It]...
Hul I. . . V. . 1.!"Fr‘i ‘ . .‘v
. u. . . .3“”
n: 1315!... ‘ flare-Ia:

.‘Ir.

\

M.

1
.' I"! .PAbefy
{IA-“a (Iii,

{31155. - Fifi!)
, ‘ ..v tn... {1 . ,
.1 it. . .

We

.2 _. ‘ , . , ‘ .‘ 3mm {inmfigibmmvmmfl

5 x.‘

 

.‘L 'f:3's

I)ate

0-7639

Illillllillll'illlllllllilllllllllllllilll

3 1293 00897 8714

This is to certify that the

thesis entitled

Pavmw m, 14 M 9m... $4.7 W

WWWW

presented by

LIA/e; AMA/ca

has been accepted towards fulfillment
of the requirements for

”75 degreein %r'g?y

u%//s; mam

Majorproef

3—5l-95

 

MS U is an Affirmative Action/Equal Opportunity Institution

 

 

 

 

LIBRARY
Mlchlgan State
L Unlverslty

 

 

PLACE IN RETURN BOX to ram ovo this checkout from your record.
TO AVOID FINES rotum on or before date duo.

DATE DUE DATE DUE DATE DUE I

n :4. 42-3-2896 l
I"; .

J

    

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

[\"fi
MSU Is An Affirmative ActiorVEqual Opportunity Institution
cMcMma-od

 

 

PERFORMANCE TESTING OF AN ONION
PEELING MACHINE USING RESPONSE SURFACE
METHODOLOGY
By

Ling Wang

A THESIS

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

MASTER OF SCIENCE

in
Agricultural Engineering
Department of Agricultural Engineering

1993

ABSTRACT

PERFORMANCE TESTING OF AN ONION PEELING
MACHINE USING RESPONSE SURFACE
METHODOLOGY

By

Ling Wang

An onion peeling machine was characterized by peeling efficiency,
peeling capacity and machine peeling loss. Onion shape, onion size, air
pressure and chain speed (each in three levels) were investigated as major
factors affecting the machine’s performance. Three types of Response
Surface Design (Hoke D6, Bos-Behnken and factorial) were conducted to test
the machine. The optimization of the process was performed to result in
maximum peeling efficiency and minimum peeling loss. The computer
generated response surfaces, the canonical analysis and the superimposed
contour plots revealed that the speed of 84 onions/min combined with the
air pressure of 517 kpa on round shape and medium size (83 mm) pungent
Machine onions should yield an optimal operating condition with a peeling
efficiency of 80% while keeping machine peeling loss as low as 25%. At
these levels, 576 kg/h peeling capacity is obtainable.

COPYRIGHTED BY
LING WANG
1993

DEDICATION

I dedicate this thesis to my parents, Xi Wang and You Xian Dong.
They gave me the love and the strength. They indicated the direction for
me and urged me to realize the ideal. My appreciation to them is beyond

words. It is right that the results of their tremendous support and persistent

encouragement, should bring success in my life.

I also dedicate this thesis to Young Ming Sun, my undergraduate
program teacher. He was the first person who guided me into the domain
of mechanical engineering with his extensive knowledge and noble-minded

morality in teaching and research.

iv

ACKNOWLEDGIVIENT

I would like first to take this opportunity to express my deep
appreciation to my major professor, Dr. Ajit K. Srivastava, for his constant
encouragement, continued moral and financial support, genuine help, and

excellent guidance. His efforts in making this learning experience

worthwhile are deeply appreciated. His sincere belief in me kept me going.

Special thanks are to my guidance committee members, Dr. Hira L.
Koul and Dr. Thomas H. Burkhardt. The valuable suggestions, advice, and
help brought this study to a successful completion. A particular obligation
of gratitude is owed to Dr. Hira L. Koul for warmly sharing his own
experience as a teacher and researcher for providing an opportunity leading
me into the domain of statistics. He made the complex statistical concepts

and theories easy to learn and understand and thereby greatly increased my

confidence to apply statistical methods in this research.

I should also express my appreciation to D.G.M. Company, Inc., the
owner of the MSU onion peeling machine. They supplied the valuable
information and the original feasible investigation report. And, they offered

onions, equipments and cooperative workers making it possible successfully

to complete sequential experiments in their plant.

TABLE OF CONTENTS

 

CHAPTER I
INTRODUCTION AND PROBLEM DESCRIPTION ......... 1
1.1 Background ............................ 1
1.1.1 Economic Significance ................ 1
1.1.2 Onion Products ...................... 9
1.1.3 Onion Peeling Methods ................ 12
1.2 The Origination of a Problem ................. 14
1.3 Objectives ............................. 16
CHAPTER H
REVIEW OF THEORY AND LITERATURE .............. 18
2.1 Introduction ............................ 18
2.2 Review of Response Surface Methodology (RSM) ..... 19
2.2.1 Selecting A Quantitative Analysis Method ..... 19

2.2.2 The Basic Assumptions and Concepts of RSM . . 20
2.2.3 The Applications of The RSM in The

Mechanical and Food Processing System ..... 23
2.3 Review of Onion Peeling Machine .............. 31
CHAPTER III
DESCRIPTION OF THE MSU ONION PEELING MACHINE . . . 95
3.1 Introduction ............................ 95
3.2 The Processing Technology of The MSU Machine . . . . 98
3.3 The Construction of The MSU Machine ........... 110
3.3.1 Multiple Slitting in The MSU Machine ...... 111
3.3.2 Self-Regulating Technique in The MSU
Machine ......................... 1 13

vi

 

3.3.3 The Application of Compressed Air Jet Peeling

in The MSU Machine ................. 117
3.3.4 Hydraulics in The MSU Machine .......... 119
3.3.5 Safety Features in The MSU Machine ....... 119
CHAPTER IV
QUANTITATIVE EVALUATION OF THE MSU MACHINE . . . 122
4.1 Introduction ............................ 122
4.2 Variable Definitions ....................... 124
4.2.1 Definition of Three Response Variables ...... 124
(1) Machine Peeling Loss .............. 124
(2) Peeling Efficiency ................. 125
(3) Machine Peeling Capacity ............ 125
4.2.2 Definition of Four Independent Variables ..... 126
(1) Onion Shape .................... 126
(2) Onion Size ..................... 126
(3) Air Pressure .................... 126
(4) Chain Speed (feeding rate) ............ 126
4.3 Design of Experiment ...................... 127
4.4 Results and Discussion ..................... 129
4.4.1 Model Fitting ...................... 129
4.4.2 Diagnostic Checking .................. 132
4.4.3 Response Surface Interpretation ........... 135
4.4.4 Response Surface Analysis .............. 149
4.4.5 Graphical Approach .................. 152
CHAPTER V
SUMMARY AND CONCLUSIONS .................... 160
5.1 Summary ............................. 160
5.2 Conclusions ............................ 162
CHAPTER VI
RECOMMENDATIONS FOR FURTHER RESEARCH ........ 163

vii

 

APPENDICES ................................. 164
APPENDIX A
The MSU Onion Peeling Machine First Experiment
The ANOVA Table and The Raw Data Table ......... 164

APPENDIX B
The MSU Onion Peeling Machine Second Experiment
The Hoke D6 Design
The Experiment Layout, The AN OVA Table And The
Response Surface Plots ....................... 166

APPENDIX C
The MSU Onion Peeling Machine Second Experiment
The Box-Behnken Design
The Experiment Layout and The ANOVA Table ....... 174

APPENDIX D

The MSU Onion Peeling Machine Second Experiment
The Full Factorial Design

The Experiment Layout and The Raw Data Table ....... 177
BIBLIOGRAPHY ............................... 180
1. General References ........................ 180

2. US Patents ............................ 188

viii

LIST OF TABLES

 

Table 1
Onion World Trade: Production and Value, 1982 - 1989 ....... 2

Table 2
The World Dry Onion: Area Harvested, Yield and Production,
1988 - 1990 ................................... 3

Table 3
The World Ten Leading Dry Bulb Onion Producers in 1990 ..... 4

Table 4
The Ten World Leading Exporter and Importer Countries
for Dry Bulb Onions in 1989 ........................ 5

Table 5
Onion Area, Production and Value in the United States, 1973-1989 . 7

Table 6
Onion Area and Production by State in The United States, 1985-1989 8

Table 7
The Independent Variables and Their Levels in The Second
Experiment ................................... 129

Table 8
The Analysis of Variance for The Machine Peeling Loss (MLoss) . . 131

Table 9
The Analysis of Variance for The Peeling Efficiency (E) ....... 132

ix

 

Table 10
The Stationary Points of The Response Models ............. 150

Table A.l
The First Experiment: The AN OVA Table for The Peeling Efficiency
(Air Pressure = 75 psi) ........................... 164

Table A.2
The First Experiment: The AN OVA Table for The Total Peeling Loss
(Air Pressure = 75 psi) ........................... 164

Table A3
The First Experiment: The Experiment Layout and Raw Data Table 165

Table B.1
The Second Experiment: The Experiment Layout and Raw Data Table
of The Hoke D6 Design ........................... 166

Table B2
The Second Experiment: The AN OVA Table for The Peeling
Efficiency (Hoke D6 design) ......................... 167

Table BS
The Second Experiment: The AN OVA Table for The Machine
Peeling Loss (Hoke D6 design) ....................... 167

Table C.l
The Second Experiment: The Experiment Layout and Raw Data
Table of The Box-Behnken Design .................... 174

Table C2
The Second Experiment: The AN OVA Table for The Peeling
Efficiency (Box-Behnken design) ...................... 175

 

Table C3
The Second Experiment: The ANOVA Table for The Machine

Peeling Loss (Box-Behnken design) .................... 176
Table D.1

The Second Experiment: The Full Factorial Design Layout and

Raw Data Table ................................ 177

Table D1 (Cont’d)
The Second Experiment: The Full Factorial Design Layout and
Raw Data Table ................................ 178

Table D1 (Cont’d)

The Second Experiment: The Full Factorial Design Layout and
Raw Data Table ................................ 179

xi

LIST OF FIGURES

 

Figure 1

Figure 2

Figure 3

Figure 4

Figure 4A

Figure 5

Figure 6

Figure 6A

Figure 7

Figure 8

Figure 8A

Onion world trade total, 1982-1989 ...........

Onion fresh market: Foreign trade, The U.S., 1973-

1988 ................................

U.S. Patent 4,457,224 Apparatus For Stripping Onions

(Kino, 1982) ...........................

U.S. Patent 2,602,480, Onion Skinning and Slicing

Machine (Taylor, 1948) ....................

The structure of knives in TAYLOR Machine (1948) .

U.S. Patent 3,724,362, Article Feeding And Treating

Apparatus (Parsons, 1970) ...................

U.S. Patent 3,515,193, Onion Orienter and Cutter

(Aguilar, 1967) .........................

The structure of oriental components of AGUILAR

Machine ( 1967) .........................

U.S. Patent 3,606.917 Peeling Machine (Orlowski,

1969) ................................

U.S. Patent 3,696,848 Method And Apparatus For

Removing Skin From Onions Or Like ............

The structure of slitting and peeling mechanism of

MELLON Machine (1970) ...................

xii ,

4

35

39

42

43

44

 

Figure 9

Figure 9A

Figure 9B

Figure 9C

Figure 9D

Figure 9B

Figure 10

Figure 10A

Figure 11

Figure 12

U.S. Patent 4,068,011 Method Of Peeling By Scalding
And Cutting (Green, 1975) .................

A Top Plan View and Side Elevation of Green Machine
(1975) (Excluding feeder and separator) ..........

A fragmentary, transverse, vertical section taken on the
line A-A of Figure 9A to show the construction of
impaling section (Green, 1975) ...............

A fragmentary view in longitudinal vertical section
as taken on the off-set line C-C of Figure 9A to show
the left-hand portion of the machine (Green, 1975) . . .

A horizontal section taken on the line D - D of Figure
9A to show the pedestal rotating Mechanism (Green,
1975) ...............................

A fragmentary, transverse, vertical section taken on the
line E-E of Figure 9C to show the construction of
gripping section ........................

U.S. Patent 4,481,875, Bulb Peeling Apparatus
(Toyosato, 1982) ........................

The construction of the cutter and cutting section of
TOYOSATO Machine (1982) ................

U.S. Patent 2,494,914, Machine For Clipping Onion
And The Like (I. R. Urschel, 1944) ............

U.S. Patent 2,961,023, Onion Trimming Machine
(Boyer, 1958) .........................

xiii

46

47

48

49

50

51

52

53

 

Figure 13 U.S. Patent 3,623,524 Machine For Preparing Onions

(Buck, 1968) ..........................

Figure 14 U.S. Patent 3,764,717, Method For Automatically

Orienting And Trimming Vegetables (Rood, 1971) . . . .

Figure 14A The structure of cutting and vibrating mechanism of

ROOD Machine (1971) ....................

Figure 15 U.S. Patent 3,861,295, Slitting and timing section of

Figure 15A

Figure 16
Figure 16A

Figure 17

Figure 18

Figure 19

Figure 20

Figure 21

Figure 22

BOYER Machine (1973) ...................

The configuration of discharge section of BUYER
Machine (1973) ........................

U.S. Patent 4,476,778, Onion Peeling (Clyma, 1980)

The Principle of Separator of CLYMAM Machine . . .

U.S. Patent 4,373,589, Harvesting Apparatus For
Onions (Hagiz, 1981) .....................

U.S. Patent 926,286, Onion-Topping Machine
(PETRIE, 1908) ........................

U.S. Patent 1,379,049, Onion Topping (Schroeder,
1920) ...............................

U.S. Patent 4,442,764, Machine For Peeling And
Cleaning Foodstuffs, Particularly Vegetables Such
As Onions (Bos, 1982) ...................

U.S. Patent 1,294,033, Onion-Cutter (Bizette, 1918) . .

U.S. Patent 1,995,694 Onion Snipper (W. E. Urschel,

xiv

55

56

57

58

62

63

64

65

 

1932) ................................ 69

Figure 22A The Orienting Mechanism of URSCHEL Machine
(1932) ............................... 70

Figure 23 U.S. Patent 2,445,881, Apparatus for Peeling
Onions, Including a Conical Jet Gas (Hemmeter,
1945) ................................ 72

Figure 24 U.S. Patent 2,766,794, Method of Removing Outer
Skin From Vegetables (Odale, 1952) ............ 73

Figure 25 U.S. Patent 2,750,977, Apparatus For Clipping
Tops From Onions (V ella, 1953) ............... 74

Figure 26 U.S. Patent 3,485,278 and U.S. Patent 3,485,279
Treatment Of Onions (Parsons, 1966) ............ 75

Figure 26A The Cutting and slitting mechanism of PARSONS
machine (1966) .......................... 76

Figure 27 U.S. Patent 3,765,320 Onion End Cutter (Raay, 1971) 78

Figure 28 U.S. Patent 3,915,083, Apparatus For Automatically
Processing Bulbs And Tuberous Plants (Spruijt,1973) . . 79

Figure 28A The structure of orienting section of SPRUIJT Machine
(1973) ............................... 80

Figure 28B The Structure of slitting and cutting section of SPRUIJT
Machine (1973) ......................... 81

Figure 29 U.S. Patent 4,361,084 Industrial Peeler For Onions Or

The Like Bulbous And Tuberous Vegetation (Raatz,
1981) ................................ 85

XV

 

Figure 30

Figure 31

Figure 32

Figure 33

Figure 34

Figure 35

Figure 36

Figure 37

Figure 38

Figure 39

Figure 40

Figure 41

U.S. Patent 3,543,824, Treatment Of Fruit And

Vegetable Crops (Parsons, 1968) ...............

A flow sheet of the technological process performed
by the MSU onion peeling machine during the onion

peeling operation .........................

A schematic general side view of the MSU onion

peeling machine .........................

The conveying-holding system of the MSU Machine

The construction of the ends-cutting and the equatorial
slitting station of the MSU Machine

A side view of ends cutting and equatorial slitting
station of the MSU Machine .................

The construction of equatorial slitting knife and build-in
air jet nozzle of the MSU Machine ............

The construction of longitudinal slitting station of the
MSU Machine

The construction of longitudinal slitting knife and build-
in air jet nozzle

station of the

The construction of the separating
MSU Machine

The construction of equatorial top slitting knife
assembly of the MSU Machine ...............

The circuit diagram of the hydraulic system of the MSU

Machine ..............................

xvi

91

96

97

101

104

105

109

116

120

 

Figure 42

Figure 43

Figure 44

Figure 45

Figure 46

Figure 47

Figure 48

Figure 49—A Superimposed contour plots for two response models . .
Figure 49-B Superimposed contour plots for two response models . .

Figure 49-C Superimposed contour plots for two response models . .

The diagnostic checking plots of the estimated models

The response surface and contour plots of peeling

efficiency and machine peeling loss as

onion shape and onion size .................

The response surface and contour plots
efficiency and machine peeling loss as

onion shape and air pressure ................

The response surface and contour plots
efficiency and machine peeling loss as

onion shape and chain speed .................

The response surface and contour plots
efficiency and machine peeling loss as

onion size and air pressure ..................

The response surface and contour plots
efficiency and machine peeling loss as

onion size and chain speed ..................

The response surface and contour plots
efficiency and machine peeling loss as

air pressure and chain speed .................

affected by

of peeling
affected by

of peeling
affected by

of peeling
affected by

of peeling
affected by

of peeling
affected by

Figure B-l The response surface and contour plots of peeling
efficiency and machine peeling loss as affected by
onion shape and onion size in the Hoke D,5 design

xvii

. 134

142

144

146

148

155

156

157

. 168

 

Figure B-2

Figure B-3

Figure B-4

Figure B-5

Figure B-6

The response surface and contour plots of peeling
efficiency and machine peeling loss as affected by
onion shape and air pressure in the Hoke D6 design

The response surface and contour plots of peeling
efficiency and machine peeling loss as affected by
onion shape and chain speed in the Hoke D6 design

The response surface and contour plots of peeling
efficiency and machine peeling loss as affected by
onion size and air pressure in the Hoke D6 design

The response surface and contour plots of peeling
efficiency and machine peeling loss as affected by

onion size and chain speed in the Hoke D6 design . . . .

The response surface and contour plots of peeling
efficiency and machine peeling loss as affected by

air pressure and chain speed the Hoke D6 design . . . .

xviii

. 169

. 170

. 171

172

173

LIST OF SYMBOLS

 

 

Symbol Definition

B, = constant coefiicient

Cp = peeling capacity

d = onion equatorial diameter

d, = space between the low and high level of 5,
BF = degree of fieedom

E = peeling efiiciency

a = random error

1; = some response

F-value = the ratio of mean square

h = onion longitudinal height

A = the chacteristic root of the B—matrix
MLoss = machine peeling loss

MS = mean square

P—value = significance level

R2 = the coefi‘icient of multiple determination
R02 = the adjusted R2

SS = sum of square

W, = initial sample weight

W", = sample weight after machine peeling
W] = final sample weight

 

xix

 

Symbol Definition

 

W” _ 4) = the coefiicients of canonical form

5, = actual value in original units of independent variable
S, = mean of high and low levels of E,

x, = coded variables

x 1 = onion shape

x2 = onion size

x 3 = air pressure

x4 = chain speed

17,: = estimated peeling efiiciency

Am, = estimated machine peeling loss

 

CHAPTER I

INTRODUCTION AND PROBLEM DESCRIPTION

1.1 Background

1.1.1 Economic Significance

Onions have been a popular food for many centuries. Today they are
valued for their flavor, aroma, and taste, being prepared domestically or
forming raw materials for a variety of food processes (dehydration, freezing,
canning and pickling). They are probably the most universally used

vegetable in most countries.

Bulbs of the common onion (Allium Cepa) and their products are an
important trade item and appear in most markets of the world throughout the
years (Table 1 and Figure 1). According to the 1990 yearbook of the Food
and Agriculture Organization of the United Nations, total production of dry
onions in that year was about 28 million tons (Table 2). Comparing the
production of 1974, the world production of dry onions increased almost
70% in 1990. The total value of dry onion world traded in 1989 was $1096
million, an increase of 54% over 1982. The main production was in Asia
(49%), followed by Europe, including the USSR. (25%), the Americas
(18%), and Africa (7%). It is estimated that the value of the world

Table 1

Onion World Trade: Production and Value, 1982-1989

   
 
 

 

    

 

 

Export

 

 

 

   

 

 

 

 

 

 

 

  
     

 

 

7 Production Value

1,726,468 416,840 | 1722,045 336,016

1983 1,741,058 378,079 1,749,620 313,750
1984 1,954,169 541,934 1,931,715 441,460
1985 1,931,206 393,265 1,895,821 289,981
1986 1,923,078 435 , 805 1,951 ,731 345 ,034
1987 2,075,274 628,385 2,124,896 489,741
1988 2,157050 603,720 2,192,990 531, 318
2, 103, 287 , 518, 026

 

 

 

production of bulb onions alone approaches $7 billion annually. Since bulb
onions are an easily transportable commodity and can be stored for a period,
approximately $500 to $600 millions worth (at 1989 prices) are traded
internationally each year. The crop is a major export earner for some
economies. The most important onion producers in 1990, with their
production, harvested area and yield are shown in Table 3. The major
exporter and importer countries in 1989, with their quantities and value are
shown in Table 4. In addition, the common onion is also an important salad
crop when eaten green. Because of onion’s economic importance, great
efforts have been made in the development of its cultural and processing

techniques. As a result, their cultivated regions, yields and productions

have increased over the years.

58.82 88.25 833 26.5.8830
233.6th 33 nook 8: \o «8938» 2: Soc 88 203 8an £5 E 8.53 23 835 05 E 358$ 88 =< ...

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

:8: 8m: 8m: :8 8:: Ram E a 89: 80 an 8m N8 .2 Bacon
88 88 88 82 80: 88: 8m: 9;: a: E o: E :8:
e8 5 :2 5 £88 28 Son 488 m w m a. «385
88 38 8:. 88 83 888 888 8: 8m 8m 8m 8m 2.23.
85 M82 82 882 men 85 MEWS :8: 82 82 82 8 s2
88 :8 88 8: 82 8: 88 m8: m2 8: 4: ms arena m
88 88 88 :2 5: 58 cs 8 88 8 P 8 s 8:95 o z
22 3: 82 :2 on: 883 8E 82 c: we § w: 8:2
was 85 ~38 88: 8:: 5: 8w: 8: E: 83 no: 32 253
82 82 82 5.82 82 82 82 5-883 82 82 82 $-83

F: 8.5 88:85 3&3: 22> a: 8.5 waste: 82

 

 

 

 

a 33 . 33 538:8...— EE 33> 638?.«mm 3.3V EoEO .3: 2.53 2:.
N “ho—nah.

 

Table 3
The World Ten Leading Dry Bulb Onion Producers in 1990

{LLLLWCAMWL _ - _ _ _
Producer . TTotal Production Area Harvested
o 7 7- 1000MT 1000 ha -
Chain '

      

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3930 248

India 3350 320
U.S. 2454 58 42.5
USSR 2200 189 11.6
Turkey 1550 79 19.6
Japan I 1280 29 44.1
Spain II 1063 30 36.0
Brazil II 984 74 11.6
Iran II 700 30 23.3
Poland II 550 30 18.4

 

 

 

 

 

1.1.80
p

 

 

 

 

 

 

 

 

 

 

 

 

 

g nee _
.° _ / 1
E sac ........ _
P 5 -1

_ I- - .
8 ; / ‘ ‘ 1
o. .
eeo , ., ..
E" I' ' "
3 . - :
2 no i
E-4 - , .
, ; 3
8°. 1 ‘ A A A J ‘ L A .I-u—I
1982 1.904 1.988 1988 1999

Years

 

 

 

Figure 1 Onion world trade total, 1982-1989

Table 4

The Ten World Leading Exporter and Importer Countries
for Dry Bulb Onions in 1989

 

 

 

 

 

    

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Exporter (ionntry Quantity 1000 MT Value 10‘$
Netherlands 415 100

Spain 233 58

Mexico 150 75

Turkey 150 17 II
U.S l 10 37

Poland 60 20

Italy 55.5 24

Australia 50 20

Pakistan 27 3 H:
Import Country Quantity 1000 MT Value 10% I
FR Germany 335 74

UK 240 57

U.S. 163 69

France 135 35

Malaysia 117 32 II
Canada 110 13

United Arab Emirates 71 31

Bel-Lax 77 18

Singapore 64 20

Kuwait 40 8

 

 

 

 

Onion cultivation in the U.S. was 1,508 million kg (42,489 ha.) in
1973 and 2,433 million kg (53,647 ha.) in 1989 (Table 5). Over this period
the production increased about 65% and the growing area increased only
about 29%. The U.S. onion yield, kg/ha, is one of two highest in the
world. The value of the crop increased about 150% from $207 million in
1973 to $502 million in 1989. Making it the third most valuable of
commercial vegetables, behind tomatoes ($1,824 million) and lettuce ($950

million).

Two crops of onions are grown each year in the U.S., as shown in
Table 6. A spring crop is grown in Arizona, California and Texas, and the
total production was 375 million kg, valued at $91 million in 1989. The
summer crop is much larger, in excess of 2,058 million kg in 1989, valued
at $462 million, comprising both non-storage produce (232 million kg,
grown primarily in New Mexico, Texas and Washington) and storage
produce (1,312 million kg, predominantly from Colorado, Idaho, Michigan,
Minnesota, New York, Ohio, Oregon, Utah, Washington and Wisconsin).
In addition, 514 million kg were grown in California and were used mainly
for processing. The summer crop occupied about 85% of annual total
production in 1989. From 1973 to 1988 the onion imports of the U.S.
increased rapidly and exports fluctuated around 100 million kg, as shown in
Figure 2. In 1982, U.S. fresh onion exports dropped from its peak of 194
million kg in 1981 to 69 million kg. Since then, the situation has improved
somewhat. However, it looks fairly week if we compare it with the trend
toward fresh onion imports, even though, in the last 15 years, the yield has
increased from 35,523 kg/ha (1973) to 42,530 kg/ha (1989). However, due

to the low cost of production in foreign countries, U.S. onion producers

7

Table 5
Onions Area, Production and Value in the United States, 1973-1989

Year Area Production Value

1 207
' 47

 

 

24.

180

120

4.

Foreign Trade (Million kg)

 

73 76 79 82 68 88

 

Years

Figure 2 Onion fresh market: Foreign trade, The U. S., 1973-1988

 

 

mmv N 3%.. N owN N NE N

NWN N

H

H

 

 

 

 

ldl

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SN NN N2 NN 88 ON NS 8 "No 2. ND
fig 23 482 SN; Na; N22. N88 81:. 81.81% 638% .soe
2N 8N aNllNdN 84 8.2 NNN; _ NINE _ _ NNINé N3 <0
2N._ 8N; N8; :2 a 88.8 N38 858 oNNNN aNoNN 33
NN N2 2 2 NN NNN NINN N44 a NNN E
N: #2 2 _ 8 8 SNN SNN NE N: 9.5 <3
N4 NN N4 N N NE NNN NS aw N8 E
NF 8 t. NN 8 :2 N} No; SN; 2.} ESQ-ma ...................................
.- SN NNN SN NNN N.N..N....- ......oo..N.N No.4 83 8N N.N.N:N ...... 28.:ng ....................................
.12 N N N : N2 NNN .2 N2 N2 mo
.1? 2.2 8N2 N: aN NNNJ. NNE. 82.2.3-..- . 2.: .N N4N.N NZ
N N .2 : a :N NoN NoN SN :N ZS
N: NE. 8 ..N 82 EN N..NN.N NBN EN EN S
.NNN.-- ....-N..o..N 4NN WNW-2.....- 8 _ - . NNNN EN 2 _ .N NNNN NNNN B
..1o.N.N.-- NNN .-...N.NN NNN NNN NN..N.N NSN 83 :3. NSN oo 8:2on
NNN EN N2 N2 N: 38118.? New N84. N88 3
N NN NN NN 8 an to SINI NNN
NN a. s. _N N... 23 83 N8; 82
N2 32 NS. 8 i N.N.o.N..-.- .-..%.N.N NN2.N Nam.—
NNN NNN NNN NNN :N 84.: 8N8 N8.N 33
3: § 2: N2 SIN EB SN 83 NNN....
5 NS.-.- .,-.N..N..N E N: N: _.N vaN 2N.N E .N
NN NN 8. N NN N2. 2%. oNN 6.8
82 NN2 82118421381 82 321% 82
"aaéé a 8:. __

 

  

 

awamwa .mD oak. 5 35m 43 55935.5 can 8.32 .mooEO
9 03:8

have met a severe challenge.

