u-~~._‘q“1‘\fi~g' awn-x um wan-.(rdvm .«w «a. v luv-I ‘v‘xnzvtvzw‘u‘fi‘ “

..- ,
f:;l’>‘- , '..

.a). .r. r"-
”M . ,. x.
V

.
. fl
, #3.. M. v

‘ . "u,
1"!I'Vl‘fri

"'13:: - - ‘ ' .A .. .fj ,.,— , ”-5.... w -r
, WV" N ‘_ Qt ' 'f.¢:y‘,‘~,.v,n.._:,w.u(5’ 'v
’ ... . ~ .1 at»)... ’;!(\P‘,‘ Wyn-- ' h V
.‘f . .- yr. n-y-O.

.n,'..«,.. ’

--;-. sum. - a ‘ 1,, ‘ . , p 1

”mum. .-‘- . ‘ m ‘7‘:
......n.,:g a.) 4‘ . '_ ' J . - r” ['5'-
. , I ,

.,,.

, ‘ ’4‘ 1. .V A .
”1”,“, . , r 'V‘.‘ V .«. ‘ ‘ rngvf/f
1.1 "1"": ' , . . . 4 a ' . .;7, 3."

. "m", u ‘ ' L-r, { .
W}? ‘” Aw _ 5;” 1,; 1%?»-

if 3'}?- , spy-193' '

. r
“gi- ' , .1 wharf}:- ”In P.-
'_:<-.,'!!ff.....,. ' -' . > I?»
,, -‘.~ vwzw-v'm'fl— gm 5' , ' I. - v
fi$fiéhflufu€wvfl If: ‘ ' f 9 y
. cab-L - 1'

‘V h»
I; , ~ - " _ "-firéi‘flé'aézryQ'eia-Zéz . *r 12?.

a

 

lHFEru

This is to certify that the

thesis entitled

IMPACT LOADING OF CORN AND ITS
EFFECT ON QUALITY

presented by

THOMAS HAROLD BURKHARDT

has been accepted towards fulfillment
of the requirements for

Ph.D. degree in Agr. Engr.

gifl 322w:

Major professor

Date February 9, 1971

0-7639

Michigan State
rsity

Unive

 
     

 

 

 

 

 

 

 

 

 

 

   

 

 

 

 

ABSTRACT

IMPACT LOADING OF CORN AND
ITS EFFECT ON QUALITY

By

Thomas Harold Burkhardt

Corn kernels are connected to the cob by pedicels which are
fractured during shelling. The phenomena associated with the initiation
of corn shelling were investigated by subjecting whole ears to radial
compressive impact loading. Impact velocities comparable with combine
cylinder peripheral velocities commonly found during field shelling were
utilized while testing two varieties of corn. Precautions were taken to
reduce biological changes in the samples between the time of collection
and the time of testing.

The impact process was recorded by high—speed cinematography.

The resulting movies were examined with a movie analyzer to estimate the
energy absorbed by corn ears during impact. Study of these movies also

revealed shelling is initiated by radial compressive impact loading as a
result of inertial forces and the wedging effect proposed by Halyk (1968).

The maximum impact force was measured for every sample tested.

Regression analysis showed that for both varieties impact force was increased

by increases in ear weight, moisture content or impact velocity. Impact
force increased as the harvest season progressed for the Michigan 250

variety, while the opposite was true for the Michigan 500 variety.

 

 

Every kernel of all samples tested were inspected and kernels with
visible mechanical damage were weighed. Mechanical damage for the
Michigan 250 variety increased with: (1) decreases in kernel moisture content
or impact velocity or (2) increases in cob moisture content or testing
date. For the Michigan 500 variety mechanical damage decreased as the
harvest season progressed until November 4, and then it increased as
the season progressed. Damage for Michigan 500 also became more severe
with increases in kernel moisture content.

Kernels removed by a single impact were weighed. Increases in impact
velocity or decreases in cob moisture content or ear weight increased
kernel removal for the Michigan 250 variety. Increases in moisture
content, testing date, impact velocity or ear weight increased kernel

removal for the Michigan 500 variety.

   

Approved
Majo Professor

Approved K

Department Chairman

 

IMPACT LOADING 0F CORN AND

ITS EFFECT ON QUALITY

Thomas Harold Burkhardt

A THESIS

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

DOCTOR OF PHILOSOPHY

Department of Agricultural Engineering

1970

 

 

 

ACKNOWLEDGMENTS

The author wishes to express a deep appreciation to Dr. B. A. Stout
(Agricultural Engineering) for his enthusiastic support and interest
while serving as Major Professor for this graduate program.

The contribution of Dr. E. C. Rossman (Crop Science) in supplying
the testing materials for this study and serving as a member of the
guidance committee was deeply appreciated.

Many thanks are due to Dr. J. L. Gill (Dairy), Dr. J. B. Holtman
(Agricultural Engineering) and Dr. G. E. Mase (Metallurgy, Mechanics
and Material Science) for serving as members of the guidance committee.

The author wishes to express his gratitude for the financial assis-
tance given by the National Science Foundation to help make this graduate
program possible.

Many thanks are due to Mr. Larry Rose for his assistance in
conducting the experiments and tabulating the data.

The Agricultural Engineering Department, University of California,
Davis, very generously provided facilities and support during the writing

of this thesis; for this aid the author expresses his appreciation.

ii

(subs-m ! i"? in .‘u Huai 13A)

 

 

 

   

To

My wife, Patricia, for her

 

assistance and encouragement.

 

iii

 

 

 

TABLE OF CONTENTS

ACKNOWLEDGMENTS . . . . . . . . . . . . . . . . . . . . .

DEDICATION.

LIST OF TABLES. . . . . . . . . . . . . . . . . . . . . .

LIST OF FIGURES . . . . . . . . . . . . . . . . . . . . .

LIST OF APPENDICES. . . . . . . . . . . . . . .

Chapter
1.

2.

6.

INTRODUCTION . . . . . . . . . . . . . . . . .

AN ANALYSIS OF PAST INVESTIGATIONS . . . . . .

2.1 Corn Harvesting. . . . . . . . . . .
2.2 Corn Quality . . . . . . . . . . . .
2.3 Shelling Theory. . . . . . . . . . .

OBJECTIVES . . . . . . . . . . . . . . . . . .

APPARATUS AND PROCEDURE. . . . . . . . . . . .

4.1 Description of Apparatus . . . . .
4.2 Comparison of Apparatus with Combine

Cylinder. . . . . . . .
4. 3 Procedure for Quality Studies.

DISCUSSION OF RESULTS. . . . . . . . . . . .

5.1 General Observation. . . . .
5.2 Regression Analysis Models . . .

SOME THEORETICAL ASPECTS OF CORN IMPACT. . .

6.1 Extension of Corn Shelling Theory to
Dynamic Range . . . .

6. 2 Analysis Utilizing High— speed Movies .

6. 3 Energy Considerations. . . . . . .

iv

Page

ii
iii
vi
vii

viii

U'Ib.)

15

16

l6

19
26

32

32
36

45

45

47
49

 

 

 

 

     

7 I CONCLUSIONS I I I I I I I I I I I I I I I I I I I I I I

7.1 Findings ofthis Study. . . . . . . . . . . .
7.2 Recommendations for Future Studies. . . . . .

 

REFERENCE 3 I I I I I I I I I I I I I I I I I I I I I I I I I I I 5 4

APPENDICES I I I I I I I I I I I I I I I I I I I I I I I I I I I 59

 

 

 

 

 

 

LIST OF TABLES

Table Page

2.1 GRADE REQUIREMENTS FOR YELLOW CORN, WHITE

CORN AND MIXED CORN. . . . . . . . . . . . . . . . 6
2.2 DAMAGE MEASUREMENT TECHNIQUES. . . . . . . . . . . . . 11
4.1 SUMMARY OF TESTING PROCEDURE . . . . . . . . . . . . . 28

A2.l SUMMARY OF REGRESSION ANALYSIS PARAMETERS
(KERNEL REMOVAL, MICHIGAN 250) . . . . . . . . . . 68

A2.2 SUMMARY OF REGRESSION ANALYSIS PARAMETERS
(KERNEL REMOVAL, MICHIGAN 500) . . . . . . . . . . 69

A2.3 SUMMARY OF REGRESSION ANALYSIS PARAMETERS
(MECHANICAL DAMAGE, MICHIGAN 250). . . . . . . . . 7O

A2.4 SUMMARY OF REGRESSION ANALYSIS PARAMETERS
(MECHANICAL DAMAGE, MICHIGAN 500). . . . . . . . . 71

A2.5 SUMMARY OF REGRESSION ANALYSIS PARAMETERS
(IMPACT FORCE, MICHIGAN 250) . . . . . . . . . . . 72

A2.6 SUMMARY OF REGRESSION ANALYSIS PARAMETERS
(IMPACT FORCE, MICHIGAN 500) . . . . . . . . . . . 73

A2.7 RANGE OF PARAMETERS FOR WHICH REGRESSION
IS VALID . . . . . . . . . . . . . . . . . . . . . 74

vi

 

 

 

 

LIST OF FIGURES

Figure Page
4.1 IMPACT TESTING MACHINE SHOWING COMPONENTS. . . . . . . 17
4.2 POSITION 0F IMPACTER AND SPECIMEN AT TIME

OF INITIAL CONTACT . . . . . . . . . . . . . . . . 18
4.3 CORN EAR WEIGHING 0.461 POUND BEING IMPACTED

AT 2500 FEET PER MINUTE. . . . . . . . . . . . . . 22
4.4 END VIEW OF CYLINDER AND CONCAVE AREA OF

A GRAIN COMBINE. . . . . . . . . . . . . . . . . . 23
4.5 CORN EAR IMPACTED BY CYLINDER BAR. . . . . . . . . . . 24
4.6 TYPICAL FORMS OF MECHANICAL DAMAGE FOUND IN

THIS STUDY . . . . . . . . . . . . . . . . . . . . 31
5.1 TYPICAL SHELLING PATTERN FROM LOW—VELOCITY

IMPACTS AT LOW AND MEDIUM MOISTURE

LEVELS . . . . . . . . . . . . . . . . . . . . . . 33
5.2 TYPICAL SHELLING PATTERN FROM HIGH—VELOCITY

IMPACT AT HIGH MOISTURE LEVELS . . . . . . . . . . 33
5.3 KERNELS REMOVED FROM AN EAR OF THE MICHIGAN

250 VARIETY WITH A SINGLE IMPACT AT

4000 FEET PER MINUTE . . . . . . . . . . . . . . . 35
5.4 KERNELS REMOVED FROM AN EAR OF THE MICHIGAN

500 VARIETY WITH A SINGLE IMPACT AT

4000 FEET PER MINUTE . . . . . . . . . . . . . . . 35

A3.l THEORETICAL MODEL OF EAR CROSS SECTION . . . . . . . . 78

vii

 

   

 

 

LIST OF APPENDICES

Appendix Page
I HIGH-SPEED CINEMATOGRAPHY . . . . . . . . . . . . . . 60
II STATISTICAL METHODS . . . . . . . . . . . . . . . . . 65
III STRESS WAVE PROPAGATION . . . . . . . . . . . . . . . 75

viii

 

 

 

1. INTRODUCTION

The most important crop grown in both Michigan and the United
States is corn. Average yearly Michigan production of corn during the
period of 1964 through 1967 was 92.2 million bushels valued at $104.2
million. Approximately 54.3 percent of this corn was fed on the farm
where it was grown with the remainder being marketed as a cash crop.

