C. A more detailed explanation of these calculations
is included in Appendix A.

The largest source of error shown in Table 4.3 is due
to the difference in thermocouple output. Accordingly, the
thermocouples were subjected to two tests. In the first the
complete measurement system was tested by inserting
thermocouples into an ice bath and then heating the water
until it boiled. Temperatures were recorded every 5 seconds

and averaged every minute using the same datalogger used in

86

Table 4.3: Temperature measurement errors.

 

 

 

ERROR SOURCE TEMPERATURE ( C)
Reference junction temperature 2 0.5
Thermocouple output 2 1.0
Thermocouple voltage measurement 1 0.05
Reference linearization t 0.001
Output linearization i 0.001
Total error i 1.5

 

the field. At the icepoint the thermocouples had an average
temperature of 0.42°C with a standard deviation of 0.l3°C.
At boiling, the average and standard deviation were 99.0°C
and 0.2°C, respectively.

The second test involved inserting the 33 thermocouples
that were actually used in the experiment into a constant
temperature water bath. Temperatures were recorded every 5
seconds and averaged over a minute. The water was agitated
with a propeller attached to a 3/4 hp drill. The water
temperature was varied between 27°C and 69°C, with four
different temperature plateaus being held. The results of
this test are shown in Table 4.4, with the average and
standard deviation of all thermocouples at a certain time
being given. This method was chosen because it was hard to
maintain the water bath temperature exactly constant over
time. Sample data from this experiment are given in
Appendix B.

As can be seen from the data in Table 4.4, the

87

Table 4.4: Water bath calibration of 32 thermocouples.

 

 

Number of Average Average
Data Temperature Standard
Points (°C) Deviation
30 27.556 0.004
30 35.554 0.271
25 42.118 0.001
25 55.742 0.008
36 68.119 0.054

 

variability between thermocouples was very low. Only the
second temperature step showed an average standard deviation
greater than 0.l°C. This is probably because heat was first
applied in this step and the induced convection currents had
not stabilized at the time of monitoring.
4.5: Gas Sampling and Analysis
4.5.1: Equipment and Method

Thin walled plastic tubing was run from each probe
diffusion chamber in slots cut in the wooden dowel to the
top of the probe (Figure 4.1). A connection was made to a
flexible plastic tubing. The total length of tubing to each
location was 4.9 m. The other end of the tube was connected
to a 3-way valve (Figure 4.4). A 3 cc syringe body plugged
with a rubber septum was attached to one of the two
remaining valve outlets. The third was open to the
atmosphere. All joints were sealed with silicon caulk and
were inspected before a sample was drawn. Joint integrity
was tested weekly when the windrow was turned and the probes

were removed. Other pieces of the apparatus consisted of a

88

Gcc syringe with

1 1/2" - 229 needle \
Rubber septum
\ .W‘
I I
I

3ccs rin ebod .1
Y 9 Y L5.
3-Way valve

[ O O J air—@—
/4

lasti t b'
/ Silicon caulk and tape p C U ”‘9 l

  

 

 

 

 

 

 

 

 
  
   

 

 

 

 

 

 

3 way pipette bulb TO probe
in compost

Glass eyedrop tube

Figure 4.4: Gas sampling apparatus. Valve is in
position to allow sample to be drawn
from diffusion chamber in pile.,

89
50 cc rubber pipette bulb, a 6 cc syringe and needle, and
evacuated test tubes with rubber septums. Samples taken
from Windrow 2 had a plastic collar around the probe to
discourage gas flow along the probe during sampling.

Gas samples were drawn at four times during each week
of treatment: on the lst, 3rd, 5th and 7th days from the
start of each turned windrow. Samples from locations closer
to the surface were taken first. The same order of sampling
was maintained throughout the experiment. A total of thirty
windrow and 3 ambient samples were drawn each day.

When gas samples were taken, the long sampling tube was
evacuated by attaching the compressed pipette bulb to the
open end of the 3-way valve, opening the valve, and allowing
the bulb to expand to its normal shape. This removed the
gas present in the tube since the previous sample. Gas from
the diffusion chamber was sucked up to the valve. This gas
had previously reached equilibrium concentration with the
gas outside the diffusion chamber. The valve was then
turned to open the passage between the diffusion chamber and
the 3 cc syringe body. A 6 cc sample was drawn from this
and discarded. Six cc was again withdrawn, approximately
0.2 cc was ejected to clear the needle, and the sample was
immediately injected into the evacuated test tube. The
‘volume of the test tube was 4 cc; gas samples were stored at
(about 1.5 atmospheres. Two cc of gas at pressures greater
‘than.one atmospheres was available for gas analysis. The

(Evacuated test tubes are commercially available and haVe

90
been successfully used to store samples for 2 weeks
(Grofman, 1986). Storage tube septums were sealed with
silicon caulk after the sample was injected into the test
tube to protect the sample for longer periods of time.

Gas samples were analyzed for carbon dioxide and oxygen
on a Carle Model 8700 Basic Gas Chromatograph. A silica-gel
column was used to analyze the carbon dioxide. A 5A
molecular sieve with a 60/80 mesh column in series with the
gel column separated and analysed oxygen and nitrogen. Two-
~to-four standard gases were used for developing calibration
curves over a range of 1 to 20 percent for both carbon
dioxide and oxygen. Concentrations were calculated based on
peak heights. A Hewlett-Packard model 3390-A reporting
integrator recorded areas under the peaks as back-up data.

4.5.2: Error Analysis

 

There are several potential sources of error in the gas
sampling and analysis system. The sources of error are

shown in Table 4.5.

Table 4.5: Potential sources of error in gas sampling,
storage, and analysis.

 

Gas drawn into sample from outside sample area.
Leaks into gas sampling tubes before sampling.
Diffusion through needle into sample before
injection into the sample tube and/or G.C.

. Flow of sample out of sample storage tube and
diffusion of gases (primarily oxygen) into sample
storage tube.

Withdrawal of sample at less than atmospheric
pressure.

Gas Chromatograph calibration.

Gas Chromatograph problems.

4:- (AMI—a
O O 0

NOW (J)
o

 

91

4.5.2.1: Effect pf Air Being Drawn into Sample from Outside
Sampling Area

 

The error caused by gas being drawn into the sample
from outside the sample area is indeterminate. A decrease
in pressure at the inlet to the sampling tube caused by
suction applied at the sampling end causes gas to move into
the tube. Gas will be drawn from those volumes with the
least resistance to flow. In an ideal situation, this gas
comes from a spherical volume immediately surrounding the
inlet. Changes in the cross-section of the probe will cause
the hole cross-section around the sampling point to be
larger than the probe itself (Figure 4.5). The air in this
void space has less resistance to flow than air in the dense
layer around the probe.

Three factors could increase this potential error.
First, if a portion of the probe has a blunt tip or a cross-
section that is much larger than average, the hole in the
windrow is larger than the probe. Second, even if the probe
and hole is initially the same diameter, over time the hole
near the surface can become larger due to windrow
settlement. As mentioned earlier, this problem was
encountered with the use of permanent probes left in the
windrows for one month or longer (Mote, 1986). Both of
these factors increase the likelihood that ambient air could
‘be sucked into the sample. Finally, the use of a vacuum
pump to draw samples exacerbates this problem.

In order to estimate and control this source of error,

three approaches were used: calculations of potential

92

 

 

  

xLlelx

" Pedorated copper tube

        

I x f1

Compost

\— Diffusion chamber

 

 

 

 

 

\

\— Wooden dowel .

     
 

 

*-

 

(.a75 1(1) ..
.4...

 

 

Figure 4.5: Cross-section of probe insertion hole
showing annular void space created by
probe insertion.

93
error, probe design, and experimentation. Calculation of
potential errors are made for two cases: (1) an air-filled
void around the probe, and (2), no void. Case 1 assumes
that air only comes from the annular void around the probe.

The affected height is calculated with the equation

1a = 0 (r02 - riz) / V5 (4.1)

length of air annulus affected by sample

withdrawal '
outer radius of annulus

where 13

r =
o . .
ri = inner radius of annulus
VS = volume of the sample
In Case 2, the air is assumed to come from a spherical

volume around the sampling point. The affected radius is

calculated with the equation

r3 = (Vs / n FAS)1/3 (4.2)

where r = radius of compost air affected
FAS = free air space

A range of annular ring cross-sections for Case 1 and free
air space for Case 2 are assumed. Sample size varies for
both cases. The estimated volume of the sampling apparatus
is also considered.

In the first Case 2 analysis, a range of large sample
volumes from 100 to 1000 cc was looked at. Assuming that 6
inches is the largest acceptable radius where no sample
contamination occurs, 300 cc samples are the largest that
can be safely taken for all potential compost free air
space. This leads to the rejection of two methods of gas
analysis: a Bacharach Fyrite tester and a flow-through gas

Chromatograph.

94

Analysis of Case 1 indicated that larger sample sizes
and small void space led to samples being drawn from further
away. Large sample size and large annular void spaces draw
air from less distance. It is easier for outside air to get
into the windrow and affect the long term composting process
in this case, however. A Case 2 analysis with smaller
sample volumes indicate that for most tube diameters and
lengths, the radius of influence is less than 6 inches.

The air sampling system incorporates two features to
reduce the possibility of contaminated air being sucked into
the system. Probes are designed and constructed with a
constant cross-section as determined by the materials and
the need for probe rigidity. Diffusion chambers are
incorporated into the probe to provide a reservoir of gas
equilibrated to compost gas concentrations. The volume of
the sampling system is shown in Table 4.6.

4.5.2.2: Effect of Withdrawing Succesive Samples 23 Changes
12 Gas CEncentration

 

 

 

An experiment was conducted to measure the changes in
gas concentration as successive samples were withdrawn from
a sample location. The probes were inserted into the
windrow in the same manner and locations as for an actual
experiment. Successive 5 cc samples were withdrawn until 75
cc was removed. Four different windrow locations were
sampled to provide a range of windrow conditions: 1, 5, 8,
and 10 (Figure 4.6). The results of one of the sampling

runs from Location 1 are shown in Figure 4.7. The

95

Table 4.6: Internal volume of sampling tubes and diffusion
chamber.

 

Volume (cm3)

 

Sampling system:

Tube ............................ 13.5
Syringe ......................... 5.0
Diffusion chamber ............... 7.2

25.7

Sample Size

Rubber Bulb... .................. 50.0
Discard....... ................ .. 6.0

6.0

Sample..........................

0‘
N
o

O

 

Sampling location

  

 

 

 

Figure 4.6: Sampling points for sequential gas
sampling.

96

 

 

 

 

22
20 - +
16 ‘ LOCATION 1
09
V 18 CI!
C -
O “ o T +
o: + + ‘ +
I.
2 1o-
0
U
S B -( o a o 0 0 a U
U
a .i
U)
(U
o ,4-
23
D
O I I I V T I f
0 2O 40 BO 80

Sequential Volume (cc)
0 C02 '9 02

Figure 4.7: Sequential sample gas concentration,
Location 1.

concentration of carbon dioxide in the first 5 cc sample is
low but rises close to its final value by the third sample.
Oxygen concentrations follow the reverse pattern. The lower
values of the 5 cc sample indicate that there might be some
leakage into the system, but that this is flushed out by the
adopted experimental method. The constant value at the 65
cc volume, which corresponds to the usual sampling point,
indicates that no contamination of the air sample occurred
from this source. Results of the other tests are shown in
Appendix C.

4.5.2.3: Effect 93 Leaks into Gas Sampling System

 

 

Leakage into the gas sampling lines was discouraged by

sealing all joints with silicon caulk. Each day before

97
samples were taken, the integrity of the lines and silicon
seals were checked and repaired if necessary. At the end of
each weekly batch, the probes were removed from the windrow
and each line was leak tested in a water bath. Blocked or

leaky lines were replaced.

4.5.2.4: Effect pf Diffusion and/or Gas Flow into Syringe
and Gas Storage Containers

Solutions to the third potential error were
incorporated into the sampling procedures. To discourage
diffusion into or out of the syringe, the syringe was kept
in the sampling apparatus for 6 to 10 seconds after the
second 6 cc sample was drawn. This allowed pressures inside
the sampling apparatus to approach atmospheric. After the
removal of the syringe from the apparatus, the ejection of
approximately 0.2 cc cleared the needle of any ambient air
that might have entered.

At the start of the experiment, the use of the
evacuated test tubes for gas storage was assumed to be
accurate based on the experience other researchers at MSU
(Groffman, 1986): analysis within 2 weeks of sampling does
not affect accuracy. Samples from the first windrow were
analyzed according to that timetable. Due to problems with
the gas chromatograph, however, analysis of the second set
of gas samples was delayed beyond the two week limit. In
addition, the oxygen levels in the center of the windrow
:seem high. An error was suspected in the gas concentration
.analysis procedure. Accordingly, two experiments were

(nanducted to assess the integrity and durability of the

98

storage system.

A possible source of error is the withdrawal of air
samples at less-than-atmospheric pressure. This occurs if
sample withdrawal caused sub-atmospheric pressures in the
syringe and sample storage tube. To test this, ten
evacuated tubes were filled with nitrogen. Successive 1 cc
samples were withdrawn and analyzed on a gas Chromatograph.
Figure 4.8 shows the effect of the sample volume withdrawn
on the measured oxygen concentration.

Syringe internal pressure was estimated based on the
volume of the sample container, the volume and number of
withdrawals and the initial pressure. Comparison (Figure
4.9) of the calculated syringe internal pressure to measured
oxygen concentrations indicates two almost linear segments.
Assuming that the initial and atmospheric oxygen
concentrations are 0 and 20.95 percent, respectively, the
influx as a function of sample number, sample volume, and
calculated internal pressure was calculated. The
relationship between influx and calculated internal pressure
is shown in Figure 4.10.

The preceding results needed to be converted from the 3
cc sample container size used in the calibration experiment
to the 4 cc size used in windrow monitoring. To accomplish
this, the influx values for the 3 cc container were
converted to volume influxes and then percent influx for the
4 cc container was calculated. The calculated 3 cc

container internal pressures that correspond to the volume

99

 

1.5 ATM, 3 CC BOTTLE 5 1 CC SAMPLES )

(3)
3
I

Oxygen

 

 

 

OdMUhUD‘IDO
l

 

I I T I

5

...-
U

Cummulative Withdrawals (cc)
0 Rep 1 + Rep 2 0 Rep 3 A Rep 4 x Rep 5 v Rep 6

Figure 4.8: Effect of cummulative gas withdrawal on
oxygen concentration.

16
15~ v
14-

 

1.5 ATM, 3 CC BOTTLE & 1 CC SAMPLES

(3)
a...“
o-‘NU
J_lll

Oxygen

 

dMUI-U‘OUOID
J

 

 

 

0
q
.4
1
4
..
.4
q
-1
..

0.2 0.4 0.6 0.8 1 1.2

Pressure (atm)
0 Rep 1 + Rep 2 o Rep 3 A Rvp 4 x Rep 5 v Rep 6

Figure 4.9: Effect of sample pressure (calculated)
on oxygen concentration.

100

 

100
so 1
30 1.5 ATM INITIAL PRESSURE, 3 cc BOTTLE

704

(1)

BO-

50-4

40--1

Influx

30+

20-

10--(

 

 

 

O I wI I I ‘I ‘—I I I I I I

o: os 07 03 L1 L3 L5
Internal Pressure (atm)

Figure 4.10: Percent influx as a function of
internal pressure.

withdrawn were used with the 4 cc influx values in a linear
regression. This equation was used to calculate the influx
at internal pressures corresponding to those occurring after
each successive 0.55 cc sample volume was withdrawn from the
4 cc container. Finally, a linear regression equation was
run on sample withdrawal number vs influx data to arrive at
an equation useful for transforming the experimental gas

concentration data. The resulting equation is

In4 0.0198 * S + 0.002 (r = 1.000) (4.3)

4 cc influx, decimal

where In4

S Sample number

Observations of sample containers that had their rubber
septums covered with silicon caulk after the samples were

injected indicated that there was some leakage out of the

101

sample tube. Since calculations of internal pressure
indicated that changes in starting pressure influenced the
speed at which a vacuum was produced in the sample
container, it was necessary to get an estimate of the error
due to variations in sample starting pressure. I assumed
that the worst case would occur if enough gas had leaked out
of the sample container so that the gas was at atmospheric
pressure. In this case, all syringe sample withdrawals
would be at less than atmospheric pressure and so subject to
an influx of air due to the pressure differential. I
assumed that this influx corresponded to the second and
steeper slope in the data shown previously in Figure 4.10.

In order to quantify this relationship a series of
calculations similar to the previous case was performed.
All calculations are shown in Appendix D. The resulting
regression equation (r = 0.998) for sample number vs. influx

is given by:

In4 = 0.0851 * S + 0.0027 (4.4)
The effect of the withdrawal of a sample of a given
volume and measured gas concentration on actual gas
concentration is described by Equations 4.5 for oxygen and
Equation 4.6 for carbon dioxide. For a given sample volume
the effect is linear and can be described as follows for

each gas:

02: O = (0m - In * Oe) / F (4.5)

c02: ca - cm / F (4.6)

102

where O = oxygen concentration, percent
C = carbon dioxide concentration, percent
a = actual
m = measured
e = ambient
In = influx gas in syringe, decimal
F = original gas in syringe, decimal

l — In
Equations 4.3 or 4.4 can be substituted into Equations 4.5
and 4.6 to achieve the desired correction.

Four different correction calculations were considered.
The first, the low influx method, involves the insertion of
Equation 4.3 into Equations 4.5 and 4.6. The second method
was accomplished by inserting the high influx equation 4.4
into Equations 4.5 and 4.6. The third method was achieved
by calculating the average of Methods 1 and 2. The fourth
method was based on two assumptions: the measured carbon
dioxide was correct and the sum of carbon dioxide and oxygen
concentrations was approximately 21 percent. With this
method, measured carbon dioxide was subtracted from 21 % to
achieve the oxygen concentration.

Method 3, based on average values, was chosen as the
most appropriate correction calculation. Method 4 does not
make the desired corrections. Methods 1 and 2 are subject
to the uncertainty of the initial sample pressure at the
start of analysis. The average method is at best a
compromise, because it is subject to the same uncertainty.

The measurement error due to the correction factor is
11.0 % oxygen at low concentrations and £0.25 % at high
concentrations. Measurement errors for carbon dioxide are

20.25 % at low concentrations and 11.0 % carbon dioxide at

103

high concentrations.

The effects of storage system type and time were also
examined. Leakage of gas through the hole in the septum due
to initial overpressure followed by diffusion of oxygen into
the sample storage tube could occur and affect measured
results.

Evacuated test tubes were filled with pure nitrogen to
1.5 atmospheres pressure. Half of these had uncovered
rubber septum tops as in Windrow 1 samples. The rubber
septum tops of the other half were covered with silicon
caulk to block the injection point, similar to Windrow 2
samples. Five tubes from each treatment were analyzed each
week for 5 weeks. The same four point calibration curve was
used to calculate the percent oxygen.

There were no observable oxygen peaks in weeks 1 to 3.
Week 4 samples showed small oxygen concentrations that were
statistically the same at the a = 0.05 level. Both covered
and uncovered samples were significantly lower than the
initial oxygen concentrations. Week 5 covered samples were
again significantly lower than week 4 covered samples. The
covered samples from week 5 showed a very large variance and
the average was significantly higher than any other time and
treatment. Two of the four samples from this group
exhibited the same concentration as the covered samples but
the other two were quite higher. The covered samples
exhibited small bubbles in the silicon, indicating that some

gas probably escaped from the uncovered tubes. This is also

104

supported by the larger variability in week 5 uncovered
samples. The results of this experiment do not elucidate the
effect of storage time and treatment although they seem to
indicate that for the covered system there is no loss of gas
and for the uncovered system that it was only after 5 weeks
of storage that a loss of gas in some storage containers may
have occurred.

Tentatively, it can be concluded that storage time has
no effect on the covered samples used in Windrow 2. While
the uncovered samples do show greater variability in week 5,
this was far beyond the actual storage time for Windrow 1
samples, and the time effect could be ignored.
4.5.2.5: Gas Chromatograph Calibration and Errors

Some potential for error is present in the procedures
used to calibrate the gas chromatograph. Initially, I
thought that point calibration would be adequate based on
the experience of other researchers using the same gas
chromatograph (Dilley, 1986). It shortly became apparent
that more points were needed and another bottled standard
was used for two point calibration; most of the samples from
Windrow l were measured using two-point calibration.

Samples from Windrow 2 were analyzed using a four point
calibration curve. The additional standards were made up
volumetrically.

The bottled standards were manufactured with a fixed,
absolute error. The volumetric standards were mixed before

each sample analysis session. The method of mixing involved

105

inserting septums into 160 cc sample bottles, alternately
evacuating the bottles with a vacuum pump and filling them
with nitrogen gas, and then filling the evacuated bottle
with precalculated amounts of oxygen, carbon dioxide, and
nitrogen to 2 atmospheres pressure. 30 and 60 cc syringes
were used to fill the bottles. Table 4.7 summarizes the
standards and variability used in this study. Error
calculations for the volumetric standards are given in
Appendix B.

Table 4.7: Gas concentration standards used in gas
chromatograph calibration.

 

Gas Concentration

 

Standard Level Std. Dev.
(% Atmos.) (% Atmos.)
coz 02 coz 02
1 5.079 20.524 $0.10 $0.61
2 0.983 4.97 $0.002 $0.15
3 10.0 10.0 $0.226 $0.226
4 1.0 20.0 $0.238 $0.167

 

The gas chromatograph itself presented a significant
source of error at times. Two types of problems arose.
When a sample had a significantly lower concentration than
the preceding sample, there was a lag in the apparent
concentration. This caused the first new subsample to
appear to have a higher concentration than it actually did.

The second problem was due to the erratic behavior of

the gas chromatograph when analysis of Windrow 2 samples

106
were to begin. As the first day's samples were analyzed,
the gas chromatograph became more erratic. Finally, not
even standards would behave consistently. The problem was
tracked down to a faulty injection valve. The gas
chromatograph behaved correctly after replacement.
Concentration data from two days were lost, however.

4.6: Core Sampling Method

 

 

4.6.1: Review pf Sampling Methods and Corer Design

Moisture content, volatile solids, coliform numbers,
bulk density, porosity and free air space are expected to
vary spatially. Accurate samples must be taken to
understand the heat and mass transfer process in DMS
composting and to provide data for future modeling efforts.
Sample location in the windrow must also be known
accurately. The sampling process, however, must disturb the
windrow as little as possible. Sample handling and storage
are also important but will be covered in later sections.

Moisture content and volatile solids samples are the
least difficult to obtain. Sampling does not change either
of these two variables, but it may disturb the windrow.
Hoyle and Matingly (1954) used a hollow auger that was 61 cm
long and 2.67 cm in diameter to take vertical samples from
61 x 61 x 61 cm composting bins. The material was analyzed
for moisture content, ash and various forms of nitrogen. A
grid was used to ensure that samples were not taken from the
same location in the windrow. No mention was made of the

effect of sampling on the composting process. Finstein et

107
a1. (1980) used a clamshell posthole digger (5 inch
diameter) to obtain samples from 7 ton windrows. Great care
must be taken with sampling for micro-organisms to prevent
contamination. Sampling for bulk density, porosity, and
free air space is more difficult. Bulk density
determinations are based on both sample weight and initial
volume; the act of sampling can change the volume of the
sample.

Disturbed samples can be used to get an estimate of
bulk density. Several methods that use disturbed samples
and the addition of water in calibrated containers have been
developed (Singley et al., 1982). These methods do not
adequately account for the effect of the overburden of the
compost and the compaction that it causes. The value of the
bulk density at a particular location over time is hard to
obtain as well. Several methods of determining soil bulk
density are described in the soil mechanics literature
(ASTM, 1986). These methods were either quite destructive
of the windrow or the required equipment was not available
for use in this study.

Core samplers are frequently used for soil bulk density
determinations. Raper and Erbach (1985) note that samplers
can either be hammered, driven in at a constant speed,
augered or inserted into the soil by some combination of the
above methods. Evidence that the length of the core sample
is shorter than the sampling depth indicates that compaction

may be a problem (Wells, 1959). In order to minimize the

108
effects of compaction, Baver (1956) recommends using at
least a 7.62 cm diameter tube.

Core sampling device design greatly affects sampling
accuracy (Raper and Erbach, 1985). Compaction can be
avoided by sampler point and interior sampler design.
Shaping and sharpening the sampler point shears the soil
instead of deforming it ahead of the sampler and also
directs unwanted soil to the outside of the sampler.
Providing a taper or interior sampler volume greater than
the interior diameter of the cutting head allows the sampled
material to expand and reduces the friction resisting the
penetration of the probe. Coating the interior with teflon
does not appear to make a significant difference in the
sampled core bulk density (Raper and Erbach, 1985).

The accuracy of standard soil corers was analyzed by
Baranowski (1983) with forest soils. He found that the core
method could reveal intertreatment densities to accuracies
no greater than :50 kg/m3, even with favorable conditions,
careful sampling and adequate replications. Under less
favorable conditions, the error was as much as 5 times
larger.

Peat sampling provides a system analogous to compost
sampling; both involve taking samples of a fibrous organic
material. Peat sampling equipment includes the use of
Macaulay samplers, core cutters and stationary piston
samplers (Jarret, 1986). Bulk density from these samplers

can be determined by the kerosene, paraffin, or cylinder

109
methods (Nat. Res. Counc. Can., 1979). The most interesting
method required a cylindrical cutter attached to a core
holder and base, powered by a hand drill. In this method,
bulk density was determined by dividing the oven-dry weight
of the peat core by the inner volume of the cylinder.
Stanek (1980) compared the results of bulk density
determinations made by the paraffin and cylinder method to
the kerosene method which was considered to be the most
accurate. He found that the cylinder method correlated well
with the kerosene method (r 5 0.979). Based on this
information a method was adapted for sampling a compost
windrow.

4.6.2: Comparison of Core Sampling Methods

 

 

The rotary corer required some skill to operate
properly. Unless the core was held very steady, it would
rotate off-course and take in material from a volume greater
than the volume intended. When the corer was removed, some
material from around the lip occasionally fell into the
hole. The sides of the cylindrical hole did not appear to
be excessively disturbed. A narrowing of the hole began
within 2 minutes and the bottom appeared to rebound
following removal of overburden pressure. The end of the
core sample showed evidence of the rotation and usually
broke off cleanly. The length of the core sample was
usually 10 to 20 % less than the hole in the windrow,
indicating that some compaction had occurred in the corer,

windrow or both. When the corer was used to take samples of

110
low density material such as that near the windrow "toe," it
was observed to push some material aside. This was not
observed in the more dense portions of the windrow. Two
different length and diameter corers were tested. The
smaller diameter corer was harder to insert into dense
portions of the windrows; it was unable to penetrate
windrows composed of long straw.

An experiment was conducted to estimate the accuracy of
several core sampling methods for bulk density, porosity and
free air space determinations. Three 55 gallon drums were
filled with 1 week old separated manure solids that had been
mixed in a silage mixer. After leveling the tops, three
sampling methods were used: (1) a 20.32 cm diameter ring,
(2) a short 7.78 cm diameter rotary corer and (3) a long
5.08 cm diameter rotary corer. The rotary corers were
powered by a 3/4 hp electric drill.

Three samples of approximately 30 cm were taken from
each barrel by the two rotary corers. The ring sampler was
then pushed and rotated into the surface and the sample was
then removed by hand. Approximately 30 cm of material was
removed from the barrel and the ring sampling procedure was
repeated. This was repeated to obtain the third ring
sample. The above steps were repeated in each of the
barrels until 3 replications of sample depth for each method
had been obtained. Samples were stored and analyzed for
weight, moisture content and volatile solids. Bulk density,

porosity and free air space were calculated. The methods of

lll
analysis and calculation are described below in the
appropriate sections. Detailed results are given in
Appendix F.

Wet bulk densities ranged from 198 to 566 kg/m3
depending on the sampling method and depth. Precision was
good with the average coefficient of variation for each
method at all depths being 8.7, 10.9 and 10.0 % for the
ring, short corer, and long corer methods, respectively.
Using the two-tailed Student's t test, the three methods
were significantly different in at least two of the three
depths at the 90 % level. The short sampler always gave the
largest bulk densities followed by the long sampler and the
ring method.

Free air space values in this experiment ranged from
0.850 to 0.471. Using the same statistical test, the three
methods were significantly different in at least two of the
three depths at the 90 % level. The average coefficient of
variation for each method at all depths was 3.2, 7.5 and 3.0
% for the ring, short corer and long corer methods,
respectively.

The ring method was not chosen for use because it could
not be used to sample windrow interior bulk densities. The
long corer with the small diameter was not chosen because of
the difficulty encountered with inserting it into some
windrows made of long straw. The short wide sampler was
selected because it would be able to penetrate a wider range

of compost material and its greater diameter would be less

112
succeptible to errors in the radius determination than the
longer, narrower sampler.

4.6.3: Description 9f Experimental Core Sampling Method

 

Samples for determining moisture content, volatile
solids, bulk density and free air space were obtained on the
lst, 3rd, and 7th days after a windrow was turned. A 7.78
cm diameter rotary corer attached to a 3/4 hp electric drill
was used to drill vertical sections out of the windrow. The
corer is shown in Figure 4.11. The cutting edge is serrated
to cut through fibrous material and slanted away from the
path of the core opening to push away any material not
directly in the path of the core opening. To relieve
friction from the surrounding windrow, the cutting head
exterior dimension is greater than the core body.

Similarly, the interior diameter of the cutting head is
smaller than that of the corer body interior to allow the
compost material to expand and give reduced resistance to
the entry of material at the cutting head opening.

The corer allowed vertical cylindrical compost samples
of variable length to be removed from the windrow. The
vertical center of each approximately 30 cm core sample was
at the same horizontal and vertical point as a temperature
and gas measurement location (Figure 4.12). At least 9
samples were taken at each time. No replicates of these
samples were taken because additional samples would have
seriously disturbed the windrow.

The sampling procedure involved first determining the

 

113

 

\
.1.
T

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1.27 cm
' a F
(0.5 In.) I V l \
‘ \N": 1.27 cm
Q2 \ I (0.5 in.)
N I
N
N
N I
N 9" \\
N . .
8.89 cm (3.5 In.) Diameter
/’ \v///\\ 1/16"1TfickroHedsteel
( LA ‘1 .
a A»
N l N
N N
. N t
74.9 cm N 1
(29.5in.) , N 0
N I j
N L‘
V
V N
N
N | h
N W
N :1"
T 0
108cn1 ¥ l \\
(4.25 in.) ($5775?) 1 1'
' ' 8.57 cm (3 3/8 in.) 0.0.
V ¢ . 1 3/16" Thick pipe
7 78 3.8) cm
.. _ D: ’ cm 1.5. .
(3.06in.) (‘— ( m)

 

 

Figure 4.11: Rotary core sampler.

114

Center of sampling location

 

-1313;

0
O
u

 

 

Figure 4.12: Core sample locations. Cylindrical
in shape, the center of the sample was
located at approximately the same
point in the yz plane as the
temperature/gas sampling p01nt.

location of each sampling point below the windrow surface.
The beginning and ending locations of each sample were
calculated. A catwalk placed directly above the sample
location provided a place for the operator to stand while
taking core samples. A colored scale on the side of the
corer guided the operator to the proper sampling depth. The
exact hole depth was measured immediately afterwards. The
compost sample was removed from the corer and placed into
either a ziplock bag for moisture content and volatile
solids analysis or a "whirlpak" for coliform analysis. If a
visible color difference was present in a sample, the sample
was split and separate analyses performed. The holes
created by the withdrawal of samples were filled with DMS
compost material of a similar age, lightly compacted and
marked.

In the experimental method adopted, a single large

sample from each location was taken at each time. The first

115

set of samples was taken from one side of the windrow after
the windrow was one day old. The second set was taken from
the opposite side on the third day and the third set was
taken on the seventh day from the same side as the first
set, approximately 1.5 feet away. This method diminished
the effect of sampling on the transport processes, but the
method could be criticized because the second samples were
taken from an area that was different.

4.7: Moisture Content and Volatile Solids

 

 

 

4.7.1: Methods and Calculations

 

 

Gravimetric moisture content (wet basis) and volatile
solids (total solids basis) were determined using Standard

Methods for the Examination of Water and Wastewater (AWWA,

 

 

 

 

1976). Gravimetric moisture content on a dry basis,
volumetric moisture content and volatile solids on an ash
basis were calculated.

Samples were removed from the pile as described in the
bulk sampler section, and were placed in ziplock bags and
stored at 2°C. Samples were first mixed by hand kneading
the bag with the sample still inside. Sub-samples were
removed by tongs and placed in ceramic crucibles. The
crucibles had previously been fired in a furnace at 550°C t
50°C for 1 hour, cooled in a dessicator and weighed. The
wet samples and crucible were weighed and dried in an oven
at 103°C i 2°C. After 24 hours, the sample were removed,
cooled in a dessicator and reweighed.

