1' $1.}, 5'4"}:{1
53:. {3' 'I‘ 3.2:.”
_ {Win -- -
n. . €§7§m

   

  
   
 

  

'1 ”‘5’? I n-u
I. " “(fa-’3. :53
"9'1, '4: § <5} }:‘I:
~ 2"" _

 
  
  

 
  
  
 
 
       
  
  
  

      

- ”'3’.“
‘J.’ .
:lq.’i“,<‘\.
.1'5’3“'.'J. '

’
a. I ‘
.\0 a . '.

   

 

   
  
 

    
 
  

”45,5: '~ w {:3
"3:: “C's . -'_;£':~ ~ «. 33:31:}: ‘51:“ ~25}; v53: f.- I“.
~ ~ 7.. ‘5: . m ...,_—- :
din} "13.33.?“ 2-0393}; . 23%? {£.£}€ {5; k . -
:7. \f MKS."

..t
«9., 7
a.
. ‘ 5%?5,

‘ “35*
w} "“.g:Z-:-q’“,-..~‘:mux..
M

  
 
  
      
   

  

     
  

~I..

   
   
 

  

-4,
*2... $3.4,»
‘m

 

 
   
  

 

. TEL}.- _.-.- . ‘W‘mt...
’~‘3"5‘:'33:E<\L.‘~‘~}5’r?a-;~.*‘JS ,
ageing, h. “‘“QVzcafij ‘
' - «xxx: 1‘ 'Wfinx. ‘ um,
r - A?» ” 1:: ”mg-“A " {$4.435
.... [ta-If}

   

 

in .;.“Z%

   

    
    

  

'7'. ,

4:10,"

1 I ' '
113'." g \ a“. —
- -:’}<r ,4‘ I." "my:
I”.“’ :‘ -“. f u 1
I. ’I I ‘

  
        
    
      
 

 

    
     

  
   
 

’ ’4
. "n «2’55“? :15: ’- . ‘- . "x :9
“V3: .27.‘ {-3. [gig " > -'
'7 {;”;?‘: ’13:”?! 'fjg,‘

 

   

 

, .12).?-

“IV/"$-

  
     

.
1""
'1 1‘: ‘2‘."

«"7591?

4'
:71" ’35: "
:1’6’.;{:M xii”, :/}:::
’Iuo—V- .2.
’E’C'ml '

    
   
 
  

          
 

 
  
  
  
   
  
  

   
 
  

 
 
   
 

     
   
         
 

K3- .7 .‘-.‘.:.-c Kath 7‘ -

  
 
 

  

    

 

  
 
       

     
    
      
         
     
    
  
  

      

  

  
 
    
   

  

  
 

      
  

     

 
     

 
 
        

        
 
 
 

         
        
       

        
 

 

  
 
  
       
  
 
 

  
    
    
  
   
    
   
 

#24"
:21: 2;},

   
 
  

\- W’il/ ' 2‘93; I ‘ ézf, 7 I a 0‘
[76$ N‘é'.‘ '1. ‘2/ 58‘): ,7 :23?! A 71...”. ,9. t.
ffif’u' 7.- fhf ‘ ‘?,g7/{.':«;-I"-
- , I: - ' -‘.' "’7
‘ ’ ' \ v v
3" “r i ,-;/I’ 1'. rag-5’5,
;. "t "a fit"? "
3/21, 2:21: A ’ ,1 “i=2; a
" . "37/? "' 7-3/7 77'.
1 :01}, -
5, r - if
/' g . -.r
..- ) r, :71. ' ‘-
n' I" " "v rt" ’7‘ '- A I
~ 1" 3r?“ . If; “Itf’q'; ; '9‘: 1’9")":
- ‘1’ . N .31}
.'I 1.. ‘2': 4, '7 /'.r .71.».- diali‘; I "M. {ff/ 'f. "n .,
«J "J; fin/1’7 if“; n .’ '~. 5/1,:"3337 ”"7: ‘13:!
- . . . J. _ .-,. , .
:1 , ’ ~le 2, fur/if. 5'7"!”
v ‘fi/fl'r £531,422, N 2’, :yj/l
. , ‘1 . »
)3 (/11; j!)
. ~,11.
' :5! .
3

Q
,1”!
l' ”(1”
12151-7317.

 
     
  
  

      

-
./ f5? ',
:j:.7",’-
f’I‘rJXJJ"’ \vaa: 3:354!"
-'

  

   
 
   
 
 
 
  

   
  
       
   
    
  
 

    
 
 
 
  

    
 

 

  
      

, .J', 1:7}: '75:}?!
'7: y" . ’. $17.17,). 7", . ‘ ' Ir '9'“
: . w /- 1.. ., ,. ,1... - .. r372, 1;. “rzw.
,, :1:- . " )‘ ' I I‘M I”, .JI/r, _)1,’ ‘1’ 7", fl JCZZI’I.’ é]?! I’é'ir;-;gl
- . "(01'1” 2 'r
é {21"1"; 7”?" 1:359 f'lé'ill ”A" I}, 'inl; JJ’" £1,536; ' (w
.1 -. ' ' . 4. ‘:
'4 fi’v'! '3' MW. I»? ~. / '3’ g‘: 42"”? {43:
J ' ’l ’7', {le‘ 71'
. r' 1'7;
‘I‘ l"
l

 
  

19;;
‘1’ .11, .J
-‘c,".;. Xh/i jgflffl' in.

  

  
  
   

       
   
    

  
     
 
 
 

. , . '33
11 4,12,, 1; ... ""335...

,1
'. /.?/5'/7‘1f;,.;
.1: :{J‘ég’ ;;/;n::","'." “5.3
$45-1"’/"v-r~
”I :1: "I”.Y/ " {W
'1

       
 
   
 
 
 

“755.4219 35" I”; a":
1&3, 1.] X411 Jr:

53:.
7,1,,” fife.“

,
1
I
I" .,
géfi' ii?
.I- I.
£3

     

     
 

  

 

           

 

 

             

fl:?.:’
.21.”. r
;,(".I‘/,’
I .

 

' '11.-
?” .'

 

 

'. .
I 1:)?" I". ’00,} ‘4' /:‘:{E' ‘
/ wry ”."W , a ‘deuwéay
' n'xvfifl- :flwcw
750/12,? "
{IL/m; //- LL17"- Jfib "$4,, ;/;
W’ I - 77/

I
/ 141%,)345’;
,1...ng " .4: 7‘2“.” if
v S,-
M .1.

h; ‘1'!" ,
..r «(a
7.4!,“ ‘ $573!,-
I.’ 91""? f.” In «cant,
I $571), It?!
1"».1:
vhf/74 {hail {115/

ZI'J”
if???"
Sufi/7. 32:51.;
' "mm

 

 

9.0-7 3 S7“?!
anmnmin at.w.WW L
31293 00 . ”B Y
Michigan State
University

CH!

W m

 

 

 

 

 

 

 

This is to certify that the

thesis entitled

 

Design Criteria for an On-line Viscometer
for Baby Food Puree

presented by
Marc Christian Marcel Vercruysse
has been accepted towards fulfillment

of the requirements for

M.S . degree inAgricultural
Engineering

%~225%

 

 

 

Date—431M

0-7639 MS U it an Ami-main Action/Equal Opportunity Institution

 

 

 

MSU

LIBRARIES
n

 

 

RETURNING MATERIALS:
Place in book drop to remove this
checkout from your record. FINES
will be charged if book is returned
after the date stamped below.

 

 

 

 

 

DESIGN CRITERIA FOR AN ON-LINE VISCOMETER

FOR BABY FOOD PUREE

by

Marc Christian Marcel Vercruysse

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

golflo'fviK

ABSTRACT

DESIGN CRITERIA FOR AN ON-LINE VISCOMETER

FOR BABY FOOD PUREE

by

Marc Christian Marcel Vercruysse

Rheological properties of four baby food products
(applesauce, carrot puree, vegetable and beef dinner, and
turkey and rice dinner) were investigated with a Bostwick
Consistometer and a concentric cylinder viscometer from 73
to 92 °C. A power law fluid behavior was found for all
products. The relationship between the two methods of
measurement, considered in terms of the Bostwick readings
and the apparent viscosity, was very poor. The effect of
temperature was product-dependent with applesauce and turkey
and rice dinner being the most temperature sensitive.

Rheological properties for turkey and rice dinner were
also investigated with a tube viscometer. Different
constants were obtained for the consistency coefficient, but
the fluid could still be characterized by a power law model.

The design criteria for an on-line viscometer were
determined based on a sensitivity analysis and the
rheological properties. This analysis showed that the flow

behavior index influenced the pressure drOp the moSt. The

the consistency coefficient on the pressure

influence of

drop depended on the values of the other variables. The

a much less sensitive variable, which was a

flowrate was
positive point, considering wear of the pump with time.

ACKNOWLEDGMENTS

I am deeply indebted to my major professor Dr. J.F.
Steffe for suggestion of the research topic and his
continuous support, interest and encouragement throughout

the course of this study.

Sincere appreciation goes to the guidance committee
members : Dr. J.B. Gerrish, Dr. R.Y. Ofoli and Dr. A.K.

Srivastava.

Special thanks and appreciation are sincerely expressed
to the Gerber Products Company for donation of the samples,
plant tour, and continuous interest in the progress of this

study.

Very special thanks are given to Bob Dail, who helped
in carrying out the experiments with the tube viscometer. I
also want to thank Brian, Kim, Larry and Robin, for their

practical help.

Most importantly, I wish to acknowledge the loving

support of my parents and Karine.

ii

TABLE OF CONTENTS
page
LIST OF TABLES .................................... vi
LIST OF FIGURES ................................... viii

LIST OF SYMBOLS O...O...OOOCOOOOOOOOOOOOOOOO0.0...O Xi

CHAPTERI INTRODUCTION OOOOOOOOOOOOOOOOOOOOOCOO...
1'01. IntrOduCtion O...OOOOOOOOOOOOOOOOOOOOOOO0.0...

1.2. An actual process line .......................

p A) ta id

1.3. Objectives OOI.....OOOOOOOOOOOOOOOOOOOOO0.0...

m

CHAPTER 2 LITERATURE REVIEW .......................
2.1. Bostwick measurements ........................ 6
2.2. Fluid models ................................. 8
2.3. Rheological measurements ..................... 13

2.3.1. Concentric cylinder viscometer ........ 13
2.3.2. Tube viscometer ....................... 15

2.4. Rheological properties of the products under
StUdy .0.0.0...OOOOOOOOOOOOOOOOOOO0.0.0000...O 15

2.5. Effect of temperature on the rheological
properties O..OOO.I.OOOOOOOOOOOOOOOOOOOOOO0.0. 19

2.6. Relationship between Bostwick Consistometer
readings and apparent viscosity .............. 21
CHAPTER 3 BOSTWICK CONSISTOMETER AND APPARENT
VISCOSITY MEASUREMENTS O... 0.. O O O O O O O I. O 23
3.1. BOStViCk anaIYSis OOOOOOOOOOOOOOOOOOOOOO00.... 23
3.1.1. Experimental procedure ................ 23

3.1.20Resu1ts O..00.0...OOOOOOIOOOOOOOOOOOOOO 25
3.1.3. Discussion ............................ 27

iii

page

3.2. Rheological parameters obtained with a

concentric cylinder viscometer ............... 34
3.2.1. Materials and methods ................. 34
3.2.2. ReSUlts ..00.0.00.0...000.00.000.000... 36
3.2030 DiscuSSion 00.0.00.00.00.00000..000000. 41
3.3. Relationship between Bostwick Consistometer
measurements and rheological parameters ...... 49
3.4. conCIUSions 000..000.0.00.00.000.00...000..... 52

CHAPTER4 TUBE VISCOMETER 0.0000000000000000000... 55

4.1. Introduction ................................. 55
4.2. Materials and methods ........................ 56
4.3. Results ...................................... 50
4.4. Discussion ................................... 60

4.5. Accuracy of the results - Sensitivity
analYSis 0.0...O..O.....0.0000.0000000..000... 71

406. conCIUSions 0.0..0.00.0..00...00...00000.0..00 78

CHAPTER 5 SENSITIVITY ANALYSIS AND

DESIGN CRITERIA ........................ 79
5.1. Introduction ................................. 79
5.2. Sensitivity analysis ......................... 79
5.3. Design criteria for an on—line viscometer .... 82
5.4. Conclusions .................................. 85

CHAPTERS CONCLUSIONS 0000.00.00.00000..0000.0.... 88
CHAPTER 7 SUGGESTIONS FOR FURTHER STUDY .......... 90

APPENDIX A BOSTWICK CONSISTOMETER READINGS FOR
. DIFFERENT PRODUCTS 0.0....0...000.0000. 91

iv

APPENDIX B

APPENDIX C

APPENDIX D

BIBLIOGRAPHY

page

RHEOLOGICAL CONSTANTS OBTAINED WITH THE CON-
CENTRIC CYLINDER VISCOMETER FOR DIFFERENT
PRODUCTS FOR THE SEPARATE RUNS ........ 93

THE APPARENT VISCOSITY AS A FUNCTION OF THE
SHEAR RATE AT DIFFERENT TEMPERATURES FOR THE
DIFFERENT PRODUCTS (RAW DATA) ......... 95

PRESSURE DROP, FLOWRATE AND TEMPERATURE DATA
FOR THE DIFFERENT TUBE VISCOMETERS .... 99

00.0.00..00.00....00000000.0.0.0000... 102

LIST OF TABLES

Table page
2.1. Rheological parameters for applesauce ....... 17

2.2. Viscometric constants for applesauce at 25 oC
(Prentice and Huber, 1983) .................. 20

2.3. Activation energy for viscous flow of applesauce
(Prentice and Huber, 1983) .................. 20

3.1. Bostwick Consistometer readings for different
prOdUCts .00....00.00000....000.000.....00.00 26

3.2. Apparent velocities and decelerations in the
Bostwick Consistometer for different
prOduCts 000000.00....00.0....0..000.....0..0 30

3.3. Viscometric constants for different products in
concentric cylinder viscometer .............. 42

3.4. Confidence limits at 95% for the power law
constants for different products ............ 43

3.5. Activation energy and constant for different
prOdUCts .0...0....0...0.0.0000.000.00...0.00 49

4.1. Viscometric constants for turkey and rice in
different tube viscometers ................... 63

4.2. Critical and generalized Reynolds number for
the different products ...................... 64

4.3. Error terms and partial derivatives of n and K
with respect to the variables for the largest .
tUbe Viscometer at 73 0C .000000000......0000 75

A.1. Bostwick Consistometer readings for different
prOdUCts 0OOO000.000.0000..0000.'0000000.00000 91

3.1. Rheological constants obtained with the concentric
cylinder viscometer for different products for the
separate runs 0.0....000000.000000.0....0.0.. 93

vi

D.3.

Pressure
the tube
at 73 oC

Pressure
the tube
at 73 oC

Pressure
the tube
at 73 oC

page

drop, flowrate and temperature data for
viscometer with a diameter of 0.0127 m
and 80 0C 0.000.00.00.00.00000000000. 99

drop, flowrate and temperature data for
viscometer with a diameter of 0.00635 m
and 80 0C 0......000000.00.00.00.0000 100

drop, flowrate and temperature data for

viscometer with a diameter of 0.003175 m
and 80 0C .00....00000.000.000.000... 101

vii

LI ST OF FIGURES

Figure page

1.1. Schematic view of a process line ............ 3

2.1. The Bostwick Consistometer
(DaVis et all, 1954) 00.00.0000..000000000000 7

2.2. Newtonian and time-independent, non-Newtonian
fIUids 0.0.0.0000.00.000.0000000000.000000... 11

2.3. Time dependent, non-Newtonian fluids :
thixotropy (A) and rheopexy (B) ............. 12

3.1. Thixotropy test for applesauce at 87 oC ..... 25

3.2. Typical flow pattern in a Bostwick Consistometer
with a product showing liquid-solid
separation 000.00.000.000.000.0..0...0000.000 27

3.3. Effect of temperature within the range of
73°C to 92°C on Bostwick Consistometer
readings (applesauce and carrot puree) ....... 32

3.4. Effect of temperature within the range of
73°C to 92°C on Bostwick Consistometer
readings (vegetable- -beef and turkey-rice) .... 33

3.5. The apparent viscosity as a function of
the shear rate at different temperatures
for applesauce .............................. 37

3.6. The apparent viscosity as a function of
the shear rate at different temperatures
for Garret puree 0.000.000.0000000000000000.. 38

3.7. The apparent viscosity as a function of
the shear rate at different temperatures
for vegetable and beef dinner ............... 39

3.8. The apparent viscosity as a function of

the shear rate at different temperatures
for turkey and rice dinner .................. 40

viii

3.9.