The corn which was sold accounted for six percent of the total Michigan
cash receipts from farm marketings (Michigan Department of Agriculture,
1968). The total disappearance of corn in the United States in 1965 was
4.392 billion bushels. Of this total, 76.2 percent was fed to livestock,
15.2 percent was exported and 0.3 percent was used for seed. The
remaining 8.3 percent was used for wet process products (4.6 percent),
corn meal and grits (2.5 percent), alcohol and distilled spirits

(0.7 percent), and breakfast foods (0.5 percent). The grain equivalent
of the starches, syrups, sugar, meal, oil, flour and other foods containing
corn that shoppers select amounts to about one bushel per person per
year. In a recent survey of the largest stores of five food Chains

in the Washington, D.C. area, 650 different items which contained one or
more of the products of the wet milling industry were counted (Liebenow,
1968).

The corn kernel is attached to the cob by means of a structure
known as the pedicel (Wolf et_al., 1952). During the shelling process
pedicels are fractured thereby freeing kernels from the cob. A tangential
force applied to kernels at the butt of the ear causes the pedicels to bend

until they fail. If a kernel is not located at the end of the ear,

 

 

 

surrounding kernels prevent tangential motion and the corn cannot be
shelled in this manner. However, high—speed photography has revealed
that when an impact load is applied by the corn harvester near the middle
of the ear, shelling is initiated. Thus, there must be forces other than
bending which cause shelling. To date, no theory has been developed to
explain the phenomena associated with shelling.

When corn was harvested by hand or mechanical picking, the ears
were dried before the shelling. Since the advent of harvesting by grain
combines equipped with corn heads, shelling takes place at high moisture
levels. Mitchell (1956) reported that much of the field shelling is
initiated at moisture levels from 25 to 30 percent, while Johnson gt El.
(1963) encountered moisture levels from 20 to 40 percent. Because of
these high moisture levels and the high velocity impact loading encountered
at the combine cylinder, many kernels are cracked or broken. Consequently,
the combine is being blamed for much of the quality loss which takes
place between the field and the feeder or processor (Bailey, 1968).

A basic understanding of the shelling process will aid in the development

of a new corn shelling process.

 

 

 

I,

 

2. AN ANALYSIS OF PAST INVESTIGATIONS

2.1 Corn Harvesting

In the eighteenth and nineteenth centuries corn ears were manually
picked, husked and tossed into wagons. The cost of hiring men for this
tedious operation became an economic burden to the American farmer
because of the competition of higher wage levels for non—farm employment.
In spite of the difficulties associated with manual corn harvesting, the
introduction of mechanical harvesting to general use was a difficult, slow
process.

The first patent issued on a corn harvester was on October 8, 1850,
to E. Quincy of Lacon, Illinois. The second corn harvester patent was
issued just one week later to W. Watson of Chicago, Illinois. Both
machines were listed as being able to strip heads from Indian corn.

The first patent for a corn picker and husker was issued to L. Devore of
Victor, Iowa on August 22, 1871, but the first picker and husker did

not appear commercially until 1909 (Aspenwall, 1924; Seferovich, 1961).
McKibben (1929) field shelled corn with a grain combine in 1928, at Iowa
State College. He reported that prior attempts at field shelling had
occurred in Kansas and Nebraska.

American farm equipment manufacturers have since developed reliable
corn heads for grain combines, and field shelling with these machines
has become a common method for corn harvesting. The farmers of Indiana,
Illinois and Iowa used grain combines to harvest 45.1, 38.0 and 12.7

percent of their corn crops, respectively, during the Fall of 1964, and

 

 

 

 

58.8, 57.0 and 34.6 percent, respectively, during the Fall of 1968.
During this same period of time the use of corn pickers dropped from
47.2 to 32.0 percent in Indiana, from 55.0 to 35.0 percent in Illinois,
and from 81.2 to 56.6 percent in Iowa. The corn not accounted for in

the above figures was harvested with a picker—sheller (Zimmerman, 1968;

Anonymous, 1969).

 

 

2.2 Corn Quality

The Federal Grain Standards Act of 1916 established the first
official federal grade standards. Prior to this legislation more than
fifty hometown versions of grain grading were recognized. One such
standard was described by Maywald (1968):

Corn was tested by picking kernels willy—nilly from boxcars,

and the kernels were put in a saucer of water. The ones that

sank represented a ”good" ear of corn, and the ones that
floated were "bad" corn.

The current Federal Grade Standards (Table 2.1) are used as the
basis for the market price in corn trading. Number Two corn is the
standard trading grade for buying and selling corn and the price quota—
tions are based on this grade. When a lot of corn is sold that does
not grade Number Two, discounts or premiums which have been established
by the purchaser are used to adjust the market price accordingly.
Normally the size of the discounts and premiums that a particular buyer
offers depends upon (1) the intended use of the corn, (2) the grade
factors of other bins of corn available for use in blending and, (3)
the prevailing market discounts and premiums (Uhrig, 1968). For example,
a lot of Number Three corn which grades Number Two on all factors except
moisture would probably be more valuable to a breakfast food manufacturer
than would another lot of Number Three corn which graded Number Two on

all factors except cracked corn and foreign material.

During this investigation mechanically damaged kernels have been
considered to be those broken into several pieces, those cracked but

remaining intact, and those with tip caps removed.

 

 

TABLE 2.1 GRADE REQUIREMENTS FOR YELLOW CORN,
WHITE CORN AND MIXED CORN.*

Minimum Maximum limits
test Cracked
weight corn and Damaged kernels
Grade per foreign Heat-
No. bushel Moisture material Total damaged
pounds ————————— percentage —————————
l 56 14.0 2 3 0.1
2 54 15.5 3 5 0.2
3 52 17.5 4 7 0.5
4 49 20.0 5 10 1.0
5 46 23.0 7 15 3.0
Sample Sample grade shall be corn that (a) does not meet
Grade the requirements for any of the grades from No. l

to No. 5, inclusive, (b) contains stones, (c) is
musty, sour or heating, (d) has any commercially
objectionable foreign odor, or (e) is otherwise of

distinctly low quality.

* From Uhrig (1968)

 

 

 

 

The first economic losses associated with mechanical damage are
the so-called invisible losses, consisting of corn chips and meal that
pass out the rear of the combine and the corn tips left in the cobs.

If these particles were to remain attached to the kernels, the producer
would realize higher returns due to the increased weight of the corn
being sold. Byg and Hall (1968) reported these invisible losses to
range from an average of 0.8 percent at 20 percent moisture to an
average of 2.9 percent at 35 percent moisture.

In his early attempt at field shelling McKibben (1929) recorded
mechanical damage of kernels as not excessive with the kernel moisture
content at 18.4 percent. However, many researchers have since reported
the mechanical damage of kernels to be excessive during field shelling
with grain combines. When temperatures are above 55 degrees Fahrenheit
and the moisture content is above 13.5 percent at harvest time, shelled
corn provides ideal ambient conditions for the growth of molds and fungi
(Bailey, 1968). Kernels which lose their tip caps during harvest have
exposed germs. According to Bailey (1968) the germ is the first portion
of the kernel to be discolored and destroyed by the growth of microflora.
Mechanically damaged kernels provide a ready source of food for microflora.
Uhrig (1968) states the problem well:

Corn damaged in the field—shelling process provides ideal growth
conditions for molds. Molds produce chemicals (toxins), known as
mycotoxins, that are toxic to livestock. One mold scientifically
termed aspergillus flavus [sic], Produces a compound known as
aflatoxin. Aflatoxin seems very dangerous because a small amount
can cause pathological changes in poultry, cattle and swine months
after aflatoxin feeding has ceased. Meanwhile, the animals may
appear in poor health without specific symptoms of a particular
disease. The Pure Food and Drug Administration personnel say there
is no way to restore moldy corn and, if seized, the corn will not
be allowed to be mixed in other grain either for human or animal

consumption. This could cause serious losses or even bankrupt
grain shippers.

 

 

Since mechanically damaged corn provides a source of food for
microflora, high moisture corn with high damage will deteriorate more
rapidly than high moisture corn with low damage. As microflora grow,
foodstuffs are oxidized and carbon dioxide is produced. Saul and
Steele (1966) measured carbon dioxide produced by corn to estimate the
loss of dry matter, and hence, the deterioration of the corn. At 65
degrees Fahrenheit a sample of corn at 2.0 percent damage and 28 percent
moisture required 600 hours for a one percent loss of dry matter. Another
sample which had 28.7 percent mechanical damage had a one percent dry
matter loss in only 200 hours at the same temperature and moisture level.

Because of its higher rate of deterioration, corn with a high
level of mechanical damage needs to be dried more rapidly than corn with
a low level of mechanical damage. If the corn iS not dried rapidly
enough, the grower will receive a lower market price due to the presence
of mold. On the other hand if the corn is dried too rapidly, the grower
will receive a lower market price due to the presence of heated kernels.

In order to properly dry highly damaged corn, equipment with a large
capacity is needed. Additional capital outlay for increased drier capacity
burdens the grower economically with higher operating expenses (Saul and
Steele, 1966). If the crop is dried commercially, these increased

costs would be passed on to the grower in the form of a lower market price.

During the shelling process some kernels are cracked internally
or externally, but remain as whole kernels. These kernels may be fractured
into several pieces during drying or subsequent handling operations.

When corn is graded, broken kernels are considered the same as foreign

material and their presence lowers the market price. Bailey (1968)

 

 

 

reported that broken kernels cost the American farmer approximately

three cents per bushel on all the corn he sells.
These broken kernels cause an additional storage problem as
described by Bailey (1968):
The increase in broken kernels is aggravated by a phenomenon
called by the trade a "spoutline". Whenever a stream of part-
icles of varying sizes is poured onto a pile, the smaller
particles tend to stay in the center and the larger particles
roll or flow down the slope. In corn, about one—third of the
mass is voids between the kernels. In the pile, the fines tend
to fill these voids and a little more so that the concentration
of fines will go to about 35 percent and form a verticle column
in the center of the pile.
In this area there is no circulation of air and so any heat
developing from mold activity on the high concentration of
broken kernels cannot escape but accumulate, causing progressive
heating and damage.
Another consequence of these "spoutlines" is the difficulty of
obtaining a representative lot of corn, or of making uniform
deliveries.
That broken and cracked kernels are not desirable for corn to
be processed is apparent; however, it may not be immediately obvious
why the presence of broken and cracked kernels in corn to be used for
livestock feed is undesirable. It appears that the feeding value of
corn will decrease more rapidly during storage when large amounts of
mechanically damaged kernels are present. Since the seed coat helps
to protect the interior portion of the kernel from the surrounding
environment, kernels with fractured seed coats should be more susceptible
to any biological activity which would lower their nutrient content. If

this is the case, even the grower who feeds all his corn to livestock

is affected by mechanical damage.

 

   

 

,-
L

10

Mechanical damage lowers the profits of corn growers by invisible
losses, accelerated deterioration, increased drying costs and lowered
market price.

Several methods have been used in past studies to quantify
mechanical damage (Table 2.2). The carbon dioxide, temperature rise and
mold growth techniques were all judged to be too time consuming for the
scope of this study. In addition to being time consuming, the water
absorption and light absorption techniques also have notable system
errors. When the germination technique is used, it is extremely difficult
to determine whether the viability reduction is due to mechanical damage
or some other form such as disease or insect damage.

When corn is graded to determine market price, corn particles
which pass through a sieve with round holes twelve sixty—fourths of an
inch in diameter are considered to be foreign material (Uhrig, 1968).
However, this method does not detect larger partial kernels or cracked
whole kernels. These undetected forms of damage probably cause losses
during subsequent handling operations. After careful consideration of
the various damage measurement techniques, it was decided that the
visual separation technique was most appropriate for this study. The
kernels were examined individually and separated into three categories:

1. No visible damage.

2. Broken kernels or visible cracks.

3. Tip caps removed.

The sum of categories two and three gives an estimate of total
mechanical damage. However, this may be an optimistic estimate since

some internal cracking may not be visible.