Volatile solids were determined by ashing the dry

116

matter in covered crucibles at 550°C 1 50°C for 1 hour. If
black or red material remained in the crucible at this
point, the crucibles were placed back in the furnace for 5
minutes, sans covers, to completely ash the samples. After
removal and cooling in a dessicator, the crucible and ash
were weighed. I

A Precision Scientific model 1254 oven, a model FD204C
Hoskins Electric furnace, and a Sartorius model 1265 MT
scale were used in this procedure. The scale is accurate to
t 0.001 g.

The following calculation for gravimetric moisture
content (wet basis) was used:

ucwb = [1 - (MCS- Mc)/(Mcsw — MC)] (4.7)

where Mcwb moisture content, wet basis, decimal
' M
x

Mass of x, g

csw = crucible, solids and water
cs = cruelble and solids
c = crucible

One of the problems with wet basis moisture content
calculations is that changes in the mass of water change
both the numerator and denominator. This makes it
impossible to directly compare the moisture contents at any
two points in the process.

If moisture contents are expressed on a dry basis,
however, only the numerator changes. The dry basis
calculation is as follows:

MCd = Mw / Md (4.8)

moisture content, dry basis, decimal

where MCdb
w water

117
d = dry solids
A comparison of wet (% wb) and dry basis (% db) moisture
contents is presented in Table 4.8. Dry basis moisture
contents are much more sensitive to changes in the weight of

Table 4.8: Comparison of moisture contents expressed on a
wet and dry solids basis.

 

Method Moisture Content (%)

 

MCw 90 80 70 60 50 40 3O 23 20
MCd 900 400 233 150 100 67 43 30 25

 

water that is present in a substance.
Volatile solids were calculated with:
vs 2 100 * [1 - (Mca - MC)/(MCS - MC)] (4.9)

where VS a volatile solids, dry solids basis, decimal
ca = crucible and ash

Expression of volatile solids on a total solids basis
suffers the same drawback as the wet basis moisture content
measurement. Both the numerator and denominator change as
the reaction progresses. This makes it impossible to
directly compare the volatile solids at any two points in
the process.

Another way of looking at substrate consumption using
the same data is to calculate the change in the volatilized
solids based on the ash fraction that remains.

vsa = MV / Ma (4.10)

where VS Volatile solids ash fraction, decimal
volatilized solids

ash solids

w <0:
u H II

118

Since the ash fraction should not change unless additional
inert material is added to the composting process, VSa
should provide a stable base for evaluating substrate
consumption.

Table 4.9 illustrates the difference between the two
methods of expressing substrate consumption. Ash solids
expressions of substrate consumption indicate a much larger
decrease in substrate than the volatile solids method.

Table 4.9: Comparison of volatile solids expressed on a
dry and ash solids basis.

 

Method Substrate (%)

 

VS 95 93 90 89 85 82 80
VSa 1900 1329 900 809 567 456 400

 

Three replications of each measurement were done initially
in Windrow 1. Five were used initially in Windrow 2.
Outlying data were rejected with the Q test (Davis, 1981).
If the variance was greater than one percent, an additional
5 samples were run.

4.7.2: Sensitivity Analysis

 

 

The sensitivity of moisture content and volatile solids
determinations to errors in their measurement was determined
using the theory of the propagation of errors (Parratt,
1961; Topping, 1951). The following paragraphs summarize
the derivation presented in Appendix G, which was orginally

given by Parratt (1961).

119

A result U, which is computed from the quantities X
and Y has the following variance for indeterminant errors:

5,.2 = (an/ax)2 5x2 + (aU/ay)z 5Y2 (4.11)
and sy2 = variances associated with
measurement of X and Y.

2

where sx

For addition or subtraction the variance of the calculated
result is given by the sum of the variances:

2 + z
5 SY (4.12)
If U is a result of multiplication or division such as

U

U = x3 ya (4.13)
then the variance is given by

= [(axa"1yb)2 5,,2 + (bxayb'l)2 sy2 1 (4.14)

2

5u

The fractional variance is given as:
suz/ u2 = [(sx/x)z + (sY/y)z] (4.15)

The fractional variance can be put in terms of percentage by
multiplying by 100.

Fractional variance for the wet basis moisture content
calculation is given by:

SMCwbz/ MCwbz = [((scs2 + scz)/ Mdz )

+ Use“2 + sc’)/ Mwsz)] * 100 (4.16)
where ws = wet solid

A similar sensitivity analysis can be run for the dry

basis moisture content calculation. The equation is

sMCdbz = (l / Md)2 SW52 + (MWS / Mdz)2 Sdz (4.17)
Where SW52 = variance of wet solids measurement
= 10.002 9
’ssz = variance of dry solids measurement

10.002 9

120
The variance of volatile solids (dry solids basis )

is given by

Z 2 2
+ (sCS + sC )/ Md ] * 100 (4.18)
where 5X2 = variance associated With measurement

Despite the large observed VSa changes, VSa is still
subject to the same errors present in the volatile solids
determination. Errors in VSa determination were analyzed
for selected data using a sensitivity analysis procedure

similar to that used for VS. The equation developed was:

SVSaz = (1 / Ma)2 sd2 + (Md / Mal)2 sa2 (4.19)
Where ss2 = variance of dry solids measurement
= 10.002 9
sa2 = variance of ash solids measurement
= $0.002 g

The average sMCwb for 75 moisture content (% wb)
determinations with measurement error variance of i 0.001 g
is 2.26 % wb moisture content. Table 4.10 summarizes the
results for six measurements selected to cover the typical
range of dry solids mass and moisture contents encountered
in the experiment.

The average standard deviation of volatile solids
(total solids basis) was much higher than that for MCw: 7.0
percent volatile solids. Four place accuracy in weighing is
necessary to assure the same level of accuracy in volatile
solids measurements as is achieved with 3 place accuracy of

moisture content determinations.

121

Table 4.10: Sensitivity analysis for moisture content
expressed on an dry solid basis.

 

 

Wet Solids Dry Solids Mw Md 52 s
(g) (g) (%) (%)
1.447 0.394 78.6 586 29.4 54.2
2.345 0.701 77.0 334 7.7 27.8
0.881 0.307 74.2 287 34.6 58.8
2.232 0.659 77.1 339 9.3 30.5
1.141 0.316 78.3 361 44.5 66.7
1.623 0.481 77.1 337 17.4 41.7

 

Table 4.11 summarizes the results for six volatile

solids content measurements selected to cover a typical

Table 4.11: Sensitivity analysis for volatile solids
expressed on a dry and ash solids basis.

 

 

Dry Solids Ash Solids VS VSa 52 s
(g) (g) (%) (%)
0.307 0.044 85.7 598 5130 716
0.461 0.083 82.0 456 925 304
0.672 0.153 77.2 339 374 193
0.415 0.087 79.4 385 628 250
0.481 0.073 84.8 559 1670 408
0.338 0.051 84.9 563 3450 588

 

range of ash mass and volatile solids. In some cases the

standard deviation of the measurement is larger than the
measurement itself. The use of scales accurate to :0.001 mg

is necessary to achieve reasonable accuracy.

122

4.8 Bulk Density

 

4.8.1: Calculations

 

Bulk density was calculated using the "cylinder"

method. The equation used in this calculation was:

0b = Mnet F / n r2 h (4.20)
where 8b = wet bulk density, kg/m3
Wnet = net mass of compost removed from corer
r = corer cutting head interior radius, in
h = sample height, inches
F = unit conversion factor

The method used to obtain quantities necessary for bulk
density determination has been partially described in the
core sampling section. The length of the sample was
determined by measuring its furthest penetration depth, and
subtracting that from the penetration depth of the previous
sample. A colored scale on the corer was used to take
initial length measurments. This was immediately confirmed
with a finer scale. Values were accurate to t 1/8 inch.

The net mass of the compost material was calculated as:

M = M15 + M - Mb (4.21)

net SS

where ls = large sample weight, 9
ss = small sample weight, g
b = bag weight, 9

4.8.2: Sensitivity Analysis

 

The variance of bulk density due to measurement errors
can be broken down into four areas: net weight of solids,
moisture content, effective core height and effective core
radius. Variance due to the net mass of the solids is
additive:

5M2 = 51 2 + s 2 + sb2 (4.22)

123
where the terms are as defined previously. For the
analytical balance used, all variances were i 0.001 and

s 2 equals 1 0.003. Moisture content errors have been

net
analyzed previously.

The effect of errors in length measurement on the
percent volume variance is shown in Figure 4.13. This is
based on the following calculation:

st12 / Vol2 = [(sh h)2 + (sr 2r)2 (4.23)
where Vol = volume of sample, m3
From these data it can be seen that height does not greatly
affect volume. Small volumes are more sensitive to errors
than larger volumes.

Radius errors effected volume variance to a much
greater degree than errors in height. Figure 4.14 shows the
effect of changes in radius variance from t 1/8 to 1 5/16
inches. The previous equation for variance was used in this
calculation as well.

Turning to the bulk density measurement, the partial
derivative of bulk density with respect to weight is given
by

Bob / awnet = (l / n r2 h) (4.24)
where the terms are previously defined. This shows that the
rate of change of bulk density with respect to weight is a
constant.

The partial derivative of bulk density with respect to

sample height is

30b / 3h = - 0b / h2 (4.25)

124

 

 

 

 

5.5
A .
SA“ RADIUS VARIANCE +/- 0.125 1n
5.3 -
a. 5;:.
“e’
a 3.1 - °
H
O .4
> S
« ASJ
a ‘.a .- +
A
g 1&7- A
Q A
'2 4.6-1 0° A
0 ° AA
>' (AS - D o A
'4'. 0° A
‘4 t. o o A.
. ‘1 9 ‘. o A 6 A
0 0° A
43~ ch°°un;"**332§§‘
"2 i I U T t“ V fir F I U T 1' V U I T U I
O 0.2 O.‘ 0.8 0.8 1 1.2 1 4 1.8 1.8 2
Volume (liters)
o t/- 0.125 in + t/. 0.25 in o t/- 0.375 in A f/- 0.5 in

Figure 4.13:

Effect of length errors on total
variance as a percent of total volume.

 

 

 

 

14
13-
12a LENGTH VARIANCE: +/- 1/4 in
11...
.5 mmeXWXXXXXXX x
E 10-
3
0.. 9—4
3 J AAAAAAAAAAAAAAAAAAA A
B
a
v 7.1
0000000000000000000 O
o 5‘
U
S 5"
.,.. H*”+.‘+4++++++++ +
1.. ‘4‘
G
> .3-
2-
1...
o I Ti—T’W t I FTiI’ Y Y T 1 I Y 1 Y T r
O 0.2 0.4 0.6 0.8; 1 1.2 1.4 1.8 1.8 2
Volume (liters) _
+ +/- 1/8 in 0 t/- 3/16 in A :/- 1/4 1r1 X */- 5/16 in

Figure 4.14:

Effect of changes in radius variance
on total variance as a percent of
total volume.

125
Changes in bulk density due to height vary with the inverse
of the square of the height.
The partial derivative of bulk density with respect to
sample radius is

Bob / 3r = - 2 0b / r3 (4.26)

Errors in bulk density with radius variations vary inversely
as the cube of the radius.

Figure 4.15 shows the effect of two different radius
errors on bulk density. The fractional variance equation

used for this calculation is
spbz / ab: = W2 Swz + 4 st2 / r2 + shz / h2 (4.27)

The relative contributions of various measuring errors
on bulk density are shown in Figure 4.16. Figure 4.17 shows
the total, radius and height percent variances. Percent
variance due to radius measurement errors makes up most of
the total percent variance. Errors due to weighing and
moisture content are smaller than that for core height.

While this analysis is generally correct, it does not
account for errors from several sources. First, if the core
does not travel in a straight line as it goes into the
windrow, larger effective volumes may be sampled. Secondly,
the effective radius may be larger or smaller than the
interior core head radius because of compost particles from
outside or inside this radius being pushed out or pulled
into the sample. Vibration of the core may shift some

compost material out of the sampler. The effect of core

126

 

 

 

 

9
VARIANCE: WEIGHT +/- 1 g,
5‘ HEIGHT +/- 1/4 in, MC Exper.
7-
3
v B '- I!) 4 + 4 + C1
3 ++ + *
c '5 ¢ 4 : +
.2 5‘ a‘ +0.. .‘ ¢ . ‘ ++ +
E ,¢w€‘+ “r “o: «1 +
"1 0:10 o 0
DO D O
3... Dub (9 00 [P O D
D txiecpmg'o °
ufihfififi5aprcp5= ‘1
2 I V I V I I
O 200 400 800
Wet Bulk Density (kq/m3)
0 +/- 1/16 in + *l- 1/8 in

Figure 4.15:

Effect of two different radius errors
on total variance as a percent of bulk
density measurements.

 

 

 

 

0.38
A
VARIANCE: WEIGHT *I- l g, HEIGHT *l- 1/4 in,
0.3.. RADIUS 9/- 1/8 MC Exper.
0 4 + g.
0.25 d f
A. 4+ +
d'
0.2 - +
8 + 6 4 4 +-
C * 4
CO 0.15 -‘ 4+ 4* 4.... 4.
o: A + A A
g A “A 6' o 4
OJ 4 g? ** 999-2» 3 4
* A
-H.+ o . ‘*¥ +
‘ngmr9 f'tv 1
0.0: ~ at :3 a A o- o A
A a 0 6
Q 35° . A ‘4 A A
a A A A
0 , 5fliaEaE3flnaya4uLJL41_ca_4nfid_;p__hr_hu___n
O 200 400 800

Figure 4.16:

Wet Bulk Density' (kg/m3)
0 Weight # height 9 MC

Relative contribution of various
sources of error to total variance as a
percent of bulk density measurements.

 

 

 

 

' 4
£332.“ x x X X x V i“ x )Q ‘(
.'I X 7 '
smegma” z: .. . . . q
4...
f: VARIANCE: WEIGHT +/" 1 g, HEIGHT t/- 1/4 in,
3" RADIUS */- 1/8 MC Exper.
m
u
c
m
'C 2—
m
>
1..(
9* 4:
444 4 4
o .Mn‘h9fln L4” + t *+ T * ’
I I Y I 1'
o 200 400 800
Wet Bulk Density (kg/m3)
0 Height + Weight 0 Total BD

Figure 4.17: Relative contribution of radius and
height errors to total variance as a
percent of bulk density percent
measurement. -

rotation speed on these errors is also not described.

4.9: Porosity and Free Air Space

 

 

4.9.1: Calculations

 

Porosity and free air space can be calculated with the
results of the moisture content, bulk density and volatile

solids content. The general equation for these calculations

is:
fx = l - (0b / 05) (4.28)
where fx = porosity or free air space
9b = wet bulk density, kg/m
05 = particle density, wet or dry, kg/m3

Based, in part, on work by Bohnhoff et al. (1984), the
particle density used in porosity determinations is:
05d = (A - B vs ) / (100 — MCwb) (4.29)

where 05d =' particle density, dry, kg/m3

128
A,B = factors associated with solid particle
density
A and B are the parameters of a linear regression of
volatile solids content vs. DMS particle density from three
dairies with different separators and rations. These values
are given in Table 4.12.

Table 4.12: Values of constants A and B from dairy manure
solids specific gravity measurements.

 

 

Source A B
Farm 1 (Centrifugal) 232000 814
Farm 2 (Stationary screen) 235600 868
Farm 3 (Perforated drum/

compressing roller) 200600 473
Average 227800 773

 

Source: Bohnhoff et al. (1984)

Bohnhoff et a1. (1984) presented the following formula

for particle density to be used in determing free air

space:
+ (105 / mcwb)] (4.30)
where 9pm = particle density, moist, kg/m3

Equations 4.23 and 4.24 are substituted into Equation
4.22 for the calculation of porosity and free air space,

respectively.

I = 1 — 0b * (100 - MCwb) / (A - B vs) (4.31)

fa = 1 — 0b * [(100 — MCwb) / (A — B vs)

129

+ (Mcwb / 105)] (4.32)

porosity, decimal
free air space, dec1ma1

where f
fa

and all terms are as previously defined.

4.9.2: Sensitivity Analysis

 

Variance due to measurement errors was analyzed in a
manner similar to that of the previous sections. The
effects of sample weight, sample radius and height, moisture
content and volatile solids were considered. The effects of
changes in the parameters A and B were also examined.

The partial derivatives of porosity with respect to the
various factors are shown in Table 4.13. The partial
derivatives with respect to bulk density, weight, and
moisture content are all linear. The partial derivatives
with respect to radius and height are similar to those for
bulk density, varying with the cube of radius and the square
of height, respectively. The partial derivatives of
porosity with respect to the volatile solids and factors A
and B are not linear but change with the volatile solids, A
or B being divided by the square of the term (A - B*VS).

The equation used to estimate the total variance in

porosity is given below:

2

sf [(C/Vol)z * 5M2 + (0b C/h)2 * shz + (2 8b C/r)z * sr2

+

(ob/(A-B*vs))2 * sMCwb2 + (0b*A/(A-B*VS)Z)2 * 5A2

+

(pb*B/(A-B*VS)2)Z) * 532

+

(0b*V/(A-B*VS)2)Z * sv2 (4.33)

where C = (100 - MCwb) / (A - B VS)

130

Table 4.13: Partial derivatives of porosity calculation.

 

f = l - 0b * [100 - MCwb) = 1 - (M / n r2 h) * (C)
A '- 3 VS
if = - ( C )
30b
8f = - ( C )/ n r2 h
8E
3f = ( C )/ w r2 h2
85
8f = 2 W ( C )/ n r3 h
8?
3f - oh (1 / A - B VS)
EfiCwb
3f = -0b(-B / (A - B VS)2)
83
at - -ob(B / (A — a v5)2)
88
at = -pb(VS / (A - B vs)z)

3;"

 

131
The variances contributed by each of the factors of
porosity variance are shown in Table 4.14. Three levels of
variance for each factor were used; they were labeled Low,
Medium and High. The above equation was applied to 75
porosity measurements. The porosity variances presented in
the following figures are given in terms of actual porosity

Table 4.14: Error levels of contributing factors to
porosity and free air space variance.

 

 

Error Error Level
Source Low Medium High
Net Mass (9) 10.001 10.01 10.1
Core Height (in) 10.125 10.250 10.375
Core Radius (in) 10.125 10.250 10.375
Moisture
Content (% wb) 11 15 110
Factor A 110 1500 11000
Factor B . 11 150 1100
Volatile
Solids (%) 11 15 110

 

values and not as a percent of the actual porosity.

The total porosity variance for the different variance
levels is shown in Figure 4.18. For all three error levels,
80 percent of all measurements had variances less than 2
percent porosity. The higher variances occurred with the
lower porosity readings. These corresponded almost entirely
to the deepest samples in the center of the windrow. One
characteristic of these samples was that they were
typically one-half the length of other samples due to the
length of the core sampler. It is also interesting to note

two linear regions in the data, one that corresponds to

132

 

 

 

 

 

A 0
00
Q
>. 5‘
a
.III 0
m
8

4...
2 ° .
v

o
0 .—
u 3 + ° 0
C
.2 . o°
1‘ o
g 2-( 4- 4‘ 6 o
1'
2: + + ..
In 1.. 1 1’ °o o
8 O... ‘Cpooox
O o "4’a°~0.6
0. 4. ' 43333
O U o o m « £
0 ' mGLYILanqfi ‘ 1

I
0.9 0.92 0.94 0.98 0.98 1

Porosity (1)
a Low 1 Medium 0 high

Figure 4.18: Total porosity variance for 3 levels
of factor variance.

porosities of 95 percent and greater and the other
corresponding to porosities between 90 and 95 percent.
Closer examination of the high error case shows the
contributions of several factors to the total porosity
(Figure 4.19). Variance in determination of Factor A
contributes almost all of the variance in porosity with the
radius variance coming in a distant second. The effect of
errors in the measurement of Factor B, height, moisture
content, volatile solids and weight are almost negligible.
Factors A and B arise from a linear regression of volatile
solids data to determine the specific gravity of the solids
that was done by Bohnhoff et al. (1984). This measurement

was not part of the current research. Given the sensitivity

133

 

 

 

:; V
i
U 5.1
n v
o A
3
A. “ V v
A
U V
_ A
3 3 A V v
on A
C A a”
2 a- .1 '
>4 5 9
t: A W:
'0 Ab‘fivq.
O 1 ‘- U A: V
L. \
a x Email
a. h
+ + A I I
0 q—a—p—I—fi—fiiifi-yt-Magmfieém— u—.———(
0.9 0.92 0.94 0.98 0.98 1
Porosity (1)
0 Height 0- Radius 0 MC A A x B v Total

Figure 4.19: Contributions to total porosity
variance due to high factor variance.

)

of porosity to Factor A, this measurement should be
performed in future work of this nature.

The partial derivatives of the free air space with
respect to the various factors are shown in Table 4.15. The
results are similar to those for porosity with a slightly
different product. The partial derivatives with respect to
bulk density, weight, and moisture content are all linear.
The partial derivatives with respect to radius and height
vary with the cube of radius and the square of height,
respectively. The partial derivatives of porosity with
respect to the volatile solids and factors A and B are not
linear but change with the volatile solids, A or B being

divided by the square of the term (A - B VS).

134

Table 4.15: Partial derivatives of free air space

 

calculation.
fa = 1 - OB * [100 - MCwb + MCwb] = 1 - (W / n r2 h) * (D)
A - B vs 105
8f = - ( D )
'56:
afa = — ( 0 )/ n r2 h
EH
afa = ( 0 )/ n r2 h2
FE
afa = 2 W ( D )/ n r3 h
8?
8f = o ( -1 + 1/105 )
mewb b A " B vs
at = -p (-A / (A - B vs)2)
5A3 b
at = -0 (B / (A - B vs)‘)
53a b
at = -pb(VS / (A - B VS)2)

a.

 

135
The equation used to estimate the total variance in
free air space is given below:
sfa2 = [(D/Vol)2 * 5M2 + (0b D/h)2 * shz + (2 9b D/r)2 * sr2
+ (pb*[(1/(A-B*vs))+1/105])2 * sMCwa
+ (0b*A/(A-B*VS)2)2 * 5A2 + (ob*B/(A-B*VS)2)Z) * 582

+ (pb*VS/(A-B*VS)2)2 * svs2 (4.34)

where D = ((100 - MCwb) / (A - B*VS)) + MCwb / 105

Free air space values have a much greater range than
porosity: 35 to 87 percent. Variance due to errors in
measurement are also muCh greater than those for porosity,
with the maximum variance for each error level being 19,
11.5 and 4.5 percent free air space for the high, medium and
low levels, respectively. This is shown in Figure 4.20,
free air space total variance. Contributions of various
factors to free air space total variance are somewhat
different with errors in radius measurement being larger
than errors in the measurement of Factor A (Figure 4.21).
Errors in the measurement of other factors are negligible.
4.10: Additional Physical Properties

A number of other physical and moisture related
parameters can be calculated from the results of the bulk
density, porosity and free air space calculations. These
parameters can provide additional insight into the internal
state of the windrow.

Gravimetric moisture content provides information on
the weight of water present in a given windrow volume, which

is important for calculating specific heat. The relative

%)

(FAS,

FAS Variance

136

 

19
18
17-
18-(
134
14-
13-
12..
11 -
10-

" 0
DO 00 ° 6 ghb

 

 

Figure 4.20:

%)

(FAS,

FAS Variance

19

I I I

07
(t)

0.5

Free Air Space

Low Medium High

‘“ Oinfiflgii E‘szzfiiggii
O u .

 

Total free air space variance for 3

levels of factor variance.

 

18
17..
18
1.1
1“
15~
12-
11—
10-
9-
ad
7—
8-
5.1
44
3...
2.:
‘—

 

0

0.3

, R—F—fl—I—ll—C
0.5

Figure 4.21:

0 Height

Free Air Space

+ Radius 0 MC

AA

(3)

X

B

 

 

v Total

Contributions to total free air space
variance due to high factor variance.

137
weight of water, as opposed to that for the dry DMS, can
mask important differences in the relative volumes of air,
water and solids. Volumetric moisture content presents
information on the volume of water present in a unit area.

The volumetric moisture content can be calculated by

0 = Mwb 8b / 0w (4.35)
where 8 = Volumetric moisture content, decimal
9w = water density, kg/m

If the values of porosity and free air space are known,

8 can be calculated as follows:

0 = f - fa (4.36)
The volumetric solids content can similarly be

calculated by
f = l - fa (4.37)

where fs = volumetric solids content, decimal

The void ratio relates the volume of the void to the
volume of solids. The definition is given as follows:

8 = (V

air + Vw) / Vd = Vp / (Vt - vp) (4.38)

where void ratio, decimal
volume of x

air

water

pores

total

#15421»: n

Void ratio can also be calculated-from the relationship
8 = (1 - f) /.f (4.39)
where f is the decimal porosity. Equation 4.35 will be used

in this study.

The degreee of saturation, that is, the volume of void

space occupied by water, is calculated as follows:

138

s = vw / vp = vw / (vair + vw) (4.40)
where s is the decimal degree of saturation. The degree of
saturation can also be calculated by

s = l - (fa / f) (4.41)
Equation 4.37 will be used in this study.

4.11: Windrow Size Changes

 

Windrow size and shape changes were measured by two
methods. The main observations were made by measuring the
distance from the windrow surface upward to a grid of
equally spaced co—planar wires stretched horizontally over
the windrow. Before each set of measurements was taken, the
height of the wire at the center was standardized by
adjusting wire tension. The distance between the grid and
windrow surface was determined by holding a plumb bob at the
grid location and measuring the distance between the wire.
and the point on the windrow surface that the plumb bob
touched. Thirty or 36 locations were measured in this
manner on the lst, 3rd, 5th and 7th days after the windrow
was turned. The distance from the top of each probe to the
top of the surface next to the probe was also measured. All
probes touched the floor. Measurement error was 1 1/8 inch.
Only one side and the top of the windrow were measured by

this method.

CHAPTER 5

EXPERIMENTAL RESULTS

5.1: General Observations

 

Two windrows were monitored: one in the summer and one
in the fall of 1986. Windrow turnings are designated A, B,
and C. Ambient conditions during the summer windrow
(Windrow 1) were generally warm and humid. The fall windrow
(Windrow 2) was preceded by cold and wet weather. Data on
ambient weather conditions are presented in Figure 5.1.
Ambient temperatures rose during the first week of Windrow 2
but the overall weather was colder and rainier than for
Windrow 1. Due to difficulties in correctly programming the
relative humidity sensor, one third of these data were not
collected.
5.2: Windrow Size and Shape

Windrow size and shape varies between windrows and
turnings. Figure 5.2 shows the cross-sections of Windrows
1A and 2A. The changes between windrows were due to the
amount of material in the windrow: Windrow 2 had half again
as much material as Windrow l. A summary of windrow weight
and sizes is given in Table 5.1. Differences that occur
when a windrow is turned are due to difficulties in

constructing successive windrows to exactly the same shape

139

.140

.mcowumpcoo umzumm3 ucmeE< "H.m muoomm

.483 2.:

ON or Nr 0 Q

0

 

 

D F D b D I P b D h D h .r hi i I b b b h

F m<qm

‘l“‘\‘

E
S

 

0

low

ION

IO”

101

TO“

[00

10K.

IOU

10¢

 

 

anniexadmai

H8 3 (3.)

(t)

.msmov mete

ON up Np O

 

 

L. 9.2

43,;

F m<qm

:1)(i(\(\\1\1.)()( \e\J\(I

E? g

 

A 30m02~3

100

.100

 

00p

ainaejadmaj

Ht! '3 (3.)

(i)

141

 

120
110- 9
'w- ./2A

90‘ }////

 

E a
U I
03- ./ 1A
70.
53-

Windrow Height

 

 

 

 

Windrow Base (cm)

Figure 5.2: Cross-sections of Windrows 1A and 2A.

Table 5.1: Summary of windrow weight and size.

 

 

 

Windrow Weight (kg) Areaa Volumea Bulk Density8
Initial Final (m2) (m3) (kg/m3)
1A 5440 3320 3.32 15.2 360
18 3320 3650 1.82 7.57 440
IC 3650 na 1.82 9.23 395
2A 7500b 6300 2.82 12.9 580
2B 6300 5740 2.37 10.8 580
2C 5740 4900- 2.44 11.1 520
5300C
3 Calculated for initial conditions.

Total amount weighed = 8860 kg.
kg not used.

Rain occurred during weighing.
amount of wetter material added.

Aproximately 1360

Difference reflects

142
and size and to material lost in the composting process.

Windrow size changes during each of the turning
periods. Figure 5.3 is a topographic representation of the
front half of Windrow 1B. The windrow is quite uniform with
a slight rounding at the edges. Changes in Windrow 1B after
1 week of composting are shown in Figure 5.4. A cross-
sectional view of average values for the same time period is
given in Figure 5.5. The center height decreases during
each windrow turning, in this case by about 2 to 3 inches.
The height of the upper part of the slope decreased slightly
while the lower face height increases slightly. These
changes are as much as 10 percent, but are typically 2 to 3
percent of windrow height at a given location. Windrow
elevation data availability are discussed in Appendix H.

The location of monitoring points varies between
windrows and with the parameter being monitored. All
temperature/gas sampling probes were inserted into the
windrow until they touched the compost bin floor. Since
windrow size and shape varies between windrows and turnings,
the distance between monitoring points and the windrow
surface also varies. This is true for both the distance
betweeen the monitoring point and surface (Figure 5.6)
measured along the probe and at right angles to the surface.
See Appendix H for information on the availability of data
on the location of the monitoring points. Due to initial
windrow sizes and changes in windrow size during the

composting period, sampling points could be exposed and give

143

.mcomum>maw
mam: acouu Hmwumcw ma aoupcwz "m.m ousmwm

A50 .x. acoum soupcmz

0.09. 0.094 0.0mm. 0.00m. 0.0mm Odom

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

II?)

”.1111
\l i
H\\\mflHHHHHH
HMWHHHHHHH
k\\\HMHHWIM p IJHIIIIF
0.5 7.0530 30.3353 II“

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I (M

 

pug MOJPUIM

(m3 'A)

144

.x003 wco 000mm

momcmno cowum>mam mam: ucoum ma soups“: «¢.m wpsmwm

.60 .x. accum 3000:“:

oodmv 00.00». 00.09.. 00.00». 00.0mm oodom Godm— oo.oo_. oodm

 

7 — a _ — J. _ q 1 A . (1‘4 4 _

      

   

 

 

 

_ P _

L

      
 

oodm

00.00—

 

 

05 23:00 30.853 H

pus MOJPUIM

‘3)

(mo

145

 

(cm)
8
I
ma

Windrow Height

 

 

 

I I l I I

I
140 1673 1:13 203

I I

o I I I I I I I I I

I
O 20 40 60 80 103

Windrow Base (cm)
Day 1 Day 3 Day 5 Day 7

I
120

Figure 5.5: Cross section of Windrow lB elevation
changes.

 

Op

“‘4

 

Cross section of windrow showing
distances from a monitoring point to
windrow surface along the probe (Dp)
and the nearest surface point (Ds).

Figure 5.6:

146
temperature or gas concentration readings that reflected
ambient conditions.
.2111 Temperature
5.3.1: Temperature Variablity

A key assumption of the experiment is that the windrow
is long enough so that transport processes along the length
of the windrow can be neglected. While this assumption was
not directly tested, data from Windrow 2 can be used to
assess whether the three replications of the 9 locations
were measuring the same phenomena. Because of the small
sample size (1 for each turning) it is hard to draw
conclusions about variability in temperature by location.

The variance of temperature between the three
replications of each location is calculated for each 15
minute average temperature for the three turnings of Windrow
2 and are shown in Figure 5.7.

Windrows 2A and 2C exhibited average variances within
the experimental measurement variance of 1 1.5°C with the
exception of 1 point in each windrow. Windrow 2B exhibits
rather large average variances, with 6 of the 9 points
having variances larger than the variance due to instrument
error.

An analysis of the patterns of variability proved
instructive. If there is a longitudinal effect, one would
expect the left (L) and right (R) replications to be closer
together than L and the center (C) or R and C. Longitudinal

effects are assumed to be present if the difference between

Hindrov 24

57.6 50.0 43.1
47.6 46.1 35.6 32.3

36.6 56.5 I

0 31
0.99 0.05 0.23
0.70 2.85 0.32 0.08

 

 

1.43 0.47 1.09
1.15 3.41 1.45 0.84»

 

u v
u u v
u n u N f

 

Figure 5.7:

1147'

 

 

 

 

 

 

 

Uindrov 28 Hindrov ZC
51.7 57.3 58.0 57.3
52.5 62.2 52.5 57.1 60.9 52.5
39.6 58.3 49.8 42.8 12.1 57.9 49.6 12.8
Average Ieaperature (’C)
1.66 0.19 .55 0.74
3.51 2.96 2.17 . 6 0.17 0.12
1.10 0.27 2.97 0.83» 7.15 1.03 0.78 0.70
Variance (’C)
2.03 0.74 1.26 0.33
3.63 2.18 2.59 1.68 .68 0.93
1.79 0.63 3.42 2"1. 4.99 1.55 1.69 I.72_
I line Sale Sign
Y N Y Y
Y N N N Y Y
n v n v f u u u v f

 

 

Sililar Shaped Curve

Measurements of temperature
variability in Windrow 2.