According to Jones and Mann (1963), the production of dehydrated
onions in the U.S. increased from 900 tons to over 9,000 tons between 1947
and 1961. Given the more recent improvements in processing technology,
the expansion of fast food outlets and the increasing stringency of quality
and microbiological safety standards, the catering industry generally and the
convenience food sector in particular, demand for these products as well as
oleoresin and essential oils has undoubtedly increased greatly since 1961.
Detailed information about the extent and value of trade in these products

is difficult to ascertain, however.

1.1.2 Onion Products

The main commercial onion products are: dehydrated onion pieces,
onion powder, onion flavoring (including onion oil and onion juice), onion

salt, pickled onions, and canned onions (Fenwick and Hanley, 1989).

To produce dehydrated onions, onion bulbs first need to be cleaned
and peeled, then cut into slices. The onion slices are automatically spread
on a continuous stainless-steel perforated belt, and hot air is blown through
the bed. The residual moisture content of the product should be 4% to 5%
to allow for good storage and acceptable flavor stability. Dehydrated onion
pieces can be put into the market not only as final outcome, but also can be
converted into powder, granules, flakes, kibbled or sliced or used to prepare
such products as onion salt, french fried onion rings or toasted onions. The

convenience and quality of today’s commercial dehydrated onion products

10

have earned them a large market. Tomato catsup contains about 1% fresh
onion equivalent by weight, and chili sauce contains up to 4%. The U.S.
catsup and chili sauce industries use approximately 454,000 kg of dehydrat-
ed onion annually; more than 227,000 kg are used annually in comminuted
meat products. Sauces, soups, mayonnaise, salad dressing, pickles and pet

food contain dehydrated onions as a component (Somogyi and Luk, 1988).

Onion powder may be obtained by grinding dehydrated onion pieces,
but a stronger flavored product may be obtained by spray drying. In this
product, the onions are peeled and washed free of debris, rinsed and blended
to a puree. For best results particle size should be below 0.3 mm. Onion

powder is used when onion appearance and texture are not requirements of

product formulation (e. g., in dehydrated soups, relishes, and sauces).

Onion oil is the most highly processed onion products. The product
is used for its solubility, lack of color and strong aroma. It can be obtained
by the distillation of minced onion which has been allowed to stand for some
hours prior to distillation. The oil itself, a brown-amber liquid obtained in
0.002 to 0.03% yield, comprises a complex mixture of (mainly) sulfur
containing volatile. The product possesses(on a weight basis) 800 to 1000
times the strength of odor of fresh onion, but its commercial value may be
many thousands of times that of the onion. The product is used for its
solubility, lack of color, and strong aroma. Onion oil has been reported to

be used in nonalcoholic beverages, ice cream and ices, confectionery, baked

goods, condiments, meats and pickles.

Sterile onion juice is obtained by the repeated expression of onion

tissue, flash heating the resulting liquor to 140°C to 160°C and immediate

11

cooling to 40°C. The juice is carefully evaporated (18°C to 40°C), usually
to 72% to 75% dry matter, which is necessary for preservation without the
need for chemical additives. The concentrated juice is pale brown in color,
possesses a strong fresh onion odor but lacks the undesirable bitterness
which characterizes the untreated juice. Further evaporation to 82 to 85 %
solids leads to a darkening of the product and to the introduction of desirable
cooked, toasted, sensory qualities. This extract is often mixed with
propylene glycol, lecithin and glucose to yield onion oleoresin having a
flavor intensity ten times that of onion powder (and a hundred times that of

the original bulb).

Onion salt is a mixture of onion powder and salt together with an
anticaking agent. Under U.S. specification it comprises the dehydrated
powder (18 to 29%), calcium stearate (1 to 2%), and sodium chloride.

Fresh onions may be preserved in vinegar as pickled onions.
Generally, silver-skin or button onions are used, as they give a more
desirable product. After trimming and peeling the onions are fermented in
a 10% salt solution, which has the effect of producing a translucent product

with the desired firmness of texture.

Canned onions are sold in volume in U.S.. The onions, which are
generally white skinned small onions (less than 38 mm in diameter), are first
peeled. After blanching (4 to 5 min) in acidified water at about 82°C, the
onions are canned, brined, and acidified. This latter process is necessary
since heat sterilization would destroy the quality of the product. Packaged,

cut onions have been stored in cartons at -l8°C for 12 months without any

observable change in appearance, flavor or aroma (Luh and Kean, 1988).

12

1.1.3 Onion Peeling Methods

From the above survey we can see that in onion processing, after
grading and curing, the first operation is peeling. There are several peeling
method applied in onion products manufacturing. The main methods used
in the modern onion processing industry are machine peeling, lye treatment

and flame peeling.

Flame peeling is usually done with natural gas at high temperatures
(about 2,000°F). The roots and outer shell are burnt off in an oven. The
loose, charred particles are drawn off with high-velocity air, and then the
onions are washed and brushed under high-pressure water sprays, to cool
them and remove the charred skins and dirt. After flaming or preliminary
washing, the bulbs are inspected for defects, and the tops and roots are
trimmed off by hand or by one of several types of mechanical de-rooters.
The onions are then given another washing, inspected, and taken to a rotary
slicer where they are cut into slices for further processing. Due to the
problems, such as scorching or the agglomeration, this method is little used

today in the dehydration process (Fenwick, 1990).

Lye peeling actually is a chemical reaction applied to onion processing
(Hanson, 1975). In this process, the step of cutting or slicing of the dried
leaf and root structure from the onion bulb is eliminated. The onions are
fed into a continuous washer mechanism which agitates the onion bulbs
while at the same time spraying them with water to remove dirt or other
foreign materials clinging to the onion bulbs. Then the onions are placed

in a caustic bath for loosening and removing the outer protective skin layer.

13

The key to this processing is to control the concentration of the conventional
caustic solution, the temperature of the solution, and the time of immersion
in the solution properly for different types of onions. From the caustic bath,
the onions are transferred into a washing apparatus in which the loosened
. protective skin layers, root structures and dried leaf structures are all
removed. The onions are then conveyed by an elevator to a stone separator
tank filled with water. In the tank any stones or non-floating debris sink to
the bottom of' the tank while the onions float and are carried out at one end
of the tank by paddles. The mechanism deposits the onions upon a
continuously moving inspection table where the onions are again inspected
and any floating debris, such as cinders or particles of wood, are manually
removed from the group of onions. From the inspection table the onions are

divided and finally fed into a number of feeding conveyors which lead to a

separate trimming machine.

Because the caustic solution penetrates between the layers and into the
inner flesh of the onion, even though the caustic solution is subsequently
neutralized, this process has already permanently harmed the texture of the
onion. Thus when such onions are packaged, they have a soft or mushy
texture, and the layers of inner flesh or meat of the onion are readily
separated from each other, so that each onion is disintegrated and fails to

retain its inherent form and shape.

Compared to the above two types of onion peeling methods, machine
peeling possesses some distinctive features. In machine peeling, the tip and
the root of the onion are cut physically and pressurized air or water are used

to score off the peel. In this processing, there is no chemical or physical

14

damage to onion products and the onions keep their natural form and shape.
With the development of food processing techniques, the quality of machine
peeling is much higher than flame peeling and lye peeling. This is
extremely important for certain onion products, such as pickled onions,
packaged onions and all kinds of canning foods which contained fresh onion
as an ingredient. In addition, processing onions in a timely manner and
keeping them as fresh as possible, are significant for fast food service
market. As a flexible processing system, the machine peeling method is
suitable not only for onion growers to pre-process before they send the
onions to professional food manufacturers, but also for food industrial plants
for vast processing. Actually, machine peeling plays an important role even
in the flame peeling and lye peeling processing system. Furthermore, in
flame peeling and lye peeling the onion’s outer layers and its roots are burnt
and damaged completely, while in machine peeling processing all parts
which are cut off or peeled by the machine can be salvaged for other uses.
Therefore, the waste in onion machine peeling processing is decreased to its
lowest level. Based on the literature, in past two decades, it is estimated
that about 98% of onion processing patents ratified in the world were

machine peeling patents.

1.2 The Origination of a Problem

In 1984, D.G.M. Co., Inc., a prominent Michigan onion producer,
and the food science department at Michigan State University investigated
the users of fresh vegetables in several states. According to their report
(D.G.M., 1984), the volume of fresh onions used by different local food

services varied from 22.7 to 681 kg per week. However, nowhere was

15

anyone using modern peeling equipment. Their research also showed that
many institutional commissaries would prefer a fresh peeled onion product
if shelf life, product quality, and supply specifications could be met. Many
expressed an interest in buying peeled onions. The Campbell Soup
Company, a large fresh vegetable processor, for instance, in order to insure
the quality and flavor of their products, reversed themselves from ordering
dehydrated onions from California at a considerable saving to go back to
fresh onions. ' The investigator found that the peeled onion product serves
as a remarkable method of discovering markets for other vegetable and food
products. Several national fast food businesses had contacted with D.G.M..
They were looking for new methods of processing and packaging their fresh
products to reduce costs and insure freshness. One of them, Campbell Soup
Co., Inc., contracted with D.G.M. Co., Inc.. Because of this situation, and
considering the potential demands of a number of food industries in
Michigan and its neighboring states, such as Wendy’s, Columbus, Ohio;
McDonald’s, Illinois; Domino’s, Ann Arbor, Michigan; Little Ceasars,
Detroit, Michigan; and Big Boy, Detroit, Michigan, D.G.M. Co., Inc.
determined to establish an onion processing center to supply a peeled fresh

onion product to the market.

Two types of onion peeling machines had been used by the D.G.M.
Co. to process fresh onions before 1988 (Srivastava, 1989). One of them
was the Martin machine developed in California, which is suitable for large
size Western onions. But it does not work well with smaller Michigan
onions. The other machine is made in Japan. It does not trim onions and
requires a large amount of hand labor to finish peeling and trimming. One

of the problems experienced with this machine is unavailability

16

of replacement parts. Both of the machines require a large amount of
compressed air which represents a substantial capital investment and

operating cost.

Thus, there was a need for a farm level onion peeling machine
suitable for smaller Michigan onions. The Michigan Department of
Commerce awarded D.G.M. Company a grant to develop an onion peeling
machine in order to exploit the market as mentioned above. The D.G.M.
Co. subcontracted with the Department of Agricultural Engineering at
Michigan State University to develop such a machine. The system was
originally designed to peel and trim onions at a rate of one onion per
second. For medium Michigan onions, this amounts to a production rate of
about 545 kg/hr.

In 1988, the first prototype was installed for testing at D.G.M. Co.,
Inc., in Stockbridge, Michigan. After one year of pilot production use and
modification, by 1990, the second prototype was finished and put into
operation. In 1991, the machine processed about 317,800 kg of onions for
Campbell Soup Company.

1.3 Objectives

The purpose of the research reported here was to evaluate the perfor-
mances of the second prototype onion machine peeling system built by the
Department of Agricultural Engineering at Michigan State University. The

specific objectives were:

17

1) To define and measure machine performance parameters

as affected by the machine operating variables and the onion properties.

2) To determine the optimum operating conditions for the machine

peeling system using Response Surface Methodology.

3) To make recommendations for design improvements in the

machine.

CHAPTER II

REVIEW OF THEORY AND LITERATURE

2.1 Introduction

Performance Evaluation of the MSU Onion Peeling Machine, is a
multi—disciplinary project. The background which needs to be reviewed is
in three different fields: (1) the products (the onion and issues related to
onion processing and utilization), (2) the equipment (the peeling machine
and issues related to mechanical systems), and (3) the research method
(Response Surface Methodology and other issues related to statistics). The
basic information on the onion and its products have been presented in
Chapter I. The background of peeling machines and statistical methods are
presented in this chapter.

Response Surface Methodology RSM) is a statistical tool used to
analyze and optimize the operating condition of the machine system. In

order to justify the use of RSM some considerations and examples of its

application are presented.

The target of the research is the evaluation of the performance of a
machine system which was newly developed to peel onions for the food

processing industry. The use of mathematics as an aid to process

18

19

understanding does not replace experience and knowledge, though it surely
acts as a significant added dimension to the qualitative approach (Harper and
Wanninger, 1969). While a review of the onion peeling machine is
necessary. However, there has been no published material was found which
relates to the evaluation or the testing of an onion peeling machine. In order
to classify this type of machine, and characterize its specific features and
functions, and, in turn, locate the new machine in a proper position for an
objective appraisal, a review of the literature regarding the onion peeling
machine is presented here to approach a qualitative analysis. Because of the
difficulties mentioned above, this review is based mainly upon the

information collected from U. S. Patents.

2.2 Review of Response Surface Methodology (RSM)

2.2.1 Selecting A Quantitative Analysis Method

The first step in evaluating the performance of an onion machine
peeling system is to build a mathematical model, that is, to use a set of
performance or objective functions to describe the relationships of all
independent variables and their responses. We want to simulate real
circumstances approximately with a set of mathematical equations. In
general, there are two kinds of models, the theoretical model and the
empirical model. The theoretical model satisfies physical phenomena, about
which we know their physical mechanisms. Usually, it is expressed by a set
of differential equations or integral equations. However, when the necessary

physical knowledge of the system is absent or incomplete and consequently

20

no theoretical model is available, an interpolation function, such as a
polynomial, could be used to provide a local empirical model in which
nothing could be assumed except that the response surface was locally
smooth. In some applications, polynomials can be used to approximate quite
complex behavior, and they are frequently applied in the case of the
examination of preliminary data to give a first insight into the form of the
model. The newly-developed onion peeling machine system belongs to this
type of situation. In this system, there are four independent variables and
a multiple-response. Because of the many variables and corresponding
interactions, there is no suitable theoretical equation which can be used to
describe the system. Therefore, a statistical technique which takes this into
account should be used to build the empirical model for the onion machine
peeling system. One of the best statistical techniques for building an
empirical model for the machine peeling system is the Response Surface
Methodology, which minimizes costs by reducing the number of
experimental formulations and seeks optimum solutions easily by using a

computer graphical approach (Floros, 1988).

2.2.2 The Basic Assumptions and Concepts of RSM

The response surface problem usually centers on an interest in some

response 17 which is a function of k independent variables 5,, £2, ..., E," that
is

n =f(€1.E2.m.E,,) (2.1)

The actual form f in Eq. 2.1 is often unknown, and perhaps extremely

21

complicated, particularly in food engineering. But RSM assumes that it can
be approximated by a polynomial function of lower order. For example, in

the case of two independent variables (k = 2), one might assume a model
of the type

y = 50 + [31x1 + [32x2 + [5113‘12 + I322":2 + 912xle + 3

where 60, 6,, .82, B, 1, are constant coefficients, x’s are the coded or
design variables, and the relationship between natural variable (E ’s) and the
coded (x’s) are simply linear; y is the measured response, and s is a random
error. The variables x1, x2, ..., x, are quantitative and are measured on

some continuous scale.

It is further assumed that the 5’s can be controlled by the
experimenter with negligible error. For example, in a food processing
system where the engineer is interested in obtaining an optimal efficiency,
17, of a machine processing system. The efficiency could be dependant upon
air pressure (5,), feeding rate (£2), material properties (£3), and so on. The
35’s can be controlled by the experimenter by developing the machine

system, or by adjusting the operating system with negligible error.

The success of the RSM is based on the approximation of f by a lower
order polynomial in some region of the independent variables. For example,
if the approximating function is linear in the variables, then we write, in

terms of the design variables,

71 = Bo + 91x1 + Bzxz + + pix]: (2'2)

22

and the second-order polynomial is

'1 = Bo + I31"1 I I32"2 + + kak + pl2xlx2 + fll3xlx3 (2.3)

+ m" Bk-ifik-ixk + [311x12 »"' 1322152 + + 5&ka
The coefficients 60, B 1, 62, . . . are parameters to be estimated from the data
collected in the experiment. For k = 2 experimental variables, these

general polynomials reduce to

'1 = 90 + [31):] + [32x2 (2'4)

and

2 2
n = so + B-x- + 5.x. + Bax-x. + m + 02.x: (25)
Strictly speaking, Equations 2.2, 2.3, 2.4 and 2.5 should not be written as
equalities. But it is usually assumed that the approximation is so close that
any lack of fit will remain undetected with some experimentation, so that for
practical purposes it is reasonable to write them as equalities and this is

common practice.

The assumptions which are fundamentally used in the RSM are

summarized as follows:

(1) A mathematical model 17 = f (x,,xz, ..., xk) exists and is either
very complicated or unknown. The variables involved in this model are

quantitative and continuous.

(2) The function f can be approximated in the region of interest by a

low-order polynomial such as Equation 2.2 or 2.3.

(3) The independent variables x,, x2, ..., x, are controlled by the

23

experimenter and measured with negligible error.

2.2.3 The Applications of The RSM in The Mechanical and Food
Processing System

In the past four decades, Response Surface Methodology, as an

experimental strategy, has been employed with considerable success in a

wide variety of situations.

RSM was initially developed and described by Box and Wilson in
1950. In their paper, a scientific approach to determining optimum
conditions was described which combined special experimental designs with
the Taylor First and Second Order Equations in a sequential testing
procedure called "Path of Steepest Ascent. " The fundamentals of RSM and
its underlying philosophy are discussed in many papers and a number of
textbooks. The most comprehensive discussion is that given in the book,
Empirical Model Building and Response Surface, by G. E. P. Box and N.
R. Draper in 1987.

From the early 1950’s to the mid-19608, a number of statisticians and
scientists published articles which described their great interests in
developing and consummating RSM as a powerful optimal method. During
this period, they confined their efforts mainly to the application of composite
design and the method of steepest ascent in the fields of chemistry and
chemical engineering, biochemical and pharmaceutical sciences, as well as

in agricultural research (Hill, 1966).

In 1957, Box introduced the idea of evolutionary operation (EVOP),

24

which assumes normal operation of the industrial process within which
systematic changes would be made giving experimental information. And,
Box and Hunter introduced the concept of rotatability (Box and Hunter,
1957). In 1959, Box and Draper discussed the various reasons for choosing
a design to investigate a response surface (Box and Draper, 1959) and Box
and Lucas discussed the criterion used for selection of a design which

minimized the variance of the parameter estimates (Box and Lucas, 1959).

During. the same period, there were three other major lines of
statistical research on RSM developed by: (1) Robbins and Monro
concerning Stochastic Approximation (1951); (2) Rao concerning Growth
Curves (1958); and (3) Kiefer concerning Optimal Design. In addition to
these major developments there were extensions in the design of
experiments, the form of response curves, the fitting of response curves and
in the general field of data checking (Mead and Pike, 1975).

The more recent work on RSM has been the emphasis on non-linear
models (Box, 1971) and the increasing use of the computer (Cady, 1970),
which has been an important factor in the choice of fitting methods as well
as in the computer graphics approach (Richard, 1979), (Floros, 1988).

According to Hill and Hunter (1966), RSM has been successfully
applied in mechanical and food engineering. In these two fields, RSM is
mainly used in machine processing systems to find a suitable approximating
function for optimizing the operating condition, and in product development
to determine what values of the independent variables are optimum as far as
the response is concerned. The optimization phase of the problem often

involves finding the values of x 1, x2, x,, which maximize the response.

25

A discussion of some successful applications follows.

In mechanical systems, Whidden applied RSM to conduct two
experiments in metals processing. One is to optimize the green tensile
strength in the sand casting process, in which the dry mulling time was
tested as an independent variable. Another is in the development of an alloy
to determine the aging time and temperature which maximized the tensile

yield strength and the elongation of aluminum (W hidden, 1956).

Ross applied RSM in research of Aeroprojects, in which a new and
unique process, ultrasonic welding, was developed for finding the
relationship between the independent results (strength of weld) and the

controllable factors (power, clamping force, etc.) (Ross, 1961).

Underwood successfully applied RSM in designing extrusion screws.
In his experiment, the length of the metering section, the channel depth in
the metering section, the channel depth in the feed section and the screw
speed were considered as independent variables. Rate of extrusion, melt
temperature, net power required, smoothness of operation and thoroughness

of mixing were dependent variables (Underwood, 1962).

Wu successfully applied RSM to optimize the metal processing in a
machine tool system, and a series of results were obtained in different
subjects. Wu indicated in his research report (Wu, 1964a and 1964b) that
with RSM, the number of tests to develop tool-life predicting equations can
be substantially reduced. The reliability of such an equation can also be

estimated. Three independent variables, speed, feed and depth of cut, were

26

investigated in the project. In his other research project (Wu, 1964c) he
indicated that empirical general cutting-tool temperature-predicting
equations, in terms of speed, feed and depth of cut, were developed by
RSM. Later, he added two new independent variables, the side-cutting-edge
angle and nose radius, into the model, and further developed the cutting-tool

temperature model to a five-variable predicting equation (Wu, l964d).

In 1979, Bemesderfer, a senior engineer and statistician at General
Electric presented an eight-point program for the approval of complex new
processes prior to their introduction to production. Data collection was based
on response surface experiments. He indicated that the use of the
procedures has resulted in greater confidence in new processes and in
demonstrably better processes, both in the development laboratory and in the
factory. His research included roll burnishing, electrochemical machining,
electrical discharge machining, laser machining, electroplating, vapor
deposition coatings, thermal spray coatings, inertia welding, brazing, and
abrasive flow machining. Particularly, he indicated that there is no reason
the procedures cannot be applied to any manufacturing process, regardless
of its nature, e. g. , mechanical machining, casting, and heat treatment
(Bemesderfer, 1979).

Geier and Hood applied RSM to build an empirical model for metal
processing. In their study, mean cutting force as a function of depth of cut
and kerf depth, as well as mechanical specific energy as a function of depth
of cut and kerf depth, were developed to describe the influence of

preweakening a rock on the cutting process (Geier and Hood, 1989).

Food engineering research has several characteristics which distinguish

27

it from other research categories. Most food research is process-oriented,
with only a limited knowledge of the mechanisms. Frequently the functional
form is unknown. The lack of a theoretical model requires efficient
experimental techniques to build an empirical model and find the Optimum
operating condition. Thus, the experimental results must estimate both a

functional form and the parameter values for predicting the response.

In 1962, Berry and his co-workers used RSM to study the production
of vinyl starch. The interrelationship of five variables (time, temperature,
pressure, base and solvent ratio) were determined by employing a central
composite rotatable second-order response surface design. A comparison of
predicted and observed value for the degree of substitution indicated that the
response surface design is a good characterization of the relationship
between the variables and degree of substitution. Two steel compression
cylinders, each containing a floating piston and an internal volume of

approximately 2 liters, were used in parallel in the experiment (Berry and
at. al, 1962).

Happer and Wanninger applied RSM to optimize a cereal toasting
manufacturing process. The objective of the study was to determine the
effect of the toaster’s operation on the finished product flavor, color and
specific volume. Raw product moisture, toaster conveying belt speed,
toaster temperature and fan speed governing the hot air velocity were tested

as independent variables (Happer and Wanninger, 1970).

Aguilera and Kosikowski successfully used RSM to analyze a soybean
extruded product process. In this experiment, the effect of three variables,
each in three levels, process temperature (120, 145 and 170° C), feed

28

moisture content (20, 30 and 40%), and screw speed (800, 900 and 1000
rpm) were studied relative to their extrusive characteristics. The objective
of this study was to explore RSM as a tool for a better understanding of the
relationship between extrusion conditions and product characteristics and as
a means for optimizing the process through the simultaneous analysis of
temperature, feed moisture content and screw speed. A fractional factorial
design with three replicates at the center point was used for this experiment.
Runs were performed randomly in a S-head Wenger X-S extruder (Aguilera
and Kosikowaski, 1975).

Box introduced Evolution Operation (EVOP) to a full-scale food plant.
As an example, the yield of the lobster manufacturing plant was studied by
means of EVOP, and the length of claws and pressure between claws were
considered as independent variables. Box indicated that the technique has
been used with particular success in the chemical industry for many years

but is capable of wider application in the process industries generally (Box,
1975).

In 1976, Smith and his co-workers used RSM as a tool to evaluate the
effect of three variables, homogenization temperature (50, 60, 70 °C),
pressure (1000, 1500, 2000 psi) and emulsifier concentration, on the
physical stability of 25% milk fat emulsions in three different series. The
equipment used for sample homogenization consisted of a high-pressure,
controlled-volume pump, a manometer and a parallel-flow heat exchanger.
The experiment used a three variable, three-level central composite design.
Stability index data were analyzed for multiple regression using a Burroughs

6700 computer and a standard statistical computer package (Dixon, 1970).

29

The ten regression coefficients were fed into a Taktronics 4010 terminal to

develop graphical plots on the display screen using the UCDRSM program
written in Fortran (Smith and et. al, 1976).

In 1980, the Response Surface Methodology and the Path of Steepest
Ascent were used in rapidly determining optimum conditions for whipping
a full-fat soy protein produced by ultrafiltration. Four independent
variables, protein concentration (3.5-4.0%), sugar (1.67% w/w), whipping
speed (150 rpm) and whipping time (3.5-4 min) were considered to optimize
the response variables, overrun and stability, respectively. A 2‘ fractional
factorial design was taken in the experiment to determine the initial Path of
Steepest Ascent (Lah, Munir and Richard, 1980).

In 1984, RSM was applied to a boneless ham processing system in
which three processing variables (tumbling, tenderization and temperature)
were optimized for cooked yield (Motycka, Devor and Bechtel, 1984). A
23 factorial experiment with replicated center points, Path of Steepest Ascent
and central composite design were performed for both pre— and post-rigor
muscle, respectively. In the experiment, a mechanical tenderizer was used
to tenderize the meat and the tumbling of the meat chunks was accomplished
with a Universal 190 Inject Star Tumbler operated continuously with
vacuum (584-660 mm Hg gauge).

Oh, Seib and Chung (1985) used RSM to examine the effects of five
variables on the quality of oriental dry noodles. A response surface design
described by Cochran and Cox (1957) was used to study the relative
contribution of a variable to noodle quality and to determine the optimum

level for each variable in the noodle-making process. Following preliminary

30

trials, five independent variables were selected, water absorption (30-80 % ,
14% mb), dough pH (4.0-10.0), mixing time (2-10 min), roll speed (4-20
rpm), and reduction percentage in roll gap (10—50 %). Seven dependent
variables were measured for each treatment: color and breaking stress for
uncooked noodles and surface firmness, cutting stress, resistance to
compression, cooked weight, and cooking loss for cooked noodles. The
optimum conditions for preparing dry noodles were obtained by
superimposing contour plots. The acceptable limits for noodle quality were

based on four commercial noodles from Japan, Korea and Singapore.

Floros and Chinnan (1987) used RSM to evaluate the effect of lye
concentration (4 to 12% NaHO), process temperature (80 to 100° C) and
time (1.5 to 6.5 min.) on the yield, peeling loss and unpeeled skin in a lye-
peeling process of pimiento peppers. Optimization of the process was
performed to maximize removal of the skin with minimum loss of edible

fruit. In the research, they used the Box-Behnken design with three

variables, each in three levels.