 

 

 

ll

 

smanoum ommama mwmnxmoz unmaumouH

Hmowufluo Hmwuo>wnb umoz amusuosuum ou mnflavcmm

ma unaudoo MH owmxmmum HMGOHuuoaoum ma bumocmum mp

unaumfloz awofiwznm xasm Hmoflwksm wwoq huflamso mmumwuo mwcflm

mamnuom

bommemm

muouum mucofiuuoz Eouw uomuuxm

awumMm ou mmmema Ou oanmmfipom “mums am

kummEoo HHnEm show on ouoz we umoo coaumuomn<

mwoaouommwn Hmwuaouwwoum xasm awowmkcm boom memosma ucwflg

muouum uwum3 maneuwx

Eoumhm ou ou oapmofipmm mommamn xn

wwnmmaoo HHmEm unoEmHsvm who: we umou GOHumuown¢

moonouowwflo o>fiwcoaxm oz Masm Hmuwm>nm comm vowmemn umumz

coausao>m

umom cu

%Hmm=m cowhxo >ufi>fluo< one Hosflmucoo

mam ousumuomEoH coaumMOflumumm HmofiMOHOHm kumHsmaH

.ousuwfloz Go Hmsuo< who: uuomasm cm aw wmflm

ucwwcoaon mno> monommoz Masm Hooflwoaowm wawcuox commemn ousumuoaaoe

hammflm cmwkxo mufl>fluo<

mam manumuomfioe coaumuOHumuon HMUHwOHOflm awuwmm boamwm

.mHSuwwoz so Hmsuo< who: uuoamom m as cofiusao>m

uawwnommn >Ho> mmusmmmz xasm HmUHmOHOHm maonuox vomemQ mvHxOHQ Conumo

meowumufiawa smou¢m>w< wwoz wmusmmmz oaafiuswum casumfiuomwo
Hmafioawum Hmmwocflum uon wwamso Hmaowuoasm ummH

 

«mdeHszMH HZMSMMDmdMZ m0<z<n N.N MAM<H

 

 

 

 

12

.Amomav mmmcw< aoum *

 

 

%ufl>fluo¢

osHH can HmuwwOHOHm oumaamoau<
wudumuoaawH sowumHOHHoumm who: uuommsm woaaonu
.oMSumHoS do Hmsuom waodumm Icoo m as
uawwaoawn huo> mousmmoz wuouumfln HmowwOHOHm wmmemQ £u3ouu vaoz
mwmama
omnfimo Esme kuflaflnmfl> ou HmUfim%sm ou
monummwz HMGOHuuoaoum mom mmoq
wswasmaoo mafia kauoopflm ououomfln HmoflwOHOflm ma zufiamso cowumsfiauoo
maoapwx
oswfiumm mapflmfl> mommama mo
was uaoamwsh Hwaumm mom oso ma wwmamm coaumumaom
amesm sw>m vsfim cmo wumuowfla Hmofiwmrm usmuuomEH Hmsmfl>
o>owm o>mflm
mwmfimn Hmuoy swsounH mwmm wumvcmum LuHB
wuawmoz waocuwx :oxoum coaumummom
uoz moon madam xadm Hmoflwmsm mo woaofluumm Havanmgooz
mGOfiumufiEfiA mmmuam>v< mwo: wwuswmwz waawonflpm soaumfiuomma

Hmawudflum Hmmfiocwum umoH ownmfio HNCOHuossm umoH

 

$625303 «$337589 Hzngmfiz mega N.N mama

 

 

 

13

2.3 Shelling Theory

When a corn ear passes between a cylinder and concave bars in a
combine harvester, it is subjected to a combination of compressive and
tangential forces. In some unknown fashion this combination of forces
initiates the shelling process. Once some kernels have been removed,
the others probably are removed by tangential forces, which cause the pedicels
to fail in bending, thus freeing the kernels from the cob.

The process through which shelling is initiated by a compressive
force has not been adequately explained. Johnson g£_al. (1969) investigated
the process of kernel removal by hand—shelling four rows of kernels on
opposite sides of corn ears, and subjecting the samples to quasistatic
compressive loading. When the load was applied to the exposed cob, no
shelling occurred; but loading the unshelled portion of the ear caused
shelling. Shelling was also successfully initiated by applying low—
velocity impact loading to whole ears.

Halyk 23 a1. (1969) investigated the mechanics of radial compressive
shelling under quasistatic and low—velocity impact loading. They reported
the initiation of shelling in a region near the point Where a radial
compressive load was applied to an ear. Under quasistatic loading
no shelling was achieved above a kernel moisture content of 15.3 percent
w.b. Using a vertical drop—weight tester shelling was achieved with kernel
moisture levels of 28 percent w.b. or lower. They developed a model

representing the shelling process after reducing the mechanics to the

 

 

 

l4

plane stress case by assuming that:

1. The kernels and cob are composed of a rigid material;

2. The kernel shape closely approximates that of an ordinary

truncated wedge; and

3. Pedicel behavior can be approximated by the reaction of an

appropriate viscoelastic connector.
It was shown that a kernel displaced radially inward acts as a wedge
and causes displacement of the surrounding kernels. As a result of this
displacement the pedicel of one of the surrounding kernels is fractured,
thereby initiating shelling.

Area under the force-displacement curves from quasistatic loading
was used by Halyk et_al, (1969) to predict the minumum energy required
to induce shelling. This procedure proved successful when the kernel
moisture level was below 15.3 percent. Their work was an important step
toward the understanding of corn shelling. However, testing needed to
be done using impact velocities and kernel moisture levels similar to
those encountered during field shelling. Hence, this theory was useful
as a foundation for the development of a corn shelling theory suitable

for high impact velocities and high moisture levels.

 

 

 

3. OBJECTIVES

When man does not understand a phenomenon, there is sufficient
reason for study. However, when a phenomenon has the economic importance
of corn shelling, there is even more justification for study. Besides
adding to man's understanding, the development of a corn shelling
theory could also be helpful to engineers in the development of more
efficient corn harvesters. Kaminski (1968) pointed out the need for
such information. Accordingly, the objectives of this study were to:

1. Investigate the initiation of shelling of high moisture

corn by radial compressive impact loading.

2. Determine the magnitudes of the impact forces and energies

applied to an ear during shelling.

3. Study the effects of impact loading on corn quality.

15

 

 

 

4. APPARATUS AND PROCEDURE

4.1 Description of Apparatus

The impact testing machine employed in this study utilizes a
rotary motion similar to that of a pendulum (Burkhardt, 1969). However,
mechanical energy is used in lieu of gravity for accelerating the impacting
arm. This rotary impact testing machine consists of two major units
(Figure 4.1). The primary unit has high rotational inertia and the secondary
unit contains an impacting arm with low rotational inertia. When the pri—
mary unit has achieved the selected rotational velocity, the two units
are coupled by an electric clutch and the impacting arm is forced to
rotate. After approximately 330 degrees of rotation, the arm strikes
a specimen. The arm is then decelerated by an electric brake and stopped
approximately one revolution after impact. The clutch and brake are
automatically controlled by an electronic circuit.

Impact force is measured by a quartz load cell mounted between the
impacting face and the arm (Figure 4.2). The signal from the load cell
is conditioned by a charge amplifier and recorded on a storage oscillo—
scope. An electronic timer measures the time required for the arm to travel
a known distance. The impact velocity can be calculated from this infor-
mation.

The impact area is surrounded by a collection box to facilitate
the retrieval of the specimen after each test. The interior of the box

is lined with foam rubber to prevent damage to particles hitting the box.

16

 

 

17

 

FIGURE 4.1 IMPACT TESTING MACHINE SHOWING THE FOLLOWING COMPONENTS:
(PHOTO 68-7l.)*

l.
2
3
4.
5.
6
7
8
9

ELECTRIC MOTOR
VARIABLE SPEED PULLEY
TACHOMETER GENERATOR
FLYWHEEL

ELECTRIC CLUTCH
ELECTRIC BRAKE

CAM AND MICROSWITCH
BRAKE CONTROL SWITCH
IMPACT ARM

* Negative filed in Agricultural Engineering Department, Michigan
State University.

 

 

 

18

COLLECTION BOX IMPACTING ARM

  
   

'//////////////j

   
  
   
   

IMPACTING

MASKING TAPE
CORN EAR MASKING TAPE

FIGURE 4.2 POSITION OF IMPACTER AND SPECIMEN AT TIME OF INITIAL CONTACT
F = FORCE MEASURED BY LOAD CELL
F = FORCE IN TAPE
F = DEFORMATION FORCE

F = INERTIAL FORCE

 

 

19

4.2 Comparison of Apparatus with Combine Cylinder

This study was solely concerned with radial, compressive impact
loading. The impacting face was a cylindrical piece of steel shafting
two inches in diameter and one-half inch long. At the time of contact
the flat surface of the face was parallel to a plane tangential to the
corn ear. The load was applied approximately halfway between the butt
and tip of each ear. Prior to impact each specimen was suspended from
the top of the collection box by two pieces of masking tape (Figure 4.2).

Johnson and Lamp (1966) reported that combine cylinders are
operated at peripheral speeds ranging from 2500 to 4000 feet per minute
during corn harvesting. In order to cover the range of cylinder peri—
pheral velocities, tests were run at impact velocities of 2500, 3000,
3500 and 4000 feet per minute.

Torsional pendulum tests were performed at the Massey—Ferguson
Harvesting Systems Laboratory in Toronto, Canada using a rasp bar cylinder
of 22 inches outside diameter. From these tests it was estimated that
the cylinder and shaft assembly have a moment of inertia of 36 lb. in.

sec2 (Cooper, 1969). The arm and flywheel of the impact apparatus used

in this study have a combined moment of inertia of 60 lb. in. secz.
When a combine is used for field shelling at a nearly constant

feed rate, the cylinder cannot have a velocity loss each time an indiv-

idual corn ear is impacted; otherwise, the cylinder would decelerate

to an inefficient velocity or a complete stop. Since the rotational

inertia of the apparatus used in this study is greater than that of a

 

 

 

 

 

 

20

typical combine cylinder, there was no measurable velocity loss during
impact of a specimen.
The impact forces measured during this study can be represented

as follows (Figure 4.2):

F = FA + FD + 2FT
where:
F = force measured by load cell
FA = inertial force
FD = deformation force
FT = force in tape

Tensile tests were run using an Instron Model TM testing machine.
It was found that for a loading rate of one centimeter per minute (.0328
feet per minute) the masking tape could withstand an average force of
approximately 22 pounds before breaking. When the loading rate was
increased to 100 centimeters per minute (3.28 feet per minute), the masking
tape fractured when the force reached only 2.5 pounds. The tape appar—
ently became more brittle with increased loading rates. During the tests
the tape did not adhere well to the moist corn ears and usually slipped
free from the specimen without breaking. Since all of the impact forces
measured were in excess of 400 pounds and the combined forces in the
supporting tape were less than five pounds, the tape forces were considered

insignificant.

21

Examination of the high-speed movies provided further evidence
for disregarding the support forces. A series of frames (Figure 4.3)
revealed that upon impact the centerline of the ends of the corn ear
remained horizontal while the middle of the ear became curved. The curva-
ture of the ear progressed outward until the total ear was deformed.
This indicated the force required to accelerate the specimen was large
compared to the support force in the tapes. If the support force were
large, the whole ear would have curved instantaneously.

