PROCESS OPTIMISATION OF MECHANICAL OIL 
EXPRESSION FROM Moringa oleifera (Lam.) (MORINGA) 
SEEDS  
 
BY 
 
FAKAYODE, OLUGBENGA ABIOLA (147132)          
B.Eng Agricultural Engineering (Akure), M.Sc. Agricultural and Environmental 
Engineering (Ibadan) 
 
A Thesis in the Department of Agricultural and Environmental Engineering, 
Submitted to the Faculty of Technology  
In partial fulfilment of the requirements for the Degree of 
 
 
DOCTOR OF PHILOSOPHY 
 
of the 
 
UNIVERSITY OF IBADAN, NIGERIA. 
 
 
 
 
Department of Agricultural and Food Engineering, 
University of Uyo, 
Uyo, Akwa Ibom State, Nigeria. 
 
 
 
JULY 2015. 
 
 
CERTIFICATION 
I certify that this work was carried out by Mr. O.A. Fakayode in the Department of 
Agricultural and Environmental Engineering, Faculty of Technology, University of Ibadan. 
 
 
 
 
 
 
……………………………………………………………………………………………….
                    
Supervisor 
Professor E.A. Ajav 
B.Sc., M.Sc. (Ibadan), Ph.D (G.B. Pant, India) 
MASABE, MNSE, MNIAE, MNIFST, R.Engr (COREN) 
 
 
 
 
 
 
 
 
 
 
                        
 
 
 
 
 
 
 
 
 
 
i 
 
 
DEDICATION 
This work is dedicated to the glory of God Almighty and the benefits of mankind. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ii 
 
 
ACKNOWLEDGEMENTS 
 This is to express my appreciation to my parents, Mr. and Mrs. Fakayode for your 
great love and care always. Your moral and financial supports as well as your encouragement 
and advice have kept me going. I promise never to let you down. I am also grateful to 
Omolola and Olufemi Fakayode; as well as Kehinde, Olayemi and Heritage Ojoawo for your 
great contributions towards the success of this work. I’m very proud of you. Special thanks to 
Mr. and Mrs. Oni for their love and care. I really appreciate you. 
I am very grateful to my supervisor, Prof. E.A. Ajav whose guidance, suggestions, 
encouragement, criticisms and assistance have contributed to this great work. I say a very big 
thank you. 
I am equally grateful to the Head, Prof. A.O. Raji and all my Lecturers in the 
Department of Agricultural and Environmental Engineering, University of Ibadan; most 
especially Prof. Lucas, Prof. Sangodoyin, Prof. Bamgboye, Prof. Ogedengbe, Prof. 
Olorunisola, Prof. Onilude, Dr. Mijinyawa, Dr. Ewemoje, Dr. Aremu, Dr. Owofadeju, Dr. 
Adefisan, Dr. Omoniyi, Dr. Adewole, Mr. Omobowale and Mrs. Ewemoje. I also appreciate 
my colleagues in the department for their great understanding and love.  
 I am greatly indebted to Dr. Akinoso (Food Technology Department, UI), Engr. 
Sadiku (Agronomy Department, UI), Mr. Al-hassan (FIIRO, Oshodi), Mr. Ola and the entire 
NOVA Technology Company, Electrical Department of the Maintenance Unit, UI and Mr. 
Uko (Food Science and Technology Department, Uniuyo).      
 I cannot forget to appreciate Dr. and Mrs. Fakayode, the Oladeles, the Olatunjis, the 
Oyelades, Dr. A.F. Alonge, Dr. Ikrang, Dr. Ajav, Prof. Aluko, Adedeji Banjo, Agbeja Tope, 
Akinade Biodun, Ojo Tunji, Ogunremi Tade, Olaoye Kunle, Aterigbade Charles, Adewale 
Niyi, Ogunlade Clement, Mr. Ogundipe, David Onwe, Femi Alege, Engr. Israel Ahuchaogu, 
Faboade Micah, Umani Kingsley and Eduke Milly. Thank you greatly.   
 To everyone that has contributed to the success of this work in any way, I really 
appreciate you in no small measure. I am also grateful to the Postgraduate School, University 
of Ibadan for awarding me Doctoral Research Grant during my programme.  
 To my precious jewel of inestimable value, I’m short of words to appreciate your 
love, care, encouragement, support, understanding and prayers. Thank you for believing in 
me. You are truly my heroine, Fakayode Omorinola Tolulope.  
 
FAKAYODE, Olugbenga Abiola (MNSE, MNIAE, MNIFST, MASABE, COREN Regd.) 
 
iii 
 
 
TABLE OF CONTENTS 
                  PAGE 
TITLE PAGE         
CERTIFICATION           i  
DEDICATION           ii 
ACKNOWLEDGEMENTS          iii 
TABLE OF CONTENTS          iv 
LIST OF TABLES           viii 
LIST OF FIGURES           x 
LIST OF PLATES           xiii 
ABSTRACT            xiv 
 
CHAPTER ONE:    INTRODUCTION 
1.1 Background Information        1 
1.2 The Research Problem         2 
1.3 Objectives of the Study         3 
 
CHAPTER TWO:   LITERATURE REVIEW 
2.1   Moringa oleifera                4 
2.2   Engineering Properties of Agricultural Materials     7 
2.3   Oil Extraction Methods        12 
    2.3.1  Water Assisted                    12 
    2.3.2  Manual Pressing         13 
    2.3.3  Oil Expellers         16 
    2.3.4  Ghanis                            16 
    2.3.5  Solvent Extraction                    18 
2.4   Effects of Processing Factors on Mechanical Expression of Oil from Oilseeds 18 
    2.4.1  Effects of Moisture Content       18 
    2.4.2   Effects of Applied Pressure                  19 
    2.4.3   Effects of Heating Temperature and Time     21 
2.5 Interaction between Process Variables and Oil Yield           22 
2.6 Effects of Processing Factors on Oil Quality     24 
2.7 Process Optimisation by Response Surface Methodology (RSM)  26 
2.8 Model Equations for Oil Expression Processes      28 
iv 
 
 
CHAPTER THREE:   MATERIALS AND METHODS 
3.1   Determination of the Engineering Properties of Moringa    34 
   3.1.1   Determination of Physical Properties of Moringa    34                             
      3.1.1.1 Dimensional Properties                  34 
      3.1.1.2 Gravimetric Properties        34 
      3.1.1.3 Frictional Properties         35 
   3.1.2   Determination of Mechanical Properties of Moringa    35 
3.2   Design Considerations for the Moringa Oil Expeller    36 
3.3   Theoretical Considerations for the Moringa Oil Expeller    38 
3.4   Design Calculations                38 
   3.4.1   Determination of the Mass of Oil extracted per batch    38 
   3.4.2   Determination of the Roaster Volume                 38 
   3.4.3   Stirring Clearance         39 
   3.4.4   Determination of Heater Capacity      39 
   3.4.5   Roasting Time         40 
   3.4.6   Determination of Expeller Volume      40 
   3.4.7   Determination of Barrel Volume       40 
   3.4.8   Determination of Shaft Volume       40 
   3.4.9   Design Capacity of the Expeller       41 
   3.4.10   Forces acting on Screw Thread       42 
   3.4.11   Torque on Screw Thread        42 
   3.4.12   Power Requirements           43 
   3.4.13   Belt Design          43 
   3.4.14   Determination of Tensions in Belt         45 
   3.4.15   The Power Transmission Shaft       49 
3.5   Description of the Moringa Oil Expeller                 52 
  3.5.1   The Roasting Chamber               52 
  3.5.2   The Feeding Assembly               52 
  3.5.3   The Expression Chamber        52 
  3.5.4   The Worms and Wormshaft Assembly      52 
  3.5.5   The Power Unit                54 
  3.5.6   The Oil and Cake Troughs                   54 
  3.5.7   The Machine Frame                          54 
3.6   Working Principles of the Moringa Oil Expeller     54 
v 
 
 
3.7   Experimental Design         54 
3.8   Response Surface Methodology (RSM)       56 
3.9   Statistical Analysis         56 
3.10  Experimental Tests of the Moringa Oil Expeller     56 
3.11  Determination of the Total Oil Yield of Moringa Seeds     58 
3.12  Physico-chemical Analysis of the Moringa Oil     60 
  3.12.1  Free Fatty Acid                               60 
  3.12.2  Colour                 61 
  3.12.3  Oil Impurity                61 
 
CHAPTER FOUR:   RESULTS AND DISCUSSION 
4.1   Physical Properties of Moringa       62 
4.2   Mechanical Properties of Moringa              74 
4.3   Effects of Processing Conditions on Oil Yield of Moringa    76          
   4.3.1   Moisture Content          81  
   4.3.2   Heating Temperature        81 
   4.3.3   Heating Time         85 
   4.3.4   Applied Pressure          85 
4.4   Optimization of the Expression Process of Moringa Seeds using Yield                    
        as Response           85      
4.5   Validation of Model                87 
4.6   Effects of Processing Conditions on the Physico-chemical Properties of                     
        Moringa Oil           87 
   4.6.1  Free Fatty Acid                87 
   4.6.2  Colour                                           90 
   4.6.3  Oil Impurity                                           94 
4.7 Predictive Models for The Physico-chemical Properties of Moringa Oil 94 
4.8   Efficiency of the Moringa Oil Expeller                            102       
4.9   Cost Analysis of the Moringa Oil Expeller                 103 
 
CHAPTER FIVE:   CONCLUSIONS AND RECOMMENDATIONS 
5.1   Conclusions                106 
5.2   Recommendations         107 
5.2   Contributions to Knowledge        107 
vi 
 
 
REFERENCES                     108 
APPENDICES          119 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
vii 
 
 
LIST OF TABLES 
No  Description                     Page 
3.1  Standard V-Belts Pitch Lengths       46 
3.2   Levels, Codes and Intervals of Independent Variables for the   
  Experiment          57 
3.3  Experimental Design (Second Order Design in the Four Variables)    57 
4.1  Dimensional Properties of Moringa Seeds      63 
4.2  Gravimetric Properties of Moringa Seeds      63 
4.3  Frictional Properties of Moringa Seeds      63 
4.4  Mechanical Properties of Moringa Seeds      75 
4.5  Oil Yield of Moringa at various Processing Conditions                             78  
4.6  Physico-chemical Properties of Moringa Oil      89 
4.7  Material Balance in terms of Oil and Cake Output    104 
4.8  Cost Analysis of the Moringa Oil Expeller     105 
 
 
APPENDICES 
No  Description                     Page 
A1  Moisture Content of Moringa Seeds      119 
A2  Linear Dimensions of Moringa Seeds     119 
A3  True Density of Moringa Seeds      123 
A4  Bulk Density of Moringa Seeds       123 
A5  One Thousand Seed Weight of Moringa Seeds    123 
A6  Dynamic Angle of Repose of Moringa Seeds    124  
A7  Static Coefficient of Friction of Moringa Seeds on different Surfaces 124 
B1  Mechanical Properties of Moringa Seeds at Peak    125 
B2  Mechanical Properties of Moringa Seeds at Break    126 
B3  Mechanical Properties of Moringa Seeds at Yield    126 
C1  Total Oil Content of Moringa Seeds by Soxhlet Method   127 
E1  Yield of Moringa Oil at different Processing     
  Conditions (Experiment 1)        129 
E2  Yield of Moringa Oil at different Processing     
  Conditions (Experiment 2)        130 
viii 
 
 
E3  Yield of Moringa Oil at different Processing     
  Conditions (Experiment 3)        131 
F1  Material Balance in terms of Oil and Cake      
  Output (Experiment 1)         132 
F2  Material Balance in terms of Oil and Cake      
  Output (Experiment 2)         133 
F3  Material Balance in terms of Oil and Cake      
  Output (Experiment 3)         134 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ix 
 
 
LIST OF FIGURES 
No  Description                     Page 
2.1  Ram Press             14 
2.2  Oil Press with Hydraulic Jack            14 
2.3  Bridge Press (NRI design)           14 
2.4  Scissor Press (IPI design)             15 
2.5  Curb Press (TCC design)           15  
2.6  Outline Drawing of a Typical Small-Scale Expeller (CeCoCo Type 52)        17 
2.7  Traditional Animal Powered Ghani          17   
2.8  Power Ghani                 17 
3.1  Effective Power of Belts as a function of rpm of small Sheaves    44    
3.2  Bending Loads on the Wormshaft        50 
3.3  Shear Force and Bending Moment Diagrams        51 
4.1  Frequency Distribution of the Length of Moringa Seeds     65 
4.2  Frequency Distribution of the Width of Moringa Seeds     66 
4.3  Frequency Distribution of the Thickness of Moringa Seeds       67 
4.4  Frequency Distribution of the Arithmetic Mean Diameter of   
  Moringa Seeds            68 
4.5  Frequency Distribution of the Geometric Mean Diameter of   
  Moringa Seeds          69 
4.6  Frequency Distribution of the Surface Area of Moringa Seeds    70 
4.7  Frequency Distribution of the Sphericity of Moringa Seeds                71 
4.8  Frequency Distribution of the Aspect Ratio of Moringa Seeds    72  
4.9  Typical Example of Force-Deflection Curves for Moringa Seeds under  
  Compressive Loading          77 
4.10  Typical Example of Force-Strain Curves for Moringa Seeds under   
  Compressive Loading          77 
4.11  Effect of Moisture Content and Heating Temperature on Oil Yield       82 
4.12  Effect of Moisture Content and Heating Time on Oil Yield        82 
4.13  Effect of Moisture Content and Applied Pressure on Oil Yield    83 
4.14  Effect of Heating Temperature and Heating Time on Oil Yield    83 
4.15  Effect of Heating Temperature and Applied Pressure on Oil Yield      84 
4.16  Effect of Heating Time and Applied Pressure on Oil Yield       84  
x 
 
 
4.17  Predicted and Actual values for Moringa Oil Yield    88 
4.18  Effect of Moisture Content and Heating Temperature on Oil FFA  91 
4.19  Effect of Moisture Content and Heating Time on Oil FFA   91 
4.20  Effect of Moisture Content and Applied Pressure on Oil FFA  92 
4.21  Effect of Heating Temperature and Heating Time on Oil FFA  92 
4.22  Effect of Heating Temperature and Applied Pressure on Oil FFA  93 
4.23  Effect of Heating Time and Applied Pressure on Oil FFA   93 
4.24  Effect of Moisture Content and Heating Temperature    
  on Oil Colour Intensity       95 
4.25  Effect of Moisture Content and Heating Time     
  on Oil Colour Intensity       95 
4.26  Effect of Moisture Content and Applied Pressure     
  on Oil Colour Intensity       96 
4.27  Effect of Heating Temperature and Heating Time     
  on Oil Colour Intensity       96 
4.28  Effect of Heating Temperature and Applied Pressure    
   on Oil Colour Intensity       97 
4.29  Effect of Heating Time and Applied Pressure     
  on Oil Colour Intensity       97 
4.30  Effect of Moisture Content and Heating Temperature  
  on Oil Impurity        98 
4.31  Effect of Moisture Content and Heating Time on Oil Impurity  98 
4.32  Effect of Moisture Content and Applied Pressure on Oil Impurity  99 
4.33  Effect of Heating Temperature and Heating Time on Oil Impurity  99 
4.34  Effect of Heating Temperature and Applied Pressure on Oil Impurity 100 
4.35  Effect of Heating Time and Applied Pressure on Oil Impurity  100 
 
APPENDICES 
No  Description                     Page 
L1  The Hopper         176 
L2  The Electric Motor        176 
L3  The Roasting Chamber       177 
L4  The Oil Barrel         177 
L5  The Wormshaft        178 
xi 
 
 
L6  The Barrel Oil Cover        178  
L7  The Oil Trough        179 
L8  The Temperature Regulator       179 
L9  Belt and Pulley Arrangement       180 
L10  The Bevel Gears        180 
L11  The Stirrer         181 
L12  The Electric Heater          181 
L13  The Pillow Bearings        182 
L14  The Machine Base        182 
L15  The Front View        183 
L16  The Right End View        184 
L17  The Left End View        185 
L18  The Plan View        186 
L19  The Schematic View        187 
L20  The Schematic Exploded View      188 
L21  The Wireframe View        189 
L22  The Main View        190 
L23  Multiple Views        191 
L24  The Exploded View        192 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
xii 
 
 
LIST OF PLATES 
No  Description                     Page 
2.1  The Moringa Plant                     8 
2.2  The Moringa Tree                     8 
2.3  Moringa Pods                               9 
2.4  Hulled Moringa Seeds               9 
2.5  Dehulled Moringa Seeds               10 
3.1  The Universal Instron Testing Machine       37 
3.2  The Moringa Oil Expeller         53 
3.3  Expressed Moringa Oil Samples        59 
3.4  Soxhlet Apparatus            59 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
xiii 
 
 
ABSTRACT 
 Moringa oleifera plant is a promising nutritional and economic resource for 
developing countries. Previous studies on moringa plant have focused largely on its 
medicinal and nutritional uses, with only a few on extraction of oil from its seeds; while 
solvent and aqueous enzymatic oil extraction methods have been reported for moringa seeds, 
literature is scarce on mechanical oil expression. This study was conducted to investigate and 
optimise mechanical oil expression from moringa seeds. 
 Selected physical and mechanical properties of moringa seeds determined using 
ASABE standards were input parameters for the design and subsequent fabrication of an oil 
expeller. Central composite rotatable design at 4 factors and 5 levels was adopted for the 
experiments. Based on moisture content at harvest, preliminary tests and literature, moisture 
o
contents (8, 9, 10, 11 and 12% wet basis), heating temperature  (50, 60, 70, 80 and 90 C), 
heating time (15, 20, 25, 30 and 35 min) and applied pressures (5, 10, 15, 20 and 25 MPa) 
were chosen. Oil yields were calculated as percentage of input materials; while oil quality 
criteria such as Free Fatty Acid (FFA), oil impurity and colour were determined using AOAC 
standard methods. The efficiency of oil expression was obtained in terms of yield and 
material output. Data analyses were done using multiple linear regression at p = 0.05. 
Predicted optimum conditions were validated using experimental values. 
 Moisture content of the moringa seeds was 7.31±0.31% (wet basis). The seed length, 
width and thickness of 8.45±0.98, 7.82±0.92 and 6.41±1.10 mm respectively were used in 
determining perforation size and barrel clearance. The true and bulk densities of 971.0±0.11 
-3
and 662.0±0.03 kgm  respectively were useful in determining the expeller capacity and the 
o 
angle of repose of 21.4±0.75 for hopper, oil and cake troughs design. The peak force, 
deformation and rupture energy of 67.0±5.47 N, 5.35±0.10 mm and 0.17±0.02 Nm 
respectively obtained were for determining energy and force requirements of the expeller. Oil 
yield ranged from 11.42-28.6% with the highest obtained at moisture content of 11%, 
o
temperature of 80 C, duration of 30 mins and pressure of 20 MPa. The FFA, oil impurity and 
colour ranged from 2.42-7.40 mg/KOH/g, 2.12-3.20% and 5.70-7.60 LU respectively and fell 
within acceptable limits. The expressed and material balance efficiencies were 81.7 and 
2
93.8% respectively. Coefficients of determination (R ) for the oil yield, FFA, oil impurity and 
colour were 0.77, 0.94, 0.98 and 0.99 respectively. Predicted optimum oil yield of 28.2% at 
o
moisture content of 11.30%, temperature of 85.57 C, duration of 27.17 mins and pressure of 
19.63 MPa was obtained. Deviations between experimental and predicted values were low 
xiv 
 
 
and ranged from 0.01-6.20, 0.01-0.68, 0.01-0.12 and 0.01-0.12 for the oil yield, FFA, colour 
and oil impurity respectively. 
 Moisture content, applied pressure, heating temperature and duration influenced the 
quality and quantity of oil recovery from moringa seeds using expeller. High oil expression 
efficiency obtained makes the expeller a potential for moringa oil expression.  
Keywords: Moringa oleifera, Oil expeller, Oil yield, Oil quality. 
Word count: 479 
 
 
 
 
xv 
 
 
CHAPTER ONE 
INTRODUCTION 
1.1 Background Information 
 Oils and fats being a natural source have extensive applications in our modern industrial 
world. Global industrialization and the increasing demand for environmentally acceptable 
materials have led to the investigation and exploitation of more vegetable oils as a renewable 
feed stock in the preparation of oleo chemicals in order to meet the growing needs of human 
society. Oils and fats are important parts of human diet and more than ninety percent of the 
world production from vegetable, animals and marine sources is used as food or as an ingredient 
in food products. Oils and fats constitute one of the three major classes of food constituents 
besides proteins and carbohydrates (Lawson, 1995). Their functional and textural characteristics 
contribute to the flavour and palatability of natural and prepared foods. They contain certain fatty 
acids which play an important role in nutrition and are also carriers of fat soluble vitamins. 
Vegetable oils are essential in meeting global nutritional demands and are utilized for many food 
and other industrial purposes (Idouraine et al., 1996). Despite the broad range of sources for 
vegetable oils, the world consumption is dominated by soybean, palm, rapeseed and sunflower 
oils with 31.6, 30.5, 15.5 and 8.6 million tonnes consumed per year respectively (Stevenson et 
al., 2007). In developing countries, Moringa oleifera (moringa) has potential to improve 
nutrition, boost food security, foster rural development and support sustainable land care 
(National Research Council, 2006).        
 Due to ever diminishing sources of fats and oils, there is the growing need for the search 
for new sources of oil, as well as exploiting sources that are currently unexploited in order to 
supplement the existing ones, since these conventional sources of vegetable oil no longer meet 
the ever increasing demands of domestic and industrial sectors. Olajide (2000) identified the 
locally available oilseeds in Nigeria that have been mostly exploited as groundnut, soybean, 
sunflower seed, melon, locust bean, conophor, beniseed, cocoa bean, palm kernel coconut, shea 
butter and cotton seeds while melon seeds, locust bean, oil bean, pumpkin seed, conophor nut, 
sheanut and moringa seeds are some of the underutilized oilseeds which are mainly used as 
ingredients in traditional food preparation even though they contain significant amount of oil. 
However, within the past few years, Moringa oleifera, a tropical, multipurpose tree has grown 
from being practically unknown, even unheard of, to being a new and promising nutritional and 
economic resource for developing countries, especially Nigeria. It is a very rapid growing tree 
found growing in a varying range of climatic conditions. The characteristics of Moringa oleifera 
1 
 
 
seed oil can be highly desirable especially with the current trend of replacing polyunsaturated 
vegetable oils with those containing high amounts of monounsaturated acids.     
1.2 The Research Problem 
 Previous studies on moringa have focused on its medicinal uses and nutritional aspects of 
the tree parts (Lowell, 1999) and on the use of the seed in the clarification of waste-water during 
treatment (Folkard et al., 1993). Though some previous researchers have reported on the oil 
extraction from moringa seeds, methods such as solvent and aqueous enzymatic extractions were 
used (Abdulkarim et al., 2005). Expellers represent one of the best choices for the extraction of 
oil from oilseeds. Methods such as water extraction and manual pressing only produce small 
amounts of oil; the extraction efficiencies (the percentage recovered from a possible maximum) 
are low and labour requirements high. Solvent extraction, while highly efficient involves very 
substantial capital cost and is only economic on a large scale. There is also a health and safety 
risk from using inflammable solvents. From the foregoing, the development of an oil expeller for 
moringa becomes highly desirable.         
 In order to design efficient equipment for the expression of oil from moringa, 
determination of the engineering properties, most especially the physical and mechanical 
properties is required. Some physical and thermal properties of moringa seeds have been 
determined (Aviara et al. 2011, Aviara et al. 2012, Adejumo and Abayomi 2012). The potential 
of moringa in producing oil and fat is far from being fully exploited as a result of lack of defined 
processing conditions necessary for optimum oil yield from the crop. Some of the needs 
triggering technology innovation in the oil extraction sector such as cost savings, environmental 
and safety concerns seem to be achievable by successful development of oil expeller. There is an 
extensive need to develop optimised and comprehensive protocols for oilseeds extraction to 
enhance greater oil recoveries. The high oil content of moringa which is between 30-40% 
(Mohammed et al., 2003; Anwar et al., 2006; Anwar and Rashid, 2007; Nzikou et al., 2009; 
Uzama et al., 2011; Adejumo et al., 2013; Ogunsina et al., 2011, Goja, 2013 and Orhevba, 2013) 
justifies the attraction of more interest to the need to optimise the oil expression process from the 
seed. To date, the optimisation of the oil expression process from moringa seeds has not yet been 
reported. In fact, a survey of literature revealed little information pertaining to the mechanical 
expression of oil from moringa seeds and no information as regards the optimisation of the 
various process parameters that influence oil expression from this crop using oil expeller. The 
present research will therefore be undertaken to optimise the oil expression process from 
moringa seeds.             
 According to Hamzat and Clarke (1993), to have an efficient process, it is necessary to 
2 
 
 
have a clear knowledge of the mechanisms of oil expression and an understanding of the 
important variables. This can be achieved by modeling oil recovery from oilseeds by using both 
theoretical and experimental methods. Prediction of oil yield can then be obtained in terms of 
process parameters. Development of mathematical models would be useful in predicting yield 
for moringa seeds at different processing conditions.    
 Overall, the research work would generate data that can serve as a useful tool in process 
and equipment design for the moringa oil processors in developing countries, most especially 
Nigeria. This will assist in improving the yield and quality attributes, thereby making more oil 
and fat available both for domestic and industrial purposes.    
1.3 Objectives of the Study        
 The primary objective of this research work is to optimise mechanical oil expression 
from moringa seeds. In pursuance of this, the following objectives arose which were to:                 
(1). study the engineering properties of moringa seeds in relation to the design of an oil     
 expeller,                             
(2). develop and evaluate the performance of an oil expeller for moringa seeds,                 
(3). develop mathematical models relating oil yield to the processing variables.                                          
(4). optimise the oil yield and process parameters for moringa using Response Surface 
 Methodology (RSM), and 
(5). investigate the effects of processing conditions on some physicochemical properties 
 of the expressed oil.  
 