Later, they applied RSM to a double-stage peeling process for
Pimiento Peppers. In this experiment, the optimum processing condition
(total removal of the skin, minimum peeling loss, maximum product yield
and highest texture values) were studied and the effect of seven factors
(pretreatment concentration, pretreatment temperature, pretreatment time,
holding time, post-treatment concentration, post-treatment temperature and
post-treatment time) on four responses (unpeeled skin, peeling loss, product

yield and texture) were studied. In this research, they applied the Box-

31

Behnken design to the seven independent variables, each in three levels and
four responses study again. The statistical package (SAS) was used to
generate response surface. Since the stationary points were not only located
outside the experimental region but also were saddle points,

superimposing contours were used to locate optimum conditions for

visible aid in both cases above (Floros and Chinnan, 1988).

2.3 Review of Onion Peeling Machines

In this review, attention is focused on outlining the development of the
onion peeling machine, defining some essential machine functions or
processing steps, and discussing their relationships, as well as, classifying
machines into certain types and pointing out their different utilizations.
Thereby, a qualitative analysis basis will be built for evaluating the
performance of a new type of onion peeling machine. In addition, the

merits and demerits of some typical mechanisms were discussed.

The generation and development of an onion peeling machine, just as
with any other machine, is determined by the requirements of human beings

and mechanical manufacturing abilities.

The onion peeling machine, if we define it as a kind of tool to
perform the function of peeling onion skins, is probably as old as the onion
itself when human beings ate it as a type of vegetable. In the earliest days,
onion peeling was done with a hand-held knife. At that time, a hand-held
knife probably satisfactory for a family’s, or even for a tribe’s requirements

to peel a small number of onions to eat as a food or used to cure diseases

32

as medicines. Later, however, with the development of human activities,
the onion played a increasingly important role in human life. All kinds of
onion products, such as, onion powder, onion oil, onion juice, fresh onions,
dehydrated onions and canned onions were brought to market. At that time,
a hand-held knife did not satisfy people’s needs to process a large number
of onions quickly. They needed equipment to treat onions in a more
effective way. Human beings’ demands determined the invention of the
onion peeling machine and the types of machine which would be developed.
Therefore, processing capacity and efficiency are two important
specifications which are closely related in the development of onion peeling

machine.

Today, in modern food processing factories and professional onion
peeling plants, onions can be processed automatically from raw onion to

onion products in a procedure in which onions are automatically loaded,

oriented, cut, slitted, peeled, washed and sliced or chopped.

As a type of tool, the onion peeling machine meets a need for certain
processing technologies in order to produce desired products. Many
technologies are applied in producing onion products. Onion peeling is the
first important step in these process technologies. The major onion peeling
processing technologies used in practice are hand peeling, lye peeling, flame

peeling and machine peeling.

Lye peeling and flame peeling are rough processing technologies
which are suitable for preparing onions only for some products. In lye
peeling and flame peeling, onions are abraded and treated in a hot solution

of caustic soda or burnt by passing through a furnace. In these processes,

33

burnt skins are removed by scrubbing them either with brusher or passing
them through a continuous type abrasive peeler or rotating them in an
abrasive drum similar in operation to a potato peeler. These methods suffer
from a number of disadvantages. First, they produce incompletely peeled
onions, that is, the core of root remains with the fleshy onion bulbs. For
some onion products this kind of onion is not acceptable and it needs to be
processed further. In addition, the peels must be thrown away; they cannot
be used to produce by-products. So, these technologies are relatively costly
and inefficient. Moreover, they tend to damage the onions and their flavor.
And, in general, they are messy, thereby, creating uncongenial working
conditions. In order to overcome the disadvantages mentioned above and
to meet the increasing demands for peeled onions, in the last 100 years,

many onion peeling methods and corresponding equipments have been

developed.

According to incomplete statistics, roughly estimated, since the
development of technologies in agriculture and industry, onion machine
peeling technology has its genesis in the early 1900s. It is a type of purely
physical treatment method, safe, low cost, and highly efficient. In the
technology, there is no pollution or chemical damage in onion products. It
is a complete peeling method in that it peels the outer layer and cuts off the
core of root from onions, and the peeled portions can be recycled to produce
other onion products. This technology not only meets the needs of large
batch processes, but also flexibly meets small and medium batch processes,

so it is widely utilized in the food processing industry.

It is estimated that about one hundred onion peeling machines and

34

related methods had been developed and granted U. S. Patents according to
the survey from 1890 to 1990. If we include many countries which did not
file patents in the United States in this period, we conservatively estimate

that in the past century, more than 200 onion machine peeling technologies

and equipment have been developed in the world.

The main reason there are so many types of onion peeling machines
in the world is due to the character of the onion itself. Because the onion
is a type of bulb-like vegetation, its numerous varieties, shapes and sizes as
well as its many different applications require onion products producers to
prepare onions in many different ways. For instance, in some Asian
countries, pickled onion is favorite food, for which small size dry onion,
just needs to be cleaned and peeled of its outer dry layer before preserving.
The Patent 4,457,224 (1982), Apparatus for Stripping Onion, filed by Kino
for Fuji Foods Engineering Co. Ltd. in Yokohama, Japan, is probably
designed just for this type of application. The schema of the machine is
shown in Figure 3. Another example is the onion ring, a type of popular
food in the United States. For onion rings, one must first take off the outer
layer and the core of the root from fresh, large-sized onions, and then slice
the cleaned flesh bulb. A machine named Onion Skinning and Slicing
Machine, Patent 2,602,480 (1948), Figure 4, filed by Taylor for Machinery
Development, Co. , in Idaho, probably was designed for preparing this type
of food. Therefore, the design of the onion peeling machines vary with the

objective in different applications.

Another reason for the variety of available machines is capacity. The

capacity refers to the number of onions the machine can process in a certain

35

 

 

\\\\\\\\\

 

 

pooooooooo/

 

 

 

 

 

 

 

Effect“:

 

 

 

 

—. _-----d

 

 

 

09..

 

 

\

 

 

L.

.. \
N. 4%..
or I

ugh .é\\$ uuu .\\\\ae

 

  

\\‘

a .

..Nl;

 

 

 

i013

Fifi. fill/I

\ - $100094
. ..u..//...///// /Ar\va\ro .\\\\\\. \\\\~I

 

M...

 

 

 

upping

Figure 3 U.S. Patent 4,457,224 Apparatus For Str

Onions (Kino, 1982)

ANNE was“: 2.582 N55 25 NEEEN 8:5 NN...NNN.N Ess— .N.: N 2:5

as 5265 .N are. N525 .N 8:5qu .N- as; Namem .N
E swam .N 023 NE .4 ass... 2&on .N £5. Nessie-N .N £5. N58 80m ._

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

\l/
0
s w;
m 11 o
6 w (\
3 ( .n
76‘ .||
Mum. . \
N ,
N il... .
Sauna /
I g . 1 a £36
N __ N
o Stag“ /
m

 

 

 

37

Na: 228: mods? 5 8:5. Na 232:: 2; 3. 2:3.

oas— wfixao 80m .m obs. 38% .658 05 SEE? :28 EN 888 2053
.5820 .3380 05 No 5:8 05 ES gouge—o E obs. wfiozm ESE 05
336% 5N .ME .3 m-m on: no :83 33> .3283. < .N 0:5. mafia m8. ._

 

 

    

C
r
J...-.-...

 

 

 

@

 

sr4

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

38

time. For example, a canning company needs a large-capacity onion peeling
machine to process large volume of onions, three shifts a day, while small
businesses, such as a fresh onion products suppliers need just a small or
medium capacity machine to process onions periodically. In general, there
are two types of onion peeling machines classified mainly by their
capacities, the industrial onion peeling machine and the farm onion peeling
machine. The best-known industrial onion peeling machines are: PARSONS
Machine, Figure 5, (Les. Parsons & Sons Limited, Burry Port, South
Wales, Great Britain, 1970). AGUILAR Machine, Figure 6, (Basic
Vegetable Products, Inc., Henry Aguilar, San Francisco, Calif., 1967).
ORLOWSKI Machine, Figure 7, (Korlow Corporation, Chicago, Ill. , 1969).
MELLON Machine, Figure 8, (Marriott Corporation, Montgomery County,
Maryland, 1970). GREEN Machine, Figure 9, (Ore-Ida Foods, Inc., Boise,
Idaho, 1975) and TOYOSATO Machine, Figure 10, (M. G. 1. Co., Ltd.,
Kanagawa, Japan, 1982). Usually, they have large and complex
construction, high automation devices, and powerful transmission and
electrical control systems. They possess the ability to deal with a large
volume of onions in a relatively short time. However, they are expensive,
and not easy to operate and they need professional maintenance. The best-
known farm onion peeling machines are: URSCHEL Machine, Figure 11,
(J. R. Urschel and G. W. Urschel, Valparaiso, Ind., 1944). BOYER
Machine, Figure 12, (Barrier Center, N. Y., 1958). BUCK Machine,
Figure 13, (Tripax Engineering Company Proprietary Limited, Victoria,
Australia, 1968). ROOD Machine, Figure 14, (Michigan Fruit Canners,
Inc., Benton Harbor, Mich., 1971). BOYER Machine, Figure 15, (4826
Oak Orchard Rd., Albion, N. Y. 14411, 1973). CLYMA Machine, Figure

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1. Slide 2. Frictionless slide 3. Conveyor 4. Transverse shaft
5. Transverse blade 6. Bevel 7. Two eccentric rollers

Figure 5 U.S. Patent 3,724,362 Article Feeding And Treating
Apparatus (Parsons, 1970)

 

E5 .534; 5:5 25 5:25 5:5 92.23“ .53.. .m.: e 2&5

882 .o 833 $530 .m EmEEooE
SEE; .v EmEEQoE 558.203 .m :8 9.9253. 560 3550 .N 28820 95 casein ._

 

 

 

.7

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

41

fig: fidsw< 2.2.2.88 .35.... ..e 9.32:... 2E. <9 95$.—

E... sac—v.20: .N. m9... .552... .: .3 wcmtoaasm .o. :25 .o 9.5m .w £30 .5 at... 02:0 6
somtom 3.80.8... 3.632633. .w .88. H.830 .v :25 .m 8...... .N 22.5.. .253 ..

 

 

 

 

 

 

 

 

 

8 WI 4 m m 3.
T < 3
m. . a. \l j
ibm W_ m , .\s ..n ...,...“ I; O
T. mam <u<

 

 

 

 

42

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 7 U.S. Patent 3,606.917 Peeling Machine (Orlowski, 1969)

 

43

85.3 56:32. 33.3%.;

8:: .O 2.55 5...... .55 wigs...“ ...h 5.33.3. ....< 2.532 wvw.eae.m 2.8.... .m.D a flaw...

...ow 8.03 .3 ENE: .033 .2 >8... .2 250.. ...m

.3. .2 5:5 .0. 2...... $5.5m .o .550 .w x2... 5...... b.5235 .b .550... $5$.$.o .850 .o .5508
.82.“... .m .388... $5.... 35.55.84 .v 8.... $5.... cam ..m 558.. =30...Eom.m 95 58.5.5.6 ..

 

 

1 5.. A ...fixQN

 

 

 

 

 

A A _ ..2l.....IIE . nl....|ullul..|.r.im!|:|\l_ Fl -.‘Nul I. ll I1. lllll //

 

 

N \f.
.4... i: E i. . A.

0.. o o o ._ . .I.. .1. ....- Hnfinlm fiflm

 

 

 

 

 

 

83: 25.2.: 2044mm... .3 E5550:— $.=8._ E... $5.5 ..c 0.5.9:..6 2E. <w 9.sz

8.83 $58.. .m 8:8. $5.... 3.5.35 .m 8.63 $5.? 3:63.33 ..

 

 

 

 

 

 

 

 

 

45

at: .829 3:5 22 9:23 E ”5.8; 5 .5592 :38... 23.3 .m.: a 2:5

couoom wazwuanow .b aozoom 339.5 .0 8:03 wEE—m .m cocoom mafia—U 88m
98 80m .v 53on mafia—E .m 85280 0&530 .N Egon 059nm ._

 

 

 

 

 

46

C8388 can .588 mafia—81nd
as: 2282 :85 a. Saga 2% Ea ES .55 a; < § 2:5

 

  
  

 

m M, m

  
  
 

 

 

 

$53 .50qu .8508 953:: ha 532.5280 2: Beam 8
<a 953% .3 < - < 0:: 2: :6 5:3 .5398 18?? 69.05%: .mufigfiwfia < mm charm

.382: 088"” 332:5 .2 9.0.2 8=om .3 Shots—m .2 32:30
530m .2 59:2: Ewen 90:55:89.: xoabAuaUd— xoahAumo mo 8qu 3585 3653a: .m
383m .w 05,—. .h 8888 033/6 coin wfiflfi bnmouom .n €253. wfiucouxm b03055;
.v Sufi—mo 533 go 03? 3:806 25; 35:8 «0 35— wacflomO.N 5982: 326:8 Emu;

£2 :32: WI-
gfbfl \

m D
.1 ........._H
_

 

2

 

 

 

 

 

 

 

 

 

47

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

GR: .526. 2.28:. 2: :c 55.5: 25.3.2 2: 32:... 8 <a v.53:
he 0 - U 2.: game: 2: .5 5:3 .3 .8503 .853.» _a::.=:w:c_ E 33> 38:053.... < U: «SE:

:25: H. .8 .23: .3 2.8.: 02.98:: .m:
owes: .2 8:235:50“: .2 8:3 8:. .688 :23: .m: m::::o:. 303: .3 :26:— .3
5:3 .330: .N: 895.8 826: .2 03.; :a 1.st: 52:5 526:6 :«fi 83:.» Shari .5
3% 3.9506 088: 5.826 xom...-:.m0 .v .5820 to:::m .m E: 530:0: :80 .N 1833: A

 

48

 

 

 

 

 

49

953862: mo 85:52.8 59. So 5on mass .9.th 03608835

GEE £8qu Egg—.82 $589.
38.x:— 2: 3.5m 8 <a 95$..— .«o G - Q 0:: 05 .5 5:3 5:08 RESIS— < Gm 25w?—

Sm .9 x5..— .m mom mama—50 .v Ea 830:8 :80 .m toga. 333m .N xoabéau A

 

 

 

 

 

 

 

 

1H

 

 

 

III

 

 

III

 

 

 

 

 

 

 

 

 

50

Egg wEE—tu .5 53933.50 2: 32m 3 0a 0.59,,—
uc m— - m— aid 2: :c :35. 558% .8559» .8355»; ..CS—BEwauh < ma 2&5

macaw? tanm 8>E.N_ 5&5”: 35500.2 33. wfiqthd Sufi—>0 ~25on cozoom
Em Ema .5 333m .0 9.0 .w .880 .v mamafibm .m 088% Huang—Em .N 088m 5:02 A

 

2~_:dflfiwlwqqnvm NH

ill. "m3... —

- -.. ..7////.,. ////..l.

 

 

 

 

15W? i3

WW4 i

G!

30/4415:
AI

 

 

 

 

 

 

i %m . ”.04.. ..V%
..N....mfilllll romou. .-
.IIII .. .Ea

__v.....m.n..m_w__-

 

 

 

 

$2 538?: @3833 win...— eam 3436 2.3.8 .m.: S 2&5
5:32‘ mufino: :<.o .238 335-8006 Baa wfisfié 38 ~68on 59:88 wEE_m.N cow—O;

 

51

 

 

 

 

 

 

 

 

 

(‘l

 

OW

o\ 0

ON

0‘ U

@

o“ o

OW U
m\\1

 

 

 

\\ \\\\\\\\\\ II III
I
I

 

 

 

 

 

\\\N\\ \\\\\\<\\\\N N\\s

i .V /I (V

o w v m

 

 

 

 

 

52

$3: 2.282 oe<mo>oe ... 558m .2.an Ea 3:5 .8 58228 2; <2 2&5

3: >53 .2 025 .o :2” Sam .m 025 s
Sm 238.3 .0 ufiNoz .m :35 .v .5582: Bamnmimmw mafia .m gov—08% .N wctaom .—

 

 

 

 

 

 

 

 

 

 

 

53

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1. Rotating wheel 2. Cradle 3. Clamp mechanism 4. Cutting knife
5. adjustable knife mechanism

Figure 11 U.S. Patent 2,494,914 Machine For Clipping Onion
And The Like (J. R. Urschel, 1944)

54

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 12 U.S. Patent 2,961,023 Onion Trimming Machine (Boyer, 1958)

55

:32 32.8 2.2.5 “3232,. Se 22%: Emaagm 28$ .3. 2 2:5

use: 2% as 8895 .w 5:6 s 5&3 mi .0
mi “558% .n 333 wcufim .v 39889“ How :< .m 8.“ch 98on .N 328 bog—om 4

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

56

:53 63%

$.afiowo> mama—EFF a: wEEoZO m:8:aES=< ..ch .3532 :fifififi 2.8a.— .m.D 3 25mg

5355 .m 8on mascara 82:0 .v 32833 =30?Eom .m 0:5— 3:50 .N :oEO A

 

 

 

 

 

 

 

 

 

 

 

 

I'Il
l. J.
kt - l o
. O .
III:

 

 

 

 

 

©©

 

 

 

I
..
A 1.. o y.
\
\
o O

 

 

 

 

57

335 2.2932 906% no Swim—.8:— wfiufifig can wag—5 he 9.32:6 2; <3 953%

8355 H 5.3% .0 =3 :Bo?2om .m
0:5— wfiasv .v uofium .m Eflamnooa mains wfiam .N “8.02. waucoto 8:5 4

 

 

 

 

 

 

58

5.3 2282 «Eon .... Eva was: 2a «:35 £3.53 .53 .m.: 2 2&5

£30.: 3325.2 >896 32:5 Hoe—oomfi 85280 988mg.
83m wEESfiBmd 8?. vaooomd Eu< 5.56 fig wagfioma 83280 EEN 3055.—

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

59

as: 2.582 mESm he 588 massage ... =2§=mc=8 2:. <2 2:5

85¢ macaw 533:, 30.0 85. .6546 comcv 85, :3on 35 :<.N $305 A

 

 

 

 

 

 

 

‘NNNNNNNNNI

II I...

 

 

.0
oil'l.

 

 

 

 

 

 

 

 

 

 

 

 

6O

16, (Spring Gully Pickles Pty, Limited, Australia,1980). They have small
and simple construction, low but practical automation devices and
corresponding transmission and control systems. They are easy to operate
and maintain. They have small to medium processing capacity and are

inexpensive.

Another complication is that onion peeling is frequently performed
together with other agricultural technologies, such as onion harvesting. The
machines are not designed specifically for peeling onions, but they do part
of the peeling job. They are operated in fields, not in houses, such as, Pat.
4,373,589, Figure 17, Harvesting Apparatus For Onions, (Hagiz, 1981).
The machine is designed for Sharnoa Ltd., Petach Tikva, Israel to harvest
field dried onions. The machine includes efficient cleaning, trimming,
sorting and bagging devices. It is even named Harvesting Apparatus and
actually about 50% of the cost of the machine is absorbed in equipment to
perform the peeling jobs, cleaning, sorting and trimming. The PETRIE
Machine, Figure 18, Onion-Topping Machine and SCHROEDER Machine,
Figure 19, Onion-Topping Machine are other examples of onion peeling

machines designed to work in fields after onion harvesting.

Another reason for onion machine variety is that even in the same
peeling technology, there are several different kinds of machines. They
employ different physical principles in design and have different structures
and processing procedures. Take for example, Pat. 4,457,224 (Kino, 1982),
Figure 3 , Apparatus for Stripping Onion. This machine employes the
principle of vacuum pressure produced as compressed air at a high speed

from a nozzle; the onion is discharged from a cylindrical path into

61

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1. Paddle 2. Stem 3. Onion Carrying Device 4. Socket 5. Wheel
6. Spike 7.Nozzle 8. Finger 9. Slitting knife 10. Swinging Arm
11. Adjustable Stops 12.Balance Weight 13. Pivot Pin 14. Spring
Arm 15.Weight 16.Stop 17. Topping and Tailing Knife 18.Feeder
and Sensing Means 19.Abutment 20.Stripper 21.Trough

Figure 16 U.S. Patent 4,476,778 Onion Peeling (Clyma, 1980)

62

 

 

 

 

 

 

1. Rib 2. Conveyor 3. Smooth Conveyor 4. Slat
5. Further Elevating Conveyor

Figure 16A The Principle of Separator of CLYMAM Machine

 

:9: £35 2.2.5 Sm “322%. ”seat“: 33%... ESE .m.: S 95%

Sumo—0-2m .w 08:8 5on 58855 .N. 588th
.0 £93 636 3298.? n 835 go :8 <6 829883 ova—n oESSOm .m 55350 .N :5 wfimwwm .—

 

63

 

 

 

 

 

 

 

 

w m: N a
\\7////\\\..\\?/ // N , I z t x x t //<\ > / , . t
_

 

 

 

 

 

 

by»? / (aw MIIDHMG .. . x
/ 2:: : a /
_. x
/./. l , w. m _

 

 

 

 

 

 

 

 

 

 

/ E

 

 

 

 

 

 

ll .
'1 a.-

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 18 U.S. Patent 926,286, Onion-Topping Machine
(PETRIE, 1908)

 

65

 

 

 

 

 

 

 

 

 

1. Pipe 2. Crank shaft 3. Collar 4. Belts 5. Fan 6. Expansion
spring 7. Cutting bar 8. Slots 9. Stationary cutting bar

Figure 19 U.S. Patent 1,379,049 Onion Topping
(Schroeder, 1920)

 

66

a separating chamber facing the path, utilizing the momentum of the onion
to realize the peeling function. In this machine, there is no cutting or
slitting. The machine has a very different structure and processing procedure
from the classical onion peeling machine. A so-called classical onion
peeling machine is a machine designed specifically to perform classical
peeling technology, which consists of loading, cutting, slitting and peeling
processing steps and works in house. Some machines might add orienting
as first step and washing as extra step after peeling. In classical processing
the onion is transferred by mechanical conveyor or like mechanical devices.
Another example which is beyond the classical onion peeling machine is Pat.
4,442,764 (Bos, France, 1982), Figure 20, Machine for Peeling and
Cleaning F oodsnwfs, Particularly Vegetables Such As Onions. This machine
uses a clothes washing machine-like principle, a container filled with water,
a rotating disc coated with an abrasive layer, and no cutting and slitting in

the processing.

Vertically viewing the history of the development of onion peeling
machine, we can see that the early-designed machines possessed only one
function, trimming or cutting, for instance. Their structures were very
simple. And often they were manpower driven or only semi-automated,
such as, Pat. 926,286 (Petrie, 1908), Figure 18, Onion-Topping Machine.
Pat. 1,294,033 (Bizette, 1918), Figure 21, Onion-Cutter. And Pat.
1,379,049 (Schroeder, 1921), Figure 19, Onion Topping Machine. From
the 1930s to about the late 19405, pockets, conveyors and rotatable cutting
knives were employed in machines to increase the processing capacity and
efficiency. The representative machines are such as, Pat. 1,995,694 (W. E.
Urschel, 1932), Figure 22, Onion Snipper. Pat. 2,494,914 (J. R. Urschel,

67

 

 

 

 

Figure 20 U.S. Patent 4,442,764 Machine For Peeling And
Cleaning Foodstuffs, Particularly Vegetables Such As Onions
(Bos, 1982)

 

68

 

 

 

 

1.0scillatory Plate 2. Blade 3. Opening 4. Blade 5. Plate

Figure 21 U.S. Patent 1,294,033, Onion-Cutter
(Bizette, 1918)

 

69

am: .385 .m .E .253 :35 39mg; Ea“.— .m.: an 2&5

EEO .o Hoxoom .m 83 “88.? .w 835— wauaooafio. .m common wcmuoou .N EflSfiooE comma A

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

o m 4 m N _
fl. 7
3 N \f i ,.
- \ qr
__ ”mm-” ... _ . .-..
_ s t ...
n... -H mm .5”
u. - . L
.r “I - ..J -
L ll. . _. ..

 

 

 

 

 

 

 

 

 

70

 

 

 

 

 

 

l. The structure of finger mechanism 2. Finger opening position
3. Finger closed position

Figure 22A The Orienting Mechanism of URSCHEL Machine (1932)

71

1944), Figure 11, Machine for Clipping Onions and the Like and Pat.
2,553,519 (Lenz, 1946), Onion Taper.

Later on, about in the 19508, the high pressure fluid jet was employed
in onion peeling. This was a very important development. It quickly
became a major peeling technique appearing in almost every classical onion
peeling machine, such as, Pat. 2,445,881 (Hemmeter,1945), Figure 23,
Apparatus For Peeling Onions, Including A Conical Jet of Gas. Pat.
2,602,480 (Taylor, 1948), Figure 4, Onion Skinning and Slicing Machine.
Pat. 2,766,794 (Odale,1952), Figure 24, Method of Removing Outer Skin
From Vegetables. Pat. 2,750,977 (Vella, 1953), Figure 25, Apparatus For

Clipping Tops From Onions.

From the late 19408 through the 1960s, the onion peeling machine
made a great leap forward. Air jet peeling technique, complex orienting
assembly, and self-regulating slitting and cutting mechanisms were added in
onion peeling machines. In this period, the machine developed into a multi-
functional machine system. Many functions, such as, cleaning, orienting,
holding, cutting and slicing, as well as peeling, were often included in one
machine. And the machines had a higher level of automation. The typical
machines are Pat. 2,602,480 (Taylor, 1948), Figure 4, Onion Skinning and
Slicing Machine (included cutting, slitting, peeling and slicing). Pat.
2,750,977 (Vella, 1953), Figure 25 , Apparatus for Clipping Tops from
Onions fincluded air flow orienting and cutting). Pat. 2,961,023 (Boyer,
1958), Figure 12, Vegetable Trimming Machine (included holding and
cutting). Pat. 3,485,279 (Parsons, 1966), Figure 26, Treatment For Onions
(included holding, slitting, cutting and peeling). Pat. 3,515,193

72

 

 

 

00":ch

/—”fi 7
'I1111 l I'G.)
&

 

 

 

Figure 23 U.S. Patent 2,445,881 Apparatus for Peeling Onions,
Including a Coniml Jet Gas (Hemmeter, 1945)

73

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 24 U.S. Patent 2,766,794 Method of Removing Outer Skin
From Vegetables (Odale, 1952)

74

a a . . 5M
39: 5:93 82.5 as..— a..... ”525 Sr.— msaaaa. 5 cm. N 33:.— m a mu 2 E

32m 053 .v 033 mat—.0 .m 33 “2:8 HF. .N 8538 £30 A

 

 

 

 

 

 

 

 

 

75

 

 

 

 

5 A 4 3
, \ /‘
2 :2 ‘
- “K‘: . ‘ 2
. ‘7" i , t
’a $325.?
. i “K

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

77777 /// /

 

1. Onion Carrier 2. Topping and Tailing Knives 3. Frictional
Restrain Star Wheel 4. Slitting Knives 5. Ejecting Star
Wheel 6. Air Blast Skinning Apparatus

Figure 26 U.S. Patent 3,485,278 and U.S. Patent 3,485,279
Treatment Of Onions (Parsons, 1966)

 

76

6:: 328:. mzoga ... 5:28.: 33% 23 $.an .53.— .md <3 28E
065% .w 8=0m .b can 025 .c 8.83 02% 832
98 895 .w 033 mafia .v macaw 5635800 .m ova—n mesa—m .N 8.8 wfifisw ._

 

 

 

 

 

 

 

 

 

 

 

 

 

 

77

(Aguilar, 1967), Figure 6, Onion Orienter and Cutter (included orienting,
holding and cutting. Pat. 3,623,524 (Buck, 1968), Figure 13, Machine For
Preparing Onions (included holding, slitting, cutting and peeling) and so on.
However, in this period, even as the machines’ automation level and
performing functions were increased, their efficiency and reliability were not
satisfactory enough to warrant mass production. For example, the machine
system was timed and intermittent. Such as, Pat. 1,995,694, (U rschel,
1932), Figure 22, Onion Snipper. Pat. 2,602,480, (Taylor, 1948), Figure
4, Onion Skinning and Slicing Machine. Pat. 3,485,279, (Parsons, 1966),
Figure 26, Treatment of Onions and Pat. 3,623,524, (Buck, 1971), Figure

13, Machine for Preparing Onions.