3.10.

3.11.

3.12.

4.1.

*4.2.

4.3.

405.

4.6.

5.1.

.the

Effect of temperature within a range of 73 °C to
92°C on the apparent viscosity (apple-
sauce and vegetable-beef) ....................

Effect oof temperature within a range of 73 0C
to 92°C on the apparent viscosity (carrot
puree and turkey- rice)

Bostwick Consistometer readings versus apparent
viscosity, evaluated at 200 5

Comparison of data in this study with data
obtained by Rao and Bourne (1977)

Schematic diagram of the tube viscometer
set-up 00.00...00.00.00000......000000000000.
Rheogram for two different tube viscometers at

73° C(turkey and rice)

Rheogram for three different tube viscometers at
80 0C (turkey and rice)

Evaluation of slip coefficient at 73 °C (turkey
and rice dinner)

Evaluation of slip coefficient at 80 oC (turkey
and rice dinner) 0000..0.......000...0.0.0.0.
Velocity profile of a power law fluid in a pipe
without slip (a) - with slip (b)

Rheograms of turkey and rice dinner obtained with

a concentric cylinder viscometer and a tube
viscometer at 73°C and 80°C

Final design criteria for the proposed on-line
viscometer

apparent viscosity as a function of
shear rate at different temperatures
applesauce (raw data)

The
the
for ....................
apparent viscosity as a function of

shear rate at different temperatures

carrot puree (raw data)

The

for

ix

page
47

48
51

53

57
61
62
66
67

68
72

86

95

96

The
the
for

The
the
for

apparent viscosity as a function of
shear rate at different temperatures
vegetable and beef dinner (raw data) ....

apparent viscosity as a function of
shear rate at different temperatures
turkey and rice dinner (raw data) .......

Page

97

98

LIST OF SYMBOLS

A : constant (Pa 5)
BR : Bostwick Consistometer reading (cm/s)
D : diameter (m)

Ea : activation energy (J/mol)

h : height of the bob (m)
K : consistency coefficient (Pa 5“)
L : length of pipe between pressure measurements (m)

Li : length of pipe between tube viscometer inlet and
first pressure measurement (m)

M : external torque (N m)
n : flow behavior index (dimensionless)

AP : pressure drop (Pa)

volumetric flowrate (m3/s)

l0

correlation coefficient

'1
N
0. .0

coefficient of determination

radius (m)

R!

R9 : universal gas constant (8.314 J/mol K)
SSE : sum of the squares of the errors
T : temperature (°C or K)

T1 : temperature of the product at the first pressure
transducer (°C)

T2 : temperature of the product at the second pressure
transducer (°C)
T3 : temperature of the steam at the outlet (°C)

xi

Q

: temperature of the steam at the inlet (°C)
: effective slip velocity (m/s)
: velocity at the wall (m/s)
slip coefficient (m/Pa s)
shear rate (s'l)
: apparent viscosity (Pa 5)
Newtonian viscosity (Pa 5)
density of product (kg/m3)
shear stress (Pa)
: yield stress (Pa)
: primary normal stress difference (Pa)

: secondary normal stress difference (Pa)

angular velocity of the bob (rad/s)

subscripts :

b

C

bob
cup

wall

xii

CHAPTER 1

INTRODUCTION

1.1. Introduction

Rheology is that field of science which studies the
deformation and flow of matter, in response to an applied
stress. Rheological properties of food materials have been
the subject of several studies. These properties are
determined for a number of purposes : quality control,
process engineering and design, understanding structure, and
correlation with sensory evaluation, are just some examples.

Quality control of a typical pureed baby food product
can be assessed at different levels of the process. First,
in the process line itself where adjustments can be made
before the product goes into the jar. Secondly, on the final
product, and finally, on the shelf (e.g. shelf-life
stability). The major concern is the consistency of the
product, more often referred to as the ”thickness”. The
consistency can be characterized and quantified by
rheological parameters.

It is a real challenge to investigate these parameters
in the process line. Typically, this can be achieved in two
ways : either on-line or in-line. An on-line measurement

requires sampling from the the main line, performing a

measurement and, if desired, returning the sample to the
main line. The in-line measurement performs the task

without sampling.

The overall focus of this study is to develop design
criteria for an on-line viscometer for pureed (not chunky)
baby food products to assess the quality of the material.
Rheology is used to determine the quality parameters and the
design criteria for the on-line viscometer. The on-line
viscometer is a tube viscometer, which is compatible with

existing industrial pipeline systems.

1.2. An actual process line

An actual process line, for pureed fruit and vegetable
products, is schematically represented in Figure 1.1. The
raw product is fed into a steamer where it is mashed and
cooked. A valve controls draining to allow a desired
consistency. Then, the product goes into a series of
finishers (screen diameter : 0.7874 mm) and is finally
dumped into a surge tank, from where it is pumped to final
batch kettles located before the jar fillers.

Usually, a sample is taken out of the batch kettles and
a consistency measurement is performed. A frequently used
instrument is the Bostwick Consistometer, which is described
in section 2.1. The reason why this is so frequently used
in the industry is that it is inexpensive, fast and simple.

However, the disadvantages are that it is labor intensive,

.mcfia mmwooua e no 3mm> ofiumEocom .H.H muammm

 

 

mumHHHm nae

 

 

 

avowedoumv umumaouma> mnaaleo u >Ac
Aamauumv mwaapmmu umumaeumwmeoo xowsumom u Mm

muesmficfim " m

 

 

 

 

 

 

 

 

 

 

 

 

 

 

mfluumx mauumx
oumm oumm
n n >40 m
1 lllll :4
HM _ _ Jana
/. / _ _
,//, / mwusm . , eweewmuo
// /. _ m . ammum
/// / m>ag
/ /,
/ .,
//,.,.,.
x mm umxooo

I amateum

1

uospoua 3mm

 

 

 

time consuming and not suitable for on-line consistency
measurements. Also, from a rheological point of view, the
Bostwick Consistometer provides only one single value, which
lumps all the rheological properties together.

It is the objective in the future to substitute the
Bostwick Consistometer with an on-line viscometer. The
location of this viscometer would be best just after the
surge tank and before the batch kettles. The rheological
information from the viscometer can then be used to make
adjustments at the point of the steamer, by opening and
closing the drainage valve, at the point of the batch
kettles, water can be added or removed. The on-line
viscometer will give more accurate rheological information
and it is also fast and simple. In addition, it will lend
itself to better control and automation of the process,

eventually leading to feedback or feedforward control modes.

1.3. Objectives

The objective of this study was to develop design
criteria for an on-line tube viscometer based on the
information from the Bostwick Consistometer and the
rheological properties measured with a concentric cylinder
viscometer and with a tube viscometer.

The specific objectives were :

- to carry out the Bostwick measurements and to obtain the
rheological properties of four baby food products

(applesauce, carrot puree, vegetable and beef dinner, turkey

and rice dinner) at processing temperatures (70-90 °C),
using a concentric cylinder viscometer;

- to investigate a relationship between the Bostwick
measurements and the rheological prOperties of the above
products;

- to obtain preliminary, on-line viscometer design
information by running one product (turkey and rice dinner)
through a tube viscometer;

- to establish design criteria for an on-line tube

viscometer for pureed baby food products.

CHAPTER 2

LITERATURE REVIEW

2.1. Bostwick measurements

The Bostwick Consistometer is a simple, cheap and fast
way of measuring flow properties of a fluid. The instrument
(Figure 2.1) consists of a rectangular trough, 25 cm long,
made out of stainless steel, whose floor is graduated 'in
increments of 5 mm. One end has a spring-loaded valve and
encloses a compartment with a capacity of 100 cm3.

At the time of the measurement the sampling container
is filled to the top and the valve is released. The fluid
starts to flow under its own weight, and movement down the
trough reflects the fluid properties. The instrument has a
fixed inclination to increase the flowrate and to overcome a
small yield stress, if present. Measurements are taken
after a fixed time and are reported in units of cm/s or
cm/min, thus the appropriate physical quantity would be a
velocity. However, since the time is fixed for a particular
test, readings are actually distances.

There is no agreement in the literature upon the fixed
testing time : Bookwalter et al. (1968) read the Bostwick
value after exactly 1 minute, Davis et al. (1954) and Luh et

al. (1954) took readings after 30 seconds, Rao and Bourne

 

Figure 2.1. The Bostwick Consistometer (Davis et al., 1954).

(1977) obtained their data after 10 seconds. A typical
industrial testing time for pureed baby food products is 5
seconds. Clearly, the testing time will be a function ‘of
the consistency of the product and of the speed one wants to
perform the measurements.
Many variables affect the readings. Some are given

below :

- the loading of the product,

- the testing time,

— the nature of the product,

— location of the reading (center or wall),

— inclination of the trough,

— geometry and surface roughness of the trough,

— temperature of the trough.
Again, although measurements are very simple and
straightforward, a lot of variables are involved and they
all influence the reading. Therefore, Bostwick readings are
only meaningful if they are totally described and specified.

Davis et al. (1954) reported that they took an average

of the values at the center and both walls. Marsh et al.
(1980) used a special consistometer with a trough, 50 cm

long, made of plexiglas acrylic plastic.

The effect of temperature on the Bostwick reading has
been investigated by Davis et al. (1954). They reported a
linear relationship in a range from 21 to 87 0C for tomato

puree.

Data, related to this study, were obtained by Rao and

Bourne (1977)

.0

- applesauce : 6.73 cm/lOs at 25 °C;

- carrot puree : 2.68 cm/lOs at 25 °C.

The Bostwick Consistometer has been and is still used
extensively in the industry. It is an empirical method, and
as long as one is consistent in taking the readings it can
be a meaningful number. However, it lumps all the
rheological properties together, since just one value is
reported, which is the distance after a fixed testing time.
A more modern approach characterizes the fluid with
different parameters. Section 2.3 will describe two methods

to accomplish this.

2.2. Fluid models

Rheological properties of fluids can be characterized
by different fluid models. An extensive review of the

official nomenclature for the material functions has been

9

carried out by Dealy (1984). The three major material

functions, which are generally accepted are (Rao, 1986) °

- the apparent viscosity : na(i)

- the primary normal stress difference - Yl(§)

- the secondary normal stress difference ° ?2(?).

In this work only the first material function will be
considered; the two other material functions are probably

zero and insignificant in this study.

The apparent viscosity can be used to classify the flow

behavior of several foods. The simplest model is a Newtonian
fluid

 

“a- "u oeoooooooooooooeoooooooo0000(201)

More complicated models involve more rheological

parameters. The Herschel-Bulkley model (Herschel

Bulkley, 1926) is

and

Q

Ila-

 

 

...............(2.2)

Equation 2.2 collapses into Equation 2.1, if n =
00

l and co =

Power law fluids are a special case of the Herschel-

Bulkley model where the yield stress is equal to zero

10

-n-1

.......................(2.3)

 

II
N
-<

"a

If the flow behavior index is smaller than 1, the fluid is
said to be pseudoplastic or shear-thinning. When n is
greater than 1, the fluid is called dilatant or shear-
thickening. An overview of the different fluid models is
given in Figures 2.2,a and 2.2,b. Figure 2.2,a is called a
rheogram and plots the shear stress versus the shear rate.
In Figure 2.2,b the shear stress is divided by the shear
rate, to yield the apparentviscosity. It shows how the
apparent viscosity is a function of the shear rate for non-
Newtonian fluids.

It must be emphasized that many other models exist and
they may even describe the fluid even more accurately, e.g.
over greater shear rate ranges (Ofoli et al., 1987).
However, the Herschel-Bulkley or power law models are most
often used in the food industry, because they are general
models, fit most foods and can be integrated to solve many

process engineering problems.

So far, only time-independent models have been
considered. A plot similar to Figure 2.2,b can be
constructed, with time as the independent variable. The
apparent viscosity, at a constant shear rate, is now a
function of time. Figure 2.3 shows two possible time-
dependent effects : thixotropy (curve A) and rheopexy (curve

B). The first effect is a decrease of the apparent

(Pa)

SHEAR STRESS

(Pa 3)

APPARENT VISCOSITY

Figure

 

 

11

Herschel—Bulkley

Bingham plastic

Pseudoplastic

ii(‘VJll()ll'iilrl

Dilatnnt

 

 

SHEAR RATE (l/s)

(a)

     
 

Dilatant

Newtonian

Pseudoplastic

 

SHEAR RATE (l/s)
(b)

2.2. Newtonian and time-independent, non-Newtonian
fluids.

12

  

 

B
’3
m
8.
>4
94
H
U)
8
3 Time-independent fluids
>.
94
2:
E
a.
Q Thixotropic
A

 

 

TIME (min)

Figure 2.3. Time dependent,‘ non-Newtonian fluids :
thixotropy (A) and rheopexy (B).

viscosity with time, while the latter is an increase. The
time-independent fluids are represented with a horizontal
line.

The rheological model for thixotropic fluids can be
based on one of the previous models, multiplied by a
structural parameter. Tiu and Boger (1974) employed the
Herschel-Bulkley model and assumed the structural parameter
to obey a second order rate equation. Other approaches have
been taken by Higgs and Norrington (1971) and Tung et al.
(1970).

In general, after a certain amount of time the
structural decay stops and an equilibrium is reached; no

further breakdown or building up will take place. Then, the

13

time effect is said to be removed.

Finally, as mentioned earlier, fluids can show
viscoelastic behavior; those fluids have both viscous and

elastic properties. They will not be considered in this

study.

2.3. Rheological measurements

There is a wide variety of instruments to measure
rheological properties of foods. It is beyond the scope of
. this work to describe all the instruments. More interested
readers are referred to Van Wazer et al. (1963) and Whorlow
(1980). Here, the concentric cylinder viscometer and the
tube viscometer will be discussed because they were used in

this study.

2.3.1. Concentric cylinder viscometer

The determination of the shear rate in a non-Newtonian
fluid, sheared between rotating coaxial cylinders, has been
mathematically solved by Krieger (1968). He showed that the
shear rate at the bob for an unknown fluid can be expressed

as

2 n (£1,927S

[ 1n + 52 5' run ....<2.4)
s (Rc)2/s-(Rb)2/S

 

 

(b =

l4

 

d(lnab)
where s = -——————-
d(lna)
d(l/s)
5':—
d(lnab)
2 R
t = -—- ln(-—E—)
s Rb
x[ex(x-2) + x + 2]
f(x) =

2(ex-1)2

while the shear stress at the bob is expressed as

M
ab = ..............................(2.5)

2 n h Rb2

 

The assumptions for this solution were :
- the flow is steady; the system is at steady state,
- the fluid is incompressible,
- the flow is laminar,
- no heat generation : isothermal flow,
- the fluid is time-independent,
- the fluid is homogeneous and isotropic,
- gravitational effects are negligible,
- no viscoelastic behavior,
- no slip,

- no end effects.

15

Keeping the above assumptions in mind, a rheogram (ob
versus lb) can be constructed and a model can be fitted

through the experimental values.

2.3.2. Tube viscometer

It is also possible to evaluate the rheological
parameters from a consideration of the pressure drop and the
volumetric flowrate in a tube. Rabinowitch (1929) and

Mooney (1931) developed the following expression

 

. 3 Q d(Q/nR3)
Yw‘: —_—3—+aw ooooooooooooo(2.6)
n R daw
and
AP R
Ow=——'— oooooooooooooooooooooooooooooo(207)
2 L

The assumptions that were made to obtain Equations 2.6
and 2.7 were the same as those mentioned to obtain

Equations 2.4 and 2.5.