A cross—sectional view of the end of the cylinder-concave area
of a grain combine is shown in Figure 4.4. Normally rasp bar cylinders
are used for field shelling. Filler plates are added to force passage
of the ears between the cylinder and concaves. Otherwise, the ears could
pass between the cylinder bars with imcomplete shelling resulting. The
center axis of the ear is shown parallel to the axis of rotation of the
cylinder. At least one manufacturer currently usus a feeder elevator
which introduces the ears oriented in this manner (New Holland, 1969).

The cylinder bar approaches the ear with peripheral velocity, V,
and upon impact a force, F1’ is applied by the cylinder bar (Figure 4.5).

The cylinder force is opposed by: (l) the concave force, F ' (2) the defor—

2,
mation force, FD; and (3) the inertial force, FA' In this case FA

includes the force required to cause rotational acceleration of the ear
as well as the linear acceleration. The complex nature of the loading

geometry prevents the writing of a simple equation relating the various

forces.

 

 

 

22

   

T = 0 microseconds T

438 microseconds

   

T = 146 microseconds T 584 microseconds

   

T = 292 microseconds T = 730 microseconds
FIGURE 4.3 CORN EAR WEIGHING 0.461 POUND BEING IMPACTED AT 2500 FEET

PER MINUTE. KERNEL MOISTURE CONTENT IS 19.6 PERCENT AND

COB MOISTURE CONTENT IS 22.4 PERCENT.

 

 

 

23

Filler plate Cylinder bar

 

FIGURE 4.4 END VIEW OF CYLINDER — CONCAVE AREA OF A GRAIN
COMBINE. THE RASP BAR CYLINDER IS EQUIPPED WITH FILLER
PLATES AND IS APPROACHING THE CORN EAR WITH A

PERIPHERAL VELOCITY, V.

 

 

 

 

 

25

Since the concave bar resists the motion of the ear, much more
deformation occurs than in the case of the impact testing apparatus.

Because of this higher deformation the impact force, F is likely to

1
be higher than the impact forces as measured during this study. If a

"mid-air" collision occurs before the ear comes in contact with the con—
caves,the loading geometry is approximately the same as that used during

this study. Thus, the corn ears encounter forces during field shelling

at least as great as those reported in this study.

 

26

4.3 Procedure for Quality Studies

Three varieties of corn, Michigan 500-2x (105 day maturity),
Michigan 250-2X (85 day maturity) and A509 X MA1334 (83 day maturity),
were‘planted May 24, 1968, on a Michigan State University farm. Flooding
because of unusually heavy rains during the month of June ruined approx-
imately eighty percent of the crop and lowered the quality of the remaining
twenty percent. As a result of this water damage the ears of the A509 X
MA1334 variety were too small and only the other two varieties were used.

In order to simulate field shelling in the laboratory, it was
important for the testing equipment to apply impact loads similar to those
applied by the combine cylinder. It was also important for the mechanical
properties of the samples to be similar to those of corn ears being
field shelled. In a real harvesting situation the ears are delivered
to the cylinder for shelling only a few seconds after they are collected.
The testing equipment used for this study were stationary and could
not be moved to the field. Consequently, it was necessary to gather the
samples in the field and haul them to the laboratory, resulting in a short
time delay between collection and testing.

After a corn ear is detached from its stalk, it is probable that
some biological activity takes place. Thus, precautions had to be taken
to minimize the biological changes of the samples during the delay between

the collection and testing to prevent changes in the mechanical properties

of the samples.

 

 

 

 

27

One biological change which seriously affects mechanical properties
of corn is a change of moisture content. In order to prevent moisture loss
of the samples before testing, each ear was manually removed from the stalk
with its husk in tact and placed in a plastic bag. The ambient temperature
in the corn field was as much as 60 degrees Fahrenheit lower than the labor-
atory temperature. It was felt that a large temperature rise caused
by indoor storage might have accelerated any biological changes initiated
by severing the ear from the plant. Consequently, the samples were stored
in a shaded area outside the laboratory where the ambient conditions were
similar to those found in the field.

Preliminary experiments were conducted on October 8, 1968, to try
out the testing procedure. Full scale testing was initiated on October 15.
The impact velocities, varieties and testing dates utilized in this study
are listed in Table 4.1.

A full scale test consisted of impacting five ears from each of
the two varieties at each of four impact velocities for a total of forty
ears impacted per test. From the statistical standpoint it would have
been more desirable to have used a random selection 0f the variety
velocity treatment combination. At this time, however, the testing equip-
ment was newly developed and its operation procedure was not well known.
The changing from one impact velocity to another was rather time consuming
at first due to lack of experience of the operators. In an effort to reduce
the delay time between collection and testing it was decided to proceed
through the impact velocities from slow to fast or fast to slow, thus
reducing the time spent changing from one velocity to another. Five

ears from each variety were tested at each velocity setting.

 

 

 

Testing
Date

October

8

October
15

October
22

October
29

November
5

TABLE 4.1 SUMMARY OF TESTING PROCEDURE

Impact velocity
(feet per minute)

4000

4000
4000

2500
3000
3500
4000

2500
3000

3500
4000

2500
3000
3500
4000

2500
3000
3500
4000

2500
3000
3500
4000

2500
3000
3500
4000

2500
3000
3500
4000

Number of
ears tested

U1U‘IU‘IU1 UTUIUI‘JI U1U'IU'IU‘I \J‘IU'IU'IUW U‘IU'IU'IUI U'IUIU‘IUI

U1U‘IU1U'I

Variety

Mich.

A509

Mich.

Mich.
Mich.
Mich.
Mich.

Mich.
Mich.
Mich.
Mich.

Mich.
Mich.
Mich.
Mich.

Mich.
Mich.
Mich.
Mich.

Mich.
Mich.
Mich.
Mich.

Mich.
Mich.
Mich.
Mich.

Mich.
Mich.
Mich.
Mich.

250

X MA1334

500

250
250
250
250

500
500
500
500

250
250
250
250

500
500
500
500

250
250
250
250

500
500
500
500

250
250
250
250

 

 

 

29

TABLE 4.1 SUMMARY OF TESTING PROCEDURE (continued)

Testing
Date

November
12

November
26

December
5

Impact velocity
(feet per minute)

2500
3000
3500
4000

2500
3000

3500
4000

2500
3000
3500
4000

2500
3000
3500
4000

2500
3000
3500
4000

2500
3500

2500
3500

Number of
ears tested

U1U1U1Ul U'IUIU1U'I U1U1UiUI U1U1U1U1 U'IU1U1U1

LJ'lU'!

U!

Variety

Mich.
Mich.
Mich.
Mich.

Mich.
Mich.
Mich.
Mich.

Mich.
Mich.
Mich.
Mich.

Mich.
Mich.
Mich.
Mich.

Mich.
Mich.
Mich .
Mich .

Mich.
Mich.

Mich.
Mich.

500
500
500
500

250
250
250
250

500
500
500
500

250
250
250
250

500
500
500
500

250
250

500
500

 

 

 

30

After an ear was randomly selected, its husk was manually removed
and its weight was recorded. It was then placed in the collection box and
held in position by a small piece of masking tape at each end (Figure 4.2).
At the time of contact the impacting arm was horizontal and the ear was
perpendicular to the arm so the load was applied normal to the middle of
the ear.

The sample was retrieved from the collection box after impact
and the impact force was recorded. The kernels which had been removed
were weighed and the remaining kernels were manually shelled and also weighed.
A portion of the manually shelled kernels from the middle of the ear and a
portion of the cob were oven—dried at 212 degrees Fahrenheit for 72
hours to determine their moisture content. All moisture contents reported
were determined on the wet basis. When there were insufficient kernels
left on the middle of the cob for a moisture sample, equal amounts were
taken from each end of the ear to estimate the moisture level at the middle.
Johnson and Lamp (1966) reported that kernel moisture content is lowest
at the ear tip and highest at the butt. Because of this variability
the middle moisture level would have been represented less accurately
if a moisture sample had been taken from only one end.

The kernels removed by the impact load were inspected and weights

were taken of those which had been broken or cracked and those which had

lost the tip cap (Figure 4.6).

31

 

FIGURE 4.6 TYPICAL FORMS OF MECHANICAL DAMAGE FOUND

IN THIS STUDY. (PHOTO 68—103)

 

 

 

 

5. DISCUSSION OF RESULTS

5.1 General Observations

After impact many of the corn ears tested at 2500 feet per minute
and 3000 feet per minute exhibited a common pattern of kernel removal
(Figure 5.1). Typically a group of kernels three rows wide which were in
contact with the impacting face during loading remained attached to the
cob. Many of the surrounding kernels were detached. Larger ears also
exhibited this pattern of kernel removal, but frequently these cobs were
fractured in the middle. As a result, the majority of the larger ears
exhibiting this pattern were broken in half during impact.

The small (0.39 pound) ear shown in Figure 5.1 is of the Michigan 250
variety. The kernel moisture content was 21 percent and the cob moisture
was 40 percent. An inspection of the exposed cob area following loading
at an impact velocity of 3000 feet per minute, revealed that the caps
of many of the shelled kernels remained attached to the cob. One kernel
at each end of the impact area suffered a broken seed coat.

Most of the Michigan 500 ears and a few of the Michigan 250 ears
impacted at 3500 feet per minute and 4000 feet per minute were broken into
many pieces. The specimen of Michigan 500 shown in Figure 5.2 was loaded
with an impact velocity of 4000 feet per minute; many tip caps remained
attached to the cob. The kernel moisture content was 26 percent and cob
moisture content 47 percent. The portion of ear in the center of Figure

5.2 has some kernels which were forced into the cob during impact.

32

 

 

 

33

fiiifi'mflam. om {ms-“r

«mm m w
«Infill Uncommon" as}.

”a «0mm 11:43:“. m“
“I 299M!» II IW on,

 

FIGURE 5.1 TYPICAL SHELLING PATTERN FROM LOW—VELOCITY IMPACT
AT LOW AND MEDIUM MOISTURE LEVELS.

(PHOTO 68-101)

,I)’|,1;"~‘*.’. Vi 1‘7 " V“
aprQiyywl m 'I fl .
;' D' ' {’fifii'

Bi
‘ “:5. V‘i'.~r";fi

 

FIGURE 5.2 TYPICAL SHELLING PATTERN FROM HIGH-VELOCITY IMPACT
AT HIGH MOISTURE LEVELS.

(PHOTO 68—102)

 

 

 

The kernels shown in Figure 5.3 were shelled from an ear of the
Michigan 250 variety by a single impact at 4000 feet per minute. The pile
of kernels in the foreground at the right were cracked and broken. The
tip caps of the two kernels in the foreground at the left were removed.

No visible damage was suffered by the kernels in the large pile. At
the time of shelling kernel moisture content was 20 percent.

The corn tested at higher moisture levels was more severely damaged.
For example, an ear of the Michigan 500 variety impacted at 4000 feet per
minute had a damage level of over 70 percent. (Figure 5.4). Cob moisture
content was 47 percent and kernel moisture content was 26 percent. The
pile in the foreground consisted of broken or cracked kernels. The tip
caps of all kernels in the left pile were removed. At these high moisture
levels more tip caps were lost during hand shelling than were lost during
shelling by impact loading. Kernels in the right pile were not visibly
damaged. Frequently pedicels of kernels attached at the point where the
cob fractured remained attached to the kernels during the shelling process.

(One such pedicel is shown at the right in Figure 5.4).