By the 1970s, the machines were built more like industrialized
products, in that their layout was more reasonable and compact, they used
more interchangeable and standardized components and had higher
adaptability. The machine’s automation had been further increased. New
techniques and new materials were widely adopted, and, also, the type of
machine, the machine’s capacity and efficiency, as well as its safety were
developed to a new level to satisfy all kinds of requirements and situations.
The typical works included, Pat. 3,696,848 (Mellon, 1970), Figure 8,
Method and Apparatus For Removing Skin From Onions or Like Vegetables.
This is an embryonic form of modern onion peeling machine. Pat.
3,765,320 (Raay, 1971), Figure 27, Onion End Cutter, electronic sensor and
pneumatic components used in the machine. Pat. 3,915,083 (Spruijt, 1973),
Figure 28 , Apparatus For Automatically Processing Bulbous And Tuberous

Plants is a very large capacity onion peeling machine. It can peel onions

78

 

 

 

 

 

 

 

 

 

 

 

 

 

1 .
5 6

l

2 ‘ I -
H19 “ J Ja":}ti"d7'i" 5

3 tw— .-

l
1 ' 4

 

 

 

 

 

1. Duct 2. Cup-shaped nozzles 3. Cup-shaped opening
4. Pneumatic cylinder 5. Knives 6. Feelers

Figure 27 U.S. Patent 3,765,320 Onion End Cutter (Raay, 1971)

 

79

Amadumaamv
map—«E 3823a. 15 3:5 wfimmooehm b—aozafiefia ..eh usawuann< 33.3.3” «5an .m.D an «Sufi

033 mass Em .8 “0:8: ages 235 .2 522. 235 .M: 220 BE .2 32.3 023 wanes
0333:; Beam .3 Sausage 33m .2 watam ~84 .3 .5355 .2 838508 magma?“ noummom
wane—m .2 5:253. 038355, .2 888 02:0 .2 and—0C. .o 83th .w :35 .h 830 0927.
-> .0 $27. 035%.?“ Emfim .m 3:3 BEE .v one 8,3250 .m goo—n 8:305 .N 3:3 988m A

 

 

S mfl/fl & mm W 2 cm

‘

 

 

 

 

2

 

 

 

80

SR: 2:82 Exam .8 858.9. 3.5:: ca 2385 2:. <3 2:5

Eon? ova—n wfiuoom 0383.82“ Beam .w 823 took .5 380m .0 coo—SE 63:00 .n
522. seas @2082 ..., cams: 58o 35.6 .m 8% ”5.352 .m one; 3:8 Em ._

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

81

RR: 0:2an EMA—m «a 553m wfifla =5“ $53? .3 0.58:.5m 2; man 0.53...—

5382 fine—u 9550 0333.2? .2 £33 mafia cum .: Stanza war—am ova—n

wEnBoZ .2 520: ova—m .a 383. wags?“ 530 .w 22o 397.5 .5 2:5 Bum .0
3:3 “mam .w 38 “98280 .v 383 8:305 .m 3:3 208m .N ova—n 33302 ._

 

 

Z a

(x
00
O\
O
v—¢

 

 

 

 

 

 

 

 

 

 

 

 

 

82

at a rate of 2.5 tons per hour. In this machine, a photoelectric device and
electronic sensor control system were considered. Pat. 4,068,011 (Green,
1975), Figure 9, Method of Peeling Onions by Scalding and Cutting, is a
high automatically operated industrial onion processing machine. It is
comprised of a special orienter and separator, and hydraulic components
were used and special materials (rubber and plastic) applied in the main
components, roller and cam-track. This machine is suitable for large batch
processing. Pat. 4,442,764 (Bos, 1982), Figure 20, Machine For Peeling
and Cleaning F oodsagfifs, Particularly Vegetables Such As Onions, is a new
type onion peeling machine, whose ideas are completely different from the
classical one. The machine is simple and safe, and special material is used
for abrasive coating. Pat. 4,481,875 (T oyosato, 1982), Figure 10, Bulb
Peeling Apparatus. This also is a new type of peeling machine whose
principle and structure are much different from existing machines. Its high
automatically operated system and reliable structure is suitable for large
batch production. Ceramic blades, plastic rollers and holders as well as

hydraulic components and systems were applied in the machine.

From this review, we can see that by the 19703, the classical onion
peeling machine had been well developed. Some fundamental mechanisms
or structures had been recognized and accepted widely by machine users and
designers, such as the endless chain-conveyor, two parallel end-cutting
rotatable disc blades, self-regulating adjuster, multiple surface slitting knives
and fluid jet peeling. These mechanisms appeared in more and more
machines. Researchers’ interests have focused on increasing the machine’s
efficiency, reliability and adaptability and decreased costs, particularly in the

industrial onion peeling machine. There are two obstructions to the further

83

development of the onion peeling machine: loading (orienting), and holding
problems. Actually, this is one interrelated problem. Previous practice
indicates that if this barrier cannot be broken down, it will be difficult to
increase the classical onion peeling machine’s efficiency, reliability and
adaptability. Before the 19703, mechanical engineers did some work to
improve loading and holding techniques. In holding, for example, spindle-
holding, pocket-holding, gripper-holding and belt-hold—down holding, had
been tried in a number of machines. The same situation emerged in loading
or orienting. Example are vibrator-pocket orienting, Figure 14, (Rood,
1971), friction-pocket orienting, Figure 6, (Aguilar, 1967), air-pressure
loading, Fig. 25 , (V ella, 1953) and mechanical-fmger orienting, Figure 22,
(Urschel, 1932), as well as gravity-rolling orienting, Figure 28, (Spruijt,
1973). But, most machines adopted hand loading or orienting. This is
because orienting onions to a proper position is a very important processing
step in the classical onion peeling machine, and it requires that the orienting
mechanisms possess high reliability. Also, obtaining a reliable onion
orienting device is not easy either in design or in manufacture. The problem
is that onions variy both in their size and shape. It requires a complex
mechanism to do the job adequately, and this adds to the machine’s cost and
reduces its useful life. This is not acceptable to consumers, particularly in
case of small or medium processing capacity, such as farm onion peeling
machines. Therefore, hand loading or orienting becomes a more and more
popular procedure in farm onion peeling machine systems. Nevertheless,
in an industrial onion peeling machine system, onion loading or orienting
still is a barrier to further increasing processing capacity and efficiency.

Solving the bottle-neck problem attracted engineers’ interest, and in the

84

1980s, the developments made in the industrial onion peeling machine
mainly centered on solving the orienting problem to increase the peeling
efficiency and capacity. The representative arts are, such as, Pat. 4,361 ,084
(RAATZ, 1981), Figure 29. This machine used rolling orienting and
flexible belt holding. Pat. 4,442,764 (Bos, 1982), Figure 20 and
Pat.4,457,224 (Kino, 1982), Figure 3. Pat. 4,470,345 (Miyata, 1983).
These three machines eliminated the orienting step by making a complete
change in processing technology. Pat. 4,481,875 (Toyosato, 1982), Figure
10. This machine employed rolling and comb-shaped pawls for orienting

and holding onions.

How many functions and what functions a machine must perform are
mainly determined by application and cost. For example, an onion grower,
if he is the supplier to the onion peeling plant or fresh vegetable market
needs a machine to perform trimming and cleaning functions only.
Commonly, the more functions a machine has, the more the buyer will
spend. Loading, that is, putting the onion in the proper place and position
for subsequent processes, such as, slitting and cutting, is the first and most
important step, particularly in large batch processes. Loading has influenced
the whole processing quality and cost. In classical peeling technology,
usually there are two kinds of loading methods, machine loading and hand
loading. In machine loading, onions are sent by a mechanical conveyor to
an orienting mechanism. The principle of gravity, vibration and physical
friction are widely applied in the design of such orienting mechanisms. The
representative arts are, such as, Pat. 1,995,694, Figure 22, URSCHEL
Machine. Pat. 2,750,977, Figure 25, VELL Machine. Pat. 3,515,193,
Figure 6, SGUILAR Machine. Pat. 3,764,717, Figure 14, ROOD Machine.

85

20% .530 803000.»
mseonah 0.2 002:3— 8—5 0:8 ..O 0:015 00h 00—09” 3.50:0:— vchan. “Exam .m.D 0N 0.53m

:0: 02500200 05 mo :00 000800 wE>02 6N :5 000.3 £802 .3 :00
buaotflm .VN :0: wEEoto 05 mo :00 80:60 wE>02 .MN Sm .NN :00 0200 AN ES anflwcfiam .ON
:00 02:0 .2 0025. wE=S 0:0 wfiamoh .3 800820 wficofinoh .5 00.82200 000—95 .2 2550 32m
.2 =0. 02:» 00329 .3 02:0 .2 @030 33$ .2 :3 0003 £802 .2 0003 mam—0:008 b33528
500E003 .9 w 80: 8 530600900 0003 £802 .a £05 303m .w :5 002m .0 @025 .0 Sm .m .200
Encoufim .v 02% 30.5095 02:0 .m 0003 9500008 300025000 3005003 .N 00.32.00 0022M A

 

2 0m 2 mm mm an“ 0:0 8

  

3 S M:

  

 

 

 

 

 

 

 

 

86

Pat. 3,915 ,083, Figure 28, SPRUIJT Machine. Pat. 3,942,428, CLAUSEN
Machine. Pat. 4,068,011, Figure 9, GREEN Machine. Pat. 4,361,084,
Figure 29, RAATZ Machine and Pat. 4,481,875, Figure 10, TOYOSATO
Machine. In hand loading, onions are put by hand to pockets or carriers in
a certain position. Such machine has a relatively simple structure and,
therefore, lower cost. However, the trade off is higher labor costs. 80,
usually, only large batch processes adopt machine loading, while small and
medium batch processes often use hand loading. Of course, there are some
peeling machines which do not need a loading function, such as, Pat.
4,457,224, Figure 3, KING Machine and Pat. 4,442,764, Figure 20, BOS
Machine, since they apply different processing technologies. Similar cases

also exist in holding, slitting and cutting functions.

Holding also is a key function, particularly in machines which are
equipped with pre-touch self-regulating devices for trimming and slitting,
because, they need onions to be held firmly to endure the impact of feelers.
In holding, onions are fixed by friction or external force to ensure efficient
slitting and cutting. There are a number of holding techniques. Which
holding method should be used in the peeling process mainly depends upon
the process technology. In earlier times, the ends of onions were removed
by manually pushing or pressing the onion against a revolving knife. In this
manner, first removing one end and then turning the onion around to remove
the other end, the holding was actually done by hand. Since this method
is slow, unsatisfactory and dangerous, later on, clamping pin holding (such
as Pat. 2,494,914, Figure 11, URSCHEL Machine and Pat. 4,476,778,
Figure 16, CLYMA Machine. Spike holding, such as, Pat. 2,445,881,
Figure 23, HEMMETER Machine. Pat. 2,602,480, Figure 4, TAYLOR

87

Machine. Pat. 2,766,794, Figure 24, ODALE Machine. Pat. 3,515,193,
Figure 6, AGUILAR Machine. Pocket holding, such as, Pat. 2,961,023,
Figure 12, BOTER Machine. Pat. 3,485,278, Pat. 3,485,279, Figure 26,
PARSONS Machine and Pat. 3,861,295, Figure 15, BOYER Machine.
Clamping holding, such as Pat. 3,623,524, Figure 13, BUCK Machine.
Pat. 3,606,917, Figure 7, ORLOWSKI Machine and Pat. 3,765,320, Figure
27, RAAY Machine. Belt-pocket holding, such as, Pat. 3,696,848, Figure
8, MELLON Machine. Pat. 3,764,717, Figure 14, ROOD Machine. Pat.
3,915,083, Figure 28, SPRUIJT Machine and Pat. 4,361,084, Figure 29,
RAATZ Machine. Wheel-pocket holding, such as, Pat.4,068,011, Figure
9, GREEN Machine and Pat. 4,481 ,875, Figure 10, TOYOSATO Machine)
was used in all kinds of classical peeling machines. The difficulty in design
of holding mechanism for a classical peeling machine is the interference
between the holding mechanism and the multi-slitting knives. The best
design of a holding mechanism for multi-slitting peeling in the previous
machines are Pat. 3,623,524, Figure 13, BUCK Machine and Pat.
3,696,848, Figure 8, MELLON Machine. In order to accomplish a
complete slitting, however, they all added a complex clamping holding

mechanism for transferring and turning onions.

Trimming means using knives to cut off the top and root portions of
the onions. In the classical onion peeling machine, the function of cutting
is performed before slitting and peeling. However, at present, there are
some machines whose trimming function is performed after peeling and
there is no slitting step in the peeling process. The key in the design of the
trimming mechanism is a self-regulating feature. The most important

characteristic of the onion peeling machine which distinguishes it from other

88

machines is that it processes objective products, that is, onions. Onions vary
in their size and shape. This is one of the most difficult points in designing
an onion peeling machine. How much of the onion will be cut depends
upon the space between the two cutting knives. If the space is fixed, the
onion will be over-cut, in the case of large-sized onions. This causes higher
weight loss. In the case of small-sized onions, the cutting will be
insufficient, thereby decreasing the peeling quality. It is need a simple
mechanism which can automatically adjust the space by itself to suit all
kinds of onions individually. Therefore, whether a machine is supplied with
a self-regulating cutting mechanism and how sensitive the mechanism is
becomes criteria for evaluating the advantages of the machine. There are
several kinds of self-regulating cutting mechanisms. The representative
structure is a pre-touch mechanism. The principle is setting two solid
feelers on the onion path. On the other end of the feelers is connected a
cutting knife assembly. When the onion passes the feelers, the onion itself
pushes the feelers away from the central path, depending upon onion’s size
or shape. The feelers always are kept in a pre-determined position by the
tension force of the spring. The sensitivity varies with different designs of
the feeler assembly. Examples of the pre-touch cutting mechanism are Pat.
3,485,279, Figure 26, PARSONS Machine. Pat. 3,515,193, Figure 6,
AGUILAR Machine. Pat. 3,623,524, Figure 13, BUCK Machine. Pat.
3,764,717, Figure 14, ROOD Machine. Pat. 4,068,011, Figure 9, GREEN

Machine and Pat. 4,476,778, Figure 16, CLYMA Machine.

Slitting is using knives to slit the surface of onions. Even though,
slitting is only a preparatory step, it seriously influences the quality of

subsequent peeling step. In early machines, there is no slitting function

89

before peeling (such as, Pat. 1,995,694, Figure 22, W. URSCHEL
Machine. Pat. 2,494,914, Figure 11, J. URSCHEL Machine. Pat.
2,445,881, Figure 23, HEMMETER Machine. Pat. 2,602,480, Figure 4,
TAYLOR Machine. Pat. 2,766,794, Figure 24, ODALE Machine). Later
on, a one-slit procedure (equatorial or longitudinal) was considered (in
machines such as, Pat. 3,485,279, Figure 26, PARSONS Machine. Pat.
3,623,524, Figure 13, BUCK Machine. Pat. 3,861 ,295, Figure 15, BOYER
Machine. Pat. 3,915,083, Figure 28, SPRUUT Machine. Later, both
equatorial and longitudinal slitting were incorporated, such as, Pat.
3,696,848, Figure 8, MELLON Machine. Pat. 4,068,011, Figure 9,
GREEN Machine. Pat. 4,361,084, Figure 29, RAATZ Machine. Now,
almost every classical peeling machine includes a slitting function.
However, since there are differences in the design of knife’s structure, the
kinematic locus and the installation of knives, the machines have different
efficiencies. An excellent slitting mechanism design is in Pat. 3,696,848,
Figure 8, MELLON Machine. In this machine one equatorial and two

longitudinal complete slits are made.

Peeling is the target function. There are several ways to peel onions.
The most popular way is using high pressure afflux to blast the outer layer
of onions. It loosens the outer layers and separates them from the fleshy
bulb first, and, then, using special mechanisms such as a roller or tripper,
cleans and further strips the loosened skin from fleshy bulb. Air and water
are a common medium. Such as, Pat. 2,445,881, Figure 23,
HBMOMETER Machine. Pat. 2,766,794, Figure 24, ORALE Machine.
Pat. 3,623,524, Figure 13, BUCK Machine. Pat. 3,606,917, Figure 7,
ORLOWSKI Machine, Pat. 3,696,848, Figure 8, MELLON Machine. Pat.

90

4,068,011, Figure 9, GREEN Machine. Pat. 4,476,778, Figure 16,
CLYMA Machine and Pat. 4,481,875, Figure 10, TOYOSATO Machine all
are typical Air-roller peeling machines. There are other peeling techniques,
such as using whirling belt or straps (in such machines as Pat. 2,602,480,
Figure 4, TAYLOR Machine. Drum peeling, Pat. 3,485,279, Figure 26,
PARSONS Machine. Pat. 3,543,824, Figure 30, PARSONS Machine. Pat.
3,724,362, Figure 5, PARSONS Machine (1970). Pat. 3,861,295, Figure
15, BOYER Machine. Pat. 4,457,224, Figure 3, KINO Machine and Pat.
4,442,764, Figure 20, BOS Machine).

Peeling efficiency is an important criterion to appraise peeling
assemblies. Peeling efficiency is directly related to the fluid pressure, the
nozzle installation and their shape as well as the distance between nozzles
and onions. In 1945, Hemometer introduced a conical, diverging, hollow
air jet through a nozzle to loosen the onion skin individually in his Patent
2,445,881, Figure 23. This is an early example of using pressurized fluid
to peel the onion skin, but its low efficiency will not serve the needs of
large batch processes. In 1948, Taylor designed a interesting peeling
mechanism, Figure 4, which consisted of two peeling stations. In the first
station, onions were held by a revolving spindle, the peeling functioned by
the combined action of a set of flexible straps which are whirled at high
speed in contact with the skin, and two steam jets assist in the removal of
the skin. The second peeling station repeats the same work done by the first
station. In every functioning time the conveyor is paused. In modern mass
production, this kind of situation (two peeling stations and one paused
conveyor) probably would not be allowed. The problem in the design is the
installation of the two steam jets. In 1952, Odale introduced an efficient

91

$02 .8830 320 033003, .5. :20 00 .558; ..Safia 22a.— .00 3. 2.0.6

0000.5; .0 swab 005000 00930.5 .0
:8 Beeswax .m 8:... .4 28:32 .0 05800 .N 025 E 832080 ._

d/In/ ..

 

 

 

 

 

   

 

.s. .....I.
... .o ,0.- a.

- 543234800083.

 

 

 

 

 

 

 

 

92

peeling way, Figure 24, in which a jet is directed against one of the cut end
surfaces at its outer layer as the onion is being rotated. Obviously, air jets
installed in this way make the peeling efficiency much higher than in the
Taylor Machine discussed above. Later, in 1968, Buck developed the idea,
Figure 13, of installing two air jets on both sides of the onion, directly
against the slits. This design is much better than any previous work.
However, the shortcoming is still that there is a distance between the air jet
and the onions, and probably this kind of installation is suitable only in an
intermittent peeling system. In 1970, Mellon designed a more effective
peeling assembly, Figure 8. He installed two sets of jet nozzles on the
onion path. One set has two jet nozzles oppositely spaced against the onion
cut end surfaces and initially lie on the path of an advancing onion and are
displaced from the path when engaged by the onion. The second set of jet
nozzles is placed to strike the longitudinal cuts in the onion skin. The
advantage of the design is that the nozzles are very close to the slits, so they
more effectively use the air force. The disadvantage is that since there are
a distance and a period of time between slitting and peeling. The distance
causes the machine to run longer, and the time allowes the onion skin to
shrink.

Summarizing the machine review, we can see that the onion peeling
machine has been developed keeping in mind the human demands and the
onions’ characteristics (which vary with their shape, size and variety) as
well as their products, processing batch size, and the working environment.
These machines can be grouped into two categories generally, industrial
onion peeling machines and farm onion peeling machines. Both of them

have valuable applications. In the case of farm onion peeling machines,

93

they run in the environment of fields or barns; the operators are farmers or
seasonal farm workers; and the machine owner usually is a farmer, which
requires that the machine be (1) inexpensive both in capital investment and
running costs; (2) simple and durable in construction and operation,
respectively; (3) small in size and multi-function; (4) and high in processing

quality and small or medium capacity.

In addition, in each category there are two types of onion peeling
machines, classical onion peeling machines and special onion peeling
machines. Each machine employs different peeling technologies. The
classical machine is specifically designed for performing classical processing
technology and working in house, which cuts onion’s top and root portion
first, then slits and peels it. Besides this technology there are other physical

ways of using the machine to peel onions, all belonging to special peeling

technologies.

Moreover, it is known from the review that the holding mechanism,
self-regulating cutting feature, multi-slitting knife assembly, air jet
installation, general machine layout and continuous operation feature as well

as power, convey and control system are all important aspects in evaluating

a classical onion peeling machine.

Furthermore, it is understood that the onion peeling machine
developed from a manpower-driven or semi-automated process to high
automation; from a single function to multi-function; from a purely
mechanical system to electronic control and hydraulic power drive system,
and that the the improvements always have been centered on increasing

efficiency, adaptability, reliability and decreasing costs.

94

Finally, it is known that it is not enough merely through qualitative
analysis to evaluate a new type of onion peeling machine. The machine
capacity, peeling efficiency and processing weight loss also are important

criteria for evaluating the performance of an onion machine peeling system.

However, these belong to a quantitative analysis which needs a
statistical experiment and response analysis. Quantitative analysis will be

presented in Chapter IV.

CHAPTER HI

DESCRIPTION OF THE MSU ONION PEELING
MACHINE ’

3.1 Introduction

The onion peeling machine described here was originally
conceptualized by Dr. Ajit K. Srivastava and fabricated in the agricultural
engineering department at Michigan State University in 1990. It was
designed for a Michigan onion grower to peel pungent Michigan onions for
use as the soup ingredient by a canning company. The process which is
performed in an onion peeling plant is shown in Figure 31 and the schematic

of the peeling machine is shown in Figure 32.

According to the initial proposal, the onion peeling machine should
be able to handle smaller Michigan onions with a 500 kg/h peeling capacity
and minimum peeling waste. The machine should be designed for farm
level use, i.e., it must be simple, durable, and require as little maintenance

as possible. It is also required that the machine use as many "off-the-shelf"

parts as possible so that parts availability would not be a problem.

This chapter evaluates the MSU Machine qualitatively. It consists
of two parts, the processing technology description and the construction

evaluation. In the first part of the chapter, the focus is on the introduction

95

96

 

 

Pungent Michigan Onion
(row, whole and unpeell

 

let Stage - DUMPING
Onions are dumped into a hopper.

 

 

 

Conveyed by an elevator into a feeding tray

 

2nd Stage - LOADING AND ORIENTING
Onion is deposited in receptacle and oriented by hand.

 

 

 

Conveyed by endless chain conveyor

 

3rd Stage - ENDS CU‘I‘I‘ING AND EQUATORIAL SLITTING
Onion is held on receptacle. Its ends are cut off and skin is
slitted equatorially. Air jets loosen skin simultaneously.

 

 

 

Conveyed

 

 

4th Stage - LONGITUDINAL SLITTING
Onion is held by a wheel on receptacle and skin is slitted
longitudinally. Air jet loosens skin from slits simultaneously.

 

 

 

Conveyed

 

 

5th Stage - SEPARATING
Onions are dropped into a Inclined separator, the loosened
skins are blown by air jets and-stripped by rollers.

 

 

 

Con ve yed

 

 

6th Stage - INSPECTING
Peeling quality is inspected by eyes.

 

 

 

Conveyed

 

 

7th Stage - WASHING
Onions are washed by water jets on a conveying belt.

 

 

 

Conveyed

 

 

8th Stage - PACKAGING
Peeled and cleaned onion bulbs are packaged for shipping.

 

 

 

 

Figure 31 A flow sheet of the technological process performed by
MSU Machine during the onion peeling operation

 

97

05:00:. $0.00: :0_:0 3m: 05 .00 33> 020 .000:0w 0:050:00 < mm 00=wE
0000000002 2000: 00.0 02.0 0000:: 0:00:00: 0:5— m:§:U.w 86:000.; 0380000 0:5— wfixg m0:m
.0 30:3 03000—0: wEE—m 05033040 000000 000% w:00_:w0:.:0m.m 0.00:3 55000—0: wEE—m
3:000:00 0:0 wfizso 00:m .0 50300000 30:? 85000—0: .m 080 35:00 .N 0.0000000: 000000-> A

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

cam w0

 

 

 

98

and discussion of the technological and processing requirements designed for
the MSU Machine to produce acceptable onion products efficiently and
economically. The second part is devoted to the description of functional

components and their distinctive features and effectiveness.

3.2 The Processing Technology of The MSU Machine

The functional processes performed by the MSU Machine in a
continuous operation flow include loading, ends-cutting, equatorial slitting,
longitudinal slitting and separating. In the processing procedure, there is no
chemical or physical damage to onion products. The technological process
which consists of eight stages, is briefly shown in Figure 31. The following
is a detailed description of the processing technology.

The first two stages were designed to feed onions properly into the
machine. The key points in these stages are the loading and orienting
quality. Improper loading or insufficient orienting will cause wrong cutting
and slitting, consequently, decreasing the peeling quality. However, slow
feeding will decrease production. Thus, both factors influence the economic
benefits. Also, in view of peeling capacity, annual production quantity, and
the budget of capital investment which have been proposed initially, a
simple, low cost and reliable feeding way is desired. For this reason, the
MSU Machine adopts a lower cost and higher reliable hand-loading and
orienting means.

In processing, preliminarily cleaned and graded raw, whole and

unpeeled onions are dumped into a hopper. An inclined elevator with

99

paddled belt is inserted into the hopper, it can be operated easily by hand
feeder to convey onions into a feeding tray. The tray is located in front of

the operator for conveniently loading and orienting onions onto receptacles.

In the Second Stage, the onions need to be deposited individually
into receptacles and positioned to lie transverse to the direction of travel.
The onions are loaded and oriented by hand in V-shaped notch receptacles
which are fixed with the links of a chain.

In the Third Stage, the technological requirements are to cut off the
onion’s top and root portions (including the tough core material) properly
and to make an equatorial slit on the onion’s outer-layer to promote easy
removal of the skin from the onion. The technique that the MSU Machine
needs to apply to accomplish the technological requirements in this
processing stage is mainly determined by the onion’s physical and chemical

characteristics.

The onion’s fibre is stratified in a longitudinal direction and, extends
from the root to the stem. By cutting off its root and top portion, one
severs the fibre. In addition, the composition of an onion is approximately
88% moisture (Ikram, 1971). Therefore, after ends cutting, because of the
severing of the fibre, there is loss of moisture, and also, due to the release
of existing tension in the outer-layer, the outer-layer shrinks towards the
middle portion or equatorial portion faster than the fleshy bulb does. In this
circumstance, if we add two slits on the onion surface, the peeling efficiency

is greatly increased.

In addition, because of the variation of onions both in their shape

100

and size, in order to cut the top and root portion properly, the space
between the two disc cutting saws should be varied to accommodate onions
individually. Furthermore, considering peeling efficiency and peeling loss,
the equatorial slitting length and the slitting depth which covers the
equatorial portion should be controlled properly.