2.4. Rheological properties of the products under study

Many workers have conducted studies to determine the
rheological properties of several foodstuffs, but

surprisingly enough, no real agreement has been found among

16

the results for one particular product. Holdsworth (1969,
1968), Rao (1977), Steffe et al. (1983) and Dervisoglu and
Kokini (1986) reviewed published values for non-Newtonian
fluid foods. Differences among results for similar products
can be attributed to the instrument of determination, the
shear rate range, the soluble solids content, the variety
and the composition (particle size), and the rheological
model fitted. An overview of applesauce, obtained by
several researchers, is given in Table 2.1. It was not the
intention to summarize all the existing data for applesauce,
but to give an idea of the variability of the reported
parameters.

Rao et a1. (1986) investigated the effect of cultivar,
firmness and processing parameters. The n values ranged

from 0.29 to 0.39 and the K values ranged from 7.38 to 19.95

Pa 5“.

To eliminate some of this variability, a collaborative
study has been carried out by Prentice and Huber (1983).
The only variable was the technique to determine the flow
curve. The same products, including applesauce, were
shipped to the different labs. Table 2.2 summarizes
applesauce results (at 25 0C). One can see that the

rheological parameters were indeed closer together.

Data for carrot puree were reported by Rao et al.
(1977). The flow behavior index was 0.228 and the
consistency coefficient was 24.16 Pa 3". ‘No yield stress

was reported for both products. Only Charm (1962) and

17

Anna.“

o.m
m..—

n®.m

00.

m.

5.5

mm.o—

mu.

#0.

C’s—om to
LauoEOumm_o

couc._>u
uvcucoucou

LouoEoumw>
cosh

h>a LauOEoumr>
O_o*$xoocm

cauosoum.>
axon:

LouoEoumv>
monk

cauoEouav>
—_*LLO!

>4 LmuoEouov>

O_O.onocm

>4 LauoEoumr>

O-O—wJOOLm

couoEoum+>
_——LLO!

p.—I—.o
ommIm.m
CON—Iow—
CON—Iow—

DON—Iow—

omlm

omIm

aha—INV.Q

Dom—Ion
com—Ion
com—tom

Dom—low

ocLJom ocm on: I nmN.
I one.
“ohm—V

mesh 0cm amount I won.
I com.
I omn.
Ammmpv moum>Mme I own.
I ocm.

Asom_v Lose:
Ucm moum>mcmm I ocm.
Ammo—v Ecmcu w.mm one.
I wow.
I mvo.
I —om.
Amom.v etmco I won.
I one.
I owe.
I mvw.

Ammmpu
..mccoz Ucm Ecmcu I mvo.
Away mmocum AIV
mucoc0&mm U_op> c

to. cozoa
so. Luzon
3a. Logan
to. Logan
30— cozoa
3m. Lmzoa
to. colon
so. cozoa
>o_x_:m
I.o£oucox
10— Logan
so. coxoa
so. cozoa
3m. meoo
;o_ coxoa
to. Lozoa
30— cores
3m. Lozoa
.ouoz

oncwccu0h

Am\_v aunt Axe mo,.om

cmocm

I ma monomm_oo<
I ma
Luau: fimw .
I om 003mmo.oo<
0... mm as mom.o
m... hm EE o.o..
0... RN EE mh—.m
o~_m coccom
OUJQmO_Qa<
_. No
_— on ouzmmo_oo<
I I condom—oa<
I m.n~ a cents
I m.n~ < ocacn
I o.om
I ~.hN ounmmm_oo<
I on
I as
I um
I NN ounuwo_oo<
nuov
o_aa_om OcaaoLOQEOh aunuoca

.ooammoaamm so“ mumumEmqu HmuwmoHoonm .H.N magma

18

—m.—_

v.m—

mp.¢

m@.m

Nn.h

¢.®

>o.x.:m
I_o£umcor

3m. Lozoo

so. Lozoo
3m. Lozoa
3m. Loxoo
>o_x_:m

I_o£umcor

>o_x.3o
I_ocomcoI

couoEoomw>
mosh

couoEoumv>
monk

couc'_>u
uvcucoocou

'I|||'I-I"'-"||"-|"'|'I|Ill1"I'II"|I""III'II'IIIIII-|I""'"l'-l"IIII-""IIIIIIII"Il'-"'-l""""-|I-""'|"'I"

m can.
Ammo—V .cpxox
Ucm 3_oom*>coo I .hw.
I owe.
I omv.
Avmm_v ouuoum I owe.
on one.
Anna—v

no.0a 0cm mmobcmm on One.
Away mmocum AI“

oucocomom U_o,> c

oaowccuoh

ooo.I_ I ma
cow—I. I ma ounmmo_ao<
oompImh. m.m on
cow—Ioh. o.w ow
own—Ion. m.o_ mm ouammo_oo<
COMI. —.m_ mm m ucmcn
DOMI— N.mp mm ( Dcmcn
ouamno_oa<
Aux—V once Axv mo._oa Auou
Lmocm o_na_om ocauMLoo50h uuzooco

.Ao.ucouv .H.~ mflnme

19

Barbosa and Peleg (1983) found a yield stress for applesauce

(see Table 2.1).

No data were found in the literature for vegetable and

beef dinner, and turkey and rice dinner.

2.5. Effect of temperature on the rheological properties

Fluid foods are subjected to different temperatures
during processing and for this reason it is often useful to
know the effect of temperature on the rheological
properties. Arrhenius, in 1912, and De Gusman, in 1913,
(Holdsworth, 1971) independently suggested the following

relationship

Ea

R9""

) ........................(2.8)

 

na = A exp(

The magnitude of the activation energy, Ba, and the constant
A can be determined by analysis of experimental data.
Holdsworth (1971) also suggested that, in general, the
higher the activation energy the greater the effect of
temperature on the viscosity. The constant A can be
interpreted in terms of the molecular weight and molecular

volume of the liquid.

From a collaborative study on applesauce (Prentice and
Huber, 1983), the activation energy for viscous flow at a

shear rate of 10 s'1 was calculated within a temperature

20

Table 2.2. Viscometric constants for applesauce at 25 °C
(Prentice and Huber, 1983).

Laboratory K n 0
(Pa 5“) (~) (Pa)
""2; """""""" 51:6 ------ 6:;;_--_-____-:-__
B 28.4 0 27 -
C 16.8 0.35 -
D 32.6 0.29 -
E 30.0 0.33 8
F 30.1 0.25 -
G 36.1 0.28 5.5

Table 2.3. Activation energy for viscous flow of applesauce
(Prentice and Huber, 1983).

21

range of 10-45 0C. These results are presented in Table
2.3.

Charm and Merrill (1959) reported an activation energy
of 5.44 kJ/mol for applesauce. Saravacos (1970) reported a
value of 5.02 kJ/mol for applesauce with a 110 Brix value at

a constant shear rate of 100 5'1.

It is interesting to note that, in all the above cases,
a logarithmic transformation of Equation 2.8 was performed

to determine the constants,

Ea
log na=logA+ — ooooooooooooooooooo(209)

RgT

A plot of log "a versus l/T yields the necessary parameters,
after linear regression. No attempt was made to perform
a non—linear regression.

Rao (1977) summarized activation energies for several
products; however, no data were found for carrot puree,

vegetable and beef dinner, and turkey and rice dinner.

2.6. Relationship between Bostwick Consistometer readings

and apparent viscosity

The investigation of the relationship between the
Bostwick readings and the apparent viscosity is one of the
objectives of this Study. Actually, one is trying to

correlate a subjective measurement with an objective

22

measurement.

Rao and Bourne (1977) have correlated Bostwick
Consistometer data to measurements on a plastometer of
Eolkin. The relationship obtained was valid for all purees

investigated and was found to be

109(HB‘HA) = 0.075 _ 0.065 BR oooooooooooo<2010)

where 0A : apparent viscosity at 400 s-1

"B : apparent viscosity at 50 3’1

BR : Bostwick reading after 10 s (cm/105)

It is important to note that this correlation was only
valid for shear-thinning purees. Also, they concluded that
the structure of the food played an important role during
the measurement of the consistency with the Bostwick
Consistometer. This method was not suitable for
pseudoplastic non-pureed foods; it was believed that
adhesion of the non-pureed foods to the gate of the Bostwick

Consistometer was the reason.

CHAPTER 3

BOSTWICK CONSISTOMETER AND

APPARENT VISCOSITY MEASUREMENTS

The objective of this chapter is to investigate a
correlation between the Bostwick Consistometer readings and
the apparent viscosity, determined with a concentric
' cylinder. The main difference between the two techniques is
that the Bostwick measurement is an empirical method, while
the determination of the apparent viscosity is more a
theoretical or fundamental method. However, both are a

means of characterizing the fluid.

3.1. Bostwick analysis
3.1.1. Experimental procedure

In this study four baby food products, applesauce,
carrot puree, vegetable and beef dinner, and turkey and rice
dinner (Gerber Products Company, Fremont, Michigan) were
examined' at four temperatures : 73, 80, 86 and 92 °C. All
materials were received in jars as finished products, except
turkey and rice dinner, which came directly from the process
line and was subsequently stored in a cold room.

Since it was the objective to simulate the product in a

23

24

process line, it can be assumed that in an actual process
line, the product‘ is fully broken down. Therefore, the
product was stirred in its jar for 12 min until the
thixotropy was removed. This was accomplished with a mixer,
turning at approximately 400 rpm. A three-blade marine-type
mixing propeller was used. The mixing occurred while the
jar remained in a temperature controlled waterbath. A lid
of a jar was attached to the rod of the mixer to minimize
evaporation of the product. This whole procedure was
checked for complete breakdown and for temperature with a
concentric cylinder (Haake Rv-12, see section 3.2.1) and' a
thermocouple attached to the bob. A thixotropy test was
performed by recording the torque over time, while the
product was sheared at a constant speed (50 rpm). A raw
curve is shown in Figure 3.1. The increase in torque in the
beginning was due to the decrease in temperature, which was
caused by contact of the colder bob with the product. The
slight decrease of the torque further on, was due to the
temperature increase of the product again. Almost the same
torque was obtained as when started (at time = 0). It can
be seen that the thixotropy was completely removed. The
temperature was controlled with a waterbath within 1 1 OC.

The fully broken down and heated sample was poured into
the compartment of the Bostwick Consistometer and the gate
was opened. Readings were visually recorded after 5, 15, 30
and 60 s. The time was measured with a regular watch.

As pointed out in section 2.1, the method to determine

25

 

 

 

 

n a 4 _.______-___,_,___ e _
E
I
1.99:1 1
LL)
3
12:
0:
D .9131
I...
an 13 . e
1%—aaa 2 91 4.51 6.02

TIME (minutes)

Figure 3.1. Thixotropy test for applesauce at 87 0C.

the consistency of a product with a Bostwick Consistometer
is very empirical and subjective. All the experiments were
carried out with the same trough and were operated by one
person; in that way, the accuracy on the readings was always
judged in the same manner. This eliminated the human and
instrument variability. The trough was always dry and clean
before each run and was washed with water at the same

temperature after each run.

3.1.2. Results

The results are presented in Table 3.1. Four
replicates were taken for each product at each temperature.
The average values are shown and raw data are given in

Appendix A.

26

m..—
m.——
m.o—
m.o—

nwacw.

..m. c.4—
m.v— ~.v.
m.v_ N.v_
m e. o v.
0.5 0.0 0.5
m.n m.0_ m.»
0.0 0.0 0.0
0.0 «.0 0.0
n... F...
0... 0.0.
0.0_ m.o.
0.0. m.m

Amoo\500 usage. Anomers
now am mom mm

m.mp
m.m—

m.o—
¢.o—
h.m.
o.m

00:6,. Amo.\er 613a,.
mm. mm

Amm\EoV
mm mm

mm
mm
om
mu

Nm
mm
om
an

mm
mm
00
nu

mm
mm
on
ms

A000
ocnumL0Q60h

more new
>OxL3h

woos 0cm
o.om~ooo>

OOLJQ
uoccmu

ouammo_oa<

.muupouQ ucmcmwufip no“ mmcmbmms umumEOumMmcou xowzumom .H.m manna

27

3.1.3. Discussion

After some time, the flow pattern of the products was
curved with the highest velocity at the center (Figure 3.2).

Measurements were taken in the center.

‘Q

 

  

—
—

——
—_—
_

 

 

 

 

 

 

 

Figure 3.2. Typical flow pattern in a Bostwick Consistometer
with a product showing liquid-solid separation.

Carrot puree was the only product which showed a
separation of the liquid and the semi—solid portion, and has
therefore two values in Table 3.1. The value for the liquid
is the average of the values at the left and the right wall.
It was observed that the semi—solid part hardly moved and,
actually, only the liquid was flowing. This would suggest a
high yield stress for the semi-solid portion of carrot

puree.

It can be seen from the raw data (Appendix A) that the
data were very close together for each replicate. It has no
meaning, however, to do a statistical analysis on four data
points. The small variances observed can be attributed to

the variable amount of product for each run, since it was

28

not accurately measured (just poured into the compartment to
the top), and the force needed to release the gate, which
produced a different shockwave with each test. This
shockwave has not been described in the literature. It was
observed that this seemed to be a very important fact. So
far, it has received little attention. Different results
can be obtained if the shockwave is reduced by holding one's
hand on the instrument. Here, the gate was released without

holding the instrument.

As expected, the Bostwick readings increased with
temperature and obviously with time. Although the liquid
part from the carrot puree - the semi-solid part was
stationary - seemed to follow the same pattern, the increase
in distance with temperature was not always evident. It was
believed that the thermal history of the product was
responsible for this.

It is known that B-hemicellulosew holds water very
strongly; however, B-hemicellulose breaks down with a heat
treatment and its water-holding capacity decreases
significantly (Uebersax, 1986). If a sample is held at a
certain temperature for a longer time than another sample,
more separation will take place in the first sample.
Consequently, the temperature will not be the only factor
determining the Bostwick reading for similar products. The
reason why this phenomenon did not occur with applesauce was
that it was known that applesauce was very susceptible to

discoloration (pink color) due to a heat treatment

29

(Uebersax, 1986), and the samples were never preheated
longer than needed.
The other two products were meat products and were

flowing regularly in the trough.

The values in Table 3.1 can be viewed as ”apparent
velocities" and can be expresSed in cm/s. This is not a
real velocity because the product first accelerates and then
decelerates. However, the transition point from
acceleration to deceleration, for the products investigated,
was believed to be less than 55. The apparent velocities
are presented in Table 3.2. Since carrot puree did (not
flow, it was omitted. Apparent velocities decreased with
time and increased with temperature.

Now, the average rate of deceleration between two
observations can be calculated ; those values are also
tabulated in Table 3.2. For the same product the
deceleration was always slightly higher at higher
temperatures. It is interesting to note that the
deceleration is almost independent of the product. This
means that, once a value at a specified time interval is
known, values at other time intervals can be estimated.
This can be very helpful in comparing data from the
literature. Unfortunately, not many, if any, data at the
investigated temperatures and for different testing times,

are published to verify this hypothesis.

As mentioned earlier, Davis et al.(l954) found a linear

30

NNNN
OODO

NNNC’)
0000

000.0
000.0
000.0
000.0

000.0
—0.0
.0.0
.0.0

¢_.0 N
0—.0 p
Nw.0 m.
pp.0 0

0—.0 v N
0—.0 ¢.N
¢—.0 m N
n—.0 N N

N0
00
00
on

N0
00

Mb

oust 0cm
>oxL3k

0000 cam
o.ou~000>

sundae—Qa<

Am\Eu0
000 an
>uvuo—o>
acocooa<

An\2u0
mom on
xusuo—o>

co.umco_ouoo «cocnoa<

00.0 0.0
no.0 0.0
No.0 5.0
No.0 5.0
00.0 m 0
00.0 0.0
00.0 0.0
00.0 m 0
«0.0 5.0
No.0 5.0
No.0 0.0
«0.0 0.0
Au\5u0

1~a\5uc mm. a.