 
 
 
 
 
 
 
 
 
 
 
 
 
3 
 
 
CHAPTER TWO 
LITERATURE REVIEW 
2.1 Moringa oleifera         
 Moringa oleifera is the most widely cultivated species of the genus Moringa, which is the 
only genus in the family Moringaceae. There are about thirteen species of moringa trees in the 
family (Price, 2007). They are native to India, the Red Sea area and parts of Africa. Of these 
species, Moringa oleifera is the most widely known. In this research work, the term “Moringa” 
refers to Moringa oleifera. Moringa is esteemed as a versatile plant due to its multiple uses. The 
leaves, fruits, flowers and immature pods of this tree are edible; and they form a part of 
traditional diets in many countries of the tropics and sub-tropics (Siddhuraju and Becker, 2003; 
Anhwange et al., 2004). The leaves are a good source of protein, vitamins A, B and C and 
minerals such as calcium and iron (Dahot, 1988). In addition to its substantial uses and 
nutritional benefits, it also has a great potential as a medicinal plant. The flowers, leaves and 
roots are used for the treatment of ascites, rheumatism and venomous bites and as cardiac and 
circulatory stimulants (Metha and Aggarawal 2008). The roots of the young tree and also root 
bark are vesicant (Hartwell, 1995; Anwar & Bhanger, 2003; Anwar et al., 2007). The seeds from 
this plant contain active coagulating agents characterized as dimeric cationic proteins. They also 
have antimicrobial activity and are utilized for waste water treatment. In some developing 
countries, the powdered seeds of moringa are traditionally utilized as a natural coagulant for 
water purification because of their strong coagulating properties for sedimentation of suspended 
undesired particles (Kalogo et al., 2000; Anwar et al., 2007).    
 Moringa seed kernels contain a significant amount of oil that is commercially known as 
"Ben oil" or "Behen oil". Moringa seeds, harvested from their pods, yield approximately 30-40% 
of non-drying moringa oil (Mohammed et al., 2003; Anwar et al., 2006; Anwar and Rashid, 
2007; Nzikou et al., 2009; Uzama et al., 2011; Adejumo et al., 2013; Goja, 2013 and Orhevba et 
al., 2013). Orhevba et al. (2013) reported that about 3000 kg of seeds could be obtained from 
one hectare, equivalent to about 900 kg oil/hectare, comparable to soybean which also yields an 
average of 3000 kg seeds/hectare but with only 20% oil yield. The oil is edible and closely 
resembles olive oil in its fatty acid composition. The free fatty acid content varies from 0.5 to 
3%. The seed oil contains approximately 13% saturated fatty acid and 82% unsaturated fatty 
acid. It has a particularly high level of oleic acid (70%). Other vegetable oils normally contain 
only about 40% oleic acid. It can be used in cooking, cosmetics, fuel and lubrications amongst 
others. It is a non-drying nutty flavoured oil with a clear or slightly pale yellow consistency. 
Moringa seed oil content and its properties show a wide variation depending mainly on the 
4 
 
 
species and environmental conditions (Ibrahim et al., 1974). The oil is obtained by mechanically 
pressing the seeds of moringa or by solvent extraction. Moringa seed oil is clear and odourless. 
Due to the numerous antioxidants in it, the oil does not become rancid for several years after it is 
produced. This makes it sought after for a number of health and beauty applications. Although it 
is viable for use as cooking oil, its high demand and low levels of production do not make it 
conducive for everyday use as a dietary product.  
 Moringa is an exceptionally nutritious vegetable tree with a variety of potential uses. 
Every part of it such as the seed, root and stem is useful. In the tropics, it is used as forage for 
livestock; in many countries, it is used as a micronutrient powder to treat diseases. The green 
pods, fresh and dried leaves are used as vegetable (Folkard et al., 2004; National Research 
Council, 2006). The seeds contain 30-40% of oil by weight which is used for cooking, soap 
manufacture, cosmetic base and in lamps. All parts of the plant are used in a variety of traditional 
medicines. The press cake, obtained following oil extraction, is useful as a soil conditioner; the 
plants are grown as live fences and windbreaks. It is also used as an intercrop with other crops 
and the wood pulp may be used for paper-making (Folkard et al., 2004). Moringa trees have 
been used to combat malnutrition, especially among infants and nursing mothers. The leaves can 
be eaten fresh, cooked, or stored as dried powder for many months without refrigeration, and 
reportedly without loss of nutritional value. It is especially promising as a food source in the 
tropics because the tree is in full leaf at the end of the dry season when other foods are typically 
scarce (Brett, 2005; Fahey, 2005). According to Oliver-Bever (1986), moringa leaves contain 
more Vitamin A than carrots, more calcium than milk, more iron than spinach, more Vitamin C 
than oranges, and more potassium than bananas, and that the protein quality of the leaves rivals 
that of milk and eggs. However, the leaves and stem are known to have large amounts of their 
calcium bound in calcium oxalate crystals (Olson and Carlquist, 2001).    
 The tree's bark, roots, fruit, flowers, leaves, seeds, and gum are also used medicinally. 
The  flowers,  leaves  and  roots  are  widely  used  as  remedies  for  several ailments. The bark 
of the moringa tree should be scraped off because of its toxicity and the flesh of the root should 
be eaten sparingly (Oliver-Bever, 1986). Moringa seeds are effective against skin-infecting 
bacteria, Staphylococcus aureus and Pseudomonas aeruginosa (Oliver-Bever, 1986). They 
contain the potent antibiotic and fungicide terygospemin. Moringa seems to have  most  of  the  
food  nutrients  required  by  the  body  to  replenish  its defensive mechanisms. The Tonga 
people of Binga District in Zimbabwe use the root powder as an aphrodisiac and when it is 
mixed with milk, it is considered useful against asthma, gout and enlarged spleen or liver 
(Maroyi, 2007). It also helps in the removal of wind from the stomach and can be used to 
5 
 
 
alleviate ear and toothache (Oliver-Bever, 1986; Maroyi, 2007). The leaf juice has a stabilizing 
effect on blood pressure. The leaf juice controls glucose levels in diabetic patients. Fresh leaves 
and leaf powder are recommended for tuberculosis patients because of the availability of vitamin 
A that boosts the immune system. If leaf juice is used as diuretic, it increases urine flow and 
cures gonorrhea. Leaf juice mixed with honey treats diarrhoea, dysentery and colitis (colon 
inflammation).  Fresh  leaves  are  good  for  pregnant  and  lactating  mothers;  they  improve  
milk  production  and  are  prescribed  for anaemia. Paste from ground bark can be applied to 
relieve pain caused by snake, scorpion and insect bites. Oil is sometimes applied externally for 
skin diseases (Maroyi, 2007).         
 Fully  mature,  dried  seeds  are  round  or  triangular  in  shape,  where  the  kernel  is 
surrounded by light wooded shell with three papery wings. When mature, the seeds from the 
pods can be extracted and treated like green peas and can be fried or roasted and eaten like 
peanuts. The seeds also contain oleic acid-type oil. The oil possesses about 75% oleic acid, a 
monounsaturated fatty acid that is less vulnerable to oxidative stress than unsaturated fats. Oleic 
acid has the ability to reduce inflammation in the system, since oleic acid appears to be one of 
the main protective agents in reducing the levels of cardiovascular disease, breast and skin 
cancer (Pauwels, 2011). The oil has high antioxidant properties, making it a valuable source of 
vitamin A, C and E. It is one of the highest naturally occurring sources of antioxidants (Dogra et 
al., 1975). The oil is good for skin formulation products because of its potent antioxidant 
inhibition which prevents bacterial infections, reduces inflammation and provides a smooth and 
healthy tone to the arms, legs and face. The oil possesses the following properties of anti-
flammatory, anti-hypertensive, anti-epileptic, anti-oxidant, anti-bacterial, antifungal and 
antipyretic (Ojiako and Okeke, 2013). It is used in all kinds of cosmetic products and soap 
(Delaveau et al., 1980). The oil can also be used as a natural source of behenic acid, which has 
been used as an oil structuring and solidifying agent in margarine, shortening, and foods 
containing semi-solid and solid fats, eliminating the need to hydrogenate the oil (Foidl et al., 
2001).           
 Moringa seeds are widely used as a natural coagulant for water treatment in developing 
countries (Santos et al., 2005). It is very good and safe for water treatment as synthetic chemical 
compounds (alum) may be carcinogenic (Ayotunde et al., 2011). Solutions of moringa seeds for 
water treatment may be prepared from seed kernels or from the solid residue left over after oil 
extraction (press cake). Moringa seeds, seed kernels or dried press cake can be stored for long 
periods, but moringa solutions for treating water should be prepared fresh each time. In general, 
one seed kernel will treat one liter (1.056 qt) of water (Doerr, 2005). Moringa seeds treat water 
6 
 
 
on two levels, acting both as a coagulant and an antimicrobial agent. Moringa works as a 
coagulant due to positively charged, water-soluble proteins, which bind with negatively charged 
particles (silt, clay, bacteria, toxins etc) allowing the resulting flocs to settle to the bottom or be 
removed by filtration. The antimicrobial aspects of moringa continue to be researched into. 
Findings support recombinant proteins both removing microorganisms by coagulation as well as 
acting directly as growth inhibitors of the microorganisms. While there is ongoing research being 
conducted on the nature and characteristics of these components, treatments with moringa 
solutions will remove 90-99.9% of the impurities in water (Doerr, 2005).    
 Plates 2.1-2.5 show the moringa plant, tree and seeds. 
2.2 Engineering Properties of Agricultural Materials  
 According to Mohsenin (1970), the ever increasing importance of agricultural products 
together with the complexity of modern technology for their production, processing and storage 
need a better knowledge of their engineering properties so that machines, processes and handling 
operations can be designed for maximum efficiency and the highest quality of the final end 
products. A rational approach to the design of agricultural machinery and equipment involves the 
knowledge of the engineering properties of the agricultural product concerned. There are several 
engineering properties of agricultural materials namely physical, mechanical, thermal, electrical, 
optical and electromagnetic properties. Often times, the physical and mechanical properties are 
the most determined. Peleg and Bagly (1982) defined physical properties of agricultural 
materials as those properties that lend themselves to description and quantification by physical 
means. Chukwu and Sunmonu (2010) defined mechanical properties as properties that have to do 
with the behaviour of agricultural products under applied forces. According to Corrêa et al. 
(2007), the knowledge of physical and mechanical properties of the agricultural products is of 
fundamental importance for proper storage procedure and for design, dimensioning, 
manufacturing, and operating different equipment used in postharvest and processing operations 
of these products. The knowledge of the engineering properties is useful for both engineers and 
food scientists, plant and animal breeders and it is also important in data collection in the design 
of machines, structures, processes and controls, and in determining the efficiency of a machine 
or an operation (Chukwu and Sunmonu, 2010).  
7 
 
 
 
Plate 2.1. The Moringa Plant 
 
Plate 2.2. The Moringa Tree 
8 
 
 
 
Plate 2.3. Moringa Pods 
 
 
 
Plate 2.4. Hulled Moringa Seeds 
9 
 
 
 
Plate 2.5. Dehulled Moringa Seeds 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
10 
 
 
 The size and shape are important in the electrostatic separation of agricultural materials 
from undesirable materials and in the development of sizing and grading machinery (Mohsenin, 
1980). The size and shape of agricultural materials also affect handling losses during cleaning 
and oil expression. According to Olayanju (2002), if the screen hole of the cleaner is too big, it 
may result in uncleaned seeds and if the screen hole is too small, it may lead to reduced 
efficiency. İf the oil barrel clearance is too big, it may result in partial crushing of seeds and if it 
is too small, it may lead to excessive choking of the discharge section as the seeds are crushed. 
For optimum performance of the cleaner and oil expeller, the size of perforations and barrel 
clearance have to be carefully selected (Olayanju, 2002).       
 Sphericity will be useful in handling operations such as conveying and discharge from 
chutes. Since agricultural materials are transferred from one placement unit to the other, the 
sphericity will be taken into consideration for designing the slope of the transfer unit. Bulk 
density, true density, and porosity (the ratio of intergranular space to the total space occupied by 
the grain) can be useful in sizing grain hoppers and storage facilities; they can affect the rate of 
heat and mass transfer of moisture during aeration and drying processes. Grain bed with low 
porosity will have greater resistance to water vapour escape during the drying process, which 
may lead to higher power to drive the aeration fans (Olayanju, 2002). Cereal grain densities are 
useful in breakage susceptibility and hardness studies (Seifi and Alimardani, 2010). The bulk 
density is important in calculating thermal properties in heat transfer processes, in determining 
Reynolds number in pneumatic and hydraulic handling of the material, in separating the product 
from undesirable materials and in predicting physical structures and chemical composition. The 
porosity gives a knowledge of the percentage void of the agricultural materials and is important 
in heat and airflow studies (Olayanju, 2002). Bulk density, grain density and porosity are major 
considerations in designing the drying, aeration and storage systems, as these properties affect 
the resistance to air flow of the mass. The static coefficient of friction is used to determine the 
angle at which chutes must be positioned in order to achieve consistent flow of materials through 
the chute. Such information is useful in sizing motor requirements for grain transportation and 
handling (Ghasemi Varnamkhasti et al., 2007). Gumble and Maina (1990) observed that angle of 
repose and coefficient of friction are important in designing equipment for solid flow and storage 
structures and the coefficient of friction between seed and wall in the prediction of seed pressure 
on walls. It is important in filling flat storage facility when grain is not piled at a uniform bed 
depth but rather is peaked (Mohsenin, 1980).      
 A vast knowledge of mechanical properties such as hardness and compression tests is 
vital to engineers handling agricultural products. According to Anazodo (1983), the 
11 
 
 
determination of mechanical properties of agricultural products under static or dynamic loading 
is aimed at the reduction of mechanical damage to agricultural produce during postharvest 
handling, processing, and storage and the determination of design parameters for harvesting and 
postharvest systems. Sitkei (1986) reported that most agricultural products are visco-elastic in 
nature, they respond differently to tensile or compressive forces and also behave differently 
when they are subjected to vibration.                             
2.3 Oil Extraction Methods        
 The common methods of extracting oil include: water assisted (finely ground oilseed is 
either boiled in water and the oil that floats on the surface is skimmed off; or ground kernels are 
mixed with water, squeezed and mixed by hand to release the oil); manual pressing (oilseeds, 
usually pre-ground, are pressed in manual screw presses); expelling (using an expeller which 
consists of a motor driven screw turning in a perforated cage; the screw pushes the material 
against a small outlet, called the choke; great pressure is exerted on the oilseed fed through the 
machine to extract the oil); ghanis (consisting of a large pestle and mortar rotated either by 
animal power or by a motor; seed is fed slowly into the mortar and the pressure exerted by the 
pestle breaks the cells and releases the oil); and solvent extraction (where oil from seeds or the 
cake remaining from expelling is extracted with solvents and the oil is recovered after distilling 
off the solvent under vacuum).   
 2.3.1 Water Assisted         
 Oilseeds in most cases are ground to a paste without removing the husk or outer covering. 
In some instances, for example, sunflower, the seeds are husked. The seeds are ground manually 
and the paste is heated alone at first and then with boiling water. The mixture is stirred and 
boiled. After boiling, the mixture is allowed to cool and the oil settles at the top where it is 
scooped off. With this method of processing, the extraction  efficiency is about 40%, that is 
percentage of oil extracted based on the total theoretical content (Ibrahim and Onwualu, 2005). 
For oilnuts, the processing methods vary because of the variation in the procedures. For a seed 
like groundnuts, they are shelled, cleaned and roasted lightly. The roasted nuts are skinned by 
placing them on a mat and rolling a wooden block over them, and winnowing them to separate 
the skin from the nuts. The skinned nuts may be pounded with a mortar and pestle or ground 
using grinding stones to a smooth paste. The paste is kneaded and pressed by hand to remove the 
oil-water mixture. Then the oil-water mixture is fried to remove most of the water (Ibrahim and 
Onwualu, 2005). 
12 
 
 
 2.3.2 Manual Pressing        
 Mechanical pressing (hydraulic and screw) is the oldest and simplest method for oil 
extraction from seeds. No chemical is used for oil extraction and therefore the residue is free of 
chemical. It is a labour intensive technique and its use declined over the years. Continuous 
screw-presses (expellers) have replaced the hydraulic equipment (Bargale 1997). It consists of an 
extruder with a perforated body (slots or holes) and a helical screw is used to convey and press 
out the oil (Balke 2006). A major drawback of this process is the lower oil extraction. Heat pre-
treatment improves the malleability of the seed, lower the shattering and denature some proteins 
to improve oil extractability (Balke 2006). Before pressing, the seed are flaked to crush the seed 
shell and for better oil diffusion (Ward 1982). The extracted oil is of superior quality, but intense 
pressure and heat damages the seed protein (Balke 2006). This method is popular in developing 
countries for low operating costs than the solvent extraction (Bargale 1997). The extreme heat 
generation during this method causes darkening of the oil (Bargale 1997). The choking and 
plugging problem results in loss of production capacity and increases in energy, labour and other 
resource requirements (Rosenthal et al., 1996).       
 Types of manual press include spindle press, bridge press (also known as a screw press), 
ram press and hydraulic press. Some different types of manual press are shown in Figures 2.1-
2.5. Manually powered spindle presses are usually small table mounted devices with a capacity 
of around 2-5 kg per hour. The bridge press comprises of a cylinder that contains the seed which 
is compressed by rotating a screw down onto the seed. The screw is held in place by a frame that 
bridges over the seed container. As the seed is compressed, the oil drains through holes in the 
cylinder onto a collection tray. The process is relatively slow as the cylinder needs to be filled, 
compressed and then the remaining cake needs to be removed. Ram presses use a lever 
mechanism to produce high pressures on a piston that forces the oil out of the seed. Hydraulic 
presses use a hydraulic pump to exert a high pressure on the seed. Hydraulic jacks from cars and 
trucks can be used. The process is similar to a screw press, in that the seed has to be loaded into a 
cylinder and then pressed to extract the oil, which runs onto a collection tray. Once the seed has 
been pressed, the remaining cake needs to be removed. Capacities are from around 1 kg per press 
(Practical Action, 2008). 
 
 
 
13 
 
 
 
Figure 2.1. Ram Press 
(Source: http://www.fao.org) 
 
Figure 2.2. Oil Press with Hydraulic Jack 
(Source: http://www.fao.org) 
  
Figure 2.3. Bridge Press (NRI design) 
(Source: http://www.appropedia.org) 
14 
 
 
 
Figure 2.4. Scissor Press (IPI design) 
(Source: http://www.appropedia.org) 
 
Figure 2.5. Curb Press (TCC design) 
(Source: http://www.appropedia.org) 
15 
 
 
 2.3.3 Oil Expellers          
 Expellers use a horizontally rotating metal screw, which conveys oil-bearing products 
into a barrel shaped outer casing with perforated walls (Figure 2.6). The products are 
continuously fed to the expeller, which grinds and presses the oil out as it passes through the 
machine. The pressure ruptures the oil cells in the product and oil flows through the perforations 
in the casing and is collected in a trough underneath (Gate, 1979). The residue of the material 
from which oil has been expressed exits from the unit, and is known as the cake. Most expellers 
are power-driven, and are able to process between 8 and 45 kg/hr of product depending upon the 
type of expeller used. Bigger units, processing greater quantities of oil are available for use in 
larger mills. The percentage of oil expressed by expellers is as high as 90% depending upon the 
type and kind of products as well as the expeller being employed (Gate, 1979). The friction 
created by  the  products  being  expressed  wears  down the  worm  shaft  and  other  internal 
parts. With small machines, this occurs often after expressing as little as 50 tons, after which 
parts must be replaced or repaired through resurfacing by welding. The use of oil expellers are 
most unlikely at the village/small town level because the maintenance requires machinery and 
equipment that are rarely found in small repair shops and local manufacture of expellers. 
 2.3.4 Ghanis         
 Ghanis originated in India where they are primarily used to express oil from mustard and 
sesame seeds, although in some cases they can be used for coconut and groundnut processing. 
Traditionally, ghanis are operated by animals and can be manufactured locally. They consist of a 
wooden mortar and wood or stone pestle. The mortar is fixed to the ground while the pestle, 
driven by one or a pair of bullocks or draught animals is located in the mortar where the seeds 
are crushed by friction and pressure (Figure 2.7). Oil runs through a hole at the bottom of the 
mortar, while the residue or cake is scooped out. Depending on the size of mortar and type of 
seeds, its capacity is approximately 40 kg a day (Practical Action, 2008); although this will vary 
depending on the size, strength and number of animals used. Animals need to be replaced after 3 
or 4 hours work as they tire. Mechanized versions of the traditional animal-powered ghanis are 
common. In these power-driven ghanis, the pestle and mortar units are usually arranged in pairs 
with either the pestle or mortar held stationary, while the other is rotated (Ibrahim and Onwualu, 
2005). Power ghanis (Figure 2.8) have a greater capacity than the traditional ghanis and can 
process about 1000 kg of seeds per day (Srikanta Rao, 1978). Power ghanis yield an oil with a 
lower pungency. 
16 
 
 
 
Figure 2.6. Outline Drawing of a Typical Small-Scale Expeller (CeCoCo Type 52)  
(Source: http://www.appropedia.org) 
 
Figure 2.7. Traditional Animal Powered Ghani  
(Source: http://www.fao.org) 
 