In addition, because the onion contains a number of acids as well as
pectin, mucilage and acid sol. ash (Ikram, 1971), the slitting blades should
be replaced easily by standard ones and, compressed air should be employed

effectively to loosen the outer-layer from their fleshy bulb.

Based upon the above considerations the technological process in
this stage is designed as follows: the onion is carried by the receptacle
passing through the ends-cutting and equatorial-slitting station. First, the
onion engages a pair of parallel vertically installed hold—down wheels. As
shown in Figure 33, these wheels are suspended on the frame of the
machine with rocking arms, and can be moved by onions up and down
freely in a vertical direction. With the help of gravity, the hold-down
wheels hold the onion in the receptacle firmly. Almost at the same time, the
onion engages a pair of feelers which are solidly connected with the ends-
cutting knife assemblies, as shown in Figure 34. The onion’s shoulder
pushes the feelers apart from their pre-determined position depending on its
size and shape. The feeler and the end-cutting knife assemblies are
suspended on the framework of the machine and can be moved by the
passing onion freely in the transverse direction of travel. When the held
onion continuously travels forward, it engages two parallel arranged

vertically rotating disc saws, and two pre—determined slitting blades, as

101

95:32 3m: 2: he E39? w:%.e:-w:_»o>:8 2:. mm ouawwm

Ease wagoEoUd BE way—ooms $23 :30?Eo: maria Eamvéwaoq .c 580 .n Engaged
.v 32?» 553.20: want—m 3:233 28 wagso 3am .m 83 way—com .N xuokofiat 05:82 A

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

102

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1. Ends cutting knife hydraulic motor 2. Self-regulating feeler
3. Ends cutting knife disc saw 4. Equatorial top slitting knife
assembly 5. Ends cutting knife assembly suspension

Figure 34 The construction of ends-cutting and equatorial
slitting station of the MSU Machine
(Bottom equatorial slitting knife assembly is removed)

 

103

shown in Figure 34 and Figure 35, first the bottom one, then the top one.
These blades are placed vertically in the center of the path of the onion and
forced by tension springs into a pre-determined position. After the onion
passes through the station, its ends are cut off by the rotational disc saws,
and two equatorial slitting are made by the slitting blades in its
equatorialportion.

There are two compressed air jet nozzles beside the slitting blade.
As shown in Figure 36, when the slitting blades cut the outer-layer of the

onion, the compressed air jets loosen the outer layers immediately.

In the Fourth stage, the technological requirement is to make a
completely longitudinal slit on the middle front of the onion. However, the
thickness of the outer layers are varied in a longitudinal direction; they are
thickest in middle portion and thinnest in each end portion. So, the slitting
depth should be varied to follow the variation of the thickness of the outer
layers. In addition, the slitting blades should be easily replaced by a
standard one. Compressed air jets should be employed effectively to further

loosen the outer layers from their flesh bulb.

During processing, the onion carried by the receptacle first engages
a vertically installed hold-down wheel, as shown in Figure 33. Gravity
helps the hold-dowm whell to hold the ends-cut and equatorially-slit onion
firmly on the receptacle, the hold-down wheel which is suspended on the
frame of the machine with a rocking arm and which can be moved by the
onion freely up and down in a vertical direction. Then, the ongoing onion
engages two slitting blades, as shown in Figure 37, one on the left side and

another one on the right side, horizontally placed in the middle of the onion

104

525:8: 2: 305: 555.20: 0:: :5: 3mm 85 55:0 «6:0 0:9
«55:2 9mg :5 .5 5:5: 55:: 3.55:3 1:: 53:9 5:: .5 33> 03: < mm 9.:wE

3a: 85 5:50 mgm .w 2.5 h: 05:85 A. £9583 0:5— wfiszm :8 353:5”:
.0 3:: ha 5835500 .m 3:503: 0:5. magma 595: 3:955 .v 22:88:
:05: conunm->.m 550 .N 30:? 56:65: 55:: 3:853 A5: wfitg mgm A

 

 

 

 

 

 

 

 

105

0:252 9m: 05
.5 2:5: 00m :5 5-25: .5: 0:5. w:_8=m 3.55:3 .5 5:55:50 05. ea 0.53,:

new: 202: .0 528%. 0:5: .m 033 a: mag:
3:833: .v owfitao 0:5: .m 57:: Ba :2 .N 250: E. :0 30.005500 5-25m A

 

 

 

 

 

 

,H

 

 

.~fl
i‘l

 

 

 

 

 

 

 

 

 

 

106

 

 

 

 

 

 

 

 

 

 

1.V-shaped notch receptacle 2.Ends cut onion 3. Equatorial top slit
4.Longitudinal right slitting knife assembly 5.Compressed air tube
6. Longitudinal left slitting knife assembly 7.Compressed air tube

Figure 37 The construction of longitudinal slitting station of the
MSU Machine
(Hold-down wheel is removed)

107

path. After the onion passes through the station, the blades slit the front
side of the onion in a longitudinal direction. The compressed air jet nozzle
is set beside the slitting blade, as shown in Figure 38, and when the slitting

blades cut the outer layers of the onion, the compressed air further loosens

the skin.

In the Fifth Stage, since the loosened outer layers are still mixed
with the fleshy bulbs, it is needed the machine to separate the broken and
loosened peels from the flesh bulbs and send the bulbs to next station. A
roller separator and three compressed air jet nozzles are employed to
accomplish this process. In the processing, the onion (its ends cut,
equatorially- and longitudinally-slit and outer layer-loosened) is moved to the
end of conveyor by receptacle and dropped into an inclined roller type
separator, as shown in Figure 39. The centrifugal effect and the friction
force, which are created by the high speed rotated rollers, combine with the
compressed air pressure, cause the loosened outer-layers of the onion to be
further peeled or stripped and, the broken peel pieces are blown off and sent
by the rollers out of the separator through the gap between the two rollers.

The bulbs are spun off from the outer layers and rolled down to the exit of

the separator.

In the Sixth Stage, the peeling quality is inspected by eye and all
damaged, diseased, bruised and discolored or defective portions are removed
by hand.

In the Seventh Stage, the qualified onions are put on a conveying

belt to pass through several water jets, where all dirt is washed off.

108

25: E ..a 5-3:... 23 £5. «5:... Evans. ... 88228 2i an 2&5

c8 org .5 @820 .0 Bug: bu 332950 .m
“2.5 as Eschaaou .... 03:8 35 .m 528% e5 .N 023 mags. savanna: .fi

 

 

 

 

 

 

 

 

 

 

 

 

109

 

 

 

 

 

 

1. Onion dropping guides 2. Friction roller 3. Stainless steel
roller 4. Rubber cover 5. Separator framework

Figure 39 The construction of separating station of
the MSU Machine
(Air jet nozzles are removed)

 

110

In the Eighth Stage, the completely peeled and cleaned onions are
packaged and weighed for shipping, or for temporary cold storage.

3.3 The Construction of The MSU Machine

The MSU onion peeling machine consists of six major mechanical
systems: (1) Hold-down and conveying system, (2) Ends-cutting system,
(3)Slitting knife system, (4) Separating system, (5) Compressed air system,
and (6) Hydraulic system.

The MSU Machine has the following advancements. First, the
MSU Machine uses simple and reliable mechanisms to accomplish multiple
outer-layer slitting (equatorial and longitudinal). Second, a self-regulating
technique is employed in holding, ends-cutting and slitting. Third, a
compressed air jet peeling technique is efficiently applied in slitting and
separating. Fourth, the whole machine system is driven and controlled by

a hydraulic system.

The MSU Machine not only is advanced, but also practical. The
practicability is mainly displayed in its compact construction, simple and

durable mechanisms, effective performance and economical cost.

Some advanced techniques often are employed in different functional
mechanisms in the MSU Machine. For instance, the technique of self-
regulating is not only applied in the ends-cutting system, but also in the
hold-down wheel and slitting mechanisms. In addition, in the machine, the
accomplishment of some process technologies employs not just one

technique, but frequently the synthesis of multiple techniques. For example,

111

equatorial and longitudinal slitting not only depend upon the slitting knife
system, but also upon the association of a specially designed receptacle and

conveying system.

3.3.1 Multiple Slitting in The MSU Machine

Equatorial and longitudinal slitting on the onion’s outer layers are
important features in a farm onion peeling machine. In previous machines,
most had only one outer layer slitting (longitudinal slitting), such as, BUCK
Machine, Figure 13, (U .8. Pat. 3,623,524, 1971). CLYMA Machine,
Figure 16, (U .8. Pat. 4,476,778, 1980) and even GREEN Machine, Figure
9, (U.S. Pat. 4,068,011, 1975). Although the MELLON Machine (Figure
8, U.S. Pat. 3,696,848, 1970) had both complete equatorial and longitudinal
slitting, it used a complex mechanical system (two sets of slitting knives plus
two separated conveying chains, a pair of holding chains and one rack

rotating mechanism). And, the procedure of processing was not continuous.

The main difficulty in realizing both equatorial and longitudinal
slitting in onion peeling is the interference between the holding system and
the multiple slitting assemblies. That is the holding system obstructed the
path of the equatorial or longitudinal knife assemblies to make free slitting
on the onion’s surface. The MELLON Machine adopted a procedure which
rotated the onions to be able to make do with equatorial slitting knives.
Consequently, the complex mechanical system not only increased the cost
and decreased the reliability, but also decreased the suitability, because, the
space between the two onion holding chains were fixed, which is suitable

only for certain size onions.

112

In order to avoid the interference between the equatorial slitting
knife assemblies and the holding system, the MSU Machine uses a pair of
equatorial knife assemblies on opposite sides in the vertical center of the
onion path. The pair can swing in a vertical plane, as shown in Figure 34
and Figure 35. And, the machine employs a specifically designed
conveying-holding system, as shown in Figure 33. The system consists of
a pair of extended pitch roller chains which are driven by a pair of cluster
sprockets. A pair of V-shaped notch plate receptacles are fixed with certain
corresponding links of the chains. A pair of parallel hold-down wheels are
separately arranged and suspended vertically on the frame by a pair of

rocking arms.

The onion is carried by the V-shaped notch receptacle from one
stage to another. When it reaches the ends-cutting and equatorial-slitting
stage, it engages a pair of hold-down wheels (being pressed on top) and is
held by the two plates of the receptacle (being supported on the bottom
firmly to engage the pre-determined slitting knife members). Between the
pair of hold—down wheels there is a space which allows the top member of
equatorial slitting knife assembly to freely swing vertically between them.
And, also, since there is a space between the pair of plate type receptacles
and the pair of chains, the bottom blade of equatorial slitting knife assembly
can freely make low-half equatorial slits from the bottom side, and the top

blade of the equatorial slitting knife assembly can freely make a top-half

equatorial slit from the top.

In order to realize the longitudinal slitting, a pair of longitudinal

knife assemblies are installed on opposite sides in the horizontal center

113

of the onion path, as shown in Figure 37. The pair swing in a horizontal
plane. Since the front comer of the V-shaped notched plate is designed
lower than the level of the longitudinal knives, the pair of blades of the
longitudinal knife assemblies can make slits on the surface of the middle
front of the onions without any obstruction when they are carried by the

receptacle to engage them.

From above description, we can see that the MSU Machine not only
obtains the desired equatorial and longitudinal slits, but also the performing

systems are simple, reliable and practical.

3.3.2 Self-Regulating Technique in The MSU Machine

The characteristics of the processing object, the onion, which varies
both in shape and size, dictates that the application of a self-regulating
technique becomes a criterion in the performance evaluation of a farm onion

peeling machine.

The MSU Machine successfully applies a self-regulating technique
in (1) Holding, (2) Ends-cutting and (3) Slitting.

In Holding, the self-regulating technique is utilized in two
progressing stages, Ends-cutting and Equatorial and Longitudinal Slitting.
In both of these stages the onion is held in place by hold-down wheels and

the receptacle.

In the Ends-cutting and Equatorial Slitting Stage, a pair of parallel

wheels are arranged vertically to hold down the onion onto the receptacle.

114

Each hold—down wheel is suspended on the frame of the machine with a
rocking arm. Ball bearings are used to connect the arms with the wheels
and the frame. When the onion passes under the wheels carried by the
receptacle, the wheel can rotate around the arm-wheel joint and the arm can
be moved by the onion up and down in a vertical direction rocking around
the arm-framework joint. Because of gravity, the hold-down wheels always
hold the onion in the receptacle firmly, whatever the size and the shape of
the onion.

In the equatorial slitting stage, a single identical hold—down wheel
is used to hold all kinds of onions firmly in the receptacles when they pass

under the wheel.

In the Ends-cutting process, a pair of feelers are installed in the
onion path which are solidly connected to a pair of end-cutting knife
assemblies and can be pre-adjusted relatively to ends-cutting assemblies. The
assemblies are suspended on the frame with a falling hinge through a
pendulum arm. The arm can sway around the arm-framework joint in a
transverse direction to the onion’s travel and is pre-forced by a tension
spring. When an onion engages the pair of feelers carried by receptacle and
held by the wheels, the onion’s shoulder pushes the feelers apart from their
pre-determined position in the transverse direction of travel, adjusting the

space between the disc cutting saws. Whatever the size and the shape of the

onion, the amount of cut off from onions always is pre-determined.

The self-regulating technique also is applied in both the equatorial
and the longitudinal slitting process, wherein two pairs of slitting knife

assemblies are installed on opposite sides in the center of the onion path.

115

The members can swing in a vertical plane for equatorial slitting and a
horizontal plane for longitudinal slitting when they are engaged by an onion,

as shown in Figure 35 and Figure 37, respectively.

The key point of a self-regulating technique in the slitting process
is its suitability, that is how well the locus of the slitting blade suits the
contour of the onion (which varies individually in size and shape). This
problem also exists in the design of the cutting and holding self-regulating
mechanisms. But in those cases, the measuring or feeling which is required
is only a point or a straight line, that is, the touch of the feeler with the
onion is only a point or straight line touching. However, in the slitting
process, it needs a curved touching or measuring. The MSU Machine has
developed successfully the technique in the onion slitting process with
specially designed slitting knife assemblies. Since the assemblies have
similar construction, only the equatorial top slitting assembly is discussed

here. The construction of the assembly is shown in Figure 40.

The equatorial top slitting knife assembly consists of one tension
spring, one knife rod component and one hydraulic buffer. In processing,
the onion is carried by a receptacle which makes a rectilinear invariable
motion, and the knife rod component functions by a tension spring which
makes the slitting blade remain in touch with the contour of the onion. The
knife rod is moved by the onion’s rocking around a rod-cantilever pivot.

The slitting blade makes a curve on the onion’s surface in a central vertical

plane from the‘middle front point to the middle back point.

In order to avoid over cutting in the process of equatorial slitting,

particularly at the beginning of the slitting when onion dashes against the

116

0:282 Dm—Z 2: he baa—8mm £5. mafia—m n3 23.—3.3.x. he 53252.8 2:. 3. «...—wt—

5833 we 0:5. 3:5... gm 3:925 .m
83m .v wEEm deacon. .m Stan 0:38 m .N ficBoEat 0:3an 4

 

 

 

 

 

 

 

 

\QN4
\\ .
—‘ fl/

 

 

 

 

 

 

117

slitting knife, a shoulder is designed on the knife carriage. And, the

beginning position of knife carriage is predetermined in a vertical position

as shown in Figure 36.

When the top slitting blade passes the top point of the onion, it
cannot engage the top slitting blade continuously. In order to help
continuing the slit on the backside of ongoing onion, the MSU Machine
adopted a disc blade in the equatorial slitting knife as shown in Figure 36.

In longitudinal slitting, the depth of slitting must vary with the
thickness of the outer layers in an equatorial direction. A specially shaped
shoulder is designed on the knife carriage. The shape of the shoulder
associated with the motion of the carriage makes the depth of the slitting

vary with the thickness of the outer layer in an equatorial direction, and thus

the depth of the cut is controlled.

A hydraulic buffer is added to produce a cushioning effect and to
absorb the shock when the downward-moving slitting rod stops quickly for

the oncoming onion.

3.3.3 The Application of Compressed Air Jet Peeling in The MSU

Machine

The compressed air jet is an important feature in onion machine
peeling both from the standpoint of the history of the development of the
onion peeling machine and the economical benefits of compressed air
technology. The MSU Machine has successfully applied the technique in

slitting and separating process, and, has developed the application of the

118

technique to a new level.

In past years, people have realized the advantages of the compressed
air jet in onion peeling, even though peeling efficiency is not high. The
problem is that the compressed air has not been utilized effectively. In
order to increase peeling efficiency using compressed air, the MSU Machine
employs a type of specially designed air jet nozzle in the onion slitting
process. The novel air jet nozzles are built within the slitting knife
assemblies and are arranged just beside the slitting blades. When the slitting
blade slits the outer layers of the onion, the compressed air jet loosens them

immediately. The construction of the slitting knife assembly with the air jet

nozzle is shown in Figure 36 and Figure 38.

In the slitting process, following the slitting blade a jet of
compressed air penetrates the outer layers of onion through the slit made by
the blade. The air pressure causes the outer layers to loosen from the fleshy
bulb. Because the compressed air jet is working together with the slitting

blade and very close to the slit, the compressed air is extremely effective.

In the separating process, three air jet nozzles function together with
the rollers. They are arranged along the length of the roller and towards the
gap between the two rollers. When the ends are cut and the outer layers are
slitted and loosened, the onion is dropped into the rollers, and the broken

peels are blown off by the compressed air jets and swept away by the

rollers.

119

3.3.4 Hydraulics in The MSU Machine

The spattering of washing water and onion juice (which contains a
number of acids) when the onion is cut, slit and peeled creates a wet and
corrosive environment in which the MSU machine must operate. A
completely hydraulic power and control system was adopted to fulfill the
requirements of safety for the machine’s operation and maintenance. In

addition, the hydraulic components and system have the advantage of

efficiency, economy and dependability in the farm onion peeling machine.

The MSU Machine is powered by a hydraulic power pack. The
hydraulic circuit is shown in Figure 41. The system utilizes a vane pump
with a priority flow divider. One flow is directed to the two ends of the
cutting motors connected in series. Another flow is used to drive the
separator rollers and the return flow is used to drive a conveying chain
motor through a low valve to control the chain speed. A pair of adjustable
pressure relief valves and a two-way direction valve are installed in both
sides of the flow divider valve to fully unload the pump in the event of an
emergency shut down. In addition, a suction screen is installed on the inlet

line of the pump and a filter is used in the return line.

3.3.5 Safety Features in The MSU Machine

Safety measures are fully considered in the design of the MSU
Machine. In addition to adopting the hydraulic power and control system,
at the end of the chain conveyor, the outside of the ends-cutting stage and

the conveying chains are all covered. The feeding station is designed far

120

0:332 3m: 95 a: 58?? 0:35;: 2: ..e =8..ch 3595 BE. 5 2&5

£028»: 85 mm .2 52¢ 8:

530m .2 888 03% Ease wfizoZ—ov .2 03? 35:8 Boa 0583:}. .m .588
03.6 8:833 .w 588 mafia—U .h o>_m> 33.025 20:28 > OS .0 “0:8 8:305
035325. .m BEE. Boa mimm .v 8on 0585 .m 955 05> .N 528 5:26 A

 

NP

 

 

 

 

121

away from the moving components. Before the ends-cutting stage a
protective gate is added, and in its design the MSU Machine widely adopts
passive performing components in loading, ends cutting and the two types

of slitting. Finally, a compressed air jet is used to increase the safety both

of the onion products and the operator.

CHAPTER IV

QUANTITATIVE EVALUATION OF THE MSU
MACHINE

4.1 Introduction

This chapter is devoted to the quantitative evaluation of the perfor-
mance of the MSU onion peeling machine. The evaluation was based

mainly on a sequential statistical investigation.

Because the MSU onion peeling machine is a newly developed
machine system and its performance is a complex multi-factor case, it cannot
be represented or analyzed satisfactorily by theory-derived models.
Consequently, two statistical experiments were conducted on the second
prototype of the machine, and Response Surface Methodology (RSM) was
used to build empirical models to simulate the machine system approximate-
ly. The purpose of this approximation was not to represent the true
underlying relationships everywhere, but merely to graduate them locally in
the experimental regions. Presented in this chapter is the second experi-
ment, which was designed and performed based upon the previous plant’s

experience and the conclusions summarized from the first experiment.

In the first experiment of the MSU onion peeling machine (Wang,
1992), a factorial design was conducted to test only two independent variab-
les, chain speed (10, 13 and 24 M/min.) and onion size (60, 82.5 and 95

122

123

mm), each in three levels. Two responses, peeling efficiency and total
peeling loss were observed. Another independent variable, air pressure, was
kept constant (517 kpa). However, the factor of onion shape was not
considered in the experiment. The raw data table and the results of the
analysis of variance are presented in Appendix A. The results of the
experiment indicated that both Chain speed and Onion Size were significant
to peeling efficiency and total peeling loss. In addition, the results showed

that the interaction of chain speed and onion size was significant in total

peeling loss.

In order to further test the onion machine peeling system, second
experiment was designed and conducted in the onion peeling plant, in Spring
of 1992. In this experiment, four independent variables, onion shape, onion
size, air pressure and chain speed were tested. Three responses, machine
peeling loss, total peeling loss and peeling efficiency, were investigated to
perform a response surface routine analysis by the analysis of variance as

well as computer generated response surfaces and their contour plots.

In this study, extra response surface designs were conducted based
upon the data collected from the second experiment. A computer interfaced
backward elimination selection procedure was conducted to determine the
"best" polynomial equations. Canonical analysis was used for judging the
location of possible limit values (maximum, or minimum, or saddle) and
computer superimposing method was used to search the optimal operating
conditions which satisfy for both of the models of peeling loss and peeling
efficiency locally. The objectives of this study were to: (1) better

understand the relationships between the factors affecting the machine

124

peeling system and the responses determining the effectiveness of the
process, and (2) determine a set of optimal operating conditions for the

existing machine system to attain the maximum processing benefits.

4.2 Variable Definitions

There were four independent variables and three response variables
considered in this study. The independent variables were onion shape,
onion size, air pressure and chain speed, and the response variables were
machine peeling loss, peeling efficiency, and machine peeling capacity.

These are defined and discussed below.

4.2.1 Definition of Three Response Variables

The performance of the onion peeling machine was characterized by
three response variables: machine peeling loss ( MLoss), peeling efficiency,
(E), and peeling capacity (Cp). Since onion varies with its weight (mass)
individually, the variables of machine peeling loss and peeling efficiency
cannot be measured directly, and so, they were transferred as the functions
of the initial weight, W, (kg), the machine peeling weight, Wm (kg), and the
final weight, W, (kg). The weights were measured by an electronic scale

directly.

(1) Machine Peeling Loss: Machine peeling loss was defined as the
amount (weight) of peels removed by the machine in proportion to the initial
onion weight. This reflects exactly the amount of peels removed by the

machine per unit weight of onion. The challenge is to have high peeling

125

effectiveness while keeping the machine peeling loss to a minimum. The
following equation was used to calculate machine peeling loss:
W, - W (4.1)

MLoss=——-—"x100
W.

where MLoss = machine peeling loss (%)

(2) Peeling Efficiency: Peeling efficiency was defined as the ratio of
the amount (weight) of peels removed by the machine to the total amount
(weight) of peels removed in the processing. This describes the peeling
effectiveness directly, that is, it describes what percentage (or portion) of the
effective or necessary peeling was done by the machine. Once again, the
objective is to obtain a high value of peeling efficiency while keeping the

weight loss to a minimum. Peeling efficiency was computed as follows:

E = " x 100 (4.2)

 

where E = peeling efficiency (%)

(3) Machine Peeling Capacity: Machine peeling capacity (kg/h) was
defined as the total weight of onions processed by the machine in an hour.
Since onions vary in weight, at a certain feeding rate different peeling
capacities will result. This can be calculated as the sum of the individual
weights of onions being processed per hour. For instance, the capacity of
the round shape and medium size onion is calculated by using an average

weight of the onions, as follows:

Cp = o,15(kg/onion) x 60(min) x Chain Speed(onions/min) (4.3)

126

Where Cp = Machine Peeling Capacity

0.15 (kg/onion) = average individual weight for round shape and
medium size of pungent Michigan onion

4.2.2 Definition of Four Independent Variables

(1) Onion Shape: Onion shape was defined as the ratio of its
equatorial diameter (d) and longitudinal height (h). There were three types

of onion shape considered in the experiment. They were flat shape: d/h 2

1.2; round shape: 0.8 < d/h < 1.2 and oval shape: d/h S 0.8.

(2) Onion Size: Onion size was defined as equatorial diameter (d).
There were three types of onion size considered in the experiment. They
were, small size: 70 mm (2.75 in); medium size: 82.5 mm (3.25 in) and

large size: 95 mm (3.75 in).

(3) Air Pressure: Air pressure was measured directly from the air
pressure meter connected to the air control valve. There were three levels
of air pressure used in the experiment. They were, low air pressure: 414
kpa (60 psi); medium air pressure: 517 kpa (75 psi) and high air pressure:
620 kpa (90 psi).

(4) Chain Speed or Feeding Rate: Chain speed was determined by
counting onions passing per minute. There were three levels of chain speed
set in the experiment. They were, low chain speed: 60 onions/min;
medium chain speed: 80 onions/min. and high chain speed: 100 on-

ions/ min. .

The equatorial diameter (d) was measured by a set of pass-type scale.

The longitudinal height (h) was measured by a general caliper. The air

127

pressure was determined by a air pressure control valve and the numerical
values were read directly from the meter combined with the valve. The
chain speed was determined by adjusting the hydraulic control valve and the
numerical values were read directly from the meter combined with the
control valve. After setting the values to each level of the independent
variables, the measure errors in the data preparation comparing to the one
of responses were very small, thus they were considered as constant

distribution and were neglected. This treatment is thought to be reasonable

is based upon the assumptions which are fundamentally used in the RSM.

4.3 Design of Experiment

In this investigation, three experimental plans were adopted from the
family of three level designs: the Hoke D6 design, the Box-Behnken design
and the complete factorial design. The Hoke D6 design and the Box-

Behnken design were used to verify and replenish the complete factorial

design.

A 34 factorial design seemed to be the natural choice for this
experimental situation, for the following reasons: (1) The factorial design is
an efficient method of experimentation, particularly in the primary research
stage for taking a complete view of the mechanism of the system. (2) It
provides a measure of interaction between the controlled variables. (3) An
additional check, lack of fit, for the inadequacy of a linear model to
represent the data can be provided by adding a center point to the design.

(4) It allows the experiment to proceed with greater sophistication sequen

128

tially, if necessary, later on. (5) The experiment may be kept within a
practical size limit by running the treatment combinations in balanced blocks
(Box, Connor, Cousins and Davies, 1956). Therefore, a well-executed,

unreplicated, full factorial experiment as the essential design is presented

and discussed in this chapter.

The independent variables (5,) were changed to coded variables

(xi) for practical convenience by the following linear equation:

x. = 2(81 - £1) (4.4)
I d.

g, = actual value in original units
5: = mean of high and low levels of £1

at = spacing (difference) between the low and high level of g!

The independent variables (g!) , the coded variables (x) and their

levels are presented in Table 7.

The decision for the levels of the independent variables was based on
preliminary work done by the author and Dr. Srivastava (Wang, Dec. 1991

and Wang, Mar. 1992) and onion peeling plant practical experience.