148
C & L and C & R showed the same time-temperature pattern and
the percent of time that they were the same sign. Figure
5.7 shows these results.

Windrow 2A is low by both measurements, with same sign
differences present at only 46 percent of the time and only
2 of 9 locations having similar patterns. Windrow 23 has 4
locations with similar patterns and the differences have the
same sign 79 percent of the time. Windrow 2C has the
highest number of locations with similar patterns; 72
percent of the time the differences have the same sign. No
pattern is obvious: Windrow 23 has high values of this
parameter and variance while 2C has slightly higher values
but low variance. This is also true for the number of
locations where temperature differences were judged to have
the same shape.

Based on Windrow 2A and 2C, it appears that the
longitudinal variability is negligible. Due to higher
variability in Windrow 28 and the effect of ambient
conditions on the composting process, however, this
conclusion must be tested further. In addition to
measurement errors, there is also the effect of the probe
placement to be considered. While each probe was placed so
that the locations were approximately the same in the y-z
plane, it was not always possible to achieve this. Thus,
differences in y-z plane location could affect the measured
temperature variability. It is also possible that the

higher variance in temperature is due to the presence of

149
different transport processes in the windrow, i.e. natural
convection as opposed to diffusion. Nonetheless, for
purposes of analysis, the replications will be considered to
be averageable.

5.3.2: Temperature Patterns

 

The maximum temperature in either windrow is 68°C;
most of the windrow tend to be in the 45°C to 55°C range.
Temperatures near 35°C are seen deep in the center and at
the outer edge of the windrow early and late in the runs,
respectively.

Temperature profiles are shown in Figures 5.8 through
5.10, for Windrow l. Temperatures in locations near the
surface, 1, 2, 4, and 8, are initally high and either rose
to a near constant level (Location 8 in Windrow 1A, Figure
5.10) or peak and fall off (Locations 1, 2, and 3 in Windrow
18, Figure 5.8). Towards the end of a given turning, the
surface temperatures tend to reflect ambient conditions.
This could be due either to an actual reduction in
respiration in those areas of the windrow or in the case of
locations 8 and 4 (Figures 5.10 and 5.9), due to the
thermocouples becoming exposed late in the runs.

Locations deeper in the windrow such as 3, 5 and
sometimes 9, show either the same pattern as the the surface
temperatures, i.e. a rise, plateau and slight fall
(Locations 3 or 5 in Windrow 1C), or develop in an "S" type
pattern (Location 3 in Windrow 2A). Locations 6, 9, and 10

tend to exhibit mainly the "S" type temperature profile

(°C)
4

Temperature

Figure 5.8:

(°C)

Temperature

Figure 5.9:

70

60

'so

4O

.90

2O

70

00

40

20

150

 

 

 

 

- 3
3
1 2 2
‘ 1

2 1 1

- 3 2
AME T

o r” 4 ‘ a . 1'; V If to 20

Time (days)

Temperature profiles for Locations 1,
2, and 3, Windrow l.

0 4 ll 13 16 1‘0

 

l

 

 

 

Time (days)

Temperature profiles for Locations 4,
5, and 6, Windrow l.

151

70

 

 

 

 

 

 

9
co -
A 8 10
a? 50 "' 9
0 SLAB
65 w _ 10
g SLAB
8
“ 5° + ‘NW/VW
s
0
{-
20 -
MB '1‘
IO - '
o r V I I I Y 1' I I V I V fit I r V T r I
O 2 4 B B 10 ‘ 12 14 16 1B 20
Time (days)

Figure 5.10: Temperature profiles for Locations 8,
9, 10 and Slab, Windrow 1.

(Location 6 in Windrow 1A).

The temperature profile in the windrow at any given
time can yield important information on the locations in the
windrow that reach a given temperature. Cross-sections of
Windrow 1A taken at noon on days 1, 3, 5, and 7 are shown in
Figures 5.11 through 5.14 for illustration. Figure 5.15
presents the same information for day 7 of Windrow 2C.

Since temperature was not measured on the windrow surface or
in the front of the toe area and since the thermal
properties of the compost and air are different, assumptions
about the temperature in those areas had to be made. The
five Figures (5.11 to 5.15) are good for illustrative

purposes only and do not reflect the actual temperatures at

152

on.

.cooc up A woo co cowuummummoLo 4H
3oupcwz L0m mmaflmoLd Auov mucumLmQEme

AEU .>v

mmmm 3oLccm3

"Ha.m musmwm

 

 

 

 

 

OO—

3ufiiaH AOqugm

42)

(mo

153

.cooc um m xmo co comuommnmmoLu 4H
zochmz L0u mmawmoua Auov musumeQEmB "NH.m mnsmmm

AEU .r. omen Jauucmz
om— . oo_.
_ .

ovr_|\.1\I/\u

 

   
 

 

 

 

OO—

3qfiian noxpugm

12)

(m3

154

.coo: an m won :0 cowuummummoLu 4H
Boupcwz LCM mwawmoLd Auov musumumdsme "ma.m mpsmmm

AEU .rv ommm zohpcmz

om_ . oo—

 

 

 

 

 

 

 

oo—

iqbgau noapugm

Iz)

(mo

155

 

.h woo co cowuowmummouu 4H 3ochfiz
L0u mmamuoLQ Auov muaumumaEmu Hmcwm ӢH.m wuzmflm

.Eu .r.
oop

q

 

 

 

 

wmmm zoLocmz

 

 

oo—

JQEEBH MOqugM

42)

(mo

156

.UN 3ocpcfiz
L0H mHMLOLQ Auov mnsumhwdsmu Hmcwm

AEu .>V amen Jaupcmz

Omp . COP

"mH.m musmmm

 

 

 

 

 

oo—

aubxaH MOJPUIM

(m3 '2)

157
the toe and surface.

In general, the temperature profile of a given windrow
tends to have the highest temperatures at some distance from
the surface, with slightly lower temperatures at the surface
and much lower temperatures deep in the interior of the
windrow. With time, the zone of higher temperatures moves
inward and the edges cool.

5.3.3: Time-Temperature Relationships

A key parameter in predicting the thermal death of
pathogenic micro-organisms is the length of time a microbial
population remains at lethal temperatures. Table 5.2
summarizes the time-temperature data for locations monitored
in Windrow 1. Similar data for Windrow 2 are presented in
Table 5.3. Figures 5.16 through 5.21 show the length of
time that a given location remains in a given temperature
range for Windrows l and 2.

5.4: Gas Concentrations

 

Concentrations of oxygen and carbon dioxide were
initially calculated using the method described in Chapter
4, unmodified. Examination of the data, however, indicates
higher oxygen concentrations than expected in the interior
of the windrow (Location 10, Figure 5.22). Under forced air
composting, oxygen depletion is very rapid and quickly falls
to levels of l or 2 percent. While the DMS composting
system is not expected to produce such a quick reduction in
oxygen concentration, because of the length of time between

turnings, oxygen levels are expected to be in the same

Table 5.2:

158

Time-temperature data for monitoring locations
in Windrow l.

 

 

Length of Time (days)
Location Temperature Range (°C)
30-35 35-40 40-45 45-50 50-55 55-60 60-65 65-70
Windrow 15 Total Time: 6.96 days
1 0.35 1.74 2.62 2.25
2 0.49 0.47 1.57 4.43
3 0. 88 0.77 1.61 3.71
4 3.15 2.56 1. 24 -
5 0. 65 0.44 0.70 1.12 4.09
6 1.74 l. 34 1.42 2.46
8 0.52 0.46 1.01 4.97
9 0.89 0.81 1.40 3.86
10 3.18 2.48 1.30
Windrow 12 Total Time: 6.88 days
1 1.44 2.28 2.70 0.47
2 0.30 0.28 2.99 3.31
3 0.39 1.09 2.38 3.03
4 1.50 2.11 2.27
5 0.65 0.56 0.72 1.33 3.62
6 2. 98 2.04 1.47 0.40
8 0. 35 6. 53
9 0. 35 0.99 2.79 2.75
10 1.69 2.15 2.74 0.30
Windrow 19 Total Time: 6.81 days
1 2.38 1.73 1.22 1.49
2 1.71 2.24 1.88 0.99
3 2.63 1.84 2.34
4 Probe exposed to air.
5 0.24 0.27 0.62 5.68
6 1.34 0.74 1.00 3.73
8 0.22 6.59
9 0.72 0.49 4.81 0.80
10 0.20 1.77 0.88 1.13 2.84

 

159

Table 5.3: Time-temperature data for monitoring locations

in Windrow 2.

 

(days)
(°C)

40-45 45-50 50-55 55-60 60-65 65-70

Length of Time
Location Temperature Range

25-30 30-35 35-40

 

Windrow 25 Total Time: 5.80 days
1 0.19 0.68 3.27 1.67
2 0.08 0.45 0.73 1.92 2.62
3 0.98 1.00 0.71 0.52 1.21 1.38
4 0.10 0.42 0.58 4.70
5 0.01 0.86 0.74 0.86 1.29 2.03
6 1.26 1.51 1.45 1.58
8 0.09 0.59 0.66 4.46
9 0.52 1.33 1.35 2.28 0.31
10 1.13 3.38 1.29
25-30 30-35 35-40 40-45 45-50 50-55 55-60 60-65 65-70
Windrow 2g Total Time: 6.842 days
1 0.01 5.12 1.20 0.51
2 0.08 0.06 1.83 2.22 2.00 0.65
3 0.26 0.15 0.12 2.99 3.32
4 0.06 0.18 2.79 3.72
5 0.14 0.56 0.67 1.06 1.01 3.41
6 0.06 2.29 1.28 1.67 1.54
8 0.21 0.32 0.59 5.71
9 0.15 1.34 1.18 1.51 2.38
10 2.36 2.85 1.62
25-30 30—35 35-40 40-45 45-50 50-55 55—60 60-65 65-70
Windrow 29 Total Time: 6.88 days
1 2.38 2.66 1.83
2 0.19 1.56 3.74 1.46
3 0.26 0.31 0.40 2.89 2.84
4 0.28 0.52 6.07
5 0.54 0.55 0.57 0.90 1.34 2.97
6 2.26 1.40 1.56 1.66
8 0.29 0.55 4.72 1.31
9 0.11 1.84 1.08 1.96 1.88
10 2.15 3.85 0.88

 

160

.mwmp mm.m mm3 zochL3 Lou memo
Hmuoe .4H 30chw3 cw comumuoH up
mmmau wusumuwdswu comm cm ucmdm mafia

"mH.m mulmwm

33333993

 

on

 

 

 

 

 

 

 

L a...333399331U 3333399330
fl 1. I
a .n .n
I In 6
rt 1' 1?
xn 3 J.
CO CO
3.6036 0.2533,.» npooflro
2833399336 E333399.33.a .K33339339ma
v— r. rp
.u .n Tn
fin 3 I
1' 1' iv
6 J. J
a 8:83 r o c883 6 n 8:83 5
2333399331. 3133339930?
.. f
Tn E
r». Yn
I. I.
J. J.
30 CO
nanoflro opooflf

 

 

«UV 59:0: «005 839.065»

Ounce-339090...

_ 8.88..

0....
IO

.—
1"

I
(“09) "no 0! own

 

.mxmp mm.m mm3 aoupcwa Lou mew»
Hmuoe .mH soups“: cm cowumuofl sh
mmmHu mLDumLmQEmu comm cm ucwdm mELB “5H.m munmwm

ADV 59:01 :00 23960:...»

 

 

 

 

or. no Do on On 01 Ov an o». or. Go 0.0 on ad 0v 0. R 9.0 we. no Do on 00 av at on one P. no on on on 9 9v 0.» o...
x IIJHI—I_ .414. .4dIIIIIII.
f. 1. T3
5 a .n

x A I” .1

T l I .1

I 1.. I. .r.

o. c2603 r 0 3.083 r n .3803 6 u covuva... r.
onnoOonaOonvgnnOn £30033anan RaconnOnoJOvnnoao

r1“ 1—14—1-1-16 .1-1-1116 I
1 l v.

#3
r3

 

 

a c383 fl n 8.503 fi n 8283

 

U
r
0

Ragaavgnng Raccoonoanvovnfl

fl!

(5
62

fi
.
1'7
’

JI‘FE

 

 

JJTFK

u 8.603 v 8.503

(1’09) "41) 0! our”

162

.mme Hm.m mm: 3ouncm3 no“ mama
Hmuoe .UH zouccwz cw cowumuoH >9
mmmau muzumquEmu comm cm ucmam meme ”mH.m musmfim

AUV IDS :65 £39509...»

 

 

 

 

 

 

 

 

 

 

 

 

 

2283399291.." P82fi399n.9no R3833??nnomu 838339338
v. _ a. I
a 7.. E
x 3 I
fio 1v 3
I I I l
38803 F 0:383 r ncavnofl 6 .233?) f
or.8333??gnawV RBfl-nmlonimv7220 2383399223
— I l -—_— f
Tu fin E
I in. 1n
I xv ..v
J. I» a
«5.603 f ncovuofl F. «8.603 F
omnwmwnoowgnnomo omfipaflponwownnoao
; r
fn .n
A ,n
T? 1'
I... I».
neon—.03 J .832: 6

 

 

(nip?) "mg 0' own

163

£823

 

OrnoOoL

£893

or c3603

0001000».

a c2503

0001000»

u 8:83

 

FIT}.

..o

'1

 

r53}.

 

.mxmv om.m mm3 30uocw3 no“ mew»

Hmuoe

OhnuOonuOnnqunOmO

 

n Cannon...

 

 

OhndOoonooflvannoh

v Con-duo...

 

.4N 30uvcm3 cm cowumuoH xn
mmmau musumumaamu comm cm ucwam meme

uhndOonnO‘uonvanu

III'|1—|‘—‘l-“-Il.-I|_lvo
w «

 

n. COVUOB

R38389332
-o

7
'-

N

 

FT};

u c333

"mH.m whammm

“UV :5 :05 83806:...»

anooflnonvoendnx.

  

. Conn X1.

1‘

J .- 3‘7 ’71"

(o’x-p) ON") (1' an”

164

07.833

39‘Ovnn

 

R3333?

2883

o— 8.503

«£383

009.an

 

Gonna,»

3.1”

 

J—JJI

On

 

u 60:83

O

 

JW‘v’stT.

.mmmc «m.m mm3 sauce“; new memo
Hmuoe .mm zopucwz cm cowumooH >9

mmmao monummmasmu comm cm ucmam memb

ObnuOonnbnninIflnOn

II: M

 

o Covuufl

283nu0001730n

 

 

n c0983 m

Ragnngflvvn

v COPBQJ

 

é a 6 A a I 63

Ragnngvgnngo

n c383

"om.m mmammm

1
Tu
vn
J

In.

 

rm

oundooanuanrvflnOn

_ — — ..p

N CODOOS

0

fl

 

F313

2

A3

«0003

gram «8.5 Bancroft.»

3300308.

 

” cone X}.

 

fl
_
r.

H...
U.
N
W
3
N.
O
I
>
w.
(n
(It.

165

Bacon...

0009092.

0. Cop—woo..—

Ragoaonnvgnn

4.314

2923

on

a £2503

01093

a $0.603

64.33

mmmau musumuwQEmu comm cm ucmam meme "Hm.m munmmm

 

9»

H

 

02

I
.-

 

6W1}:

.mamc mm.m mm: 30uccmz pew mama
HmuOB .Um zopocwz cw cowumuoH an

RagnuRnVOOnnOm

I:

n covuofi

O

vw

F.

 

VIII

RnooonnonnvOoflan

—_..

n copuos

 

RQSnnoonvowonon

>

v Spoon.—

 

Razongflvn‘nn

n. Congo...—

 

anoonnquvgnng

‘

N C0303

 

ADV :5 .905 Kain...»

fl 3 2 3

399304

 

 

_ €09.3sz

i
f!

 

JIIA

(“99) "1‘3 u' uugl

166

20
19 0

18 a

17 a 8
1B-
151
14-1
13-
:3“ WINDROW 1, LOCATION 10
10-
9-4
a-
7.1
81:
5d
4-1 * 4.
3‘ + +
2-4
1%

 

on
a
D
a
+4»
an

%)

(

1+?
+
(II

Gas Concentration
+
++

 

 

-D

 

I
1 3 5 7 9 11 13 15 17 19 21

Sample Date
a C02 + 02

Figure 5.22: Gas concentrations at Location 10
showing suspect high oxygen levels.

range.
The experimental methods used in the gas sampling and
analysis were re-examined as described in Chapter 4. In the
course of this examination, it was found that ambient gases
could have entered the sampling syringe. This was probably
due to gas flowing into the syringe needle because of either
diffusion or a pressure differential between the syringe
body and the ambient air. The effect of calculated internal
sample bottle pressure on the percent influx into the
syringe was studied and a calculation procedure was derived
to correct for this, based on the number of samples that had
been withdrawn from the sample bottle. The effect of
leakage during several weeks' storage time was found to be

negligible, although further work in this area is necessary.

167

The difference between original and corrected oxygen data is
shown in Figure 5.23 for selected data.

The gas concentration data are presented in Figures
5.24 through 5.27. These data represent the average values
for each location. Several trends are present in this data.
At any given time, locations near the surface (Locations 1,
2, 4, and 8) have higher oxygen and lower carbon dioxide
concentrations than those in the center (Locations 6, 9, 10,
and sometimes 5 and 3). This is shown in Figure 5.28.

Over time the relationship between carbon dioxide and
oxygen depends on the location. Near the surface carbon
dioxide starts high and decreases over time, while oxygen

concentrations do the inverse (Location 1, Figure 5.29).

 

 

 

 

 

A 2
,9
v x x
.. X x
c ‘ x x
O
... x x
u
2 o 1;x x § x xx “
.9 ..* O A )1 4A
0 W X+ A ‘ 404
c: -‘ " A 5 °
0 ‘4‘ x x. A 0 o
U ‘ o 0
AA“ A o o
1. ‘co °
:1
m ‘ .30
O

I o
. co
1: "‘
o o
a)
u
e -5 I I I I ‘I I r I I I I I 1 Iii
‘5 :5 s 7 9 11 1:3 15 17 19
U .

Measured Concentration (2)

+ Low 0 High 11 Average x C02 Correct

Figure 5.23: Difference between original and
corrected oxygen concentration for
selected data.

168

.H saunas:

"¢N.m mLomwm

 

 

 

 

 

 

 

 

 

 

 

Lo“ womb comumLucmucoo cmmxxo
- so.” sm.o _~.a_ o~._~ . Kn.m_ ~_.o_ a_.o~ a_._~ _o..- _~._ on.». ~m.¢_
...w .a... -.o~ m_.m_ o..o_ oa.o~ n..o- Ls... o~.m_
ao.o~ mm._~ mm.¢_
no._ on.“ .n.~_ vn.m_ - .m.n o~.. ~m.. .n.¢_ ,mm.~ n~._ n~.~ vs."—
«v.. ~n.u ma.¢_ oo.~ a~.n ~n.m_ mo.n g..~ om.m_
an... on.n_ mm.~_ ...n. n..°_ ~_.m.
- a~.~ as." «n.~ No.”— -o_.v um." nn.o_ .a.w_ .~_.n o_.. av.” mm.o_
a~._ «..n n_.¢_ no.” .m.n mo.c_ oo.o. Ln.n we...
n«.o_ aw.e_ aa.o ca.n_ aw.” Lo.~_
h gag m an: n >a=

l ox.~ an.~ w~.~ ~¢.m_
an._ n_.. m..._
o_.¢_ om.a_
- a¢.. mm.” .o.m an.”
...N .~.~ ~_.o
oa.a .~.n_
.a.¢ «1.. an.. em.”
m..a ma.. a~.°L
.°.a _~.¢.
_ sac

169

Log womb

 

.m soaocwz

cowompucmocou comwxo "mm.m mLommm

 

 

 

 

 

 

 

.9." a... we... ...oa 1.x... ...a ._.__ _n.._ ..N._ ._.N .... ”a... - ...N ..._ an.” no...
no.“ a... _..._ n..~ mu.” .m.._ a... x... mm.m_ a... a.._ ...o.
a... mm... a... a... ma.. .n.. a... m...
-..._ ...o .n... ...ou (a... ...m .N... ...._ -m... ...n m... xx... -..._ .~._ N..~ m..m_
"a.“ a... a..._ °~.~ N... 5°... ¢_.n ...N 5.... ~_.N wa.~ ...o.
a... a... .0.» x..n_ m... a... ~_.~ a...
.n._ a... m... ...m. ...n a... a... _..._
...o n... N... N... ~_.~ ...o
9..“ s... .o.~ a...
5 us: n ».a n >~= a xaa

170

MUMU COM umhuchCOU

.H 3oLoc.3 Lo.
wowaMU coanU

.mm.m muommm

 

 

 

 

-2... a... .N.. 2.... 3.. .... 8.. ...o .3. .2. .... .... ..... ....N ..N. ....
..... ...o. 8.. .... o... .3. s... .... .... N... a... 2..
.3. .... 3.. 2.. a...
3... ...N. .... ..N -..... 8... .N... 3.. ,3... ..... 2... .... -2... ...2 ..... ...2
..... 2... o... ..... z... .... s... 2.... N... s... ..... 3....
2.. .... ...2 .... .N... .... s... N...
-...... ...2 ..N. N... ...o. 2.2 2.... .... -..... ....N NE. .3. -..... ..... ..... 2...
..... ..... N... .N..N ...2 3.. ...N. ..... 9... .9... ...2 N...
s... 2.. a... 2... ..... o... ...2 ....

s an: n ..a n ..a . ..a

1'7].

.N 30.U:.3 no.
mumn coquHucmucou momxowo conumu unm.m whammm

 

 

 

~°._~ on... ow.m vn.o n... mm.a_ aw.m nw.o on..~ sv._~ .m.¢_ .a.o v_.n~ ow.m~ vo.o~ mn.~
mn._~ -.o_ mo.m ~_._~ gs... .o.n ~a.o~ o~.a_ o~.n ao.- ~v.- no.0—
~n.__ «m... .o.m_ ao.o_ e“... am.n~ cm.a_ ~m.o_

 

 

 

 

- on..~ ow.o~ mm.m mm.o . oo.o~ av.m. .m.~ oo.ou .m._~ on... _~.~_ a... v¢.m~ mo.m~ mo.o~ .5.—
mo... oa.~_ _~.~ an... an... .m._ o~._~ m~.a. vo.c v«.v~ vm.- vo.o_
.a.n_ .c.o~ no... on.v av.a_ ~a.~_ v~.v~ on.o_

 

 

. .~.- ~o.- ov.o~ so." so._~ aw.n~ -._~ nw..
mm..~ ~n..~ av... n~._~ ao..~ .N..—
~o.o~ ow... .o.o~ .n.n.

. ..a n ..o n ..a . ..a

172

22
21— E
20- +

19- U

‘8- U
17- a D
15' a

15' 4'
14-
13-4

12-1 + I
11-
1O‘
9" D

B- B

 

(%)
+4-
DD
4.
on
o

040

0+

3.
s—
4..

Gas Concentration

+4”!-
4.

2-
1.. +

 

 

 

T I

O 2 4

ca +4

Windrow Location
0 C02 4 02

Figure 5.28: Gas concentrations by location,
Windrow 13, Day 3.

 

 

 

 

22 4»
+ J
20- . + ;
¢
A ‘84 '. + 0 1
” 8
‘3‘ o + '0 + o
c O
o 14- °
3 a , WINDROW 1, LOCATION l
m 12.: O
L. +
u
5 10-4 9
U +
C
o B- D D
U 4-
m 3" "'
8 B D + D
+
_. U
‘ .a a
E
2...
a a a T
D I:
O T r I ‘ 1 .1 I r I I I I I I Y I I T I 1 1
O 2 4 8 B 10 12 14 15 1B 20

Sample Date
a C02 + 02

Figure 5.29: Gas concentrations over time, Location
1. Lines indicate general trends in
data, not intermediate values.

173
There is a partial recovery towards the initial
concentration after the windrow is turned but the pattern
then repeats itself. Locations in the middle zone (3 and 5,
Figure 5.30) Show much longer durations of high carbon
dioxide or low oxygen than those near the surface.
Locations 6, 9 and 10 consistently show high carbon dioxide
and low oxygen for the duration of the composting (Figure
5.31).

Some gas samples from Locations 9 and 10 taken during
Windrows 23 and 2C were analyzed for methane concentration.
Low concentrations of between 0.5 % and 1.5 % methane were
found. This result suggests that future studies include
methane analysis as a means of more accurately determining
the windrow reaction and transport processes.

5.5: Moisture Content

 

 

Drying of the windrow surfaces was evident within a few
hours after windrow formation. The compost material turned
light brown, almost white, as it dried. The greatest
penetration of the dry layer was only 0.5 to 1.0 cm,
however. Blooms of small mushrooms appeared between 3 to 5
days after each windrow turning.

Figures 5.32 and 5.33 present the wet basis moisture
contents (% wb) by location for Windrows l and 2,
respectively. Figure 5.34 shows the average moisture
content by location for each turning.

Moisture content remains essentially constant

over the 3 week duration of each windrow, decreasing from

8)

Gas Concentration

Figure 5.30:

(t)

Gas Concentration

Figure 5.31:

22

174

 

20-
18-
18-
14.-
12-
10-

a—

 

O WINDROW 1,

LOCATION 3

+
'f-\
+

 

 

 

3.
data,

Gas concentrations over time,

12
Sample Date ,
a C02 +-02

Location

Lines indicate general trends in

not intermediate values.

 

 

 

 

22 0 D
D O
20- ;///5‘\\\\ ////o"—__““a B
D
0 a
a
18- a c: g \\\\ 1
18- fl
0
‘4‘ waoRow 1, LOCATION 10
‘2-
10-1
8-
a- * -
\+ + "
A- + + '0 \‘° * /.,
\ + + }
2.. Y\____*
+ +
o I I I l I I I l I l I Y I I I I I Y I I
O 2 4 8 8 1O 12 14 18 18 20
Sample Date
cacoz . 02
Gas concentrations over time, Location

10.
data,

Lines indicate general trends in
not intermediate values.

175

 

 

 

.... .... .... ....
.... .... ....
....

-.... .... .... ....11
.... .... .... ....
.... ....

...... .... .... ....
.... .... .... ....
.... o...

. ...

.H aochMz Lo. :oMumooM xb
..bzw. .cmucou muoumMoe uMLumEM>mLU

 

 

 

.... .... .... ....
.... .... ....
.... ....
g. so..=.=
-.... .... .... o...
.... ..... .... ....
.... ....
2 305.2:
.... .... .... ....
o... .... .... o...
.... ....
.. .......
. ...

.mm.m musmMm

 

 

 

,o... .... .... ....
.... .... ....

.... ....

.... .... o... ....
.... .... .... ....
.... ....

-..o. .... .... ....
.... .... ....

o... ....

. ...

176

.N aoupcMz Lo. coMumqu xn
.n3w. acmucou muoumMoe uMLumEM>mLu .mm.m mLDmMm

 

 

 

 

 

 

 

 

 

-.... .... .... .... -.... .... .... .... .... .... .... ....
.... .... .... .... .... .... .... .... ....
.... .... .... .... .... ....
.. .......
-.... .... .... .... -.... o... .... .... .... .... .... ....
.... .... .... .... .... .... .... .... ....
.... .... .... .... .... ....
.. .......
o... .... .... .... ..... .... .... .... .... .... N...
.... .... .... .... .... .... .... .... ....
.... .... .... .... .... ....
.. .......

. ... . ... . ...

 

 

 

177

 

 

 

Hindrov 1A Hindrov 2A
80.4 79.6 81.6 79.5
78.8 79.9 79.7 79.1 81.1 79.4 77.1
80.4 79.7 79.4 79.6_ 53.5 85.2 78.6 76.3
Hindrov 18 Hindrov 28
77.9 78.1 78.2 77.4
77.8 78.1 77.8 78.0 78.3 77.0 78.8
78.9 78.2 77.9 78.4 78.1 77.5 77 3 77.9
Hindrov It Hindrov ZC
51.5 77.0 77.9

78.1 77.4 77.3
77.4 77.4 77.9 77.9

77.5 78.0 77.8

77.6
77.3 77.6 77.8 77.9

 

Figure 5.34: Average gravimetric moisture content

(%wb) by location and turning for
Windrows l and 2.

only from 81 to 77 % in both windrows. Average standard
deviations of the measurements are 2.2 percent. Moisture
content sampled after day l of each turning tends to have
the smallest differences between locations. Samples taken
on days 3 and 7 showed a greater range of moisture content.
Locations near the surface and base (1 and 2) tended to be

the driest in Windrow 1. This pattern is not present in

Windrow 2, perhaps due to the accidental flooding with rain.

5.6: Volatile Solids
Volatile solids measurements show small decreases
within each one—week turning and over the entire three week
composting period for each windrow. Figure 5.35 presents
the average by location for each turning in Windrows l and

2. Figures 5.36 and 5.37 show the volatile solids content

 

 

 

Hindrov 1A
94 95
94 95 95 96
95 95 96 96_
Hindrov 18
94 95
94 94 95 94
94 95 95 64
windrow 1C
62 94
93 94 94
93 93 94 94

 

 

Figure 5.35:

1778

 

 

 

 

Hindrov 2A
89 85
87 85 82
84 86 86 84,
Hindrov 28
. 83 86
83 85 86
84 85 87 86 _
windrow 2C
82 82
82 83 84
83 83 84 84

 

 

Average volatile solids content (%db)
by location and turning for windrows

l and 2.

 

 

179

 

 

 

.. .. .. ..
.. .. ..
..

. .. m. .. ..1.
.. .. .. ..
.. ..

.. .. .. ..
.. .. .. ..
.. ..

n»:

.M zoupcMz LCM coMumqu
.n .no.. .cmucou we..om m...m.o>

 

 

 

.. .. .. ..
.. .. ..
.. ..

.. .......
.. .. .. ..
.. .. .. ..
.. ..

.. .......

, .. .. .. ..
.. .. .. ..
.. ..

.. .......
n .2.

"mm.m musoMm

 

 

 

.. .. .. ..
.. .. ..
.. ..
.. .. .. ..
.. .. .. ..
.. ..

l .. .. .. ..
.. .. ..
.. ..

. ...

180

...m aoupch Lo. coMumqu

 

>2 .npw. ucmucou mpMMom mMMumMo> .mm.m mLDmMm

 

 

 

 

 

 

 

 

- .. .. o. .. -.. .. .. .. -.. .. .. ..
.. .. .. .. .. .. .. .. ..
.. .. .. .. .. ..
.. ....._.
. .. .. .. .. -.. .. .. .. .. .. .. ..
.. .. .. .. .. .. .. .. ..
.. .. .. .. .. ..
.. ....._.
. .. ... .. .. . -.. .. .. .. .. .. .. ..
o. .. .. .. .. .. .. .. ..
.. .. .. .. .. ..
.. .......
. ... . ... . ...

181
by location for Windrows l and 2, respectively.

The magnitude of the decreases are 2 % and 4 % VS for
Windrows l and 2, respectively. Due to the volatile solids
standard deviation of 7 % VS, however, this is not
significant. Differences between the maximum and minimum
values of volatile solids in each sampling period increase
from day l to day 7, but due to the lack of replication and
relatively high measurement error, these values are not
significant.

5.7: Wet Bulk Density

 

There are several difficulties with sampling for bulk
density and comparing the results. Windrow bulk density can
have large inherent variation. Too few samples can lead to
results that reflect this variability rather than the
underlying bulk densities. Taking too many samples can
affect the transport processes in the windrow.

Direct comparison of bulk densities is hampered because
successive samples are rarely taken from exactly the same
depths. Therefore, even if the variation of bulk density
with depth is the same, differences in sampling depth lead
to apparently different bulk densities for locations that
are nominally the same. The data are best compared when
used in conjunction with a model of bulk density. This has
not been done, so the comparisons are somewhat crude.

Initial experimentation to evaluate the method used to
sample for bulk density indicated that the coefficient of

variation is approximately 11 percent (See Chapter 4).

182

Figure 5.38 presents bulk densities by location for
Windrow 1. Similar data from Windrow 2 are shown in Figure
5.39. The availability of data on the position of each
sampling location in the windrow is given in Appendix H.

The average bulk density by location in each windrow turning
is shown in Figure 5.40.

Comparison of the bulk densities for Windrow 18 with
the overall bulk density calculated from the weight of the
windrow and its measured dimensions provides another check
on the bulk density sampling method. The gross bulk density
of 442 kg/m’ compares favorably with the bulk densities
shown in Figure 5.38. Even though the gross bulk density is
greater than the bulk density at most locations, this is not
a problem. The bulk density sampling method is not able to
sample deep in the interior of the windrow where the
greatest bulk densities are located.