 

 

 

 

FIGURE 5.3 KERNELS REMOVED FROM AN EAR OF THE MICHIGAN 250 VARIETY

WITH A SINGLE IMPACT AT 4000 FEET PER MINUTE. UPPER
PILE — SOUND KERNELS. LOWER LEFT — TIP CAPS REMOVED.
LOWER RIGHT - CRACKED AND BROKEN KERNELS. (PHOTO 68—98)

 

FIGURE 5.4 KERNELS REMOVED FROM AN EAR OF THE MICHIGAN 500 VARIETY

WITH A SINGLE IMPACT AT 4000 FEET PER MINUTE. LOWER
PILE - CRACKED AND BROKEN KERNELS. UPPER RIGHT -

SOUND KERNELS. UPPER LEFT — TIP CAPS REMOVED. (PHOTO
68-99)

 

 

 

36

5.2 Regression Analysis Models

Important relationships among the measured parameters were
estimated by means of regression analysis. The ”Multiple Regression
Program" used for this analysis was run on an International Business
Machine Corporation Model 7044 computer at the Computer Center, Univer-
sity of California, Davis. A sub—routine was written for making the
appropriate transformations while reading in the raw data. Estimates
of the regression coefficients were determined by standard matrix inver-
sion tenchiques.

For studying kernel removal the following general second order

statistical model was used:

Y = 80 + 81x1 + ..... + 86x6 + BllXT + ..... + B66X6
+ 812X1X2 + ..... + 856X5X6 + e [5.1]

where:

Y = percent kernels removed by a single impact (observed value)

x1 = cob moisture content (percent w.b.)

x2 = kernel moisture content (percent w.b.)

x3 = weight of ear (pounds)

x4 = impact velocity (feet per minute)

x5 = testing date (day of year)

X6 = impact force (pounds)

E = error of observation

6

i and Bii are regression coefficients.

37

This model can be written as:

A
_ 2 2
Y — bo + lel + ..... + b6xb + blle + ..... + b66X6 +
b12X1X2 + ..... + b56X5x6 [5.2]
where:

A
Y is an estimator of Y
b are estimators of B.

i 1
b.. are estimators of 8..

ij ij

After calculating the values of the bi and bij the percent of kernels
removed by a single impact is expressed as a function containing twenty—
eight terms. The term with the lowest standard partial regression
coefficient, and hence, the last important term, was dropped from the
model and then the data were reprocessed. This procedure was repeated
until all regression coefficients were significantly different from

zero at the 95 percent level. In this manner a less cumbersome function
was obtained and the important coefficients were estimated with greater
precision.

The kernel removal model for the Michigan 250 variety was deter—

mined to be:

A
Y = —95.6 + 0.0809X4 - 9.23 x 10—6Xi — 49.8X3 — 0.340X1 [5.3]
A
By taking the partial derivatives of Y with respect to the Xi’
A

it can be determined how a change in one parameter will affect Y.
In this case an increase in cob moisture content, X1, or ear weight,

X3, will reduce the amount of kernel removal. For the range of velocities

 

 

 

 

38

used in this study an increase in impact velocity is accompanied by an
increase in kernel removal. The kernel moisture content, impact force
and testing date did not significantly affect the kernel removal.

The kernel removal model for the Michigan 500 variety is as follows:

A
Y = -78.5 — 1.77X1X + 2.76X2X + 1.85X — 0.0247X X +

3 3 l 1 2

—3
4.93 x 10 XAX5 [5.4]

Determination of the consequences of parameter changes in this
model is not as straight forward because of the cross—product terms. The
rate of change in kernel removal due to a unit change in cob moisture
content is a function of kernel moisture content and ear weight. In this
and similar cases the partial derivatives were evaluated by substitution

of the mean values for the included parameters. As an example:

A
33— : 185—177X — 0247x — 106
ax . C 3 O 2- I
1
for: X3 = .566
x2 = 30.8

An increase in kernel moisture, cob moisture, impact velocity, or
testing date was accompanied by an increase in kernel removal. A decrease
in ear weight reduced the kernel removal while impact force had no
significant effect.

According to Halyk (1968) shelling can be initiated more easily
when ears have small spaces between rows of kernels than when ears have

large spaces, During the testing it was noted that the small ears of

 

   

 

 

39

the Michigan 250 variety were more tightly packed with smaller spaces
between kernel rows than were the large ears. Hence, it was not surprising
that the regression analysis showed that for Michigan 250 the smaller

ears shelled more easily than the larger ears.

For the range of impact velocities studied, both varieties shelled
more easily with increased velocity. This result agrees with Morrison
(1955) that combine—type cylinders shell corn more effectively with increased
peripheral velocities.

Burrough and Harbage (1958) reported shelling became more effective
as the harvesting season progressed. Testing date had the same influence
on the Michigan 500 variety, but had no significant effect on the Michigan
250 variety.

According to Johnson and Lamp (1966) shelling effectiveness increases
as the kernel moisture (or cob moisture) decreases. For the Michigan 250
variety the cob moisture content had the same effect, but the kernel
moisture content had no significant effect. Both kernel moisture and
cob moisture had opposite effects on kernel removal for the Michigan
500 variety. An increase in either moisture level significantly increased
the kernel removal. The difference appears to stem from the fact that
the moisture levels were much higher for the Michigan 500 variety than
for Michigan 250. Many seed coats of the Michigan 500 kernels were broken
leaving tip caps attached to the cob, whereas the impact affected very
few of the Michigan 250 kernels in this manner. Even though the kernel
removal improved with increased moisture levels, the mechanical damage

increased to levels which would make field shelling impractical.

 

 

 

 

40

The regression analysis for total mechanical damage (kernels
visibly cracked and broken plus kernels which lost the tip caps) utilized
the same dependent variables, Xi’ as used previously for kernel removal.

The resulting damage model for the Michigan 250 variety is as follows:

A
_ 2 -4
Y — 3.38 + .251X1X2 — .0672Xl - .00193X2X4 + 1.14 x 10 X4X5
— 0.0112X2X5 [5.6]
where:
A
Y = total mechanical damage (expressed as a percent of the kernels

removed by a single impact)

Total damage is increased by (1) decreases in impact velocity, X4,

or kernel moisture level, X2; or (2) increases in cob moisture level , X1,
or testing date, X

Ear weight, X , and impact force, X6, had no sign—

5.

ificant effect on total damage.

3

The subsequent damage model was developed for the Michigan 500
variety:

A 2

Y = 3680 - 20.9X5 + .0285X5 — 29.7X2 + .lO7X2X5 [5.7]

Cob moisture level, X1, ear weight, X3, impact velocity, X4,

and impact force, X had no significant influence on total damage. Increases

6’
in kernel moisture content, X2, are accompanied by increases in damage.
The rate of change in damage with respect to a unit change in testing
date, X5, reaches a minimum at November 4. Before this date damage

decreases with an increase in testing date, but after this date the opposite

is true.

 

 

 

41

That Michigan 250 suffered less damage at high impact velocities
than at low velocities seemed strange initially. However, after careful
consideration of the available information, this result does not appear
so unusual. Since the collection box was lined with foam rubber, only
the kernels which come in contact with the impacting face can be broken
or cracked, the most prevalent form of mechanical damage for this variety.
Since kernel removal was significantly improved by increases in impact
velocity, increasing the impact velocity reduced the percentage of shelled
kernels which could be damaged by coming in contact with the impacting
face.

Earlier it was shown that kernel removal for the Michigan 250
variety was hampered by increases in cob moisture levels. These shelling
difficulties due to cob moisture increases were accompanied by greater
incidence of damage.

Kernel moisture level had an opposite effect on damage for the two
varieties. Using the data from these tests there is no way to deter—
mine whether this discrepancy was due to the differences in average moisture
level or varietal differences. The damage level increased as the harvest
season progressed for Michigan 250 and after November 4 for Michigan 500.

For studying the effect of several parameters on impact force the

following statistical model was used:

A
2 2
Y — bO + blxl + ..... + bSX5 + blle + ..... + b55x5 + b12XlX2 +

.. + b45x4x5 [5.8]

 

 

 

 

42

where:
A
Y = estimator of Y
Y = maximum impact force (pounds)
X1 = cob moisture content (percent w.b.)
X2 = kernel moisture content (percent w.b.)
X3 = weight of ear (pounds)
X4 = impact velocity (feet per minute)
X5 = testing date (day of year)
b1 and bij are estimators of the regression coefficients

After elimination of unimportant terms using the system described

above,the impact force model for Michigan 250 was found to be:

A
_ 2 2
Y — 4000 437x2+ .826X2X5 + 4.25x2 — .0232):5 +
-5 2
5.22x10 x4 +855x3 [5.9]

Cob moisture, X1, was the only parameter which had no signifi—
cant influence on impact force. Increases in ear weight, X3, impact

velocity, X4, or testing date, X , caused an increase in impact force.

5
The rate of change of impact force with respect to a unit change in

kernel moisture, X2, reached a minimum within the range of moisture levels
studied. For example, on day 317 the force decreased with an increase

in moisture level up to 20.6 percent. Above that level increases in
moisture content caused an increase in impact force. For days 289, 303

and 331 the critical moisture levels were 23.3, 22.0 and 19.2 percent,

respectively.

 

   

 

43

The impact force model determined for Michigan 500 was:

A
Y = 366 - 489OX3 + 15.8X3X5 + 45.6Xl —.l73XlXS +

—5
27.7X3X2 + 4.53 x 10 X4 [5.10]

Impact force increased with: (l) a decrease in cob moisture, X1,
or testing date, X5, or (2) an increase in kernel moisture, X2, ear
weight, X3, or impact velocity, X4.
Before the impacting face and corn ear collide the ear is at
rest. For a perfectly plastic collision the final velocity of the
specimen would be equal to the impacting velocity, and for a perfectly
elastic collision the final velocity would be twice the impacting
velocity. Since the collision is neither perfectly plastic nor perfectly
elastic, the final velocity of the ear must be a constant multiple of the
impacting velocity and that constant must be a real number greater than
one and less than two. Because of this relationship, increasing the
impacting velocity will increase the final velocity of the ear. This,
in turn, will increase the acceleration the ear undergoes.
According to Newton's Second Law of Motion an external force,

>
F, is required to change the velocity of a body of mass, m, with an

>
acceleration, a, where:

>

F = ma [5.11]
The regression for both varieties showed that an increase in ear weight
or impacting velocity increased the maximum impact force.

The testing date had opposite effects on impact force for the
two varieties. Increased force followed increased testing date for Michigan

250 and decreased testing date for Michigan 500. The testing date was

 

 

 

 

44

more highly correlated with kernel moisture and ear weight for Michigan
500 than for Michigan 250. Interaction of the parameters could have caused
the discrepancy in results.

For the investigator accustomed only to quasistatic loading,
the effect of kernel moisture level on impact force may seem contrary
to nature: force should be greater for a dry ear than for a wet ear.
However, since corn is a viscoelastic material, force becomes a function
of displacement rate as well as displacement. Increasing the moisture
level increases the dependence of the force upon loading rate because
the material becomes more viscous and less elastic under stress.
Thus, the indicated influence of moisture level on impact force and
the influence of impact velocity on force can be substantiated.

Earlier it was indicated that a decrease in cob moisture caused
an increase in impact force for Michigan 500; the opposite of the effect
of changes in kernel moisture. Kernel and cob moisture were positively
correlated; however, a one percent change in cob moisture resulted in
a —7.9 percent change in impact force while a one percent change in kernel
moisture resulted in a 15.7 percent change in force. Since kernel moisture
had more influence than cob moisture, the net result is that an increase
in ear moisture caused an increase in impact force for Michigan 500. Cob

moisture had no significant influence on impact force for Michigan 250.

 

   

 

6. SOME THEORETICAL ASPECTS OF CORN IMPACT

6.1 Extension of Corn Shelling Theory to Dynamic Range

Prior to this study the investigations attempting to explain
corn shelling have utilized quasistatic and low—velocity impact
loading. No researcher has reported the use of impact loading at
velocities comparable to the cylinder peripheral velocities commonly
found during field shelling.