A sample size of 30 onions was used for each experimental run. In
the experiment, well prepared onion samples were run in randomized
blocks. The blocks were built according to different onion shipping and
storage conditions. First, the sample was weighed before feeding into the
machine for recording the initial weight, Wi. Then, the sample was

randomly fed into the machine. After the machine peeling, the machine

129

Table 7
Independent Variables and Their Levels

 

 

 

 

 

 

 

 

 

 

 

 

 

 

“figfiém IL Coded Uncoded n Coded Uncoded
l Oval (d/h < = 0.8)
Onion Shape x1 Shape O Round(0.8 < d/h < 1.2)
-1 Flat (d/h > = 1.2)
1 Large (95 mm)
Onion Size X, Size 0 Medium (82.5 mm)
-1 Small (70 mm)
1 620 (kpa)
Air Pressure x, Pressure 0 517 (kpa)
-1 414 (kpa)
1 100 (onions/ min)
Chain Speed x4 Speed 0 80 (onions/ min)
-1 60 (onions/ min)

 

 

 

 

 

 

 

peeling weight, Wm was recorded. Then the onions were inspected and
additional peels were removed as necessary by hand. Finally, after the hand
peeling sample was weighed for the final weight, W,. The response values
of Wm, B and C, were computed from the data collected in the experiment
using equations 4.1, 4.2 and 4.3.

4.4 Results and Discussion
4.4.1 Model Fitting

The Response Surface Design routine of the Statistical Graphics
System (STATGRAPHICS Plus, 1992) was used to fit the second order
polynomial equation. The experimental data and response surface design are

shown in Appendix D.

130

A computer interfaced backward elimination selection procedure was
conducted for selecting the "best" regression equations. The two terminated

polynomial equations are shown as following:

Machine Peeling Loss Estimating Equation:

MLoss = 26.27 - 1.59 x2 + 5.86 x3 — 5.58::4
+ 1.04 x11:2 - 1.13 )ch3 - 2.46 x15:4 - 2.73 x2x3

- 0.87 x,x, - 1.66 xi + 6.55 x3

(4.5)

Machine Peeling Efficiency Estimating Equation:

E = 80.21 - 1.17 x2 + 5.34 x3 - 7.45 x4
- 3.60 x1):4 -1.6l x2x3 + 1.38 :ch4 + 2.6lx3x4 (4.6)

- 2.50 x12 + 3.68 x,2 + 1.64 x}

The results of the analysis of variance for the two terminated response

models are presented in Table 8 and Table 9, respectively.

From the tables, one finds that almost all the remaining independent
variables and their interactions are significant at a high level. The values
of R2, R3 and MSR were adjusted to the best conditions. The treatment of
blocking was significant only in the model of machine peeling loss.
However, the blocking for the model of peeling efficiency still exhibited
some effects to improve the model, so they still remain in the model. Onion
shape (x,) was eliminated, since it was not significant independently in
bothmodels. But, its significance effectiveness appeared in combination with

other variables. Some variables and their interactions were still kept in the

 

13 1
Table 8

 

       

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MS F
136.3267 11.83:?
x3: Pressure ll 1853.86963 1 1853.8696 160.45 ***
x4: Speed 1682.25852 1 1682.2585 145.59 ***
xlxz 38.85444 1 38.8544 3.36 * 0.0711
xlx, 45.78778 1 45.7878 3.96 ** 0.0505
xlx4 218.05444 1 218.0544 18.87 *** 0.0000
xzx3 268.96000 1 268.9600 23.28 *** 0.0000
sz4 27.21361 1 27.2136 2.36 * 0.1295
x,2 49.55654 1 49.5565 4.29 ** 0.0422
x,2 772.68173 1 772.6817 66.87 *** 0.0000
block 43.65932 1 43.6593 3.78 * 0.0561
block 4.56691 1 4.5669 0.40 0.5384
Total Error 785 .70552 68 11.5545
Error (Com) 5924.74321 80 * Significant at the level of 0.1.
% Variability explained (R2) = 88 ** Significant at the level of 0.05.
% of R3 (adj. for d.f.) = 84 *" Significant at the level of 0.01.

 

 

 

models, even though their P-value was greater than 0.05 (such as, x2x4 and
x,2 in peeling efficiency model). Because eliminating them would decrease
the value of R3 and increase the value of MSR. The results presented in the
tables of the analysis of variance will be further discussed in section 4.4.3.
These two estimated models were used later on by computer to generate the

response surfaces and their contour plots.

132

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 9
74.20167 I 4.01
10,: Pressure 1537.06685 1 1537.0669 39.70 *** 0.0000
x,: Speed 2994.15574 1 2994,1557 77.33 *** 0.0000 '
x,x, 467.28028 1 467.2803 12.07 *1“ 0.0009 n
xzxa 93.12250 1 93.1225 2.41 * 0.1256 l
x,x, 68.89000 1 68.8900 1.78 0.1867 ||
x,x, 245.44444 1 245.4444 6.34 ** 0.0142 11
x,2 112.33340 1 112.3334 2.90 * 0.0931
x2z 243.71414 1 243.7141 6.29 ** 0.0145
x} 48.12895 1 48.1290 1.24 0.2688
block 21.63358 1 21.6336 0.56 0.4653
block 11.41358 1 11.4136 0.29 0.5947 II
Total error 2632.84352 68 38.7183
Total (corr.) . 8582.19580 80 * Significant at the level of 0.1.
% Variability explained (R2) = 69 ** Significant at the level of 0.05.
% R.2 (adj. for d.f.) = 64

 

4.4.2 Diagnostic Checking

 

The models of peeling efficiency and machine peeling loss were che-
cked in two aspects: (1) the relationship between the size of the residuals
and the expected value of the responses; and (2) the normal distribution of

residuals. The results are shown in Figure 42 (A through D).

First consider the plot of the residual and the predicted value.

133

Through a careful examination of each plot Figure 42 (A and C), that the
model of peeling efficiency predicted the medium and the high values quite
well, whereas, there were some slight larger residual values appearing in the
area of low peeling efficiency. In the model of the machine peeling loss
(plot C of Figure 42), only a few wild points irregularly appeared. In
general, however, no specific pattern could be found between the size of the

residuals and the estimated value. This indicates that the assumption of

residuals distributed independently was tenable.

The normal probability checking, Figure 42 (B and D), shows some
slight evidence that the parent distribution in the model of machine peeling
loss (plot B of Figure), was not normal and probably had a slight heavy-
tailed distribution at the left side of the curve and a light-tailed distribution
at the right side. However, most residuals were distributed normally in
Figure 42 (B and D); they are very closed to a straight line. This

demonstrated that the assumption of normality of residual distribution was

tenable.

Furthermore, the observed response value vs. the predicted value
plots, the residuals vs. the run order plots, and the residuals vs. the
individual factor plots were performed by screen checking on computer. No

defective evidence has been found from checking these plots which indicates

that the estimated models should not be acceptable.

Summarizing the diagnostic checking, and also, considering the results
of the analysis of variance, R2, R3, MSR, P-value and F-value (presented
in Table 8 and Table 9), as well as the lack-of-fit checking in the Box-
Behnken design (presented in Appendix C), the conclusion is that these two

134

223.: 83:58

2: 3 33 Eat—ooze oumecwflv 25. N9 9.sz

 

 

>ocm_o_tm £328”.
a I DI

 

U

 

 

 

 

 

 

Da 0.- ..7 ma:
.... a ,. a ..l< : a . t. ”id.-
«
6 m
w
on n
pl.
on u.
A
on 9
d
9
ma m
a
no W
>oco_o_tm 9:83 ..2 5E .mEBZ
EmaEmem
on .1 on on on o
l .. 1. ..
I . 1. 1 .H‘ I1. 01
H . . H w
1 .1 . . 1 an my:
r I m
, r . a
I .. . . .. .. . . l a D.
x . . . 1. .
. i
r I
.I m P

 

 

 

>oc2oEm 9586 .8 38626 .m> 23231

335. 2323....
a N1 DI

 

 

 

 

 

 

 

 

 

v 01 «an
. .
O..— on
.111
V M n
y H .
. . .
I . .l
v , . . . +1
I .\ u . A
I I.
4 . w . .
r . . n .
I h ~ U
I .. . . 1
. v. . .
. 1
I . . . M

 

 

< 80.. 956?. 0528.2 Loe 8859.". .m> $6323”.

.1
.

IUBOJBd aAueInumg

pamgpaid

 

 

135

estimated models are accurate enough to be acceptable for further use in the

study.

4.4.3 Response Surface Interpretation

The computer generated response surfaces and their contour plots
presented in Figure 43 through Figure 48 were obtained using the estimated
models of MLoss and E presented as Equation 4.5 and 4.6, respectively.
Such three-dimensional response surfaces and associated contour plots
supplied accurate geometrical representation and provided useful information
about the performance of the system within the experimental region. An
analysis
of the surfaces and contour plots can enable us not only better understand
visually the relationships between the factors affecting the machine peeling
system, but also can further demonstrate the results obtained from the

analysis of variance.

In the presentation of response surface, every time only two
independent variables and one corresponding response can be generated by
computer. Other two independent variables were held as constants.
Theoretically speaking, they can be held at any level individually within the

experimental region. In this study, they were always held at middle-level.

Figure 43 shows the plots for the model of peeling efficiency and that
of machine peeling loss as affected by onion shape and size. From the
surface plot of peeling efficiency, it is noted that the surface slopes from flat

to oval, this indicates that the flat shape onions have a higher efficiency than

136

on? :25 :5 2:26 :25. .3 Exact: mg 82 w:=oo: 2522: .5: 5:3qu wE—ooa
he 32: 58:8 :5 88:3 8:88.. 2; mv PSME

 

 

camcw coEO 395 SEC
.350: uiat

 

 

TING

. vi-l
l...1
IT'US

 

'"IDCH
azgs uoguo

III...l...
“IF."

azgs uoguo

 

 

 

 

 

 

 

d
m a
a 2.“.
N U
U F5
9 3
% 00m
9 8m.
m.
6 com
m ...)
s m
m

 

 

 

137

slopes in opposite direction, which indicates that the flat shape onions have
lower peeling loss than the one of the oval shape onions. This is a desirable
situation, because it indicates that the optimal area is located in the area

between the flat shape and the round shape. This area has the highest

peeling efficiency and the lowest peeling loss.

On the other hand, in both models, the surface raises from the middle
to the sides, so that, the limitation (maximum) in both sides is out of the
experimental region, or, there is no practical feasibility. However, the least
peeling efficiency and the lowest peeling loss in terms of size can be found
in the middle of the surface, which means that the optimal search area for
the least peeling efficiency and the minimum peeling loss in terms of size

are in the area around the medium size onions regardless the onion shape.

The above analysis shows that for the feasible peeling efficiency and
the lowest peeling loss the optimal area in terms of onion shape and onion

size is in the area of medium size and between flat shape and round shape.

Furthermore, in the efficiency plot, one finds that the surface appears
parallel to the axis of the shape, which confirms the result of the analysis of
variance, that is, that the interaction, xlxz (shape and size), is not
significantin the efficiency model. And, the curvature of the surface is not
sensitive in terms of the onion shape and the onion size (notice the density

of contour lines) confirming that both xl (shape) and x2 (size) are not

significant.

The orientation and the distribution of the loss contour lines in plot B

indicate that the variable x, (shape) is not significant. And, the loss surface

138

raises more precipitously in both sides of the onion size than the one in the
efficiency model (notice the density of contour lines) indicating that the
variable of x2 (size) is more significant. In addition, the surface has a little
warp, thus, the interaction xlxz (shape and size) is significant. This
confirms the results of the analysis of variance. Peeling loss is significantly
affected by the onion size (this point will be discussed in more detail later),
this may be due to the ends-cutting and slitting mechanism’s ability to adopt

to a different size onion.

Figure 44 shows the plots for the peeling efficiency and the machine
peeling loss models as affected by onion shape and air pressure. Generally,
they are two slightly curved planes and are inclined in the same direction in
terms of the air pressure. Thereby, it is resulted, first, the air pressure is
very significant to both models regardless of the onion shape; second, since
the response surfaces are inclined in the same direction in terms of
pressure,the possible compromise for a choice which satisfies both models
is the medium air pressure. In addition, the F—value of x, (pressure) in the
model of machine peeling loss is 160.45, which is much higher than the one
(39.70) in the efficiency model. This confirms the above analysis that air
pressure has a telling on effect to machine peeling loss; second, that

machine peeling loss is more sensitive than peeling efficiency to air

pressurechanges.

It is noted that the machine seems to perform better on the flat-shaped
onion than on others. This impression results from the following: the
highest value for peeling efficiency is located in the corner (86%) of the

flat-shaped onions and at the highest air pressure, while the lowest value

139

2.78...— .m.. “E: 0%.? :28 .3 88b: 8 m8. ”.502— .EEQE: :5. 88658 $58:
he 32: 33:8 c...“ 8&3... 8:38.. 0.5. 3. 9:55

 

 

camcm coEO oamzw coEO
a 350..

 

 

 

 

 

 

 

 

 

 

 

«.30 3:30 0'4“. «'30 I41
1' in
.00' V
M.
.0: 9
S
S
n
...-u» M
m
.m
. . . own
0 $2 80.. 958: 9.282
.4.
98 :8 3.25. 8:5
00 v In
..., ...... ., .
'44 1on a . . . ..... , Ia W
rev o... u _. ...... 3 w
0.2.» M.
30 ., .. . «N “U:
. on d
, : 9
O“ m
U
\ ...n 5
- .I
1 3 o
x w
- m
x (

 

 

 

3

s E e
(econ BJnSSOJd le

fl 8

,.
(%) Momma Buneed

 

 

140

formachine peeling loss is in the corner (18) of flat shape onion and at the
lowest air pressure. This indicates that the machine is more suited to flat
shape onions. On the other hand, the lowest value for peeling efficiency is
in the corner (71%) of the oval shape and at low air pressure, the highest
value for machine peeling loss is in the area of the oval-and-round-shaped
onion and at high air pressure. This indicates that the machine is less suited
to the oval shape onions. Furthermore, from contour plots, it is also noted
that flat shape onions are better than others for the machine. For instance,
for certain air pressure, in plot D, say 537.6 kpa, the oval-and-round-shaped
onion has a value of 27% for peeling loss, while for the flat-shaped onion
the value is only 25% or lower, and, in plot C, one finds that the oval-

shaped onion has a value of 76% for peeling efficiency, while the value for
the flat shape onion is 82%. This feeling is confirmed by the analysis of
variance. In the loss model the interaction of x1x3 (shape and pressure) is
significant at the level of 0.05, and the second order of XI (shape) is

significant for both models.

By further examining the surface and contour plots one finds that the
optimal value for peeling efficiency is located in the area between the flat
and the round shape with medium and high air pressure. For peeling loss
the minimum value is located in the area between the flat and the
roundshape with medium and low air pressure. Consequently, the possible
optimal area for satisfying both models is between flat and round shape

onion with a medium air pressure.

Figure 45 shows the plots for the peeling efficiency and the machine

peeling loss models as affected by onion shape and chain speed. Both

141

response surfaces appear a little warped. They are diagonally inclined in a
similar direction, that is, their lowest values are located in the same area,
namely, the oval shape and the high speed area. But their highest values are
located in different areas. The highest value for peeling efficiency is located
in the region of the low speed regardless of onion shape, while the highest
peeling efficiency and machine peeling loss as affected by onion shape and
chain speed value for peeling loss is located in the corner of the low speed
but oval shape.

There are two practical inferences that could be drawn from above
observation. First, speed has an inverse effect to peeling efficiency, that is,
higher speed causes lower peeling efficiency, particularly for round- and
oval- shaped onions. When speed exceeds 84 onions/min, peeling efficiency
worsens quickly. Therefore, in practice, one should avoid operating in this
area. However, in the same area opposite happens to peeling loss, that is,
when the speed is higher than 84 onion/min. the peeling loss getting down.
Consequently, the possible conpromise for a choice which satisfies
bothmodels can be found only in the area around medium speed. If the
speed is too high, it will result in low peeling efficiency even though we
would get higher peeling capacity and lower peeling loss. If the speed is too
low it will cause high peeling loss and low peeling capacity even though we
could get higher peeling efficiency. Speed significantly affects the
responsemodels, which is also confirmed by the results of the analysis of
variance. It is significant at a level of 0.01 or higher for both models.

Therefore, this investigation suggests keeping medium or a slight high speed
to fulfill both models.

142

 

 

 

Machine Peeling Loss (%)

 

88"

  

Peeling Efficiency (%)

Onion Shape

3833.183

(%) Aouelouia Buneed

 

F

 

1 '1' .“
Round

Onion Shape

 

 

 

 

 

    

.. 1.; .l. ..l .1. ll

....--h.,uz.ru mm-.. H;
A I 1
Round

Onion Shape

 

F1.

 

i

O
0
q

a .2 s :
(Ull—U/SUOWO) needs Uleuo

Flat

 

 

f“

peeling efficiency and machine pee mg loss as affected by onion shape and chain speed

Figu

nse surface and contour plots of

re 45 The res

143

Second, the analysis of Figure 43 and Figure 44, is confirmed, that
in general, oval-shaped onions have higher peeling loss than flat-shaped
onions for all speeds, pressures and sizes. The effect of onion shape is
mainly caused by the interaction, x,x4 (shape and speed), and its second
order, x12. Even though its first order x1 is eliminated from the estimated
equations. Furthermore, the efficiency surface is warped in a diagonal
direction, the possible optimal efficiency should locate in the area nearby the

straight line of 80% in the contour plot of the peeling efficiency model.

The fact that oval-shaped onions have higher peeling loss and lower
peeling efficiency and flat-shape onions have lower peeling loss and higher
peeling efficiency may be explained by the following: (1) more top and root
potions are cut in the case of oval-shaped onions; (2) the cut portions
mentioned above are no help for increasing the peeling efficiency; (3) the

slitting is not sufficient as in the case of flat shape onions.

Figure 46 shows the plots for the peeling efficiency and machine
peeling loss models as affected by onion size and air pressure. The
response surfaces are two curved surfaces and are inclined in the same
direction in terms of pressure. The inclination indicates the compressed air
jet system is highly effective to both models. The curvature of the surfaces
indicates the interaction, x2x3 (size and pressure) is significant to both
models. And, it is noted from the plots that the hollow of the surfaces were
shifted slightly to the right of the medium size, which is caused by the effect

of onion size.

By comparing two surfaces, one finds that the loss surface is curvier

than the surface of efficiency, which indicates that the loss model is more

144

9:58.... ..ma :5 on? :25 .3 88...... 8 m8— wfimxa 0:58... 9.: 8:295“. $58..
he 82: ...—3:8 E... 88...... 8:38.. 2E. we 2&5

 

 

 

 

 
 

 

 

 

 

 

 

 

 

 

      

  

  

   

 
  

 

      

         

35 SEC 35 5:5
on... I53: .5 I?! 5:8: «one
.1 .... «n 1. 1...... ,. NH . .. . 1
. 1 W. o .3. W
M. w
. % .... a
m m
...n 9 ...-B M.
m m
u m1 e m1
.3
1.2.1.. ems;
OVA] v . one . . ...
a W r0917? u.F.. . .. kwafig?‘ 1... z. W
a N . 2.1.15... ..‘!%Q%Q%.§A 2. w.
a m... on u n ........ §§§%%% «o ..N.
. m §%%0s00 .. 1
. m. i5h§00000w¢¢4§0i0 .. ...
B .81.... ...:V Q m
#030860
m 000%00 .1.

 

 

145

sensitive than the efficiency model to size change. This confirms the result
of the analysis of variance, that the former has much higher F-value (11.80)
than the later (1.92). The two surfaces are inclined in terms of pressure.
This also confirms that the result of the analysis of variance, x3 (pressure)
was significant at a 0.01 level or higher for both models. Comparing the
curvature and the distribution of the contour lines in plots H and G, it is
noted that the effect of onion size and/or air pressure in the loss model is
more significant than in the efficiency model. This was confirmed by the
results of the analysis of variance, i.e., x2 (size), x3 (pressure) and x2x3 (size

and pressure) are more significant in the loss model than in the efficiency

model.

From surface and contour plots one can see that the feasible limited
value (minimum) for both peeling loss and peeling efficiency models in
terms of size is located in the area between medium and large. Thereby, the
feasible optimum size which satisfies both models is between medium and

large onions, i.e., 87 mm in detail.

Figure 47 shows the plots for the peeling efficiency and the machine
peeling loss models as affected by onion size and chain speed. The surfaces
are backward inclined in the same direction in terms of speed, and the
sufaces have the similar features with the case discussed in Figure 46. The
inclination of the surfaces is caused by the speed. The curvature of the
surfaces is mainly determined by the effect of the interaction, x2x4, but in
this case, the influence of the interactions to the models are much weaker,
particularly in the model of peeling loss. And, also, the hollow of the
surfaces were shifted slightly to the right of the medium size. All above

 

 

    

838833?-

Onion Size

 

 

J

  

 

 

‘ i
Lara-

 

Be
tn
0) :
O
.1 ..
g :
"3:: :5
a: .1!
D. 1!
a, ..
.E
.c
o
m
E
.0
4d
:1... 3
w

 

’2”
g a a : s :
(in/suowm paads Uieuo

 

Peeling Efficiency (96)

m1:-

 

 

._ .. . .. .
1
an“.

 

 

Onion Size

Onion Size

 

 

Figure 47 The response surface and contour plots of
peeling efficiency and machine peeling loss as affected by onion size and chain speed

147

observations are confirmed by the results of the analysis of variance.

From contour plots one can see that the possible limited value in terms
of size for both models is located in the area between medium and large.
That the contour lines in the peeling loss model appear much curvier than
the one in efficiency model indicates that loss model more sensitive to the
size change. This also confirms the result of the analysis of variance, that
is, that the size is significant at level of 0.05 for the loss model, while it is

not significant for the efficiency model.

Figure 48 shows the plots for the peeling efficiency and the machine
peeling loss as affected by air pressure and chain speed. The response
surfaces are similarly diagonally-inclined planes and are sloped in the same
direction in terms of the interaction of x3x4 (pressure and speed). Their
highest response values are located in the same corner of the high pressure
and the low speed; and their lowest values are located in the same comer of

the low pressure and the high speed.

From the contour plot of the efficiency model one can see that it is a
curved surface, which indicates that the model is affected by the interaction,
x3):4 (pressure and speed). This feature is confirmed by the results of the
variance, in which x3x4 was found to be significant at a level of 0.05,
whereas, the surface of the peeling loss model is a flat plane, and all
contours are straight lines. This feature is also confirmed by the results of
the analysis of variance for the loss model, in which x3x4 was not

significant.

148

 

 

   
 

Air Pressure (KPa)

3355833

(%) 3501 Buileed eugqoew

  
   
   
   
    

’1’...
1".

...v‘ -..
. O

’-
- a- 7

AirPressure (KPa) ea - -'

. Q
' it
'

83333.13

(96) A9U9F’UE Bugleed

Machine Peeling Loss (%)

Peeling Efficiency (%l

 

 

    

 

.. a
..E...z....~...

 

0‘
0.4.9

a w
0 O
I

1.. —

3 3 o
(aim/snow) paads meuo

 

 

IrerlrnyerI

 

 

 

no —'

3 3 2 8 3

(vim/suowm peads Uieuo

570.0

537..

Air Pressure (KPa)

‘1‘

578-0

490.4 631..
Air Pressure (KPa)

455.2

1‘

 

 

Figure 48 The response surface and contour plots of
peeling efficiency and machine peeling loss as affected by air pressure and chain speed

149

4.4.4 Response Surface Analysis

One of the objectives of this study was to find a set of operating
conditions which could optimize the responses, namely, to search a set of
values from chain speed, air pressure, onion shape, and onion size which
could maximize peeling efficiency and minimize machine peeling loss. Prac-
tice and experiments indicated that peeling efficiency and machine peeling
loss are two mutually conditioned responses. They are both opposite and
complementary to each other. Therefore, this is a multiple response
problem. The computer generated second order polynomial equations, as
shown in Eq. 4.5 and Eq. 4.6, were used as empirical models for searching

the optimum in the experimental region. Peeling capacity response was

considered as chain speed in the optimal analysis.

The optimal condition searching procedure consisted of: (1)
calculating the stationary points (points of zero slope or points where the
first derivative is zero); (2) performing a canonical analysis, and (3)

superimposing corresponding contour plots.

The method of calculating stationary points was introduced in detail
by Myers (1971) and Draper (1963). In this study, MATHEMATICA
(1991) was used to calculate the stationary points and find the characteristic
root of the B-matrix for building canonical forms. The stationary points are
presented in Table 10. Something can be learned from the stationary points.
In the model of peeling efficiency, the stationary point indicates that the
optimal onion shape is between oval and round, and justshort of round. The
optimal size is 100 mm. The air pressure is 1000 kpa. The chain speed is

56 onions/min. In the peeling loss model, the stationary point indicates

150

Table 10
Stationary Points for Response Models

 

 

 

 

 

 

Stationary Points 7"

Variables Coded Experimental Region E fficiency . Loss I!

t Onion Shape from -1 to 1 0.698 - 1.476 .
Onion Size from -1 to 1 1.349 -0.609
Air Pressure from -1 to 1 "4.658 -5 328
Chain Speed from -1 to l - 1.243 3.988

 

 

 

 

 

 

that the optimal onion shape is flat. The optimal size is 76 mm. The air
pressure is O kpa. The chain speed is 160 onions/min. The values of shape
and size still remain in the experimental region, but, speed and pressure are
totally outside it. Particularly, the pressure in the two models seems
antithetical to each other. Subsequently, it is noted that in both cases, the
four-dimensional stationary points are located outside the experimental
region. Therefore, from the point of view of practical feasibility, it is
necessary to move away from these points and to locate the optimum inside

the experimental region.

Two B-matrices can be obtained from the estimated quadratic
polynomial equations. They are as following:

Peeling efficiency B-matrix

' -2.49815 - 11 0.32917 -0.25139 -l.80139 1
0.32917 3.67963 - 1. -0.80417 0.69167
0.25139 -0.80417 0.04630 - 1 1.30556 z 0
—l.80139 0.69167 1.30556 1.63519 - 1 .

 

 

151

Peeling loss B-matrix

 

' -1.65926 - 1 0.51945 0.56389 -1.23056 ‘
0.51945 6.55185 - 1 -1.36667 -0.43477 = 0
0.56389 -1.36667 0.78519 - 1 -0.10278

1 -1.23056 -0.434722 —O.102778 0.04074 - 1 .

 

Expanding the determinants and solving the two quartic equations
yields the characteristic roots of the B—matrix. Then, the canonical form

results in the following equations for peeling efficiency and machine peeling

loss.
7, = 96.6 - 3.25 w? - 0.72 w: + 2.89 w: + 3.94 w}
13,, = -1.55 - 2.40 w? + 0.13 w: + 1.08 w: + 6.91 w:
Where,

W1, W2, W3, and W4 (eigenvalues or coefficients) are linear

combinations of the xi.

From the equations, it is noted that the characteristic roots
(eigenvalues or coefficients) have mixed signs. This indicates that the
stationary points exist in terms of saddle points, which in turn suggests that
movement away from these points would cause an increased or decreased

response, depending upon the direction of the movement (Myers, 1971).

The location of the stationary points as well as the form of the
canonical equations suggest that a complicated ridge system exists for both

of the response surfaces. Analysis of such a ridge system is not easy to

152

perform and is not always successful, especially when multi-response
problems are involved in the system (Floros and Chinnan, 1987). Hence,
a simpler approach, computer graphical superimposition approach, was taken
to further explore and explain the system. A series of contour plots of equal
response were generated by computer which provide useful information for

further investigation.

4.4.5 Graphical Approach

In the graphical approach, the estimated models were used to create
superimposed contour plots within the experimental region by the computer.
The superimposed contour plots are usually generated on two-coordinate
diagrams. These plots present information for two factors and one or more
responses and are reasonably accurate (depending upon the representational
accuracy of the model) within the experimental region. The regions of
optimum response(s) are judged by visual inspection to the superimposed
contour plots. This method reduces the possibilities of "unrealistic"
solutions, since the regions within the experimental space are examined, and
allows simultaneous optimization of several competing responses by simple
superimposition. Therefore, this method particularly suits the situations in
which optimal conditions are searched in a multi-response system and the
stationary points are outside the experimental region.