Experimental wet bulk densities range from 150 to 250
Kg/m’ near the surface to 300 to 650 Kg/m3 in the center of
the windrow. Windrow bulk density shows the expected trend
of increasing with windrow depth. Only 5 of the 82 sampling
pairs (6 %) do not show this trend. Location 6 exhibited
the greatest incidence of decreasing bulk density with depth
with 3 of the 5 sampling pairs.

Changes in density with respect to time are
inconsistent. Figure 5.41 presents information on whether
bulk density at each location increases from the beginning

to the end of a windrow turning. With one exception,'

183

 

.H zochMs Lo.
coMumqu an Mme\mx. qumch stn umz

 

 

 

 

 

- ... ... ... ...
... ... ...

...

-... ... ... ...
... ... ... ...
... ...

-... ... ... ...
...w ... ... ...
... ...

. ...

 

... ... ... ...
... ... ...

... ...
.. ......3

-... ... ... ...
... ... ... ...
... ...

.. .......

-... N.. ... ...
... ... ... ...
... ...

.. .......
. ...

“mm.m mLDmMm

 

 

 

... ... ... ...
... ... ...
... ..N

-... ... ... ...
... ... ... ...
... ...

.... ... ... ...
... ... ...
... N..

. ...

184

.m soupcMz Lo.

:oMumqu so .mE\mxv qumch stn umz

 

 

 

 

 

 

... ... ... ... -... ... ... ...
... ... ... ... ...
... ... ... ..N
.. .......
- ... ... ... .... ... ... .N.
... ... ... ... ..N ...
... ... ... ...
.. .......
- ... ... ... ... -... ... ...
... ... ... ... ... ...
... ... ... ...
.. .......
. ... n ...

.mm.m musmMm

 

 

 

-... ... ... ...
... ... ...
... ...

, ... ... ... ...
... ... ...
... ...

. ... ... ... ...
..N ... ...
... ...

. ...

 

 

 

 

 

 

.1835

 

 

 

 

 

Hindrov 1A Hindrov 2A
226 179 176 237
168 241 281 286 265 338 285
240 303 442 501 228 260 438 393_.
Hindrov 18 Hindrov 28
170 131 ' 194 195
200 224 291 305 235 293 272
321 187 471 514' 219 315 436 462
Hindrov 18 Hindrov 28
147 206 200 181
214 333 276 239 409 310
214 313 320 511 253 351 375 477
Figure 5.40: Average wet bulk density (kg/m3)

by location and turning for
windrows l and 2.

186
approximately 68 % (s = 6.5 %) of the locations experienced
increases in bulk density over time. Locations 6 and 9
(Figure 5.42) have the lowest incidence of bulk density
increases over time with only 17 % (l of 6) turnings showing
increases.

When the effect of the 11 % coeffient of variation
encountered in preliminary experiments is considered, the
effect of time on windrow bulk density becomes even less
conclusive. When this criterion is applied, decreases over
time greater than 11 % were present as much as 56 % in some
cases and were seen in all windrow turnings. 37 percent of
the locations in each turning show changes in wet bulk
density less than 1 ll % of the coefficient of variation.
Increases in wet bulk density greater than 11 % occur at 36
percent of the locations while decreases in wet bulk
densities greater than 11 % of initial values occur at 27
percent of the locations. Figure 5.43 summarizes these
results.

Until the core sampling method can be better
calibrated, densities should be considered accurate in a
relative as opposed to an absolute sense. Bulk density does
appear to change with depth, but the effect of the process
on bulk density changes with time was not clear with the
method used in this study.
§4§i Porosity

Porosity at each sampling location is calculated using

the results of the bulk density and volatile solids

187

.coMumooM
an mEMu Lm>o mmmcmcu xuMmcmp xMSQ box

 

.N¢.m wusmMm

 

m=o_.~.o.

.. . . . - o. ..
. . . .. .. ..
. w . ... ..

.mEMu Lm>o mmmcmcu xuMmcmp stn .83

 

 

 

 

a.» u a. a.» a ..
> > > > z z > z
z > z » > >
> > r >
.. ....... .. .......
... . .. . ... . ..
z z z z z z > >
x z > z z > >
r z > >
.. ....... .. .......

.o oo—
nm co

.M¢.m musmMm

mo» a .o

 

cw so.ch=

a.» u om

 

Z
>-
>..

(M ao.ucM=

 

188

 

 

 

 

 

 

Hindrov 14 Hindrov 18 Hindrov 18
Y N N N N Y
N N I I Y Y -Y N Y -Y -Y
Y Y -Y N Y Y -Y -Y -Y N -Y -Y
- 10 50 4 40 - 30 40 4 30 - 56 22 1 22
Uindrov 2A Uindrov 28 81ndrcv 2C
8 N Y Y Y -Y
Y N N N N Y N Y -Y
Y N -Y Y» N Y -Y -Y ‘7 N Y Y Y
- 22 56 4 22 - 22 33 4 4S - 22 22 + 56

Figure 5.43: Wet bulk density changes over time
greater than 11 % coefficent of
variation.

calculations according to the method described in Chapter 4.
The porosity measurement is therefore subject to all of the
uncertainties described in the bulk density and volatile
solids section. In addition, the method does not
distinguish between porosity within the compost particles
and that between the compost particles.

Porosities for Windrows l and 2 are given in Figures
5.44 and 5.45, respectively. Average porosities for each
turning are shown in Figure 5.46.

Porosities vary from 0.98 to 0.97 at the surface of the
windrow, to 0.95 to 0.92 in the windrow interior. As
expected, porosity generally decreases with depth. The
accuracy of the porosities cannot be evaluated until a bulk

density model is developed. The effect of time on porosity

189

.H 3ochM3 LoM coMumqu up auMmoLOQ

 

 

 

-..... ..... ..... .....

..... ..... .....

.....

..... ..... ..... .....
..... N.... ..... .....
..... .....

..... ..... ..... .....
..... ..... ..... .....
..... .....

. ...

 

 

 

-..... ..... ..... .....
..... ..... .....

..... .....
.. .......

-..... ..... ..... .....
..... ..... ..... .....
..... .....

.. ....=..

-..... ..... ..... .....
..... ..... ..... .....
..... .....

.. .......

. ...

.¢¢.m mLDmMm

 

 

 

..... ..... ..... .....
..... ..... .....

..... .....

-..... ..... ..... .....
..... ..... ..... .....
..... .....

..... ..... ..... .....
..... ....o .....
..... .....

. ...

190

.N zoLocMz Lo. :oMumqu Mb qumoLom

 

 

 

 

 

 

-..... ..... ..... ..... ..... ..... ..... .....
..... ..... ..... ..... .....
..... ..... ..... .....
.. .......
-..... ..... ‘4 ..... - ..... ..... ..... .....
..... ..... ..... ..... ..... .....
..... ..... ..... .....
.. .......
N.... ..... ..... ..... , ..... ..... .....
..... ..... ..... ..... ..... .....
..... . ..... ..... .....
.. .......
. an: a an:

 

.m¢.m muamMm

 

 

 

-..... ..... ..... .....
....o ..... .....
..... .....

-..... ..... ..... .....
..... ..... .....
..... .....

...... ..... ..... .....
..... ..... .....
..... .....

. ...

 

 

 

 

 

191

 

 

 

81ndrov 1A Hindrov 24

0.972 0.976 0.980 0.970

0.977 0.969 0.964 0.961 0.969 0.959 0.960

0.969 0.961 0.941 0.933 , 0.973 0.967 0.942 0.944
windrow 18 Hindtov 28

0.976 0.974 0.974 0.973

0.971 0.973 0.959 0.956 0.968 0.959 0.966

0.956 0.959 0.933 0.928 _0.971 0.956 0.941 0.938
Hindrov 18 Hindrov 28

0.645 0.969 0.974 0.975

0.969 0.953 0.960 0.967 0.944 0.958

0.969 0.955 0.955 0.928 0.965 0.951 0.937 0.936

 

 

 

 

Figure 5.46: Average porosity by location and

turning for Windrows l and 2.
is similar to that for bulk density. Five of the six
turnings had an average of 70 percent of their locations
showing decreases over the 7 day compost turning period.
Figure 5.47 presents this information. Porosity changes
over time by location are shown in Figure 5.48.
§;21 Free Air Space

Free air space calculations are made using the results
of the wet bulk density, moisture content and volatile
solids content for each location as discussed in Chapter 4.
Figures 5.49 and 5.50 give the free air space for each
location in Windrows l and 2, respectively. Figure 5.51
presents the average free air space values by location for

each windrow turning.

Free air spaces of 0.87 to 0.71 are found near the

192

.coMumooM
mo mEMu Lm>o mwmcmcu xuMmoLom .ma.m muomMm

 

 

.8333
, o. . . . - o. o. .. ...
. . . .. .. o. .. ..
. . .. ...

.mEMu Lm>o mmmcmcu muMmoLom ....m mLDmMm

 

 

 

 

 

 

... . .. ... . .. ... . ..

. . . z z . z . z z

x . . . . . . . z .
. . . . .

.. ....... .. ....... .. .......

... . .N ... . .. ... . o.

a z . - z z . . - . z .

z a . . z . . . . .

2 3.3M: 3 3.2:: 5 3.2.5

193

 

-__~.°

.H aouncfiz
pew cofiumuoH an mumam umm mmum

 

 

 

 

cnh.o ¢m~.° ”No.0
¢o~.° ..~.° n-.°
m-.°

-.mu.o "no.0 ~.~.o ~am.¢
”05.: aw~.° ~_o.o ¢a~.°
m,¢.° mma.o

,cmn.° oug.o _.~.° m.~.o
_n~.° c.~.° ”$5.0 ~.a.°
n_a.o nn~.°

N xaa

 

5 322::

m an:

m...° ouh.o mac.o .oa.o
¢m~.° ..m.° «$5.0
._o.o coa.°
u. ,o.u=_=
~wn.o ¢°¢.° .nh.o m¢¢.°
.w~.° .n~.° mwa.¢ ~o~.o
“no.0 a.m.o
a. ,o.u=_.
, avm.o oan.° osm.° m-.°
o_~.o ~n~.° __~.° ~.o.o
mnm.o mah.o

"m¢.m musmwm

 

 

 

.an..o mn¢.° 5.5.o _m~.o
._~.° mao.° ..w.o
n.o.° mo~.°

- omm.o mn..° ~.~.° ~m~.°
_¢c.o no~.° ¢m~.° www.°
guc.c n.¢.o

, ”Nn.o am..o mo~.° °_a.c
~n~.o _.~.° owo.o
onm.° .«m.o

_ ».a

194

 

-.N...

.N sauna“;

new cowumuoH up mumam ham mmpm

 

 

 

 

ann.° ~n¢.¢ .a~.°
_a~.° na..° ~a~.°
.oa.o -o.o
_o¢.° aom.° nou.o
~_~.. -~.o .o~.°
.°¢.. .an.°
. omm.° ¢~u.o ~m~.° nmh.o
_.~.o _n~.° a_~.°
na~.° .Na.°
h ».a

 

 

 

.mnm.o .nn.° a.w.o anh.o
._~.° ao¢.o
moo.° o¢~.°
ow .o.==_=
.nn.° ”No.0 ~°~.° ~a~.°
”$5.0 ~.~.° .-.°
nnm.o oaa.°
.N ,o.u=_.
~°¢.o amn.o .a~.°
~n~.o ~am.o a¢~.°
.m~.° ~.¢.°
«N ,o..=_=
n ».a

"om.m muzofim

 

 

 

-~¢m.o .om.o o-.° °a~.o
$.m.° °°~.° w-.o
nam.° m.c.°

-n_n.o nhn.o _¢h.o mm~.o
c.~.o -~.o ¢a~.o
oNo.¢ a.a.°

- aag.o nsn.° om~.° n.m.°
a-.° nn~.° a-.°
oa~.° snm.°

_ >.¢

195

Hindrov 18 Hindrov 2A

. 0.790 0.834
0.845 0.776 0.739 0.735
0.777 0.718 0.589 0.535

0.754 0.688 0.740
0.527 0.760 0.598 0.643

0.935 0.792,

 

Hindrov 18 Uindrou 28 '

0.843 0.833
0.815 0.825 0.732 0.719
0.703 0.733 0.566 0.526

0.785 0.734 0.749
0.799 0.712 0.601 0.576

0.822 0.322,

 

 

Hindrov 18 Hindrov ZC
0.531 0.811 0.818 0.835
0.803 0.694 0.746 0.783 0.627 0.717
0.804 0.712 0.705 0'529. 0.769 0.670 0.571 0.564

 

 

Figure 5.51: Average free air space by location
and turning for Windrows 1 and 2.

windrow surface. In the interior of the windrow, free air
space ranges from 0.36 to 0.69. Free air space decreases
with depth. No changes over time were observed (Figures
5.52 and 5.53).

The coefficient of variation of the free air space
measurements from the preliminary experiment is 7.5 %.
Applying this i 7.5 % factor to free air space changes over
time led to the result that for 69 percent of the locations,
free air space does not show a change greater than 1 7.5 %,
with 16 % showing an increase and 15 % showing a decrease in

free air space for the 1 week compost turning period (Figure

5.54).

.cofiumooH wo
mew» um>o mwmcmcu mumam ppm ovum "mm.m mpsmmm

 

 

m=o_.~uod

-0. m a _ mm 2. cm co.
a n N ¢_ 5. on mm on
a . no ”a

 

.mEMu um>o mwmcmnu mumam cam mmum "mm.m whomfim

196

 

 

 

 

 

 

mo» u 50 a.» n mm «o» u so

> > > > z z > z > z z

z > z > z > z p >
z > > > > >

ow :o.u=_: mm accec_: <~ so.ec_=

no» u an no» u on no» m cm

x a z z z a > > z z >

z z > z z > > z > >
> z > > > >

a. .o.u=_= a. so.==_= <_ .o.u=_=

197

 

 

 

 

 

Nindrov IA Hindrov 18 Hindrov 1C
1 Y I N H Y
N N N N Y N -Y N Y N N
N N -1 N I N -Y -Y N N -Y -Y
- 10 90 9 O - 30 60 9 10 - 22 56 4 22
Hindrov ZA Hindrov 28 Hindrov 28
N N H N N N
Y I H N N N N Y -1
N N -Y Mr N Y -Y N Y N Y
- 11 78 O 11 - 12.5 75 0 12.5 - 11 56 O 33

Figure 5.54: Free air space changes over time
greater than 7 % coefficent of
variation.

5.10: Additional Moisture and Physical Properties

 

Volumetric moisture content ( 8 ), volumetric solids
content ( fs ), void ratio ( e 1, and the degree of
saturation ( s ) were calculated according to the methods
described in Chapter 4. The results are presented in

Figures 5.55 to 5.62. The implications of these results are

discussed in Chapter 6.
§4lli Windrow Color Changes

When the windrows were pulled apart, three separate
color zones could be seen. Profiles of the front half of
Windrow l cross-sections are shown in Figure 5.63. Figure
5.64 shows cross-sections of the left and right side of the
fron half of Windrow 2. The method of note-taking for these

observations was not standardized until Windrow 2 was

 

 

 

 

..~.° a-.° -~.° m.~.°
sa_.° s_~.° cm_.°
om_.o

8 -

mu oo~.° mom.o °.~.° -~.°
a.~.° 3°~.° °¢_.o ~a..o
nn_.° am_.°
No..o _-.o a.~.° ~_~.c
._~.° o.~.o n~_.o _n_.°
an..° m_~.°

s ens

.H 3ouocw3 no“ comumuoH

an ucmucoo monumwos umbumEDHo>

 

"mm.m whomwm

 

 

 

 

-mm..o _m~.o hs~.o .s_.o
~o~.° .om.o s~_.o
mm_.° _m_.o

a. ,o.u=_.

.m.n.° .mn.° n-.o ~n~.o
oa_.° n-.° m...° .~_.o
m~_.o -~.o

._ ,0..=_=

-mam.o ~nm.o a-.¢ so_.o
~.~.o ._~.o o.~.o mn_.o
-_.o -_.o

._ .o..=_=
m ».a

 

-mn..¢ ma~.° ¢m~.° mm_.o
.w~.o awN.o .m_.°
sn~.° _m~.°

lasn.° mo..o m_~.° ¢-.o
n¢~.o. an~.° ¢m_.o ~__.°
a._.° ~m_.o
so..o mm..o ~°~.o ms_.o
-~.° .-.° wn_.o
s._.° .w_.a

_ ..a

1599

 

 

 

-m...o osm.° aon.° ma_.°
s~_.o -..¢ ~o_.o
a__.° a...°

-~a~.° a-.° “a..°
«.~.° _n~.°
hg..° .5..°

-mon.o m_n.° a°~.° m°~.°
a_~.° an~.° n.~.o
n¢_.o .m_.o

a >..

.m Boupcwz u0m cowumuoH

 

 

 

 

 

 

an ucmucou muzummos uwuumESHo> "mm.m whommm
oom.o ~am.o mo~.° s-.o -owm.o o-.o mm~.o m~_.o
mv~.o om~.o «on.o ww~.o ~m_.o
mm..o oa_.o _v_.o ~m_.o
UN soevc_a
-ncm.o m_n.o mv~.o vh~.° -n_v.c ~mn.¢ vo~.o v~_.o
nm_.o n_~.o ~o_.o e_~.o ~m~.o m~_.o
ov~.o nn—.o mv_.o a~—.o
an aoeac_=
-nmm.° a.n.o °m_.o -.¢~.° osn.o nm~.° ~m_.o
o_~.e nmn.o mo~.o vm~.o -~.o nw_.o
no~.o mn~.o ~n_.o nv_.o
cu socuc_a
a sun _ x»:

200

.H 30hmqu3 hOw COMHMUOH

an ucmucou mofiaom Hmcowuumpm

 

 

 

 

-m.°.o ~.°.° ~.°.o a~°.o
om°.° °.°.° mn°.°

~m°.°

-nno.o ono.c ..o.o n.o.o
a.°.° om°.° «No.0 «no.9
a~°.° ¢~°.°

-_~°.° o.°u° _.°.° nm°.°
__~.° _.°.° «No.0 n~°.°
an_.° .mo.°

~ 1.:

unm.m whammm

 

 

 

 

 

 

-amo.o m.o.o ono.o ~mo.o moo.o _mo.o nwo.c vmo.o
omo.o mno.o ~mo.o avo.o mvo.o w~o.o
omo.c .no.o muo.o vmo.o
u_ socuc_=
. soo.o moo.o _vo.o svo.o -mwo.o m~c.o smo.c wmo.o
wmc.o mvo.o -o.o vmo.o vvo.o avo.o n~o.o owo.o
-o.o v~o.o mmo.o m~°.o
a. scene—a
-mmo.o mno.o neo.o ~mc.o woo.o omo.o mmo.° wwo.o
«vo.o ~no.o _vo.o muo.o omo.o mno.o m~°.o
ouo.c -o.o vao.o vuo.o

c. soccc_=

n 1.:

_ 1.:

201

 

 

 

-:::.: :::.: 1::.: :::.:
.::.: 111.: :::.:
.::.: :1:.:

-:::.: 1::.: :::.:
:.:.: :1:.: 1:..:
:::.: :::.:

-:::.: :::.: :::.: :::.:
:.:.: :::.: :::.:
1::.: :::.:

1 1::

.N 30uocmz uON cowumuofl
mo ucmucou mpwaom Hmcomuomum 1mm.m mcommm

 

 

 

 

 

 

-.::.: :::.: :::.: :.:.: - :.:.: .::.: 1.:.: 1::.:
:1:.: :.:.: 1::.: :.:.: .::.:
:::.: :::.: .::.: ~1:.:
:: 1:.::_1 .
4:::.: :::.: :.:.: :::.: - :::.: :::.: .::.: :::.:
:_:.: :.:.: 1::.: :::.: 1.:.: :::.:
1::.: 1::.: :::.: :::.:
:: 1:.::_1
-:::.: :::.: :::.: - :.:.: m::.: 1::.: :::.:
:::.: :.:.: :::.: :::.: :::.: :::.:
:::.: :_:.: :::.: :::.:

:: 1:_::_:

m 1.:

1 1.:

202

.H 30poc13 10m cowumuoH >2 ofium: owo>

 

 

 

.:.:1

 

 

..1: :.1: .... ....
.... :..~ ...:

...:

-:.:1 1.:1 ..1: :.1:
...1 ...: ...: ...:
.... :.:.

-..:1 ...1 ...: ...:
...: ...: ...: :.:.
...: ...:

1 1.:

 

...: :.:1 ..::
...: :.11 1...
...: 1.1:

:1 .....11
-... ...1 ...: ..1:
N..: ..:: ...: ..1.
1... n.:.

:1 .....1:
:.:1 1..1 ...: ..1:
...: ...: ...: ..1.
.... ...:

11 .....1:

: 1.:

 

 

 

"mm.m musmmm

:.11 1.:1 :..: ...:
..N: :.1: ....

...: :.:N

-...1 1.11 ...: ...:
..1: :.:: ...: 1.1.
:.1: :.:.

1... :.:1 :..~ 1...
...: ...: ...:

:.1. :.:.

1 1.:

 

203

.m 3010:13 no. comumuofi >2 011mb U1o>

 

 

 

 

..11 1..1 ..11 ...:
..1: 1.11 1.::

..1. 1...

1.:1 :.:1 ..1.
1.11 1.11 ...

1.:. 1...
.1.:1 1.11 ..11 ..:.
...: ...: ...:

..1: ....

1 1.:

 

 

 

 

...1 ...1 1.11 :..1
:.:1 1.:. ..11
.... ...:

1: 1.1.11:
...1 1.11 n.11 ....
..:. 1.:. ...1
1.:: 1.::

1. .....11
-...1 ...1 ..1: ...:
...: ...1
...: 1.::

.1 1.1.:11

: 1.:

 

 

 

 

“om.m whomwm

,1.:1 .... ...: :.:.
:.:1 ..1. ...:

:.:. 1.:.

1.:1 ...1 1.:: ..1:
...: ...: 1.::

:.:: ....

41.11 ..11 ...: ....
...: ...: ....

...: 1.:.

1 1.:

2(34

 

.H 3oupc13
pom cowumuoH xn co1umpsumm mo mwummo

 

 

 

 

 

 

1.1.: .11.: 1:1,: .11.. 1...: 1.1.: 1.1.: :11.:
.:1.: .11.: 1.1.: :11.: 111.: 1.1.:

:11.: 111.. 1.1.:
:1 1.1.11:

,1::.: 111.: 1.1.: :11.: - .:1.: .:1.: .11.: :11.:
1.1.: 111.: ..1.: :11.: .:1.: .11.: 1.1.: :.1.:
:11.: 1.1.: .11.: ::1.1

:1 1.1..1:

,1...: .:1.: ..1.1 .11.: - .1..: .11.: 1.1.: 1.1.:
.11.: 1.1.: :11.: .11.: 1.1.: .11.: 1.1.: :11.:
1.1.: .11.: .11.. 1:1.:

.1 1.1.11:
s 1.9 m 11:

"Hm.m musmwm

 

 

 

1::.: .:1.: :w1.: 1:1.:
.m1.: 111.. :11.:
.:1.: 111.:

- 1:... 1:1.: .11.: .:1.:
111.: 1.1.: 1.1.: .11.:
1.1.: .:1.:

1...: :...: .:1.: .11.:
.11.: 1:1,: .11.:
:11.: ::1.:

1 1.:

 

205

101 :o1umu01 xn :OMumpoumm mo wmummo

 

-11...

.1 301oc13

 

 

1mm.m musmmm

 

 

 

 

 

 

1.... 111.: 1.1.. -..... 1.... 111.. 111.. -..... 111.. 1.1.. 111..
.11.. 1.... 111.. 111.. .11.. 111.. 1.1.: .11.:
111.. 111.. 111.. ..1.. ..1.. 111..
11 1.1.111
.111.. .11.: 111.. -.1... 111.. .11.. 111.. -..1.. 111.. 111.: 111..
.11.. 1.1.. 111.. 111.. 111.: 111.. 1.1.. .11..
111.. 111.. ..1.. 111.. 1.1.. 111..
11 1.1.111
- 111.. 111.. 111.. .11.: , 111.. 111.. 111.. -111.. 111.. 1.1.. 111..
.11.. 1.1.. .11.: 111.. 111.. 111.. 1.1.. .11.. 111..
111.. 111.. .11.. 111.. 111.. 1.1.:
.1 1.11.11
1 1.. 1 1.: 1 1.:

206

 

 

 

 

 

 

 

E 63 X
D 37 Z
Windrow 1A
A
£11 0 542
D 46 z
Windrow 1C
88 1
E
D 12 X
Windrow 10

Figure 5.63: Windrow l cross-section showing color
zones.

207

 

 

 

 

 

 

 

Windrow 2A

 

 

 

 

 

 

 

 

 

 

 

 

 

Windrow 2C

Figure 5.64: Windrow 2 cross-section showing color
zones.

208
started, so the profiles do not contain the same
information.

At the outside surface and in the toe area, the compost
is dark brown color. Temperatures in this zone are
moderately hot. A second zone of mixed brown and white
colored material is underneath this first zone. This second
zone is typically 15 cm thick near the center line of the
windrow and 30 cm thick in the toe area. A well-defined
color gradient is present in this zone with the white
particles increasing towards the center of the windrow.
There is a distinct white line at the interface between the
second zone and the yellow third zone. High temperatures
tend to correlate with the occurrence of the white
particles; temperatures are very high in the white line
interface. The white line is most noticeable in areas where
there was a bloom of mushrooms on the windrow surface.

The third zone is the color of the original uncomposted
DMS, usually a yellow to yellowish-brown.

The location of the three zones varies between the end
of each successive windrow turning. This can be quantified
by looking at the change in the percent of the total area
represented by each color and by the depth of penetration of
the black and white zone into the windrow at the centerline
and bottom of the windrow.

Table 5.4 presents the percent cross-sectional area for
each windrow turning profile and color zone. With each

successive turning, the average Windrow 2 Zone D area

209

Table 5.4: Windrow color changes.

 

 

Windrow Color (% Area)
Yellow Brown and Brown
White

1A 37 63a

13 46 54

1C 12 88

2A Left 42 40 18
Right 45 28 27
Average 43 35 22

23 Left 33 38 29
Right 39 46 15
Average 36 42 23

2C Left 27 59 15
Right 34 56 10
Average 31 57 12

 

a Could not calculate brown and brown/white areas from
notes. Area percentages represent non-yellow areas.

 

210
decreases. The percent area in Zone D in Windrow 1
decreases dramatically from Windrow 1A to Windrow 1C. The
left and right halves of Windrow 2 are quite different.
This is probably due to differences in windrow size and
construction. Zone B in Windrow 2 expands inward with the
interior edge moving faster than the exterior edge.
Examining the penetration depth data presented in Table 5.5,
the side and top penetration depths are very different. Top
penetration depths stayed the same or increased slightly
with time in both windrows. Average side penetration depths
for Windrow 2 increased almost linearly with each turning.
The top penetration depths are very consistent when compared
to side penetration depths. Variances for top penetration
depths are 0.33, 0.25 and 0.33 inches for turnings A, B, and
C, respectively, while side penetration depth variances are
4.75, 98 and 108 inches. Side penetration is much more
sensitive to the conditions encountered in each windrow:
ambient temperature and relative humidity, substrate, and

windrow size and shape.

Table 5.5:

211

Windrow penetration depths.

 

 

Windrow Penetration Depth (cm)
Side Top
1A 89 33
13 76 36
1C 152 41
2A Left 79 36
Right 80 33
Average 80 34.5
28 Left 114 36
Right 69 36
Average 91.5 36
2C Left 107 40.5
Right 107 38
Average 107 39.25

 

 

CHAPTER 6

DISCUSSION

6.1: Experimental Methods

 

 

6.1.1: Probe Design

Probe design was generally successful. Probe insertion
did not cause the windrow to split apart during the one week
turning periods. The probes could be inserted and were
durable. Some difficulty was encountered with the repair of
damaged gas collection lines; this was solved by taping new
lines to the outside of the probe. New designs should allow
easier probe assembly/disassembly.

6.1.2: Temperature

 

Temperature monitoring equipment had a calculated
measurement error of 11.5°C. Measured variations during
calibration were much smaller than this on the order of
10.07°C.

In retrospect, the ability of the experiment to test a
key assumption could have been improved. Temperature and
gas monitoring points were located in the windrow front half
center section to allow replicate samples to be taken. This
placement did not allow testing the assumption that
longitudinal edge effects were negligible. While two of the

three windrows with replicate monitoring locations did not

212

213
show evidence of longitudinal effects, these effects cannot
be ruled out. The longitudinal effects could have been
determined if probes had been placed in one quarter of the
windrow.

Additional measurements of temperatures at the windrow-
air and windrow-concrete boundries would have improved the
accuracy of temperature contour figures and aided future
modeling efforts.

6.1.3: Gas Concentration

Unexpectedly high oxygen concentrations in the windrow
interior suggested that leakage of air had occurred during
the experimental procedures. The gas sampling probe and
withdrawal equipment functioned very well with no influx of
air attributable to these components of the gas sampling and
analysis system.

The storage system that was initially adopted was able
to function adequately over time periods of less than 2
weeks. Covering the septum top with silicon caulk stopped
the sample from leaking out due to overpressures inside the
storage tube and greatly extended the storage time. If this
type of storage system is to be used in the future, longer
term experiments should be conducted to determine the exact
length of intact storage. The injection of a fluid into the
tube while air samples are being withdrawn would also keep
internal pressures closer to atmospheric pressure.

The largest error in the gas system occurred when gas

was withdrawn from either the sampling probe or from the

214
storage system. Correction factors based on experimental
work (See Appendix D) accounted for most of the differences
between the expected and measured values.

Several correction factors were evaluated: low, high,
and average influx. The low and high influx corrections
were based on the observation of two separate linear
segments in the influx vs. calculated sample pressure data.
The low influx correction was based on the initial slope
which occurred while calculated sample pressures were less
than 1.0 atmospheres. The high influx correction was based
on the steeper sloped influx vs. calculated internal
pressure data. Calculated pressures in this case were less
than 1.0 atmosphere. The physical explanation for these
two sloped lines is most likely diffusion for the low influx
segment and a pressure gradient for the high influx
condition.

The average influx approach represents a compromise due
to the uncertainty of the actual initial pressure in the
sample tube during sample removal. If this pressure were
known, more accurate estimates of gas concentrations could
be made.

From a parameter estimation and modeling standpoint,
the gas concentration data are probably accurate enough to
illustrate basic trends and to do initial estimation and
modeling work. Additonal data must be collected, however,
to justify more rigorous parameter estimation and model

validation efforts. Carbon dioxide data are subject to less

 

215
error than the oxygen concentrations because the influx
mainly caused the addition of oxygen. The use of a gas
valve on the sampling syringe would directly reduce influx
and should be used in future experiments.

A large quantity of gas samples were analyzed in this
experiment. The available analytical method, a gas
chromatograph, is quite time consuming. The use of
equipment such as an infared C02 analyzer and a paramagnetic
oxygen analyzer could speed up the analysis process
considerably, particularly if it is onsite.

6.1.4: Gravimetric Moisture Content

Gravimetric moisture content determinations had
acceptable levels of measurement error with 12 % wb standard
deviation. Measurement errors can be reduced with the use
of scales accurate to 10.01 mg. Sample sizes with wet
weights between 2.5 and 3.0 grams can further reduced the
range of measurement error. Gravimetric moisture content
determinations are also very time consuming.

6.1.5: Volatile Solids

Volatile solids determinations made in this experiment
were subject to unacceptable measurement errors with an
average standard deviation of 17 percent. Dry sample
weights of 0.7 grams coupled with the use of a scale
accurate to 0.01 mg could produced standard deviations of

approximately 1 percent.

216

6.1.6: Dairy Manure Solids Sampling and Bulk Density

The rotary corer sampling method proved to be an
acceptable way to sample compost materials for insitu bulk
density, porosity and free air space determinations.
Initial experimentation indicated that bulk density
replicates taken at the same depth had coefficents of
variation of 11 % for several depths. Windrow bulk
densities obtained with the corer were in reasonable
agreement with the overall windrow bulk density. This
agreement could be more readily determined if samples were
taken from deeper in the windrow's interior and if a windrow
consolidation model was developed. Examination of the
source of bulk density measurement variance indicated that
the variance in effective radius measurement was the most
significant factor. Future study of the core sampling
method for composting materials should focus on the effects
of corer speed, radius and compost material on bulk density
determinations.

6.1.7: Porosity and Free Air Space

 

Examination of the sources of porosity measurement
variance indicated that errors in determining the particle
density could affect porosity determinations. This was also
an important error source in free air space determinations,
but radius measurement errors were the largest source in

this case.