Halyk SE 31. (1969) reported that a kernel with moisture level
below 15.3 percent subjected to compressive loading acted as a wedge and
caused shelling of one of the surrounding kernels. By using a vertical
weight-drop machine they found this wedging effect to hold for kernel
moisture levels of 28 percent or lower.

During this study kernel moisture levels as high as 43.6 percent
were encountered. Radial compressive impact loading at 2500, 3000, 3500
and 4000 feet initiated shelling on all of the ears tested regardless
of moisture content. The amount of kernels removed by a single impact
was found to depend upon moisture content, ear weight, impact velocity
and testing date. The nature of this dependence is discussed in Section
5.2.

The pedicels of kernels on the top half of the corn ear and on the
outside of the contact area were subjected to a tensile force during impact
(Figure 4.2). This tensile force was the inertial force due to the

cob's accelerating downward away from the top kernels.

45

 

 

 

46

When the pedicel of one kernel fractures to initiate shelling,
forces applied to that kernel must be great enough to exceed the maximum
tensile strength of the pedicel while overcoming the frictional forces
between kernels. Thus far, two mechanisms for applying tensile force
to the pedicel during impact have been described. The wedging effect
originally reported by Halyk (1968) causes tensile stress in pedicels
of kernels adjacent to the loaded area because of tangential displacement
of the adjacent kernels. Pedicles of the top kernels were also subjected
to a tensile force because of the kernel's resistance to downward
acceleration. The tensile forces resulting from these two actions is
greatest near the loaded area of the ear. It was observed that the first
kernels removed were always adjacent to the loaded area. Shelling is
initiated during radially inward impact loading because of the inertial
forces on the shelled kernels and the wedging effect.

The reflected tensile stress wave discussed in Appendix III
could also be a factor in the initiation of shelling; however, the stress
wave propagation route and velocity through the ear must first be

determined.

 

 

 

47

6.2 Analysis Utilizing High—Speed Movies

The result of impacting five corn ears tested at 2500 feet per
minute showed the cob of each specimen broken into two pieces; and approx—
imately 50 kernels shelled with the remaining kernels still attached
to the cob halves. The high-speed cinematography techniques used to
record these tests are described in Appendix I. The movies were analyzed
with Vanguard Instrument Corporation Model M—16C motion analyzer to
determine the kinetic energy of the specimens.

The movies were projected on the screen of the motion analyzer
one frame at a time. The coordinates of the center of mass of every
particle were recorded for each frame. Since the ear halves were rotating
after impact, the coordinates of two points on each half were also re—
corded to measure angular displacement. Only the kernels shelled from
the camera side of the ear were visible, so it was assumed there was
one kernel not in view for each visible kernel. Consequently, the kinetic
energy determined for the visible particles was doubled to compensate
for the kernels not in view.

A Hewlett—Packard Model 9100A calculator was programmed to
calculate the linear and angular velocities of the particles. The
coordinates measured on the analyzer screen were adjusted to reflect the
true displacements. Between frames number one and two the average linear
velocity, v, of the center of mass of a given particle can be expressed as:

2 2 1/2

v = [ (xl - XI) + (yl - yl’)] / T [6.4]

 

   

 

48

where:

(x ) are the adjusted coordinates from frame number one.

1’ V1
(xi, yi) are the adjusted coordinates from frame number two.

T is elapsed time between frames one and two.

The data need not be adjusted for determining the angular

velocity: if (x1, yl) and (x2, y2) are coordinates of two points on a
rigid body in frame one and (xi, yi) and (xi, yé) are coordinates of the

same points in frame two, then the average angular velocity, w, of the

body between frames one and two can be written as:

w = O / T [6.5]
where:
O = arc cos [(2d2 — c2) / 2d2]
2 , , 2 , , 2
c — (y2 - yl — y2 + yl) + (x2 - xl - x2 + x1)
d2 = 2 2

(y2 - yl) + (x2 - X1)
T is elapsed time between frames one and two.
To determine the kinetic energy of the ear halves after impact
the linear and angular velocities were calculated over a group of fifteen
successive frames. The average of these values was used to calculate
the linear and rotational kinetic energies. Only linear kinetic energy

was determined for the shelled kernels. The kinetic energy of all the

particles was totaled to estimate the total kinetic energy of the ear.

 

 

49

6.3 Energy Considerations

When two bodies collide at a high relative velocity, some of
the original kinetic energy is transformed into heat energy during the
collision and absorbed by the bodies. If the total kinetic energy lost
by both bodies during collision can be determined, this value can be used
to estimate the energy absorbed by the bodies. However, in the case
of this study this type of energy balance could not be run to determine
the energy absorbed by the corn during impact. The impact tester was
powered by an electric motor to prevent a loss of velocity of the impacting
arm while in contact with the arm.

Haliday and Resnick (1963) discussed the topic of impact and
some of their results were used to estimate the energy absorbed by the

corn during impact. If a body of mass m traveling at a velocity u

l

and a body of mass m2 traveling at a velocity u

l

2 impact with a perfectly

elastic collision, the velocities after the collision can be written as:

m — m 2m
1 2 2
V = ( ) u + (———————) u [6.6]
1 ml + m2 1 ml + m2 2
2m m — m
l 2 l
v = (—-————-) u + (———————) u [6.7]
2 ml + m2 1 ml + m2 2

If m2 is considered to be negligible in comparison to ml, and 1112

is initially at rest, equations 6.6 and 6.7 reduce to:

v u [6-8]

1 1

V2 = 2u2 [6.9]

 

 

 

 

50

If the corn ear is considered to be m2, this precise situation
existed during the impact. For a perfectly elastic collision at 2500 feet
per minute the ear should have moved downward with a velocity of 5000
feet per minute after impact. A corn ear traveling at this velocity
and weighing 0.468 pound would have a kinetic energy of 50.6 ft—lbs.

High-speed movies were analyzed for five corn ears with average
kernel and cob moisture contents of 20.5 percent and 27.9 percent,
respectively, and weighing an average of 0.468 pound. The kinetic energy
measured for these ears averaged 17.2 ft—lbs after impact. The observed
kinetic energy differed from the theoretical value because the collision
was inelastic and not perfectly elastic. The discrepancy of 33.4 ft—lbs
from the theoretical value of 50.6 ft—lbs is considered to be the energy
dissipated during impact. Negligible amounts of energy were absorbed
by the supporting tape and impacting machine. An insignificant portion
of energy was also used to form sound waves. Hence, the energy discre—

pancy of 33.4 ft-lbs is an estimate of the energy absorbed by the ear during

impact.

 

   

 

 

7. CONCLUSIONS

7.1 Findings of This Study

Whole corn ears were tested to investigate the initiation of
corn shelling by radial compressive impact loading. It was found that
radial compressive impact loading would initiate shelling on corn ears
with kernel moisture levels as high as 43.6 percent and impact velocities
as low as 2500 feet per minute.

Halyk (1968) reported that a kernel displaced radially inward
by quasistatic or low velocity impact loading acted as a wedge and shelled

one of the surrounding kernels. During this study this wedging effect

was also shown to be valid for impact velocities ranging from 2500 to 4000
feet per minute and is illustrated in Figure 4.3. Under the loading

configuration utilized by Halyk (1968) the acceleration of the cob was
insignificant and was disregarded. However, in this study the cob was
accelerated downward away from top kernels not in contact with the
impacter (Figure 4.2). Consequently, the shelling observed in this study
was initiated by forces resulting from the wedging effect and the downward
acceleration of the cob.

High—speed movies were utilized to determine the kinetic energy
of five corn ears with average kernel and cob moisture contents of 20.5
percent and 27.9 percent, respectively, and weighing an average of .468
pound. This information was used to estimate that these ears absorbed

an average of 33.4 ft-lbs of energy during impact.

51

 

 

 

52

The maximum impact force was measured for every sample tested.
A regression analysis showed impact force was dependent on kernel moisture
content, ear weight, impact velocity and testing date for both varieties.
Cob moisture content had a significant effect on impact force for only
the Michigan 500 variety.

The percentage of kernels removed by a single impact was shown
by regression analysis to depend on cob moisture content, ear weight and
impact velocity. Kernel moisture content and testing date significantly
affected kernel removal for the Michigan 500 variety.

Regression analysis showed mechanical damage of both varieties was
influenced by kernel moisture content and testing date. Cob moisture
content and testing date had a significant effect on mechanical damage for

only the Michigan 500 variety.

7.2 Recommendations for Future Studies

Several ideas which could be the subject of other investigations
have arisen during the preparation of this thesis. The effect of
storage conditions and time between collection and testing of samples
should be studied for important grains. No researcher has reported an
investigation of the dependence of feeding value of corn upon mechanical
damage and storage time. As stated in Appendix III determination of the
stress wave propagation velocity and route could reveal the role of reflected
tensile waves in corn shelling.

A study in cooperation with the Crop Science Department could

utilize the impact testing equipment to evaluate the resistance of corn

 

   

 

 

to mechanical damage and the ease with which they can be

 

shelled.

program.

It appears this could be a valuable part of a corn breeding

 

 

   

 

Agness, J. B.
1968

Anonymous

1968

Anonymous
1969

Aspenwall, C.
1924

Bailey, J.
1968

Barkstrom, R.
1955

54

REFERENCES

Measuring mechanical damage to corn kernels. Paper No.
68-620 presented at the 1968 Winter Meeting, American
Society of Agricultural Engineers, Chicago.

A new combine. Implement and Tractor 83(6): 36—37.

What's ahead. Implement and Tractor 84(7): 18.

0.
Economics of the corn picker—husker. Agricultural Engi-
neering 5(9): 199, 202, 208.

Problems in marketing damaged grain and corn. Proceedings
of a Symposium on Grain Damage Held Under the Auspices

of the American Society of Agricultural Engineers, Iowa
State University, Ames.

Attachments for combining corn. (International Harvester
Co.). Agricultural Engineering 36(12): 799-800.

Birth, G. S. and R. M. Johnson

1968

Burkhardt, T.
1969

Burkhardt, T.
1969

Detecting mold contamination in corn by optical measure—

ment. Paper No. 68—566 presented at the 1968 Winter
Meeting, American Society of Agricultural Engineers, Chicago.

H.
A high—velocity, high—momentum impact testing device for
agricultural materials. A technical problem report for the

degree of M.S., Michigan State University, East Lansing.
(Unpublished).

H. and B. A. Stout

A high—velocity, high—momentum impact testing device for
agricultural materials. Paper No. 69—344 presented at the
1969 Summer Meeting, American Society of Agricultural
Engineers, Purdue University, Lafayette.

Burrough, D. E. and R. P. Harbage

1953

Performance of a corn picker—sheller. Agricultural Engi—
neering 34(1): 21—22.

 

   

 

 

55

Byg, D. M.
1968 How do you measure grain damage? Proceedings of a

Symposium on Grain Damage Held Under the Auspices of the
American Society of Agricultural Engineers, Iowa State
University, Ames.

Byg, D. M. and G. E. Hall
1968 Corn losses and kernel damage in field shelling of corn.
Transactions of the ASAE 11(2): 164—166.

Cooke, J. R. and J. W. Dickens
1967 A centrifugal gun for impaction testing of seeds. Paper
No. 67—330 presented at the 1967 Summer Meeting, American
Society of Agricultural Engineers, Saskatoon, Saskatchewan.

Cooper, G. F.
1969 Supervisor, Harvesting Systems Lab, Massey—Ferguson
Industries Limited, Toronto, Canada. Personal Corres—
pondence, December 15.

Cross, H.
1952 Engineers and Ivory Towers. McGraw-Hill Book Company,
Inc., New York. 141 pp.

Ewing, W. M., W. S. Jardetzky and F. Press
1957 Elastic waves in Layered Media. McGraw—Hill Book Company,
Inc., New York. 380 pp.