Numerous successful applications of graphical Optimization pertaining
to mechanical and food systems were reported in the literature (Kissell,
1967; Kissell and Marshall, 1962; Floros and Chinnan, 1987, 1988; Johnson
and Zabik, 1981; Lind, Goldin and Hickman, 1960; Henselman et al., 1977;

153

Myers, 1985; Oh, Seib and Chung, 1985; Terhune, 1963; Underwood,
1962; Wilson and Donelson, 1964; Wu, 1964; Wu and Meyer, 1964).

In this study, the optimization of peeling efficiency and machine
peeling loss was realized by superimposing a series of computer generated
contour plots. The procedure of superimposition was performed in three

sequential steps.

First Step: In the first step, as shown in Figures 43 through 48, two
response surfaces and their contour plots were presented in terms of
independent variables with six different combinations. They explored the
MSU machine peeling system with two variables at a time and showed the

responses of peeling efficiency and machine peeling loss separately. A

general interpretation was taken to the graphics in section 4.4.3.

Second Step: In order to analyze the two response models
simultaneously and to find a set of operational conditions satisfying both
models, in the second step, two corresponding contour plots in terms of
peeling efficiency and machine peeling loss were superimposed on each
other from Figures 43 through 48. The results are shown in Figure 49-A
through 49-C.

Figure 49-A is the superimposition of shape vs. size as well as
pressure vs. speed. The shaded area in plot M of Figure 49-A indicates the
region of onion shape and onion size in which the peeling efficiency is
between 80% and 81%, and the machine peeling loss is between 25% and
27%. The shaded area in Plot R shows the region of air pressure and chain

speed in which the peeling efficiency is between 79% and 82%, and the

154

machine peeling loss is between 23% and 26%.

Figure 49-B is the superimposition of shape and size vs. speed,
respectively. The shaded area in plot 0 is the region of onion shape and
chain speed in which the peeling efficiency is between 80% and 82%, and
the machine peeling loss is between 24% and 25 %. The shaded area in plot
Q shows the region of onion size and chain speed in which the peeling
efficiency is between 79% and 81 %, and the peeling loss is between 24%
and 26%.

Similarly, in Figure 49-C, the shaded areas in plot N is the region of
onion shape and air pressure in which the peeling efficiency is between 80%
and 81%, and the peeling loss is between 25% and 27%. In plot P, the
shaded area shows the region of onion size and air pressure in which the
peeling efficiency is between 80% and 82% , and the peeling loss is between
24% and 27%.

Third Step: In the third step, contour plots in terms of speed and
pressure were further superimposed by computer from O and Q in Figures
49-B for speed and from N and P in Figure 49-C for pressure. The results
are shown as S in Figure 49-B and as T in Figure 49-C. In these plots, S
and T, the immaterial contour lines were eliminated for clarity, the small
cross-shaded areas show the region of optimal feeding chain speed and air
pressure which satisfies the conditions of the peeling efficiency being about
80% and the peeling loss being about 25 %. It is noted from plot S in
Figure 49-B, the optimal chain speed is around 84 onions/min., the optimal
onion shape is between flat and round shape, and just short of round. From

plot T in Figure 49-C, the optimal air pressure can be found in the region

155

 

 

M: A superimposed over B

Larg-

2:!

Onion Size
flndaun

 

Small

Flat Round 00.1

Onion Shape

R: K superimposed over L
' ' 72 3’ f 71" I 17 ' '

 

 

100

 

 

92

 

Chain Speed (Onions/min)

 

414 455.3 498.4 637.8 578.8 629

Air Pressure (KPa)

 

Figure 49-A Superimposed contour plots for two response models

 

156

 

 

 

O: E superimposed over F

.-
C
O

_--'/

0
ID
I

 

. 3 . . e .
as as . N-

I i . . 7
5° L . 2g .... .. H19 ... . 3 ‘ . . L .. 1..

1

Flat Round Oval

Onion Shape

 

 

 

Chain Speed (Onions/min)

 

 

 

Q: I superimposed over J

,9. . ..... . ‘ .. . .‘ .. .. . . 5.-

 

.—
.
.

Chain Speed (Onions/min)

 

 

 

 

 

Onion Size

S: O superimposed over 0

Chain Speed (Onions/min)”

 

Flat Round Oval

Onion Shape

 

Figure 49-B Superimposed contour plots for two response models

 

157

 

 

 

N: C superimposed over D
5 SE a: 1 BL sq

 

  
     

   

Onion

B
578.8

 

 

 

 

EST-6 ‘
I
498.4

456.2 _

 

 

Air Pressure (KPa)

 

 

 

Flat Round Ou-l

Onion Shape

 

P: G superimposed over H

 

 

 

537.3

498.4

Air Pressure (KPa)

466.2

 

 

 

 

 

 

 

 

Hadlun Larg-

Onion Size

T: N superimposed over P
628 : 81

578.5
537.6

493.4

Air Pressure (KPa)

465.2

 

Small H-dlum Lara.

Onion Size

 

Figure 49-C Superimposed contour plots for two response models

 

158

between 517 kpa and 537.6 kpa and the optimal onion size is medium.

This conclusion can be verified from plot M and R in Figure 49-A.
In plot M, the small square represents the region of the optimal shape and
size obtained from the plots S and T in Figure 49-B and C, it is just located
in the shaded area of plot M in Figure 49-A. In plot R of Figure 49-A, the
small square represents the region of the optimal pressure and speed
obtained from the plots S and T in Figure 49-B and C, it is overlapped on
the shaded area of plot R in Figure 49-A.

According to the above conclusion, at the speed of 84 onions/min for

round and medium size onion, the optimal peeling capacity for the machine

is computed as 756 Kg/h.

The Hoke D6 design (Hoke, 1974) meets the Wheeler (1972) criteria
and has high efficiencies. Its economical feature particularly suits research
that includes independent variables greater than three, and, therefore, was
suggested by Thompson (1982) and Lucas (1976). In this experiment, only
19 experimental runs were required to obtain the highest value of coefficient
of multiple determination (R2). The value of R2 of the Hoke D6 design was
0.88 for peeling efficiency and 0.94 for machine peeling loss, respectively.
The general conclusion of the Hoke D6 design was the same as the factorial
design. But, in some aspects, the Hoke D6 design resulted in a better
conclusion than did the complete factorial design. For example, the optimal
chain speed of 84 onions/min obtained from the Hoke D6 design, resulted
in a peeling loss of lower than 24% combined with a peeling efficiency of
higher than 86%. This value (optimal speed) was more precise than the one

obtained in the factorial design. More detailed information can be found

159

in Appendix B. One suspects there may be a possible disadvantage is that
there is the bias due to the existence of higher order terms. However, for
this four factors design it should not be a problem. But, as a preliminary
investigation using only 19 points, finding 15 parameters and 2 responses

seems too week to support the conclusion.

The Box-Behnken design was also tried for the analysis of variance.
The advantage of the design is that it needs only a total of 27 runs and the
lack of fit can be tested by adding only two extra central points. The
analysis of the fitting-test indicated that the lack of fit was not significant at
5% for both models of machine peeling loss and peeling efficiency. The R2
values were 0.69 for the peeling efficiency model and 0.88 for the machine
peeling loss model, respectively. The experiment layout and the results of
analysis of variance by the Box-Behnken design are presented in Appendix
C.

CHAPTER V

SUMMARY AND CONCLUSIONS

5 .1 Summary

In this study, the MSU second prototype onion peeling machine

and its performance were evaluated qualitatively and quantitatively.

In the qualitative evaluation, the background of onions’
economic significance as well as their products and manufacturing methods
were reviewed; onion peeling methods and equipments which were granted
U.S. Patents in the last 100 years (1890-1990) were investigated and
summarized; and the applications of the Response Surface Methodology in
mechanical systems and food engineering were selectively presented. Based
upon the above information, the technology process and the construction of

the MSU onion peeling machine were evaluated.

The MSU machine is farm level onion peeling machine. It was
designed to peel pungent Michigan onions and performs an eight stage
classical peeling technology in a manner of continuous flow. The machine
consists of six major mechanical systems. It applies a novel approach of
using four scoring blades assisted by compressed air jets to make four

mutual perpendicular slits in the outer layers of the onion skin. The

160

161

compressed air penetrates through the slits and causes the skins to loosen
and/or dislodge from the onion bulb. A pair of parallel blades rotating at
a high speed trim the ends of the onion. A pair of Co-rotating rollers
assisted by three compressed air nozzles are utilized to spin the scored and
trimmed onions for final separation and removal of the onion peels.

In the quantitative evaluation, the machine performance was
characterized by peeling efficiency, machine peeling loss and peeling
capacity. Onion shape, onion size, air pressure and chain speed were
investigated as major factors influencing the machine’s performance. Two
sequential statistical experiments were taken to the MSU onion peeling
machine. Three types of Response Surface Design (Hoke D5, Bos—Behnken
and factorial) were conducted to test the machine’s performance and
properties. Two second order polynomial equations were created and
modified by computer as the empirical models to simulate the performance
characteristics (peeling efficiency and machine peeling loss) and their
explanatory variables (shape, size, pressure and speed). Computer
generated response surfaces and contour plots as well as the analysis of
variance were used to analyze the machine’s performance. In addition,
canonical analysis was applied in judging the location of stationary points.
Because searching for a maximum peeling efficiency and minimum peeling
loss is a multi-response problem, and also, the stationary points were outside
the experimental region, consequently, a computer graphical superimposing
method was used to locate the optimal operating condition in the

experimental region. The obtained optimum satisfies both of the models of

machine peeling loss and peeling efficiency simultaneously .

162

5 .2 Conclusions

The results of machine performance testing and statistical
analysis indicated that the MSU machine is an efficient onion peeling
machine. For round shape and medium size (82.5 mm) pungent Machine
onions, the peeling efficiency of 80% or higher is obtainable while keeping
the machine peeling loss as low as 25%, and an average peeling capacity of

576 kg/h or higher is possible.

The results of the experiment also indicated that the air
pressure, the feeding chain speed and the interaction of onion shape and
feeding chain speed are highly significant for both models of peeling
efficiency and machine peeling loss. The onion size as well as the
interaction of onion size and air pressure significantly affected the model of
machine peeling loss, whereas the interaction of air pressure and feeding

chain speed is significant to the model of peeling efficiency.

Because the MSU onion peeling machine possesses compact
construction, reliable conveying and holding system, passive performing
components and full hydraulic power and control system as well as
performing successive flow processing procedure, it also is a safe and
economical farm onion peeling machine. Particularly, the application of the
novel air jet nozzles greatly increases the peeling efficiency. The self-
regulating technique relevantly applied in the machine system greatly
enhances the peeling adaptability.

CHAPTER VI
RECOMNIENDATIONS FOR FURTHER

RESEARCH

(1) The adequacy of the model equations should be examined in the

onion peeling plant.

(2) If new experimentation needs to be conducted for the machine,

the replications should be taken to each testing point.

(3) It would be valuable to see if grading onions in size first then
feeding them to the machine with different chain speeds will decrease the

machine peeling loss.

(4) It would be also valuable to test whether the graded-onion helps
to improve the feeding or loading quality, thereby increasing the peeling

efficiency and decreasing the machine peeling loss.

(5) If the moisture and the storage time can be considered as factors
in an experiment, it would be helpful to discover their influence on the

peeling efficiency, and, in turn, it would be profitable to handle the pre-

processing conditions in onion peeling.

163

APPENDICES

APPENDIX A

MSU Onion Peeling Machine First Experiment

The ANOVA Table and The Raw Data

Table A.l

The First Experiment: The ANOVA Table for The Peeling Efficiency
(Air Pressure = 75 psi)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Effect __ d.f. ss MS F
Total (corr.) 8 1381.6290 I
x,: Chain Speed “1 900.37500 900.37500 91.91 ...... II
x,: Onion Size 1 11.20667 11.20667 1.14 II
x,x2 1 14.06250 14.06250 1.44
x12 1 204.69390 204.69389 20.90 **
x; 1 221.90222 221.90222 22.65 ** II
Total Error 3 29.38861 9.79620== II
R2 = 0.978729 122 (adj. for d.f.) = 0.943277
** Significant at the 0.05 level.
Table A.2

The First Experiment: The ANOVA Table for The Total Peeling Loss
(Air Pressure = 75 psi)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

II Effect __]l d.f. ss MS F I
Total (corr.) 1 8 104.16000 I
x1: Chain Speed 1 4.1666667 4.166667 1.67 H
Xi: Onion Size 1 42.135000 42.135000 16.89 **
xix2 1 44.890000 44.890000 18.00 **

x,2 1 1.2800000 1.280000 0.51

x; 1 4.2050000 4.205000 1.69
Total Error 3 2.4944440 2.494444 =

R2 = 0.928155 R2 (adj. for d.f.) = 0.808414

** Significant at the 0.05 level.

164

The First Experiment:

Table A.3

The Experiment Latout and The Raw Data Table

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

E“ Size Speed Pressure I] w, Wm w, J] E wL
1 ‘ L M M " 25.4 20.8 17.0 '1 55 33.0
2 M M M 34.6 26.2 21.0 62 39.3
3 S M M 27.2 21.8 18.4 61 32.4
4 M S M 43.6 30.6 28.4 86 34.9
5 M M M 36.4 28.6 22.4 56 38.5
" 6 s s M 44.0 30.0 25.4 82 42.3 """
7 M M L 38.4 30.4 25.4 62 33.9
"""" 8 M M H 41.2 31.6 25.0 59 39.3
""" 9 L s M 29.0 20.6 18.6 81 35.9
10 S S M 27.2 20.0 18.4 75 32.4
11 M F M 41.4 31.6 26.8 67 35.3
12 L F M 37.4 29.0 20.6 50 44.9
1.13—l. S F L 28.2 25.8 19.4 27 31.2
Notes Onion Size Air Pressure Chain Speed
8: Small = 70 mm L: Low = 65 psi S: Slow = 10 M/min
M: Medium = 82.5 mm M: Medium = 75 psi M: Medium = 13 M/min
L: Large = 95 mm H: High = 85 psi F: Fast = 24 M/min

165

APPENDIX B

The MSU Onion Peeling Machine Second Experiment
The Hoke D6 Design
The Design Layout and The AN OVA Table

Table 8.1

The Second Experiment: The Experiment Layout and Raw Data
Table of The Hoke D6 Design

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Shape Size Pressure SpeerflIMachine Lo; Efficiency II

1 -1 o 0 0 " 23.4 66.7 '
2 0 -1 0 0 39.6 93.3

3 0 0 -1 0 21.2 87.8

"4' 0 0 0 -1 27.1 95.8
5 i -1 -1 -1 -1 28.3 72.2
6 -1 1 1 1 26.3 79.0

7 1 -1 1 1 27.9 77.3

8 1 1 -1 1 17.5 68.6

" 9 1 1 1 -1 44.3 91.2

10 1 1 -1 -1 33.3 88.2

11 1 -1 1 -1 45.2 96.6

12 1 -1 -1 1 19.4 60.0

13 -1 1 1 -1 40.7 90.6

14 -1 1 -1 1 20.9 70.6

....15 -1 -1 1 1 44.6 89.1

16 0 1 1 1 28.2 84.1

17 1 0 1 1 21.9 77.8
....1§ 1 1 0 1 24.3 81.0
.2. 1 1 1 0 27.0 82.2

 

 

166

Table B.2

The Second Experiment: The ANOVA Table for The Peeling
Efficiency (Hoke D6 design)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

II Effect II ss DF MS F P II
x,: Pressure I 638.209991 1 638.209991 25.56 0.0003 I
x4: Speed 603.107646 1 603.107646 24.16 0.0004
x,x, 26.333999 1 26.333999 1.05 0.3247
x,x, 241.942343 1 241.942343 9.69 0.0090
71,2 510.316822 1 510.316822 20.44 0.0007
x22 62.468149 1 62.468149 2.50 0.1397
Total error 299.606673 12 24.96722
Total (corr.) 1988.37789 18
R2 = 0.849 8.2 = 0.774
Table B.3

The Second Experiment: The ANOVA Table for The Machine Peeling
Loss (Hoke D‘i design)

‘7

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

II Effect II ss DF MS F P I
I x1: Shape 41.599238 1 41.599238 2.76 0.1407
x2: Size 70.289409 1 70.289409 4.66 0.0677
x,: Pressure 372.488433 1 372488433 24.70 0.0016
11,: Speed 200.746235 1 200.746235 13.31 0.0082
x,x, 47.054408 1 47.054408 3.12 0.1207
x,x, 32.437576 1 32.437576 2.15 0.1859
x2x3 37.053878 1 37.053878 2.46 0.1610
x271, 41.923738 1 41.923738 2.78 0.1394
x,x, fl 30.012514 1 30.012514 1.99 0.2012
x,2 29.552949 1 29.552949 1.96 0.2043
x} 193.223458 1 193.223458 12.81 0.0090
Total error 105.554954 7 15.07928
Total (corr.) 147422105 18
R2 = 0.928 R} = 0.816

 

 

 

 

167

dwfioc an axe: 2: E 9% :25 BE 2.2—m :35 .3 “Seats
3 m8. 958.. 2::an 1.5 55659 $58.— .3 $2.— .5355 can 88.5w 8:38.. 2:. Tm «2&5

 

 

camcm coEo 255 SEC
3.508 in Ulul

dial «'3

 

 

   

 

 

«I30

. . .4

.1" Ar

1 1

0 .
Um. .m
0 1am.
mu ”mu
nnlb .13
2 1mm.
9 r. 8

1.1

1-

a

w

 

 

 

 

 

 

 

$3 3:225 9:86

. .8 395 coEO
.. p.30:

  

w w
a P9
«0 Hi.
u. U
«m. 2.5
N9 B
an

:3. ......
WP 009.
”I. U
U

5 00M
1 v.)
n

O

s m
S

m

 

168

 

:wfioc .9 8...: 2: E 9:583 ha can 23% :35 .3 138%“
m5 m8— w—Euon 2.392: 25 55650 3qu .3 $2.— .ESEE Ea oust—5. 8:88.. 2:. Ni 9.53%

 

 

82m coEO
use: «1.. 326 5:5

4
F p

«'30

 

 

I D 1‘ V
(9:1)!) SJnSSBJd 1w

5 8 3 9
(edx) GJnSSGJd ng

 

 

 

 

 

 

 

 

 

     

...,...... . \\\\ OP
. .:‘\\ 3
g
I‘-W\\\\
I... \ «a

(%) Aouesoma Buuaed

 

      

(%) sso1 Bunaad eugqaew
8 8

 

 

169

 

 

 

Machine Peeling Loss (%)

 

 

Onion Shape

; ‘i‘
Hutu

 

Plat

..i... AAAL

 

0
F

O
”(in/suowm paads ugeuo

 

Peeling Efficiency (%)

Hutu Inh' .
an.

Onion Shape

 

 

 

...liii ...

 

1.. _L.H

(%) Aouepma Buueed

 

O P O
(in/suowm P99dS Uieuo

 

affected by onion shape and chain speed in the Hoke D6 design

f" . i'
flulfl

Onion Shape

 

F1.

Figure B-3 The response surface and contour plots of peeling efficiency and machine peeling loss as

 

5?... on use: 2: 5 0.539:— .._a ER on? :25 .3 teapot“
8 m8. 958.. 2:..an was 5:3qu wagon he 32.— .5859 EB 8&5; 8:38.. 25. win PEwE

 

 

cum coEO

 

 

      

     
 

 

 

 

 

0N5 COEO
sssss lino: ... 4!...» .0le line: 3!.»
M. . DMZ , . 4 lb V . . _
, 1. ~6va .23 W
i la la
. H W
1.001 9 m S
s . s
n ., n
I. . In
0.59 . . , ,2. \. 1 ....“ 3
) . . H mm
a . w . d

 

 

 

 

 

 
   

«new v...” ....... .. :33:
9a a m3...) \k‘:: =8.»
0v v.03 ., . ‘
«a, . h§§0 ..
0.2.. ..
O . _

  

        
 

ififioefo.

        

(%) Aoueioma 5U!l99d

W

. .. .«mwflflhgooo‘ .. ..
\ggOOg ... m
\WM“..%§MIN%0O'O" M
r m

 

 

 

171

 

 

 

Lara-

; i
findin-

mdluli ' ..

Onion Size
Machine Peeling Loss (%)

  

 

. A .1 IL .1 Al .I .. .l i. ii

Snail

 

 

 

 

" " " ° “’ " ‘3 3 l! 8 3
‘ (%) sso1 Buuaad euguoew (ugw/suoguo) paads ugeuo

 

J
Larg-

 

 

Peeling Efficiency (%)
m.“- ; . - .
Onion Size

1a; o
'.=' .L‘
(I)
,5:
”.9
:5:
f O

   

 

 

Small ‘

 

i...1...i...i...i...i
g 8 8 3' 3 8 8 g g 3 ,2 g g
(%) Aoueiouia 5U!l99d (U!w/SUO!UO) P99dS Uieuo

 

172

Onion Size

Figure B-S The response surface and contour plots of peeling efficiency and machine peeling loss as
affected by onion size and chain speed in the hoke D6 design

 

:38: on 3.5 o... E .50.... Ease ...... 2.58.... ....w .3 389......
we «we. 3:9... 9.....an ...... 55.95.. 2.28.. .... 82.. 53:8 ...... 8&2; 8:88.. 25. cum 95$..—

 

 

.mmv: 2:32... :4.

ONO P40

 

 

pl

   

 

 

$3 mwo. uc=mmm 2.2022

 

on
3
u.
on w.
u
S
2. M
S
D.
vo m
m.
0
«a w
/
w.
..«iUi

3 S 3 3 8 a S
(93) 5301 Buuaad euiqoew

Ens: Summoi .2.

.N0 Pun

 

.‘3 3 8
(in/suowm peeds Uieua

3

  

 

 

 

 

 
   

.. ww. .. _Nnr . 00d
33 6:225 8:8“.
co. . . ., 3%: 2330.5 :4
z . ; ,. ., ......
Ivy/”0060 v...
o

             

i.
l
l.
. .
I
i
I

l

.....- \‘i‘i‘- ---
..' - 'i.-l-“§-l.‘\-
--.. . .‘ii. 'i 1-- .-..
...\~': in-.. ii iii
...-x.-i‘|n1i.'t 5‘.
......~.D‘L'..‘
---- -----.NILKI

; a" 5 8 5 a F
(%) Aoueiama Buyieed

 

 

173

APPENDIX C

The MSU Onion Peeling Machine Second Experiment
The Box-Behnken Design
The Experiment Layout and The ANOVA Table

Table C.1
The Second Experiment: The Experiment Layout and The Raw Data
Table of The Box-Behnken Design

Size Pressure

0 22.5
29.6
28.3
23.3
37.7
37.9
33.3
29.4
23.2
19.5
20.8
26.5
21.9
23.7
41.1
37.0
18.3
26.2
31.0
21.1
23.5
32.6
35.1
27.4
16.7
41.7
24.3

a:
5

Chm-5mm}—
HOOHOHO
I
V-‘O

FLOt—i
Loooo

v'-o
~'—-o»—-o--oo

I
p—e
v—Or—‘Ot—Ov—b

HOOH

' o
ot—v—Loooooooo

I

p—a

OHHp—up‘

oi—oov—Loo—u—o

0
O
1
1
1
0
1
O
1
O
0
O
O
O

OOF‘HOOHOOOH

 

174

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

83 n N... .88 u ...
8 $88.88 3.8. .89..
88.88 8 $8.8m 8:0 2....
$8... 28 88.8 E 88.8.. 2.33.8.
$8... 88 $8.8 . $8.8 N...
888 88 22.8 . 3:8 N...
83.8 a... 58.88 . 58.08 N...
3%... 9.8 88.8. . 88.8. .3
3:3 8.. 8888 . 88.8 ......
888 2 .o 88.8. . 88w .8 ......
:5... 8... .8882 . 88.38. Exam 3.
88.8 8.8 88.28 . 88.28 8.58.... n...
83.8 88 85.8 . 88.8 SE n...
.83 8 .o 88.8 . 88.8 2.2m u;
.. .. m2 ...: mm .88..

 

 

.88.. 5.5.2.58 85.2.... M8.8.. 2:. a. 2...... $52... 2... ".5558... .588 2?

Nd oEfih

175

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8.3 n ..m 38.8 n .m
8 $88.82 2.58 :38.
8888 a 8884 8...; 25m
$8.8 8.: 888.2 2 $8.82 2.8.8.83
88.8 8.8 3.3.8 H 3.228 ..x
:88 8.8 888.8 8 883: ..x
888.8 83.: 3.838 _ 388.8 ..x
88.8 8.8 8888 _ 8888 .3
$8.8 8.8 8838 8 888.8 .3
R88 8.: 888.2 8 888.: .3
28.8 8.8 888.8 8 888.8 {a
$88.8 8.8 888.8 _ 888.8 .3
28.8 88$ 888.8 H 8888 .5on ".2
9.8.8 8.8 888.88 8 888.88 2:82.“ u;
82 .o :8 888% _ 8888 825 ax
.— m m2 .5 mm 888.

 

 

 

 

 

 

 

Summon :oxusom-xce

33 Mason 8:82 2:. 3.. 2.3. <>oz< 2.9 3888.8 2.88 2;

nd «Earn.