Figure 2.8. Power Ghani  
(Source: http://www.fao.org) 
17 
 
 
 2.3.5 Solvent Extraction       
 Solvent extraction of oil from oilseeds is the most efficient and attractive method for low 
oil content seeds (Anjou 1972; Caviedes 1996), for high oil content seeds like sunflower, peanut 
and canola, and also for medium oil content seeds like cottonseed and corn germ (Ward 1976). 
Therefore, vegetable oils are mostly extracted by using solvent extraction (Kemper 2005). 
Solvent extraction method involves the use of organic solvents such as straight chain 
hydrocarbons, chlorinated hydrocarbons, alcohols and ketones to recover the oil from the 
sources. The process for solvent extraction of nut like groundnut is similar to that of seeds like 
soyabean. Generally, nuts or seeds are shelled and winnowed to remove fibre-rich shells. Next, 
o
the nuts or seeds are cracked into pieces and conditioned to 10-11% moisture at 70 C or more, 
and then flaked by passing through rolls. Sometimes the nuts or seeds flakes are cooked before 
they are conveyed to the extractor. In the extractor, the oil is removed by means of a solvent. The 
solvent ladden flakes are then passed through a desolventizer, which recovers the solvents. The 
defatted and desolventized cake may undergo further treatment before it is used as feed. The 
crude oil may be clarified by passing it through a filter press (Ibrahim and Onwualu, 2005). 
Solvent extraction is capable of removing nearly all the available oil from oilseeds or nuts. About 
98% of the oil is being extracted by solvent method (Ngoddy and Iherokoronye, 1985; Cecoco, 
1988). In addition to the high yield of oil, the method produces oil with better qualities, and a 
higher protein meal (Khan and Hanna, 1983). The method generally requires more capital 
expenditure, and refining the oil before use. There are also possibilities of toxicity from the 
solvent used and danger of fire explosion from the use of volatile organic solvent. 
2.4 Effects of Processing Factors on Mechanical Expression of Oil from Oilseeds 
 Efficient expression of vegetable oils from oil bearing seeds by mechanical method 
requires optimum preliminary processing before the expression (Ibrahim and Onwualu, 2005; 
Mwthiga and Moriasi, 2007). Moisture content, heating temperature, heating time and applied 
pressure affect the yield of fats and oil during expression (Khan and Hanna, 1984). For 
maximum oil recovery and least residual oil in the cake, it is necessary to control these factors 
during the oil or fat expression process. Inability to control them could lead to failure in getting 
high yield and good quality fats and oil during expression (Bamgboye and Adejumo, 2011a). 
Therefore, efficient processing of the oil bearing materials prior to expression is necessary in 
order to achieve higher yield and subsequently cheaper end product.  
 2.4.1 Effects of Moisture Content       
 The optimum moisture contents for oil expression have been established for different 
18 
 
 
oilseeds. Investigations by Abidakun et al. (2012) while working on dika nut revealed that oil 
yield was higher at the higher moisture contents of 6 and 9% wet basis, with the maximum oil 
recovery obtainable at 6% moisture content. Olajide (2000) reported that the increase in moisture 
content showed a substantial improvement in oil recovery in both groundnut and sheanut kernels. 
However, further increase in moisture content resulted in a decline in oil expression from both 
the groundnut and sheanut kernels. Optimum moisture content of 6.6% w.b. and 13% w.b. were 
obtained for groundnut and sheanut kernels respectively. Bamgboye and Adejumo (2011a) while 
working on roselle seeds observed oil yield increase of 7-8% as the moisture content increases. 
Farsaie and Singh (1983) showed that the maximum oil recovery was obtained when sunflower 
seeds were expressed at 6% moisture content and increasing the moisture content to 14% 
decreased oil recovery by 16%. According to Bongiriwar et al. (1977), as the moisture content of 
groundnut increased up to 6%, the percentage oil removed increased and the yield decreased 
beyond 6% moisture content. Southwell et al. (1990) reported that a moisture content of 9% was 
found to be optimum for expression of oil from avocado fruit. In the case of canola seed, 
moisture content in excess of 9% adversely affected the oil yield (Blake, 1982); while 10% 
moisture content and 11% moisture content gave the maximum oil recovery for unsieved rice 
bran and sieved rice bran respectively (Sivala et al., 1992). It was reported that increase in 
moisture content led to increase in yield of oil for sunflower kernels from 53.2% to 58.2% 
(Southwell and Harris, 1992); while maximum oil could be obtained from grated coconut if the 
moisture content of expression falls within 10.9% to 12.9% (Hammonds et al., 1991). According 
to Fasina and Ajibola (1989), maximum oil yield was obtained from conophor nuts when 
conditioned to 11% moisture content. For melon seeds, Ajibola et al. (1990) found the maximum 
oil yield when the seeds were conditioned to a moisture content of 9.2%. Dedio and Dorrell 
(1977) observed that a moisture content of 8% (wet basis) was found to be the optimum for 
expression of oil from flaxseed. Williams and Rathod (1983) reported that a moisture content of 
7-8% gave the best oil yield from soybean. For mechanical oil expression from mustard seeds, 
maximum oil yield was obtained when the seeds were conditioned to a moisture content of 8.7% 
(Reddy and Bohle, 1993). Akinoso et al. (2006) while working on sesame seeds observed that 
decreased moisture content caused increased oil yield. Maximum oil yield of 50.4% at an 
optimum moisture content of 4.6% wet basis was obtained.  
 2.4.2 Effects of Applied Pressure       
 The optimum pressure necessary for maximum oil expression from the different oilseeds 
have been determined as the application of more pressure than required reduces the pressing 
19 
 
 
efficiencies, leads to a higher cost of expression and decreases oil yield. Ward (1976) reported 
that during the process of oil expression from oilseeds, increasing the pressure applied during 
screw pressing tends to decrease the size of the capillaries through which oil flows and further 
increase in pressure may eventually lead to the sealing of the capillaries. Bamgboye and 
Adejumo (2011a) observed a decrease in oil yield as the applied pressure increased from 30 to 
37.5 MPa for roselle seeds. In their investigation, there were steady increases in the oil yield up 
to 30 MPa before the yields started decreasing. This was attributed to the fact that increasing the 
pressure applied on oilseeds during oil expression could narrow, shear, and may eventually seal 
the capillaries through which oil drains out during oil expression causing reduction in the oil 
yield despite the fact that there has been an increase in the applied pressure according to Ward 
(1976). According to Abidakun et al. (2012) on dika nut, it was observed that oil yield increased 
for all pressure levels ranging from 5-25 MPa with the highest at 25 MPa. This was also 
observed for soybean as reported by Mwthiga and Moriasi (2007). Reporting from studies 
conducted on oil expression from groundnut, Pominski et al. (1970) found out that the amount of 
oil expressed from the seed tends to level off at a pressure of 13.8 MPa. Adeeko and Ajibola 
(1990) equally observed that oil yield increase with increase in pressure up to 20 MPa beyond 
which the yield levelled off, but decrease in oil yield was observed at 25 MPa. Olajide (2000) 
observed significant increase in oil yield when applied pressure was increased from 5-15 MPa, 
but decreased when the pressure was increased to 25 MPa. A similar phenomenon was observed 
while working on sheanut kernels too. This decrease was attributed to the sealing of some inter-
kernel voids at that increased pressure. Ajibola et al. (1990) observed that there was a significant 
increase in oil yield from melon seeds when applied pressure was increased from 5-18 MPa, but 
oil yield either levelled off or decreased slightly when the pressure was increased to 25 MPa. 
Sivala et al. (1992) observed that increase in applied pressure increased the oil yield up to 25.5 
MPa beyond which there was a decrease in oil yield. For unsieved and sieved bran samples, 
maximum oil yields of 55.9% and 50.4% respectively were obtained at a pressure of 25.5 MPa. 
Adekola (1991) reported that oil yield increased with increase in applied pressure during oil 
expression from coconut. Fasina and Ajibola (1989) while working on oil expression from 
conophor nuts observed an increase in oil yield as pressure increased from 10 MPa to 25 MPa. 
Maximum oil yield of 39.6 % was obtained at an applied pressure of 25 MPa. For mustard seeds, 
Reddy and Bohle (1993) reported that the lowest and highest oil yields of 22.78% and 28.93% 
were obtained at 3.924 MPa and 7.848 MPa applied pressures respectively. Increased applied 
pressure resulted in increased oil yield.      
20 
 
 
 2.4.3 Effects of Heating Temperature and Time    
 The optimum heating temperature and time for oil expression have been established for 
different oilseeds. Adekola (1991) reported that oil yield increase with increase in heating 
o
temperature subject to duration of heating for coconut. Highest oil yield was obtained at 120 C. 
o
However, at heating times longer than 15 mins at 120 C, oil yield reduced by 12.3%. This 
indicated an optimum heating time of 15 mins for maximum oil yield from coconut. Adeeko and 
Ajibola (1990) reported that oil yield increased by an increase in heating temperature and heating 
time for groundnut. An increase in heating temperature increased the oil yield of samples heated 
for 15 and 25 mins. For samples heated at 35 and 45 mins, there was no significant difference in 
o
the yield of oil samples heated at different temperatures. Oil yield from samples heated at 135 C 
was highest at 15 mins and further increase in heating time decreased the yield. Samples heated 
o
at 160 C increased in yield from 33% to 37% when heating time was increased from 15 to 25 
mins. A further increase in time of heating led to a decrease in yield. Hammonds et al. (1991) 
o
reported that moderate heating of groundnut to about 60 C resulted in an improvement in oil 
yield as compared to expression without heating under room conditions. South-well and Harris 
(1992) observed that sunflower kernels which had been heated and conditioned gave a higher oil 
yield when compared with unheated sample. Reports from the investigation carried out on 
mechanical expression of conophor nut by Fasina and Ajibola (1989) showed that oil yield 
o o
increased from 50-65 C, but decreased with increase in heating temperature to 110 C. It was 
o
observed that oil yield obtained from samples heated at 50 C were considerably lower than those 
obtained from the other heating temperatures. Oil yield increased with increased heating time for 
o
samples heated at 110 C. High oil yields were obtained at heating times of 20 and 28 mins for 
o
samples heated at 65, 80 and 95 C. Highest oil yield of 39.6% was obtained when sample was 
o
heated at 65 C for 28 mins. Ajibola et al. (1990) observed an increase in oil yield from melon 
seeds with increase in temperature and heating time for the range of variables considered in their 
o
study. Maximum oil yield of 41.6% was obtained when samples were heated at 130 C for 20 
mins. Bamgboye and Adejumo (2011a) observed an increase in the oil yield as the heating 
o
temperature is increased to 100 C for roselle seeds. Increase in the heating temperature caused 
the protein to coagulate at a very fast rate, thus reducing the viscosity significantly and adjusting 
the moisture content which led to the release of the oil bound to them. Akinoso et al. (2006) 
while working on sesame seeds observed that increased heating temperature and time caused 
increased oil yield. Maximum oil yield of 50.4% at optimum heating temperature and time of 
o
124.2 C and 13 mins respectively was obtained. Abidakun et al. (2012) while working on dika 
nut observed an increase in oil yield with increase in heating temperature and heating time. 
21 
 
 
However at higher heating temperatures, the oil yield decreased continuously with increase in 
o
heating time. Oil yield reached the maximum at 5 mins when it was heated at 150 C and declined 
o
continuously thereafter. The highest oil yield was obtained from sample heated at 100 C for 10 
mins. The continuous decrease in oil yield at higher heating temperatures could be due to the fact 
that the moisture content decreased further as the heating increased according to Mwthiga and 
Moriasi (2007). Abidakun et al. (2012) observed that the combination of heating temperature and 
heating time is an integral factor affecting oil yield. This is explicable as it represented the 
quantity of heat energy added to the sample. Therefore, if the heat energy exceeded the latent 
heat of vaporisation, the moisture content of the meal is readjusted and if it falls below a 
threshold, the oil expression is obstructed.         
2.5 Interaction between Process Variables and Oil Yield   
 Processing variables in oilseed expression affect the yield drastically, no matter how 
efficient the methods and machines employed. It has been established that there are interactions 
between process variables and oil yield from oilseeds.     
 Singh et al. (1984) reported that applied pressure, time, temperature and moisture content 
were the variables that affect oil yield from sunflower seeds. It was observed that at moisture 
content above 6%, the oil yield decreases. The highest oil yield of 83.3% was obtained at a seed 
o
moisture content of 6%, applied pressure of 42 MPa and temperature of 20 C. Khan and Hanna 
(1984) observed that the oil yield from soybean was improved by increasing the temperature and 
pressure with a sample moisture content of 9.5-10%. Maximum oil yield of 85.71% was obtained 
o
when soybean flakes conditioned to 9% moisture was heated at 60 C at a pressure of 45 MPa. 
Fasina and Ajibola (1989) observed that oil yield from conophor nut at any pressure was 
dependent on the moisture content of the sample, heating temperature and time. A maximum oil 
yield of 39.6% was obtained when milled sample, conditioned to 11% moisture was heated at 
o
65 C for 28 mins and expressed at a pressure of 25MPa. Ajibola et al. (1990) reported that oil 
yields from melon seeds were affected by the seed moisture content, heating temperature and 
time. From the range of variables considered in their study, a maximum oil yield of 41.6% was 
o
obtained when samples conditioned to a moisture content of 9.2% were heated at 130 C for 20 
mins and expressed at 25 MPa. Adekola (1991) reported that moisture content, applied pressure, 
heating temperature and time are factors affecting oil yield from coconut. Highest oil yield of 
51.9% was obtained at an expression pressure of 25 MPa when samples were conditioned to 
o
moisture content of 7.31% wet basis and heated at 120 C for 15 minutes. Reddy and Bohle 
(1993) observed that the oil yield from mustard seeds was significantly affected by moisture 
22 
 
 
content of milled samples and applied pressure. For the range of variables considered in their 
study, a maximum oil yield of 28.93% was obtained when milled mustard seed was conditioned 
to a moisture content of 8.7% and a pressure of 7.848 MPa was applied. Olajide (2000) observed 
the combinations of process variables which are moisture content, applied pressure, heating 
temperature and time for maximum oil recovery from sheanut kernels. Optimum oil yield of 
o
46.7% when samples were conditioned to a moisture content of 13% wet basis, heated at 120 C 
for 50 minutes and expressed at 20 MPa was obtained. Tunde-Akintunde et al. (2001) 
investigated the effect of moisture content, heating temperature, heating time, applied pressure 
and pressing time on mechanically expressed soybean oil. Results showed that oil yield increased 
o
as moisture content was varied from 7.3 to 10.2%, heating temperature from 70 to 80 C and 
heating time from 15 to 30 mins. The highest oil yield of 10.4% was obtained when sample at 
o
10.2% moisture content was heated for 30 mins at a temperature of 80 C. Akinoso et al. (2006) 
investigated the effects of moisture content, roasting duration and temperature on the oil yield 
from sesame seeds. The optimum moisture content, roasting duration and roasting temperature 
o
were 4.6 % wet basis, 13 mins and 124.2 C respectively. These combinations gave 50.4 % oil 
yield. Bamgboye and Adejumo (2011a) investigated the effect of moisture content, applied 
pressure, heating temperature and time on roselle seed. The highest oil yield was obtained when 
o
sample was conditioned to a moisture content of 6.4%, heated at 100 C for 20 mins at an applied 
pressure of 30 MPa. Abidakun et al. (2012) observed that the various factors responsible for 
proper oil expression from dika kernel include moisture content, applied pressure, heating 
temperature and time. Maximum oil yield of 72.2% was obtained at 6% moisture content, 
o
heating temperature of 100 C, heating time of 10 mins and applied pressure of 25 MPa. 
Sivakumaran et al. (1985) revealed that moisture content, heating temperature and time were the 
interactive factors that influenced the expression of oil from runner type groundnut grown in the 
United States of America. Findings from the investigations showed that the maximum oil 
o
extraction of 92% was achieved at seed moisture content of 5.42%, heating temperature of 20 C 
and heating time of 27.4 mins. Adeeko and Ajibola (1990) identified heating temperature, 
heating time and pressure as interactive factors that influenced oil expression from groundnut. 
o
Maximum oil yield of 37.8% was obtained when sample was heated to 100 C for 25 mins and 
pressed at 25 MPa. Olajide (2000) investigated the combinations of process variables which are 
moisture content, applied pressure, heating temperature and time for maximum oil recovery from 
groundnut kernels. Optimum oil yield of 33.4% when samples were conditioned to a moisture 
o
content of 6.6% wet basis, heated at 95 C for 30 minutes and expressed at 20 MPa was obtained. 
Ajav and Olatunde (2011) examined the influence of moisture content, heating time and heating 
23 
 
 
temperature on the percentage oil yield of groundnut in order to establish the optimum process 
conditions. The optimum moisture content, heating duration and heating temperature were 6% 
o
wet basis, 25minutes and 100 C respectively which gave 41.6% oil yield.   
2.6 Effects of Processing Factors on Oil Quality      
 Processing factors affect oil quality and vary with several conditions which include 
differences of environmental conditions, climate cultivation, soil composition, maturity level of 
oil bearing seeds and the harvesting time  (Compaoré et al., 2011). The method of extraction also 
has significant effect on the oil quality (Ogunsina et al., 2011). Applied pressure, heating 
temperature, heating duration, moisture content, particle size, handling and storage are factors 
influencing yield and quality of vegetable oil expression (Weiss, 2000). The degree of influence 
varies with the type of oilseed and the method of oil expression used (Akinoso, 2006). Ohlson 
(1976) reported that processing conditions can have a rather negative effect on oil quality. In 
order to obtain high quality oils, it is very important to avoid deleterious effects such as long 
processing times, high temperatures, light and other catalysts. The effect of functional processing 
parameters on oil quality therefore becomes very important.    
 Pearson (1981) reported that for oil mechanically expressed from conophor nuts, the 
processing conditions considered did not have noticeable effect on the colour, specific gravity 
and iodine value. Ajibola et al. (1990) observed in their study on mechanical oil expression from 
melon seeds that processing conditions did not have significant effects on the colour and specific 
gravity of the expressed oil. The colour of the oil obtained was pale yellow. Adeeko and Ajibola 
(1990) while working on groundnut observed that increasing temperature and time of heating 
increased the colour intensity of the oils. This was attributed to the fact that with increase in 
temperature or time of heating, the quantity of moisture lost from the sample increased. Highest 
o
oil yield was obtained from samples heated at 160 C and 25 mins, but the colour of the oil was 
dark. According to Adeeko and Ajibola (1990), dark coloured oils require special refining, 
o
thereby increasing the cost of production. The colours of oils from samples heated at 135 C and 
o
160 C for longer than 25 mins were dark and unattractive. Increasing the heating temperature 
and time increased the free fatty acid. This was attributed to the fact that increasing the 
temperature increased the lipase activity, thus leading to increase in free fatty acid. 
Saponification value, iodine value and specific gravity of the oil were not affected by variations 
in processing conditions. Olajide (2000) investigated the effect of processing conditions on the 
colour, specific gravity, melting point, peroxide value, iodine value and saponification value of 
expressed oil from groundnut kernel. The results of the investigation revealed that increase in 
24 
 
 
o
heating temperature of groundnut kernel from 65 to 105 C increased the colour intensity of the 
expressed oil, while other properties were not significantly affected by processing conditions. 
Ajav and Olatunde (2011) in their study on groundnut oil expression observed that the colour of 
oil extracted was affected by the heating time and temperature. Increased heating temperature 
and time gave dark red oil whose intensity increased as the heating temperature and time 
o
increased. At 120 C, the oil extracted was of offensive burnt smell, with no increased oil yield 
observed. Adekola (1991) observed that the processing conditions studied on oil expression from 
coconut did not have noticeable effect on colour. Reddy and Bohle (1993) observed in their 
study of oil expression from mustard seed that processing conditions considered in the study did 
not have significant effects on the colour, specific gravity, free fatty acid and iodine value of oil. 
The mustard seed gave dirty yellow colour. Takeoka et al. (1997) investigated the effect of 
heating on the characteristics and chemical composition of seven commonly used frying oils and 
fats (beef tallow, canola oil, partially hydrogenated canola oil, corn oil, cottonseed oil, soybean 
oil, and partially hydrogenated soybean oil). It was observed that during heat treatment, a 
progressive decrease in unsaturation was observed in all oils by measurement of iodine value. 
This decrease was attributed to the destruction of double bonds by oxidation, scission, and 
polymerization (Cuesta et al., 1991). It was also observed that oil colour darkens as heating or 
frying proceeds. Olajide (2000) observed the effect of processing conditions on the colour, 
specific gravity, melting point, peroxide value, iodine value and saponification value of 
expressed oil from sheanut kernel. It was found out that increase in heating temperature of 
o
sheanut kernel from 90 to 130 C increased the colour intensity of the expressed oil, while other 
properties were not significantly affected by processing conditions. Olayanju (2002) reported 
that the colour intensity of beniseed oil increased as the moisture content increased. This was 
attributed to the fact that colours are formed from carbohydrates in food where there is a loss of 
one or more molecules of water from the carbohydrate. Akinoso (2006) studied the effects of 
moisture content, roasting duration and time on the free fatty acid, colour and oil impurity of 
sesame seed oil. It was observed that free fatty acid increased with moisture content, decreased 
with roasting duration and temperature. Oil impurity increased with increased moisture content, 
decreased with roasting duration and temperature. Also, the colour intensity increased with all 
the processing conditions. Akinoso (2006) also investigated the effects of moisture content, 
roasting duration and time on the free fatty acid, colour and oil impurity of palm kernel oil. It 
was observed that free fatty acid increased with moisture content, decreased with roasting 
duration and temperature. Oil impurity increased with increased moisture content, decreased 
with roasting duration and temperature. Also, the colour intensity increased with all the 
25 
 
 
processing conditions. Olaniyan (2010) while working on castor bean observed that an increase 
in the heating temperature increased the free fatty acid, saponification value, iodine value, total 
acid, peroxide value and the colour intensity of the oil expressed. However, increase in the 
heating temperature decreased the ester value. Specific gravity initially remained constant at 
o o o
30 C and 60 C but later decreased at 90 C, while refractive index remained constant at all 
o
temperatures. Viscosity decreased when the heating temperature was increased from 30 to 90 C. 
o
The colour of the oil was golden yellow and light brown at 30 and 60 C respectively, but 
o
changed to dark brown at 90 C. Ogunsina et al. (2011) investigated the physicochemical 
characteristics of cold pressed and hexane extracted moringa seed oils namely iodine value, 
saponification value and unsaponifiable matter. Results of the findings showed 67.8 and 68.5 
gI2/100g oil, 190.4 and 191.2 mg KOH/g oil and 0.59 and 0.65% in that order for cold pressed 
and hexane extracted moringa seed oils respectively. Adejumo et al. (2013) studied the effect of 
heat treatment on the quality of moringa oil extracted by soxhlet method. It was discovered that 
an increase in the heating temperature decreased the saponification value, free fatty acid, acid 
value, peroxide value and iodine value; while heating temperature had no significant effect on 
the specific gravity of the oil. Orhevba et al. (2013) investigated the effect of moisture content on 
some quality parameters of mechanically expressed neem seed kernel oil. The quality parameters 
include saponification value, iodine value, fatty acid, acid value and colour. It was discovered 
that the highest saponification value of 262.46 mg/KOH/mol was obtained at moisture content of 
8.1%, thereafter the saponification value decreased as the moisture content was increased. Iodine 
value increased as the moisture content increased from 6.1% to 8.1%, decreased sharply at 
13.2% and then increased sharply at 16.6%. Moisture content had increasing, decreasing and 
then increasing effects on the values of the free fatty acid. Moisture content had an upward and 
downward effect on acid value. It was also observed that moisture contents at higher levels 
affected the colour of the oil, which changed from brown to dark brown.       
2.7 Process Optimisation by Response Surface Methodology (RSM) 
 Response Surface Methodology (RSM), based on the combination of statistical and 
mathematical tools, is considered to be a valuable technique for the development, modification 
and optimisation of various processes (Montgomery, 2005; Raymond and Douglas, 2002). RSM 
has proven to be a useful tool for the analysis of problems during which a certain response of 
concern is usually influenced by different reaction variables with the purpose of optimising a 
defined response of interest. Modeling of experimental response was the main objective of using 
RSM, but later on, applications of RSM were extended to develop models for the optimisation of 
26 
 
 
numerical experiments (Box and Norman, 1987). When treatments are based on continuous array 
of values, then RSM can be used for the improvement, development and optimisation of 
response variables mathematically expressed as:  
𝑦 = 𝑓  𝑥1 𝑥2  + 𝑒             … (1) 
 Central Composite Design (CCD) has gained much attention in the recent years as the 
most acceptable second order design for the comprehensive estimation of response surfaces 
based upon second order models. Box and Wilson first introduced CCD in 1951 for response 
surface optimisation (Box and Wilson, 1951). CCD uses either a full factorial design with two 
k k–f
levels (2 ) or fractional factorial designs (2 ) fabricated with numerous design points. CCD is 
comprised of three types of design points including; factorial points nf, axial points na and central 
points nc. The following expression can be used for the cumulative design points:     
𝑛 = 2k  𝑛f  + 2𝑘 𝑛a  + 𝑘 𝑛c           … (2) 
 RSM is aimed at topographical understanding of response surfaces and region finding 
where we can find optimal responses (Montgomery, 2005).     
 Olajide (2000) used RSM to optimise the applied pressure, pressing time, heating 
temperature, heating time and moisture content for groundnut kernels and sheanut kernels oil 
expression processes. Optimum oil yield of 33.4% was obtained when coarsely ground 
o
groundnut kernels conditioned to a moisture content of 6.6% (wet basis) were heated at 95 C for 
30 minutes and expressed at 20 MPa for 6 minutes. For sheanut kernels, optimum yield of 46.7% 
was obtained when finely ground sheanut kernels conditioned to a moisture content of 13% (wet 
o
basis) were heated at 120 C for 50 minutes and expressed at 20 MPa for 6 minutes. Shridhar et 
al. (2010) used RSM to optimise the dilution level and agitation time for castor oil extraction. 
The percentage recovery of oil was investigated with respect to two variables including dilution 
level and agitation time. Optimal dilution level and agitation time of 7.3 and 2.38 hr respectively 
were obtained and the maximum extraction was found to be 48.75%. Cvjetko et al. (2012) 
optimised three operating parameters namely pressure, temperature and extraction time of 
supercritical CO2 extraction of oil from rapeseed using RSM to obtain high yield of oil. Optimal 
conditions for oil yield within the experimental range of their studied variables were 29.7 MPa, 
o
52.14 C and 3.36 h, and oil yield was predicted to be 28.27%. Results showed that data were 
adequately fitted into the second-order polynomial model. The linear and quadratic terms of 
independent variables of temperature, pressure and extraction time had a significant effect on the 
oil yield. Awolu et al. (2013) optimised neem oil production and its characterization using RSM. 
They studied the effects of three factors viz sample mass, particle size and extraction time on the 
27 
 
 
neem oil volume. 49% (47 ml) of neem oil was obtained at 45 g mass of sample, 1.39 mm 
particle size and 2 h extraction time. The Analysis Of Variance (ANOVA) results of the RSM 
showed that all the linear coefficients and the quadratic coefficients were significant (p = 0.05). 
2 2
R  and adjusted R  values of 0.9966 and 0.9935 were obtained respectively.      
2.8 Model Equations for Oil Expression Processes     
 Hamzat and Clarke (1993) stated that it is necessary to have a clear understanding of the 
mechanisms of oil expression and the important process variables in order to be able to design an 
oilseed press in an optimal manner and also to have efficient process. This can be achieved by 
modelling the oil recovery from oil-crops by using both theoretical and experimental methods. 
Mathematical models are also useful in predicting process efficiencies for a wide range of 
processing conditions. Several investigations have been carried out to obtain predictions of oil 
yield in terms of process variables.         
 Singh et al. (1984) developed mathematical models to predict oil yield from sunflower in 
terms of process variables such as moisture content, applied pressure, pressing duration, heating 
temperature and pressing time. Mathematical models were developed for different types of seeds 
and are presented as follows:                 
For whole seed:                            
𝑅0 = −77 + 13.8𝑀 + 0.25𝑃 + 0.47𝑇 − 0.35𝑀
2 − 0.0038P2 + 0.002𝑇2  − 0.056MT      … (3) 
R = 0.93, s.e. = 1.00            
Where,                           
Ro = residual oil left in the cake, %                     
M = moisture content of the seed, % wet basis                  
P = pressure applied during the pressing, MPa                              
o
T = seed temperature before pressing, C       
 The multiple correlation coefficient R was 0.97 and the standard error of estimate was 
2.07. The model revealed that moisture content was the most important factor affecting cake 
residual oil content. The duration of pressing had no effect on cake residual oil content, thus it 
o
did not appear in the model. At 6% moisture content and 20 C, the least amount of residual oil 
was obtained.                                      
For dehulled seed:                              
𝑅0 = 23 + 4.6𝑀− 2.3𝑡 + 0.17𝑇 − 0.180𝑀
2 − 0.0008P2 + 0.10t2 + 0.0060MP + 0.09Mt −
0.013MT                              … (4)            
R = 0.93, s.e. = 1.00                               
28 
 