>uvuo_o>

covumco_ou00 «cocaoo<

Auxsuv

Aunxsuv am «I
xuwuo_o>
cosuoco_ouoo «cocmoo<

Aoov

ocauocooEOh

"-"""I'l'|'-"'-ll"-"|'|'|'II'III'"'l'l'""-"|'II"II'I-"""|'II‘|"l"5'---"'--"'-'I"l'|'lll"

xumsumom 0:» cm mcomumuwaouwv 0cm momumuoam> acmumamd .N

.muosooua ucmuouuflo an cmquOumwmcou

.m «same

31

relationship between the Bostwick measurements and the
temperature in a range from 20-90 0C for tomato puree.
Relying on this linearity, the following equations were
obtained for the different products at 5 5, including all

the raw data (see Appendix A),

applesauce : BR = .1062 T + 0.5957 ; r = .93
......(3.1,a)
vegetable and beef : BR = .056 T + 6.9814 ; r = .80
......(3.1,b)
turkey and rice : BR = .1019 T + 1.6526 ; r = .87
......(3.1,c)
carrot puree (solid) : BR = .06 T + 2.18 ; r a .88

......(3.l,d)

where T : temperature (°C)

BR : Bostwick Consistometer reading at 5 s (cm/55)

These regression lines are presented in Figure 3.3 and
Figure 3.4. No effort has been made to fit other models,
such as non-linear models, since only four points were
available. This regression is given to have a better idea
of the effect of temperature within the range of 70-92 0C on
the Bostwick Consistometer readings. Turkey and rice dinner
and applesauce were the most temperature sensitive, while
the other products were only half as sensitive.

The Bostwick reading for applesauce, reported by Rao

and Bourne (1977), 6.7 cm/lOs at 25 0C, is, considering the

32

.Ammusq
ocuumo can mosmmoaqmmv mocwvmwu souoEOmemcou xowsumom :0

00 mm o» 00 ms “0 moans on» cflnuwa ousumuoasou mo uomuum .m.m wusmfim

Gov mmazmmasfi

 

 

 

N0 mm .60 om 0N. NB
B — b n p L b L b — L n k — b b b — p b h m

. 8

I0 0

S

. M

1m“ mm

. VA

1mw mm

s G

10 INI

. 9

-2 m;

w

.. /

o . 9

.I—F as

3.50 09:00 I (K

I

00332000 0

 

 

Cd

33

. .Aooquamxssu
ocm mmmanHbmuomo>v mmcfiomou umDoEOBmflmcou gumsumom so
00 Nm Ou 00 mm «o omens on» cmnumz mssumcmqsmu mo uummum .¢.m ousmwm

Gov $351525
mm mm *0 . mm 05 Na.

[PL - b L — L b (r — n b h L F b h (P m

 

 

I0—

1
N
r-

 

—r

j r
L”. I:
(SQ/W3) ONIovaa MOM/(1808

8:185:30. <
fimmnlmfigmmg b

 

 

d-

34

regression, not in good agreement. Possible reasons are
that the applesauce used in their study was unstrained as
opposed to strained (used in this study), the extrapolation
to room temperature is not acceptable or the confidence in
the regression model is too small. The value for carrot
puree was 2.68 cm/lOs at 25 °C; no separation was mentioned,
which supports the thought that B-hemicellulose was

responsible for this phenomenon.

3.2. Rheological parameters obtained with a concentric

cylinder viscometer
3.2.1. Materials and methods

A Haake Rv-12 concentric cylinder viscometer was used
to collect data for the same baby food products considered
in the Bostwick Consistometer. The inner cylinder, the bob
(MVI), was rotating, while the outer cylinder, the cup, was
stationary. The height of the bob was 60.32 mm, the radius
was 20.04 mm. The radius of the cup was 21.00 mm. The
torque was measured and transformed into a proportional
electrical signal by the measuring drive unit (M150) .

A Haake PG-142 was connected to the measuring drive
unit to manually control the speed of the bob. The whole-
system was connected to a HP-3497A (HP meaning Hewlett-
Packard) data acquisition system, which was connected to a
HP-85 via a 82937A HP-IB interface. This was developed in

the food rheology lab at Michigan State University. The

35

whole system was calibrated over the full scale of the load
cell. Known weights were used to generate a known torque.
The voltage was displayed on the data acquisition system and
the calibration factor was found to be .00290 N m/V (the
previous factor was .00291). Then calibration fluids were
run with a Newtonian viscosity of 0.0505 Pa 3 and 0.985 Pa 3
in a shear rate range of, respectively, 5-350 5"1 and
5-100 s'l. No further adjustments had to be made; the
values were accurate to 1 0.75%.

Samples were treated in exactly the same way as they
were treated for the Bostwick measurements. They were
heated in a waterbath and stirred until thixotropy was
removed (12 min), while the product remained in the jar.
Then, the product was poured in the preheated cup and an
additional check for thixotropy was performed for each test.
The torque was recorded at a speed of 50 rpm and a pattern
similar to Figure 3.1 was observed. The rheological
measurements were only taken when equilibrium was reached.
Four replicates were performed for each test. The content
of each jar provided two samples.

Once the product was in the cup, the temperature
control was established with a temperature vessel (Haake FC-
3) built around the cup. The vessel had three functions :
it connected the sensor system to the measuring drive unit,
centered the bob and the cup and controlled the temperature
of the tested material. The vessel had two hose fittings,

which were connected to the waterbath. The temperature

36

could be controlled within 1 l 0C. A small protective lid
(Haake) covered the samples and prevented evaporation.
Tests were performed at four temperatures : 73, 80, 86 and
92 °C. The speed ranged from 2-150 rpm, resulting in a
shear rate range of approximately 10-400 s‘l, depending on
the product and temperature. Twenty five data points were
taken in this range for each test. Soluble solids content
was measured with an Abbe-3L Refractometer (Bausch and
Lomb). Values for applesauce and carrot were 9.0% and 6.8%,
respectively. The total solids content was determined with
a, drying oven at 103 0C for 24 hours. Values 'for
applesauce, carrot puree, vegetable and beef, and turkey and
rice dinner were 20.7%, 18.1%, 25.3% and 20.1%,

respectively.

3.2.2. Results

A computer program on the HP-85 provided the raw data
for each test : the speed (rad/s) and the corresponding
torque (N m). They were, together with the geometric data,
substituted into Equations 2.4 and 2.5 to yield data for the
shear stress, ab, and the shear rate, lb- A power law model
was then fitted to obtain a rheogram. Following this, the
shear stress was divided by the shear rate to yield the
apparent viscosity. Then, the apparent viscosity was
plotted against the shear rate.

Figures 3.5 through 3.8 plot the regression lines for

the different products at the different temperatures. The

37

.mosmwoaddm uOH mocsumsodsou acoumuufio
um mums ummzm on» no cowuocsu n no auwmoumw> ucmumddm one .m.m wusmmm

“Q 0 02m ”3%

 

 

 

 

com 00¢ 00AM OON OO— O
P H n F n h h (— b 000
”vaw
"XuNm
”Xvom
”XVNN

 

(8 Dd) MISOOSIA .LNEIEIVddV

 

 

 

38

.mwcsd bohemu ~00 mmssumuoaaou ucmuwmuwc
um mums umonm onu mo cofiuucsu m mm Nu0m0umm> ucmummam one .0.m musmfim

Am\ 0 BE mfixm

 

 

 

(8 Dd) MISOOSIA mam/adv

 

 

 

39

.uoccwo woos new oanouomo> so“ mousuouomsou acououuwc
um ouou soocm onu mo cowuucsw o no mummoomw> acouomdo one .>.m ousmwm

Am\ 0 BE mfizm

 

 

 

(8 Dd) MISOOSIA 1N3dVddv

 

 

 

40

.uoccwv oofiu cco Noses» ecu mousuosoQEou acououuflo
um ouou noonm onu mo comuoczm o no sawmoomw> acouoaao one .m.m ousmwm

Am\ 0 BE mfizm

 

 

 

(8 Dd) MISOOSIA 1N3dVddv

 

 

 

 

N;

41

rheological parameters and their limits, evaluated at 95%,
are given in Table 3.3 and Table 3.4, respectively. Each
regression line represents approximately 100 data points.
The rheological parameters for the individual runs are given

in Appendix B and the raw data in Appendix C.

3.2.3. Discussion

All the assumptions made in section 2.3.1 were
considered. The only assumption which could not be checked
was laminar flow. A criterion to predict the transition
point in a concentric cylinder has been proposed (for
Newtonian fluids by Dealy (1982). When rotating the inner

cylinder, secondary flow will occur if

RR ph h
b [ 11/2 > 41

...................(3.2)

 

u Rb

This can be checked for the available data. With h = 60.32
mm, Rb = 20.04 mm, o = 1055 kg/m3 (turkey-rice), Equation

3.2 becomes

 

> 18.5 ..............................(3.3)

The Idifficulty is now to substitute for the Newtonian
viscosity. One can use the apparent viscosity, but it is

known that this is usually not a good approach. If this is

42

.00.0 005.0 000.0I 050.0 0v.0 000.0 «05.0 «00.0 0¢.0 N0

N¢0.0 550.0 N.0.0 00¢.0 v0.N 0V0.0 550.0 00v.0 00.. 00

0¢0.0 .m0.0 5.0.0 0vv.0 0N.c 0Q0.0 .N0.0 00v.0 N..¢ 00 ou.c 0:”

000.0 500.0 N.0.0I 00¢.0 0Q.0 000.0 500.0 cmv.0 c0.m 05 >042.»

0N0.0 ..5.0 0.0.0 000.0 00.0 0N0.0 005.0 00v.0 N¢.0 N0

N00.0 000.0 000.0 00v.0 00.N N00.0 000.0 000.0 0..N 00

00..0 000.0 500.0 0.v.0 05.N N...0 0v0.0 .0?.0 5¢.N 00 0000 0:0

00v.0 550.0 5w..0 00..0 «0.0 500.0 0v0.0 c0v.0 «v.0 05 o.omuo0o>

.00.0 000.0 000.0 00..0 00.0 «00.0 000.0 0NN.0 00.N N0

00..0 500.0 N00.0 .QN.0 0N.0 00..0. 500.0 0vN.0 0N.0 00

5N0.N N00.0 .00.0 00N.0 00.0 .50.N 000.0 ¢0N.0 00.0 00 oocao

000.. 000.0 000.01 050.0 0N.v .00.. V00.0 00N.0 00.¢ 05 uOLLmu

000.0 050.0 .00.0 5¢0.0 .N.N 000.0 050.0 0Q0.0 0..N N0

000.0 000.0 000.0 N.0.0 00.N 000.0 000.0 c~0.0 00.N 00

000.0 5V0.0 500.0 000.0 50.0 00..0 5¢0.0 000.0 «0.0 00

N.0.0 000.0 050.0 000.0 00.N 5.0.0 000.0 va.0 00.N 05 ooamwo.oo<
Andy “IV 5cm may AIV Acm adv .000

000 Ne ob c x 000 «c c x oLJuMcoQth. uosooca
>o4x42mI4mrommor 244 mozoa

.souosoomw> noocwaao
ofluucoocou cw mDUDUOuQ ucououwdv so“ mucoumcoo omsuosoomw> .m.m oanoe

Table 3.4. Confidence
Product Temperature

(°C)
Applesauce 73

80

86

92
Carrot 73
puree 80

86

92
Vegetable 73
and beef 80

86

92
Turkey 73
and rice 80

86

43

limits at 95%

for

constants for different products.

2.33

3.12

0.51

0.56

the

law

44

l

done, transition points ranging from 4-7 rad/s (38-67 rpm)
are predicted, depending on the product and temperature.
However, it was thought that the flow was still laminar.
The problem is very complex and not well understood for non-
Newtonian materials.

Applying Equation 2.4 required a polynomial to
calculate the derivative ,5 , from a plot of 1nR versus lno.
A third order polynomial was fitted (the third degree
coefficient was always zero). The resulting correlation
coefficient was always higher than 0.99. Equation 2.4 was
further simplified by neglecting the correction term

2

s s' f(t). As pointed out by Krieger (1968) :

"Since f(t) has a maximum value of about 0.1, it is
evident that even the first correction ... will

always be small and frequently negligible."

Here, typical values for s were 0.25 and for s' a value of
0.8. This makes the correction term even smaller giving it

a magnitude of approximately 0.02.

Instead of calculating the rheological parameters for
each run separately - those results are given in Appendix B
- and then averaging, a non-linear regression analysis has
been carried out on the combined data from the four.
replicates (100 data points). The same technique was used
by Barbosa and Peleg (1983). Both the power law model and
the Herschel-Bulkley model were fitted. The sum of the

squares of the errors (SSE) for the Herschel-Bulkley model

45

was always less than or equal to the one for the power law
model, indicating a better fit with the Herschel-Bulkley
model (Table 3.3).

In three cases, the yield stress was calculated to be
negative, which is physically not possible. In the other
cases, the calculated yield stress was so small, that it can
be said to be zero. In practice, it is not even possible to
measure such small yield stresses. Therefore, from a
rheological standpoint, the power law model obtained by a
non-linear regression model was thought to be the best
model.

From the Bostwick Consistometer observations it was
expected that carrot puree would have a yield stress, since
the semi-solid part stayed stationary. The data obtained
here did not show a yield stress. In the concentric
cylinder, the product was homogeneous and no separation
could take place; in the Bostwick Consistometer, on the
other hand, separation could readily occur and, essentially,
a different product was evaluated. This is an example of
how a Bostwick Consistometer can provide totally different

information, because it reports just one value.

A higher temperature, generally, results in a lower
apparent viscosity, which can be observed in Figures 3.9 and
3.10. However, it seems that this is a function of the
product under consideration.

As opposed to the usual linear fit after a logarithmic

transformation, a non-linear regression was carried out on

46

the Arrhenius equation (Equation 2.8). The apparent
viscosity was evaluated at 200 5'1. Figures 3.9 and 3.10
represent the regression lines through the apparent
viscosities of the four replicates at each temperature for
the different products. The parameters obtained by this
method are presented in Table 3.5.

Results show that the average flow behavior index is
proportional to the activation energy. The higher the value
of n (less pseudoplastic fluid), the higher the activation
energy and, following Equation 2.8, the greater the effect
of temperature on the apparent viscosity. This is. in
agreement with earlier observed data on fruit purees and
juices obtained by Holdsworth (1971).

The fit. was worst for carrot puree, indicating that
there was almost no temperature dependency in this range.
Applesauce and vegetable and beef dinner showed a small
temperature effect, while turkey and rice dinner. was the
most sensitive product. The same trend was observed with
the Bostwick Consistometer, although applesauce was a little
more sensitive in those low shear rate ranges (Figures 3.3
and 3.4).

Table 3.3 indicated that the flow behavior index was
not a function of temperature, but the consistency
coefficient seemed to be affeCted significantly by
temperature variations. This effect is the same as that
reflected in the apparent viscosity, because the latter was

evaluated at a constant value and n was not a function of

47

..uoonIoHnouo0o> 0cm oosomoaadov 5u.moomw> acouoddo on» :o

00 «0 00 00 m5 no omcos o cwnu.3 ouSDosodsou mo uuouum .m.m ouommm

030 ..\.

 

 

 

wmmood ¢0N000 00000.0 05N000 05000.0
L I— L L L b L h L (— L L L L L L L o
000
. .
n00.0
INTO
T0..0
#-
I¢Nd
hoonIoBBomo> p .
ooaomofiao o a

 

 

('8 Dd) MISOOSIA 1N3dVddv

48

..ooMuIaoxusu 0cm oousq nouuoov 5ummoumw> ucouoaao on» co
00 N0 0» 00 M5 00 oocou o cwnufia ossuouodsou uo uoommm .o..m ousmmm

oi; K.