Forth, M. and M. Zimmerman
1968 Don't blame it all on the combine. Implement and Tractor
83 (10): 22-24, 26.

Garrett, R. E.

1970 Velocity of propagation of mechanical disturbances in
apples. Thesis for the degree of Ph.D., Cornell Univer—

sity, Ithaca. (Unpublished).

Goldsmith, W.

1960 Impact — The Theory and Physical Behavior of Colliding
Solids. Edward Arnold (Publishers) ltd., London. 379 pp.

Goss, J. R., R. Bainer and R. G. Curley
1955 Field tests of combines in corn. Agricultural Engineering

36(12): 794-796.

Hall, G. E. and W. H. Johnson
1968 Corn kernel crackage induced by mechanical shelling.
Paper No. 68-112 presented at the 1968 Summer Meeting,
American Society of Agricultural Engineers, Logan.

 

   

 

 

56

Halliday, D. and R. Resnick
1963 Physics for Students of Science and Engineering. John
Wiley and Sons, Inc., New York. 1177 pp.

Halyk, R. M.
1968 A theory of corn shelling based on a quasistatic analysis.
Thesis for the degree of Ph. D., Michigan State University,
East Lansing. (Unpublished)

Halyk, R. M., B. A. Stout and E. R. Norris
1969 A theory of corn shelling based on quasistatic and low—
speed impact loading. Paper No. 69—342 presented at the
1969 Summer Meeting, American Society of Agricultural Engi—
neering, Purdue University, Lafayette.

Hurlbut, L. W.
1955 More efficient corn harvesting. Agricultural Engineering
36(12): 791—792.

Johnson, W. H. and B. J. Lamp
1966 Principles, Equipment and Systems for Corn Harvesting.
Agricultural Consulting Associates, Inc., Wooster. 370 pp.

Johnson, W. H., B. J. Lamp, Jr., J. E. Henry and G. E. Hall

1963 Harvesting performance at various dates. Transactions of
the ASAE 6(3): 268-272.

Johnson, W. H., M. L. Jain, M. Y. Hamdy and P. F. Graham
1969 Characteristics and analysis of corn ear failure. Trans—
actions of the ASAE 12(6): 845—838, 852.

Kaminski, T. L.

1968 Need for standards for evaluation of grain damage.
Proceedings of a Symposium on Grain Damage Held Under
the Auspices of the American Society of Agricultural
Engineers, Iowa State University, Ames.

Kolsky, H.
1963 Stress Waves in Solids. Dover Publications, Inc., New
York. 213 pp.

Liebenow, R. C.

1968 Corn in the supermarket. Corn 24(1): 1—7.
Logan, C. A.
1931 The development of a corn combine. Agricultural Engineering

12(7): 277,278.

 

   

 

57

Maywald, F.

1968 Do grain grading systems meet grain marketing needs?
Proceedings of a Symposium on Grain Damage Held Under
the Auspices of the American Society of Agricultural
Engineers, Iowa State University, Ames.

McKibben, E. G.
1929 Harvesting corn with a combine. Agricultural Engineering
10(7): 231, 232.

Michigan Department of Agriculture
1968 Michigan Agricultural Statistics. Michigan Crop Reporting
Service, Marketing Division, Lansing, Michigan. 38 pp.

Miles, S. R.
1956 Maturity of corn in relation to field shelling. Pro—

ceedings of Conference on Field Shelling and Drying of
Corn. United States Department of Agriculture, Chicago,
Illinois.

Mitchell, N.
1956 Heated air drying of shelled corn-batch and continuous driers.
Proceeding of Conference on Field Shelling and Drying of
Corn. United States Department of Agriculture, Chicago, Ill—
inois.

Morrison, C. S.

1955 Attachments for combining corn (Deere and Company).
Agricultural Engineering 36(12): 796—799.

New Holland
1969 Corn combines. Advertising brochure by New Holland
Division of Sperry Rand Corporation. 6 pp.

Pickard, G. E.
1955 Laboratory studies of corn combining. Agricultural Engin—
eering 36(12): 792—794.

Pickard, G. E. and H. P. Bateman
1954 Combining corn. Agricultural Engineering 36(7): 500,504.

Richey, C. B., J. F. O-Donnell, J. T. Ashton and R. J. Groves

1956 Corn picker features new principle. Agricultural Engin—
eering 37(2): 93—97.

Saul, R. A. and J. L. Steele
1966 Why damaged shelled corn costs more to dry. Agricultural
Engineering 47(6): 326—329, 337.

 

 

 

58

Scranton, C. J.
1955 Attachments for combining corn (Allis—Chalmers Mfg. Co.).
Agricultural Engineering 36(12): 800,802.

Seferovich, G. H.
1961 Symbols of modern mechanization. Implement and Tractor

76(4): 31.
Seferovich, G. H.
1969 Trajectory of a technology. Implement and Tractor
84(1): 28—30.

Skelton, R. F. and H. P. Bateman

1942 Field shelling of corn. Agricultural Engineering 23(4):
131-133.
Uhrig, J. W.
1968 Economic losses of damaged grain. Proceedings of a

Symposium on Grain Damage Held Under the Auspices of the
American Society of Agricultural Engineers, Iowa State
University, Ames.

United States Department of Agriculture
1967 Agricultural Statistics. U. S. Government Printing Office,

Washington, D. C.

Waelti, H. and W. F. Bushele
1967 Factors affecting corn kernel damage in combine cylinder.
Paper No. 67-631 presented at the 1967 Winter Meeting,
American Society of Agricultural Engineers, Detroit.

Wolf, M. J., C. L. Buzan, M. M. MacMasters and C. E. Rist
1952 Structure of the mature corn kernel. Cereal Chemistry 29:
321—382.

Mmmnmn,M.
1968 Which machine is best for field shelling? Implement and

Tractor 83(11): 24—27.

 

 

 

 

 

 

APPENDICES

59

 

 

 

 

 

 

APPENDIX I

HIGH—SPEED CINEMATOGRAPHY

60

 

 

 

 

APPENDIX I

HIGH—SPEED CINEMATOGRAPHY

The high~speed movies utilized in this study were obtained with
a continuous—type 16mm Fastax camera using Kodak Type 7278 TRI—X
reversal film in a lOO—foot roll. A Wollensak-Fastax Goose Control Unit
synchronized the actions of the camera and the testing apparatus.

The large amount of preparation time before filming and the high
cost of purchasing and processing the film inspired the recording of
useful information on each roll of film. To achieve this the impact had
to occur after the film was accelerated to a desired framing rate and
before the end of the film passed through the camera. For proper
coordination of the equipment, it was necessary to obtain the following
information:

1. Elapsed time between activation of start switch and initial

contact between impacting face and specimen;

2. Time required for impact;

3. Time required for film to reach desired velocity;

4. Time required to run 100 feet of film through camera;

5. Time available for filming impact process.

By measuring the appropriate distance on the trace of a storage
oscilloscope, it was possible to determine the time which elapsed between
activation of the start switch and initial contact between the impacting

face and the specimen. For an impact velocity of 2500 feet per minute,

 

   

 

 

62

the elapsed time was 0.42 second. Using the same method it was found that
the time required for impact was much less than one millisecond. This

was not a critical parameter since it was so much smaller than the

other parameters being considered.

The rate of film passing through the camera was controlled by the
magnitude of the voltage supplied to the camera motors by the Goose
Control Unit. The camera manufacturer provided curves from which the per—
tinent data was estimated. For example, on the first filming, the voltage
setting was adjusted to 40 volts. From the curves it was estimated 3.3
seconds would be required to accelerate the film to the desired velocity
and 5.0 seconds would be required to run the 100 feet of film through the
camera.

During the filming the manual start switch on the impact appara-
tus was replaced by a normally open contact switch on the control unit.
The "four position" switch was set at position number three. The event
timer was set at 2.9 seconds and the camera timer was set at 5.0 seconds.
Under this adjustment the impact apparatus was automatically started
2.9 seconds after the camera was started,and the camera was stopped
after running 5.0 seconds. This setting provided an average framing
rate of 1860 frames per Second during the impact process.

The timing for stopping the camera was quite important. Premature
stopping would have caused the camera to masticate film that was not on
the take-up reel. Late stopping would have caused the take-up reel to

rotate at a high speed with the end of the film striking the interior of

the camera and breaking off small pieces of film.

 

 

 

63

The camera was located four feet from the specimen. A 50mm lens
was used with a f/4.0 aperture setting. The impact velocity was 2500
feet per minute. The first attempt at high—speed cinematography proved
successful.

For another filming on the same day, an average framing rate of
4100 frames per second was achieved by using the following adjustments:

1. Voltage to camera — 90 volts.

2. Event timer — 0.93 second.

3. Camera timer - 2.0 seconds.

4. Aperture setting - f/2.7.

The camera was then moved to six feet from the specimen to increase
the field of view. An average framing rate of 6470 frames per second
was achieved using the following adjustments:

1. Voltage to camera - 200 volts.

2. Event timer - 0.30 second.

3. Camera timer — 1.0 second.

4. Aperture setting — f/2.7.

5. Lens - 50 mm.

An average framing rate of 6830 frames per second was achieved
by changing the event timer to 0.38 second. In both cases the light meter
reading was 1800 foot—candles. This amount of light was barely adequate,
and an increased amount would have improved the quality of the movies.
For lighting this intense it is very easy to burn a biological specimen
or dry it excessively. Consequently, the lamps cannot be lighted for
a period longer than necessary.

The light colored corn ears showed up very well against the dark

background. The impacting face should have been painted a lighter color

 

 

 

 

64

because it did not show up very well.

For any reader who is contemplating the use of high-speed

cinematography the following information should prove useful:

1.

There should be at least one object of known size in the
field of view to use as a reference.

There should be a method for determining the framing rate.
a. A light flashing at a known rate.

b. An object moving at a known constant velocity.
c. A timer in the field of view.

There should be one point which is stationary throughout the
filming sequence to use for removing the error of camera
movement.

A number should be placed in the field of view for later
identification.

The objects to be studied should be a light color against
a dark background.

Read the directions carefully.

 

 

 

 

 

APPENDIX II

 

STATISTICAL METHODS

65

 

 

 

 

APPENDIX II

STATISTICAL METHODS

The data collected in this study were used to estimate the co—
efficients of six different regression models. In Section 5.2 effects
of changes in the dependent variables on the independent variable for
each model were discussed. However, no mention was made of the t-tests
and analysis of variance tests carried out during the regression analyses.

For the sake of illustration consider the following regression

model:

A 2

Y = b4X4 + b44X4 + b3X3 + lel + be
where:

A

Y = percent of kernels removed by a single impact

X4 = impact velocity (feet per minute)

X3 = ear weight (pounds)

X1 = cob moisture content (percent w.b.)

b4, b44, b3, bl’ bO are estimates of the regression coeffic1ents,
84' 844’ 83' 81’ 80'

An analysis of variance was carried out to test the hypothesis that

A

the dependent variable, Y, is unrelated to the independent variables, X4,
X3, X1, in a linear or quadratic fashion (HO: 84 = 844 = 83 = 81 = O).

This analysis is summarized in Table A2.1. The calculated F value was
significant at the 95 percent level. Thus, it was possible to conclude
with at least 95 percent confidence that at least one of the regression

coefficients was different from zero.

66

 

 

 

67

To determine which of the regression coefficients were different
from zero t—tests were utilized. For example, it was hypothesized that
the independent variable, X1, was unrelated to the dependent variable

A
Y (H1: 8 = 0). The calculated t value was significant at the 95 percent

1
level. Thus it was possible to conclude with at least 95 percent con—
fidence that 81 was different from zero. The same procedure was used

to determine that 83, B and 844 were also different from zero.