 
Where,                       
t = duration of pressing, mins                          
For finely ground seed:                     
𝑅0 = −10 + 4.5𝑀 + 0.29𝑃 − 1.7𝑡 + 0.13𝑇 − 0.130𝑀
2 − 0.001T2 + 0.0011MP + 0.11Mt −
0.012MT − 0.012Pt − 0.012PT +  −0.012tT               … (5)            
R = 0.98, s.e. = 0.68          
 The above model revealed that all the independent variables and their interactions were 
significant in the case of finely ground seed.                 
For coarsely ground seed:                     
𝑅0 = −70 + 11.5𝑀 + 0.26𝑃 + 1.5𝑡 + 0.53𝑇 − 0.347𝑀
2 − 0.0025P2 − 0.13t2 − 0.0014T2 −
0.038MT − 0.014P𝑇                   … (6)            
R = 0.99, s.e. = 0.76         
 Sivakumarran et al. (1985) carried out a study on expeller optimisation for peanut oil 
production and the operating conditions of a small expeller for maximum expression of oil from 
U.S. runner type peanuts were determined. The moisture content, temperature, period of pre-
heating and pressure applied were interactive factors that influenced the oil expression when the 
pressure of expression was kept at a maximum. Functional relationships between these variables 
and the meal oil content were established from experimental data. The response surface equation 
was:                                      
Y = 376.661 − 8.214𝑋1 + 7.449𝑋2 − 29.072𝑋3 − 0.118𝑋1𝑋2 + 0.271𝑋1𝑋3 − 0.302𝑋2𝑋3 +
0.0052𝑋 21   + 0.100𝑋
2
2 +  1.056𝑋
2
3                   … (7)            
Where,                    
o
X1 = peanut temperature, C                                       
X2 = pre-heat time, min                      
X3 = moisture content, %                                   
Y = meal oil content, %                    
o
 The best oil expression achieved was 92% and conditions for this were 95.6 C, 5.42% 
moisture content and 27.4 mins preheat time respectively.      
 Khan and Hanna (1984) also obtained oil yield prediction models for different samples of 
soybean.                                               
For ground soybean (with hulls):                  
Y = 6.78 + 1.68T - 0.004T2 – 3.83M + 0.422M2 + 0.3917P + 1.39t – 0.1335TM + 0.007TP – 
0.0017MPt                                                         … (8) 
29 
 
 
For flakes with hulls (Brady crop cooker):                 
Y = 199.16 + 2.81T - 0.0077T2 + 32.26M - 1.20M2 + 1.399P + 1.23t – 0.1243TM - 0.013TP 
+ 0.005Tt – 0.076MP           … (9) 
For ADM soybean flakes (without hulls):                 
Y = 62.89 + 0.45T - 0.002T2 – 0.04251P + 1.318t      … (10)  
Where,                      
o
T = temperature, C                                                     
P = applied pressure, MPa                         
t = pressing time, mins                                  
M = moisture content, % wb                              
Y = predicted percent oil expressed        
 In general, the trends of the experiment showed that the best oil yield results were 
achieved by increasing the temperature, pressure and pressing time with a moisture content of 
9.5-10%.           
 Fasina and Ajibola (1990) also developed some equations to predict the yield of oil 
expressed from conophor nuts. Empirical equations relating oil yield to processing variables such 
as pre-heating moisture content, heating temperature, heating time, applied pressure and post 
heating moisture content and duration of pressing. An equation was developed which relates the 
post-heating moisture content to the pre-heating moisture content, heating temperature and 
heating time as:                    
Mf = Mi – aMib  Tpc  tid                                … (11)                  
Where,                    
Mf = post heating moisture content, % db                                                
Mi = pre heating moisture content, % db                     
o
Tp = heating temperature, C                                    
t = heating time, mins          
 a, b, c and d are equation parameters and their values were determined. Another equation 
relating the oil yield to the post heating moisture content at each of the pressure levels considered 
in the study was also developed. The equation was of the form:                  
Y = 40 M b e-cm3            f                        … (12)           𝑓
Where,                      
Y = oil yield, %                                                   
Mf = post heating moisture content, % db                           
30 
 
 
 An equation relating the oil yield to post heating moisture content and applied pressure 
was developed. The equation was of the form:                           
Y = a Mfb Pc edMf                     … (13)                    
Where,                      
Y = oil yield, %                                                   
Mf = post-heating moisture content, % db                 
P = applied pressure, MPa                              
 Equation (12) below was also developed to relate the applied pressure at each of the pre-
pressing conditions.                 
𝑦
log (1 - ) = m – kp                     … (14)                      
40
Where,                      
Y = oil yield, %                                                     
P = applied pressure, MPa                                    
m and k are equation parameters.         
 A relationship between the volumetric oil yield and time of pressing was also presented 
in the form of equation and given as:                      
Y = 100 (1 - e-kt 𝑛)                              … (15)                   
𝑓
Where,                    
Yv = volumetric oil yield, %                                                
tp = pressing time, mins                    
k and n are equation parameters.                          
 Low standard error of estimates was obtained for all the equations and this shows that 
there was a good fit between fitted curves and the experimental points.  
 Hamzat and Clarke (1993) obtained the following equations to predict oil yield from 
groundnuts using the concept of quasi-equilibrium oil yield.                                     
For whole seed,                        
y = 40.45M-0.0063 P0.014 H-0.005 (1 – 0.270e-0.004t)                            … (16) 
s.e. = 0.90                  
For coarse sample,                      
y = 39.65M-0.31 P0.004 H-0.003 (1 – 0.483e-0.004t)                 … (17) 
s.e. = 0.87                  
For fine sample,                      
y = 40.45M-0.067 P0.011 H-0.001 (1 – 0.346e-0.004t)                 … (18)  
31 
 
 
s.e. = 0.89                                              
Where,                       
y = oil yield, %                                                    
M = seed moisture content, % wb                  
tp = pressing pressure, MPa                   
H = thickness (bed depth) of pressed seed, mm                
a, b, c, d, f are constants.         
 Reddy and Bohle (1993) developed oil yield prediction equation for mustard seed. The 
equation is of the form:                   
Y = 19.85214 – 1.74831M + 7.33665P + 0.024707T + 0.08542M2 – 0.29723P2 + 
0.00136T2 – 0.35917MP – 0.01391MT + 0.005708PT      … (19)              
Where,                      
Y = oil yield, %                                                     
M = moisture content, % wb                   
P = applied pressure, MPa                   
T = Pressing time, mins        
 Akinoso et al. (2006) developed oil yield prediction equation for sesame seed. The 
equation is of the form:                   
Y = 38.485 – 4.944X1 + 0.446X2 + 0.207X3 + 0.118X12 – 0.01722X22 – 0.00009972X32 + 
0.03032X1X2 + 0.005413X1X3+ 0.0007251X2X3                              … (20) 
Where,                      
Y = oil yield, %                                                     
X1 = moisture content, % wb                   
X2 = roasting duration, mins                   
o
X3 = roasting temperature, C        
 Shridhar et al. (2010) obtained the following equation in terms of coded factors for the oil 
yield of castor seed:                     
Y = 47.50 + 7.41X1 + 2.08X2 + 0.63X1X2 – 16.62X12 – 2.87X22                                        … (21) 
Where,                    
X1 = dilution level                     
X2 = agitation time, hr         
 Awolu et al. (2013) obtained the following equation in terms of coded factors for the oil 
yield of neem seed:                                         
32 
 
 
Y = 47.30 + 10.75X1 + 1.58X2 + 3.50X3 + 2.50X1X2 + 2.00X1X3 – 0.50X2X3 – 14.47X12 –
12.33X22 – 11.32X32                     … (22)                    
Where,                      
Y = oil yield, %                                                     
X1 = mass of sample, g                               
X2 = particle size, mm                     
X3 = extraction period, hr          
 Jazie et al. (2013) obtained model equation based on the coded values for the biodiesel 
yield from rapeseed oil  as given below:                
Y = 95.85 + 1.089X1 – 9.15X12 + 2.64X2 – 9.87X22 – 0.63X3 – 8.9X32 + 0.81X1X2 + 2.07X1X3 + 
3.2X X  2 3                                   … (23)               
Where,                      
Y = oil yield (%)                                                     
X1 = catalyst concentration (% wt)                               
X2 = methanol/oil molar ratio                           
o
X3 = reaction temperature ( C)        
 The empirical equations and models that have been developed thus far are usually crop 
specific because different oil crops respond differently to changes in process parameters. At 
present, no model is known to relate the yield of mechanically expressed moringa seed oil to 
process parameters using Response Surface Methodology, hence the need for this research work. 
 
 
 
 
 
 
 
 
33 
 
 
CHAPTER THREE 
MATERIALS AND METHODS 
3.1 Determination of the Engineering Properties of Moringa  
The engineering properties (physical and mechanical) of moringa seeds were determined 
using ASABE standards and used as design parameters for the development of an oil expeller. 
 3.1.1 Determination of Physical Properties of Moringa  
Physical properties namely length, width, thickness, arithmetic mean diameter, 
geometrical mean diameter, surface area, sphericity, aspect ratio, moisture content, true and bulk 
densities, porosity, one thousand seed weight, coefficient of static friction and angle of repose of 
moringa seeds were determined following standard procedures.    
  3.1.1.1 Dimensional Properties    
 One hundred moringa seeds were randomly selected and the length, width and thickness 
were measured using a digital caliper with an accuracy of 0.01 mm. The arithmetic mean 
diameter (Da), geometric mean diameter (Dg), surface area (S) and sphericity (Φ) were calculated 
from equations (24-27) respectively as given by Mohsenin (1986), while the aspect ratio (Ra) 
was calculated from equation (28) as given by Maduako and Faborode (1990). 
(L+W+T)
𝐷a =                                  … (24) 3
1
𝐷g = (𝐿𝑊𝑇)3                                 … (25) 
S = 𝜋𝐷 2g                        … (26) 
Φ = (𝐷g  ∕ L) × 100                      … (27) 
𝑅a = (W∕ L) × 100                                … (28) 
Where; L= length, W= width, T= thickness. 
  3.1.1.2 Gravimetric Properties      
 The initial moisture content of the moringa seeds was determined in the laboratory using 
the oven-dried method. 50 g of moringa seeds were oven-dried at a constant temperature of 
o
130 C for 6 hrs. The moisture content was calculated using equation (29) below: 
𝐼𝑛𝑖𝑡𝑖𝑎𝑙  𝑤𝑒𝑖𝑔𝑕𝑡 𝑜𝑓  𝑝𝑟𝑜𝑑𝑢𝑐𝑡 −𝐹𝑖𝑛𝑎𝑙  𝑤𝑒𝑖𝑔 𝑕𝑡 𝑜𝑓  𝑝𝑟𝑜𝑑𝑢𝑐𝑡
Moisture Content (% wet basis) =  × 100 … (29)
𝐼𝑛𝑖𝑡𝑖𝑎𝑙  𝑤𝑒𝑖𝑔 𝑕𝑡 𝑜𝑓  𝑝𝑟𝑜𝑑𝑢𝑐𝑡
 The true density of moringa seeds was determined by the water displacement method. A 
3
known weight of samples was poured into a 500 cm  fractionally graduated measuring cylinder 
3 
containing 250 cm distilled water. The rise in water indicated the true volume of the seeds. The 
true density was calculated as:                      
𝑊𝑒𝑖𝑔 𝑕𝑡  𝑜𝑓  𝑡𝑕𝑒  𝑠𝑎𝑚𝑝𝑙𝑒
True Density =                                                                               … (30) 
𝑉𝑜𝑙𝑢𝑚𝑒  𝑜𝑓   𝑤𝑎𝑡𝑒𝑟  𝑑𝑖𝑠𝑝𝑙𝑎𝑐𝑒𝑑  
34 
 
 
The bulk density of samples was determined by using a regular container of known mass. 
The container was filled to the brim with the samples and was gently tapped ten times for the 
samples to consolidate. The weight of the samples were noted and the volume of the container 
3
was estimated by filling with water, which was then poured into a 500 cm  fractionally graduated 
measuring cylinder to determine the volume. The bulk density was calculated as: 
𝑊𝑒𝑖𝑔 𝑕𝑡 𝑜𝑓  𝑠𝑎𝑚𝑝𝑙𝑒
Bulk Density =                                                                                             … (31) 
𝑉𝑜𝑙𝑢𝑚𝑒  𝑜𝑐𝑐𝑢𝑝𝑖𝑒𝑑  
The porosity was calculated from the bulk and true densities as given by Mohsenin 
(1986). 
𝑇𝑟𝑢𝑒  𝐷𝑒𝑛𝑠𝑖𝑡𝑦 −𝐵𝑢𝑙𝑘  𝐷𝑒𝑛𝑠𝑖𝑡𝑦
𝑃𝑜𝑟𝑜𝑠𝑖𝑡𝑦 =   × 100                            … (32) 
𝑇𝑟𝑢𝑒  𝐷𝑒𝑛𝑠 𝑖𝑡𝑦
 The one thousand seed weight was determined using a digital electronic balance having 
an accuracy of 0.001 g. One hundred moringa seeds were weighed and multiplied by 10 as 
adopted by Tarighi et al. (2011).           
  3.1.1.3 Frictional Properties        
 The coefficients of static friction of moringa seeds were determined with respect to six 
different surfaces namely glass, stainless steel, mild steel, galvanized steel, rubber and plywood 
using an inclined plane apparatus as described by Dutta et al. (1988). The inclined plane was 
gently raised and the angle of inclination at which the samples started sliding was measured. By 
measuring the angle of surface (ϕ) at this point, the static coefficient of friction (μ) is determined 
as the tangent of the angle (Dutta et al., 1988). 
μ = tanϕ                      … (33)  
For the dynamic angle of repose determination, the method described by Taser et al. 
(2005) and Garnayak et al. (2008) was adopted. A bottomless cylinder was used, the cylinder 
was placed over a plain surface and the seeds were filled in. The cylinder was raised slowly 
allowing the sample to flow down and form a natural slope. The dynamic angle of repose was 
calculated from the height and diameter of the pile as: 
-1  2𝑕  
θ = tan (  )                      … (34) 
𝑑
Where θ = angle of repose, h = height of the pile (mm) and d = diameter of the pile (mm).  
 3.1.2 Determination of Mechanical Properties of Moringa  
Mechanical properties namely force at peak, deformation at peak, stress at peak, energy 
to peak, force at break, deformation at break, stress at break, energy to break, force at yield, 
stress at yield, energy to yield and the Young‟s modulus were determined following standard 
procedures.            
35 
 
 
 An Instron Universal Testing Machine (Plate 3.1) equipped with a 25 kg compression 
load cell and integrator was used to determine the various mechanical properties of the moringa 
seeds. The measurement accuracy was 0.001 N in force and 0.001 mm in deformation as given 
by Mohsenin, 1970. The height of the seed when loaded in the machine was taken as the 
thickness of the seed as measured with a digital caliper. The individual seed was compressed by 
the machine until rupture occurred as is denoted by a bio-yield point in the force-deformation 
curve. The  rupture  point  is  a  point  on  the  force-deformation curve at which the loaded 
specimen shows a visible or invisible  failure  in  the  form  of  break  or  crack.  This point was 
detected  by  a  continuous  decrease  of  the  load  in  the  force-deformation  diagram. Once the 
bio-yield was detected, the loading was stopped. The various parameters were measured at a 
speed of 20 mm/min. 
3.2 Design Considerations for the Moringa Oil Expeller 
 In designing the oil expeller for moringa, some design considerations which would 
increase the performance level of the machine were made. In the design, simple manufacturing 
and engineering techniques were adopted to produce a machine which is relatively cheap, easy to 
operate and maintain, spare parts available and a machine that will not fail. In the material 
selection, considerations were given to the techno-economic status of the small scale vegetable 
oil processors who are the intended users of the machine. Other important considerations 
included: 
(i) Ready availability of the materials of construction for the design of the various 
 components in the local markets to facilitate easy fabrication and purchase of spare parts.  
(ii) Since the expeller is a food processing equipment, the parts that will get in contact 
 with the oil were made of simple iron to prevent contamination of the oil. Chromium 
 and nickel in steel may react with the oil at high temperatures and lower the oil 
 quality. 
(iii) For the expeller to work efficiently and to prevent avoidable failures, the selection of 
 materials for fabrication considered the appropriate material strength properties and 
 sizes. This was achieved by using established standard data and calculations. 
(iv) The design took into consideration the easy transportation of the machine where 
 necessary. Based on this, the design made it possible to be able to dismantle some 
 parts of the machine. 
 
 
36 
 
 
 
Plate 3.1. The Universal Instron Testing Machine  
 
 
 
 
 
 
 
 
 
 
37 
 
 
3.3 Theoretical Considerations for the Moringa Oil Expeller 
 The following assumptions were made for design considerations as given by Olayanju 
(2002):  
(i) No pressure development would take place in the feed section. The pressure 
 development and the expression of oil start at the beginning of the press section.  
(ii) The maceration of oilseed was complete in the press section leaving the homogenous 
 mixture of oil and solids in the discharge section. 
(iii) The temperature of oilseed mass remained constant in the ram section. In reality, the 
 temperature would increase along the ram section due to shearing action of the shaft.  
(iv) As the oil-solid mixture passes through the section, it is subjected to radial pressure 
 exerted by the wormshaft. The pressure causes flow of oil in the radial direction 
 through the solid matrix and out through the barrel slots. The oil-flow in turn changes 
 the flow rate of mixture in the axial direction. 
3.4 Design Calculations 
 Several calculations were made based on the results of the measured physical and 
mechanical properties of the moringa seeds. 
 3.4.1 Determination of the Mass of Oil extracted per batch 
 Assuming 2 kg of oil is extracted per hour. The oil content of moringa seed is between 
30-40% (Mohammed et al., 2003; Anwar et al., 2006; Anwar and Rashid, 2007; Nzikou et al., 
2009; Uzama et al., 2011; Adejumo et al., 2013; Ogunsina et al., 2011, Goja, 2013 and Orhevba, 
2013); assuming oil content of 35%             
𝑀𝑎𝑠𝑠 𝑜𝑓 𝑜𝑖𝑙
 = 0.35                             … (35) 
𝑀𝑎𝑠𝑠 𝑜𝑓 𝑚𝑜𝑟𝑖𝑛𝑔𝑎
𝑀𝑎𝑠𝑠  𝑜𝑓  𝑜𝑖𝑙 2
 𝑀𝑎𝑠𝑠 𝑜𝑓 𝑚𝑜𝑟𝑖𝑛𝑔𝑎 = = = 5.71 𝑘𝑔  
0.35 0.35
Therefore, about 6 kg of moringa is required.  
 3.4.2 Determination of the Roaster Volume   
 In order to estimate the roaster volume, the volume of moringa (v) with bulk density (ρ) 
662 𝑘𝑔𝑚−3 (as determined by the physical properties of moringa) and mass (m) 6 𝑘𝑔 is 
calculated as;             
𝑚
 𝑣𝑜𝑙𝑢𝑚𝑒, 𝑣 =                           … (36) 
𝜌
6
 𝑣𝑜𝑙𝑢𝑚𝑒, 𝑣 = = 0.0091 𝑚3 
662
The volume of the cylindrical roaster, VR is given by  
38 
 
 
   V 2R = 𝜋𝑟 h                   … (37) 
Where, 
r = radius of roasting cylinder, m 
h = height of roasting cylinder, m 
   𝜋𝑟2h = 0.0091 𝑚3    
Assuming a height of 0.5𝑚, the radius of the roasting cylinder is,     
  𝑟2
0.0091
=           
𝜋𝑕
2 0.0091  𝑟 =           
3.142×0.5
 𝑟 =  0.00579         
 r = 0.08 𝑚                                                               
The diameter of the roasting chamber is calculated as                                     
  𝑑 = 2𝑟 = 2 × 0.08𝑚 = 0.16 𝑚 
 3.4.3 Stirring Clearance         
 The roaster volume should be 1.5 times greater than the calculated volume VR to allow 
for stirring clearance. Hence the total volume of the roasting chamber will be 1.5 × 
0.0091 𝑚3 = 0.0137 𝑚3. 
 3.4.4 Determination of Heater Capacity       
 To determine the heater capacity of the roasting chamber, the average moisture content of 
the seeds and moisture content of roasted seeds will be used (about 2-3 % wb) in evaluating the 
quantity of heat needed to properly roast moringa seeds for oil expression. 
Mass of Moringa = 6 𝑘𝑔 
The amount of initial water content at 7.31% moisture content  = 0.0731 × 6 𝑘𝑔 = 0.439 𝑘𝑔 
Amount of dry matter and oil = 6 − 0.439 = 5.56 𝑘𝑔 
0.0731
Initial moisture content (db) =  × 100 = 7.89%                … (38) 
1−0.0731
0.02
Final moisture content (db) =  × 100 = 2.04%                 … (39) 
1−0.02
The amount of water removed equal to product of initial dry matter and difference in dry basis 
moisture content i.e. 
7.89 – 2.04
The amount of water removed = 5.56 × = 0.325 𝑘𝑔               … (40) 
100
𝑎𝑚𝑜𝑢𝑛𝑡  𝑜𝑓  𝑤𝑎𝑡𝑒𝑟  × 𝑠𝑝𝑒𝑐𝑖𝑓𝑖𝑐  𝑕𝑒𝑎𝑡  𝑜𝑓  𝑎𝑖𝑟
The quantity of heat required, Q  =               … (41) 
𝑟𝑜𝑎𝑠𝑡𝑒𝑟   𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑐𝑦
0.49 × 1.005 𝑥 103
   = 408.28 𝑘𝑐𝑎𝑙/𝑕𝑟  
0.8
1 kcal/hr = 1.622 watts 
39 
 
 
408.28 kcal/hr = 662.23 W = 0.662 kW 
For adequate supply of heat, a heating element of 1.0 kW was chosen. 
 3.4.5 Roasting Time 
 The time needed for the roasting is computed from Newton‟s law as;              
𝑑ω
=  𝑘𝑎𝑟(𝑇2 − 𝑇1)                     … (42) 𝑑𝜃
Where, 
dω = Amount of water removed, kg 
d𝜃 = Time needed for roasting, min 
k = Constant = 0.2 
2
ar = Area of Roaster, m  
o
T1 = Initial temperature, C 
o
T2 = Final temperature, C   
𝑑ω 0.325𝑘𝑔
𝑑𝜃 = =      
𝑘𝑎r (𝑇2 −𝑇1) 0.2 × 320 𝑥  10−3  × 250 𝑥  10−3  × (100−40)
 𝑑𝜃 = 0.34 𝑕𝑟 ~ 20 𝑚𝑖𝑛s 
 3.4.6 Determination of Expeller Volume  
In determining volume of expeller, consideration was given for volume of hollow 
cylindrical barrel and the volume of the tapered screw shaft. Assuming a 40% decrease in 
volume as oil is processed, the volume of expeller is calculated as:  
60 3
 Volume of expeller (VE)             =   × 0.0137 𝑚     … (43) 100 
        = 0.00822 𝑚3  
 3.4.7 Determination of Barrel Volume 
 The barrel volume is calculated using equation (44) below:  
The volume of cylindrical barrel (V 2B) = 𝜋𝑟 𝑙             … (44) 
= 3.142 × 0.082 x 0.97 = 0.0195 𝑚3 
Where,                       
r = radius of cylinder 
 
l = length of cylinder
 3.4.8 Determination of Shaft Volume 
 The shaft volume is calculated using equation (45) below:  
The volume of desired tapered screw shaft, VS = VB − VE            … (45) 
VS = 0.0195 − 0.00822 = 0.01128 𝑚3  
The volume of tapered shaft, (VS) 
40 
 
 
𝜋𝑙
   𝑉𝑆 =  𝑅
2
 +  𝑅𝑟 + 𝑟2                   … (46) 
3
Where,                       
r = minimum radius of tapered shaft. 
R = maximum radius of tapered shaft.    
2 3 × 𝑉
  𝑅  +  𝑅𝑟 + 𝑟2  =   s                  … (47) 
𝜋 × 𝑙
3 × 0.01128
        =    
3.142 × 0.97
2
  𝑅 +  𝑅𝑟 + 𝑟2  =  0.0111 𝑚3 
Using Goal Seek; 
Minimum radius r, = 0.015 𝑚 
Maximum radius R, = 0.098 𝑚 
 
 Barrel Shaft Active volume 
Length   (m) 0.40 0.97  
Radius   (m) 0.08 0. 015 - 0.098  
3
Volume (m ) 0.0195 0.01128 0.00822  
  