 

 

 

 

 

mmNood .ummood 00N00.0 05N00.0 N5N00.0
F — L Lr - — F L n b b L n L b L L o
000

. V

I0

.0; w.

v0

3

. N

III—

I . IA!

0N0 S

O

.. O

.6

I03 M

\l/

- d

D

S

. ootlxovtau < 10.....0 /I\

oocaa uotoo I

 

 

49

Table 3.5. Activation energy and constant for different

products.
Product A Ea r
(Pa 5) (kJ/mol)

Applesauce 9.2 10‘6 40.758 0.92
Carrot 7.4 10‘3 20.869 0.50
puree

Vegetable 3.2 10‘6 45.157 0.78
and beef

Turkey 4.6 10‘12 85.539 0.93
wand rice

temperature. Consequently, it is sufficient to relate the

temperature to the apparent viscosity only.

3.3. Relationship between Bostwick Consistometer

measurements and rheological parameters

In this section the data from the two previous sections
are brought together. It is known that the Bostwick
Consistometer yields data at very low shear rates, while the
data from the concentric cylinder viscometer are evaluated
at much higher shear rates. The two instruments are in a
sense not compatible. For the Bostwick Consistometer a
distance is reported and for the concentric cylinder
viscometer two or more parameters are investigated,
eventually leading to the apparent viscosity. However, the
apparent viscosity was evaluated at a constant shear rate of

200 5"1 (see section 5.3) to investigate a relationship.

50

First, each product was considered separately.
Previously, all relationships were developed with the raw
data; here, the Bostwick readings and the rheological
parameters were measured using different samples and
consequently, an average value had to be used. This brings
the number of data points down to four for each product.
Those four points represent the four different temperatures.
Figure 3.11 illustrates the pattern of the relationship, for
each product (the temperature is only indicated for one
product). Data points between the different temperatures
are only connected for the purpose of illustration.

As expected, the apparent viscosity tended to be
inversely proportional to the Bostwick readings. Applesauce
and turkey and rice dinner showed the best relationship and
suggested a straight line. The activation energy of turkey
and rice was very high, and the slope of both products in
the Equations 3.1 was higher than the other products. Both
are indications of their sensitivity to temperature. I The
other products were much less sensitive and only a trend was
visible : a decrease in the apparent viscosity with

increasing temperature.

Next, all products can be considered at one particular
temperature. This has been investigated by Rao and Bourne
(1977) for 28 different products at room temperature. They
found a single relationship for all non-pureed foods
(Equation 2.10).

It has already been shown (Figure 3.11) that no such

51

.mummoomm> ucouomdo msmuo> mmcfivoou pomm

m com um nouosao>o

an}... ozafim x0258

 

 

 

N. 0. 0 0
5 a _ _ L 00.0
A. 00 N0 I
06.00 .
.0 0 1I050
o 0 06 H5
I0N0
ootlxoxcfl «In I00 0
u.oonio33o0o> I
oocaa 09:00 I ..
ooaomoaao @Io

 

 

odxo

sobmmmcou sowsumom .Ha.m ousmflm

(8 Dd) MISOOSIA lNadVddv

52

relationship existed between the Bostwick readings and the
apparent viscosity. Connecting each first point of every
product shows the pattern at 73 °C; connecting each second
point at 80 °C, etc... One would expect a lower apparent
viscosity for a higher Bostwick reading and vice versa.
Here, the points were scattered. To illustrate} this, the
results of this study, for the lowest temperature and at a
testing time of 5 s, were plotted versus the results from
Rao and Bourne (1977) (Figure 3.12). As seen, their results
were not reproducible in this study.

. Finally, the slope of the regression lines (equations 3.1)
can be plotted versus the activation energies. Again a

scatter was obtained.

Summarizing, it can be said that the more the product
was temperature sensitive, the better the correlation with
the Bostwick reading. A single relationship for all

products was not found.

3.4. Conclusions

In this chapter the rheological properties of four baby
food products have been determined in two ways : empirical
and theoretical. No real agreement has been found between
the two methods, except for the turkey and rice dinner and
applesauce.

The combination of the temperature insensitivity of the

products, the small number of different products, and the

53

Abbmflv OCH—Jon 0C0

0mm an woodcuno pump no”: hogan was» as pump mo comduoasou .NH. m ouomnm

OED/mm xo_>>._.mOm

 

 

NF P P. n: m m n m
_ L L h b _ h . _ L b h . Foo
00 mm " ocnuocanB. ..
Ame—\Es mm
CkmC occaom oco oom IN.O
Into
00 mm " 839380... ITO
Amn\Eov mm

 

mto

 

(s Dd) Va _ Bu

54

small temperature range investigated, made it extremely
difficult to build a true relationship. The results at a
constant temperature, obtained by Rao and Bourne (1977)
could not be reproduced.

Although no relationship has been found, the target
change from Bostwick readings to apparent viscosities can
still be made. Quality control can now be based on the
apparent viscosity, however, this will be product specific.
How this can be done in an on-line viscometer will be

outlined in the next chapter.

CHAPTER 4

TUBE VISCOMETER

4.1. Introduction

Rheological properties of materials can be evaluated
from the pressure drop and the volumetric flowrate in a
pipe. This technique is very useful in the industry,
because it provides on-line information and thus on-line
quality control.

There is no need to work with delicate laboratory
instruments, which are often expensive, complicated and
time-consuming. No extra cleaning is required as opposed to
commercially available on-line rheometers. Cheng et al.
(1985) reviewed and evaluated the current applications of
some commercial process control viscometers. In their

discussion they point out that :

"Commercial instruments may exist to do a good job;
but the cost has a high profile. ... Homemade

instruments have been successfully developed ... '

This study is a first step toward a self-made on-line
viscometer. In this_ chapter tests were performed with
different sizes of tube viscometers. In the next chapter 'a

sensitivity analysis will be presented along with some

55

56

criteria.

4.2. Materials and methods

A schematic view of the total system is given in Figure
4.1. Two tanks were available : one tank (D = 0.7 m; h a
0.7 m) held water for cleaning purposes, while the other
tank (D = 0.8 m; h = 0.7 m) contained the product. A mixer
stirred the product continuously at a low speed (40-50 rpm)
and the whole system was made out of stainless steel.

Prior to testing, the product was stored in a cold
room. A volume of 230 liters of turkey and rice dinner was
divided into three equal parts, to run three different
tests. Three tube viscometers were available with an inside
diameter of 0.0127 m, 0.00635 m and 0.003175 m.

A positive displacement rotary pump, variable speed,
size 60 (Waukesha, Abex Corp., Wisconsin), was installed in
the system. The displacement of this pump was 0.568 1/min
with a maximum allowable pressure to insure seal integrity
of 1379 kPa.

A bypass was constructed just after the pump to allow
for a lower flowrate. An air-to-close valve was used to
control the flowrate. An additional air-to-close valve was
mounted at the end of the tube viscometer, to pressurize the
system and hence control the flowrate. When air was
applied, the orifice at the end was physically more closed,
putting more fluid through the bypass.

Flexible hoses were attached to the end of the bypass

57

.QSIuom souoeoumfi> was» mnu mo snowman owumEonom .H.¢ ouamwm

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ouommou . ~.H
m . m
in as m. o
Emmum mo ousumummaoh “ H 1\ . a
~.~ umumz
uospoua wo ououmummaoe u H
o>~m> mmoaououuufi< “ 094
monocovcoo
amummm cofiufimaoaom «own “ m<n
L
o>Hm> amalmouna u 0 fl ‘1
a
to E
a
«H Howcmnoxo boom ma m
M“, 09¢
w// _ .
/
I!/1.III:.I! umumaoomfi> 0939 N\ r ,
H a." // \ ‘ N «N fi
H m I!!! a m
{ J .x
m , w<a U
ammum

m>oo

58

and the tube viscometer which made it easy to switch from
vat to vat. These hoses, just like the leading pipes in the
whole system, had an inside diameter of 0.0381 m.

Tests were performed at two different temperatures :
73 °C and 80 °C. The product was heated in a double tube
heat exchanger with steam flowing countercurrently in the
outside annulus. The inside diameter of this inner pipe was
0.022 m. Two thermocouples were inserted in the steam
jacket to monitor (then manually control) the temperature.
The steam pressure controlled the temperature.

The tube viscometer was connected directly after 'the
heat exchanger. Two pressure transducers and two
thermocouples were installed. The tubes were insulated to
reduce the heat losses and to keep a constant temperature at
the wall. The distance between the two transducers was
4.594 m for the tubes with a diameter of 0.0127 m and
0.00635 m; the length was 5.193 m for the tube with a
diameter of 0.003175 m. The distance from the entrance of
the tube viscometer to the first pressure transducer was
1.33 m, 1.33 m and 0.73 m respectively; the distance from
the second pressure transducer to the end was always 0.73 m.

The volumetric flowrate was measured by collecting and
weighing samples after a fixed period of time. Due to the
failure of the available flowmeter, no continuous recording
was possible. Product was always collected for exactly one
minute.

Density measurements were taken at the time of the test by

59
weighing a sample and measuring its volume.

The four thermocouples and the two pressure transducers
were connected to a data acquisition system. A 21x
micrologger (Campbell Scientific Inc., Logan, Utah) was used
to collect the data every second and average them over 1
minute. It was carefully noted down at which minute the
flow rate was taken, to completely agree with the average
values, over the same minute, of the micrologger. An inside
clock reported the time and was set exactly at the same time
as a wrist watch. The execution time which was needed. to
average and store the data from the six transducers, and to
store the real time was, according to the manual, 0.5142 s.
This was perfectly in the safe range. Data were downloaded
on a tape recorder and from there to a floppy disk, using

the C20 Cassette Interface (Campbell).

All the thermocouples were checked in ice water and
boiling water and were within 0.5 0C of the expected values.
The pressure transducers, model 808 (Sensotec, Columbus,
Ohio) were a flush diaphragm type and were temperature
compensated up to 160 oC. The transducers were bonded
strain gage type; they were wired into four aCtive arms of a
Wheatstone Bridge and bonded to the sensing element of the
transducer. The excitation voltage was 5 VDC. The range
covered was from 0-1379 kPa.

The pressure transducers were calibrated with a dead

weight tester, type #1300 (Ashcroft, Manning Maxwell and

60

Moore, Bridgeport, Conn.). Known weights produced a known
pressure. Calibration was performed before and after the
experiments with differences less than 0.1% found. The
multipliers as an input in the data acquisition system were
calculated to be 67.235 and 67.032; the off-sets were
-8.6573 and 7.9845 for the serial numbers 161287 and 161290,

respectively.

4.3. Results

The rheograms obtained by the tube viscometers _are
given in Figures 4.2 (at 73 oC) and 4.3 (at 80 0C) for two
and three different diameters, respectively. The power law
coefficients, and their limits at 95%, are given in Table
4.1. A Herschel-Bulkley fit always gave a negative yield
stress, which was physically not acceptable. The average
temperature for the whole run is also indicated.~ The raw

data for each run are given in Appendix D.

4.4. Discussion

Using the Rabinowitch-Mooney equation (Equation 2.6),
required fulfillment of the assumptions made in section
2.3.2. The most important criterion was that of laminar
flow. The prediction of the transition point for non-
Newtonian fluids in pipes is well established (Hanks and
Ricks, 1974). Their equations were programmed by Garcia and

Steffe (1987).

61

0000

.Amuflu Ucm Moxuouv

00 ms um muouosoomm> was» ucmuomuwp 03D ecu Emumoonm .~.¢ muawwm

Am\ 5 BE «5%

 

 

 

oowm oow¢ oown oowm 030— o
h n b n n p O
-3
ob
. m
low no
as
. ll
ud
[0.2 E
ob
r \II
no
10m~ MW
E nolw mtfi ..u m .
E nolm and n m loom

 

 

 

62

ooom

.Amowu cam hoxuouv
uo om um muouoeoomw>.mn:u ucououwwo moucu no“ Emumoogm .m.o ouomflm

Am\ 5 ME»... mfizm

 

 

ooom ooots coon ooom 009 o
t _ . L . b b _ L _ . O
.
10¢
S
. H
3
V
10m 8
S
. I.
mm
Iom— 8
Co
I \l/
d
:2: m\
E nolm nnmné n... m o r
E nolm mph—...... n m I
E nolm and n m a 100m

 

 

 

63

Table 4.1. Viscometric constants for turkey and rice in
different tube viscometers.

Inside Average K n r
Diameter temperature K10w “high n10“ ”high
(m) (°C) (Pa 3") (-)

0 0127 72 7 0 86 1 06 1 26 0.567 0 595 0 624 0 996
78.9 1.53 2.13 2.74 0.448 0.491 0.534 0.992

0.00635 71.5 1.30 1.55 1.81 0.547 0.567 0.587 0.998
83.6 3.05 3.42 3.79 0.450 0.463 0.476 0.999

0.003175 78.6 1.02 3.30 10.71 0.310 0.454 0.599 0.831

The transition point (critical Reynolds number) and the
generalized Reynolds number were calculated for the worst
case (highest flowrate) for each tube viscometer at each
temperature. The rheological parameters from Table 4.1.
were used to perform the calculations. As seen in Table
4.2, all tests were performed in the laminar region, far

from the transition point.

The next assumption considered end effects, which can
cause turbulence. As a rule of thumb, end effects are
usually negligible if Li/D > 100 (Cho and Hartnett, 1982).
This tells that either a long length or a small diameter
causes less problems, because in those cases a higher total
pressure drop is obtained, making the end pressure error
relatively small. The smallest ratio in this study was 105.

End effects were therefore negligible.

Equation 2.6 assumed no slip at the wall. A slip

coefficient 8 can be calculated for each ow, using various

64

Table 4.2. Critical and generalized Reynolds number for the
different products.

R = 6.35 10'3 m n = 3.175 10‘3 m R = 1.5575 10‘3 m
T=73 °c 7:30 °c 7:73 °c T=80°C 7:13 °c

Generalized 275 223 588 575 122
Reynolds

Number

Critical 2340 2385 2354 2392 2394.
Reynoids

Number

tubes with different radii. Then, an additional term shows
up in the generalized form of the Rabinowitch-Mooney

equation, as follows

3
“1‘ 0" 2
I aw f(aw)dow + nRZUs ..........(4.1)
0

 

ow3

where Us = Bow : effective slip velocity (m/s)

f(ow) = iv : appropiate rheological model
Equation 4.1 can be rewritten as

l a

Q 2
[wow f(aw)dow .....(4.2)
0

nR3aw

 

8
= -—— +
R ow4

From equation 4.2 it can be observed that 8 is obtained as
the slope of a plot of Q/nR3aw versus l/R. However, a slip
coefficient was desired for each particular value of ow,

thus a curve for each ow was needed. To do this, first a

65

plot of Q/nR3aw versus 0w was made. Figures 4.4 and 4.5
illustrate this for the two different temperatures. The
curves obtained in those figures were determined by linear

regression. The regression equations were :

for 73 oC :
radius 5.35 E-03 m : Q/nn3aw = 0.0403 aw + 0.73 ; r2=.972
radius 3.18 E-03 m : Q/nn3aw = 0.0231 0w + 0.91 ; r2=.983
for 80 oC :
radius 5.35 E-03 m : o/nn3ow = 0.0534 aw - 0.21 ; r2-.963
radius 3.18 E-03 m : Q/na3ow = 0.0313 aw - 0.97 ; r2=.99l
radius 1.59 E-03 m : Q/nn3aw = °°°323_°w + 1.07 ; r2=.513

Disregarding the results for the smallest tube
viscometer, because of the bad correlation, one can see that
the slopes were different for the different radii.

Next, the plot of Q/nR3owr versus l/R should be
constructed for constants values of aw, using the above
regression lines. However, no effort has been made to do
this, because negative slip coefficients would have been
obtained, for the slope of the above regression lines was
higher for larger values of the radius (smaller 1/R).

The velocity profile for a power law fluid in a pipe is
presented in Figures 4.6,a and 4.6,b. In the first case
there is no slip at the wall. In the latter, a positive
slip is illustrated; the flow rate is higher and this was_
accounted for by an additional term in- Equation 4.1.