217

6.1.8: Dairy Manure Solids Placement

 

Dairy manure solids placement and operator skill were
observed to have a large impact on the apparent bulk density
of a windrow. Pushing DMS into position, dropping DMS in a
large mass (as opposed to ”shaking" it out of the skid steer
bucket), and leveling the windrow with the bucket all
appeared to increase compaction. Although efforts were made
to standardize the placement procedure, placement of DMS
into the windrow probably had a large and unpredictable
effect on bulk density and therefore on the transport
properties in the experimental windrows.

6.1.9: Windrow Size Measurements

The method of measuring windrow size changes had
acceptable accuracy. Measurement errors were due to sag in
the lines, problems with determining the windrow surface,
and parallax.

6.2: Experimental Results

6.2.1: Temperature

 

The temperatures encountered during both DMS composting
runs were lower than those reported for uncontrolled sewage
sludge windrows. A number of factors could have influenced
this. The large quantity of water, as indicated by high
moisture contents, could have absorbed the heat output of
the reaction. This type of behavior was observed for
decomposing oat straw in an adiabatic calorimeter by
Bartholomew and Norman (1953) Conversely, the windrow size

might have been small enough so that heat dissipated before

218
the temperature could increase. Finally, if the reaction
rate was low, the heat output would be low and high
temperature would not have occurred.
6.2.2: Gas Concentration

There are two possible explanations that can account
for the gas concentrations observed in this experiment, in
particular the higher than expected oxygen concentrations in
the windrow interior. The first hypothesis is that the
reaction rate is limiting. In this case the flux of oxygen
to the reacting sites is large enough so that oxygen is not
rate limiting. The second hypothesis is that the reaction
is transport limited. Oxygen flux, in this case, is great
enough to supply all the oxygen demanded by the reaction.
The high oxygen concentrations would be due to experimental
error, in particular the influx mentioned earlier.

The reaction limited hypothesis occurs under two
different scenarios: (1) high transport rates/medium
reaction rates (HTmR, small windrow), and, (2) medium
transport rates/low reaction rates (MTLR, relatively non-
reactive substrate).

The system studied in this experiment is not a high
rate system: oxygen transport is either by diffusion or
natural convection, neither of which is a high-rate process.
Forced aeration or even enhanced natural convection were not
studied in this experiment. Therefore the possibility that
the observed high oxygen concentrations are due to HTMR is

not realistic.

219

The MTLR scenario is more plausible in light of the
previous points. The lack of change in volatile solids
levels indicates an unreactive substrate.

There are several arguements against MTLR, however.

The first argument is that volatile solids is not a good
indicator of substrate consumption. The volatile solids
measurement is not very specific: it cannot distinguish
between the original substrate, added microbial mass and
material that may be organic but cannot be degraded
(Finstein, 1986). As is discussed later, the use of an ash
solid basis is a better indication of substrate consumption.
There is evidence that substrate is being consumed when
volatile solids are expressed on an ash basis.

If the possibility exists that the substrate is not
limiting, then the low oxygen concentrations in the windrow
interior become more troublesome for MTLR. In this case,
these low concentrations could be limiting the reaction. In
many microbial systems, including composting, the transport
of oxygen across the water film is the limiting rate
(Finger, 1975; Haug, 1980). While Haug's assertion is based
on forced aeration systems that have high air flow
velocities, Finger studied diffusion/natural convection
systems that are similar to what is studied in this
research. Furthermore, Shell's (1955) study of diffusion in
windrows also pointed to the limited ability of the
effective oxygen diffusion rate to supply oxygen to an

actively composting material.

220

The strongest challenge to the short-term reaction
limited hypothesis is that if the interior measured oxygen
levels were not rate limiting, why were higher interior
temperatures not observed? During temperature measurements
for Windrow 1A, Locations 6 and 10 Figures 5.9 and 5.10)
have an "S" shaped profile. Locations 5, 8 and 9, on the
other hand, have steep increases followed by a leveling off
of temperature.

Substrate is initially the same at the all locations:
it is not the cause of the different temperature histories.
Possible high temperature limitations do not apply at
Locations 6 and 10 as both these temperatures are solidly in
the mesophilic range. Referring to the moisture content
data presented in Figures 5.32 and 5.33, the wet basis
gravimetric moisture contents at all locations are
approximately the same, so no difference in the effect of
moisture on the reaction rate or heat storage is expected.
Examination of wet bulk densities (Figure 5.40) and free air
space (Figure 5.51) indicate that the rate of transport of
gases, whether by diffusion or natural convection, might be
lower in the interior locations (6 & 10) than the locations
closer towards the surface of the windrow (5, 8, 9).

Given that other measurements do not provide a ready
explanation for the shape of the temperature curves, oxygen
transport rate limitations might provide an answer. In
Figure 5.10, while Locations 8 and 9 are showing large

temperature increases, Location 10 does not increase in

221
temperature. Only when the temperatures at Locations 8 and
9 level off does the Location 10 temperature increase. The
increasing temperatures at the two higher locations increase
the reaction rates and most of the available oxygen is
consumed, leaving little to diffuse down to Location 10. As
reactions at Locations 8 and 9 leveled off due to
temperature inhibition, more oxygen diffused down towards
Location 10. Oxygen ceased to become a limiting factor in
the reaction at Location 10 and the increased reaction began
to put out more heat, causing temperatures to increase.

The course of the reaction at Locations 5 and 6 (Figure
5.9) is somewhat different. Temperatures begin to increase
before the maximum temperature plateau is reached at
Location 5. This could be due to a higher transport rate
near the edge of the windrow (Location 6) as opposed to the
windrow interior (Location 10). Evidence for this
possibility will be discussed in a later section.

Another feature of the temperature profiles that could
be explained by oxygen limitations is the leveling off of
Locations 6 and 10 before the higher temperatures of 5, 8
and 9 are reached. In this case, while the reaction rate at
the higher temperatures is reduced, it is still occurring
and consuming oxygen. Therefore, while more oxygen is being
transported into the windrow interior, not enough is present
in the free air space at Location 10 to permit a reaction
that is not limited by low oxygen concentrations. If higher

oxygen transport rates were present at Location 6, the

222
oxygen concentration might be higher than that at Location
10, although still rate limiting. The higher temperatures
would be due to a higher rate of reaction caused by oxygen
concentrations which are less limited in the free air space.

A potential objection to the temperature level argument
for oxygen transport limitations that heat conduction from
the bottom of the windrow was enough to limit the
temperature rise. This could indeed be an explanation,
particularly if the ground was very cold, because of the
large thermal mass of the earth.

Measured slab temperatures during this period are very
close to those at Location 10 and increase along with that
temperature. In addition, one would expect this problem to
occur in the winter when the ground was very cold. Both
tests were conducted in the summer and fall, however,
presumably after the ground had warmed up. Furthermore, DMS
had been composted there for several months previously, so
heat stored in the ground would have limited this effect.

Based on the evidence presented above, I feel that a
oxygen transport limitation is the most probable explanation
for the gas concentration data.

A common assumption in forced air composting is that
the sum of measured oxygen and carbon dioxide concentrations
equal 21 percent. This is roughly the concentration of
oxygen in the ambient environment. The sum of oxygen and
carbon dioxide at several locations in the compost windrows

I studied was greater than 21 %, however. Initially, I

223
thought that I had encountered another indication of
experimental error. Re-examination of the mushroom
composting literature revealed that carbon dioxide
concentrations could become as high as 30 % in windrow
interiors (Lambert and Davis, 1934). The high carbon
dioxide concentration in Lambert and Davis' study was
correlated with zones of ”green" manure and they suggest
that manure decomposition is retarded in this area.

6.2.3: Volatile Solids

 

Volatile solids measurements had several problems
associated with their use. The method chosen to measure
volatile solids was subject to large measurement errors.
Future work that uses volatile solids will need to use more
accurate techniques, as described in Section 6.1.5, above.

Expression of volatile solids on an ash solids basis
is more sensitive to changes in volatile material than
expression of volatile solids on a dry solid basis.
Expressed as a percent decrease in the original ash
fraction, the dry solids basis decreased only 8 %, while the
ash solids basis decreased between 30 and 44 percent.

The conversion to measuring substrate and substrate
consumption on an ash solids basis correlates more closely
with other indications that a composting reaction was
occurring. The large changes in temperature, gas
concentration and compost color can therefore be supported
by evidence of substrate consumption.

While the volatile solids determinations do provide a

224
rough measurement of biological activity, its usefulness is
limited because it lacks both specificity and sensitivity.
Finstein et a1. (1986) note:

The test fails to discriminate among readily
metabolized, putrescible material, less readily
metabolized material, and organic material that is not
metabolized during any reasonable composting period.
Process performance is primarily concerned with the
first of these, secondarily with the second, and not at
all with the third. Also,...both fresh organic waste
and stabilized organic residue are included in the VS
test, which decreases sensitivity.

Similarly, sensitivity is poor because a high
percentage of the dry wei ht is volatile matter. This
means that the decomposition of a large amount of VS
may result in only a small change in percent VS.

In the last 10 years most of the theoretically sound
work with composting has been done with sewage sludge and
garbage (Haug, 1980; Finstein and Morris, 1975). Volatile
solids has been used successfully as an indicator of
substrate consumption (Higgins et al, 1982) in those
reactive substrates. I based my work closely on this
experience. I neglected to consider the relatively inert
nature of the DMS substrate, however, and volatile solids
expressed on a dry basis showed little change. The mushroom
composting industry deals with straw based substrates that
are very similar to DMS. They express volatile solids
consumption on an ash basis in their work (Burrows, 1951).
It would appear that in dairy manure solids composting,
volatile solids should also be expressed on an ash basis.

Even though the use of the ash solids fraction allows

substrate level changes to be observed, for detailed work

225
more specific measurements of the substrate are useful. The
adoption of the procedures used in the mushroom composting
industry as described by Mueller (1962) present an
alternative. He recommended analyzing cellulose, lignin,
pentosans (90 % of hemicellulose), ether and hot water
extracts, and soluble and insoluble ash contents.
Determination of the total carbon, total nitrate and ammonia
nitrogen and nitrogen in lignin and alpha cellulose is also
important for research work.

6.2.4: Moisture Content

 

Moisture content changes expressed as a percent wet
basis do not exhibit major changes in either of the two
windrows: 81 to 77 % in Windrow l and 80 to 77 % in Windrow
2. This observation can be misleading, however. The wet
basis expression can mask changes in the moisture content
that are significant.

Both windrows had final moisture contents of
approximately 77 % wb. Initially Windrows l and 2 had 81
and 80 % wb, respectively. The corresponding percent dry
basis values are 335, 426 and 400 % db. A 21 % decrease in
the intial weight of water of Windrow 1 was observed;
Windrow 2 had a 16 % decrease from the initial weight of
water.

Another advantage of the dry basis expression is that
at the lower moisture contents it relates more directly to
changes in the drying rate. Bohnhoff and Converse (1986)

found that the point where the DMS drying rate changed from

226
a constant rate to a falling rate was at approximately 30
% db. This corresponds to 23 % wb.

While the expression of gravimetric moisture contents
on a dry basis is more sensitive to changes in water weight,
the denominator is affected by the change in solids weight
due to composting. Therefore it should also be corrected to
an ash solids basis, although this will result in very large
percentages.

In addition to the problems mentioned with the wet
basis moisture content, considering moisture contents on a
gravimetric basis hides changes in water relationships that
affect compost windrow heat and mass transfer.

6.2.5: Bulk Density

 

Wet bulk density changed with depth but did not show a
change over time. The change with depth is not suprising.
In-situ measurments of bulk density at specific points in
compost windrows have not been previously reported in the
literature, however. Disturbed sample bulk density and
gross bulk density have been measured but given the
variation with depth found in this experiment, they do not
adequately describe the local conditions inside the windrow.
Given the importance that bulk density plays in a variety of
yheat and mass transfer parameters, accurate bulk densities
are crucial if distributed parameter heat and mass transfer
modeling is to be conducted.

Bulk density appeared to be relatively constant over

time because of the relative lack of precision in the method

 

227
of measurement. The coefficent of variation of
approximately 11 percent lead to the uncertainty of the bulk
density determination of :17 kg/m3 for lower bulk densities
and :51 kg/m3 for the largest bulk densities encountered.
If longer term or high rate systems are to be studied,
windrow consolidation and bulk density increases will need
to be measured. Therefore, an improvement in the core
sampling method to improve its accuracy and precision is
imperative.
6.2.6: Additional Derived Physical Parameters

A wide range of physical parameters are derived from
the moisture content, volatile solids and wet bulk density
measurements.

The porosities encountered in this experiment are
rather high, typically above 90 percent. This indicates
that the pore spaces in this material occupy a high
proportion of the total volume. Given that the compost
substrate is a fibrous material with a high internal
porosity and that the particle density of the DMS is fairly
high, this is not too suprising.

The volume of water also has important implications.
This can be expressed in several fashions: free air space,
volumetric moisture content and degree of saturation. When
compared with the wet basis moisture contents that were
fairly constant, all three of these parameters show that on
a volume basis, water content increases with depth.

The free air space measurement indicates that as depth

228
increases, there is less space occupied by voids where gases
(:an be transported through. This can significantly reduce
'the transport of oxygen either through the void spaces or
across the thicker liquid films that are present at greater
depths. Experiments by Shell (1955) on diffusion of oxygen
through a composting mass support this observation.

The volumeteric moisture content (0) can affect thermal
conductivity. Increasing the volume of water in the series-
parallel models of thermal conductivity will increase k as
water displaces air. Bohnhoff et al. (1983) observed just
such a relationship between k and e. e is also more
convenient to use when calculating fluxes of water through
the composting mass.

In comparison with other composting systems, the dairy
manure solids system composts at very high moisture contents
between 76 and 81 % wet basis. The 9 and high fa values
point towards an explanation as to why composting could take
place. Free air space values measured in the DMS windrow
did not even approach the free air space limitations

observed in other substrates at FAS values below 30 percent.

 

This is probably due to the relative structural stability of
the dairy manure solid particles that were able to form a
porous windrow and the high moisture content of the
particles themselves.

The degree of saturation is useful in predicting the
effective stress on a material when bouyancy forces are

considered. Both the degree of saturation and the void

 

 

229

ratio are preferred expressions in soil mechanics where
prediction of compaction and consolidation are important.

Many of the properties discussed in this section are
derived directly from the soil mechanics and soil physics
literature. While they provide insight into the phySical
processes in composting, a significant difference exists.
Unlike most soil materials, compost particles are
compressible. This means that under enough pressure, the
internal porosity of a compost particle can be reduced.
Therefore the nature and size distibution of pores may
change significantly at greater depths. Intra-particle
water may be expressed into inter-particle pore spaces.
Particles that have been crushed during one windrow will
respond differently to stress when redistributed and
reloaded. Models of windrow bulk density must deal both
with the effects of compressibility and changes in
compressibility over time.

6.2.7: Thermal Properties

 

An important simplifying assumption made by Finger
(1975) was that thermal and physical properties in a compost
windrow were neither spatially or temporally variable. From
the results of this investigation, it appears that some of
the physical properties do vary spatially. The question
remains whether this physical property variance affects heat
transfer properties.

In order to evaluate the potential thermal property

spatial variability, I calculated a number of thermal

 

230
properties using temperature, moisture content and bulk
densities at 9 locations in Windrow 1B on days 1, 3, and 7.
Empirical equations by Bohnhoff et a1. (1984) for thermal
conductivity, specific heat and thermal diffusivity were
used. The results of the calculations are shown in Table
6.1.

Specific heat variability is very small in this
windrow. The maximum difference between the largest and
smallest values at any time is 2 percent. Thermal
diffusivities exhibit greater spatial variability.
Comparison of locations at any time reveals differences of
16 to 30 percent.

Spatial variability in thermal conductivity is very
large. The maximum percent difference ranges from 150 % to
320 percent. There is a pattern to this variability as
well: the outer surface and windrow interior have low and
high thermal conductivities, respectively. The low exterior
thermal conductivities provide insight into why the interior
the windrow tends to heat up. It would appear that the
assumption of constant thermal and physical properties is
not accurate.

6.2.8: Windrow Size Changes

 

There was evidence of windrow size changes both during
the windrow composting periods and between different
periods. Typically, during a one week composting period,
the center of the windrow would have the greatest decrease

and the windrow toe would show a slight increase. The

 

231

Table 6.1: Effect of windrow physical property
variability on the calculated thermal
properties in Windrow 1B.-

 

 

Day 1 Day 3 Day 7
50.8 50.8 49.7 52.4 42.8 51.2
57.5 53.9 53.4 62.7 64.9 58.5 55.7 68.1 61.4
59.4 53.3 45.7 44.9 65.1 61.5 51.0 48.7 46.0 57.7 60.1 55.6
Telperature, C (s = +l-1.5 C)
78.6 78.8 77.3 78.3 77.8 77.3
78.5 78.8 79.4 77.6 77.2 77.4 78.2 77.4 77.1

79.2 79.0 79.1 78.8 78.0 77.7 77.3 78.5 79.5 78.0 77.4 77.8
Hoisture Content, Ivb (s = 4!-2.2 106)
168 189 164 157 178 198
177 321 334 189 290 257 204 262 324
290 273 585 477 330 288 428 696 343 307 399 368
Bulk Density, kg/I‘3 (s = +l-11 I)
0.132 0.149 0.127 0.123 0.138 0.153
0.139 0.259 0.265 0.146 0.223 0.198 0.160 0.204 0.249
0.229 0.215 0.463 0.376 0.257 0.225 0.331 0.546 0.272 0.240 0.308 0.286
Volunetric Hoisture Content
0.074 0.082 0.071 0.072 0.070 0.084
0.083 0.140 0.142 0.092 0.137 0.114 0.092 0.130 0.146
0.132 0.116 0.239 0.186 0.157 0.133 0.174 0.301 0.134 0.136 0.178 0.157

Therlal Conductivity, HII‘Z C
3.51 3.52 3.47 3.50 - 3.48 3.47

3.5 3.54 3.48 3.46 3.47 3.50 3.47 3.46
3.53 3.52 3.53 3.52 3.49 3.48 3.47 3.51 3.54 ~ 3.49 3.47 3.48

p.
0

Specific Heat, Kj/kg C

. 4 . . 1.40 . .
1.29 1.21 1.16 1.11 1.36 1.33 1.17 ‘ 1.23 1.10

Thernal Diffusivity, 10‘SXI‘2/s

 

232
largest changes in the center of the windrow occurred from
day 1 to day 3. The experimental design did not include
detailed sjze measurements immediately after windrow
construction. If this had been included in the design,
large changes might have been observed.

The small changes in windrow size are not suprising due
to the shortness of the experiment and the type of
composting system studied. Other researchers that have
noticed larger changes either monitored long term changes in
windrow systems (Mears et al. 1975), or studied higher rate
systems (i.e. forced aeration) (Steniford et al., 1984).

The difference between the small changes that I
observed in the experimental DMS windrows and those reported
by Steniford et al. (1984) in co-composted forced aeration
systems may indicate that high rate composting leads to
changes in particle size that affect consolidation behavior.
Chang and Rible's (1972) analysis of the particle size
distribution and the nutrient content of fresh and composted
dairy manure indicates that composting decreases the
proportion of small particles in the overall particle size
distribution. If longer term composting systems are to be
studied, the effect of composting on particle size and
windrow consolidation should be examined.

6.3: Reaction and Transport Characteristics
6.3.1: Type pf Reaction
The oxygen concentration profiles presented in Figure

5.26 and 5.27 provide insight into the type of reaction that

233
is occurring. By analogy with the diffusion of a gas into a
reactive spherical particle, when the rate of gas diffusion
is less than the reaction rate, the gas concentration
profile falls very quickly from the ambient level near the
particle surface to very low levels that are maintained in
most of the particle (Findlayson, 1980). This description
is almost identical to the oxygen concentration profile
mentioned earlier.

In such cases, the Thiele modulus, or dimensionless
ratio of the characteristic time for transport to the
characteristic time of reaction, is much greater than 1.
This occurs with a fast reaction and relatively slow
transport of oxygen and heat (Findlayson, 1980). The exact
value of the Thiele modulus is hard to determine because of
the non-isothermal composting reaction. Difficulties that I
encountered with the gas sampling and analysis system
prevented accurate determination of this modulus.

The active presence of fungal masses in the advancing
color zone, the deeper penetration of the color zone with
each windrow turning, and the relatively poor nutrient
status of flushed and separated dairy manure solids suggests
that substrate limitations were present and may have caused
the classic microbial population progression to be short
circuited to the fungal recolonization period.

6.3.2: Unreacted Core Model
The overall course of the composting reaction behaved

like an unreacted core model of a reactant being transported

234
into a reactive mass (Levenspiel, 1972). According to the
unreacted-core model, the reaction proceeds at a narrow
front which moves into the solid particle. The substrate is
completely consumed as the front moves through the particle,
leaving behind an unreactive mass.

As applied to the composting windrow, the windrow
itself would correspond to the particle. Evidence of a
moving reaction front is provided by the observed movement
of the temperature, oxygen, carbon dioxide and color
profiles through the windrow over the course of the
composting period.

The rate at which each of the observed fronts moves
through the windrow can also be accounted for with the
unreacted core model. Initially, the reaction consumes the
oxygen very rapidly. The controlling step quickly becomes
the oxygen transport rate, and zones with low oxygen
concentrations have lower or non-existant aerobic reactions.

In those areas where aerobic reactions are occurring,
the rate of heat transfer begins to affect the reaction. In
areas with low heat transfer, the temperature builds up and
reactions increase. This process continues until inhibitory
temperatures are reached and the reaction rate decreases.
With this rate decrease, less oxygen is consumed and more is
available for transport to locations deeper in the windrow.

Finally, as the reaction proceeds, substrate at the
initial location is consumed to where it cannot support high

rate reactions. The front between the "consumed" and fresh

 

235
dairy manure solids moves inward.

The observed composting reaction was more complex than
the classic unreacted core model. Additional transfer
resistances due to transport inside the windrow and across
the water film surrounding the DMS compost particles must be
included. Another complication is that the composting
reaction occurs across a diffuse front rather than a sharp
interface between the spent and fresh solid zones. Finally,
the presence of temperature gradients influences both
reaction and transport.

These complications can be dealt with, however.
Levinspiel (1972) describes how the combination of
resistances can be handled and what implications the
different resistances have for what the controlling rate is.
The diffuse front problem is considered by Wen (1968) and
Ishida et al. (1971a). Wen and Wang (1970) treat the effect
of non-isothermal reactions.

6.3.3: Pore Space Heat and Mass Transfer

In the composting literature there is a consensus that
heat and mass transfer in the pore spaces of windrow
composting systems is primarily due to natural convection.
Support for this comes from a variety of sources and
includes anecdotal observations of steaming windrows, higher
oxygen concentrations over buried tile inlets (Lambert and
Davis, 1932), comparison of the maximum observed composting
reaction rates to oxygen diffusion rates (Shell, 1955) and

simple models of natural convection caused by temperature

 

236
and moisture content gradients (Haug, 1980).

While there is no reason to doubt that these inferences
are incorrect, there is no direct proof that natural
convection as opposed to diffusion is the major heat and
mass transfer mechanism. A major reason for this lack of
proof is the difficulty in measuring natural convection in
uncontrolled environments.

In my observations of the depth of penetration of the
color zones in the compost windrow, there is some evidence
that two different types of transport were present.
Comparison of the depth of penetration of white colored zone
at the top and sides of each windrow turning at completion
indicated that the side penetration was 2.4 to 3.1 times
greater, depending on the length of composting. The depth
of penetration of a reactive front over time is a rough
indication of the ratio of the rate of transport to the rate
of reaction or Thiele modulus.

The observed difference between the top and side of the
windrow could be due to different reaction rates or methods
of heat and mass transfer. Reaction rates are affected by
the substrate, oxygen availability, and temperature. After
the DMS is mixed and placed into a windrow, the substrate is
assumed to be the same in both locations.

Oxygen concentrations and heat removal are affected by
transport. Higher transport rates increase reaction rates
by providing more oxygen. Higher heat transport lowers

reactions initially but prolongs a reaction by delaying

 

237
inhibitory temperatures. The effect of higher transport
rates on the Thiele modulus for oxygen is therefore somewhat
ambiguous.

If the rate of reaction of the substrate is considered,
the effect is not ambiguous, however. Because of the higher
oxygen concentrations and lower temperatures due to higher
transport rates, the substrate is consumed more quickly.

The reaction front moves inward at a more rapid pace where
transport rates for oxygen and heat transfer are larger.

This experimental observation provides strong evidence
that two different rates of transport processes are present
in windrow composting systems. These observations need to
be better quantified before they are considered proven. The
evidence presented above is also not sufficent to conclude
what types of transport are present. Higher free air space
in the side could lead to greater rates of diffusion in that
area as compared with the top, for instance. Other
potential mechanisms include diffusion or roll type
convective cells in the top portion and in the windrow side,
diffusion, natural convection, or even heat conduction
through the slab.

6.4: Pathogen Survival

 

The ultimate concern with studying DMS composting is to
determine whether it can produce a bedding material that is
sufficiently free of mastitis-causing organisms that it will
not cause outbreaks of mastitis when used by dairy cattle.

This concern can be addressed by considering the effect of

 

238
observed composting conditions on the survival of pathogenic
organisms.

Microbial survival and growth is affected by a number
of factors including time-temperature conditions, gas
concentrations, moisture content, water activity, pH and
substrate availability. Time-temperature profiles are the
most important factor in composting and will be the subject
of the rest of this discussion.

The discussion in the Literature Review indicated that
there are a variety of standards for composting time-
temperature curves. A common one is the maintenance of 55°C
for three days. This criterion is based on refuse
composting where there is a concern with micro-organisms
that have higher sterilization temperatures than the
mastitis-causing organisms. Temperatures from 50°C to 55°C
are minimally to moderately inhibitory for E. coli. 35 to
45°C encompasses the optimum growth temperature range for E.
coli. Accordingly, these three criteria were applied to the
time-temperature data in Tables 5.2 and 5.3. The number of
hours spent in each temperature range are presented in Table
6.2. An assessment of the impact of the time spent in each
temperature range criteria as presented in Figure 6.1.
Looking at Figure 6.1 it is apparent that while there are
zones of temperatures lethal to coliforms there are also
areas that have sublethal or even optimal temperature for
growth. Generally, the toe area and the center of the

windrow seem to have the largest potential for E. coli

 

239

 

 

Table 6.2: Length of time spent in selected temperature
ranges by location and windrow.
Time (hours)
Windrow 1A 18 1C 2A 28 2C
LOC
T > 55 C
1 117 76 0 0 0 0
2 144 151 0 109 117 76
3 128 130 0 33 155 83
4 0 0 na 112 156 88
5 125 136 151 49 148 76
6 0 45 114 0 117 24
8 119 0 0 107 151 88
9 93 133 19 0 100 56
10 0 7 95 0 0 0
T > 50 C
1 159 131 0 40 12 0
2 155 158 24 127 161 99
3 146 156 56 62 158 89
4 0 54 na 126 160 96
5 142 150 158 80 162 84
6 0 94 132 0 148 47
8 143 157 0 123 159 96
9 126 157 135 8 128 72
10 0 73 116 0 38 13
35 < T < 45
1 0 0 71 21 123 73
2 0 0 95 2 2 0
3 0 0 63 41 0 4
4 137 36 0 2 2 0
5 14 0 0 38 0 8
6 74 0 0 73 2 33
8 12 0 164 2 0 0
9 21 0 17 64 4 2
10 136 41 5 31 57 31

 

 

214(3

 

 

 

 

63 L62
L L61 L61
L L 62 63
Nindrov 18
61 N
L L L
L L N N61 _
Hindrov 18
63
62 LN N61
62 62 LN L
Nindrov 16

= T ) 55 C for 3 days.

 

 

L L
L 862 62
61 62 62 62
Uindrov 24
L L
L L L
63 L L 627
Nindrov 28
L L
L L N
62 L 62 62

 

Hindrov 20

61 = 0pt1:UI regrowth temperatures ) 10 hours.
62 = 30 ( Optinul regrowth teaperatures ( 100 hours.
63 = Dptinuo regrowth temperatures ) 100 hours.

Figure 6.1: Summary of assessment of time-

N = T ) 50 C for 3 days.

temperature effects on microbial
survival and growth.

 

241
survival and regrowth. Based on the number of locations
that had temperatures greater than 50 or 55°C, both windrows
seemed to be equally effective. When the length of time
that a location had optimal E. coli growth temperatures is
considered, Windrow 1 had only half as many survival zones
as Windrow 2.

Another way to evaluate the effect of a time-
temperature curve on organism survival is to calculate the
number of log reductions that it causes. Following the
procedUre described by Haug (1980), kd as a function of
temperature was calculated for "coliform bacteria" with
time-temperature survival data originally developed by Ward

and Brandon (1977). The equation developed was

kd = 0.216 exp [0.145 * (TC—50)] (6.3)
where TC is the material temperature in degrees C. Equation
6.3 was substituted into

1n (no / nt) = Rd (6.4)
and converted into base 10 logs to achieve
log (no / nt) = (0.216 / 2.303) *
exp [0.145 * (TC - 50)] (6.5) )

 

Equation 6.5 can be used with time-temperature data to
predict the number of decimal reductions (Dr) that are
possible at a given location. The 9 locations in Windrow 18
have been analysed in this manner.

Figures 6.2 through 6.4 show the Dr at each location
over the one week composting period. Most locations have Dr

occurring throughout the composting period. Locations 4, 6,

 

242

 

 

 

 

 

c m
2
u
3 no —
'o
o
m
o 50 ‘
1.4
0'
o
a-‘

40-
‘U
c
a
1. so 1
U
C
O 20 -
h
3
3 '0 LRZ
o ' LR3
2' LRI
g 0 LR3 LR]. R2

I U r U I r
O 2 4 O 8
Time (Days)

Figure 6.2: Windrow lB temperature and log reduction
at Locations 1, 2, and 3.

 

 

 

 

 

C 70
O
"‘ T5
‘J
U
'3 60 -
m T6
m

50 —
3 v
(7'
3

4O -« 7:.
‘U
I:
(U
U
o
1. LR6
q: 20 -4
h
D
u
2 1O -(
3
Q)
E" O l I r 1-l I I I

0 2 4 O 8

Time (Days)

Figure 6.3: Windrow lB temperature and log reduction
at Locations 4, 5, and 6.

 

 

243

 

 

 

 

 

r: 70
O
or.
3 79
'8 00 -
Q . T10
I:
3 so '1 .ar-f' T8
0"
.9. M
'0 7° "
r:
to
8 “’5
O
:3 1 -
:1
U
«I
:3 IO .1 LR9
O.
5 LR8 L110
(— o — fi‘ 7 - t“
I I I I T I I
O 2 4 0 I
Time (Days)

Figure 6.4: Windrow lB temperature and log reduction
at Locations 8, 9, and 10.

 

244

8, and 10 show very low values. Figure 6.5 summarizes the
Dr at all 9 locations for the entire composting period.
Even Location 4, which has the lowest Dr, still has a total

of 380 over the seven day period.

380 1180
3870 7480 3250
1890 3380 1310 510

Figure 6.5: Decimal reductions in Windrow 18 by
location.

Calculations of total Dr do not tell the entire story,
however. The effect of micro-organism growth is not
included. Problems with unheated, anaerobic zones exist.
The effect of mixing on organism survival is similarly not
covered. Both of these problems will now be discussed.

Assuming that an nonreacting volume is represented by a
sphere, the amount of time that it would take for the
temperature at the center of the sphere TO to heat to 90
percent of the surrounding mass is calculated. The

dimensionless time is given by the Fourier ratio (F0)

F0 = (k t / 0 cp r2) (6.6)
where Fo = Fourier ratio, dimensionless
k = thermal conductivity, W / m-s
t = time, s
p = density, kg/m3
c = specific heat, Kj / kg °C

radius, m

From the Heisler charts (Holman, 1981), the value of the

 

245
Fouier number at a dimensionless temperature of 0.9 is 0.3.
This is substituted into Equation 6.6 and the equation
solved for the time t:

t = 0.3 p c R2 / k (6.7)

P

Selected values of thermal and physical properties covering
the range encountered in Windrow 18 are substituted into the
above equation and the effect of radius is examined.

Examination of Table 6.3 indicates that unreactive
spheres with radii of 5 cm would require less than 2 hours
to heat up to 90 % of the surrounding material. Spheres
with a radius of 20 cm would take slightly more than a day
for the same degree of heating. Large anaerobic compost
balls are not expected in DMS solids composting, but it
appears that even fairly large spheres could be heated
adequately.

Thermal inactivation during composting may also be
limited because of non-uniform heat distribution followed by
mixing of the compost. Cold pockets could allow pathogenic
micro-organisms to survive or regrow. Mixing would then
redistribute the organisms throughout windrow or bedding
material.

Haug (1980) described the effects of thermal
inactivation of pathogens in a two zone windrow. One zone
is assumed to have sublethal time-temperature profiles and
the other is assumed to have lethal conditions. The
windrows are turned at time intervals of At and thoroughly

mixed so that a random redistribution of compost occurrs.