 3.4.9 Design Capacity of the Expeller       
 The capacity of an expeller is controlled by the drag flow, pressure flow and leak flow in 
the barrel assembly. Varma (1998) gave the theoretical capacity of an expeller with single flight 
in feed section as: 
Q = 𝜋DNH cosα  Pcosα - e)                    … (48) 
Where,                      
Q = Theoretical capacity of the expeller, kg/hr 
D = Mean diameter of screw, mm 
N = Speed of rotation, rpm 
H = Depth of worm, mm 
P = Worm Pitch, mm 
e = Thickness of worm, mm 
o 
α = Helix angle, 
Using measured data for moringa seeds as input values, 
Q = 3.142 × 50 × 45 × 6.25 × cos10 (50cos10 – 6.25)  
    = 1.87 x106 mm3/min  
41 
 
 
    = 0.1122 m3/hr  
3
For an average bulk density of 662 kg/m   
Q = 0.1122 m3/hr × 662 kg/m3 = 74.30 kg/hr 
 3.4.10 Forces acting on Screw Thread      
 The two main forces acting on the screw thread are the frictional force required to 
translate and compress the moringa seeds as well as the frictional force resulting from the 
screw‟s motion. According to Olayanju (2002), by taking equilibrium of forces; the following 
equations (49) and (50) will be solved simultaneously 
 W = Kcos  α + ϕ)                          … (49) 
 F = Kcos  α + ϕ)                          … (50) 
𝐹 Ksin (α  + ϕ)
Therefore, =   = tan α + ϕ)                        … (51) 
𝑊 Kcos  (α  + ϕ)
 F = W tan (α + ϕ)                          … (52) 
But the friction angle, ϕ = tan-1µ 
Where µ = coefficient of static friction = 1.376 (as determined by the physical properties of 
moringa) 
ϕ = tan-1(1.376) = 53.99 ~ 54o 
W = 67 N (maximum force at peak as determined from the mechanical properties of moringa)  
F = 67 tan (10 + 54) = 137.37 N 
 At a point in time, a minimum of 50 moringa seeds are crushed at the feed end portion of 
the machine. Therefore, a minimum average force of 6.87 kN will be required to express the oil. 
 3.4.11 Torque on Screw Thread       
 Torque (T) and axial load are related to each other by equation (53) as given by Hall et 
al., (1980): 
T = W [rm tan α + ϕ) + fcrc]                   … (53) 
 With the use of a well lubricated bearing, the frictional force at the collar will be neglected. 
Thus, the quantity fcrc becomes zero. Hence, the equation becomes 
T = Wrm tan α + ϕ)                     … (54) 
But from equation (52), F = W tan  α + ϕ) 
Therefore, the torque transmitted by worm action is given by 
T = Frm                                                                                                                                                                     … (55) 
Where,                      
T = Torque on screw thread, Nm 
F = Axial force required to expel oil, N 
42 
 
 
rm =  Radius screw, mm 
D 54
rm =
m  =  = 27 mm  
2 2
Where, Dm = Mean diameter of screw, mm 
T = 6.87 x 103 × 27 × 10-3 
T =  185.5 Nm 
 3.4.12 Power Requirements          
 The power drive mechanism incorporates the use of a reduction gear motor coupled to the 
expeller shaft by pulley and belts arrangement. The chosen speed for the expeller Nc is 45 rpm. 
2πN
The angular speed we =                                                                                        … (56) 
60
2 × 3.142 × 45
we  =  = 4.71 rad/s 60
The power input to the expeller is computed from equation (57) 
P  Tw                                               e  = e … (57) 
= 185.5 × 4.71 = 873.7 W = 0.874 kW  
The power of the electric motor to drive the expeller was estimated using the equation (58) 
below as given by Onwualu et al. (2006): 
P
P em =                 … (58) 
ƞ
Where, 
Pm = Power of the electric motor 
Pe = Power requirement of the expeller 
ƞ = Drive efficiency 
Assuming a drive efficiency of 75% = 0.75  
873.7
Pm =  = 1165 W = 1.165 kW = 1.55 hp 
0.75
 To give allowance for power used in driving pulleys and the shaft, a 2.0 hp (1.50 kW) 
electric reduction gear motor with a speed of about 180 rpm is chosen. 
 3.4.13 Belt Design         
 For the chosen 1.50 kW, 180 rpm electric gear motor, the belt type is a B-section (Figure 
3.1). Diameter, d = 75 mm is used at the gear motor shaft. The pulley is designed by considering 
the power to be transmitted between the electric motor and the screw expellant shaft. The ratio of 
the pulley for the electric motor to that of the expellant shaft would be 1:2 (Ajao et al., 2009) and 
the allowable diameter of the pulleys is calculated as given by Olaomi (2008) in equation (59) as:  
N1D1=N2D2                             … (59) 
43 
 
 
 
Figure 3.1. Effective Power of Belts as a function of rpm of small Sheaves 
Source: Mubeen (1998) 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
44 
 
 
Where, 
N1 = Speed of the driving motor, rev/min 
N2 = Speed of the expellant shaft, rev/min  
D1 = Diameter of driving pulley, m  
D2 = Diameter of driven pulley, m  
The expeller pulley‟s diameter is calculated as: 
N D
D2 =  
1 1                        … (60) 
N2
180 x 75
D2 =  = 300 𝑚𝑚 45
The total length of the belt is calculated as given by Khurmi and Gupta (2004) in equation (61):  
π (D2−D1)
2
L = 2C +  (D1 + D2) +                     … (61) 2 4C
Where, 
L = Total length of belt, mm 
C = Minimum centre distance, mm 
The minimum centre distance is calculated as given by Ajao et al. (2009) in equation (62):  
D +D
 C = 1 2 + D1               … (62) 2
75+330
C = +  75 = 262.5 𝑚𝑚 
2
From equation (61), the total length of the belt is calculated as:  
3.142 (300−75)2
L = 2 x 262.5 +  (75 + 300) +   =  1162 mm 
2 4 x 262.5
From Table 3.1, the nearest standard pitch is 1110 mm for which the nominal length is 1067 mm. 
A 2 B42- synchronous (toothed) belt arrangement which combines the characteristics of belts 
and chains will be used. This will guard against slippage, hence maintaining a constant speed 
ratio between the driving and the driven shafts. 
3.4.14 Determination of Belt Tensions        
 The angle of wrap is calculated using equation (63) as given by Hall et al. (1980): 
α = 180 + 2sin-1[(R2 − R1)/C]                   … (63) 
R1 = Radius of the smaller pulley = 37.5 mm 
R2 = Radius of the larger pulley = 150 mm 
C = Centre distance = 262.5 mm 
α = 180 + 2sin-1[(150 − 37.5)/262.5] = 230.8o = 4.03 rad 
α = 180 - 2sin-1[(150 − 37.5)/262.5] = 129.2o = 3.86 rad 
 
 
45 
 
 
Table 3.1. Standard V-Belts Pitch Lengths 
Nominal Length mm (inches)             Standard Pitch Length, mm (inches) 
 A – Section B – Section C – Section 
660 (26) 696 (27.4) -------         -------         
787 (31) 823 (32.4) -------         -------         
838 (33) 874 (34.4) ------- ------- 
889 (35) 925 (36.4) 932.2 (39.7) ------- 
965 (38) 1001 (39.4)     1008.4 (39.7)         ------- 
1067 (42)     1102 (43.4) 1110 (43.7)             -------- 
1168 (46)     1204 (47.4)     1212 (47.7)            -------- 
1219 (48)     1252 (49.4)         -------- -------- 
1295 (51)     1331 (52.4)     1339 (52.7) 1351 (53.2) 
1295 (53)     1382 (54.4)     1389 (54.7)            -------- 
1397 (55)     1433 (56.4)     1440 (56.7)            -------- 
1524 (60)     1561 (61.4)     1567 (61.7)       1580 (62.2) 
1575 (62)     1610 (63.4)     1618 (63.7)            -------- 
1625 (64)     1661 (65.4)     1669 (65.7)            -------- 
1727 (68)     1762 (69.4)     1770 (69.7) 1783 (70.2) 
1905 (75)     1941 (76.4)     1948 (76.7)       1961 (77.2) 
1981 (78)     2017 (79.4)     2024 (79.7)            -------- 
2032 (80)     2067 (81.4)         -------- -------- 
2057 (81)        -------- 2101 (82.7)       2113 (83.2) 
2108 (83)     2144 (84.4)     2151 (84.7)            -------- 
2159 (85)     2195 (86.4)     2202 (86.7)       2215 (87.2) 
2286 (90)     2322 (91.4)     2329 (91.7)       2342 (92.2) 
2438 (96)     2474 (97.4)         -------- 2499 (98.2) 
2464 (97)     2499 (98.4)     2507 (98.7)            -------- 
2667 (105)     2702 (106.4)     2710 (106.7)     2723 (107.2) 
2845 (112)     2880 (113.4)     2888 (113.7)       2901 (114.2) 
3048 (120)     3084 (121.4)     3091 (121.7)       3104 (122.2) 
    Source: Mubeen (1998). 
 
 
 
 
 
 
 
 
 
 
46 
 
 
 To obtain the tensions in the tight and slack sides, the following equations (64) and (65) 
are solved simultaneously 
(T1– T2) V = P                                 … (64) 
T1−𝑚𝑣
2
= µα/sinɵ/2                           and 2 e … (65)T2−𝑚𝑣
P = Power transmitted (kW) 
T1 = Tension in the tight side (kN) 
T2 = Tension in the slack side (kN) 
But m = bte                                 … (66) 
b = belt width = 17 mm 
t = belt thickness = 11 mm 
3
e = belt density = 970 kg/m  for leather belt 
-3
m = 17 × 11 × 10  × 970 = 0.18 kg/m 
µ = coefficient of friction between belt = 0.15 for leather belt on steel 
v = belt velocity 
v = rwe                       … (67)  
2πN
But we =  from equation (56) 60
2πrN 2 × 3.142 × 37.5 × 0.001 × 180
Therefore, v = rwe =  =  = 0.71 m/s 
60 60
𝜃 o = 40  (most common angle of groove) 
ɵ        
For small pulley, eµ1α1/sin /2               … (68) 
 e0.15 x 3.86/sin20 = 5.44 
For big pulley, eµ2α2/sin
ɵ/2                     … (69) 
 e0.15 x 4.03/sin20 = 5.86 
According to Olayanju (2002), the pulley with the smaller value governs the design. Therefore, 
the smaller value (5.44) will be used in the design. Equation (65) then becomes  
T1− 𝑚𝑣
2
= 5.44                         
T − 𝑚𝑣 22
T1− 0.18 × 0.71 × 0.71 = 3.27                                                                                                
T2− 0.18 × 0.71 × 0.71
T1−0.0907 = 3.27 
T2−0.0907
T1 − 0.0907 = 3.27T2 − 0.297 
3.27T2 − T1 = 0.297 − 0.0907 
47 
 
 
3.27T2 − T1 = 0.2063                    … (70) 
But from equation (58), 
(T1 −T2) V = P 
P = 1 hp = 0.746 kW 
V = 0.71 m/s 
P
Therefore, T1 − T2 =                      … (71) V
0.746
T1 − T2 =   = 1.05 0.71
T1  = 1.05 + T2                            … (72) 
By substituting equation (72) into equation (70), equation (70) becomes 
3.27T2 − (1.05 + T2) = 0.2063 
3.27T2 − T2 = 0.2063 + 1.05 
2.27T2 = 1.2563 
1.2563
T2 = = 0.553 kN 2.27
By substituting the value of T2 into equation (66), 
T1 = 1.05 + 0.553 = 1.603 kN 
The bending load on the wormshaft is due to the weight of the pulley, the summation of 
tensions on the belts acting vertically downward and the weight of the threaded shaft as shown in 
Figure 3.2. The shaft will be supported at the two ends A and B by two bearings. The reactions 
RA  and RB at the two supports are determined by equation (73) 
RA+ RB  = Ws  + (T1 + T2) + Wp                                 ... (73) 
W = Weight of threaded shaft = 50 N 
T1 + T2 = Sum of tensions on vertical belts = 2156 N 
Wp  = Weight of pulley = 50 N 
RA + RB  = 50 + 2156 + 50 
RA + RB  = 2256                                           … (74) 
Taking moment about B, 
48 
 
 
RA (0.605) − 50 (0.3025) − 2206(0.545) = 0 
0.605RA = 1217.395 
RA = 2012.22 N  
Taking moment about A, 
RB  (0.605) − 50 (0.3025) − 2206(0.06) = 0 
0.605RA = 147.485 
RB  = 243.78 N  
 The shear force and bending moment diagrams are shown in Figure 3.3. 
3.4.15 The Power Transmission Shaft       
 The power transmission shaft is designed to satisfy the strength criterion. 
 The required diameter for a solid shaft having combined bending and torsional loads is 
obtained from ASME code equation as adopted by (Hall et al. 1980) as: 
16
D3 =  (K M 2 2b b) + (KtMt)                     … (75) 𝜋Ss
Where, 
D = Diameter of solid shaft, mm 
Ss = Allowable combined shear stress for bending and torsion = 40 MPa for steel shaft with               
        keyway 
Kb  = Combined shock and fatigue factor applied to bending moment = 1.5 to 2.0 for minor    
         shock 
Kt  = Combined shock and fatigue factor applied to torsional moment = 1.0 to 2.0 for minor    
         Shock 
Mb  = Bending moment, Nm = 147.5 Nm 
Mt = Torsional moment, Nm = 185.5 Nm  
16
D3 = 6  (2.0 𝑥 147.5)
2 + (2.0 𝑥 185.5)2  
40 𝑥  10 𝜋
D3 = 6.04 x 10-5 m 
D = 0.0392 m = 39.2 mm 
The calculated diameter is less than the chosen diameter of 40 mm. Therefore, strength criterion 
is satisfied. 




49 
 
 
     
  
    
    
 
 
 
 
 
 
 
 
 
 
 
 
 
 
        
   
        R A                                            RB 
 
Figure 3.2. Bending Loads on the Wormshaft 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
50 
 
 
 
 
 
  
 
 
 
 

 
  
 
 2012N

 

 
 2012N
 
 
 
 
  


 

 

 
 
 
 
 

 
 
 
 

 
 
 
 
  

 
Figure 3.3. Shear Force and Bending Moment Diagrams. 
 
51 
 
 
 
3.5 Description of the Moringa Oil Expeller 
 The moringa oil expeller is shown in Plate 3.2. It consists of an electric gear motor [5], 
feed hopper [4], roasting chamber [3], expression chamber, stirrer, heater, wormshaft [8], shaft 
housing, oil barrel, oil barrel cover [7], belts and pulleys [6], bevel gears [1], bearings [9], 
temperature regulator [2], oil outlet [10], cake outlet and machine frame [11].    
 3.5.1 The Roasting Chamber 
 This is the unit where the seeds are being heated. It consists of a roasting drum with 
dimensions 320 mm × 250 mm, 1 kW rated capacity heater which will supply the energy to 
vapourize moisture from the wet products in the roasting drum and a temperature regulator 
which controls the temperature of the air inside the drying chamber. The heater normally causes 
a difference in relative humidity and temperature of the inlet air and that of the air in the drying 
chamber and if not controlled, it leads to overheating and subsequent burning of products. The 
temperature regulator therefore controls the heater. 
 3.5.2 The Feeding Assembly 
 This consists of a hopper which is directly located under the roasting chamber. It is made 
of mild steel. The hopper has dimensions of 200 mm × 220 mm × 225 mm. The roasted seeds 
flow down from the roasting chamber into the feed hopper from where they enter the expelling 
chamber. 
 3.5.3 The Expression Chamber 
 This was made of 90 mm diameter, 790 mm long and 12.7 mm thick stainless steel pipe. 
This was split into two equal halves (top and bottom parts). Perforations were provided on the 
bottom part of the barrel so that the expressed oil can drain through them. The two halves were 
bolted together using bolts and nuts. The expression chamber is enclosed in a cover to prevent 
expressed oil from coming in contact with dust and foreign materials.  
 3.5.4 The Worms and Wormshaft Assembly 
 The wormshaft assembly consists of a special wormshaft fitted with six worms of 
different pitches. The worm flight design along pressure and discharge section is such that the 
o
materials do not wrap around more than 320 . This leaves an axial gap in the flight that enables 
the compressed material to slide in either direction relative to velocity generated by worm pitch. 
This balances the pressure over a group of worm section and reduces the tendency of material to 
lock in individual section and rotate with the shaft. The configuration of the worm section is such 
that the volume displacement at the feed end of the press is greater than the discharge end. The 
52 
 
 
whole assembly rotates in the barrel. The wormshaft was made from a mild steel solid rod of 
diameter 80 mm and length 970 mm, which was machined on the lathe at a decreasing screw 





 






 
Plate 3.2. The Moringa Oil Expeller 
1: Bevel Gear     2: Temperature Regulator   
3: Roasting Chamber    4: Hopper   
5: Electric Gear Motor   6: Belt and Pulley Arrangement 
7: Oil Barrel Cover    8: Wormshaft      
9: Pillow Bearing    10: Oil Trough  
11: Machine Frame 
53 
 
 
 
pitch and decreasing screw depth. The wormshaft is housed in the cylindrical barrel at a 
clearance of 1.5 mm between the screw diameter and inside diameter of the barrel. Two pillow 
bearings (90 mm diameter) are elevated to a height where the wormshaft can rotate 
concentrically. The worms have a dual role of conveying the oilseeds through the barrel and at 
the same time exerting pressure on the material. Besides increasing the pressure in the barrel, the 
shear action on the barrel breaks the oilseeds into smaller particles. Due to pressure created by 
the worms and choke, the oil flows out of the oil-solid matrix through the holes in the oil barrel. 
 3.5.5 The Power Unit 
This consists of a 2.0 hp, 180 rpm, 3-phase AC electric gear motor, a set of belts and 
pulleys (one connected from the electric gear motor to the main shaft and the other from the 
main shaft to the roaster).  
 3.5.6 The Oil and Cake Troughs       
o
 These are made of mild steel. They are inclined at an angle of 60  to the horizontal to 
allow for the free flow of oil and cake. The oil trough is located at the front of the machine for oil 
collection, while the cake trough is located at the back of the machine for cake collection.  
 3.5.7 The Machine Frame 
 This supports all the other components and units of the oil expeller. It consists of a base 
and supports which are made from mild steel.      
 The AUTOCAD drawings of the various component parts showing the orthographic, 
isometric, schematic, wireframe, exploded and main views of the expeller are shown in 
Appendix H. 
3.6 Working Principles of the Moringa Oil Expeller 
 The seeds are introduced into the machine through the roasting drum where they are 
heated. Thereafter, the seeds flow down to the feed hopper through a sliding gate provided on the 
roasting drum. The machine conveys, grinds and presses the seeds inside the cylindrical barrel 
with the aid of the wormshaft until oil is squeezed out of the seed. The oil extracted is drained 
though the oil barrel into the oil trough where it is collected, while the residual cake is 
discharged to the cake trough at the cake outlet where it is collected. The machine is powered by 
a 2.0 hp three-phase electric motor. 
3.7 Experimental Design  
 The effects of the processing variables namely moisture content, heating temperature, 
heating time and applied pressure on the oil yield were investigated. The four independent 
variables considered are very important factors affecting the yield and quality of oil expression 
54 
 
 
from oilseeds using oil expellers. The experimental design adopted was 4 factors, 5 levels, 
factorial Central Composite Rotatable Design (CCRD) of Response Surface Methodology (Box 
et al., 1978) as adopted by Olajide (2000) while working on groundnut and sheanut kernels; 
Shridhar et al. (2010) while working on castor seed oil, Narayana (2012) while working on 
coconut and Awolu et al. (2013) while working on neem seed.    
 Central Composite Rotatable Design is comprised of three types of design points namely 
factorial points (nf), axial points (na) and central points (nc). According to the Central Composite 
Rotatable Design, the total number of treatment combinations is 𝑛 = 2k  𝑛f  + 2𝑘 𝑛a  + 𝑘 𝑛c  
where „k‟ is the number of independent variables and n is the number of repetition of 
experiments at the center point. The total number of design points is thus 𝑁 = 2k +  2𝑘 + (𝑛o). 
4
Therefore, the CCRD involved 30 experiments consisting of 2  factorial CCD, with 8 axial 
points (α = 2) and 6 replications at the center points. For each independent variable, the levels 
were chosen with respect to moisture content of moringa seeds at harvest, preliminary 
experiments, observations and previous reports by various researchers on various oilseeds since 
there is no information as regards the optimization of the various process parameters that 
influence oil expression from moringa using oil expeller. Five levels of applied pressures (5, 10, 
15, 20 and 25 MPa) and moisture content levels of (8, 9, 10, 11 and 12 %wet basis) were chosen. 
o
Five heating temperature levels (50, 60, 70, 80 and 90 C) for heating times of (15, 20, 25, 30 and 
35 minutes) were also chosen. Coded values of the independent variables (-2, -1, 0, 1, 2) were 
used; where -2, 0 and 2 represent the lowest, medium and highest levels respectively. The coded 
values are set out in Tables 3.2 and 3.3. 
𝑁𝑎𝑡𝑢𝑟𝑎𝑙  𝑣𝑎𝑙𝑢𝑒 −𝐵𝑎𝑠𝑒  𝑙𝑒𝑣𝑒𝑙  (𝑙𝑒𝑣𝑒𝑙  0)
𝐶𝑜𝑑𝑒𝑑 𝑣𝑎𝑙𝑢𝑒 =                       … (76) 
𝐼𝑛𝑡𝑒𝑟𝑣𝑎𝑙  𝑜𝑓  𝑣𝑎𝑟𝑖𝑎𝑡𝑖𝑜𝑛
The relationships between the coded and natural values in the RSM design for the 
experiments are given below:   
(𝑀c−10)
𝑋1 =                                … (77) 1
(𝐻T−70)
𝑋2 =                                … (78) 10
(𝐻t−25)
𝑋3 =                                … (79) 5
(𝐴p−15)
𝑋4 =                                … (80) 5
X1 = Coded value for moisture content 
Mc = Natural value for moisture content 
55 
 
 
X2 = Coded value for heating temperature 
HT = Natural value for heating temperature 
X3 = Coded value for heating time 
Ht = Natural value for heating time 
X4 = Coded value for applied pressure 
Ap = Natural value for applied pressure 
3.8 Response Surface Methodology (RSM)      
 A Design Expert (version 6.0.6) software package for design of experiments was used to 
analyze and generate model equations for the expression process. Data obtained through the 
experimental matrix were computed for the determination of regression coefficient of the second 
order multiple regression models. The analysis of regression and variance was performed by the 
Design Expert. To validate the optimal parameters, the experiment was repeated at the optimal 
conditions as suggested by Islau et al. (2002). The obtained results were compared with 
predicted values. 
3.9 Statistical Analysis  
 The data obtained in the experiments were analyzed using Response Surface 
Methodology so as to fit the quadratic polynomial equation generated by the Design-Expert 
software version 6.0.6 (Stat-Ease Inc., Minneapolis, USA). In order to correlate the response 
variable to the independent variables, multiple regression was used to fit the coefficient of the 
polynomial model of the response. The quality of the fit of the model was evaluated using 
Analysis Of Variance (ANOVA). ANOVA was carried out to determine the significance and 
fitness of the model as well as the effect of significant individual terms and their interaction on 
the chosen responses. The P-value (probability of error value) was used as a tool to check the 
significance of each regression coeffiecient which also indicated the interaction effect of each 
cross product. Data obtained from the experiments were also statistically analyzed to determine 
the significant difference in the process conditions and their interactions at 5% probability level 
using SPSS window 20.0 software package.  
3.10 Experimental Tests of the Moringa Oil Expeller  
Moringa seeds were procured from the market, dehulled and thereafter, manually cleaned 
to remove foreign materials. A factorial experimental design (4 × 5) was used in the test at 
various moisture contents, heating temperatures, heating times and applied pressures. Initial 
moisture content of the sample was determined using the standard method as described in section 
3.1.1.2. One kg each of moringa seeds was conditioned to desired levels of moisture contents (8, 
9, 10, 11 and 12 %wet basis) using equation (58) as adopted by Olajide (2000): 
56 
 
 
 
Table 3.2. Levels, Codes and Intervals of Independent Variables for the Experiment 
Factors Codes -2 -1 0 1 2 Interval of Variation 
Moisture Content (%wet basis) X1 8 9 10 11 12 1 
o
Heating Temperature ( C) X2 50 60 70 80 90 10 
Heating Time (min) X3 15 20 25 30 35 5 
Applied Pressure (MPa) X4 5 10 15 20 25 5 
 