Considering this,‘ a negative slip coefficient would give

66

OFN

.Aumccflo wows
can moxDSDV Uo mu um acoquuuooo awam mo cowumaam>m .¢.¢ ouzmwm

Gav. mmmEm mfizm
cup on— mm on OF

 

 

 

 

1N no
/
n
. mm
D
A
T¢ \/
/
r d
nu
rm (SK
Enoum mtdnm _. .
881m and nm a

 

 

67

ohm

.Auoccmp mowu

0cm moxusuv uo on an acowowuuooo Qwam mo coflumoam>u .m.¢ musmfim

Gav. mmmmem mfizm

on?

on—

. F

cm on

_ . _

 

 

E nolm mhmné n
E nolm mnfin n
E nolm and .11.

0:11:01

G.

l
N

(8 Dd/ L) "bran/O

l
(D

 

 

 

68

<1
ll
O
<
‘iL
O

 

 

]; g
A r

 

 

 

 

 

 

 

 

 

 

 

 

i ‘\
i \
I
'r 1
I I
i ./
P ;7

(a) (b)

Figure 4.6. Velocity profile of a power law fluid -in a

pipe : without slip (a) - with slip (b).

a lower flowrate. Physically, this has no meaning. As a

consequence, the flowrate was not corrected for slip.

With the data of the flowrate and the pressure drop in
a tube viscometer, one has now sufficient information to
construct a rheogram, according to Equations 2.6 and 2.7.
The shear stress at the wall was easy to calculate. The
shear rate at the wall involved the determination of the
derivative d(Q/0R3)/d0w, which was accomplished by fitting a
polynomial. A software package, developed at MichiganState
University, calculated the Rabinowitch-Mooney equation and
fitted a Herschel-Bulkley or power law model through the
data. The Rabinowitch-Mooney equation and the regression
were both checked with other packages to verify the program
in the range of this study; practically no differences were
observed. The Michigan State University program was used to

solve Equations 2.6 and 2.7, but the same non-linear

69

regression model as in the previous chapter was used to fit
the model. Only for the smallest tube viscometer the non—
linear regression failed (bad correlation) and therefore a
linear regression was performed on the logarithmic
transformed data. This result was also close to the fit
given by the Michigan State University program (least square
curve fit).

The confidence limits for the results of the tube
viscometer are of the same order of magnitude as in the
previous chapter, except for the smallest tube viscometer;
this is probably due to the bad fit and the logarithmic

transformation.

Temperature control was very hard to achieve in the
system. The steam had to be continuously turned on and off,
based on the temperature of the product and the steam as
displayed on the micrologger. The product would eventually
heat up to 100 °C or more, if this had not been done.
Secondly, the pipe inside diameter from the heat exchanger
was rather small, so the product heated up too fast.
Finally, in order to obtain low shear rates at the wall, a
low flowrate was required, which increased the rate of
heating even more. Technical problems prevented heating of
the vat. This is a much larger volume to heat up and the
temperature control would have been improved. Most problems
were encountered with the smallest tube viscometer because
of the low flowrates; as seen in Figure 4.3, data do not

correlate as well as they do for the other sizes.

70

According to the temperature sensitivity of turkey and
rice dinner, an effort was made to stay in the neighborhood
of the target values of 73 and 80 °C. Higher temperatures
were too difficult to control. The average temperatures
over the whole run were 72.7 and 71.5 0C, and 78.9, 83.6 and
78.6 0C, which allowed comparison with the data obtained

from the concentric cylinder viscometer at 73 and 80 °C,

respectively.

Lack of a better control over the temperature prevented
experimentation at low flowrates. Consequently, shear rates
covered with the tube viscometer were higher than in the

previous chapter.

The results from the tube viscometer showed that the
consistency coefficient increased with temperature and that
the flow behavior index decreased. Both effects are very
unusual. The most probable reason was evaporation, which
caused the fluid to thicken and to change properties. AThe
time between the two experiments at 73 and 80 °C was around
30 minutes. During that time the product was constantly hot
and could evaporate (only half the tank was covered due to
returning hoses).

The data at the same temperature for the different tube‘
viscometers were reasonably close. How close the results
were, is outlined in the next section.

Comparing the data with the previous chapter revealed

that the fluid had a lower consistency coefficient and was

71

less pseudoplastic at 73°C in the tube viscometer. At 80 0C
the same flow behavior index was obtained, but again a lower
c0nsistency was observed. It was not believed that the
different shear rate range played an important role in the
discrepancy of the results. Comparing data that did fall in
the same range (40-400 s‘l) were not compatible. Figure 4.7
illustrates the rheograms obtained with the concentric
cylinder viscometer and the largest tube viscometer at two

different temperatures.

4.5. Accuracy of the results - Sensitivity analysis

The rheological parameters were. obtained with the
Rabinowitch-Mooney equation. This involved a polynomial fit
and a curve fit to obtain a rheogram. To assess the
accuracy on the rheological parameters, the program
developed at Michigan State University was perturbed by
giving small variances to the variables and observing‘ the
changes in the rheological parameters.

The accuracy can be determined by combining the errors

as follows in the root-sum square formula (Scarborough,

1955)

on the flow behavior index :

 

 

 

 

 

J/fan an an an 1
An = ( A(AP))2 + ( 1012 + ( AL)2 + ( A012
3(AP) ' 80 BL 8Q

72

come

.uo om 0:0 00 me am
swuweoomw> was» cam souoeoomfi> umocmaao owuucoocou

m nuw3 oocwmuno umccwo ooflu 6cm woxuou mo manhooonm .h.¢ ouaowm

Am\ 5 BE mfizm

 

 

OOOF 00m 000 00¢ CON 0
u _ . _ . [F1 . h b _
00 HR III
00 cm .. .. ..

  

Ezoom5 mma.

EZOOMS
mmozfiio QEZmozoo

 

 

(Dd) 583818 8173115

73

on the consistency coefficient :

 

/// 8K 8K 8K 8K fl
1x = ( 1119112 + ( 1012 + (— AL)2 + <——A0)2
8(AP) an 8L 8Q

 

 

The partial derivatives were obtained by perturbing the
program. . Their magnitude is not so important, however,
multiplied by the accuraCy of the variable, it reflects the

importance of the contribution to the overall error.

First, the accuracy on each variable was determined.
The accuracy on the pressure drop was obtained as follows.
Since AP is a difference of two pressures, the accuracy on
each pressure transducer must be multiplied by /5: Using
the calibration accuracy, the error on AP was calculated to
be 1 300 Pa. The accuracy on the diameter of the pipe was
based on the accuracy of the measurements at both ends and
its uniformity, which was estimated; AD was then set to be 1
0.0004 m. The length between the two pressure transducers
was measured three times and the difference between the
average and the highest measurement was thought to be the
accuracy; AL was then found to be 1 0.002 m for all three
tube viscometers. The accuracy on the flowrate consisted of
a chain of equations, and was based on density measurements
— which were volume and weight measurements - time
measurements and weight measurements. The worst case was
calculated and found to be 1 9.36 10'6 m3/s, which was

approximately 5%.

74

Next, the program was perturbed with the data from the
largest tube viscometer at 73 oC. The partial derivatives
are given in Table 4.3. After multiplying and squaring, the

following numbers were obtained

 

An /'1.41 10‘5 + 0 + 0 + 40.5 10'4 1s 1 0.064

 

AK /F5.54 10‘4 + 9.92 10‘3 + 1.9 10'7 + 147 10'37

1 0.4 Pa sn

The derivative of n with respect to L and D was zero.
In other words, changing the geometry did not influence the
value of n. This means that the velocity profile stays
identical, regardless of the geometry.

Since the above ‘calculations were obtained by
perturbing, it was only valid in the neighborhood of the
chosen values. Identical calculations for the middle size
tube viscometer yielded a value of An of 1 0.056 and for
AK 8 value of 1 0.62 Pa s“.

As can be seen from the above equations, the flowrate
contributed most in the error and was thus the most
sensitive parameter in this study. This was expected,

because its accuracy was lowest.

Considering these accuracies, the results for identical
temperatures, presented in Table 4.1, come closer together
and are more acceptable. However, the increase of K and

decrease of n with temperature can still not be explained.

75

«no.0 mmao.o dwm.o mmm.o oooav oomm 0

 

 

 

 

 

 

oso.o o a-o~ a o -.o o d
~mo.o o ooH.o o mam o o
mmo.o -H0.o mmo.o muss ms.m m-o~ m.m m-o~ m~.H 54
9m 9m em as an an
x< x< x4 x< manmwum>
sm an em as em as
wH um hOth hOhhw HMDHU<

. ms um youweoomw> was» ummmuma onu so“ moanmwum> on» on
uuommmu :ufla & cam c mo mo>fium>wumo Hmwuuma 0cm mEuou Ecuum .m.¢ manna

76

The discrepancies of the tube viscometer results with
the results from the concentric cylinder viscometer are
still too high. A similar error analysis on Equations 2.5
and 2.6 revealed an error on K of 1 0.25 Pa 5n and 1 0.0005
on n. This higher accuracy could be expected, because they
were carried out on a laboratory scale. As pointed out
earlier, the confidence limits were similar.

A possible reason for the discrepancy in Figure 4.7 is
that there was a two week delay between the measurements
with the concentric cylinder viscometer and the tube
viscometer. Enzymatic, chemical or microbial activity c0u1d
have influenced the properties, however, no visible changes
were detected during this period. Another reason could have
been the extended heat treatment or the fact that the
product was more broken down in the tube viscometer.
However, it was more than likely that the rheological
parameters were instrument-dependent. Work from other
researchers (Tables 2.1 and 2.2) demonstrated this already
for applesauce. Different instruments shear the product
differently. Saravacos (1968) pointed out that the shear
rates in rotational viscometers can cause separation of the
suspended solids in a puree, resulting in an erroneous
measured consistency (usually lower). Charm (1963) believed
that aggregates are broken and that fluid is released in the
area of the rotating cylinder surface. Less torque is
exerted on the cylinder wall, resulting in a lower

consistency. In a .tube viscometer, the whole fluid is

77

subjected to high shear rates and the measured value is more
representative of the true consistency. On the other hand,
particle migration can take place in a tube viscometer, also
predicting a lower consistency. In any case, rheological
parameters obtained by a concentric cylinder viscometer,
could not, within the limits of this study, be used to

predict the pressure drop in a tube viscometer.

Finally, a sensitivity analysis can be performed by
attributing the same accuracy (1%) to all the variables in
Equation 4.3. A similar analysis has been carried out by
Mohamed and Steffe (1985) to predict the flowrate for a
Herschel-Bulkley fluid.

Calculations (Table 4.3) showed that the pressure drop and
the flowrate were equally sensitive with respect to n. As
mentioned earlier, the geometry had no effect on n. For the
consistency coefficients, it was shown that the pressure
drop was the most sensitive parameter, closely followed by
the flowrate. Errors in the diameter and the length

contributed much less to the total error.

In summary, one can state that the determination of the
rheological parameters in a tube viscometer is equally
sensitive to errors in pressure drop and flow rate. The
accuracy on the geometry can be considered negligible. If
one wants to establish the target values for the rheological
parameters with a tube viscometer, this should be kept in

mind.

78

4.6. Conclusions

In this chapter the rheological properties of turkey
and rice were measured by means of a tube viscometer. Two
important conclusions can be drawn.

The results from the tube viscometer seemed not to
agree with what was obtained with the concentric cylinder
viscometer. The flow behavior indices were in one case
higher than, and in the other case, equal to the earlier
observed values. The product) seemed to have a alower
consistency in the tube viscometer. Calculations of the
accuracies on the parameters did not bring the results from
the two different instruments together. Consequently,
target values should also be investigated with a tube
viscometer, despite a higher obtainable accuracy with a
concentric cylinder viscometer.

Secondly, a sensitivity analysis showed that for the
calculations of the rheological parameters, the flowrate and
the pressure drop were equally sensitive variables, while
the geometry was negligible. This is exactly what is
preferred, namely the pressure drop will now be the new
target to predict a change in the rheological parameters.
The next chapter will investigate the sensitivity of the

pressure dr0p, as a criterion for quality control.

CHAPTER 5

SENSITIVITY ANALYSIS AND DESIGN CRITERIA

5.1. Introduction

In this chapter the design criteria for an on-line
viscometer are determined. This will mainly be based on the
experiments of the two previous chapters. In addition, to
this, a sensitivity analysis will be carried out to
investigate the sensitivity of the pressure. drop, as a
variable to assess changes in the rheological parameters.
This analysis is different from the one in section 4.5,
because now, the variables which affect the pressure drop
will be of interest. It is not the intention to estimate
the rheological parameters, for this is not possible with

the measurements of one flowrate and one pressure drop.

5.2. Sensitivity analysis

The Rabinowitch-Mooney equation can be transformed to a
much simpler equation, which involves no polynomial fitting,
since the model of the fluid materials was shown to be a

power law. Rearranging Equation 2.6 for a power law yields

79

80

.............(5.1)

II
A
v

:1
A

 

AP

 

The accuracy on AP can then be expressed as

 

 

 

 

///3(AP) 8(AP) 3119)
A(AP) = ( AL)2 + ( Ax)2 +( AR)2 +
an ax an

 

fi

3(AP) 8(AP)
A0)2 + (
80 an

( An)2 ........(5.2)

 

 

Since just one equation is used, the derivatives can be
calculated mathematically, without the need for perturbing.

The partial derivatives are given below

 

 

 

 

 

 

 

 

 

 

 

 

 

8(AP) 2K Q 3n+1
= ( 3 )n ( )n
3L R “R n
8(AP) 2L Q 3n+1
= ———- < 3 )n ( 1"
3K R 0R n
8(AP) Q 3n+l —l-3n
= 2LK ( 3 )n ( )n (-——-§——)
3R 0R n R
8(AP) 2Lx 1 3n+l ’
= ( )n ( )n n Q“-1

30 R 0R3 n

81

8(AP) 2LK Q 3n+1 Q(3n+1) 1
.____. )“( )"[ln( 3 ) - (——————)1
an R 0R3 n nnR 3n+1

 

 

 

II
A

The derivatives were checked with a numerical solution
and were found to be in good agreement. The above
derivatives were then incorporated in Equation 5.2.

To investigate the sensitivity of the pressure drop as
a function of the variables L, K, R, Q and n, the following

representative values were used :

L = 3 m

K = 5.54 Pa 5n
n = 0.48

R = 0.0127 m

Q

2.53 10‘4 m3/s

The pressure drop is then 33 kPa.
The precision level was chosen.to be 1% to 'represent

reasonable numbers. Equation 5.2 then becomes

 

A(A)P = /6110842 + 110842 + 559913 + 25539 + 510294‘
L [A~ 1“ u .1 0 1
It can be seen that the flowrate term (25539)
represents the least sensitive variable, while the radius
and the flow behavior index are the most sensitive
variables. An error in the consistency coefficient and the
length is less important.

The fact that the flowrate is not sensitive turns out

82

to be very practical. A less accurate flowrate, produced by
the pump, will not influence the pressure drop too much.
This allows more wear on the pump. The lower sensitivity of
K, on the other hand, is not so favorable.

If a fixed system is used, there will be no changes in
R and L; then, the pressure drop is only a function of n, K
and Q. A change in n will be easier to detect.
Fortunately, it has been shown that n is not very
temperature sensitive; however, further research must be
carried out to find out how n changes with the consistency
of the product. If this is not significant, then 'the
pressure drop reflects only the changes in K.

Changing the neighborhood of the calculations to n =
0.2 (e.g. carrot puree), yields the following numbers for

Equation 5.2

 

1(1)? = /E5838 + 5838 + 17504 + 273 + 71891
L V\ ('3'
This gives essentially the same results, except for the
sensitivity of n, which has become approximately equal to

the sensitivity of K.