176

The MSU Onion Peeling Machine Second Experiment

The Experiment Layout and The Raw Data Table

Appendix D

The Full Factorial Design

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table DJ
The Second Experiment: The Full Factorial Design Layout and The
Raw Data Table
flggflgfll Shape Size Pressure Speed ll w1 wL
1 1 T..—1 -1 -1. ............... -1 ..... ’ 9.2 61?...
2 1 1 0 -1 -1 ..... 20.4 14.6
3 ___1 __ 0 1 -_1 -1 ..... 25.8 17.0
..... 1 9........1 9 '1 91 1112 6-6
-11.-1 -1 0 0 -1 ..... 15.0 10.0
12 ___1 ______ 1 1 0 -1 ......... 28.2 17.4
..... 1 2 1 1 '1 1 '1 12-4 6-3 -
--..20 ...1 ...... 0 O 1 '1 ..... ”'0 10'6
.2-....1 ........... 1 ................ t .1 ...................... 1 11 ......... 23-6 149.
..-.29..-1 ....... 1 0 -1 ----1 -1 ........... 0 10.6 7.6
...29 l '1 0 ..........'..1. ............... O ..... 15'4 124.
“__30 1 1 1 -1 ..... 0 28.2 20.3
37 1 1_ _-1 0 0 12.2 9.0
..--3.1.3 1 0 0 0 0 16.6 12.2
.-39 1 '1 ..... 1 9. .................... 9. ........ .293 17-3
.-.9.9-.1 ....... 1 ................ t .1 ..... '1 1 9 ......... 9.9 5-1
47 1 ............ 1..- 9 1 ..... 9 ..... 19-0 14-0
__41; 1 0 1 - 1 ..... 0 ..... 26.2 17.0
....55 1 .1- '1 .........:1 ............... 1 ...12-4 10-9 _- -:
56 1 ............ 0 0 -1 1 16.8 14.0
....57 1 _____ -1 1 ____-__1_ 1 23.0 18.2
-..64 1 -1 —1 1 0 1 9.2 6.2
765 1 1_ 0 0 1 ..... 20.8 17.0
66 ...1 ___________ 0 1 0 1 25.0 19.2
....7..3.... ........ 1 .................. 9 '1 ............ 1 1 10.8 6-3
...-7.9. ............ 1 ................ t .1. .... 9 1 ..................... 1 ......... 1 5-2 19-6 -
75 12... 1 1 1 1 29.2 21.0

 

 

177

 

 

 

 

 

 

 

Table D.1 (Cont’d)
The Second Experiment: The Full Factorial Design Layout and The

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Raw Data Table
Ign Block Shape Size Pressure Speed II w! w! “'1 II E _11/IL
4 2 0 -1 -1 -1 10.8 7.8 7.4 88.2 27.8
5 2 -1 0 -1 -1 15.4 12.0 10.6 70.8 22.1
6 2 1 1 -1 -1 27.0 18.0 16.8 88.2 33.5"
13 2 1 -1 0 -1 12.2 7.2 7.0 96.2 41-9...
14 2 0 0 0 -1 17.0 12.4 12.2 95.8 27.1
15 2 -1 1 0 -1 24.0 16.0 15.0 88.9 33.3
22 2 -1 "-1 1 -1 9.4 5.0 4.7 93.6 46.8
23 2 1 0 1 -1 18.8 11.8 11.4 94.6 37.2
224 "2 0 1 1 -1 26.6 15.4 14.4 91.8 42.1
_31 2 1 -1 -1 0 12.0 9.0 8.4 83.8 25.0
32 2 0 0 -1 0 17.0 13.4 12.9 87.8 21.2
33 2 -1 1 -1 0 23.4 17.0 14.8 74.4 27.4
40 2 -1 -1 0 0 9.2 6.2 5.7 85.7 32.6
N .... 9.1 2 1 0 0 0 19.8 15.2 13.4 71.9 23.2__
42 2 0 1 0 o 26.8 18.8 16.2 75.5 29.9
49 2 0 -1 1 0 10.8 6.3 5.4 83.3 41.7
50 2 -1 0 1 0 16.2 11.4 10.0 77.4 29.6
51 11111111111 2 ..... 1 1 1 0 27.4 20.0 18.4 82.2 27.0
...-5.9 2 -1 -1 -1 1 10.2 8.0 7.2 73.3 21.6
59 2 1 0 -1 1 18.6 16.0 14.2 59.1 14.0_
...99 2 0 1 -1 1 24.8 20.6 18.8 70.0 16.9
.-.9? 2 0 -1 0 1 10.6 8.4 7.4 68.8 208...
-..9-8 2 -1 0 0 1 14.6 11.4 10.2 72.7 21.9
W69 2 ..... 1 1 0 1 28.0 21.2 19.6 81.0 24.3__
76 2 1 11111 -1 1 1 12.2 8.8 7.8 77.3 27.2"
W22____2 0 0 1 1 16.8 12.2 11.4 85.2 27.4
78 2 -1 1 1 1 22.8 16.8 15.2 79.0 26.3

 

 

 

178

Table D.l (Cont’d)
The Second Experiment: The Full Factorial Design Layout and the

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

II

Raw Data Table
IEMI Shape Size Pressure sfll wl wIII “'1 II E _w___L_
_____7 _________ 3 1 -1 -1 -1 12.2 8.8 8.6 94.4 27.19"”.
...... 8 3 0 0 -1 -1 17.0 13.0 12.2 83.3 23.5
9 3 -1 1 -1 -1 23.2 16.4 15.2 85.0 29.3
16 3 -1 -1 0 -1 9.6 6.6 6.3 90.9 31.3
17 3 1 o 0 -1 20.0 12.6 11.8 _90.2 37.0
18 3 0 1 0 -1 26.4 16.4 14.2 82.0 37.9
"25 3___ 0 -1 1 -1 10.4 4.7 4.4 95.0 54.8
..... 268 -1 o 1 -1 15.4 11.6 11.2 90.5 24.7
273 ..... 1. 1 1 -1 28.0 15.6 14.4 91.2 44.3
”34 3 _____ -1 -1 -1 0 9.6 7.4 6.6 73.3 22.9
35 __5 1 0 -1 0 18.0 13.8 12.2 72.4 23.3
__"36 3 0 1 -1 0 25.8 17.8 16.0 81.6 31.0
43 3 0 -1 0 0 10.6 6.4 6.1 93.3 39.6
44 3__ -1 0 _____ 0 0 15.4 11.8 10.0 66.7 23.4__
45 3 1 1 0 0 __ 27.2 19.2 15.7 69.6 294
52 3 1 -1 1 111111111111 0 12.4 7.2 7.0 96.3 41.9
533 0 0 1 0 17.2 11.8 10.8 84.4 31.4
...9.9 3 -1 1 1 0 _1 23.6 14.6 12.4 80.4 38.1
61 3 0 -1 -1 1 10.8 8.6 7.2 61.1 20-9...
62 3 -1 0 -1 1 14.8 12.2 11.4 76.5 17.6
63 3 _____ 1 1 -1 1 27.4 22.6 20.4 68.6 17.5
70 3 1 -1 0 1 12.4 9.8 8.8 72.2 21.0
21 11111111111 5 ..... 0 0 0 1 17.0 13.4 12.0 72.0 21.2
72 _"3 -1 1 0 1 25.0 19.4 18.0 80.0 22.4
798 -1 -1 1 1 9.2 5.1 4.6- 89.1 446.
80 5 1 0 1 1 19.2 15.0 13.8 77.8 21.9
81 3 0 1 1 1 26.2 18.8 17.4 84.1 28.2

 

 

 

179

BIBLIOGRAPHY

BIBLIOGRAPHY

1. General References

Aguilera, J. M. and F. V. Kosikowski, 1976, Soybean Extruded Product:
A Response Surface Analysis, V. 41, J. of Food Science, 647-651

Baasel, W. D., 1965, Exploring Response Surfaces to Establish Optimum
Conditions, Chemical Engineering, 10, 25, 147-152

Baldwin, G., 1991, 1992, personal communication

Berry, J. W., H. Tucker and A. J. Deutschman, Jr., 1963, Starch
Vinylation-Determination of Optimum Conditions by Response Surface
Designs, I & EC Process Design and Development, Vol. 2, No. 4,
October, 318-322

Bhattacharyya, G. K. and R. A. Johnson, 1977, Statistical Concept and
Method, John Wiley & Son

Biles, W. E., 1975, A Response Surface Method for Experimental
Optimization of Multi-Response Process, Ind. Eng. Chem., Process
Des. Dev., V. 14, No. 2, 152-158

Blake, A. and J. Mazumder, 1985, Control of Magnesium Loss During
Laser Welding of Al—5083 Using a Plasma Suppression Technique, J.
of Engineering for Industry, V. 107, 8, 275-280

Box, G. E. P., 1954, The Exploration and Exploitation of Response

Surfaces: Some General Considerations and Examples, Biometrics, 3,
17-60

180

Box, G. E. P. and P. V. Youle, 1955, The Exploration and Exploitation of
Response Surfaces: An Example of the Link Between the Fitted Surface
and the Basic Mechanism of the System, Biometrics, 9, 287-323

Box, G. E. P. and J. 8. Hunter, 195 7, Multi-Factor Experimental Designs
for Exploring Response Surfaces, Ann. Math. Statist., 28, 195-241

Box, G. E. P. and D. W. Behnken, 1960, Some New Three Level Designs
for the Study of Quantitative Variables, Technometrics, V. 2, No. 4,
11, 455-475

Box, G. E. P., W. G. Hunter and S. J. Hunter, 1978, Statistics for
Experimenters, John Wiley & Sons

Box, G. E. P. and N. R. Draper, 1987, Empirical Model-Building and
Response Surface, John Wiley & Sons

Bradley, R. A., 1958, Determination of Optimum Operating Conditions by
Experimental Methods-Part 1: Mathematics and Statistics Fundamental
to the Fitting of Response Surfaces, Industrial Quality Control, V.
XV, No. 1, 16-20

Burkhardt, T. H., 1993, personal communication

Cady, F. B. and W. A. Fuller, 1970, The Statistics-Computer Interface in
Agronomic Research, Agronomy J ., V. 62, 9-10, 599-604

Chinnan, M. 8., personal communication, 1993

Daniel, C., 1976, Applications of Statistics to Industrial Experimentation,
John Wiley & Sons

De Baun, R. M. , 1959, Response Surface Designs for Three Factors at
Three Levers, Technometrica, V. 1, No. l, 2, 1-8

De Baun, Robert M. , 1956, Block Effects in the Determination of Optimum
Conditions, Biometrics, 3, 21-22

181

Derringer, G. and R. Suich, 1980, Simultaneous Optimization of Several
Response Variables, J. of Quality Technology, V. 12, No. 4, 10,
214-219

Derringer, G. , 1983, A Statistical Methodology for Designing Elastomer
Formulations to Meet Performance Specifications, Kautschuk Gummi
Kunststoffe, V36 (5), 349-352

Draper, Norman R. and H. Smith, 1981, Applied Regression Analysis, John
Wiley & Son, Inc.

Draper, N. R. and D. K. J. Lin, May 1990, Small Response-Surface
Designs, Technometrics, Vol. 32, No. 2, 187-194

Evans, L. B., 1982, Optimization Theory and Its Application in Food
Processing, Food Technology, 6, 88-96

Fenwick, G. R. and A. B. Hanley, 1990, Processing of Alliums; Use in
Food Manufacture, in Onions and Allied Crops, Chapter 4, Vol 111,
73—91

Floros, J. D. and M. S. Chinnan, 1987, Optimization of Pimiento Pepper
Lye-Peeling Process Using Response Surface Methodology, Trans. of
the ASAE, Vol. 30(2), 3-4, 560-565

Floros, J. D. and M. S. Chinnan, 1988, Seven Factor Response Surface
Optimization of a Double-Stage Lye (NaOH) Peeling Process For
Pimiento Peppers, J. of Food Science, V. 53, No. 2, 631-638

Floros, J ._ D. and M. S. Chinnan, 1988, Computer Graphics-Assisted
Optimization for Product and Process Development, Food
Technology, 2, 72-84

Fox, R. L., 1971, Optimization Methods for Engineering Design, Addison-
Wesley Publishing Company, Inc.

Geier, J. E. and M. Hood, 1989, The Efiect on Cutting Toll Forces of Pre-
Weakening a Rock Surface by Waterjet Kerfing, J . of Engineering for

182

Industry, V. 111, 2, 1-6

Hahn, Gerald J. , 1985, More Intelligent Statistical Software and Statistical
Expert Systems: Future Directions, The American Statistician, 2, V.
39, No.1, 1-8

Harper, J. M. and L. A. Wanninger, Jr., 1969, Process Modeling &
Optimization, 1. Experimental Design, Food Technology, V. 23, 9,
1153-1155

Harper, J. M. and L. A. Wanninger, Jr., 1970, Process Modeling &
Optimization, 2. Modeling, Food Technology, V. 24, 3, 266-268

Harper, J. M. and L. A. Wanninger, Jr., 1970, Process Modeling &
Optimization, 3. Optimization, Food Technology, V. 24, 5, 590-595

Heldman, Dennis R., 1975, Food Process Engineering, The AVI Publishing
Company, INC.

Heller, N. B. and Glenn E. S., 1973, Response Surface Optimization when

Experimental Factors are Subject to Costs and Constraints,
Technometrics, V. 15, No. 1, 113-123

Henika, R. G., 1972, Simple and Efi’ective System for Use with Response
Surface Methodology, Cereal Science Today, Vol. 17, No. 10, 10,
309-314

Henselman, M. R., S. M. Donatoni and R. G. Henika, 1974, Use of
Response Surface methodology in the Development of Acceptable High
Protein Bread, J. Food Science, V. 39, 943-946

Hill, W. J. and W. G. Hunter, 1966, A Review of Response Surface
Methodology: A Literature Survey, Technometrics, V. 8, No. 4, 11,
571-590

Hoke, A. T. , 1974, Economical Second-Order Designs Based on Irregular
Fractions of the 3" Factorial, Technometrics, V. 16, No. 3,

183

Hunter, J. S., 1958, Determination of Optimum Operating Conditions by
Experimental Methods-Part II-I: Models and Methods, Industrial
Quality Control, V. XV, No. 6, 12, 16-23

Hunter, J. S., 1959, Determination of Optimum Operating Conditions by
Experimental Methods-Part II-2: Models and Methods, Industrial
Quality Control, V. XV, No. 7, 1, 7-15

Hunter, J. S., 1959, Determination of Optimum Operating Conditions by
Experimental Methods-Part II-3: Models and Methods, Industrial
Quality Control, V. XV, No. 8, 2, 6-14

Hunter, J. S. and T. H. Naylor, 1970, Experimental Design for Computer
Simulation Experiments, Managment Science, V. 16, No. 7, 3, 422-
434

Hunter, J. S. , 1985, Statistical Design Applied to Product Design, Journal
of Quality Technology, V. 17, No. 4, 210-221

Jones, H. A. and Louis K. Mann, "Onion and Their Allies ", 1963

Johnson T. M. and M. E. Zabik, 1980, Response surface Methodology for
Analysis of Protein Interactions in Angel Food Cakes, J. of Food
Science, V. 46, 1226-1230

Kannard, R. W. and L. A. Stone, 1969, Computer Aided Design of
Experiments, Technolometrics, V. 11, No. 1, 137-148

Khuri, A. I. and J. A. Cornell, 1987, Response Surface Design and
Analysis, Marcel Dekker, Inc.

Kissell, L. T. and B. D. Marshall, 1962, Multi-Factor Responses of Cake
Quality to Basic Ingredient Ratios, Cereal Chem., 39: 16-30

Kissell, Leo T., 1967, Optimization of White Layer Cake Formulations by
a Multiple-Factor Experimental Design, Cereal Chem., V. 44, 1967,
253-268

184

Koul, H. L., 1993, personal communication

Lah, C. L., Munir C. and Richard E. D., 1980, A Response Surface
Methnology Approach to the Optimization of Whipping Properties of
an Ultrafiltered Soy Product, J. of Food Science, V. 45, 1720-1731

Lind, E. E., J. Goldin and J. B. Hickman, 1960, Fitting Yield and Cost
Response surfaces, Chemical Engineering Process, V. 56, No. 11, 62-
68

Lopez, A. , I 98 7, A Complete Course in Canning and Related Processes,
Book III - Processing Procedures for Canned Food Products, 12th
Edition, The Canning Trade INC.

Lucas, J. M. , 1976, Which Response Surface Design is Best-A Performance
Comparison of Several Types of Quadratic Response Surface Designs
in Symmetric Regions, Technometrics, V. 18, No. 4, 411-417

Luh, B. S. and J. G. 1988, Woodroof, Commerical Vegetable Processing

Mead, R. and D. J. Pike, 1975, A Review of Response Surface Methodology
From A Biometric Viewpoint, Biometrics, 31, 803-851

Min, D. B. and E. L. 1980, Thomas, Application of Response Surface
Analysis in the Formulation of Whipped Topping, J. of Food Science,
V. 45, 346-348

Motycka, R. R., R. E. Devor and P. J. Bechtel, 1984, Response Surface
Methodology Approach to the Optimization of Boneless Ham Yield, J.
of Food Science, V. 49, 1386-1389

Myers, R. H., 1971, Response Surface Methodology, Allyn and Bacon, Inc.

Nakai, S. , 1981, Comparison of Optimization Techniques for Application to
Food Product and Process Development, J. of Food science, V. 47,
144-176

Norback, John P. , 1980, Techniques for Optimization of Food Processes,

185

Food Technology, 2, 86-113

Oh, N. H., P. A. Seib and D. S. Chung, 1985, Noodles. III. Efi’ects of

Processing Variables on Quality Characteristics of Dry Noodles,
Cereal Chemistry, V. 62, No. 6, 437-440

Ross, E. G. , 1961 , Response Surface Techniques as a Statistical Approach
to Research and Development in Ultrassonic Welding, ASQC
Convention Transactions, 445-456

Saguy, Israel, 1982, Optimization Theory, Techniques and Their
Implementation in the Food Industry: Introduction, Food Technology,
6, 87

Saguy, Israel, I 982, Computer-Aided Techniques in Food Technology,
Marcel Dekker, New York

Saguy, Israel, 1984, Martin A. Mishkin and Marcus Karel, Optimization
Methods and Available Software, Part 1, CRC Critical Reviews in
Food Science and Nutrition, V. 20, Issue 4, 275-299

Salunkhe, D. K. and B. B. Desai, 1984, Onion and Garlic, in Postharvast
Biotechnology of Vegetables, Vol 11, Chapter 3, 23-38

Schoney, R. A., T. F. Bay and J. F. Moncrief, 1981, Use of Computer
Graphics in the Development and Evaluation of Response Surfaces,
Agronomy J ., V. 73, 5-6, 437—442

Skeggs, P. K., 1985, Mechanical Dough Development-dough Water Level
and Flour Protein Quantity, Cereal chemistry, V. 62, No. 6, 458-462

Smith, L. M., M. B. Carter, T. Dairiki, A. Acuna-Bonilla and W. A.
Williams, 1977, Physical Stability of Milk Fat Emulsions After
Processing as Evaluated by Response Surface Methodology, J. Agric.
Food Chem,. V. 25, No. 3, 647-653

Snee, R. D. , 1985, Computer-aided design of Experiments-some Practical
Experiences, J. of Quality Technology, V. 17, No. 4, 222-236

186

Srivastava, A., VanEe, G., Ledebuhr, R., Welch, D., and Wang, L., 1992,
Design and Development an Onion Peeling Machine, ASAE 1992
International Summer Meeting, June, 1992, Charlotte, North Carolina

Srivastava, A., personal communication, 1990, 1991, 1992 and 1993

Srivastava, A., VanEe, G., Ledebuhr, R., Stuecken, T, 1989, Development
of A Mechanical System to Pell and Trim Onions-A Final Report,
Department of Agricultural Engineering, Michigan State University,
East Lansing, Michigan

Swanson, A. M., J. J. Geissler, P. J. Magnino Jr. and J. A. Swanson,
1967, Use of statistics and Computers in the Selection of Optimum
Food Processing Techniques, Food Technology, V. 21 , 99-102 (1513-
1516)

Terhune, H. D., 1963, A Ststistical Analysis of the Efiect of Four Variable
Operating Conditions on Beleaching Results, American Dyestuff
Reporter, April 1, 33-38

Thompson, D., 1982, Response surface Experimentation, J. of Food
Processing and Preservation, V. 6, 155-188

Underwood, W. M., 1962, Experimental Method for design Extrusion
Screws, Chemical Engineering process, V. 58, No. 1, 59-65

Wang, L. , 1991 , Onion Peeling System Machine Efi'iciency Testing Report-
First experiment, Agricultural Engineering Department, Michigan
State University, dec., unpublished

Wang, L. , 1992, Onion Peeling System Machine Efi'iciency Testing Report-
Second experiment, Agricultural Engineering Department, Michigan
State University, Mar., unpublished

Welch, W. L., 1984, Computer-Aided Design of Experiments for Response
Estimation, Technometrics, V. 26, No. 3, 217-224

187

Whidden, P., 1956, Design of Experiment in Metals Processing, Trans.
Tenth Ann. Com., ASQC, 677-683

Wu, S. M. , 1964, Tool-Life Testing by Response Surface Methodology-Part
I, J. of Engineering for Industry, 5, 105-110

Wu, S. M., 1964, Tool-Life Testing by Response Surface Methodology-Part
2, J. of Engineering for Industry, 5, 111-116

Wu, 8. M. and R. N. Meyer, 1964, Cutting-Tool Temperature-Predicting

Equation by Response-Surface Methodology, J. of Engineering for
Industry, 5, 150-156

Wu, S. M. and R. N. Meyer, 1964, A First-Order Five-Variable Cutting-
Tool Temperature Equation and Chip Equivalent, J. of Engineering
for Industry, 11, 395-400

Yutaka K. and et al., 1978, Studies on the Physical Properties of Onions,

Research Bulletin of the Hokkaido National Agricultural Experiment
Station, Japan, 1, 1-15

2. U.S. Patents
(1) U.S. Patent, 899,340, 1/1906, Leon G. Simpson, Paring-Machine,
Rochester, New York, USA

(2) U.S. Patent, 926,286, 7/1908, Charies J. Petrie, Onion-Topping
Machine, Cleveland, Ohio, USA

(3) U.S. Patent 1,294,033, 5/1918, Arthur Felix Bizette, Onion-Cutter
Morganza, Louisiana, USA

(4) U.S. Patent 1,379,049, 10/1920, Frank J. Schroeder, Onion-Topping
Machine, Rochester, New York, USA

188

(5) U.S. Patent 1,720,468, 8/1928, Ross Combest, Rotary Fruit And
Vegetable Stemming, Feeding, And Trimming Head, Beaumont, Texas,
USA

(6) U.S. Patent 1,902,506, 7/1929, Herbert L. Johnston, Food Handing
Apparatus, Troy, Ohio, USA

(7) U.S. Patent 1,995,694, 12/1932, William E. Urschel, Onion Snipper,
Valparaiso, Indiana, USA

(8) U.S. Patent 2,286,649, 10/1941, Harrison W. Rogers, Fruit Peeling
Machine, Miami, Florida, USA

(9) U.S. Patent 2,445,881, 12/ 1945, George T. Hemmeter, Apparatus For
Peeling Onions, Including a Conical Jet of Gas, Berkeley, Calif., USA

(10) U.S. Patent 2,494,914, 9/1944, Joe R. Urschel and Gerald W.
Urschel, Machine For Clipping Onions And the Like, Valparaiso, Indiana,
USA

(11) U.S. Patent 2,553,519, 7/1946, E. F. Lens, Onion Tipper, Larsen,
Wisconsin, USA

(12) U.S. Patent 2,602,480, 4/1948, David L. Taylor, Onion Skinning And
Slicing Machine, Calif., USA

(13) U.S. Patent 2,750,977, 5/1953, Salvatore S. Vella, Apparatus For
Clipping Tops From Onions, Albion, New York, USA

(14) U.S. Patent 2,766,794, 6/1952, Erneat D. Odale, Method of
Removeing Outer Skin From Vegetable, Calif., USA

(15) U.S. Patent 2,858,866, 5/1955, Elmer Hendry, Vegetable Trimming
Machine, San Jose, Calif., USA

(16) U.S. Patent 2,906,308, 4/1957, Leon A. Genettl Sr., Apparatus For
Removing Shell Jackets From Dried OnionsHazieton, Pennsylvania, USA

189

(17) U.S. Patent 2,961,023, 1/1958, Emanuel F. Boyer, Vegetable
Trimming Machine, Barrie Center, New York, USA

(18) U.S. Patent 3,134,413, 1/1962, Frank J. Dorsa, Apparatus For Peeling
Fruits or Vegetables, San Jose, Calif., USA

(19) U.S. Patent 3,212,545, 1/1963, Henry Aguilar, Method of Orienting
And Pitting Fruit Having Stems, San Francisco, Calif., USA

(20) U.S. Patent 3,485,278, 11/ 1966, Leslie A. Parsons, Treatment of
Onions, Burry Port, Wales, Great Britain

(21) U.S. Patent 3,485,279, 11/ 1966, Leslie A. Parsons, Treatment of
Onions, Burry Port, Wales, Great Britain

(22) U.S. Patent 3,515,193, 10/1967, Henry Aguilar, Onion Orienter And
Cutter, San Francisco, Calif., USA

(23) U.S. Patent 3,543,824, 12/1968, Leslie A. Parsons, Treatment of Fruit
And Vegetable Crops, Berry Port, South Wales, Great Britain

(24) U.S. Patent 3,602,279, 4/1969, Albertus V. Raaij, Machine For
Skinning Onions And Like Bulb-Type Vegetable, Ulft, Netherlands

(25) U.S. Patent 3,606,917, 5/1969, Gerald J. Orlowski, Peeling Machine,
Chicago, Illinois, USA

(26) U.S. Patent 3,623,524, 12/1968, Edwin J. Buck, Machine For
Preparing Onions, Victoria, Australia

(27) U.S. Patent 3,696,848, 7/1970, Eugene P. Mellon, Method And
Apparatus For Removing Skin From Onions or Like Vegetables,
Clarksburg, Maryland, USA

(28) U.S. Patent 3,724,362, 10/1970, Leslie A. Parsons, Article Feeding
And Treating Apparatus, Burry Port, South Wales, England

(29) U.S. Patent 3,764,717, 8/1971, Walter E. Reed, Method For

190

Autumatically Orineting And Trimming vetetables, Stevensville, Michigan,
USA

(30) U.S. Patent 3,765,320, 1/1971, Albertus V. Raay, Onion End Cutter,
Ulft, Netherlands

(31) U.S. Patent 3,861,295, 6/1973, Emanuel F. Boyer, Onion Skinning
Machine, Albion, New Kork, USA

(32) U.S. Patent 3,886,951, 8/1973, Leon R. McRobert, Trash Separator,
Ocoee, Florida, USA

(33) U.S. Patent 3,915,083, 9/1973, Hendrik Spruijt, Apparatus for
Automatically Processing Bulbous And Tuberous Plants, Emmeloord,
Netherlands

(34) U.S. Patent 3,942,428, 11/1973, Claus Clausen, Apparatus For
Removing Roots From Bulbs or the Like Corms, Faarevejile, Denmark

(35) U.S. Patent 3,954,032, 12/1973, Franklin K. Holbrook, Citrus Peel Oil
Extractor For Whole Fruit, Whittier, Calif., USA

(36) U.S. Patent 3,959,506, 5/1974, Paul Kunz, Method For Peeling Fruit
and Vegetables, Doettesfeld, Germany

(37) U.S. Patent 4,023,477, 9/1975, Kastsuli Hirahara, Flexible Cable Dry
Peeler With Rubber-Like Cords, Santa Clara, Calif., USA

(39) U.S. Patent 4,068,011, 9/1975, Glen R. Green, Method of Peeling
Onions By Scalding and Cutting, Ontario, Oregon, USA

(40) U.S. Patent 4,125,066, 1/1977, Thomas P. Stokes, Apparatus For
Facilitating Separation of Peel From Produce, East Wenatchee, Washington,
USA

(41) U.S. Patent 4,132,162, 12/1977, Robert M. Magnuson, Apparatus For
Peeling Fruits And Vegetables, Saratoga, Calif., USA

191

(42) U.S. Patent 4,258,069, 2/1979, John H. Amstad, Method For
Continuous Produce Surface Treament, Alameda, Calif., USA

(43) U.S. Patent 4,373,589, 2/1981, Yitzchak Hagiz, Harvesting
Apparatus For Onions, Petach Tikva, Israel

(44) U.S. Patent 4,361,084, 1/1981, Gabor J. Raatz, Industrial Peeler For
Onions or the Like Bulbous and Tuberous Vegetation, Wageningen,
Netherlands

(45) U.S. Patent 4,442,764, 9/1982, Pierrie H. Bos, Machine For Peeling
and Cleaning Foodstuffs, Particularly Vegetables Such As Onions,
Aubusson, France

(46) U.S. Patent 4,448,118, 5/1982, Paul Kunz, Cleaning and Peeling
Machine, Dottesfeld, Fed. Rep. of Germany

(47) U.S. Patent 4,450,762, 6/1982, Thomas P. Lustig, Onion Peeling
Means, Wellington, New Zealand

(48) U.S. Patent 4,457,224, 4/1982, Mitsutaro Kino, Apparatus For
Stripping Onions, Yokoharma, Japan

(49) U.S. Patent 4,470,345, 4/1983, Hiroyuki Miyata, Apparatus For
Peeling Skins Off the Bulbs of Onions, Hoya-shi, Tokyo, Japan

(50) U.S. Patent 4,476,778, 7/1980, Malcolm R. Clyma, Onion Peeling,
Holden Hill, Australia

(51) U.S. Patent 4,481,875, 12/1982, Kaku Toyosato, Bulb Peeling
Apparatus, Yokohama, Japan

192