Table 3.3. Experimental Design (Second Order Design in the Four Variables) 
 X1 X2 X3 X4 Mc HT Ht Ap 
1  0.000  0.000  0.000  0.000 10.00 70.00 25.00 15.00 
2  1.000 -1.000 -1.000  1.000 11.00 60.00 20.00 20.00 
3  1.000  1.000  1.000 -1.000 11.00 80.00 30.00 10.00 
4 -1.000  1.000  1.000 -1.000 9.00 80.00 30.00 10.00 
5  0.000  0.000  2.000  0.000 10.00 70.00 35.00 15.00 
6  0.000  0.000  0.000  0.000 10.00 70.00 25.00 15.00 
7  1.000  1.000 -1.000 -1.000 11.00 80.00 20.00 10.00 
8  2.000  0.000  0.000  0.000 12.00 70.00 25.00 15.00 
9  1.000  1.000 -1.000  1.000 11.00 80.00 20.00 20.00 
10 -1.000 -1.000 -1.000  1.000 9.00 60.00 20.00 20.00 
11  0.000  0.000  0.000  2.000 10.00 70.00 25.00 25.00 
12  1.000 -1.000  1.000  1.000 11.00 60.00 30.00 20.00 
13 -1.000  1.000  1.000  1.000 9.00 80.00 30.00 20.00 
14 -1.000 -1.000  1.000  1.000 9.00 60.00 30.00 20.00 
15 -2.000  0.000  0.000  0.000 8.00 70.00 25.00 15.00 
16  1.000 -1.000 -1.000 -1.000 11.00 60.00 20.00 10.00 
17  0.000  0.000  0.000  0.000 10.00 70.00 25.00 15.00 
18  0.000  0.000  0.000 -2.000 10.00 70.00 25.00 5.00 
19  1.000  1.000  1.000  1.000 11.00 80.00 30.00 20.00 
20  0.000  0.000 -2.000  0.000 10.00 70.00 15.00 15.00 
21  0.000  0.000  0.000  0.000 10.00 70.00 25.00 15.00 
22  0.000  0.000  0.000  0.000 10.00 70.00 25.00 15.00 
23  1.000 -1.000  1.000 -1.000 11.00 60.00 30.00 10.00 
24 -1.000 -1.000 -1.000 -1.000 9.00 60.00 20.00 10.00 
25  0.000  0.000  0.000  0.000 10.00 70.00 25.00 15.00 
26  0.000 -2.000  0.000  0.000 10.00 50.00 25.00 15.00 
27 -1.000  1.000 -1.000  1.000 9.00 80.00 20.00 20.00 
28  0.000  2.000  0.000  0.000 10.00 90.00 25.00 15.00 
29 -1.000  1.000 -1.000 -1.000 9.00 80.00 20.00 10.00 
57 
 
 
30 -1.000 -1.000  1.000 -1.000 9.00 60.00 30.00 10.00 
100−Mi
𝑄 = (  - 1) ×  𝑊                     … (81) 
100−𝑀𝑓 s
Where, 
Q = Volume of water to be added, ml.            
Mi = Initial moisture content of the sample, % wet basis  
Mf = Final (desired) moisture content of sample % wet basis   
Ws = Weight of the sample, g  
 The conditioned samples were kept in cloth wrapped with polythene bags and stored in a 
 o
refrigerator set at a temperature of (4 + 1) C for a period of 15 hrs to allow for even distribution 
of moisture. The samples to be expressed were removed from the refrigerator and kept in a 
desiccator to prevent moisture loss prior to heating. They were then heated at 50, 60, 70, 80 and 
o
90 C for 15, 20, 25, 30 and 35 mins duration. Applied pressures of 5, 10, 15, 20 and 25 MPa 
were used in the experiment. Experiments were done in triplicates. The expressed oil (Plate 3.3) 
was collected and left to stand for 96 hours after which it was weighed (Weiss, 2000). The cake 
output was also weighed.       
3.11 Determination of the Total Oil Content of Moringa Seeds   
 In order to determine the initial total oil content of the moringa seeds, the method as 
described by Orhevba et al. (2013) was adopted. The samples collected were washed with 
distilled water. They were then air-dried and ground. 10 g of the ground sample was weighed and 
placed on a filter paper which was folded carefully. The filter paper containing the sample was 
then inserted into the Soxhlet apparatus (Plate 3.4). The weights of the filter paper and sample 
were recorded. 200 ml of the solvent (hexane) was measured using a measuring cylinder and 
o
then poured into a 500 ml round bottom flask with the sample and heated at 60 C for 5 hours 
o
after which the sample was removed and transferred into the air oven to dry at 105 C for 15 
minutes. The sample was then weighed and the difference was calculated using equation (82) 
below. The oil was recovered by solvent evaporation. It was heated at a low temperature until the 
solvent finally evaporated leaving behind the oil extracted. 
Weight  of  sample  before  extraction − Weight  of  sample  after  extraction
𝑂𝑖𝑙 𝑌𝑖𝑒𝑙𝑑 = × 100              ... (82)
Weight  of  sample  before  extraction
 This was taken as the total oil content of the moringa seeds and was used in the 
calculation of the percentage extraction efficiency in terms of oil yield.   
 The oil yield was determined from equation (83) below as adopted by Ajav and Olatunde 
(2011):   
58 
 
 
 
Plate 3.3. Expressed Moringa Oil Samples 
 
 
Plate 3.4. Soxhlet Apparatus 
   
  
59 
 
 
W
Y = o × 100                             ... (83)       
W g
Where, 
𝑌 = Oil yield, %            
Wg = Weight of Moringa, kg            
Wo = Weight of oil expressed, kg  
 The extraction efficiency was determined from equation (84) as adopted by Ajav and 
Olatunde (2011). The expressed oil was weighed and results were expressed as a percentage of 
the total weight of sample used for the expression. Results obtained for each experiment at 
different processing conditions were used to calculate the percent oil yield in comparison with 
the initial oil in the sample estimated by the Soxhlet method. 
O
ƞ = Y × 100                                ... (84)       
OT
Where, 
ƞ = Extraction efficiency, % 
OY = Oil yield, %                         
OT = Total oil content, %   
 The material balance efficiency in terms of the oil and cake output was determined from 
equation (85) below:    
W + W
φ = o  c  × 100                        ... (85) 
W T
Where, 
φ = Material balance efficiency, %          
Wo = Weight of oil, g                          
Wc = Weight of cake, g 
WT = Total weight of moringa seeds used for the experiment, g 
3.12 Physico-chemical Analysis of the Moringa Oil 
 The Free Fatty Acid (FFA) content, oil impurity and colour were used as criteria for 
assessing oil quality using AOAC standard methods.  
 3.12.1 Free Fatty Acid          
 10 g of the oil sample was accurately weighed into a 250 ml stopper flask. In a second 
o
flask, 50 ml of ethanol was heated to the boiling point and while still over 70 C, it was 
neutralized with 0.1 KOH using phenolphthalein indicator. The neutralized ethanol was poured 
in the first flask containing the oil and the content of the flask mixed. They were boiled, and 
while still hot, titrated with 0.1 KOH, shaking vigorously during the titration. The end point of 
60 
 
 
the titration was reached when the addition of a single drop of 0.1 KOH produced a slight, but 
definite colour change persisting for at least 15 seconds. 
The FFA was then calculated as: 
(5.61 × V × N)
FFA =                       … (86) 
𝑊
V = Volume in ml of 0.1 KOH used 
N = Normality of the ethanolic KOH (0.1) 
W = Weight of the oil sample 
 3.12.2 Colour  
 The oil was filtered through a filter paper to remove any impurity. A glass cell was 
cleaned with carbon tetrachloride and dried. The glass cell was filled with the oil and placed in 
position in the tintometer. The colours were matched with sliding red, yellow and blue colours. 
The colour was reported in terms of Lovibond units. 
Colour =  𝑌 + 5𝑅                        … (87) 
Y = Sum total of yellow slides used 
R = Sum total of red slides used 
 3.12.3 Oil Impurity         
o
 A filter paper and a beaker were dried in the oven between 103-105 C to constant weight. 
The filter paper was cooled in the dessicator for 30 mins and weighed (W1). The filter paper was 
then folded and placed in the beaker. 10 ml of oil sample was accurately measured and spread on 
the filter paper. Small quantity of petroleum ether was added to the filter paper containing the oil 
to enhance filtration. When the filtration was complete, the filter paper was removed and dried in 
an oven till all the oil evaporated. The filter paper was quickly transferred to the dessicator and 
cooled for 30 minutes. The filter paper was weighed (W2) and the oil impurity calculated as: 
W − W  
Oil Impurity =  2 1  × 100        … (88) 
ml  of  oil
 
 
 
 
 
 
 
 
 
61 
 
 
CHAPTER FOUR 
RESULTS AND DISCUSSION 
4.1 Physical Properties of Moringa  
The values of the physical properties of moringa seeds measured are presented in Tables 
A1 to A7 in Appendix A, while a summary of the results is shown in Tables 4.1-4.3.   
 The average moisture content was 7.31% wet basis and all the other experiments were 
conducted at this moisture content. The moisture content is very important as it influences the 
size, shape and angle of repose of the seeds; which in turn determines the hopper capacity and 
the free flow of seeds.          
 The average seed length, width and thickness were 8.45±0.976 mm, 7.82±0.922 mm and 
6.41±1.092 mm respectively. Reported values for the length, width and thickness (respectively) 
of other oil-bearing crops revealed 56.18±8.64 mm, 37.89±3.82 mm and 12.01±1.66 mm for oil 
bean seeds (Asoegwu, 2006); 10.63-11.09 mm, 7.27-7.99 mm and 6.02-7.41 mm for castor bean 
seeds (Shafiee et al., 2009); 12.81-14.50 mm, 7.02-8.42 mm, 2.22-2.49 mm for melon seeds 
(Davies, 2009); 2.5 mm, 1.6 mm and 0.94 mm for sesame seeds (Arafa, 2007); 11.21±1.60 mm, 
7.56±0.94 mm, 6.93±0.77 mm for groundnut kernels (Olajide, 2000); 31.50±0.28 mm, 
23.70±0.20 mm and 22.00±0.24 mm for sheanut kernels (Olajide, 2000); 3.34 mm, 2.13 mm 
and 0.80 mm for beniseeds (Olayanju, 2002); 58.87 mm, 18.96 mm and 15.64 mm for fennel 
seeds (Ahmadi et al., 2009); 12.14-12.57 mm, 5.79-6.38 mm, 3.86-4.09 mm for sunflower seeds 
(Seifi and Alimardani, 2010); 7.27-7.81 mm, 3.50-3.79 mm and 2.80-3.50 mm for safflower 
seeds (Aktas et al., 2006); 7.27-8.25 mm, 6.48-6.97 mm, 5.41-5.94 mm for soybean grains 
(Tavakoli et al., 2009) and 17.01-17.71 mm, 10.74-11.23 mm, 8.19-8.16 mm for jatropha seeds 
(Bamgboye and Adebayo, 2012). Comparatively, it would be observed that moringa seeds are 
smaller in length to all the above mentioned seeds except sesame, safflower, soybean and 
beniseeds. Also, the width is in the same range as that of castor bean and melon seeds, but 
slightly wider than groundnut kernels. Furthermore, the thickness is in the same range as that of 
castor bean seeds, but slighty lower than groundnut kernels. The thickness is also lower than 
sheanut kernels, fennel, jatropha and oil bean seeds, but higher than melon, sesame, sunflower, 
safflower, soybean and beniseeds. The oil bean seeds are approximately seven times longer, five 
times wider and two times thicker than moringa seeds. The sheanut kernels are approximately 
four times longer, three times wider and three times thicker than moringa seeds, while the fennel 
seeds are approximately seven times longer, two times wider and thicker than moringa seeds. 
The values of the linear dimension of the moringa seeds were used in the calculation of the 
62 
 
 
Table 4.1. Dimensional Properties of Moringa Seeds 
Physical No of Unit of Minimum Maximum Mean Standard Coefficient of 
Properties observations measurements Value Value Value Deviation Variation 
Length 100 mm 6.44 10.68 8.45 0.976 11.55 
Width 100 mm 5.98 9.91 7.82 0.922 11.79 
Thickness 100 mm 4.11 7.97 6.41 1.092 17.04 
Arithmetic 100 mm 5.813 8.963 7.560 0.866 11.46 
Mean Diameter 
Geometric Mean 100 mm 5.668 8.782 7.490 0.880 11.75 
Diameter 
Surface Area  100 mm2 100.93 242.29 177.47 40.08 22.58 
Sphericity  100 - 0.750 0.969 0.888 0.052 5.86 
Aspect Ratio 100 - 0.735 0.986 0.927 0.050 5.39 
 
Table 4.2. Gravimetric Properties of Moringa Seeds 
Physical No of Unit of Minimum Maximum Mean Standard Coefficient of 
Properties observations measurements Value Value Value Deviation Variation 
Moisture 3 % wet basis 6.97 7.58 7.31 0.312 23.43 
Content     
True Density  10 g/cm3 0.721 1.071 0.971 0.105 9.25 
Bulk Density  10 g/cm3 0.630 0.692 0.662 0.026 25.46 
Porosity - % - - 68.18 - - 
One Thousand 10 g 234.00 246.00 239.20 3.084 1.29 
Seed Weight 
 
Table 4.3. Frictional Properties of Moringa Seeds 
Physical No of Unit of Minimum Maximum Mean Standard Coefficient of 
Properties observations measurements Value Value Value Deviation Variation 
Coefficients of 10 o 0.932 1.192 1.027 0.072 14.26 
Static Friction 
[Glass] 
Coefficients of 10 o 1.036 1.150 1.111 0.036 30.86 
Static Friction 
[Stainless steel] 
Coefficients of 10 o 1.235 1.600 1.376 0.111 12.40 
Static Friction 
[Mild steel] 
Coefficients of 10 o 1.000 1.428 1.234 0.174 7.09 
Static Friction 
[Galvanized 
steel] 
Coefficients of 10 o 1.881 2.605 2.199 0.265 8.30 
Static Friction 
[Rubber] 
Coefficients of 10 o 1.327 1.963 1.607 0.257 6.25 
Static Friction 
[Plywood] 
Angle of Repose 10 o 20.46 22.66 21.44 0.745 28.78 
 
 
 
63 
 
 
amount of seeds that will be crushed at the feed end portion of the machine. It also assisted in 
determining the force that will be required to express the oil based on the number of seeds to be 
processed per batch.          
 The arithmetic mean diameter and geometric mean diameter of the moringa seeds were 
2
7.560±0.866 mm and 7.51±0.880 mm respectively. The surface area was 177.19±40.08 mm ; 
while the sphericity and aspect ratio were 0.889±0.052 (88.9%) and 0.9257±0.050 (92.57) 
respectively. Garnayak et al. (2008) considered any grain, fruit or seed as spherical when the 
sphericity value is above 70%, thus, the high sphericity of the moringa seeds is indicative of the 
shape towards being a sphere. This value is higher than those reported for other oil-bearing 
crops; 0.523±0.065 for oil bean seeds (Asoegwu, 2006); 0.69-0.73 for castor bean seeds (Shafiee 
et al., 2009); 0.47-0.53 for melon seeds (Davies, 2009); 0.63 for sesame seeds (Arafa, 2007); 
0.76 for groundnut kernels (Olajide, 2000); 0.80 for sheanut kernels (Olajide, 2000); 0.52-0.55 
for beniseeds (Olayanju, 2002); 0.45 for fennel seeds (Ahmadi et al., 2009); 0.53-0.55 for 
sunflower seeds (Seifi and Alimardani, 2010); 0.47-0.49 for safflower seeds (Aktas et al., 2006) 
and 0.672-0.684 for jatropha seeds (Bamgboye and Adebayo, 2012). However, it is in the same 
range as soybean grains (0.847-0.873) and canola seeds (0.82-0.93) as reported by Tavakoli et 
al., (2009) and Razavi et al. (2009) respectively.       
 Figures 4.1-4.8 show the frequency distribution of the dimensional properties for the 
hundred samples of moringa seeds measured.        
-3 -3 -3
 The average true and bulk densities were 0.971 gcm  (971 kgm ) and 0.662 gcm  (662 
-3
kgm ) respectively; while the porosity was computed from the values of the true density and 
bulk density using equation (26) and was found to be 68.18%. The true density of the moringa 
seeds showed that the seeds are slightly less dense than water and therefore will float on water. 
Comparatively, the true density is in the same range as those reported for sunflower seeds, (740-
-3 -3
980 kgm ) and jatropha seeds (863-1035 kgm ) (Seifi and Alimardani, 2010; and Bamgboye 
and Adebayo, 2012). It is higher than the values reported for castor bean seeds (704.75-761.74 
-3 -3 -3
kgm ), melon seeds (816.09-847.47 kgm ), sesame seeds (542.1-607.3 kgm ), fennel seeds 
-3 -3  
(889.08-937.98 kgm ) and safflower seeds (776-780 kgm ) (Aktas et al., 2006; Shafiee et al., 
2009; Davies, 2009; Darvishi, 2012 and Ahmadi et al., 2009). Also, it is lower than those 
-3 -3
reported for groundnut kernels (1010 kgm ), beniseeds (981-1042 kgm ), sheanut kernels (1170 
-3 -3
kgm ) and soybean grains (1126.43-1147.86 kgm ) (Olajide, 2000; Olayanju, 2002; Olajide, 
2000 and Tavakoli et al., 2009). The bulk density is higher than those reported for oil bean seeds 
-3 -3 -3
(588 kgm ), castor bean seeds (434.75-447.50 kgm ), soybean (625.36-650.95 kgm ), melon 
64 
 
 
40
35
30
25
20
15
10
5
0
5 6 7 8 9 10 11 12
Length (mm)
 
Figure 4.1. Frequency Distribution of the Length of Moringa Seeds 
 
65 
 
Frequency (%)
 
50
45
40
35
30
25
20
15
10
5
0
3 4 5 6 7 8 9 10 11
Width (mm)
  
Figure 4.2. Frequency Distribution of the Width of Moringa Seed 
66 
 
Frequency (%)
 
45
40
35
30
25
20
15
10
5
0
2 3 4 5 6 7 8 9
Thickness (mm)
                      
Figure 4.3. Frequency Distribution of the Thickness of Moringa Seeds 
67 
 
Frequency (%)
 
50
45
40
35
30
25
20
15
10
5
0
3 4 5 6 7 8 9 10
Arithmetic Mean Diameter (mm)
 
Figure 4.4. Frequency Distribution of the Arithmetic Mean Diameter of Moringa Seeds 
 
68 
 
Frequency (%)
 
45
40
35
30
25
20
15
10
5
0
3 4 5 6 7 8 9 10
Geometric Mean Diameter (mm)
 
Figure 4.5. Frequency Distribution of the Geometric Mean Diameter of Moringa Seeds 
 
69 
 
Frequency (%)
 
45
40
35
30
25
20
15
10
5
0
0 50 100 150 200 250 300
Surface Area (mm2)
   
Figure 4.6. Frequency Distribution of the Surface Area of Moringa Seeds 
70 
 
Frequency (%)
 
60
50
40
30
20
10
0
0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1
Sphericity
 
Figure 4.7. Frequency Distribution of the Sphericity of Moringa Seeds 
71 
 
Frequency (%)
 
 
90
80
70
60
50
40
30
20
10
0
0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1
Aspect Ratio
 
Figure 4.8. Frequency Distribution of the Aspect Ratio of Moringa Seeds 
72 
 
Frequency (%)
 
-3 -3 -3
seeds (405-543 kgm ); sesame seeds (640 kgm ), groundnut kernels (479.28 kgm ), beniseeds 
-3 -3 -3
(613-688 kgm ), fennel seeds (413.51-352.39 kgm ), sunflower seeds (380-410 kgm ), 
-3 -3
safflower seeds (520-547 kgm ), and jatropha seeds (428-474 kgm ) (Asoegwu, 2006; Shafiee 
et al., 2009; Tavakoli et al., 2009; Davies, 2009; Arafa, 2007; Davies, 2009; Olayanju, 2002; 
Ahmadi et al., 2009; Seifi and Alimardani, 2010; Seifi et al., 2010; and Bamgboye and Adebayo, 
2012). The densities were utilized in the estimation of the capacity of the expeller in feed section. 
The average one thousand seed weight was 239.20 g. Corresponding values reported for 
other oil bearing seeds showed that moringa seeds weigh more than soybean grains (171.50-
219.04 g), sunflower seeds (80.3-96.8 g), safflower seeds (52.68-68.8 g) and melon seeds (94.0-
110.0 g) (Tavakoli et al., 2009; Seifi and Alimardani, 2010; Aktas et al., 2006; and Davies, 
2009). However, other oil bearing seeds which weigh more than moringa seeds include castor 
bean seeds (401.0-468.8 g) and jatropha seeds (515.4-692.6 g) (Shafiee et al., 2009; and 
Bamgboye and Adebayo, 2012). One point worthy of note however is that the one thousand seed 
weight is a function of the individual mass (weight) of the seed/kernel/grain.  
The average coefficients of friction on six different surfaces namely glass, stainless steel, 
mild steel, galvanized steel, rubber and plywood were 1.027, 1.111, 1.376, 1.234, 2.199 and 
1.607 respectively. It was observed that the static coefficient of friction was highest on rubber 
and lowest on glass. This was in agreement with earlier reports on other oil bearing seeds by 
Shafiee et al. (2009) on castor bean seeds, Davies (2009) on groundnut kernels, Davies (2010) on 
melon, Olayanju (2002) on beniseeds, Tavakoli et al. (2009) on soybean grains and Karaj and 
Müller (2010) as well as Bamgboye and Adebayo (2012) on jatropha seeds. It was observed that 
the smoother the structural surface, the lower the coefficient of friction of the moringa seeds on 
the surface. The knowledge of the coefficient of friction was utilized during the calculations of 
the force required to translate and compress the moringa seeds and the frictional force resulting 
from the expeller screw‟s motion.        
 The mean angle of repose was 21.44º±0.745. Comparatively, it is in the range of sesame 
seeds, 20.16-28.67º ((Darvishi, 2012). Also, it is higher than the values reported for groundnut 
kernels, 17.0º (Olajide, 2000); and lower than those values reported for castor bean seeds (30.2-
34.8º), fennel seeds (36.33-48.66º), melon seeds (29.7-36º), sheanut kernels (34.0º), soybean 
grains (24.56-29.93º), sunflower seeds (41-57º), safflower seeds (47-56º)  and jatropha seeds 
(28-36º) (Shafiee et al., 2009; Shafiee et al., 2010; Davies, 2009; Olajide, 2000; Tavakoli et al., 
2009; Seifi and Alimardani, 2010; Seifi et al., 2010; and Bamgboye and Adebayo, 2012). The 
lower angle of repose was due to the high sphericity value obtained for moringa seeds compared 
to the other oil bearing seeds, making them to flow more easily than most oil bearing crops. The 
73 
 
 
angle of repose determined the angle at which chutes were positioned in order to achieve 
consistent flow of materials through the chute. To ensure free flow, an angle of repose which was 
modestly higher than the average angle of repose (21.44°±0.745) obtained for the moringa seeds 
was used.  
4.2 Mechanical Properties of Moringa  
The values of the mechanical properties of moringa seeds measured are illustrated in 
Tables B1 to B3 in Appendix B, while a summary of the results is shown in Table 4.4. 
The force-deflection curves are shown in Figures 4.10 and 4.11, while Figures 4.12 and 
4.13 show the force-strain curves for the samples tested. The force-deflection curves obtained in 
this investigation were similar to those obtained by previous researchers on different oil bearing 
crops viz: Ozumba and Obiakor (2011) on palm-kernel seed, Manuwa and Muhammad (2010) on 
shea kernel and Bamgboye and Adejumo (2011b) on roselle seeds. The average force to rupture 
the seed was obtained as 58.535 N varying from 49.9 N to 67.0 N, while the average rupture 
energy of the seed was 0.1344 N.m ranging from 0.1043 N.m to 0.1708 N.m. The average 
deformation was 5.099 mm varying from 4.85 mm to 5.354 mm. Maximum strain of 85% was 
recorded for the moringa seeds. The force was used in the calculation of the main forces acting 
on the screw thread of the expeller.        
 The bio-yield point in the force-deformation curves indicates the seed rupture point and 
this point was determined by a visual decrease in force as deformation increased. This point 
reflects the sensitivity of the moringa seeds to damage. The rupture point indicates failure over a 
significant volume of the material. Beyond the rupture point, the stress decreases rapidly with 
increasing deformation. The amount of energy required to bring about rupture in the seeds 
indicates its toughness. It is the area under the force-deformation curve before the rupture point. 
The Young‟s Modulus is the gradient of the initial straight line portion of the stress-strain curve 
2
and has been found to be 195.32+17.85 N/mm . The yield is the point at which the initial straight 
line portion of load/deformation curve dips (drops off). The mean value for the deformation at 
peak was found to be 5.0990+0.0974 mm and it represents the distance travelled at point of 
yield. The energy to peak is the workdone to the point of yield and was found to be 
0.1344+0.0185 N.m. The force at break is the force at which maximum deformation was reached 
and was found to be 58.420+5.479 N. The deformation at break represents the maximum 
deformation and was found to be 5.1241+0.0779 mm. The energy to break is the energy at the 
point of maximum deflection and was obtained to be 0.1708+ 0.0184 N.m.  
 In comparison with other oil bearing crops, Manuwa and Muhammad (2010) obtained 
maximum values of 588.55 N, 8.822 mm and 1.9999 N.m for small size shea kernel and  
74 
 
 
Table 4.4. Mechanical Properties of Moringa Seeds  
Physical No of Unit of Minimum Maximum Mean Standard Coefficient of 
Properties observations measurements Value Value Value Deviation Variation 
Force at Peak  20 N 48.900 67.000 58.535 5.472 10.70 
Deformation at 20 mm 4.8500 5.3540 5.0990 0.0974 52.57 
Peak 
2
Stress at Peak 20 N/mm  42.500 55.800 49.26 4.403 11.19 
Energy to Peak 20 N.m 0.1043 0.1708 0.1344 0.0185 7.26 
Force at Break 20 N 48.900 67.000 58.420 5.479 10.66 
Deformation at 20 mm 5.0530 5.3540 5.1241 0.0779 65.78 
Break 
2
Stress at Break  20 N/mm  42.500 55.800 49.12 4.412 11.13 
Energy to 20 N.m 0.1056 0.1708 0.1357 0.0184 7.38 
Break 
Force at Yield 20 N 19.800 54.300 39.000 7.1700 5.44 
2
Stress at Yield 20 N/mm  27.300 38.900 33.66 4.011 8.39 
Energy to 20 N.m 0.0277 0.1215 0.0584 0.0224 2.61 
Yield 
2
Young 20 N/mm  66.65 147.95 195.32 17.85 10.94 
Modulus 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
75 
 