 

246

 

Table 6.3: Temperature penetration time as a function
of clump diameter and thermal properties
(6 = 0.9).
Thermal Cond. (W/m—C): 0.301 0.174

Bulk Density (Kg/m“3): 696 428
Specific Heat (KJ/Kg-C): 3.51 3.47

 

 

 

 

Radius Penetration Time Radius Penetration Time
(cm) (hours) (days) (cm) (hours) (days)
0.5 0.02 0.001 0.5 0.02 0.001
1.0 0.07 0.003 1.0 0.07 0.003
5.0 1.69 0.070 5.0 1.78 0.074
10.0 6.76 0.282 10.0 7.11 0.296
15.0 15.22 0.634 15.0 16.00 0.667
20.0 27.05 1.127 20.0 28.45 1.185
25.0 42.27 1.761 25.0 44.46 1.852
50.0 169.09 7.045 50.0 177.82 7.409
100.0 676.35 28.181 100.0 711.28 29.637
Thermal Cond. (W/m-C): 0.133 0.071

Bulk Density (Kg/m‘3): 288 164
Specific Heat (KJ/Kg-C): 3.48 3.47

 

 

 

Radius Penetration Time Radius Penetration Time
(cm) (hours) (days) (cm) (hours) (days)
0.5 0.02 0.001 0.5 0.02 0.001
1.0 0.06 0.003 1.0 0.07 0.003
5.0 1.57 0.065 5.0 1.67 0.070
10.0 6.28 0.262 10.0 6.68 0.278
15. 14.13 0.589 15.0 15.03 0.626
20.0 25.12 1.047 20.0 26.72 1.113
25.0 39.25 1.635 25.0 41.75 1.739
50.0 156.99 6.541 50.0 166.98 6.958
100.0 627.97 26.165 100.0 667.93 27.831

 

247 ,

The resulting equation is:

nt = no (£1 + fh expL"kd AtJN ) (6.8)
where nt = number of organisms surviving
no = number of organisms initially present
fl = fraction of composting material in the low-
temperature, sublethal'zone
fh = fraction of composting in the high-
temperature zone
At = time interval between windrow turnings
kd = thermal death coeffiecient
N = number of windrow turnings
BDd fl + fh = 1

Various assumptions about inital pathogen population
levels, lethal zone fractions and time temperature thermal
death factors are evaluated. Table 6.4 presents the results
of such an analysis.

Table 6.4: Effect of initial decimal reduction, fraction

of lethal temperature and thermal death
coefficent on the number of windrow turns

 

 

required.
Number of Turns
Rd A t f1 / fh Dr
-2 -3 -4 -5 —6 -7
a 0.11 2 3 4 5 6 7
a 0.25 3 5 6 8 9 10
m 0.50 7 10 14 17 20 24
m 1.00 8 12 16 20 24 27

 

A kd At value of infinity is assumed for this analysis
because of the long one week residence times. The kd At
value used in the previous time-temperature curve analysis

was approximately 1.5 decimal reductions for a 15 minute

248

period. A range of fl/fh values was used to correspond to
different levels of the lethal fraction. The lethal
fraction was based on the assumption that each temperature
measurement location represented an equal volume in the
composting windrow. The 1.0 value results from the
assumption that 3 days at 55 C are necessary for safe
composting. The 0.25 and 0.11 values for fl/fh correspond
to the assumption that temperatures over 50 C resulted in
thermal inactivation in Windrows l and 2, respectively.

From the results of the analysis presented in Table
6.4, it is apparent that a large number of turns are
necessary to achieve thermal destruction by complete mixing
of the windrow. If 5 Dr are desired, 5 turnings are
required even under the most optimistic assumptions. This
analysis would predict that the 3 turnings result in at best
3 Dr'

A number of factors contribute to potential errors in
the number of turns analysis. More accurate measurement of
the lethal areas improves accuracy. Data developed in this
study could provide a first approximation of lethal areas.
However, it should be remembered that temperatures in the
toe area and on the surface were not adequately monitored.

A major confounding factor is the observation that the
degree of substrate consumption could reduce the ability of
a windrow to self heat on successive turnings. Finally, the
effect of pathogen growth, both at optimal temperatures and

on reduced substrates,is not directly included.

249

Analysis of time-temperature patterns in the compost
windrows indicated that there were regions that would have
significant reductions of coliform organisms due to long
periods of high temperature. There were also windrow
regions, however, that had episodes that were in the optimal
temperature range for coliform growth. The locations of
both regions changed with time, windrow size and windrow
shape. The rationale for an improved heat and mass transfer

model to predict windrow temperatures is justified.

 

6.5: Application of Com ostin to Dairy Manure Solids for
Production gI—Bedding MageFTaI

 

 

At this point I find it hard to advise dairy operators
on how best to handle the manure solids, because I collected
no microbial evidence to support the deductions based on the
time-temperature curves. From a consideration of the time-
temperature history alone, I would recommend that dairy
operators using this type of windrow composting system
should not mix the windrow immediately before using the
composted dariy manure solids as bedding material.

Based on the results of the time-temperature curve
analysis, I would suggest that at the end of a three week
composting period material from the toe area and the
unreacted windrow interior should be set aside and the hot
composted material should be mixed and used for bedding.

The material set aside should be discarded or recycled back
into the first week's composting material.

In the long run, enhanced natural or forced convection

 

 

250
may prove to be the methods of choice because of the greater
control offered by the forced convection, and the higher
composting rate and smaller land requirements of both when
compared with a windrow composting system.

Microbial evidence that confirms conclusions based on
time-temperature data should receive high research priority.
If this evidence becomes available, research to optimize
pathogen and substrate reduction in natural or forced
convection systems should receive priority over further
windrow studies.

6.6: Prelimipary Assessment pf Heat and Mass Transfer
Models For Composting Dairy Manure Solids

 

 

 

 

 

Without conducting a detailed analysis, the complex
unreacted core model discussed earlier appears to describe
the observed behavior of this type of composting system very
well. The chemical engineering literature contains a large
body of information on this subject. This should prove to
be a fruitful avenue of exploration to understand the
behavior of this type of composting system.

As a first approximation of a heat and mass transfer
model, a diffusion based model with spatially varied bulk
density and constant moisture content should be considered.
The effect of the latent heat of evaporation would be
contained in a non-linear keff term. Respiration modeling
should include the effects of substrate, oxygen
concentration, and temperature on micro-organism growth and
inactivation. An accurate unsaturated compaction model for

the determination of the spatially varied bulk density,

 

 

251

volumetric moisture content and free air space variables
that affect heat and mass transfer is essential.

A second generation model that explicitly includes
moisture generation and transport relationships would
provide a theoretically more accurate picture of windrow
processes. Longer term studies would require an unsaturated
consolidation model to determine windrow bulk density. The
effect of composting on particle size and structural
stability should be included in such a model.

Due to observed differences in the penetration depth of
the treated dairy manure solids, a model that includes both
natural convection and diffusion may be the only acceptable
alternative. Due to the formidable requirements of applying
this type of model to a windrow, it is suggested that such a
model be attempted only after the previously mentioned

diffusion based models prove to be inadequate.

CHAPTER 7

CONCLUSIONS

1. The probe design and methods of determining temperature
(5 a :1.5°C), gravimetric moisture content (s = t 2.0 % wb),
and wet bulk density (CV = 11 %) had acceptable levels of
accuracy. Gas concentration determinations were subject to
errors caused by gas influx into storage containers. With
the use of an empirical correction equation, gas
concentrations were accurate to within :2 percent. Volatile
solids, whether expressed on a dry or ash basis, had

unacceptable measurement errors (sd = 7 %).

2. The assumption of constant physical and thermal
properties used in previous models is not warranted in all
cases. Bulk density, volumetric moisture content and the
degree of water saturation increase with depth. Porosity,
free air space and the void ratio decrease with depth.
Spatial changes in the moisture content (% wb), volatile
solids content (% db) and calculated specific heat were not
observed. Since many heat and mass transfer properties are
related to bulk density, volumetric moisture content and
free air space, they should be considered spatially varied.

With the methods of measurement used in this experiment,

252

 

 

253
there was no evidence indicating that bulk density varied
with time for this type of compost substrate and windrow

system-

3. The dairy manure solids behaved similarly to other
composting substrates. Differences exhibited by dairy
manure solids composting included the very high moisture
contents (between 76 and 81 % wb) and low maximum
temperature of 68°C. High moisture content composting
occurred because of the high free air spaces caused by dairy
manure solid particles. The low temperature could be
caused by small windrow sizes, high moisture contents, or
long-term substrate limitations. Experimental evidence
points toward high moisture content or substrate
limitations. Unlike sewage sludge and garbage composting,
the expression of moisture content as percent dry basis and
volatile solids on a percent ash basis is necessary to

detect changes in these parameters.

4. The overall course of the composting reaction behaved
like an unreacted core with a reactant being transported
into a reactive mass. The observed composting reaction was
more complex than a classic unreacted core model as evidenced
by the lack of a sharp interface between the spent and fresh
solid zones and because of a fast exothermic reaction that

produced temperature gradients in the windrow.

5. An estimate of the magnitude of the Thiele modulus was

made. The Thiele modulus was not constant throughout the

 

 

 

 

254
windrow.

6. Analysis of time-temperature patterns in the compost
windrows indicated that there were regions that would have
significant reductions of coliform organisms due to long
periods of high temperature. There were also windrow
regions, however, that had periods of temperature that were
in the optimum range for coliform growth. The locations of
both regions changed with time, windrow size, and windrow

shape.

7. Dairy operators using a windrow compost system should
not mix the entire windrow immediately before using the
composted dairy manure solids as bedding material. A safer
method of handling a three week old windrow of composted
dairy manure solids is to set aside material from the
windrow toe and central unreacted core for recycling or
disposal. The remaining hot composted material can be used

for bedding.

8. As a first approximation, a heat and mass transfer
model of the dairy manure solids windrow composting process
should have the following characteristics:

Diffusion-based

Spatially varied bulk density, predicted by an
unsaturated compaction model

Constant moisture content

Respiration modeled with terms accounting for the
effect of substrate consumption, temperature and
free air space oxygen concentration.

 

 

CHAPTER 8

DIRECTIONS FOR FUTURE RESEARCH

8.1: Experimental Methods

1. Bulk Density. The in situ bulk density of the

 

composting material is a critical parameter that must be
accurately measured in any futher work. Together with
moisture content, it affects a number of heat and mass
transport processes in the compost windrow. Two areas
warrant particular study: (a) methods of in-situ
measurement of bulk density, and (b), the spatial variance
of bulk density as a function of the methods of mixing,
placement and turning.

a. In-situ Measurements. There is little discussion

 

of insitu bulk density measurements in the composting
literature. The soil and peat sampling literature
provided the most fruitful sources of information that
lead to the adoption of the bulk density sampling
method used in this study. While the rotary corer
method gave good repeatability and appeared to give
good estimates of bulk density, this needs to be
further substantiated. In particular, the effect of
drill speed, the possible inclusion or exclusion of dms

material in the sample, the effect of methods, rates

255

 

 

256
and vibration of insertion and the effect of dms
particle size and shape on the effective height and

radius should be examined.

b. Spatial Variance pf Bulk Density. Once accurate

 

methods for in-situ bulk density measurement have been
developed, the spatial variance of bulk density as
affected by compost substrate, methods of mixing,
placement and turning should be investigated. Other
investigators have observed that tremendous variations
in bulk density beyond those expected due to compost
compaction can be present in windrows. Variance
estimates the variability are particularly necessary
because of the difficulty in taking many bulk density
samples during an experiment without affecting heat and

mass transfer in the experimental windrow.

2. Moisture Content. Moisture contents expressed on a dry

 

basis give more realistic indications of moisture content
changes than wet basis moisture content. The moisture
content measurement technique used in this experiment was
very time consuming and did not provide frequent
measurements. A quicker method could greatly reduce
experimental efforts in this area. The development of real
time moisture content determinations could complement
temperature data measurements. This becomes more crucial in
1Windrow environments with enhanced natural convection or

f0 rced aeration where more rapid moisture content changes

 

257

can be expected to occur.

3. Substrate. While volatile solids is a commonly used

 

parameter in engineering studies of waste treatment systems,
it is a rather crude measurement of substrate consumption.
If volatile solids are to be measured in future studies,
they should be expressed on an ash solid basis. Future
studies should include additional substrate measurements in
order to more completely reflect substrate degradation.
Levels of amino acids, amino sugars, nitrate and ammonia
nitrogen, hemicellulose, cellulose and lignin have been used

to measure the degradation of other compost substrates.

8.2: Windrow Compost Processes

 

1. Respiration. The respiratory activity of micro

 

organisms on dms needs to be better understood and
quantified. It is doubtful that research on other
substrates such a sewage sludge and ground garbage will be
transferable beyond providing the general shape of
respiration equations. If the dms respiration rate is not
quantified, the difficulty in obtaining reasonably accurate
estimates of respiration as well as dms heat and mass
transfer parameters is greatly increased. The effect of
temperature, oxygen and moisture content on respiratory
activity needs to be examined. The consumption of various

substrate components needs to be measured concurrently.

2. Moisture Content. If moisture content is to be included

 

in a compostwindrow model, the role of water in the

 

 

 

258

composting process must be much better understood. Areas of
investigation include the degree of free vs bound water on
dms particles as a function of particle size and moisture
content; the role of dms compressibility on the extrusion of
water from particles to void spaces: and the quantification
of water lost to seepage from the windrow base and
evaporation from windrow surfaces. The use of

tensiometers and relative humidity probes should be

evaluated for this purpose.

8.3: Compost Heat and Mass Transfer Modeling

 

 

 

1. The development of an unsaturated model of dms windrow
compaction is essential to the development of an in-

situ bulk density sampler and to future modeling efforts.
Since bulk density is a key parameter in predicting several
heat and mass transfer properties, such a model must be

available to predict bulk density.

2. Once the model described above is available, work should
proceed with the development and validation of a heat and
mass transfer model along the lines suggested in Chapter 7.
A second generation model that explicitly includes moisture
content would provide a more theoretically sound model. If
longer term studies on windrow composting are to be
conducted, an unsaturated consolidation model needs to be
developed. The effect of the composting process on particle
size distributions and dms particle compressibility needs to

be investigated and included in the model.

 

 

259

3. Due to the observed differences in the penetration depth
of the treated dms, a model that includes both diffusion and
natural convection may be the only acceptable alternative.
Because of the formidable requirements of applying this type
of model to a windrow, it is suggested that such a model be
attempted only after the previously mentioned diffusion
based models prove to be inadequate. Model development
should procede from simpler l and 2 dimensional cases before
expanding to the complex 3 dimensional situation in dms

windrows.

APPEND I CES

 

~n-r‘,‘ 1%.- -._r1 ...—7' :1

APPENDIX A

TEMPERATURE MEASUREMENT ERROR CALCULATIONS

APPENDIX A

TEMPERATURE MEASUREMENT ERROR CALCULATIONS

The total thermocouple temperature measurement error is
the sum of the errors in the reference junction temperature,
the thermocouple output, the thermocouple voltage
measurement, and the linearization error (difference between
standard and polynomial approximations).

A Fenwal Electronics UUTSlJl thermistor mounted in the
center of the analog input terminal strip measures the 21X
panel temperature. According to the Campbell literature,
the "worst case" example adds a : 0.3°C error in the
range of -35°C to +50°C. In CSI's experience, the overall
accuracy is typically better than t 0.2°C. The thermistor
error increases drastically if temperatures greater than
50°C or less than -35°C are imposed on the CR21X.

Differences between the thermistor and actual reference
junction temperature will become errors in measurement. The
terminal cover provides a shield to reduce the gradient
along the terminal strip. The CR21X manual (Campbell
Scientific, 1985) gives an example in which the 21x was
brought from -25°C to an ambient temperature of 20°C.

After 100 minutes the temperature gradient between the air

and the battery is still 15°C, but the deviation from the

260

261
measured temperature along the reference strip is only
0.3°C. This taken as the maximum error that could occur in
a field situation with proper radiation shielding.

The ANSI standards for thermocouple limits of error are
given in the National Bureau of Standards Monograph 125
(1974). For copper constantan thermocouples, they are as
follows:

Table A.1: Limits of Error for Thermocouple Wire
(Reference Junction at 0°C)

 

Limits of Error

 

Thermocouple Temperature (Whichever is greater)
Type Range °C Standard Special
T -200 to 0 1 l.0°C or 1.5 %
0 to 350 t l.0°C or 0.75 % t 0.5°C or
0.4 %

 

In order to quantitatively evaluate thermocouple error when
the reference junction is not fixed at 0°C, the limits of
error for the Seebeck coefficient (slope of thermocouple
voltage vs temperature curve) for the various thermocouples
must be available. According the CSI, if this information
is missing, a reasonable approach is to apply the percentage
errors, with perhaps 0.25 % added on, to the difference in
temperature being measured by the thermocouple. The
temperatures in this experiment will range from 15°C to
70°C. The error due to the thermocouple would be l.0°C.
The accuracy of a 21X voltage measurement is specified

as 0.1 % of the full scale range being used to make the

262

measurement. The error in the temperature due to inaccuracy
in the measurement of the thermocouple voltage is worst at
temperature extremes, where a relatively large scale is
necessary to read the thermocouple output. .In the
environmental temperature range with voltage measured on an
appropriate scale, error in temperature due to the voltage
measurements is a few hundredths of a degree (Anonymous,
1985).

Voltage to temperature conversions are accomplished
using a proprietary 6th order polynomial. The limit of
error on 21x thermocouple output linearization, relative to
NBS Standards, is 10.001 over a range of -100°C to 100°C.

When external reference junction boxes are used errors
can arise if the reference junction temperature is outside
of the linearization range. The reference temperature
compensation range and linearization error relative to the
NBS Standards for type T thermocouples are -100°C to 100°C
and t0.001°C.

The sources of error discussed above are summarized in

Table A.2.

263

Table A.2: Summary of the sources of error in
thermocouple measurements.

 

Source Error (°C) Percent of Total Error

 

Reference junction

temperature 0.5 32.2
Thermocouple output 1.0 64.4
Voltage measurement 0.05 3.2
Reference linearization 0.001 0.1
Output linearization 0.001 9:1

 

Total Error 1.552 100.0

 

APPENDIX B

SAMPLE THERMOCOUPLE WATER BATH
CALIBRATION DATA

APPENDIX B

SAMPLE THERMOCOUPLE WATER BATH
CALIBRATION DATA

 

 

Table B.l Water bath thermocouple calibration-~Run 1.
Thermocouple

Hour:Mi 1 2 3 4 5 6 7
1418 -6999 27.86 27.84 27.86 27.86 27.84 27.86
1419 -6999 27.86 27.86 27.87 27.85 27.87 27.86
1420 -6999 27.84 27.82 27.85 27.82 27.84 27.83
1421 -6999 27.82 27.80 27.81 27.80 27.80 27.80
1422 -6999 27.75 27.74 27.75 27.73 27.76 27.75
1423 —6999 27.69 27.68 27.70 27.70 27.70 27.69
1424 -6999 27.64 27.64 27.67 27.65 27.68 27.66
1425 -6999 27.60 27.61 27.62 27.61 27.63 27.62
1426 -6999 27.56 27.57 27.58 27.57 27.59 27.59
1427 -6999 27.53 27.53 27.53 27.52 27.55 27.54
1428 -6999 27.50 27.49 27.50 27.49 27.51 27.51
1429 -6999 27.46 27.46 27.47 27.47 27.47 27.48
1430 -6999 27.41 27.42 27.42 27.42 27.43 27.42
1431 -6999 27.38 27.37 27.38 27.38 27.39 27.38
1432 27.35 27.34 27.33 27.35 27.33 27.36 27.34
1433 -6999 27.30 27.30 27.31 27.30 27.32 27.31
1434 27.26 27.26 27.24 27.27 27.26 27.25 27.26
1435 27.22 27.22 27.20 27.23 27.21 27.22 27.23
1436 27.19 27.17 27.17 27.19 27.18 27.19 27.19
1437 27.12 27.12 27.11 27.12 27.12 27.13 27.13
1438 27.08 27.08 27.08 27.10 27.08 27.09 27.09
1439 27.03 27.04 27.03 27.05 27.03 27.05 27.06
1440 27.03 27.04 27.03 27.06 27.03 27.04 27.06
1441 27.40 27.46 27.42 27.44 27.39 27.41 27.45
1442 27.69 27.71 27.70 27.67 27.71 27.73 27.68
1443 27.66 27.66 27.66 27.67 27.66 27.68 27.68
1444 27.62 27.62 27.61 27.63 27.62 27.64 27.63
1445 27.57 27.58 27.57 27.58 27.57 27.59 27.58
1446 27.52 27.53 27.51 27.53 27.53 27.54 27.54
1447 27.49 27.49 27.49 27.50 27.48 27.50 27.51

Ave. 27.35 27.48 27.48 27.49 27.48 27.49 27.49

Var. 0.05 0.06 0.06 0.06 0.06 0.06 0.06

 

78.1

265

Table 8.1 (Cont.).

 

 

Thermocouple

8 9 10 ll 12 13 14 15
27.84 27.89 27.91 27.91 27.94 27.96 27.94 27.94
27.85 27.89 27.92. 27.92 27.94 27.95 27.95 27.94
27.84 27.87 27.88 27.89 27.92 27.92 27.92 27.91
27.80 27.84 27.85 27.87 27.88 27.89 27.89 27.88
27.75 27.78 27.80 27.81 27.84 27.84 27.85 27.84
27.68 27.73 27.76 27.77 27.79 27.80 27.80 27.80
27.66 27.70 27.72 27.73 27.77 27.77 27.78 27.78
27.62 27.66 27.69 27.70 27.73 27.73 27.74 27.75
27.58 27.62 27.65 27.66 27.69 27.69 27.72 27.72
27.53 27.58 27.61 27.62 27.64 27.66 27.66 27.67
27.50 27.55 27.57 27.58 27.61 27.62 27.63 27.64
27.47 27.52 27.54 27.55 27.57 27.59 27.61 27.60
27.41 27.46 27.48 27.50 27.52 27.54 27.54 27.55
27.38 27.42 27.45 27.47 27.48 27.49 27.51 27.51
27.33 27.38 27.37 27.38 27.44 27.45 27.45 27.45
27.30 27.35 27.37 27.35 27.40 27.41 27.43 27.40
27.26 27.30 27.32 27.33 27.36 27.38 27.38. 27.39
27.21 27.26 27.29 27.30 27.32 27.35 27.34 27.35
27.19 27.22 27.25- 27.26 27.28 27.29 27.31 27.31
27.12 27.17 27.18 27.19 27.23 '27.24 27.27 27.26
27.09 27.12 27.15 27.16 27.18 27.21 27.22 27.22
27.05 27.09 27.11 27.12 27.15 27.16 27.18 _27.19
27.05 27.09 27.08 27.08 27.13 27.13 27.15 27.17
27.45 27.45 27.40 27.37 27.41 27.41 27.45 27.45
27.69 27.74 27.71 27.73 27.76 27.77 27.81 27.80
27.66 27.71 27.73 27.74 27.78 27.78 27.81 27.80
27.63 27.67 27.68 27.69 27.73 27.74 27.76 27.76
27.57 27.61 27.64 27.65 27.68 27.69 27.71 27.71
27.54 27.58 27.60 27.61 27.64 27.65 27.68 27.68
27.49 27.53 27.55 27.57 27.60 27.60 27.63 27.64
27.48 27.53 27.54 27.55 27.58 27.59 27.60 27.60
0.06 0.06 0.06 0.06 0.06 0.06 0.05 0.05

 

 

266

Table 8.1 (Cont.).

 

 

Thermocouple

16 17 18 19 20 21 22 23
27.91 27.91 27.99 27.94 27.94 27.94 27.91 27.91
27.95 27.92 28.00 27.95 27.93 27.92 27.94 27.92
27.91 27.91 27.97 27.93 27.91 27.90 27.91 27.91
27.91 27.88 27.96 27.91 27.89 27.87 27.89 27.89
27.85 27.84 27.91 27.86 27.84 27.83 27.84 27.83
27.82 27.80 27.87 27.82 27.80 27.80 27.80 27.79
27.79 27.78 27.85 27.80 27.77 27.76 27.77 27.76
27.75“ 27.75 27.81 27.75 27.75 27.72 27.74 27.72
27.72 27.72 27.79 27.72 27.70 27.69 27.70 27.69
27.69 27.67 27.74 27.69 27.66 27.65 27.65 27.63
27.65 27.63 27.71 27.64 27.64 27.62 27.62 27.60
27.62 27.60 27.66 27.63 27.59 27.58 27.59 27.57
27.56 27.56 27.63 27.56 27.56 27.53 27.53 27.52
27.53 27.51 27.59 27.53 27.51 27.49 27.50 27.47
27.47 27.47 27.53 27.50 27.47 27.46 27.46 27.44
27.45 27.42 27.49 27.45 27.45 27.41 27.43 27.41
27.40 27.39 27.46 27.41 27.39 27.38 27.38 27.36
27.37 27.35 27.43 27.38 27.36 27.34 27.34 27.32
27.32 27.31 27.38 27.35 27.33 27.31 27.32 27.28
27.28 27.27 27.35 27.31 27.29 27.25 27.26 27.24
27.26 27.23 27.31 27.27 27.24 27.22 27.22 27.21
27.21 27.20 27.27 27.23 27.21 27.18 27.18 27.17
27.21 27.17 27.27 27.22 27.19 27.18 27.17 27.15
27.56 27.46 27.57 27.64 27.59 27.56 27.57 27.54
27.84 27.81 27.90 27.90 27.87 27.86 27.85 27.84
27.84 27.82 27.90 27.85 27.83 27.81 27.81 27.79
27.79 27.78 27.86 27.81 27.78 27.76 27.76 27.74
27.75 27.74 27.81 27.76 27.73 27.71 27.72 27.70
27.70 27.70 27.78 27.70 27.68 27.68 27.66 27.65
27.67 27.65 27.73 27.67 27.64 27.63 '27.63 27.60
27.63 27.61 27.68 27.64 27.62 27.60 27.61 27.59
0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05

 

 

 

267

Table 8.1 (Cont.).

 

 

Thermocouple

24 25 26 27 28 29 30 31
27.94 27.91 27.94 27.91 27.91 27.84 31.20 31.37
27.95 27.93 27.92 27.89 27.88 27.84 31.17 31.39
27.91 27.90 27.89 27.88 27.85 27.82 31.17 31.34
27.89 27.87 27.88 27.84 27.82 27.79 31.25 31.44
27.84 27.82 27.82 27.78 27.77 27.73 31.28 31.53
27.78 27.78 27.77 27.72 27.71 27.67 31.26 31.47
27.75 27.74 27.75 27.69 27.68 27.63 31.24 31.47
27.71 27.70 27.71 27.65 27.64 27.60 31.31 31.50
27.68 27.66 27.66 27.60 27.58 27.54 31.25 31.45
27.64 27.62 27.62 27.56 27.55 27.50 31.24 31.40
27.60 27.58 27.58 27.54 27.51 27.47 31.21 31.41
27.57 27.55 27.56 27.50 27.48 27.45 31.25 31.45
27.52 27.51 27.51 27.45 27.44 27.39 31.32 31.50
27.47 27.45 27.47 27.42 27.39 27.35 31.25 31.51
27.45 27.43 27.44 27.37 27.35 27.31 31.35 31.55
27.41 27.39 27.39 27.34 27.31 27.28 31.30 31.58
27.35 27.34 27.34 27.28 27.28 27.22 31.46 31.66
27.33 27.30 27.30 27.25 27.24 27.19 31.44 31.69
27.28 27.28 27.27 27.20 27.19 27.14 31.51 31.64
27.24 27.21 27.21 27.16 27.13 27.09 31.46 31.55
27.20 27.17 27.17 27.12 27.10 27.06 31.49 31.55
27.15 27.14 27.14 27.07 27.06 27.02 31.46 31.49
27.15 27.13 27.12 27.10 27.07 27.01 31.48 31.51
27.54 27.53 27.53 27.50 27.47 27.44 31.48 31.54
27.83 27.82 27.80 27.76 27.74 27.70 31.51 31.58
27.78 27.77 27.76 27.70 27.69 27.64 31.51 31.58
27.74 27.71 27.71 27.65 27.63 27.60 31.51 31.56
27.68 27.67 27.67 27.61 27.59 27.56 31.56 31.61
27.64 27.61 27.62 27.57 27.54 27.50 31.59 31.64
27.60 27.58 27.57 27.52 27.50 27.47 31.51 31.57
27.59 27.57 27.57 27.52 27.50 27.46 31.37 31.52
0.06 0.06 0.06 0.06 0.06 0.06 0.02 0.01

 

 

 

268

Table 8.1 (Cont.).

 

 

32 Ave. Var. 8.0.
31.59 27.91 0.002 0.040
31.61 27.91‘ 0.002 0.040
31.61 27.88 0.002 0.039
31.63 27.86 0.002 0.043
31.63 27.81 0.002 0.046
31.63 27.76 0.003 0.053
31.64 27.73 0.003 0.058
31.66 27.69 0.003 0.058
31.66 27.65 0.004 0.064
31.66 27.61 0.004 0.063
31.68 27.57 0.004 0.062
31.70 27.54 0.004 0.061
31.70 27.49 0.004 0.061
31.71 27.45 0.004 0.061
31.72 27.41 0.004 0.060
31.73 27.37 0.003 0.058
31.74 27.33 0.004 0.062
31.75 27.29 0.004 0.064
31.76 27.26 0.004 0.064
31.76 27.20 0.005 0.070
31.77 27.17 0.005 0.068
31.78 27.13 0.005 0.070
31.78 27.12 0.004 0.065
31.79 27.48 0.005 0.069
31.80 27.78 0.005 0.068
31.80 27.75 0.005 0.070
31.81 27.70 0.005 0.070
31.82 27.66 0.005 0.069
31.83 27.61 0.005 0.072
31.83 27.57 0.005 0.070
31.72 27.56 .00 0.06

0.01 0.06 .00 .00

 

 

 

APPENDIX C

EFFECT OF SEQUENTIAL GAS SAMPLES AND LOCATION ON CARBON
DIOXIDE AND OXYGEN CONCENTRATION

 

 

APPENDIX C

EFFECT OF SEQUENTIAL GAS SAMPLES AND LOCATION ON CARBON
DIOXIDE AND OXYGEN CONCENTRATION

 

 

Table C.1: Effect of sequential gas samples and location
on C02 and 02 concentrations.
Sequential
Volume Location 1 Location 5 Location 8 Location 10
Withdrawn C02 02 C02 02 C02 02 CO2 02
5 0.63 19.54 0.80 18.54 0.27 16.64
15 7.85 12.64 9.02 10.17 3.40 16.06 16.38 5.82
25 7.80 13.01 15.95 4.65 4.40 16.40 19.08 3.96
35 7.82 13.11 16.87 3.81 4.00 15.73
45 7.61 12.75 15.73 5.63 4.22 16.33 18.63 4.99
55 7.37 12.95 15.44 2.96 4.03 16.30 19.32 4.06
65 7.78 13.49 15.66 4.36 20.05 3.49
75 7.94 13.23 16.04 4.61 2.62 16.87 19.12 3.87
AVG * 7.74 13.03 15.95 4.34 4.16 16.19 19.24 4.07
SD 0.18 0.27 0.46 0.82 0.16 0.27 0.46 0.50
CV 2.29 2.05 2.86 18.87 3.86 1.65 2.41 12.18

 

 

 

*Statistics are calculated from the linear
portion of the data.

 

APPENDIX D

GAS CONCENTRATION CORRECTION FACTOR CALCULATIONS

APPENDIX D

GAS CONCENTRATION CORRECTION FACTOR CALCULATIONS

The procedures used to determine the gas concentration
correction factor due to withdrawal of samples from the
"venoject" sample containers with a syringe prior to
injection into the gas chromatograph were described in
general in Chapter 4. This appendix describes in more
detail the procedures and calculations used to arrive at
these results.

The experiment consisted of evacuating six 3 cc
venojects and filling them to an initial pressure of 1.5
atmospheres. This duplicated the inital conditions of
sample storage in the main experiment. 1.05 cc withdrawals
were made from each venoject and injected into the gas
chromatograph to check for the oxygen concentration. The
lengths were recorded and oxygen concentrations were
calculated based on a four point calibration curve. The
presence of any oxygen was assumed to be due to oxygen
present in the syringe tip before sample withdrawal,
diffusion of oxygen into the syringe needle after withdrawal
or the flow of ambient air into the syringe due to a
pressure difference. The initial lengths of the gas

chromatograph traces and the calculated oxygen

270

 

271
concentrations are shown in Table D.l and Table D.2. The
calibration curve is given at the bottom of Table D.2.