4
The regression parameters for the six models are summarized in

Tables A2.l through A2.6. All of the t and F values listed are signifi—

cant at the 95 percent level. The standard partial regression coefficients

were calculated for all of the parameters of every model. For a given

model the most important parameter had the partial regression coefficient

with the largest absolute value. The next most important parameter

had the second largest partial regression coefficient. The parameters

were listed in the tables in descending order of importance as indicated

by the absolute values of the partial regression coefficients. The

ranges of parameters for which the models are valid are listed in Table

A2.7.

 

 

 

68

l,

 

 

 

 

I I ss.~wm.mm Nae sauce
I me.moe Hq.oaw.me wNH Hours
m.om Hm.mmq.m mo.Nsa.HN a coammmrmmm
m .m.z .m.m .w.w condom
scammoumou ou one oosmfium> wo mfimuams<
Hm.NI I o.mml usmumsoo
so.MI mH~.I osm.I ex
. . . m
Ha «I new I w as- x
om.NI wo.NI oIos x mm.mI we
am.m ow.~ some. as
usowoflmmooo seammwummm ucmfloflmwooo
osaw>lH Hmwuumm mumvsmum soawmouwwm poumEmumm
case. I coeumanenmums no remnanmmmoo omN essences I sumsnm>

uommefl camsfim m an mw>osou waosuox mo unmouom n w

 

 

mMMHmz¢m<m mHmwA<z¢ ZOHmmewmm mo wm<zzbm H.N< mqm<a

 

 

 

 

69

 

I I mm.qmw.qa mmH HmuoH

 

I No.0m mN.HHs.q mNH House
0.0m Ha.mso.N Ho.HHN.OH m aOHmmemmm
a .m.z .m.m .m.s mouaom

scammoummn on map oufimaum> wo mammams<

 

 

 

 

qw.ml I m.wml udMumcoo
. . . m a
a 0H me mIOH x mm a x x
mH.MI OH.HI asNo.I NxHx
. . . H
Hm m os H mm H x
Na.m ms.H aa.N mxwx
os.sI Hm.HI HH.HI mxHx
usmflowwwooo GOHmmouwom usoflUfiwwooo
osam>IH HMfluumm pummsmum cowmmwummm umumamumm
ammo. I cOHHmaHaHmums Ho uaoHoHHHmoo com smmHHon I aHmHHm>

uomafifi meaflm m an mo>ofimn mamsuwx wo usoouom w

 

 

mMMHME<M<m mHmwA<z< ZOHmmmMUmM mo MM¢ZZDw N.N< mam<H

 

 

 

 

 

70

I

 

I I mm.oma.mm NMH Hmuoe
I Ho.amN om.mHa.mN HNH Hornm
ms.m mN.smm ms.HHN.s m :OHmmmHmmm
m .m.z .m.m .m.p wousom

soawwouwou ou mam ousmfium> mo mfiwxamsm

 

 

 

oHN. I mm.m newsmaoo
so.NI mHs.I NHHo.I mama
mH.m mN.H GIOH x «H.H qux
sm.m- mw.HI mmHoo.I axmx
Ha.NI H¢.NI Naso.I Ma
wH.N Hm.q Hmm. NxHx
usmflofimmwoo scammmummm uamfloflwmooo
oDHm>IH Amwuumm pummsmum cosmwwuwmm Housemumm
amNH. I coHumcHapmums Ho HamHoHHHmoo omm smeson I aHmHHm>

Avm>oeon wHosumx mo pawouoav owmfimp fireflsmnomz u w

 

 

mMWBmz¢M<m mHmwA<z< ZOHmmmmwmm mo wm<zzbm m.~¢ mam<fi

 

 

 

71

 

 

I I ma.maw.mo mMH HmHoe
I ma.moN Ho.mmm.qm mNH Houum
o.¢N Hm.omw.a mo.NNm.Hm a aOHmmmeom
a .m.z .m.m .H.s 00H30m

soamwmuwou on one mosMfiHm> mo mammamcm

 

 

 

 

no.4 I .omsm Hawumcoo
mo.s mw.m HOH. mxmx
so.m- ma.sI H.aNI Nx
am.m N.MH memo. We
so.aI «.mHI a.omI ma
usmflofimmwoo sowmmmuwom unofiofiwwooo
o=Hm>IH Hmwuumm mummsmum cowmmwumom Hmuoamumm
mass. I EOHumaHanmuma Ho HamHUHHHwoo oom pmeHon I sumHHm>

Amw>oaou mascuox mo unmouomv ommamm Hmoflsmaomz u N

 

 

mMMHmz<m<m mHmWA¢z< onmmmmomm mo WM<ZZDm q.N¢ MAM<H

 

 

 

 

 

i-

72

 

I I H.oom.wao.a NNH Hmuoe

. |.I|..II|IN| III
I a.oom ON N mHm 0N0 N sNH Hours
m.Hm H.omN.HHo.H m.omm.st.s a aOHmmmemm
a .m.: .m.m .H.s mousom

sOHmwwuwwu on use mosmflum> mo wwwhams<

 

 

 

 

ms.m I .ooos samumaoo
No.m Nom. .mmm mx
a

. . x .

o mH NNH mIOH NN N Na
. . . N
am NI mHm I NmNo I Nx
. . . N
an N we N mN a Na
mm.N Ha.N GNN. mex
wa.mI Ho.mI .NNNI Nx

unmaowwmooo seammouwwm unoHOHmwooo
wSHm>IH Hmfluumm pudendum seawmouwwm nouoamumm
NOHH. I EOHumnHanuuma No HamHUHNNmoo omN cmeHUHz I suwHHm>

Ammcsoav ooH0w uummafi Esawxmz n w

 

 

mmmamz¢m<m mHmVA<z< ZOmemmomm mo NM<ZZDm m.N< mam<H

 

   

 

 

73

 

 

I I w.mmm.mHs.NH mmH Hmuoe
IIIIIIIIIIqI IIII

I H.HsN.mm N.NmH.sNa a NNH Horne

«.mm o.omm.amN.H o.oma.omo.m o coawmwuwmm

e .w.z .w.m .H.a muusom

scammmuwou ou mam mommaum> we mammams<

"l'Ill'lll'I‘llli'lll

oo.N I .oom ucmumsoo
Hm.a ass. OH x mm.a ax

mI N
as N mom N NN Nxmx
NN NI mH HI MNH I mxHx

. . . H
00 N wH H a me x
HN.N oq.N w.mH mama
NH.NI mm.NI .oamsI ma
IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII
IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII

uawwoflmwoou soammwuwom ucmfloflmwooo
oDHm>IH Hafiuumm mummsmum soammwuwmm Houmamumm
IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII
IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII
wqowu H nOHUNGHEmUQU MO unmfiUHMNUOU Com. Gmmflgoflz I \fluwfiHNNV

Awmssoav mouow uomasa adaflxmz u w

"l"['llll'l
I'll!"

mmmymz<M<m mHmVA<z< ZOHmmmmomm mo Vm<szm ©.N< mam<H

 

   

 

TABLE A2.7

Cob Moisture
Content (percent)

Kernel Moisture
Content (percent)

Ear Weight
(pounds)

Maximum Impact
Force (pounds)

Total Mechanical
Damage (percent)

Kernel Removal
(percent)

74

RANGE OF PARAMETERS FOR WHICH REGRESSION IS VALID

minimum
maximum

minimum
maximum

minimum
maximum

minimum
maximum

minimum
maximum

minimum
maximum

Michigan
500

.256
.853

440
2000

2.06
100.00

Michigan
250

.248
.710

350
1640

0.28
93.99

 

   

 

 

 

 

APPENDIX III

STRESS WAVE PROPAGATION

75

 

 

   

 

 

APPENDIX III

STRESS WAVE PROPAGATION

Stresses and deformation in a viscoelastic body under quasi—
static conditions of loading can be determined by applying the theory of
viscoelasticity. These calculations are based on the assumptions that
the deformations due to applied external forces have reached their static
values and the body is in equilibrium. In such analysis inertial forces
due to deformation or motion are neglected. These methods are suitable
when the external forces are applied for a period of time longer than
that required to establish equilibrium and make the necessary experimental
observations. However, when external forces are applied for only a few
microseconds, inertia stresses attain values of importance and the stresses
and deformations may be analyzed in terms of stress wave propagation.

The impact loading of corn ears discussed in this thesis caused
stresses and deformations which may require stress wave considerations.
Moisture content significantly affected response of the corn to impact
loading, suggesting the possibility of modeling corn ears as viscoelastic
material. The complex geometry of corn ears and the viscoelastic modeling

of corn cause the mathematics of this problem to be very difficult and
outside the scope of this thesis. However, a discussion of a procedure
for developing necessary descriptive equations is in order.

The corn ear can be considered as consisting of cylindrical layers

01? isotropic material with the glumes, lemmas and paleas represented by

76

 

 

 

77

viscoelastic compressive elements and the pedicels by viscoelastic tensile
elements (Figure A3.l). Conceivably a viscoelastic model could be developed
for each layer in the model by experimentation.

To arrive at a governing partial differential equation for stress
wave propagation, Kolsky (1963) considered the stress wave problem for
a continuous elastic body through application of Hooke's Law, strain—

displacement equations and the following equations of motion:

 

 

 

p azu = 80xx + 30xy + Biff
3:2 3x By 32
a2 80 30 80 2
DJ =_E{. +_XX+__Y_ [A3.l]
2 3x By 32
3t
2 Bo 80 80
p 3 w = zx + zy + 22
at2 3x By 82
where:
p = mass density of the medium
u = displacement in the x—direction
v = displacement in the y—direction
w = displacement in the z—direction
t = time
Oxy = stress (The first letter in the suffix denotes the direction

of the stress and the second letter defines the plane

in which it is acting.)
For wave motion in a viscoelastic media the appropriate constitutive
equation must be used in place of Hooke's Law to arrive at a governing

partial differential equation for each cylindrical layer of the corn ear

model.

WOODY
MIDCOB

COMPRESSIVE
ELEMENT

KERNEL

TENSILE
ELEMENT

FIGURE A3,1 THEORETICAL MODEL OF EAR CROSS SECTION. TENSILE

ELEMENT REPRESENTS PEDICEL.

COMPRESS IVE ELEMENT

REPRESENTS GLUMES, LEMMAS AND PALEAS.

 

 

 

79

When the disturbance associated with a stress wave reaches the
interface between two layers, the boundary conditions will be satisfied
if the following quantities in both layers at the interface are identical:

1. Normal displacement

2. Tangential displacements

3. Normal stress

4. Tangential stresses
These boundary conditions can be used to determine the nature of the waves
refracted into the second layer and those reflected back into the first
layer at any interface.

The only way stress wave propagation could have an effect on corn
shelling would be if a tensile wave were propagated through pedicels of
top kernels not in contact with the impacter (Figure 4.2). Under this
condition the tensile wave would help fracture pedicels of the top
kernels. The mathematical manipulation described above could answer
three pertinent questions:

1. Could a tensile wave be reflected from one of the interfaces

to the top of the ear?

2. Would the propagation velocity be sufficient to allow a tensile
wave to reach the top pedicels while the impacting face is
still in contact with the ear?

3. Would the amplitude of such a wave be great enough to be an
important force in the initiation of corn shelling?

If the answers to any or all of these questions is no, stress wave propaga-

tion would not be an important factor in corn shelling.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

    

 

 

 

 

   

 

 

   

 

 

   

     

   

   
       
    

 
     
    

      
      

  
 
 
 

                     

         
 
 
 
    

 

 
 
 
   
   

 

   
  
 

 
          
    

     
 

' ' ' ‘ PICHIGAN STATE UNIV. LIBRARIES
‘ 293107587382
- i I