 
940.61 N, 10.086 mm and 2.8946 N.m for large size shea kernel; for the applied force, 
deformation and rupture energy of shea kernel respectively. Olayanju (2002) obtained values of 
8.7 N and 18.6 N for the force required to rupture and express oil from Yandev-55 variety of 
beniseeds at rupture and oil points respectively, and 9.0 N and 20.8 N for the force required to 
rupture and express oil from E-8 variety of beniseeds at rupture and oil points respectively. The 
corresponding deformations ranged between 0.123-0.494 mm and 0.46-0.54 mm for rupture and 
oil point respectively. Karaj and Müller (2010) obtained maximum values of 0.96±0.25 mm, 
1.11±0.30 mm and 0.92±0.26 mm for the deformation at rupture point in horizontal (x), 
transverse (y) and vertical (z) directions respectively for a fraction size greater than 0.69 for 
jatropha seeds. Bamgboye and Adejumo (2011b) while working on roselle seeds obtained 23.45-
-2
49.05 N, 23.38-35.48 Nmm  and 0.41-0.50 mm for the compressive force, rupture stress and 
-2 -2
deformation respectively. Also, 27.0-38.15 Nmm , 33.3-52.56 N and 216.03-374.11 Nmm  
were obtained for the yield stress, rupture force and Young‟s Modulus respectively. Gupta and 
Das (2000) found that the compressive force required to cause the rupture of sunflower kernel 
ranges from 26.86-33.94 N. The sunflower seeds loaded in a vertical orientation absorbed more 
3 3
energy (144.7-222.9 J/m ) prior to rupture than those loaded in the horizontal(95.21-84.2 J/m ) 
orientation; while the sunflower kernels loaded in a vertical orientation required less energy 
3 3
(18.1-54.3 J/m ) to rupture than those loaded in the horizontal (38.9-65.8 J/m ) orientation.  
 Figures 4.9 and 4.10 show typical examples of force-deflection and force-strain curves 
for moringa seeds under compressive loading. 
4.3 Effects of Processing Conditions on Oil Yield of Moringa  
 The average summary of the oil yield results at the various processing conditions 
combinations using 4 factors, 5 levels, factorial Central Composite Rotatable Design (CCRD) of 
Response Surface Methodology is presented in Table 4.5 below:    
 For the range of variables considered in this study, the highest oil yield of 28.58% was 
obtained when moringa seeds were conditioned to a moisture content of 11% wet basis and 
o
heated at 80 C for 30 mins at an applied pressure of 20 MPa. In comparison, Adejumo et al. 
o
(2013) obtained an optimum oil yield of 33.7% at a heating temperature of 100 C and a heating 
time of 30 mins using soxhlet extraction method. Mohammed et al. (2003) obtained a yield of 
31% using organic solvent for extraction. Anwar et al. (2006) obtained an oil yield of 30.36, 
35.26 and 38.37% respectively from one drought (Layyah) and two irrigated regions (Rahim Yar 
Khan, Jhang) of Punjab, Pakistan using hexane as solvent for extraction. Anwar and Rashid    
76 
 
 
`  
Figure 4.9. Typical Example of Force-Deflection Curves for Moringa Seeds under Compressive 
Loading 
 
 
Figure 4.10. Typical Example of Force-Strain Curves for Moringa Seeds under Compressive 
Loading 
77 
 
 
Table 4.5. Oil Yield of Moringa at various Processing Conditions  
o
S/No MC (% wb) HT ( C) Ht (mins) AP (MPa) Oil Yield (%) 
1 10.00 70.00 25.00 15.00 24.34
2 11.00 60.00 20.00 20.00 23.26
3 11.00 80.00 30.00 10.00 26.16
4 9.00 80.00 30.00 10.00 18.88
5 10.00 70.00 35.00 15.00 18.23
6 10.00 70.00 25.00 15.00 24.09
7 11.00 80.00 20.00 10.00 25.33
8 12.00 70.00 25.00 15.00 17.27
9 11.00 80.00 20.00 20.00 27.79
10 9.00 60.00 20.00 20.00 16.32
11 10.00 70.00 25.00 25.00 23.66
12 11.00 60.00 30.00 20.00 24.55
13 9.00 80.00 30.00 20.00 20.68
14 9.00 60.00 30.00 20.00 17.20
15 8.00 70.00 25.00 15.00 13.68
16 11.00 60.00 20.00 10.00 19.36
17 10.00 70.00 25.00 15.00 24.08
18 10.00 70.00 25.00 5.00 11.42
19 11.00 80.00 30.00 20.00 28.58
20 10.00 70.00 15.00 15.00 14.23
21 10.00 70.00 25.00 15.00 24.20
22 10.00 70.00 25.00 15.00 24.24
23 11.00 60.00 30.00 10.00 20.20
24 9.00 60.00 20.00 10.00 14.91
25 10.00 70.00 25.00 15.00 24.30
26 10.00 50.00 25.00 15.00 17.46
27 9.00 80.00 20.00 20.00 17.00
28 10.00 90.00 25.00 15.00 20.34
29 9.00 80.00 20.00 10.00 16.10
30 9.00 60.00 30.00 10.00 14.46
   
MC = Moisture content of moringa seed  
HT = Heating temperature of moringa seed    
Ht = Heating time of moringa seed  
AP = Applied pressure on moringa seed 
 
78 
 
 
(2007) found out that the moringa seeds harvested from the forests of Kohat district of North 
West Frontier Province exhibited an oil yield of 34.80%. Nzikou et al. (2009) reported an oil 
yield of 38.5% and 40% using two oil extraction methods viz extraction with petroleum ether 
(soxlhet) and extraction with a mixture of chloroform:methanol (1:1) (blye and dyer) 
respectively. Goja (2013) reported the hexane-extracted oil yield of moringa to be 34.5%. This 
suggested that method of extraction has significant effect on oil yield as observed by Orhevba 
(2006). Also, Anwar et al. (2006) observed that genetic and environmental factors can affect the 
oil yield and quality of a vegetable oilseed crop. 
 Increase in moisture content showed a substantial improvement in oil recovery up to 11% 
wet basis. However, further increase in moisture content up to 12% led to a decline in the oil 
yield. Sivala et al. (1992) explained that any oil bearing material subjected to compression is not 
completely saturated with oil at the beginning. The oil emanates from the interior of particles and 
within a short time, if drainage is not permitted, it saturates the system. If at this stage an opening 
is provided, oil drains from the system. Addition of moisture helps to reach saturation point 
early. Following initial compression when the entrapped air is pushed out, the liquid, if present in 
optimum quantity, transmits the applied pressure in all directions and more oil cells are formed, 
thereby releasing a greater quantity of oil. As compression progresses and the oil drains out from 
the system, the load is slowly transferred to the rigid structure of the solid cake. Even if some oil 
is entrapped in the cake matrix, it cannot be expelled because the particles form a solid skeleton 
and prevent the load from being transferred to the oil. When excess moisture is present, the 
liquid phase takes the entire load; itself being incompressible and does not exert any pressure on 
the oil bearing particles, thus showing a decrease in oil yield. At higher moisture level, mucilage 
is developed in the outer cell and the addition of more water causes swelling of the mucilage; 
and this produces a cushioning effect which prevents the rupturing of the oil cells. Therefore, the 
optimum moisture content for moringa was found to be 11% wet basis for which an increase in 
moisture content causes a decline in oil yield. This conformed to earlier reports by Farsaie and 
Singh (1983) on sunflower, Bongiriwar (1977) on groundnut, Blake (1982) on canola seed, 
Olajide (2000) on groundnut and sheanut, Bamgboye and Adejumo (2011a) on roselle and 
Abidakun et al. (2012) on dika nut amongst others.     
o
 Increase in temperature from 50-80 C increased oil yield, but oil yield decreased with 
o
increase in heating temperature up to 90 C. Increase in heat treatment leads to hardening of 
samples, thereby offering increased resistance to pressure application during expression, leading 
o
to decrease in oil yield of samples. Oil yield obtained from samples heated at 50 C were 
considerably lower than those obtained from the other heating temperatures. This could be 
79 
 
 
attributed to insufficient heat treatment given to samples during heating. Oil yield increased with 
increasing time for samples up to 30 mins after which further increase in time caused a decline in 
o
oil yield. Samples heated at 50 C needed more time to allow for the coagulation of protein, 
increasing fluidity of the oil, breakdown of oil-cells and adjustment of moisture content to the 
o
optimum level, while samples heated at 80 C needed a relatively shorter time to achieve all these 
and further heat treatment decreases the oil yield. Increase in heating time increased the oil yield 
o
up to 80 C, beyond which further increase in heating time decreased the yield. This was due to 
the heat coagulation of the proteins which leads to the reduction of the viscosity of the oil being 
expressed while moisture loss takes place simultaneously. At higher temperatures, protein 
coagulation and viscosity reduction take place at a faster rate leading to increased yield at short 
durations; while extending the heating duration at higher temperatures caused substantial 
moisture loss leading to hardening of samples which consequently leads to a decrease in oil 
yield. Oil flow was found to be inversely proportional to the kinematic viscosity which decreases 
with increase in heating temperature, thus increase in the ability of the oil to flow (Kagwacie and 
o
Anozie, 1995). The highest oil yield of 28.58% was obtained when sample was heated at 80 C 
for 30 mins. This agrees with the findings of Adekola (1991) on coconut, Adeeko and Ajibola 
(1990) on groundnut, Olajide (2000) on groundnut and sheanut, Tunde-Akintunde et al. (2001) 
on soybean, Ajav and Olatunde (2011) on groundnut, Bamgboye and Adejumo (2011a) on 
roselle and Abidakun et al. (2012) on dika nut amongst others. It was observed that moringa 
seeds do not require a very high temperature during expression process as compared to some 
other oil bearing seeds like groundnut, melon, soybean, sheanut, coconut, roselle, cashew, dika 
nut amongst others which were reported by other researchers. At high temperatures, oil recovery 
becomes extremely low due to hardening of samples.      
 There was a significant increase in oil yield when applied pressure was increased up to 20 
MPa, but oil yield decreased when the pressure increased to 25 MPa. This was due to the sealing 
of some inter kernel voids at this higher pressure. According to Ward (1976), increasing the 
pressure applied on oilseeds during pressing tend to narrow, shear and may eventually seal the 
capillaries through which oil is expressed. This observation is in agreement with oil expression 
from other oilseeds as reported by various researchers viz Adeeko and Ajibola (1990) on 
groundnut, Sivala et al. (1992) on rice bran, Ajibola et al. (1990) on melon seeds, Reddy and 
Bohle (1993) on mustard seeds, Olajide (2000) on groundnut and sheanut, Abidakun et al. 
(2012) on dika nut, Bamgboye and Adejumo (2011a) on roselle amongst others.    
 The interactions between the process variables namely moisture content, heating 
80 
 
 
temperature, heating time and applied pressure on the yield of moringa oil during expression 
process are shown in Figures 4.11-4.16. 
 4.3.1 Moisture Content        
o
 Increase in moisture content from 9-11% wb at a heating temperature of 60 C, heating 
time of 30 mins and applied pressure of 20 MPa increases oil yield by 30%; while increase in 
o
moisture content from 9-11% wb at a heating temperature of 60 C, heating time of 20 mins and 
applied pressure of 10 MPa increases oil yield by 23%. Also, increase in moisture content from 
o
9-11% wb at a heating temperature of 80 C, heating time of 20 mins and applied pressure of 10 
MPa increases oil yield by 36.4%, while increase in moisture content from 9-11% wb at a 
o
heating temperature of 80 C, heating time of 30 mins and applied pressure of 10 MPa increases 
oil yield by 28%. Similarly, increase in moisture content from 9-11% wb at a heating 
o
temperature of 80 C, heating time of 30mins and applied pressure of 20 MPa increases oil yield 
by 27.6%. Finally, oil yield increases from a moisture content of 8-10% wb at a heating 
o
temperature of 70 C, heating time of 25 mins and applied pressure of 15 MPa by 30%, while it 
decreases by 29% by increasing the moisture content from 10-12% wb. This shows that the 
optimum moisture content for oil expression from moringa is 11%, above which there is a 
decline in the oil yield of moringa oil. 
 4.3.2 Heating Temperature 
o
  Increase in heating temperature from 60-80 C at a moisture content of 9% wb, heating 
time of 20 mins and applied pressure of 10 MPa increases oil yield by 7.4%; while increase in 
o
heating temperature from 60-80 C at a moisture content of 9% wb, heating time of 20 mins and 
applied pressure of 20 MPa increases oil yield by 4%. Similarly, increase in heating temperature 
o
from 60-80 C at a moisture content of 9% wb, heating time of 30 mins and applied pressure of 
o
20 MPa increases oil yield by 16.8%. Also, increase in heating temperature from 60-80 C at a 
moisture content of 11% wb, heating time of 20 mins and applied pressure of 20 MPa increases 
o
oil yield by 16.3%; while  increase in heating temperature from 60-80 C at a moisture content of 
11% wb, heating time of 30 mins and applied pressure of 20 MPa increases oil yield by 14%. 
o
Finally, oil yied increases from a heating temperature of 50-70 C at a moisture content of 10% 
wb, heating time of 25 mins and applied pressure of 15 MPa by 28.3%, while it decreases by 
o
16.4% by increasing the heating temperature from 70-90 C. This shows that the optimum heating 
o
temperature for oil expression from moringa is 80 C, above which there is a decline in the oil 
yield of moringa oil. 
  
81 
 
 





27.123   

 22.623   
18.123   
13.623   
  Oil Yield (%)   
9.123   
90.00   
  12.00 
80.00   
  11.00 
  Heating Temperature (oC)  70.00   
    10.00 
60.00     9.00 
Moisture Content (%wb)  
50.00    8.00 

Figure 4.11. Effect of Moisture Content and Heating Temperature on Oil Yield 
 





25.496   

 20.231   
14.966 
9.702   
  Oil Yield (%)  
4.437   
35.00   
  12.00 
30.00   
  11.00 
25.00   
  10.00 
  Heating Time (mins)   20.00     9.00 
     Moisture Content (%wb)  
15.00     8.00   
  
Figure 4.12. Effect of Moisture Content and Heating Time on Oil Yield 
82 
 
 





26.601   

 21.438   
16.276   
11.113   
  Oil Yield (%)
5.95 0  
25.00   
  12.00 
20.00   
  11.00 
15.00   
  10.00 
  Applied Pressure (MPa)  10.00     9.00 
    Moisture Content (%wb)
5.00    8.00 

Figure 4.13. Effect of Moisture Content and Applied Pressure on Oil Yield 





25.226   

 22.068   
18.911   
15.754   
    Oil Yield (%)  
   12.597   
35.00   
  90.00 
30.00   
  80.00 
25.00   
  70.00 
  Heating Time (mins)  20.00     60.00 
  Heating Temperature (oC)  
15.00     50.00 
Figure 4.14. Effect of Heating Temperature and Heating Time on Oil Yield 
83 
 
 





 25.558   

 21.461   
17.364   
13.267   
  Oil Yield (%)  
9.170 
25.00   
  90.00 
20.00   
  80.00 
15.00   
  70.00 
Applied Pressure (MPa)10.00     60.00 
  Heating Temperature (oC) 
5.00    50.00 

Figure 4.15. Effect of Heating Temperature and Applied Pressure on Oil Yield 
 





25.155   

 21.319   
17.484 
13.649   
  Oil Yield (%)  
9.813   
25.00   
  35.00 
20.00   
  30.00 
15.00   
  25.00 
Applied Pressure (MPa)10.00     20.00 
   Heating Time (mins)  
5.00    15.00   


Figure 4.16. Effect of Heating Time and Applied Pressure on Oil Yield 
  
84 
 
 
 4.3.3 Heating Time 
 Increase in heating time from 20-30 mins at a moisture content of 9% wb, heating 
o
temperature of 60 C and applied pressure of 20 MPa increases oil yield by 5%; while increase in 
o
heating time from 20-30 mins at a moisture content of 9% wb, heating temperature of 80 C and 
applied pressure of 20 MPa increases oil yield by 17.8%. Also, increase in heating time from 20-
o
30 mins at a moisture content of 11% wb, heating temperature of 60 C and applied pressure of 
10 MPa increases oil yield by 4.2%; while increase in heating time from 20-30 mins at a 
o
moisture content of 11% wb, heating temperature of 60 C and applied pressure of 20 MPa 
increases oil yield by 5.3%. Similarly, increase in heating time from 20-30 mins at a moisture 
o
content of 11% wb, heating temperature of 80 C and applied pressure of 10 MPa increases oil 
yield by 3.2%. Finally, oil yied increases from a heating time of 15-25 mins at a moisture content  
o
of 10% wb, heating temperature of 70 C and applied pressure of 15 MPa by as high as 41.5%, 
while it decreases by 25% by increasing the heating time from 25-35 mins. This shows that the 
optimum heating time for oil expression from moringa is 30 mins, above which there is a decline 
in the oil yield of moringa oil. 
 4.3.4 Applied Pressure        
 Increase in applied pressure from 10-20 MPa at a moisture content of 9% wb, heating 
o
temperature of 60 C and heating time of 20 mins increases oil yield by 8.6%; while increase in 
o
applied pressure from 10-20 MPa at a moisture content of 9% wb, heating temperature of 80 C 
and heating time of 20 mins increases oil yield by 5.3%. Similarly, increase in applied pressure 
o
from 10-20 MPa at a moisture content of 9% wb, heating temperature of 80 C and heating time 
of 30 mins increases oil yield by 8.7%. Also, increase in applied pressure from 10-20 MPa at a 
o
moisture content of 11% wb, heating temperature of 60 C and heating time of 30 mins increases 
oil yield by 17.7%; while increase in applied pressure from 10-20 MPa at a moisture content of 
o
11% wb, heating temperature of 60 C and heating time of 20 mins increases oil yield by 8.9%. 
Finally, oil yied increases from an applied pressure of 5-15 MPa at a moisture content of 10% 
o
wb, heating temperature of 70 C and heating time of 25 mins by as high as 53%, while it 
decreases by 2.8% by increasing the applied pressure from 15-25 MPa. This shows that the 
optimum applied pressure for oil expression from moringa is 20 MPa, above which there is a 
decline in the oil yield of moringa oil. 
4.4 Optimization of the Expression Process of Moringa Seeds using Yield as Response  
  Four different models namely linear, two factorial interaction (2FI), quadratic and cubic 
were used to analyze the expression process for moringa seeds and the models were fitted to the 
experimental data using Design Expert software. The appropriate model was chosen based on the 
85 
 
 
selection of the highest order polynomial where the additional terms are significant and the 
model is not aliased, insignificant lack-of-fit and the maximization of the “Adjusted R-Squared” 
and the “Predicted R-Squared”. Considering these, the linear and quadratic models were 
2
suggested. In terms of higher coefficient of determination (R ) and lower standard deviation 
values, the quadratic model was finally chosen ahead of the linear model to predict the oil 
expression process for moringa seeds. The final equation is given below:  
Y = -203.79542 + 29.44125Mc + 0.40712HT + 2.74250Ht + 1.11525Ap – 1.57708 Mc2 – 
0.007208HT2 – 0.055533Ht2 – 0.042433Ap2 + 0.067000McHT – 0.039250McHt + 
0.078500McAp + 0.006900HTHt – 0.006025HTAp + 0.006600HtAp        … (89)  
(Std. Dev. = 3.08, R-Squared = 0.7691, Mean = 20.41, Adj R-Squared = 0.5536, 
C.V. = 15.07, Pred R-Squared = -0.3294, PRESS = 817.61, Adeq Precision = 6.952). 
Where, 
Y = Oil yield, % 
Mc = Moisture content, %wb  
o
HT = Heating temperature, C 
Ht = Heating time, mins 
Ap = Applied pressure, MPa 
 The positive terms in the equation represent a direct relationship between processing 
conditions and interactions with yield, while the negative terms represent an inverse relationship 
between them. It was observed that moisture content, heating temperature, heating time and 
applied pressure all have a direct relationship with oil yield. Increase in moisture content, heating 
temperature, heating time and applied pressure leads to increase in oil yield. Also, it was found 
that moisture content is the most important factor affecting oil yield from moringa seeds. This 
conforms with the findings of Khan and Hanna (1984) on soybean, Sivakumarran (1985) on 
peanut, Olajide (2000) on groundnut and sheanut and Akinoso (2006) on sesame seeds. 
 The Model F-value of 3.57 implies that the model is significant. Values of "Prob > F" 
less than 0.0500 indicate model terms are significant. In this case A, B, D, A2, C2 are significant 
model terms (Appendix G). This implies that the moisture content, heating temperature, heating 
time and applied pressure all have significant effects on oil yield with the moisture content 
having the greatest influence on oil yield. Therefore, the four procesing conditions influenced the 
quantity of oil recovery from moringa seeds using expeller. Values of "Prob > F" greater than 
0.1000 indicate the model terms are not significant. The "Lack of Fit F-value" of 1238.86 implies 
the lack of fit is significant. It was found that the model was significant with a very low 
86 
 
 
2 
probability value (< 0.0001) and a satisfactory coefficient of determination (R = 0.7691). The 
high coefficient of determination showed excellent correlations between the independent 
variables. This value indicates that the response model can explain 76.91% of the total variability 
in the responses.           
 The statistical analysis of the oil yield showed that all the factors were significant except 
the interaction among the heating temperature, heating time and applied pressure. This shows 
that all the processing variables have significant effects on the yield of moringa oil during 
expression process. It has a mean of 19.651, standard error of 0.021 and lower and upper bounds 
of 19.597 and 19.705 respectively. The results of the statistical analysis are presented in 
Appendix H. 
4.5 Validation of Model 
 An excellent agreement between the observed and predicted values for oil yield from 
moringa was obtained from the parity plot between the predicted and the actual values as shown 
2
in Figure 4.17. There is a high correlation (R  = 0.7691) between the predicted and experimental 
values for moringa oil yield which indicated that the predicted values and experimental values 
are in reasonable agreement. It means that the data fitted well with the model and gave a 
convincingly good estimate of response for the expression process in the range studied. In the 
o
range of 8-12% wet basis for moisture content, 50-90 C for heating temperature, 15-35 mins for 
heating time and 5-25 MPa for applied pressure, predicted optimum oil yield of 28.2% at 
o
moisture content of 11.30%, temperature of 85.57 C, duration of 27.17 mins and pressure of 
19.63 MPa was obtained. Under these optimal conditions, the experimental value was 28.22% 
which was in agreement with those predicted by computation. Deviations between experimental 
and predicted values were low and ranged from 0.01-6.20. This shows that the model chosen 
predicted the oil yield adequately.  
4.6 Effects of Processing Conditions on the Physico-chemical Properties of Moringa    
            Oil 
 The results of the Free Fatty Acid (FFA), colour and oil impurities of moringa oil at 
various processing conditions are presented in Table 4.6.      
 4.6.1 Free Fatty Acid        
 The free fatty acid value measures the extent to which the glycerides in the oil have been 
decomposed by lipase action. This decomposition is accelerated by light and heat, hence, 
rancidity is usually accompanied by free fatty acid formation. The FFA of the expressed moringa 
oil ranged between 2.42-7.40 mg/KOH/g for all the processing conditions. It was observed that 
 
87 
 
 

 DESIGN-EXPERT Plot Predicted vs. Actual
Yield
28.58
24.29
Predicted 
20.00
15.71
11.42
11.42 15.71 20.00 24.29 28.58
Actual 
Actual
 
Figure 4.17. Predicted and Actual values for Moringa Oil Yield 
88 
 
Predicted
 
Table 4.6. Physico-chemical Properties of Moringa Oil 
S/No MC HT Ht AP FFA CL OI 
1 10.00 70.00 25.00 15.00 5.61 6.6 2.74 
2 11.00 60.00 20.00 20.00 7.16 6.1 3.07 
3 11.00 80.00 30.00 10.00 2.93 7.3 2.32 
4 9.00 80.00 30.00 10.00 2.76 7.2 2.26 
5 10.00 70.00 35.00 15.00 4.16 6.8 2.61 
6 10.00 70.00 25.00 15.00 5.67 6.6 2.72 
7 11.00 80.00 20.00 10.00 3.88 7.0 2.57 
8 12.00 70.00 25.00 15.00 6.07 6.7 2.81 
9 11.00 80.00 20.00 20.00 3.65 7.0 2.52 
10 9.00 60.00 20.00 20.00 7.03 5.9 2.94 
11 10.00 70.00 25.00 25.00 4.83 6.6 2.66 
12 11.00 60.00 30.00 20.00 6.93 6.2 2.89 
13 9.00 80.00 30.00 20.00 2.64 7.2 2.21 
14 9.00 60.00 30.00 20.00 6.74 6.1 2.85 
15 8.00 70.00 25.00 15.00 4.66 6.4 2.64 
16 11.00 60.00 20.00 10.00 7.21 6.0 3.11 
17 10.00 70.00 25.00 15.00 5.55 6.6 2.71 
18 10.00 70.00 25.00 5.00 5.98 6.5 2.76 
19 11.00 80.00 30.00 20.00 2.82 7.4 2.30 
20 10.00 70.00 15.00 15.00 6.41 6.3 2.83 
21 10.00 70.00 25.00 15.00 5.45 6.5 2.75 
22 10.00 70.00 25.00 15.00 5.36 6.6 2.72 
23 11.00 60.00 30.00 10.00 7.05 6.2 2.91 
24 9.00 60.00 20.00 10.00 7.06 5.9 2.98 
25 10.00 70.00 25.00 15.00 5.27 6.5 2.74 
26 10.00 50.00 25.00 15.00 7.40 5.7 3.20 
27 9.00 80.00 20.00 20.00 3.34 6.9 2.44 
28 10.00 90.00 25.00 15.00 2.42 7.6 2.12 
29 9.00 80.00 20.00 10.00 3.45 6.9 2.46 
30 9.00 60.00 30.00 10.00 6.84 6.1 2.86 
 