The internal pressure of the sample container and
syringe were calculated for several pressures, sample
container volumes and syringe volumes. The equation used in
this calculation was

Pf = (Pi * vC)/(vC + vs) (0.1)

where Px = internal pressure of x, atmospheres

V

volume of y, cubic centimeters

final

initial

sample container
sample syringe

MOI-"H: *4

For example, if the initial conditions were 1.5 atmospheres,
3 cc and 1.05 cc for the initial pressure, container volume
and sample volume, respectively, the final internal pressure
is

Pf = 1.50 * 3.00 /(3.00 + 1.05) = 1.11 atmospheres
The results of these calculations are shown in Table 0.3.

Plotting the calculated internal pressure for each
sample withdrawn against the measured oxygen concentration
indicated two linear segments. These were shown in Figure
4.9 of the text.

Influx was calculated by assuming that the initial and
atmospheric oxygen concentrations were 0.0 and 20.9 percent,
respectively. Equation D.2 was used:

om = In * oe + F * 0a (0.2)

where 0x = oxygen concentration at x, percent

 

 

 

 

Table 0.1:

272

Gas chromatograph trace
withdrawal number.

length as a function of

 

 

 

 

 

 

With. Replication

Number 1 2 3 4 5 6 AVG STD CV
1 1.90 1.30 1.55 1.40 2.30 1.75 1.70 0.34 19.73
2 2.25 2.30 2.95 2.70 3.70 2.85 2.79 0.48 17.29
3 5.85 6.50 6.50 6.45 7.15 6.85 6.55 0.40 6.09
4 8.40 9.00 9.10 9.10 8.70 9.90 9.03 0.46 5.10
5 11.10 11.80 11.45 11.45 0.29 2.50
6 12.40 13.00 12.70 12.70 0.24 1.93

Table .2: Oxygen concentrations as a function of

withdrawal number. *

With. Replication

Number 1 2 3 4 5 6 AVG STD CV
0 0 - 0 0 0 0 0 0.00
1 1.56 0.99 1.22 1.08 1.97 1.42 1.37 0.33 23.89
2 1.92 1.97 2.66 2.39 3.49 2.55 2.50 0.52 20.97
3 6.06 6.88 6.88 6.82 7.72 7.33 6.95 0.51 7.32
4 9.37 10.19 10.32 10.32 9.78 11.43 10.24 0.63 6.16
5 13.11 14.12 13.61 13.62 0.41 3.01
6 14.99 15.86 15.42 15.43 0.36 2.32

* Calibration equation: 02 = 0.7217 * L ‘ 1.2048

 

273

Table 0.3: Effect of initial pressure, sample container
volume, and sample size on sample container and

syringe internal pressure after sampling.

 

 

 

 

 

 

Init. Press.: 1.500 1.375 1.250 1.125 1.000
Stor. Vol.: 4.00 4.00 4.00 4.00 4.00
Sample Vol.: 0.55 0.55 0.55 0.55 0.55
SAMPLE PRESSURE AFTER SAMPLE
NUMBER (Atmospheres)
0 1.500 1.375 1.250 1.125 1.000
1 1.319 1.209 1.099 0.989 0.879
2 1.159 1.063 0.966 0.869 0.773
3 1.019 0.934 0.849 0.764 0.679
4 0.896 0.821 0.747 0.672 0.597
5 0.788 0.722 0.656 0.591 0.525
6 0.692 0.635 0.577 0.519 0.462
7 0.609 0.558 0.507 0.457 0.406
8 0.535 0.491 0.446 0.401 0.357
Init. Press.: 1.500 1.375 1.250 1.125 1.000
Stor. Vol.: 4.00 4.00 4.00 4.00 4.00
Sample Vol.: 1.05 1.05 1.05 1.05 1.05
SAMPLE PRESSURE AFTER SAMPLE
NUMBER (Atmospheres)
0 1.500 1.375 1.250 1.125 1.000
1 1.188 1.089 0.990 0.891 0.792
2 0.941 0.863 0.784 0.706 0.627
3 0.745 0.683 0.621 0.559 0.497
4 0.590 0.541 0.492 0.443 0.394
5 0.468 0.429 0.390 0.351 0.312
6 0.370 0.340 0.309 0.278 0.247
7 0.293 0.269 0.245 0.220 0.196
8 0.232 0.213 0.194 0.174 0.155

 

 

 

274

Table D.3: (cont'd.).

 

 

 

 

 

 

Init. Press.: 1.500 1.375 1.250 1.125 1.000
Stor. Vol.: 3.00 3.00 3.00 3.00 3.00
Sample Vol.: 0.55 0.55 0.55 0.55 0.55
SAMPLE PRESSURE AFTER SAMPLE
NUMBER (Atmospheres)
0 1.500 1.375 1.250. 1.125 1.000
1 1.268 1.162 1.056 0.951 0.845
2 1.071 0.982 0.893 0.803 0.714
3 0.905 0.830 0.754 0.679 0.604
4 0.765 0.701 0.638 0.574 0.510
5 0.646 0.593 0.539 0.485 0.431
6 0.546 0.501 0.455 0.410 0.364
7 0.462 0.423 0.385 0.346 0.308
8 0.390 0.358 0.325 0.293 0.260
Init. Press.: 1.500 1.375 1.250 1.125 1.000
Stor. Vol.: 3.00 3.00 3.00 3.00 3.00
Sample Vol.: 1.05 1.05 1.05 1.05 1.05
SAMPLE PRESSURE AFTER SAMPLE
NUMBER (Atmospheres)
0 1.500 1.375 1.250 1.125 1.000
1 1.111 1.019 0.926 0.833 0.741
2 0.823 0.754 0.686 0.617 0.549
3 0.610 0.559 0.508 0.457 0.406
4 0.452 0.414 0.376 0.339 0.301
5 0.335 0.307 0.279 0.251 0.223
6 0.248 0.227 0.206 0.186 0.165
7 0.184 0.168 0.153 0.138 0.122
8 0.136 0.125 0.113 0.102 0.091

 

 

275

m = measured

e = environmental

a = actual

In = influx fraction, decimal

F = original fraction, decimal = l - In

Substituting known values and rearranging, we get

In = O / 20.948 (D.3)

m
Figure 4.9 of the text shows influx for the 3cc sample

container as a function of internal pressure. Table D.4
includes the calculated average 3cc influx values (In3) for

this experiment.

 

Low error intermediate
calculations.

Table 0.4:

 

 

Sample Vol P4 02M In3 In4

Number Removed
0 0.00 1.500 0.0 0.000 0.000
1 1.05 1.111 1.3 0.066 0.049
2 2.10 0.823 2.5 0.119 0.089
3 3.15 0.610 7.0 0.332 0.249
4 4.20 0.452 10.2 0.489 0.366
5 5.25 0.335 13.6 0.650 0.487
6 6.30 ‘0.248 15.4 0.736 0.552

 

The 3 cc results were converted to those for a 4 cc

 

container by (a) converting the 3 cc fractional influx to a
volume influx and (b) calculating the fraction of 4 cc that

was represented by (a). For samples of 1.05 cc, the 3 cc

fractional value was 0.066, the volume was 3 * 0.066 =

0.198, and the 4 cc fractional volume was 0.198 / 4 =

 

276
0.0495. These results are also shown in Table D.4.

The calculated internal pressures that corresponded to
the volume withdrawn from the 3 cc sample container were
used with the 4 cc influx values in a linear regression. In
Table D.4, the pairs of numbers corresponding to the
sample numbers 0, 1 and 2 in columns 3 and 6 were used for Y
and X, respectively. The calculated pressures from the 3 cc
container were used for the same sample number from a 4 cc
container because it was the pressure conditions in the 3 cc
container that produced the calculated pressure. The
equation developed (r = 0.9996) was

In4P = -7.619 * Pf + 1.495 (0.4)

Next, a linear regression was run with the withdrawal
number and the assumed internal pressure of a 4 cc container
(P4) as x and Y respectively. The resulting equation (r =
0.9971) was

P4 = -0.1508 * S + 1.4802 (D.5)

Using Equation 0.5, P4 values for the main experiment
sample numbers were caclutaed. The results of this
calculation are shown in Table D.5. The influx into a 4 cc
container was calculated using the output of the previous
step and Equation D.4. A linear regression was performed on
the In4 as X and S as Y. The equation developed (r = 1.000)
was

In4 = 0.01977 * S + 0.002 (D.6)
This equation was used to calculate influx values as a

function of the sample withdrawal number.

277

 

 

Table 0.5: Final gas correction
calculations.

Sample Low Error High Error

Number P4 In4 P4 In4
0 1.480 0.0020 0.9868 0.0000
1 1.329 0.0218 0.8862 0.0914
2 1.179 0.0415 0.7856 0.1726
3 1.028 0.0613 0.6850 0.2588
4 0.877 0.0811 0.5844 0.3417

 

In order to develop a worst case analysis, it was
assumed that the worst situation would occur if the sample
gas leaked out of the container to the point where the
internal pressure was at one atmosphere. In this situation,
the removal of gas with the syringe would immediately
produce a relative vacuum in the syringe and influx of
ambient air would occur. It was further assumed that the
worst case was represented by the second linear section of
the pressure vs. measured oxygen curve (Figure 4.9).

The calculation procedure was similar to that of the
previous error calculation. The fractional influx was
determined and the 3 cc influxes were converted to 4 cc
influxes. As in the small error case, the calculated P3 was
assumed to be the cause of the gas transport and a linear
regression was run on In4 and P4 (= P3). The data used
corresponded to the rows with sample numbers 3 to 6 in Table
0.9976) was

D.4. The resulting equation (r =

94 = -1.160 * In4 + 0.8934 (0.7)

 

278

Since all gas under pressure was assumed to have
escaped the sample container in this worst case scenario,
the initial pressure was assumed to be 1.0 atmospheres. The
linear regression of sample numbers and internal pressure
used the values corresponding to the calculated pressures
in a 4 cc sample container initially at 1.0 atmosphere with
sequential 0.55 cc syringe samples being withdrawn. These
values are shown in Table D.3.

The equation that resulted from this regression was

In4 = -l.160 * S + 0.9868 (r = -0.9971) (D.8)
New values of the 4 cc internal pressure were calculated
using this equation. The results are shown in Table D.5.

The slope of Equation D.8 was assumed to be accurate
and the "a" term was adjusted so that at S = 0, the influx
would be zero. The resulting equation was

In4 = (P4 - 0.9868) / -l.166 (D.9)
New values of the influx were calculated using the above
equation and the results are shown in Table 0.5.

A linear regression was run on the sample number and
the In4 influxes and the following equation (r = 1.000) was
developed:

In4 = 0.0851 * S + 0.0027 (D.10)
This equation was used to assess the high error case.

For a given sample volume the effect is linear and can

be described as follows for each gas:

0 = O

a - In * Oe / F (D.ll)

m

Ca = c:m / F (v.12)

 

279
where all terms are as defined before.

Data on the sample number of each gas measurement were
available. With this, the measured concentrations,
Equations 0.6 and D.10 for the low and high influx
calculations, and Equations 0.11 and D.12 for the actual
oxygen and carbon dioxide, the actual gas concentrations
could be calculated.

In order to provide sample calculations a number of
combinations of gas concentration and sample withdrawal
numbers are shown in Table D.6. Three combinations of
carbon dioxide and oxygen concentrations with high, medium
and low levels of each gas are shown in this table.
Withdrawal numbers ranging from 1 to 3 are included. Most
sample withdrawal numbers in the study were either 1 or 2;
only a few were 3. As can be seen from the calculated
concentration columns in Table D.6, the correction equations
do adjust the concentrations in the correct direction and
larger withdrawal numbers do lead to greater influxes and
larger adjustments to the concentrations. One drawback is
that the sum of the calculated concentrations is affected by
the starting concentration and the number of withdrawals,
with lower C02 and higher withdrawal numbers lowering the
sum of the two calculated concentrations.

Two additional estimates of gas concentration
corrections were examined. The third estimate was to take
the "average" of the low and high influx error estimates.

This was intended to approximate the uncertainty that the

 

280

 

 

Table D.6: Gas concentration calculations.

Withdrawal Influx Influx Caluclated Sum

C02 02 Number Eqn C02 02
17 4 1 D.6 0.022 17.4 3.5 20.9
12 9 1 D.6 0.022 12.3 8.5 20.8
2 19 1 D.6 0.022 2.0 18.5 20.6
17 4 2 D.6 0.042 17.7 3.1 20.8
12 9 2 D.6 0.042 12.5 8.1 20.6
2 19 2 D.6 0.042 2.1 18.1 20.2
17 4 3 D.6 0.061 18.1 2.6 20.7
12 9 3 D.6 0.061 12.8 7.6 20.4
2 19 3 D.6 0.061 2.1 17.6 19.8
17 4 1 0.10 0.088 18.6 2.0 20.6
12 9 1 0.10 0.088 13.2 7.0 20.1
2 19 1 D.10 0.088 2.2 17.0 19.2
17 4 2 0.10 0.091 18.7 1.9 20.6
12 9 2 D.10 0.091 13.2 6.9 20.1
2 19 2 0.10 0.091 2.2 16.9 19.1
17 4 3 D.10 0.093 18.7 1.8 20.6
12 9 3 D.10 0.093 13.2 6.8 20.1
2 19 3 0.10 0.093 2.2 16.8 19.1
17 4 1 Average 18.0 2.8 20.8
12 9 1 Average 12.7 7.8 20.5
2 19 1 Average 2.1 17.8 19.9
17 4 2 Average 18.2 2.5 20.7
12 9 2 Average 12.9 7.5 20.4
2 19 2 Average 2.1 17.5 19.6
17 4 3 Average 18.4 2.2 20.7
12 9 3 Average 13.0 7.2 20.2
2 19 3 Average 2.2 17.2 19.4
17 4 1 C02 Correct 17.0 3.9 20.9
12 9 1 C02 Correct 12.0 8.9 20.9
2 19 1 C02 Correct 2.0 18.9 20.9
17 4 2 C02 Correct 17.0 3.9 20.9
12 9 2 C02 Correct 12.0 8.9 20.9
2 19 2 C02 Correct 2.0 18.9 20.9
17 4 3 C02 Correct 17.0 3.9 20.9
12 9 3 C02 Correct 12.0 8.9 20.9
2 l9 3 C02 Correct 2.0 18.9 20.9

 

 

 

281
exact initial pressure in the sample container was unknown.
The fourth estimate was based on the assumption that the
measured carbon dioxide value was correct and that the
oxygen concentration was equal to 20.95 % minus the measured
carbon dioxide concentration.

Increases in carbon dioxide concentration at high
measured concentrations were 1.6 % for the high influx
correction, 1.0 to 1.4 % for average influx, and 0.4 to 1.0
% for the low influx correction. Increases at low measured

carbon dioxide concentrations were less than 0.2 percent.

 

Oxygen concentration corrections achieved by the high,
average, and low influx calculations were all less than the
measured values. The largest corrections occurred at the
lower measured concentrations. The high influx correction
resulted in an approximately 2 % calculated oxygen decrease
from measured concentrations of 4 percent. The
corresponding decreases for the average and low influx
calculations were 1.2 to 1.8 % and 0.5 to 1.4 %
respectively. Corrected oxygen concentrations obtained from

the fourth method tended to be higher than the measured

 

concentrations by as much as 1 % oxygen.

The low influx correction is based on the assumption
that all gas samples were placed in the cylinders at the
same pressure and no gas leaked out. The high influx
assumes that all gas at pressures over 1 atmosphere leaked
out. As indicated in Chapter 4 of the text, the assumptions

underlying both these methods are open to question. The

282
initial pressure under which gas was inserted into the
sample containers was unknown. Preliminary experiments also
indicated that there was some leakage out of the storage
system but that uncovered samples did not decrease to 1
atmosphere pressure in the two week period before they were
analyzed. All that can be said, therefore, about the
internal sample container pressure at the start of analysis
is that it was between 1.0 and 1.5 atmospheres. The fourth
correction method was abandoned because it did not result in
the desired corrections.

The average influx calculation was chosen for use
because of the uncertainty over the internal sample pressure
at the start of analysis. This method is not without its
problems, however. In realistic conditions, the actual
influx would be governed by the low influx correction until
the internal pressure fell below 1.0 atmospheres, at which
time the high influx equation would apply. Since the
initial pressure is unknown and the number of samples that
could be withdrawn before the high influx conditions would
prevail is therefore uncertain, the average influx

calculation is the best compromise that can be achieved.

 

APPENDIX E

VOLUMETRIC GAS STANDARDS ERROR CALCULATIONS

APPENDIX E

VOLUMETRIC GAS STANDARDS ERROR CALCULATIONS

Two of the four standards used in calibrating the gas
chromatograph were made up volumetrically. The following
discussion describes the procedure that was used and the
error calculations that were made.

The standards were made using 160 cc sample bottles
with rubber caps. The capped bottles were evacuated using a
vacuum pump and then filled with pure nitrogen gas. This
procedure was repeated 3 times with the sample bottles being
left in the vacuum state. Oxygen, carbon dioxide and
nitrogen were then added by gas syringes of different sizes
to make up the desired concentrations at 2 atmospheres of
pressure. Two syringe sizes were used: 30 cc and 60 cc.
The two syringes had variances of 0.5 and 1 cc respectively.
Table E.1 summarizes the concentrations and numbers of
syringe fillings that were needed for each concentration.
The calculation for gas concentration in percent has the
form

CC = 100 * V1 / Vt (3.1)

where GC gas concentration, percent

Volume, cc
ith gas

283

284

Table E.1: Concentration and syringe use data.

 

 

 

Standard Gas Concentration Number of Syringe Uses
Number C02 02 C02 02 N2
(percent) 30 60 30 60 30 60
3 10 10 2 0 2 0 l 4
4 20 l l 1 l 0 l 4
t = Total

Applying the product rule for the propagation of errors

(See Appendix G), we have

59 = [ (aGC2/ avi’) * Si: +
(aGCZ/ avt2 ) * st: 1°-5 (3.2)

Taking the paritial derivatives of Equation E.l with
respect to Vi and Vt we get values of (lOO/Vt) and
(lOOVi/TZ) respectively. Values of the variance for Vi and
Vt are arrived at by multiplying the individual variance of
each syringe by the number of times that syringe is used.
The variance of the 10 % C02 standard can be calculated

as follows:

Vi = 2 * 0.5 cc = 1.0 cc
Vt = [2 * 0.5 cc] + [2 * 0.5 cc] + [4 * 1.0 cc
+ 1 * 0.5 cc] = 6.5 cc
SCOZZ = [(100/320)2 * (1.0) + (100*32/(320z ))2 * (6.5)
= 0.104 %

The variance for the oxygen concentration of standard 3 is

285

the same as that for carbon dioxide. Variances of standard
4 are 0.1719 % and 0.049 % for carbon dioxide and oxygen,

respectively.

APPENDIX F

PRELIMINARY BULK SAMPLER EXPERIMENTS

APPENDIX F

PRELIMINARY BULK SAMPLER EXPERIMENTS

 

 

 

Table E.1: Effect of sampler type and sample depth on
calculated bulk density, porosity and
free air space.
TYPE DEPTH REPLICATION AVG STD CV
4 5 7 DEV
Moisture Content
RING 1 80.6 81.4 81.8 81.3 0.49 0.006
2 82.1 83.0 81.2 82.1 0.72 0.009
3 81.6 81.1 81.7 81.5 0.26 0.003
AVG 81.4 81.8 81.6
STD 0.61 0.83 0.24
CV 0.007 0.010 0.003
SHORT 1 81.0 81.4 81.9 81.4 0.36 0.004
2 80.8 81.0 81.8 81.2 0.43 0.005
3 82.6 80.7 80.9 81.4 0.85 0.010
AVG 81.5 81.1 81.5
STD 0.83 0.26 0.42
CV 0.010 0.003 0.005
LONG 1 79.6 80.9 81.9 80.8 0.91 0.011
2 80.3 81.2 81.6 81.0 0.54 0.007
3 80.1 80.7 81.4 80.7 0.53 0.007
AVG 80.0 80.9 81.6
STD 0.28 0.19 0.18
CV 0.004 0.002 0.002

 

 

 

 

Table F.1 (cont'd.).

 

 

TYPE DEPTH REPLICATION AVG STD CV
4 5 7 DEV
Volatile Solids
RING 1 93.5 94.3 94.4 94.0 0.42 0.004
2 94.8 94.6 94.8 94.7 0.09 0.001
3 93.8 94.9 94.7 94.5 0.50 0.005
AVG 94.0 94.6 94.6
STD 0.56 0.25 0.17
CV 0.006 0.003 0.002
SHORT 1 94.1 95.0 94.3 94.5 0.36 0.004
2 94.9 94.9 94.3 94.7 0.26 0.003
3 94.8 95.0 94.2 94.7 0.37 0.004
AVG 94.6 95.0 94.3
STD 0.33 0.06 0.07
CV 0.004 0.001 0.001
LONG 1 94.3 94.6 94.4 94.4 0.15 0.002
2 94.5 94.0 94.5 94.3 0.21 0.002
3 94.4 94.6 93.7 94.3 0.38 0.004
AVG 94.4 94.4 94.2
STD 0.09 0.28 0.34
CV 0.001 0.003 0.004

 

 

Table F.1 (cont'd.).

288

 

 

TYPE DEPTH REPLICATION- AVG STD CV
4 5 7 DEV
Full Bag Weight (gm)
RING 1 467.50 495.80 413.75
2 460.70 556.20 492.50
3 500.20 446.50 508.80
SHORT 1 _296.33 317.43 311.67
2 425.91 409.90 371.03
3 447.10 408.50 454.30
LONG 1 221.08 192.38 189.91
2 237.26 224.72 216.72
3 291.44 312.68 276.08
Net Solids Weight (gm)
RING 1 458.00 486.30 404.25
2 451.20 546.70 483.00
3 490.70 437.00 499.30
SHORT 1 286.83 307.93 302.17
2 416.41 400.40 361.53
3 437.60 399.00 444.80
LONG 1 211.58 182.88 180.41
2 227.76 215.22 207.22
3 281.94 303.18 266.58
Length (in)
RING 1 2.0 2.0 2.0 2.0 0.00 0.000
2 2.0 2.0 2.0 2.0 0.00 0.000
3 2.0 2.0 2.0 2.0 0.00 0.000
SHORT 1 12.0 12.0 12.0 12.0 0.00 0.000
2 11.5 12.5 12.0 12.0 0.41 0.034
3 9.5 7.0 6.5 7.7 1.31 0.171
LONG 1 13.0 12.5 14.0 13.2 0.62 0.047
2 10.5 12.5 10.0 11.0 1.08 0.098
3 10.0 10.0 10.0 10.0 0.00 0.000

 

 

 

289
Table F.1 (cont'd.).
TYPE DEPTH REPLICATION AVG STD CV
4 5 7 DEV
Bulk Density (kg/m‘3)
RING 1 278 295 245 273 20.7 0.076
2 274 332 293 300 24.1 0.080
3 298 265 303 289 16.7 0.058
AVG 283 297 281
STD 10.5 27.2 25.2
CV 0.037 0.092 0.090
SHORT 1 198 213 209 206 6.1 0.030
2 300 265 250 272 21.0 0.077
3 382 472 567 474 75.7 0.160
AVG 293 317 342
STD 75.1 112.0 160.1
CV 0.256 0.354 0.469
LONG 1 202 182 160 181 17.2 0.095
2 270 214 258 247 23.9 0.097
3 350 377 331 353 18.7 0.053
AVG 274 258 250
STD 60.6 85.4 70.1
CV 0.221 0.331 0.281

 

 

 

 

Table F.1 (cont'd.).

290

 

 

TYPE DEPTH REPLICATION AVG STD CV
4 5 7 DEV
Porosity
RING 1 0.967 0.964 0.971 0.967 0.003 0.003
2 0.884 0.859 0.876 0.873 0.010 0.012
3 0.964 0.969 0.964 0.965 0.002 0.002
AVG 0.938 0.930 0.937
STD 0.039 0.050 0.043
CV 0.041 0.054 0.046
SHORT 1 0.976 0.974 0.975 0.975 0.001 0.001
2 0.962 0.967 0.970 0.967 0.003 0.003
3 0.957 0.941 0.930 0.942 0.011 0.012
AVG 0.965 0.961 0.959
STD 0.008 0.014 0.020
CV 0.008 0.015 0.021
LONG 1 0.973 0.977 0.981 0.977 0.003 0.003
2 0.965 0.974 0.969 0.969 0.003 0.004
3 0.955 0.953 0.960 0.956 0.003 0.003
AVG 0.964 0.968 0.970
STD 0.008 0.011 0.009
CV 0.008 0.011 0.009

 

 

 

 

Table F.1 (cont'd.).

291

 

 

TYPE DEPTH REPLICATION AVG STD CV
4 5 . 7 DEV
Free Air Space
RING 1 0.741 0.724 0.770 0.745 0.019 0.026
2 0.743 0.688 0.726 0.719 0.023 0.032
3 0.721 0.752 0.716 0.730 0.016 0.022
AVG 0.735 0.721 0.737
STD 0.010 0.026 0.023
CV 0.013 0.036 0.032
SHORT 1 0.815 0.801 0.805 0.807 0.006 0.007
2 0.720 0.752 0.766 0.746 0.019 0.026
3 0.641 0.559 0.471 0.557 0.070 0.125
AVG 0.726 0.704 0.681
STD 0.071 0.104 0.149
CV 0.098 0.148 0.219
LONG 1 0.812 0.830 0.850 0.831 0.015 0.019
2 0.749 0.800 0.759 0.769 .0.022 0.029
3 0.674 0.648 0.690 0.671 0.017 0.026
AVG 0.745 0.760 0.766
STD 0.057 0.080 0.065
CV 0.076 0.105 0.085

 

 

 

292

Table F.2: Statistical analysis of bulk density and
free air space data.

 

COMPARISON STD DEV F RESULT POOLED STUDENT SIGNIF.
RATIO * (95%) STD DEV T ‘ LEVEL

 

Bulk Density

Ring/ 25.3/7.529 11.3 nsd 40.00 2.03 nsd 95
Short 29.519/25.772 1.3 nsd 29.13 1.18 nsd 95
20.414/92.956 16.0 nsd 114.60 1.98 nsd 95

Ring/ 25.3/21.07 1.4 nsd 54.20 2.07 nsd 95
Long 29.519/29.258. 1.0 nsd 38.98 1.65 nsd 95
20.414/22.845 1.2 nsd 40.16 1.96 nsd 95

Short/ 7.529/21.070 7.8 nsd 19.83 1.54 nsd 95
Long 25.772/29.258 1.3 nsd 28.08 1.07 nsd 95
92.956/22.845 16.6 nsd 89.61 1.65 nsd 95

Free Air Space

Ring/ 0.0233/0.007 11.1 nsd 0.037 2.04 nsd 95
Short 0.0282/0.0236 1.4 nsd 0.028 1.20 nsd 95
0.0195/0.085 19.0 sd 0.110 1.94 nsd 95

Ring/ 0.0233/0.019 1.2 nsd 0.058 2.08 nsd 95
Long 0.0282/0.027 1.1 nsd 0.037 1.66 nsd 95
0.0195/0.212 1.2 nsd 0.037 1.95 nsd 95

Short/ 0.007/0.019 7.4 nsd 0.018 1.61 nsd 95
Long 0.0236/0.027 1.3 nsd 0.026 1.08 nsd 95
0.085/0.0212 16.1 nsd 0.083 1.68 nsd 95

 

 

* Larger standard deviations placed on top
for calculations. n = 3 for all treatments.

 

 

APPENDIX G

DERIVATION OF ERROR PROPAGATION EQUATIONS

 

 

 

APPENDIX G

DERIVATION OF ERROR PROPAGATION EQUATIONS

Assume that u = f(x,y) and that all errors are
independent and may be treated as random. Further assume
that all deviations Gxi = xi - x and Gyi = Yi - y are
relatively small. Applying the Taylor series expansion,

neglecting higher order terms, we have

“i = f([x + 6xi], [y + Gyil)
= f(x,y) + an * 6x- + au * Gy- (G.l)
3; 1 5y 1
and,
6u- = u- - u = au 6x. + an Gy- (6.2)
1 1 5; 1 5? 1

By definition, the square of the standard deviation in

u, SD, is
su2 = 2 (Gui)z / n (6.3)
Squaring 6.2 gives
(Gu-)z = (8u)z(6x-)z + 2 an au 6x- 6y- + (au)2(6y-)2
1 8i 1 5i 5? 1 1 5? 1
(6.4)

Placing this expression into 6.3, we have

s 2 = (Bu)2 2(bx-)z + 2 au 8u £(bx- by-) + (an)2 £(by-)2
u 3? 1 5i 3? 1 1 5? 1

n (G.5)

As n increases, the sum (dxiéyi) goes to 0 if xi and

 

Yi are independent because any bxi byi is likely to be

positive as negative. Since

293

294

5,,2 = £(6xi)2/ n and 5Y2 = £(6yi)2/ n (6.6)
can be substituted into G.5 to get
2 z 2 z z
s = (an) s + (an) s (6.7)
u 5; x §§ y
For more than 2 terms, 6.7 can be generalized to
s 2 = 2(3u )2 s -2 (6.8)
u 5;. X]

1
The fractional variance in U is written by dividing su
by u:

(su / a )2 = (2(au)z sijJ / a: (0.9)
X

The fractional variance should not be used unless the 0 of

each xo

1 and u scales are physically significant (Parrot,

1961).

 

Sum or Difference

Let u = x i y
then
' 8u = l and Bu = :1
’52? ”517
substituting into G.8, we have
su’ = 5,,2 + syz (6.10)

The fractional standard deviation is
- 2 - 2 2 "'2
(sn / U) - (5x + sy ) / U (6.11)

Product 95 Quotient

 

Let '
u = x8 + Yb (G.12)
with a and b assumed to be exact constants. Then
au = axa_lyb and Bu = bxayb'l (G.13)
F? F?

Substituting the above equation into equation G.8, we

295

have

2 =

U a

s 2 x2(a-1)Y2bs 2 + b2x2ay2(b—l)sy2 (0.14)

X

The fractional variance of a product is given by

(s / U)2 = (a2x2a-2Y2bsx2/ 02) + (bZXZaYZb-25y2/ U2)

U

2X2a-2Y2bsx2/ §2a§2b) + (bZXZaYZb—Zsy2/ §2a§2b)

(a

= (aZsz/ 22) + (bzsyz/ Y2) (0.15)

 

 

APPENDIX H

WINDROW SIZE, TEMPERATURE, GAS CONCENTRATION
AND PHYSICAL PROPERTY SAMPLING LOCATIONS
AND DATA AVAILABILITY

 

APPENDIX H

WINDROW SIZE, TEMPERATURE, GAS CONCENTRATION
AND PHYSICAL PROPERTY SAMPLING LOCATIONS
AND DATA AVAILABILITY

Complete data on windrow size, temperature and ambient
conditions, gas concentration and physical properties are
available from the author at the following address:
Stephen E. Ferns
c/o Dr. John Gerrish
Department of Biological and Agricultural Engineering
A.W. Farrall Hall
East Lansing, Michigan 48823
USA

The data has the following formats and requirements:
Temperature: Symphony (1) 7 360 KB floppy disks;
Windrow Elevations, Gas Concentrations, and Physical
Properties: Lotus 123 (1A) 2 360 KB floppy disks.

Please send the necessary disks and disk mailer with your

request.

296

L I ST OF REFERENCES

J--" '-

 

 

LIST 93 REFERENCES

 

Acott, K. M. and T. P. Labuza. 1975. Microbial Growth
Response to Water Sorption Preparation. J. Food
Technol. 10:603-611. Cited in Christian, 1980,
op cit.

Adams, D. M. 1978. Heat Injury of Bacterial Spores.
Advances in Applied Microbiology 23:245-261.

Allen, S., M. Wallentine, S. Austin, P. Burch and K. Hoopes.
1979. Recycled Manure Solids as a Free-stall
Bedding for Lactating Dairy Cows, and its
Association with Mastitis. Report from the National
Mastitis Council. pp. 121—127.

Allwood, M. C. and A. D. Russell. 1970. Mechanisms of
Thermal Injury in Nonsporulating Bacteria. Advances
in Applied Microbiology 15:245-261.

Andrews, J. F. and K. Kambhu. 1973. Thermophilic Aerobic
Digestion of Organic Solid Wastes. Office of
Research and Development, 0.8. EPA NTIS PB-222-396.
Cited in Haug, 1980. op. cit.

ASTM. 1986. Annual Book of ASTM Standards: Soil & Rock,
Building Stones. SEction 4, Vol. 04.08. -
American Society of Testing and Materials.
Washington, D.C.

 

 

 

 

AWWA. 1976. Standard Methods for the Examination of Water
and Wastewater. “IIfE'EaT‘1976. AWWA, APHA, afia“‘
WPCF TJointly published), Washington, DC.

 

 

 

Atlas, R. M. and R. Bartha. 1981. Microbial Ecology:
Fundamentals and Applications. Addison-Wesley
Publishing Company, Reading, Mass.