5.3. Design criteria for an on-line viscometer

Figure 1.1 gave a schematic view of a process line and
indicated the position of the on-line viscometer. The

inside diameter of the main pipe is 0.0635 m and the

83

flowrate ranges from 3.7 10'3 to 4.3 10'3 m3/s. The purpose
is to connect a smaller pipe to the main line, having its
own pump, to produce a more stable flowrate, one temperature
sensor and one or two pressure transducers. To reduce the
costs, the product must not necessarily be returned to the
main line, but it can be pumped directly into a surge tank.
This would require only one pressure measurement, because
the other is at atmospheric pressure. However, this is not
advised, because of uncontrollable end effects.

The smallest commercially available diameter is 0.0254
m. Keeping the rule of thumb in mind to reduce the' end
effects (Li/D > 100), an entranCe length of 2.5 m is
advised.

It is best to keep the shear rate in the same range as
in the main pipeline to maintain the same texture. The
shear rate at the wall can be calculated with the
Rabinowitch-Mooney equation, valid for a power law, as

follows

 

YW = .........................(5.3)

The flow'behavior index is assumed to be 0.5. Accordingly,
the shear rate range in the main pipe becomes 184-215 5.1.
The flowrate in the on—line viscometer can be back
calculated 'using Equation 5.3. This yields a flowrate in

the range of 2.36 10"4 to 2.76 10'4 m3/s. Variation of n

84

will not affect the shear rate range significantly.

The critical Reynolds number is in both cases 2382.
The generalized Reynolds number is 27 for the tube
viscometer and 174 for the main line. The flow is well
within the laminar region.

The assumption of no slip at the wall is acceptable,
because the error on the flowrate can be much greater than
the error in K, before the same effect is observed on the
change in pressure drop.

The pump that suits the requirements is a positive
displacement, variable speed pump, size 15 (Waukesha). . The
displacement of this pump is 4.92 10"5 m3/revolution
(Waukesha, 1980). To displace the volume calculated above,
the pump will have to turn at 287—336 rpm. A reducer from
the port size 0.0381 m to 0.0254 m will be required.

The pressure at the discharge should be calculated to
be certain that the upper pressure limitations (1379 kPa)
are not exceeded. The pressure was calculated by means of a
mechanical energy balance and found to be 37 kPa, which is
well within the pump design limitations.

The pressure at the inlet side should be sufficient to
fill pump cavity. For the above pump this net inlet
pressure required is only 20.7 kPa, at a speed of 300 rpm.
No sufficient information was available to calculate the net
inlet pressure available, but it is believed that this
pressure will be easily attainable.

Pulsations will be very small due to the, small

85

discharge volume. Measuring over a long enough period of
time and frequently enough will average this out.

The distance between the two pressure transducers can
be chosen freely, since the pressure drop varies linearly
with distance. A longer length will increase the magnitude
of the pressure drop. A length of 3 m is advised, to have a
reasonable pressure differential.

Equation 5.2 dictates that the accuracy of the pressure
transducers will be a function of the accuracy on the
flowrate measurements and the determination of the
rheological properties. In the worst case, the flow ‘rate
can be measured by recording the speed of the pump, assuming
slip of the pump is negligible. This is true for viscous
materials ( > 0.1 Pa s") (Waukesha, Abex Corp., 1980); baby
food can be considered viscous enough.

The final design criteria for the proposed on-line

viscometer are presented in Figure 5.1.

5.4. Conclusions

Some preliminary design criteria have been developed,
according to a sensitivity analysis of Equation 5.1. It was
found that, in a fixed system, the flow behavior index was
the most sensitive parameter. The consistency was,
depending on the values of the other parameters, less
sensitive. Both characterize the fluid and can therefore be

used as quality control parameters. The flowrate was much

86

 

.uouosoomw> onwauco oomomoEQ on» LOu oflumuwuo commoo Hmcwm .H.m ousmflm

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

moH>oa
wafivuoomm
\
\ /
\ \ ///
\.
\\ \ a o.muq fill. a n.~uHA
Smcoauoov \ \ /
\ \ // 2 13m
.5008 ill,
a .IIJ.
soak mm nouoaoomu> oboe e.gm a.-~o.olm a 1
3
m a 0 n
\m «log n N o m
Emu oom
wcfia cfimz a Hmoo.onm

87

less sensitive than the rheological parameters, which was a
positive point, considering possible wear of the pump with
time.

The final design parameters for a start up on-line
viscometer for baby food puree were found to be : L = 3 m, Q
= 2.5 10"4 m3/s, D = 0.0254 m, a variable speed pump, size
15, turning at a speed of 300 rpm, two pressure transducers,

one temperature sensor and a flowmeter, the latter being

optional.

CHAPTER 6

CONCLUSIONS

Rheological parameters for four baby food products were
obtained by using a Bostwick Consistometer and a concentric
cylinder. Unfortunately, the results could not be related
to each other. Rao and Bourne (1977) were able to correlate
their Bostwick Consistometer readings with a plastometer.
In this study, an analogous relationship could not be
reproduced.

This important fact showed that the Bostwick readings
could not be translated into rheological parameters. As a
consequence, target values for a desired quality are
recommended to be determined in terms of the rheological
parameters, for each specific product.

Rheological experiments with a tube viscometer,
simulating an on-line viscometer, revealed that the
rheological properties tare significantly different from
those obtained with a concentric cylinder viscometer. This
means that the determination of the target values are
recommended to be carried out in a tube viscometer.

Finally, design criteria were set up for a start up on-
line viscometer for baby food puree that would best simulate

the actual on-line viscometer. This was based on a

88

sensitivity analysis and on the measured rheological data.
The entrance length was 2.5 m, the length between the two
pressure transducers was 3 m, the flowrate was 2.5 10"4 m3/s
and the (diameter was 0.0254 m. Other dimensions can be
obtained, if a similar procedure, as outlined in chapter 5,

is followed.

89

CHAPTER 7

SUGGESTIONS FOR FURTHER STUDY

1. To build the proposed on-line viscometer.

2. To investigate the effect of total solids on the
Bostwick Consistometer readings and tube viscometer data.

3. To determine the target values of the rheological
parameters for a desired consistency, with the above ondline

viscometer.

4. To determine the range of the pressure drop as a

quality control point.

90

APPENDI CES

APPENDIX A

BOSTWICK CONSISTOMETER READINGS

FOR DIFFERENT PRODUCTS

91

mN... om... mn.o. om.m 0.00 c

om.__ 0N... mh.0. 0m.0 0.00 0

oo.__ 0N... 00.0. 00.0 0.—0 N move 0:0
oo.N_ mn.p— mN.__ 00.0. 0.00 _ Nexus»
mN.m_ 00.5— 0N.m— om... 0.00 v

mN.v_ 00.5. oo.n_ mN._. o._m m

oo.m. mN.¢. oo.m_ mN.__ o..m N +500 0:5
ma.vp om.m_ mN.N. mN... m.mN F o_omuumm>
mN.N m~.0_100.0 00.N 00.0.100.0 00.N mn.m 100.5 00.N 0.00 v

oo.N oo.o_1mN.m oo.N mN.m nom.N oo.N 00.0 uoo.N oo.N m.oN m

om.0 0m.m nmN.m 00.0 00.0 smN.N 00.0 00.0 100.0 0m.0 0.00 N 00030
oo.N om.o,-om.m oo.N mN.m 1mN.N oo.N om.m 100.5 00.5 0.00 . 551150
05.3. mN.o_ 00.0 00.0 0.00 v

so... om.o. mN.o 00.0 m.mN m

ma.o_ mN.o_ mN.m o_.m 0.00 N

mN.o_ mN.o_ 00.0 oo.m 0.00 F 003810.00<
00... 0N... mh.o_ mN.0 0.¢N 0

mn... mN.__ mN.o. mN.m 0.5N N coat 0:5
cm... mN... mN.o_ mN.m o.vN _ >oxuah
om.¢. 00.5. oo.m_ mN._. m.mN v

m~.¢_ 00.¢_ om.N. m~.0_ 0.0N m

mN.mF mN.v. mN.m_ om... 0.0N N 4000 0:0
mN.v_ mN.N_ mN.N_ oo... o.vN 1 0.009000)
mN.o oo.o.-oo.m mN.o om.m 10m.» om.o om.m ioo.N om.o o.NN e

mN.o om.o.uom.m mN.o oo.o.-om.N om.o 00.0 100.5 om.o o.NN n

ma.@ m5.o ms.o ma.o o.na N 55130
00.5 oo.o.1oo.m co.» om.m nom.N oo.N oo.m 100.5 00.5 o.NN . gottmo
mn.0 05.0 00.0 mN.0 0.0m v

00.0 0¢.m 00.0 mN.0 0.0» m

00.0. 00.0 00.0 00.0 0.05 N

mN.o. mn.0 on.m om.m 0.NN . ouammo_oo<
Amoo\550 0,35,. “mom\500 01:0,. Rum—\Euo 51:5,. Aum\500 Aooo

now am mom mm mm. mm mm mm oLJuMLoQEob o_QEmm pounced

.mouopoua bambooufiv LON mmcmvmou poumEoDmflmcou xufisumom .H.¢ oHan

92

00.N. 0N.N. 05... 05.0. 0.N0 0

00.0. 00.0. 05.N. 00... 0.N0 0

0N.N. 00.N. 00... 0N.0. 0.N0 N 00.1 0:0
00.0. 0N.0. 05.N. 0N... 0.N0 . 501135
05.0. 0N.0. 00.0. 05.N. 0.N0 0

00.0. 00.0. 00.0. 05... 0.N0 0

0N.0. 05.0. 00.0. 00.N. 0.N0 N 1000 0cm
05.0. 0N.0. 00.0. 05... 0.N0 . 0.00.000>
05.5 00.0.105 0 05.5 05.0 10N.0 05.5 0N.0 105.5 05.5 0.N0 0

00.5 00.0.105.0 00.5 00.0.100.0 00.5 05.0 105.5 00.5 0.N0 0

05.5 00.0.100.0 05.5 00.0 100.0 05.5 00.0 100.0 05.5 0.N0 N 00130
0N.0 00 0.10.0. 0N.0 00.0 100.5 00.0 00.0 100.0 00.0 0.N0 . 301100
0N... 0N... 00... 00.0. 0.N0 0

00.N. 00... 0N... 05.0. 0.N0 0

00 .. 05.0. 00.0. 00.0 0.N0 N

00.0. 05.0. 00.0. 00.0. 0.N0 . 003000.000
0N.N. 00.N. 05... 00.0. 0.00 0

00.N. 05... 00... 0N.0. 0.50 0

00.N. 05.N. 0N.N. 00... 0.00 N 00.1 0:0
00.N. 05.N. 0N.N. 00... 0.00 . 500135
05.0. 0N.0. 00.0. 00.N. 0.00 v

05.0. 0N.0. 00.0. 00.N. 0.00 0

05.0. 0N.0. 00.0. 00.N. 0.00 N 1000 0:0
05.0. 0N.0. 00.0. 00 N. 0.00 . 0.00.000>
00.5 05...105.0 00.5 05.0.105.0 00.5 00.5 0.00 v

00.5 00.0.100.0 00.5 00.0.100.0 00.5 00.5 0.00 0

0N.5 0N.5 0N.5 0N.5 0.00 N 00130
00.5 00.5 00.5 00.5 0.00 . 101100
00... 05.0. 00.0. 00.0. 0.00 v

0N... 00... 05.0. 00.0. 0.00 0

05.0. 00.0. 00.0. 00.0 0.00 N

00... 05.0. 0N.0. 00.0 0.00 . 003000.000
.000\Eu. 0.33.. .000\Eu. 0.30.. .00.\500 0.30.. .00\Euo .00.

wow mm mOm mm mmw— mm mm mm OLJuMLQQEOP m—QEMW HUDUOLQ

..0.Dcouv .H.< magma

APPENDIX B

RHEOLOGICAL CONSTANTS

OBTAINED WITH THE CONCENTRIC CYLINDER VISCOMETER

FOR DIFFERENT PRODUCTS FOR THE SEPARATE RUNS

93

Table 8.1. Rheological constants obtained with the
concentric cylinder viscometer for different
products for the separate runs.

Product Sample Temperature. K n r2
°C) (Pa 5") (-)
Applesauce l 72 2.81 0.428 0.997
2 73 2.54 0.451 0.972
3 72 2 43 0.447 0.995
1 80 3.37 0.335 0.999
2 80 4.23 0.330 0.999
3 80 2.23 0.400 0.999
4 80 2.99 0.364 1.000
1 87 2.20 0.333 0.999
2 86 2.11 0.329 0.999
3 86 2.36 0.327 1.000
4 86 2.38 0.324 1.000
1 92 1.65 0.391 1.000
2 92 2.64 0.346 1.000
3 92 1.99 0.357 1.000
4 92 2.24 0.342 0.999
Carrot 1 73 2.06 0.346 0.985
Puree 2 73 6.46 0.244 1.000
3 73 3.92 0.280 1.000
4 73 7.33 0.231 1.000
1 81 3.82 0.254 0.999
2 80 4.63 0.245 0.999
3 80 4.00 0.312 0.998
4 80 10.20 0.148 1.000

“NH DOOM!“
m
01
p
O O
O
O
O
O
u
N
(A
O
O
\D
\D
u

94

Table 3.1. (cont'd)

Product Sample Temperature K n r2
(°C) (Pa 5") (-)
Vegetable 1 73 1 24 0.643 0.999
and beef 2 72 3.28 0.496 1.000
3 72 4.00 0.461 0.999
4 72 1.07 0.672 0.995
1 80 2.52 0.490 0.998
2 79 2.63 0.500 1.000
3 80 1.88 0.495 0.999
4 80 2.47 0.475 0.994
1 85 2.12 0.512 1.000
2 85 2.15 0.504 0.998
3 85 2.19 0.511 1.000
1 92 0.87 0.686 0 989
2 92 0.53 0.469 0.992
3 92 0.23- 0.576 0 989
4 92 0.54 0.515 0.999
Turkey 1 74 6.80 0.461 1.000
and rice 2 74 6.29 0.466 1.000
3 74 6.47 0.454 1.000
4 72 3.84 0.527 0.995
1 80 '3.72 0.459 1.000
2 81 4.75 0.482 1 000
3 80 4.48 0.451 1.000
4 80 3.50 0.472 1.000
1 86 1.73 0.470 0.995
2 86 2.13 0.460 0.998
3 87 1.89 0.482 0.997
4 86 1.92 0.432 0.996

11>me

KO
...:
O
O

.p
...:
O
.9
(II
on
O
O

\O
\O
m

92 0.75 0.450 0.999

APPENDIX C

THE APPARENT VISCOSITY AS A FUNCTION

OF THE SHEAR RATE AT DIFFERENT TEMPERATURES

FOR THE DIFFERENT PRODUCTS (RAW DATA)

95

com

..0000 300. wusmmmaqam 000 mouzumuwosmu accumuuwn

00 0000 00000 0:» 0o 00.00030 0 00 auflmoumw> acoumddm 0:9 .H.U 0030.0

Afisflé 00000
2.00 . own 2.0“

LI

 

 

b h
9 ‘

on no 3:03.39

 

 

.206
w
166

1nd

 

o.—

 

con

A0}. 0500 00000
2.00 can 0mm

n P

 

 

 

00 on 00303330

 

 

o.—

(s Dd) msoosm lNI-IHVddV

(s Dd) MISODSIA lNBdeV

01>. 0.0.. 0.000

 

 

 

 

 

 

 

com 00* 00». CON 8— o
u h L b h L- P 0
1!» o o
rNd
n
1.0.0
r
T
10.0
u. 00 838.08 010.0
f
o.—
..1.\ c 0.0.. 00000
com 000 can CON oo— o
p n h p — r h 0.0
T
[11111111514114.0177 1nd
. fl
.
10.0
1nd
.
00 mm 33002000 Ind
1

 

 

 

(s Dd) MISOOSIA lNEdeV

(s 00) wsoosm maavadv

96

..0000 300. 000:0.000000 000 000000003300 000000000
00 0000 000:0 0:0 00 00000090 0 m0 aummoumM> 0:000Q00 0:9 .~.U 009000