 

Baranowski, R. 1983. Evaluation of the Useability of
Procedures for the Determination of Soil Physical
Properties in Tillage Experiments. Institute of
Soil Science and Plant Cultivation. 55-230.
Luskowice Olawskia, Poland. Cited in Raper, R. L.
and D. C. Erbach. 1985. Accurate Bulk Density
Measurements Using a Core Sampler. ASAE Paper
85-1542.

297

 

 

298

Bartholomew, W. V. and A. G. Norman. 1953. Microbial
Thermogenesis in the Decompositon of Plant
Materials IV. Influence of Moisture Content and
of Initial Temperature. J. Bacteriol. 65:228-232.

Bauer, L. D. 1959. Soil Physics. Third Edition. John
Wiley and Sons. Cited in Raper, R. L. and D. C.
Erbach.. 1985. Accurate Bulk Density Measurements
Using a Core Sampler. ASAE Paper 85-1542.

 

Bailey, J. E. and D. F. Ollis. Biochemical Engineering
Fundamentals. McGraw-Hill, San FranciSco, CA.

 

 

 

Beckwith, C. P. and J. W. Parsons. 1980. The Influence of
Mineral Amendments on the Changes in the Organic
Nitrogen Components of Composts. Plant and Soil
54:259-270.

Bejan, A. 1984. Convection Heat Transfer. John Wiley and
Sons. New York, NY.

 

Bell, R. G. and J. Pos. 1971. Design and Operation of a
Pilot Plant for Composting Poultry Manure. Trans.
of the ASAE 14(6):1020-1023.

Bishop, J. R., J. J. Jansen, and A. B. Bodine. 1980.
Effect of Ambient Environments on Survival of
Selected Bacterial P0pulations in Dairy Waste
Solids. J. Dairy Sci. 63(3):523-525.

Bishop, J. R., J. J. Janzen, A. B. Bodine, C. A. Caldwell
and D. W. Johnson. 1981. Dairy Waste Solids as a
Possible Source of Bedding. J. Dairy Sci. 64(4):
706-711.

Blosser, T. H. 1979. Economic Losses from and the National
Research Program on Mastitis in the United States.
J. Dairy Sci. 62(1):ll9-127.

Bohnhoff, D. R., J. C. Converse and J. B. Petersen. 1984.
Thermal and Physical Properties of Separated Manure
Solids. ASAE Paper No. 84—4077. Am. Soc. of Agr.
Eng., St. Joseph, MI 49085.

Bohnhoff, D. R. and J. C. Converse. 1986a. Engineering
Properties of Se arated Manure Solids. Agricultural
Wastes (in press .

Bohnhoff, D. R. and J. C. Converse. 1986b. Water Desorption
Properties of Separated Manure Solids. Agricultural
Wastes (in press).

Braithwaite, R. L. 1956. A Study of Garbage Composting in
Controlled, Insulated Barrels. Unpublished M.S.

 

 

 

 

299

Thesis. Michigan State University, E. Lansing, MI
48824.

Bramely, A. J. 1974. The Aetiology and Control of
Coliform Mastitis in Dairy Cattle. PhD
Dissertation, University of Reading, Reading,
England. Cited in Carroll, L. J. 1977.
Environmental factors in bovine mastitis. JAVMA
l70(10):1l43-ll49.

Bramley, A. J. and F. R. Neave. 1975. Studies on the
Control of Coliform Mastitis in Dairy Cows. Brit.
Vet. J. 131:160-169.

Brander, G. C. 1972. Dairy Herd Environment and the
Control of Mastitis. Vet. Rec. 92:501-506.

Brandon, J. R. 1976. Sandia's Sludge Irradiation Program.
In Slud e Mggagement, Disposal and Utilization.
Information Tranéfer. Rockville, Md. Cited in
Haug, 1980, op cit.

 

Bretzloff, C. W. and M. S. Fluegel. 1962. Chemical
Compositon of Mushroom Compost During Composting
and Cropping. Mushroom Sci. 5:46-60.

Brooker, D. B., F. W. Bakker-Arkema and C. W. Hall. 1981.

Dr in Cereal Grains. AVI Publishing Company.
Wesfporf, CN.

Brown, M. H. and O. Emberger. 1980. Oxidation-reduction
Potential. In Microbial Ecolo of Foods I. Factors
Affecting Life and Death of MicroEFganismST— J.H.
SilliRer et al., eds. *Acidemic Press. New York,
NY.

 

 

Browne, C. A. 1929. The Spontaneous Combustion of Hay.
U.S. Dept. of Agric. Tech. Bull. 141:1-38.
Cited in Cooney and Emerson, 1964, op. cit.

Burge, W., D. Colacicco, W. Cramer and E. Epstein. 1978.
Criteria for Control of Pathogens During Sewage
Sludge Composting. In Proceedings of the National

Conference on Desi n of*Munici a1 SIfidge Compost
Facilities. InformatTEn TransIer. RockvilIe, Md.
Busta, F. F. 1978. Introduction to Injury and Repair of

Microbial Cells. Adv. Appl. Microbiology
23:195-201.

 

 

 

 

 

Burrows, S. 1951a. The Chemistry of Mushroom Composts. I.
General Introduction and Methods of Investigation.
J. Sci. Food Agric. 2(9):395-403.

 

300

Burrows, S. 1951b. The Chemistry of Mushroom Composts.
II. Nitrogen Changes During the Composting and
Cropping Processes. J. Sc1. Food Agric. 2(9):

Campbell Scientific. 1985. CR21X Microlo er
Instruction Manual. CampEeII Scientific, Inc.
Provo, Utah.

 

Carroll, E. J. 1977. Environmental Factors in Bovine-
Mastitis. JAVMA 170(10):1143—1149.

Carroll, E. J., N. C. Jain, O. W. Schalm, and J. Lasmanis.
1973. Experimentally Induced Coliform Mastitis:
Innoculation of Udders with Serum-sensitve and
Serum Resistant Organisms. Am. J. Vet. Res.
34(9):1143-1146.

Carroll, E. J. and D. E. Jasper. 1978. Distribution of
Enterobacteriaceae in Recycled Manure Bedding on
California Dairies. J. Dairy Sci. 61:1498-1508.

Carroll, E. J. and D. E. Jasper. 1979. Coliform
Populations in Bedding Materials and Coliform
Mastitis Incidence. University of California,
Davis, CA.

Chang, A. C. and J. M. Rible. 1975 Particle-size
Distribution of Livestock Wastes. In Managing
Livestock Wastes--Proceedings of the 3rd
International Symposium on LivEEtEEK WEEtes.
American Society ongriculturaI Engineers, St.
Joseph, MI. pp. 339-342.

Chang, C. S., F. S. Lai and B. S. Miller. 1980. Composting
of Grain Dust. Trans. of the ASAE 23(3):709-711.

Chang, Y. and H. J. Hudson. 1967a. The Fungi of Wheat
Straw Compost I. Ecological Studies. Trans. Br.
mycol. Soc. 50(4):649-666.

Chang, Y. 1967. The Fungi of Wheat Straw Compost II.
Biochemical and Physiological Studies. Trans. Br.
mycol. Soc. 50(4):667-677.

Chatt, E. M. 1953. Cocoa: Cultivation, Processin ,
Analysis. Interscience. New York, NY. C1ged in

Cooney and Emerson, 1964, op.cit.

 

Chau, K. V., J. J. Gaffney and S. Bellagha. 1984.
Simulation of Heat and Mass Transfer in Products
with Internal Heat Generation and Transpiration.
ASAE Paper No. 84-6513. Am. Soc. Agr. Eng. St.
Joseph, MI 49085.

301

Chen, Y. R. 1982. Engineering Properties of Beef Cattle
Manure. ASAE Paper No. 82-4085. Am. Soc. Agr.
Eng. St. Joseph, MI 49085.

Chen, Y. R. 1983. Thermal Properties of Beef Cattle
Manure. Agricultural Wastes 6:13-29.

Chen, W. W. H, R. E. Zimmerman and A. G. Franklin. 1977.
Time Settlement Characteristics of Milled Urban
Refuse. In Geotechnical Practice for Disposal of
Solid Waste Materials. Am. Soc. Civil Eng., New
York, NY. pp. 136-152.

Cohn, F. 1889. Uber thermogene Wirkung von Pilzen.
Jahresber. Schles. Geselsch. Vaterl. Kult. 66:
150-156. Cited in Cooney and Emerson, 1964, op.
cit.

Cooney, D. G. and R. Emerson. 1964. Thermophilic Fungi:
An Account of Their Biolo , ActiVities, and
CIass1f1cati5n. W. H. reeman andiCompany, San
Fiansisco, CA.

 

 

 

 

Corlett, Jr., D. A. and M. H. Brown. 1980. pH and Acidity.
In Microbial Ecology of Foods 1; Factors Affecting
Life and Deat g_ 1croorganisms. J.H. SiIliker et
aI., eds. Academic Press, New York, NY.

 

 

 

 

Christian, J. H. B. 1980. Reduced Water Activity. In
Microbial Ecolo of Foods 1. Factors Affecting Life
and Death 9: M1cr05?gan1sm§:— J.H. S111iker éf al.,
53—. Academic Press, New York, NY.

 

 

 

Cussler, E. L. 1984. Diffusion Mass Transfer in Fluid
S stems. Cambridge University Press, Cambr1dge,
EngIand.

 

Davis, M. L. 1981. Water and Wastewater Analysis.
Department of C1v11 Engineering, MiChiganiState
University, East Lansing, MI.

 

Dilley, D. 1986. Professor, Department of Horticulture,
Michigan State University, East Lansing, Michigan
48824. Personal Communication.

Douglas, F. R. Hambleton, and G. J. Rigby. 1973. An
Investigation of the Oxidation-Reduction Potential
and of the Effect of Oxygen on the Germination and
Outgrowth of Clostridium Butyricum Spores, Using
Platinum Electrodes. J. Appl. Bacteriol. 36:623-
633. Cited in Brown and Emberger, 1980, op. cit.

Eberhart, R. J. 1977. Coliform Mastitis. JAVMA 170(10):

302
1160-1163.

Eberhart, R. J. and J. M. Buckalew. 1972. Evaluation of a
Hygine and Dry Period Therapy Program for Mastitis
Control. J. Dairy Sci. 55:1683-1690.

Farnsworth, R. J. 1977. Significance of Fungal Mastitis.
JAVMA l70(10):1173-1174.

Finger, 5. M. 1975. Aerobic Microbial Growth in Semi
-Solid Matrices. Ph.D Dissertation. University of
Maryland, College Park, Md.

Finger, S. M., R. T. Hatch, and T. M. Regan. 1976. Aerobic
Microbial Growth in Semi-Solid Matrices: Heat and
Mass Transfer Limitation. Biotechnology and
Bioengineering 18:1193-1218.

Finnish Peatland Society. 1982. Peatlands and Their
Utilization in Finland. Finnish Peatland
Society, Hel§inEi, FinIand.

 

 

Finstein, M. S. 1980a. Composting Microbial Ecosystem:
Implications for Process Design and Control.
Compost Science/Land Utilization 21(4):25-27.

Finstein, M. S., J. Cirrello, S. T. MacGregor, F. C. Miller,
and K. M. Psarianos. 1980b. Slud e Composting and
Utilization: Rational Approacfi to Process Contr6i}
Department of Env1ronmenta c1efiEe,iRutgers--The
State University of New Jersey, Cook College. New
Brunswick, New Jersey 08903. USEPA P882-136243.

 

 

Finstein, M. 8., F. C. Miller, P. F. Strom, S. T. MacGregor
and K. M. Psarianos. 1983. Composting Ecosystem
Management for Waste Treatment. Bio/Technology
1(4):347-353.

Francis, P. G., J. Sumner and D. A. Joyce. 1981. The
Influence of the Winter Environment of the Dairy Cow
on Mastitis. The Bovine Practitioner 16(11):24-27.

Freese, E., C. W. Sheu and E. Galliers. 1973. Function of
Lipophilic Acids as Anitmicrobial Food Additives.
Nature (London) 241:321-325. Cited in Corlett and
Brown, 1980, op. cit.

Frost, A. J., D. D. Wanasinghe and J. B. Woolcock. 1977.
Some Factors Affecting Selective Adherence of
Micro-organisms in the Bovine Mammary Gland.
Infection and Immunity 15(1):245-253.

Garner, W. W. 1946. The Production of Tobacco. Blakiston,
Phiadelphia. Cited in Cooney and Emerson, 1964,

 

 

 

 

303
op.cit.

Geankoplis, C. J. 1983. Transport Processes and Unit
Operations. Allyn and Bacon, Inc. Boston, MA.

 

Gerrits, J. P. G., H. C. Bels-Koning and M. Muller. 1965.
Changes in Compost Constituents During Composting,
Pasteurization and Cropping. Mushroom Sci. 5:
225-243.

Gerrits, J. P. G. 1972. The Influence of Water in Mushroom
Compost. Mushroom Sci. 8:43-57.

Glathe, H. 1960. Selbsterhitzung und selbstenzundung von
erntestoffen und ihre verhutung. Ergeb.
landwirtsch. Forsch. 3:83-98. Cited in Cooney
and Emerson, 1967, op. cit.

Gregory, P. H., M. E. Lacey, G. N. Festenstein and F. A.
Skinner. 1963. Microbial and Biochemical Changes
During the Moulding of Hay. J. gen. Microbiol. 33:
147-174. Cited in Finger and Morris, 1975, op. cit.

Griffis, C. L. and C. R. Mote. 1978. Cotton Gin Trash
Composting Studies. Ark. Farm Res. Ark. Ag. Exp.
Stn. 27(4):3.

Groffman, P. 1986. Personal Communication. Research Associate.
Department of Crops and Soils Sciences. Michigan
State University, East Lansing, MI 48824. 6/86.

Hansen, N. H. and H. Riemann. 1963. Factors Affecting the
Heat Resistance of Nonsporing Organisms. J. Appl.
Bact. 26(3):314-333.

Hajny, G. J. 1966. Wood Chip Storage: A Review and
Bibliography. TAPPI 49(10):97A-105A.

Haug, R. T. 1979. Engineering Principles of Sludge
Composting. JWPCF 51(8):2189-2206.

Haug, R. T. 1980. Compost Engineering: Principles and
Practice. Ann Arbor Science Publishers. Ann Arbor,
MI.

 

 

Hauser, E. and J. W. Sinden. 1953. The Nature of the
Composting Process and its Relation to Short
Composting. Mushroom Sci. 2:123-130.

Hermanson, R. E. 1985. Flush Cleaning Dairy Barns: Case
Studies. In Agricultural Waste Utilization and
Management--Proceedings of the Fifth International

Sym osium on AgriculturaI_Wastes. December 16-17,
1985. Afi.-Soc. Agr. Eng., St.iJOSeph, Michigan.

 

 

 

 

304
pp. 590-597.

Hewitt, L.F. 1950. Oxidation-reduction Potentials in
Bacteriology and Biochemistry. Sixth EditiofiT
Livingston, Edinburgh. Cited in Brown and Emberger
1980, op. cit.

 

Higgins, A. J., A. J. Kaplovsky and J. V. Hunter. 1982.
Organic Composition of Aerobic, Anaerobic, and
Compost-stabilized Sludges. JWPCF 54(5):466-473.

Hillel, D. 1982. Introduction to Soil Physics. Academic
Press, New York: NY.

 

 

Hoitink, H. A. J. and H. A. Poole. 1980. Bark Compost Use
in Container Media. Compost Science/Land
Utilization 21(3):38-4l.

Houkom, R. L., A. F. Butchbaker and G. H. Brusewitz. 1972.
Thermal Properties of Beef Manure. ASAE Paper No.
72-316. Am. Soc. Agr. Eng. St. Joseph, MI 49085.

Hoyle, D. A. and G. E. G. Mattingly. 1959. Studies on
Composts Prepared from Waste Materials I:
Prepartion, Nitrogen Losses and Changes in 'Soluble
Nitrogen.‘ J. Sci. Food Agric. 5(1):54-65.

Hummel, J. W. and G. B. Willson. 1975. High Rate
Composting of Dairy Manure. In Proceedings of the
International Symposium on Livestock Wastes,_AmT_—
Soc. Agr. Efi§., St. Josepfi, MI pp. 4ST3434.

 

 

 

 

Hurst, A. 1980. Injury and Its Effect on Survival and
Recovery. In Microbial Ecolo of Foods I. Factors
Affecting Life and Death of M1cr6:organ1sޤ. J.H.
Silliker et al., eds. Academic Press. New York,
NY.

 

 

 

 

Ishida, M. and C. Y. Wen. 1971a. Comparison of Zone-
Reaction Model and Unreacted Shrinking-core model
in Solid-gas Reactions--I. Isothermal Analysis.
Chem. Engng Sci. 26(7):1031-104l.

Ishida, M., C. Y. Wen and T. Shirai. 1971b. Comparison of
Zone-reaction Model and Unreacted Shrinking-core
Model in Solid-gas Reactions--II. Non-isothermal
Analysis. Chem. Engng Sci. 26(7):1043-1048.

Jackson, H. and M. Woodbine. 1962. The Effect of Sublethal
Treatment on the Growth of Staphylococcus Aureus.
J. Appl. Bact. 25, viii. Cited in Hansen and
Riemann, 1963, op. cit.

Jackson, R. D. 1964. Water Vapor Diffusion in Relatively

 

305

Dry Soil: 1. Theoretical Considertions and Sorption
Experiments. Soil Sci. Soc. Am. Proc. 28:172-176.
Cited in Hillel, 1982, op.cit.

Jain, N. C. 1979. Common Mammary Pathogens and Factors in
Infection and Mastitis. J. Dairy. Sci. 62(1):

James, L. H., L. F. Rettger and C. Thom. 1928. Microbial
Thermogenesis II. Heat Production in Moist Organic
Materials with Special Reference to the Part Played
by Micro-organisms. J. Bacteriol. 15:117-141.

Jarrett, J. A. 1981. Mastitis in Dairy Cows. Veterninary
Clinics of North America: Large Animal Practice
3(2):447-455.

Jarret, P. 1986. Personal Communication. Professor of Civil
Engineering. Royal Military College of Canada.
Kingston, Ontario K7L2W3. 5/86.

Jasper, D. E. 1977. Mycoplasma and Mycoplasma Mastitis.
JAVMA l70(10):1167-ll72.

Jasper, D. E., J. D. Dellinger, and R. B. Bushnell. 1975.
Herd Studies on Coliform Mastitis. JAVMA 166(14):

Jeris, J. S., and R. W. Regan. l973a. Controlling
Environmental Parameters for Optimum Composting,
Part I: Experimental Procedures and Temperature.
Compost Science. Jan/Feb: 22-28.

Jeris, J. S., and R. W. Regan. 1973b. Controlling
Environmental Parameters for Optimum Composting,
Part II: Moisture, Free Air Space, and Recycle.
Compost Science. March/April: 8-22.

Jeris, J.S., and R.W. Regan. 1973c. Controlling
Environmental Parameters for Optimum Composting,
Part III: pH, Nutrients, Storage and Paper Content
Relative to Composting. Compost Science.
July/August: 16-22.

Kane, B. E. and J. T. Mullins. 1973. Thermophilic Fungi
in a Municipal Waste Compost System. Mycologia 65:
1087-1100.

Karak, S. and Y. Yener. 1979. Heat Conduction. Middle
East Technical University Press. Ankara, Turkey.

 

Kaufmann, C. W., L. G. Harmon, O. C. Pailthorp and I. J.
Pflug. 1959. Effect of heat treatment on the
growth of surviving cells. J. Bact. 78:834. Cited

306
in Hansen and Riemann, 1963, op. cit.

Keys, Jr., J. E., L. W. Smith, and B. T. Weinland. 1977.
Response of Dairy Cattle Given a Free Choice of Free
-sta11 Location and Three Bedding Materials. J.
Dairy Science 56(6):1157-1162.

Lacey, J. 1973. In Actinomycetales: Characteristics and
Practical Importance. G. Sykes and F.A. Skinner,
eds. pp. 231-251. Academic Press. New York. Cited
in Finstein and Morris, 1975, op. cit.

Lambert, E. B. 1941. Studies on the Preparation of
Mushroom Compost. J. Agric. Res. 62(7):415-422.

Lambert, E. B. and A. C. Davis. 1934. Distribution of
Oxygen and Carbon Dioxide in Mushroom Compost Heaps
as Affecting Microbial Thermogenesis, Acidity, and
Moisture Therein. J. Agric. Res. 48(7):587-601.

Latif, N. and E. Lissik. 1986. Respiration Model for Heat
and Moisture Release During Grain Storage. ASAE
Paper No. 86-6508. Am. Soc. Agr. Eng., St. Joseph,
MI 49085.

Lembke, A. 1937. Die Hitzewiderstandsfaghigkeit der
Colibakterien und die verwendbarkeit dieser
eigenschaft als vergleichmasstab fur die Beurteilung
von Milcherhitzern. Zbl. Bakt. (Abt. 2) 96:92.
Cited in Hansen and Riemann, 1963, op. cit.

Lindemann, E. R., J. M. Sweeten, and J. P. Burt. 1985.
Sludge Removal from Dairy Lagoons. In Agricultural
Waste Utilization and Management--Proceedings of*the
Fifth InternationaI—S osium on Agriculturai WESEEE,
Am. Soc. Agr. Eng., St. UosephT—MI. pp.

 

 

 

Lombardo, P. 1977. Septage Composting. Compost Science
18(6):12-14.

Marshall, B. J., D. F. Ohye and J. H. B. Christian. 1971.
Tolerance of Bacteria to High Concentrations of
NaCl and Glycerol in the Growth Medium. Appl.
Microbiol. 21:363-364. Cited in Christian, 1980,
op. cit.

Martin Jr., J. H., M. Decker Jr., and K. C. Das. 1972.
Windrow Composting of Swine Wastes. In Waste
Management Research: Proceedings of the I972
Cornell Agricultual Waste Managemefit Conference.
Giaphics Management Corp., Washington, D.C. pp.
159-172.

 

 

 

 

McKinley, V. L. and J. R. Vestal. 1984. Biokinetic

 

307

Analysis of Adaptation and Succession: Microbial
Activity in Composting Sewage Sludge. Appl.
Environ. Microbiol. 47(5):933-941.

Mears, D. R., M. E. Singley, G. Ali and F. Rupp III. 1975.
Thermal and Physical Properties of Compost. In
Ener , Agriculture and Waste Management:
Proceedings ofthe 1975 COrnelI AgricuItural Waste
Management Conference. W.J. JewelI, ed. Ann
Arbor Science Publishers. Ann Arbor, MI.

 

 

 

 

 

 

Meijering, A. G., W. H. Clifford and F. W. Bakker-Arkema.
1972. Drying a Bed of Composted Waste. Trans. of
the ASAE. 15(1):116-120.

Miehe, H. 1907. Die Selbsterhitzung des Heus. Eine
Biologische Stuie. Gustav Fischer. Jena. ppl-127
Cited in Cooney and Emerson, 1964, op. cit.

Miehe, H. 1930. Die Warmebildung von Reinkulturen im
Hinblick auf die Atiologie der Selbsterhitzung
pflanzlicher Stoffe. Arch. Mkrobiol., 1:78-118.
Cited in Cooney and Emerson, 1964, op. cit.

Moats, W. A. 1971. Kinetics of Thermal Death of Bacteria.
J. Bacteriol. 105(1):165-l71.

Moore, A.F. 1958. Oxygen Uptake Rates and Respiratory
Quotients of an Aerobically Decomposing Synthetic
Garbage. M.S. Thesis. Sanitary Engineering
Department, Michigan State University, E. Lansing,
MI 48834.

Mote, C. R. and C. L. Griffis. 1980. Variations in the
Composting Process for Different Organic Carbon
Sources. Agricultural Wastes 2(3):215-223.

Mote, C. R. and C. L. Griffis. 1980. Water Management for
Sustaining the Composting Process in Windrows of
cotton gin trash. Report Series 256, Agricultural
Experiment Station, Division of Agriculture,
University of Arkansas, Fayetteville, Arkansas.

Mote, C. R. 1985. Personal Communication. Professor.
Department of Agricultural Engineering, University of
Tennesee--Knoxville. Knoxville, Tennesee. 12/85.

Muller, M. 1965. Some Thoughts About Composting. Mushroom
Sci. 5:213-223.

Nat. Mastitis Counc. 1978. Current Concepts of Bovine
Mastitis. National Mastiti§ Council, THC.
Washington, DC.

 

 

 

 

 

308

Nat. Res. Counc. Can. 1979. Peat Testing Manual. Nat.
Res. Counc. Can., Assoc. Comm. Geotech. Res.
[Muskeg Subcomm.]. Technical Memo 125. pp 1-8.

Neave, F. K. and J. Oliver. 1962. The Relationship Between
the Number of Mastitis Pathogens Placed on the Teats
of Dry Cows, Their Survival, and the Amount of
Intramammary Infections Caused. J. Dairy Res
29:79-93.

Olson Jr., J. C. and P. M. Nottingham. 1980. Temperature.
In Microbial Ecolo of Foods I. Factors Affectin
Life and Deatfi g: 31CFB-6?§35i§fisT_—UTHT SiiiiREFS
et al., edS. Academic Press, New York, NY.

 

 

 

Packer, R. A. 1977a. Bovine Mastitis Produced by
Cornebacteria. JAVMA l70(10):1164-1165.

Packer, R. A. 1977b. Bovine Mastitis Caused by
Pseudomonas Aeruginosa. JAVMA l70(10):1175.

Parratt, L. G. 1961. Probability and Experimental Errors
In Science. John Wiley and Sons, New York, NY.

 

Pizer, N. H. 1950. Horse Manure Composts. Mushroom Sci.
1:46‘510

Poincelot, R. P. 1977. The Biochemistry of Composting.
In Composting of Munici a1 Residues and Sludges,
Proceedings g£_fhe 1977 National Conference.
Information Transfer. RockvilIe, MD. pp. 33-39.

 

 

Pratt, A. W. 1969. Heat Conduction in Low Conductivity
Materials. In Thermal Conductivity, Vol 1. R.P.
Tye, ed. Academic Press. New York, NY.

Raper, R. L. and D. C. Erbach. 1985. Accurate Bulk
Density Measurements Using a Core Sampler. ASAE
Paper No. 85-1542, Am. Soc. Agr. Eng., St. Joseph,
MI 49085.

Rendos, J. J., R. J. Eberhart and E. M. Kesler. 1975.
Microbial Populations of Teat Ends of Dairy Cows and
Bedding Materials. J. Dairy Sci. 58(10):l492-1500.

Rodriguez, A. C., G. H. Smerage, A. A. Teixeira, and F. F.
Busta. 1986. Kinetic Effects of Lethal
Temperatures on Spore Population. Paper No. 86-
6533. Am. Soc. Agr. Eng., St. Joseph, MI 49085.

Rose, W. W., J. E. Chapman, S. Roseid, A. Katsuyama, V.
Porter and W. A. Mercer. 1965. Composting Fruit
and Vegetable Refuse. Compost Science 6(3):l3-25.

309

Rosenberg, S. L. 1975. Temperature and pH Optima for 21
Species of Thermophilic and Thermotolerant Fungi.
Can. J. Microbiol. 21:1535-1540.

Ryan, D., R. G. Carbonell and S. Whitaker. 1981. A Theory
of Diffusion and Reaction in Porous Media. AIChE
Symposium Series 77(202):46-62.

Schalm, O. W., E. J. Carroll and N. C. Jain. 1971. Bovine
Mastitis. Lea and Febiger, Philadelphia, PA. pp.

Schulze, K. L. 1962. Continuous Thermophilic Composting.
Applied Microbiology 10:108-122.

Scott, W. J. 1957. Water Relations of Food Spoilage
Microorganisms. Adv. Food Res. 7:83-127. Cited
in Christian, 1980, op. cit.

Selna, M. and D. Smith. 1976. Pathogen Inactivation During
Composting. Technical Services Department, Los
Angeles Count Sanitation Districts, Los Angeles, CA.
Cited in Haug, 1980, op cit.

Shell, B. J. 1955. The Mechanism of Oxygen Transfer Through
a Compost Material. Ph.D Dissertation. Sanitary
Engineering Dept., Michigan State University, E.
Lansing, MI 48824.

Sinclair, C. G. and D. N. Ryder. 1975. Models for the
Continuous Culture of Micro-organisms under Both
Oxygen and Carbon Limiting Conditions.
Biotechnology and Bioengineering 17: 375-398.

Singley, M. E., A. J. Higgins and M. Frumkin-Roesengaus.
1982. Sludge Composting and Utilization: A Desi n
and Operat1n Manual. New Jersey AgricuItuFaI
Experiment tation, Cook College, Rut ers--The State
University of New Jersey. New Brunsw1ck, NJ.

 

 

Snell, J. R. 1957. Some Engineering Aspects of High-rate
Composting. J. San. Eng. Div., ASCE, Paper 1178.
Cited in Haug, 1980, op. cit.

Stanek, M. 1972. Micro-organisms Inhabiting Mushroom
Compost During Fermentation; Mushroom Sci. 8:797-
811.

Stanek, W. 1980. Comparison of Three Methods for
Determining the Bulk Density of Peat. Bimonthly
Research Notes, Forestry Service, Environment
Canada. 36(6):29;

310

Steniford, E. I., D. D. Mara and P. L. Taylor. 1984.
Forced Areation Co-composting of Domestic Refuse and
Sewage Sludge in Static Piles. In Composting of
Agricultural and Other Wastes. Elsevier AppliEd
ScienceiPUinshers. London, England. pp. 42-55.

 

Strom, P. F., M. L. Morris and M. S. Finstein. 1980. Leaf
Composting Through Appropriate, Low-level,
Technology. Compost Science/Land Utilization
21(6):44-48.

Suzuki, M. and K. Kumada. 1977. Nitrogen Transformation
During the Rotting Process of Rice Straw Compost.
Soil Sci. Plant Nutr. 23(2):163-174.

Topping, J. 1951. Errors of Observation and Their
Treatment. Tfie Insiitute of Physics, London,
England.

 

Troller, J. A. and Christian, J. H. B. 1978. Water
Activity and Food. Academic Press. New York, NY.
Cited in Cfiristian, 1980, op. cit.

 

Waksman, S. A., T. L. Gordon and N. Hulpoi. 1939.
Influence of Temperature Upon the Microbiological
Population and Decomposition Process in Composts of
Stable Manure. Soil Sci. 47:83-113.

Ward, R. L. and J. R. Brandon. 1977. Effect of Heat on
Pathogenic Organisms Found in Wastewater Sludge. In
Composting of Munic a1 Residues and Sludges, '
Pioceedings_3f tfie I977 National Conference.
Information T?ansfer,Rockvi11e, Md. pp. 122-127.

 

 

Wei, J. H. and S. L. Chang. 1975. A Multi-poisson
Distribution Model for Treating Disinfection Data.
In Disinfection: Water and Wastewater. J.D.
Johnson, ed. Ann Arbor SEience PUinShers, Ann
Arbor, MI. Cited in Haug, 1980, op. cit.

 

 

Wells, C. B. 1983. Core Samplers for Soil Profiles. J.

Wen, C. Y. 1968. Noncatalytic Heterogeneous Solid Fluid
Raction Models. Ind. Eng. Chem. 60(9):34-54.

Wen, C. Y. and S. C. Wang. 1970. Thermal and Diffusional
Effects on Noncatylitic Solid Gas Reactions. Ind.
Eng. Chem. 63(8):30-51.

Whang, D. S. and G. F. Meenaghan. 1980. Kinetic Model of
Composting Process. Compost Science/Land
Utilization 21(3):44-46.

311

Wiley, J. S. 1957. Progress Report on High-rate Composting
Studies. In Proceedings gf the 12th Industrial Waste
Conference. PurdueUniversity, West Lafayette, IN.

 

 

Wiley, J. S. and G. W. Pierce. 1955. A Preliminary Study of
High Rate Composting. Proceedings ASCE 81:846,
American Society of Civil Engineering. New York, NY.

Willson, J. F., E. Parr, E. Epstein, P. B. Marsh, R. L.
Chaney, D. Colacicco, W. D. Burge, L. J. Sikora, C.
F. Tester and S. Hornick. 1980. Manual for
Composting Sewage Sludge by the Beltsville Aerated-
Pile Method. Municipal Environmental Research
Laboratory, Office of Research and Development,
US-EPA, Cincinnati, Ohio, 45268. EPA-600-8-80-022.
May, 1980. p. 31.

Zimmerman, R. E., W. W. H. Chen and A. G. Franklin. 1977.
Mathematical Model for Solid Waste Settlement. In
Geotechnical Practice for Dis osal of Solid Waste
Materials. American Society of CiviI Engineers.
New York. pp. 210-226.

 

 

 

Ipll I !.|.‘ l. I: ll...

nICHIan STnTE UNIV. LIBRARIES
(Iimm]mW(WWW,(W("IIIHIIWWI
31293016279816