Am\ 5 0000 000:0

 

 

 

 

 

 

 

000 co.
__0 1..
u
_
.0
r
00 N0 0.130 .8100 rod
['0' 1| .1 ' I|I1 1QI1| IIIIIvIII' [0.—
Am\ C 0500 00010
000 00.. 000 00N 00. 0
. k . F L . . 0.0

00 on 00030 .0008

 

(Ill 1"!1

 

 

 

(S 00) ulsoosm maavadv

(s 00) MISOOSIA lNBBVddV

A0\ .0 000... 00000

 

 

 

 

 

 

 

 

 

 

oom 00¢ con CON oo— o
b b b h bf h % Ll o o
v
.d
.d
v
ud
3
N
I;
M
S
3
O
B
M
AW
00 00 8130 0818 . 10.0 0
fl S
(
T
I 1 O.—
3 5 0000 0010
Con 00¢ 00” CON OO— O
b 0 1p, h I n b P F 0.0
V
d
M
3
N
l
M
S
3
O
K.
. u
.. \I
V o d
00 2 0.030 .250 / .10 O O
0 S
T (
f
o.—

97

.00000 3000 000000 000: 0:0 00000000> 000 000000000300 000000000
00 0000 000:0 0:0 00 00000000 0 m0 >0mmoumw> 00000000 0:9 .m.u 000000

7.\ 5 use mfixm

 

 

 

 

 

 

 

com oov 8n SN 8— O
. r r L r 0 r 0 w 0.0
r
r¢.o
0.
Fa
w
r
on «a .93. 033.02. .190
fi
_ r
_ 3
Am\ 5 02m «firm
con 00* com CON OO— o
b n p n L h n L! r 0.0
00 on .oonloZoaoog 10.0
I o.—

 

 

 

(s 0d) wsoosm mauvadv

(s 0d) wsoasm iNBHVddV

Am\ 5 E5. «firm

 

 

 

 

 

 

com 00+ con 00N 00— 0
P — b h h h n p p 0.0
1
0.0.0
10.0
0
0
10.0
. w
T
r
00 00 $310383! 10.0
[I 0.—
?\ 5 02¢ «firm
can 00¢ con 000 00. 0
r h L n p r if r p .
0 0
r
_ 0o 2 .80. .3832, 10.0

 

 

 

(s 0d) wsoosm iNBHVddV

(s 0d) wsoosm lNBBVddV

98

.00000 3000 000000 0000 0:0 000000 000 000000000500 000000000
00 0000 000:0 0:0 00 00000000 0 00 >000ou0w> 00000000 0:9 .¢.U 000000

 

 

 

 

 

 

 

 

 

 

 

 

A0\3 MED. «firm Am\$ M58 «610
com 00? con CON 2: o
_ 0 0 L 0 . 0 . 0 . 0.0 can . on»: Don . Omw L om. . o 0.0
_ w a
M d d
V
a w
3 3
. m m
ltd
T w m
. S
. «00 m
10.0 m B
. A M
) )
I d d
0.. «a 801033 .10 0 0 0o 8 3:10.03 . 10.0 o
,0 . ,0
f
[iiilllll 1.11.! o: :1 o._
E 0 05. $010 3 0 03.. «firm
con 0 owe T own . owN » 0m: F o . can 00* con CON oo— o
o o k — b h b — + h h . .
fi 0 o
w W 1 V
1~.0 d . . w
. V 10o v
. m 0 a
. N . m...
l I
10.0 . l
A 11¢ 0 A
1 g 1 I
0 o . s
o . m
de pl.) 10.0 S
w l H.
. 1 H 1
. . \Alu . hm
no 00 ootl§3 HQ 0 0 Do an 080108.030 a r06 0
.6 , S
O (
r o.— o.—

 

 

 

 

 

 

APPENDIX D

PRESSURE DROP , FLOWRATE AND TEMPERATURE DATA

FOR THE

DIFFERENT TUBE VISCOMETERS

99

Table D.1. Pressure drop, flowrate and temperature data for
the tube viscometer with a diameter of 0.0127 m
at 73 °C and 80 °C.

T T T3 T4 AP

(0c) (°C) (°C) (0c) (Pa) (m3/s)*10'4
73.5 74.7 99.4 99.4 97284.2 1.8036
73.2 74.5 99.4 99.4 95284.8 1.8352
72.8 74.0 99.0 98.9 96112.1 1.8688
73.0 74.1 99.1 99.0 99559.5 1.8896
73.5 77.0 92.0 96.1 81357.5 1.3121
70.1 72.4 91.2 101.9 81978.0 1.2311
73.1 77.0 106.4 106.3 76531.2 1.3221
69.5 72.6 109.6 109.6 59570.2 0.8492
72.6 78.0 99.0 100.7 57846.5 0.8183
69.9 74.0 110.4 110.3 60259.7 0.7847
73.2 74.6 100.9 103.9 35163.0 0.3303
74.8 75.0 104.3 104.3 29647.2 0.2809
75.8 75.9 103.7 103.8 25510.4 0.1627
80.9 82.7 105.1 106.1 84391.1 1.5063
81.0 84.9 110.7 110.6 82391.7 1.4346
80.8 85.8 111.6 111.5 79495.9 1.3508
77.5 81.0 110.3 110.2 96112.1 1.6589
77.4 80.4 105.1 105.0 81426.4 1.1630
77.4 81.2 109.7 109.6 78599.6 1.1243
78.3 83.6 117.8 117.7 78185.9 1.1437
74.9 78.3 109.7 110.4 50745.0 0.4758
80.5 84.5 113.7 114.5 77841.2 1.0125
77.3 79.9 114.8 114.7 44057.1 0.3483

81.9 81.2 124.1 124.0 20959.9 0.0609

100

Table D.2. Pressure drop, flowrate and temperature data for
the tube viscometer with a diameter of 0.00635 m
at 73 °c and 80 °C.

T T T T4 AP

(0%) (06) (00) (°C) (Pa) (m3/s)*10"4
68.6 69.5 102.5 102.5 604251.5 1.1981
69.5 70.8 103.1 103.1 602665.7 1.2017
70.4 72.0 103.8 103.8 593633.7 1.2110
70.8 73.2 98.3 98.3 542543.9 0.9652
69.7 72.3 98.3 98.2 545370.8 0.9645
68.9 72.4 98.5 98.3 531650.3 0.9172
71.0 74.9 104.1 104.0 393687.4 0.5668
72.8 76.8 105.9 105.8 397617.3 0.6019
73.9 76.1 106.6 106.6 389826.3 0.5847
75.6 76.4 95.7 98.3 299092.1 0.3447
73.8 74.6 88.7 98.2 287302.1 0.3096.
72.8 73.3 83.9 98.2 297161.6 0.3124
83.2 83.4 120.7 120.6 543164.5 1.1272
83.4 84.1 120.9 120.8 545508.7 1.1394
83.7 84.8 121.2 121.1 545439.7 1.1387
81.0 84.7 119.2 119.1 481112.2 0.8183
80.8 84.6 118.9 118.9 490764.7 0.8535
80.6 84.8 118.9 118.8 497866.3 0.8606
84.0 85.8 123.6 123.5 373141.2 0.5396
85.5 86.8 124.2 124.1 387551.1 0.5482
85.7 86.8 124.0 123.9 397203.7 0.5496
85.9 80.5 109.7 109.6 215942.0 0.1727
84.7 81 4 109.6 109.5 190431.6 0.1254

84.8 83:8 109.5 109.4 191190.0 0.1376

101

Table D.3. Pressure drop, flowrate and temperature data for
the tube viscometer with a diameter of 0.003175 m

at 73 OC and 80 °C.

T T2 T3 T4 AP
<°c> (°C) (°c) (°C) (Pa) (m3/s)*1o-6
78.3 76.0 89.5 98.3 893718.6 9.2439
77.0 74.6 84.3 98.2 897965.7 8.5273
75.2 72.8 79.0 98.2 900358.2 8.3840
74.2 70.3 113.4 113.3 862575.2 8.4557
81.7 78.7 105.4 105.3 858955.5 10.8921
82.4 79.7 101.0 100.9 863457.8 9.8172
80.9 78.3 103.3 103.2 986762.6 11.9669
81.0 78.7 98.5 98.4 963389.5 10.6054
79.0 76.6 98.2 98.1 958942.5 10.3905
77.1 74.4 99.3 99.3 789808.6 6.8792
77.8 75.0 100.1 100.1 806211.1 7.7391

78.6 75.9 100.8 100.7 797013.5 7.7391

BIBLIOGRAPHY

B I BL I OGRAPHY

Barbosa, C.G.V. and M. Peleg. 1983. Flow parameters of
selected commercial semi-liquid food products. J. of
Text. Stud. 14 : 213-234.

Bookwalter, G.N., A.J. Peplinski and V.F. Pfeifer. 1968.
Using a Bostwick Consistometer to measure consistencies
of processed corn meals and their CSM blends. Cereal
Science Today 13 (11) : 407-410.

Charm, S.E. and E.W. Merrill. 1959. Heat transfer
coefficients in straight tubes for pseudoplastic food
materials in stream line flow. Food Res. 24 : 3194331.

Charm, S.E. 1962. The nature and role of fluid consistency
in food engineering applications. Adv. Food Res. 11 :

Charm, S.E. 1963. The direct determination of shear stress
- shear rate behavior of foods in the presence of a
yield stress. J. of Food Science 28 : 107-113.

Cheng, D.C-H., J.A. Hunt and P. Madhvi. 1985. Status report
on process control viscometers : current applications
and future needs. Warren Spring Laboratory, Stevenage,
Hertfordshire SGlZBX, UK.

Cho, Y.I. and J.P. Hartnett. 1982. Non-Newtonian fluids in
circular pipe flow. In : J.P. Hartnett and T.F.
Irvine, Jr. (editors). Advances in Heat Transfer (Vol.
15). Academic Press, New York.

Davis, R.B., D. De Weese and W.A. Gould. 1954. Consistency
measurements of tomato puree. Food Technology 8 : 330-
334.

Dealy, J.M. 1982. Rheometers for Molten Plastics. A
Practical Guide to Testing and Property Measurement.
Van Nostrand Reinhold Company, New York.

Dealy, J.M. 1984. Official nomenclature for material
' functions describing the response of a viscoelastic
fluid to various shearing and external deformations.

J. Rheology 28 (3) : 181-195.

102

103

Dervisoglu, M. and J.L. Kokini. 1986. Steady shear
rheology and fluid mechanics of four semi-solid foods.
J. Food Science 51 (3) : 541-546, 625.

Garcia, E.J. and J.F. Steffe. 1987. Comparison of friction
factor equations for non-Newtonian fluids in pipe flow.
J. Food Process Engr. 9 : 93-120.

Hanks, R.W. and B.L. Ricks. 1974. Laminar-turbulent
transition in flow of pseudoplastic fluids with yield
stresses. J. Hydronautics 8 (4) : 163-168.

Haugen, P. and M.A. Tung. 1986. Rheograms for power law
fluids using coaxial cylinder viscometer and a template
method. Can. Inst. Food Sci. Technol. J. 9 (2) : 98-

104.

Herschel, W.H. and R. Bulkley. 1926. Konsistenzmessungen
von Gummi-Benzollosungen. Kolloid. Zeitschrift 39 :

291-300.

Higgs, S.J. and R.J. Norrington. 1971. Rheological
properties of selected foodstuffs. Proc. Biochem. 6(5)
: 52-54.

Holdsworth, 8.0. 1968. Processing of non-Newtonian foods.
Proc. Biochem. 4 : 15-21, 33.

Holdsworth, 8.0. 1971. Applicability of rheological models
to the interpretations of flow and processing behaviour
of fluid food products. J. of Text. Stud. 2 : 393-418.

Krieger, I.M. 1968. Shear rate in the Couette viscometer.
Trans. of the Soc. of Rheology 12 (1) : 5-11.

Luh, B.S., H.D. Wesley and S. Leonard. 1954. Consistency
of pastes and puree from Pearson and San Marzano
tomatoes. Food Technology 8 : 576-580.

Marsh, G.L., J.B. Buhlert and S.J. Leonard. 1980. Effect
of composition upon Bostwick Consistency of tomato
concentrate. J. of Food Science 45 (3) : 703-710.

Mohamed, I.O. and J.F. Steffe. 1985. Sensitivity analysis
of flow rate calculations of the rheological properties
of Herschel-Bulkley fluids. Trans. of ASAE 28 (4) :
1349-1352.

Mooney, M. 1931. Explicit formulas for slip and fluidity.
J. of Rheology 2 (2) : 210-222.

Ofoli, R.Y., R.G. Morgan and J.F. Steffe. 1987. A
generalized rheological model for inelastic fluid
foods. J. of Text. Stud. : in print.

104

Prentice, J.H. and D. Huber. 1983. Results of the
collaborative study on measuring rheological
properties of foodstuffs. In : R. Jowitt, F. Escher,
B. Hallstrom, H.F.T. Meffert, W.E.L. Spiess and G. Vos
(editors). Physical Properties of Foods. Applied
Science Publishers, London and New York.

Rabinowitch, B. 1929. Uber die Viskositat und Elasticitat
von Solen. Zeitschrift fur Physikalische Chemie 145A :
1-26 0

Rao, M.A. 1977. Rheology of liquid foods. A review. J. of

Rao, M.A. and M.C. Bourne 1977. Analysis of the
plastometer and correlation of Bostwick Consistometer
data. J. of Food Science 42 (1) : 261-265.

Rao, M.A. 1986. Rheological properties of fluid foods. In
: Rao, M.A. and S.S.H. Rizvi (editors). Engineering
Properties of Foods, Marcel Dekker, Inc. New York. ‘

Rao, M.A., H.J. Cooley, J.N. Nogueira and M.R. Mc Lellan.
1986. Rheology of applesauce : effect of apple
cultivar, firmness, and processing parameters. J. of
Food Science 51 (1) : 176-179.

Saravacos, G.D. and J.C. Moyer. 1967. Heating rates of
fruit products in an agitated kettle. Food Technology
21 : 372-376.

Saravacos, G.D. 1968. Tube viscometry of fruit purees and
juices. Food Technology 22 : 89-92.

Saravacos, G.D. 1970. Effect of temperature on viscosity
of fruit juices and purees. J. of Food Science 35 :122-
125.

Scarborough, J.B. 1955. Numerical Mathematical Analysis,
Third edition. The John Hopkins Press, Baltimore, Md.

Steffe, J.F., 1.0. Mohamed and E.W. Ford. 1983.
Rheological properties of fluid foods. ASAE-Paper No.
83-6512, American Society of Agricultural Engineers,
St. Joseph, Mi.

Steffe, J.F., 1.0. Mohamed and E.W. Ford. 1984. Pressure
drop across valves and fittings for pseudoplastic
fluids in laminar flow. Trans. of ASAE 27(2) : 616-
619.

'Tiu, C. and D.V. Boger. 1974. Complete rheological
characterization of time-dependent food products. J.
of Text. Stud. 5 : 329-338.

105

Tung, M.A., J.F. Richards and B.C. Morrison. 1970.
Rheology of fresh, aged and gamma-irradiated egg white.
J. Food Science 35 : 872-874.

Uebersax, M.A. 1986. Personal communication. Department
of Food Science and Human Nutrition. Michigan State
University, East-Lansing.

Van Wazer, J.R., J.W. Lyons, K.Y. Kim and R.E. Colwell.
1963. Viscosity and Flow Measurements : A Laboratory
Handbook of Rheology. Interscience, New York.

Waukesha Foundry Division, Abex Corporation. 1980.
Waukesha Pump Engineering Manual, Third Edition. 1300
Lincoln Avenue, Waukesha, Wi, 53186.

Whorlow, R.W. 1980. Rheological Techniques. Halstad
Press, New York.

HICHIGQN STRTE UNIV. LIBRRRIES
1|llllmllH))llllllM)I)ll)MIMI!)INIIIIMHHHI
31293005770015