FFA = Free Fatty Acid, mg/KOH/g 
CL = Colour, LUY 
OI = Oil Impurity, % 
 
 
 
89 
 
 
the FFA increased with increase in moisture content. This as explained by Akinhanmi and 
Akintokun (2008) and Suganya et al. (2000) was due to the fact that during extraction of oil from 
oilseeds, water inside the seed bed flows out along with the oil and reacts with triglycerides to 
form free fatty acids and glycerols. Conversely, the FFA decreases with other processing 
conditions. This conforms to earlier reports by Olayanju (2002), Akinoso (2006) and Adejumo et 
al. (2013). Orhevba (2013) obtained values of 8.27±0.19 mg/KOH/g at a heating temperature of 
o
60 C for 5 hrs after which the moringa sample was removed and transferred into the air oven to 
o
dry at 105 C for 15 minutes using solvent (hexane) extraction. Adejumo et al. (2013) obtained 
o
values of 2.74, 2.71 and 2.70 mgKOH/g for moringa samples heated at 100, 130, 150 C 
respectively for 30 mins using soxhlet extraction; and 5.80 mgKOH/g for unheated sample. 
Ojiako and Okeke (2013) obtained a value of 2.51 mg/KOH/g for moringa seeds using soxhlet 
extraction. These showed that FFA of moringa decreases with increase temperature. Increase in 
temperature inactivates lipolytic enzymes which could cause rapid degradation of oil, hence, 
leading to a low FFA. The interactions between the process variables namely moisture content, 
heating temperature, heating time and applied pressure on the FFA of moringa oil during 
expression process are shown in Figures 4.18-4.23. 
 4.6.2 Colour 
 One of the key determinants of quality is the colour of the moringa oil. Colour is an 
indirect measure of product quality or processing performance. Though appearance properties 
have only aesthetic value with little influence on performance, it is the most obvious product 
characteristic evaluated by any consumer of vegetable oil. Dark coloured oils require special 
refining, thereby increasing the cost of production. The colour of the moringa oil was found to be 
pale yellow at all the processing conditions. Weiss (1983) stated that high quality oils must be 
pale or colourless and therefore, dark coloured oils are undesirable. This is because they will 
need to be specially refined and this invariably will increase the cost of production. This 
conforms to earlier reports by Mohammed et al. (2003) and Anwar and Rashid (2007) who 
observed a pale yellow colour for moringa oil. Akinoso (2006) reported that in practice, most 
consumers are attracted to a colour range of between 5.4-7.7 LU. The colour intensity of the 
expressed moringa oil ranged between 5.7-7.6 LU for all the processing conditions. Even though 
the colour at all processing conditions are within tolerable range, it was observed that colour 
intensity increased with increase in all the process variables. This conforms to earlier findings of  
 
90 
 
 





8.07   

 6.15   
4.23   
2.30   
  FFA (mgKOH/g)  
  0.38   
90.00   
  12.00 
80.00   
  11.00 
70.00   
  10.00 
  Heating Temperature (oC)  60.00     9.00 
  Moisture Content (%wb)  
50.00     8.00 
 
Figure 4.18. Effect of Moisture Content and Heating Temperature on Oil FFA 
 





8.07   


 6.15   
  
4.23  
2.30 
    FFA (mgKOH/g)  
     0.38  
35.00   
  12.00 
30.00   
  11.00 
25.00   
  10.00 
Heating Time (mins)  20.00     9.00 
Moisture Content (%wb)  
15.00    8.00 
 
Figure 4.19. Effect of Moisture Content and Heating Time on Oil FFA. 
 
91 
 
 





8.07 

 6.15 
4.23 
2.30 
    FFA (mgKOH/g)  
     0.38 

25.00   
  12.00 
20.00   
  11.00 
15.00   
  10.00 
  Applied Pressure (MPa)  10.00     9.00 
  Moisture Content (%wb)  
   5.00    8.00 
 
Figure 4.20. Effect of Moisture Content and Applied Pressure on Oil FFA 
 





8.07 

 6.15 
4.23 
2.30 
    FFA (mgKOH/g)  
     0.38 


35.00   
  90.00 
30.00   
  80.00 
25.00   
  70.00 
Heating Time (mins)  20.00     60.00 
 Heating Temperature (
oC)
15.00     50.00 
 
Figure 4.21. Effect of Heating Temperature and Heating Time on Oil FFA 




92 
 
 





8.07 

 6.15 
4.23 
2.30 
      FFA (mgKOH/g)  
     0.38 

25.00   
  90.00 
20.00   
  80.00 
15.00   
  70.00 
  Applied Pressure (MPa)   10.00     60.00 
   Heating Temperature (
oC)
5.00    50.00 
 
Figure 4.22. Effect of Heating Temperature and Applied Pressure on Oil FFA 
 





8.07 

 6.15 
4.23 
2.30 
  FFA (mgKOH/g)  
  0.38 
25.00   
  35.00 
20.00   
  30.00 
15.00   
  25.00 
  Applied Pressure (MPa)  10.00     20.00 
 Heating Time (mins)  
5.00    15.00 
 
Figure 4.23. Effect of Heating Time and Applied Pressure on Oil FFA  
 
 
93 
 
 
other researchers on oil bearing seeds. Akinoso et al. (2006) also observed that colour intensity 
of sesame oil increased with increase moisture content. Adeeko and Ajibola (1990) and Olajide  
 (2000) observed that increasing temperature and time of heating increased the colour intensity of 
o o
oils. They observed respectively that at temperatures above 135 C and 105 C respectively, the 
colour of groundnut kernel oils were getting darker and unattractive. Olajide (2000) also 
o
observed the colour of sheanut kernel oil turning white at temperatures above 110 C and getting 
o
darkish at 130 C. Ajav and Olatunde (2011) observed an offensive burnt smell for groundnut 
o
expression at 120 C. Adeeko and Ajibola (1990) reported that increase in heating temperature or 
heating time causes moisture loss from the sample, thereby increasing the colour intensity of oils. 
These showed that increase heating temperature at prolonged heating times have adverse effects 
on the colour of expressed oil. Oil colour is due to the presence of carotenoid and chlorophyll 
pigments. It is noteworthy that co-oxidation of carotein and fatty acid causes significant colour 
formation during thermal process. High temperature accelerates oxidation with a consequent 
o
colour rise; but since the highest temperature was 90 C in this experiment, adverse colour change 
was not observed. The interactions between the process variables namely moisture content, 
heating temperature, heating time and applied pressure on the colour intensity of moringa oil 
during expression process are shown in Figures 4.24-4.29. 
4.6.3 Oil Impurity 
  This has to do with the quality of the oil. The impurity ranges between 2.12-3.20% for all 
the processing conditions. It was observed that the impurity increases with increase in moisture 
content. Moisture constitutes about 87% of volatile impurities in vegetable oil, this account for 
the noticeable influence of moisture content on impurities. Conversely, impurity decreases with 
other processing conditions. Increase in heating temperature and time results in a decrease of 
moisture, therefore, impurities decrease with increase in heating temperature and time. This was 
in agreement with earlier reports by Akinoso (2006) who obtained a similar trend while 
investigating the effects of processing conditions on volatile impurities of palm kernel and 
o
sesame seed oils at 105 C. The interactions between the process variables namely moisture 
content, heating temperature, heating time and applied pressure on the moringa oil impurities 
during expression process are shown in Figures 4.30-4.35. 
4.7 Predictive Models for The Physico-chemical Properties of Moringa Oil   
 Four different models namely linear, two factorial interaction (2FI), quadratic and cubic 
were used to analyze the physico-chemical properties of moringa seeds namely free fatty acid, 
colour intensity and oil impurity. The models were fitted to the experimental data using Design  
94 
 
 





7.64   

 7.07   
6.49   
Colour (LU)  5.92 

5.34 
90.00   
  12.00 
80.00   
  11.00 
70.00   
  10.00 
  Heating Temperature (oC)60.00     9.00 
 Moisture Content (%wb)  
50.00     8.00 
   
Figure 4.24. Effect of Moisture Content and Heating Temperature on Oil Colour Intensity 
 





7.64 

 7.07 
6.49 
5.92 
Colour (LU)  

5.34 

35.00   
  12.00 
30.00   
  11.00 
25.00   
  10.00 
Heating Time (mins)  20.00     9.00 
   Moisture Content (%wb)  
15.00    8.00 
   
Figure 4.25. Effect of Moisture Content and Heating Time on Oil Colour Intensity 
 
 
95 
 
 





7.64 

 7.07 
6.49 
Colour (LU)  5.92 

5.34 


25.00   
  12.00 
20.00   
  11.00 
15.00   
  10.00 
Applied Pressure (MPa)  10.00     9.00 
Moisture Content (%wb)  
5.00    8.00 
 
Figure 4.26. Effect of Moisture Content and Applied Pressure on Oil Colour Intensity 
 





7.64 

 7.07 
6.49   
  5.92 Colour (LU)  

5.34   
35.00   
  90.00 
30.00   
  80.00 
25.00   
  70.00 
Heating Time (mins)  20.00     60.00 
 Heating Temperature (
oC)
15.00    50.00 
 
Figure 4.27. Effect of Heating Temperature and Heating Time on Oil Colour Intensity 
 
 
96 
 
 





7.64   

 7.07  
6.49   
  Colour (LU)  5.92   

5.34 
25.00   
  90.00 
20.00   
  80.00 
15.00   
  70.00 
Applied Pressure (MPa)   10.00     60.00 
   Heating Temperature (
oC)
5.00    50.00 
 
Figure 4.28. Effect of Heating Temperature and Applied Pressure on Oil Colour Intensity 
 





7.64  

 7.07   
6.49   
  Colour (LU)  5.92  
5.34   
25.00   
  35.00 
20.00   
  30.00 
15.00   
  25.00 
  Applied Pressure (MPa)   10.00     20.00 
 Heating Time (mins)  
 5.00    15.00 
 
Figure 4.29. Effect of Heating Time and Applied Pressure on Oil Colour Intensity 
 
  
97 
 
 





 3.17   

   2.86   
  2.55   
Impurity (%)    2.24   
  1.93 
90.00   
  12.00 
80.00   
  11.00 
70.00   
  10.00 
  Heating Temperature (oC) 60.00     9.00 
  Moisture Content (%wb)  
50.00     8.00 
   
Figure 4.30. Effect of Moisture Content and Heating Temperature on Oil Impurity 





 3.17 

   2.86   
  2.55 
  Impurity (%)    2.24  
 
 1.93 
35.00   
  12.00 
30.00   
  11.00 
25.00   
  10.00 
Heating Time (mins)   20.00     9.00 
   Moisture Content (%wb)  
15.00    8.00 
 
Figure 4.31. Effect of Moisture Content and Heating Time on Oil Impurity 
98 
 
 





 3.17   

   2.86   
  2.55 
  Impurity (%)  2.24  

 1.93   
25.00   
  12.00 
20.00   
  11.00 
15.00   
  10.00 
Applied Pressure (MPa)   10.00     9.00 
Moisture Content (%wb)  
5.00    8.00 
 
Figure 4.32. Effect of Moisture Content and Applied Pressure on Oil Impurity 
 





 3.17   

   2.86  
  2.55   
  Impurity (%)   2.24   

 1.93   
35.00   
  90.00 
30.00   
  80.00 
25.00   
  70.00 
  Heating Time (mins)   20.00     60.00 
    Heating Temperature (oC)
 15.00    50.00 
 
Figure 4.33. Effect of Heating Temperature and Heating Time on Oil Impurity 
99 
 
 





 3.17  

    2.86   
  2.55   
Impurity (%)  2.24   
 1.93   
25.00   
  90.00 
20.00   
  80.00 
15.00   
  70.00 
Applied Pressure (MPa)  10.00     60.00 
o
      Heating Temperature ( C)
o
5.00    50.00 ( C)

 
Figure 4.34. Effect of Heating Temperature and Applied Pressure on Oil Impurity 
 





3.17   

   2.86 
  2.55   
Impurity (%)2.24 
  1.93  
25.00   
  35.00 
20.00   
  30.00 
15.00   
  25.00 
  Applied Pressure (MPa)  10.00     20.00 
   Heating Time (mins)  
5.00    15.00 
 
Figure 4.35. Effect of Heating Time and Applied Pressure on Oil Impurity 
 
 
100 
 
 
Expert software. The appropriate model was chosen based on the selection of the highest order 
polynomial where the additional terms are significant and the model is not aliased, insignificant 
lack-of-fit and the maximization of the “Adjusted R-Squared” and the “Predicted R-Squared”. 
Considering these, for the FFA, only the linear model was suggested and therefore chosen. For 
the colour intensity, the linear and 2FI models were suggested. In terms of higher coefficient of 
2
determination (R ) and lower standard deviation values, the 2FI model was finally chosen. For 
the oil impurity, only the linear model was suggested and therefore chosen. The final equations 
for the FFA, colour and oil impurity are given below:  
FFA = +18.18808 + 0.10375Mc - 0.16004HT - 0.088917Ht - 0.043917Ap         … (90) 
(Std. Dev. = 0.53, R-Squared = 0.9052, Mean = 5.14, Adj R-Squared = 0.8901 
C.V. =  10.35, Pred R-Squared = 0.8574, PRESS = 10.65, Adeq Precision = 29.468). 
 
CI = +2.42250 + 0.020833Mc + 0.075417HT - 0.054167Ht - 0.058333Ap - 0.002500McHT + 
0.005000McHt + 0.007500McAp + 0.000250HTHt - 0.000500HTAp + 0.001000HtAp  … (91) 
(Std. Dev. =  0.064, R-Squared = 0.9884, Mean = 6.59, Adj R-Squared = 0.9822 
C.V. = 0.97, Pred R-Squared = 0.9614, PRESS = 0.26, Adeq Precision = 51.018). 
 
OI = +4.76817 + 0.033333Mc - 0.026917HT - 0.018000Ht - 0.005667Ap      … (92) 
(Std. Dev. = 0.063, R-Squared = 0.9529, Mean = 2.68, Adj R-Squared = 0.9454 
C.V. =  2.33, Pred R-Squared = 0.9300, PRESS = 0.15, Adeq Precision = 42.168). 
Where, 
Mc = Moisture content, %wb  
o
HT = Heating temperature, C 
Ht = Heating time, mins 
Ap = Applied pressure, MPa 
FFA = Free Fatty Acid, mg/KOH/g 
CI = Colour Intensity, LU 
OI = Oil Impurity, % 
 Deviations between experimental and predicted values were low and ranged from 0.01-
0.68, 0.01-0.12 and 0.01-0.12 for the FFA, colour and oil impurity respectively. This implies that 
for the range of variables studied, the models chosen could predict the FFA, colour intensity and 
oil impurity adequately.         
 The statistical analysis of the FFA showed that only the moisture content, heating 
temperature, heating time, applied pressure and the interaction between the heating temperature 
101 
 
 
and heating time are significant; while the other interactions are non-significant. This implies 
that all the processing variables have significant effects on the FFA of moringa oil during 
expression process, more especially, the heating temperature and heating time. This testifies to 
the fact that free fatty acid value measures the extent to which the glycerides in the oil have been 
decomposed by lipase action and the decomposition is accelerated by light and heat. It has a 
mean of 5.156, standard error of 0.030 and lower and upper bounds of 5.079 and 5.234 
respectively. 
 The statistical analysis of the colour intensity showed that only the moisture content, 
heating temperature and heating time are significant, while the applied pressure and all the 
various interactions were non-significant. Even though the heating temperature and heating time 
have significant effects on the colour intensity of moringa oil, their interactions do not have 
because of the careful selection of medium heating temperature and heating time values used in 
this experiment. At higher heating temperatures and heating times, the effect of colour change 
becomes pronounced; therefore, having a significant effect. It has a mean of 6.583, standard error 
of 0.010 and lower and upper bounds of 6.557 and 6.609 respectively.    
 The statistical analysis of the oil impurity showed that only the moisture content, heating 
temperature, heating time, applied pressure and the interaction between moisture content and 
heating time as well as heating temperature and heating time are significant; while the other 
interactions are non-significant. This implies that all the processing variables have significant 
effects on oil impurity of moringa during expression process, more especially, the heating time. 
Moisture constitutes about 87% of volatile impurities in vegetable oil and increase in heating 
temperature and time results in a decrease of moisture. It has a mean of 2.682, standard error of 
0.003 and lower and upper bounds of 2.674 and 2.690 respectively.       
 The results of the statistical analysis are presented in Appendix H.  
4.8 Efficiency of the Moringa Oil Expeller      
 The efficiencies of the moringa oil expeller in terms of extraction and material balance 
were calculated from the oil yield, and the oil and cake output respectively as given in equations 
(84) and (85). The average summary of the material balance in terms of oil and cake output is 
presented in Table 4.7.         
 The total oil content was found to be 35% (Appendix C). This was in agreement with 
most researchers who observed that moringa seeds have 30-40% (w/w) of oil content 
(Mohammed et al., 2003; Anwar et al., 2006; Anwar and Rashid, 2007; Nzikou et al., 2009; 
Uzama et al., 2011; Adejumo et al., 2013; Ogunsina et al., 2011, Goja, 2013 and Orhevba, 2013 
amongst others). From equation (84), the maximum oil yield of 28.58% corresponds to an 
102 
 
 
efficiency of 81.7%. This was taken as the extraction efficiency of the machine in terms of the 
oil yield. To determine the material balance efficiency of the machine in terms of the oil and 
cake output, average efficiency was calculated from the oil and cake output results as shown in 
Table 4.7. An efficiency of 93.8% was obtained. Overall, the performance of the moringa oil 
expeller was satisfactory as evident from the efficiencies in terms of the extraction and material 
balance. 
4.9 Cost Analysis of the Moringa Oil Expeller 
 The cost of fabrication materials, bought-out components and others for the moringa oil 
expeller is shown in Table 4.8. At an exchange rate of N160 equivalent to $1, the most expensive 
component was the electric gear motor which costs N40,000 ($250), while the least expensive 
components were bolts and nuts N60 ($0.38). The labour cost was N20,000 ($125) for the 
fabrication and electrical connection of the machine. Overall, the total cost of the oil expeller 
was estimated to be N160,540 ($1003.38). Since solvent extraction method is capital intensive, 
this expeller serves as a suitable option for small to medium scale moringa farmers and oil 
processors for efficient extraction of oil from moringa seeds. The maintenance of the expeller is 
cheap, as it involves cleaning of the expression chamber daily after use to prevent the hardening 
of the cake stuck to the wormshaft. Apart from the electric motor, the various parts are relatively 
cheap and easily affordable in case of failure of any part of the machine.   
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
103 
 
 
Table 4.7. Material Balance in terms of Oil and Cake Output  
S/No MC HT Ht AP Oil Output Cake Output (g) Total Output (g) Efficiency (%) 
1 10.00 70.00 25.00 15.00 243.4 703.8 947.2 94.72 
2 11.00 60.00 20.00 20.00 232.6 694.9 927.5 92.75 
3 11.00 80.00 30.00 10.00 253.3 677.0 930.3 93.03 
4 9.00 80.00 30.00 10.00 188.8 755.7 944.5 94.45 
5 10.00 70.00 35.00 15.00 182.3 756.9 939.2 93.92 
6 10.00 70.00 25.00 15.00 240.9 693.6 934.5 93.45 
7 11.00 80.00 20.00 10.00 261.6 677.7 939.3 93.93 
8 12.00 70.00 25.00 15.00 172.7 766.2 938.9 93.89 
9 11.00 80.00 20.00 20.00 277.9 652.6 930.5 93.05 
10 9.00 60.00 20.00 20.00 163.2 784.1 947.3 94.73 
11 10.00 70.00 25.00 25.00 236.6 705.8 942.4 94.24 
12 11.00 60.00 30.00 20.00 245.5 678.7 924.2 92.42 
13 9.00 80.00 30.00 20.00 206.8 725.9 932.7 93.27 
14 9.00 60.00 30.00 20.00 172.0 771.7 943.7 94.37 
15 8.00 70.00 25.00 15.00 136.8 794.9 931.7 93.17 
16 11.00 60.00 20.00 10.00 202.0 733.2 935.2 93.52 
17 10.00 70.00 25.00 15.00 240.8 696.6 937.4 93.74 
18 10.00 70.00 25.00 5.00 114.2 821.6 935.8 93.58 
19 11.00 80.00 30.00 20.00 285.8 661.8 947.6 94.76 
20 10.00 70.00 15.00 15.00 142.3 779.5 941.8 94.18 
21 10.00 70.00 25.00 15.00 242.0 712.7 954.7 95.47 
22 10.00 70.00 25.00 15.00 242.4 695.5 937.9 93.79 
23 11.00 60.00 30.00 10.00 193.6 746.4 940.0 94.00 
24 9.00 60.00 20.00 10.00 149.1 779.6 928.7 92.87 
25 10.00 70.00 25.00 15.00 243.0 690.3 933.3 93.33 
26 10.00 50.00 25.00 15.00 174.6 760.2 934.8 93.48 
27 9.00 80.00 20.00 20.00 170.0 776.0 946.0 94.60 
28 10.00 90.00 25.00 15.00 203.4 741.4 944.8 94.48 
29 9.00 80.00 20.00 10.00 161.0 768.1 929.1 92.91 
30 9.00 60.00 30.00 10.00 144.6 795.2 939.8 93.98 
    
 
 
 
 
 
 
104 
 
 
Table 4.8. Cost Analysis of the Moringa Oil Expeller  
    
  

    
 

    
  
 
    
  
  1200mm x 600mm x  
 5mm 
  500mm x 25mm,  
Ѳ 

    
 Ѳ 

Driven Pulleys    


    
 
  Ѳ  
 
    
  
    
  
 
    
 

    
  
  Ѳ  
 
  Ѳ  
 
    
  
105 
 
 

    
  

    
 
    
 
    
 
    


Transportation    

Labour -  N200 N20000 
00 
    
$1,00
3.38
 
Year 2014 
 
 
 
CHAPTER FIVE 
CONCLUSIONS AND RECOMMENDATIONS 
5.1 Conclusions 
 The engineering properties (physical and mechanical properties) of moringa seeds were 
established as design parameters for the development of an oil expeller for the crop. 
Subsequently, the moringa oil expeller was successfully designed and fabricated based on the 
determined physical and mechanical properties.      
 For the range of variables considered in the study, highest oil yield of 28.58% was 
obtained when moringa seeds were conditioned to a moisture content of 11% wet basis, heated at 
o
80 C for 30 mins at an applied pressure of 20 MPa. Predicted optimum oil yield of 28.2% at 
o
moisture content of 11.30% wet basis, temperature of 85.57 C, duration of 27.17 mins and 
pressure of 19.63 MPa was obtained. Under these optimal conditions, the experimental value was 
28.22% which was in agreement with those predicted by computation. Deviations between 
106 
 
 
experimental and predicted values ranged from 0.01-6.20. It was established that moisture 
content, applied pressure, heating temperature and duration influenced the quantity of oil 
recovery from moringa seeds using expeller. A model equation was generated with a satisfactory 
2 
coefficient of determination (R = 0.7691). The coefficient of determination showed excellent 
correlations between the independent variables. For the range of variables studied, the model 
chosen adequately predicted the yield for moringa oil expression.     
 Extraction and material balance efficiencies of 81.7% and 93.8% were obtained. High oil 
expression efficiency obtained makes the expeller a potential for moringa oil expression since 
the use of solvent is only economic on large scale and it is not affordable by most farmers. Also, 
the cake obtained thereafter can be dried and used as animal feed, thereby making the expeller a 
more viable option for oil expression.         
 The FFA, oil impurity and colour ranged from 2.42-7.40 mg/KOH/g, 2.12-3.20% and 
5.70-7.60 LU respectively and fell within acceptable limits. It was found out that moisture 
content, applied pressure, heating temperature and duration have significant effects on the quality 
of moringa oil expressed using expeller. Deviations between experimental and predicted values 
were low and ranged from 0.01-0.68, 0.01-0.12 and 0.01-0.12 for the FFA, colour and oil 
2
impurity respectively. The coefficients of determination (R ) for the FFA, oil impurity and 
colour were 0.94, 0.98 and 0.99 respectively.     
 Overall, the data generated in this study will serve as a useful tool in process and 
equipment design for moringa oil processors in developing countries, most especially Nigeria. 
This will help in improving the yield and quality attributes, thereby making more oil and fat 
available both for domestic and industrial purposes.       
5.2 Recommendations 
The following recommendations are made for further studies:  
(i) The effects of particle size on oil yield of moringa seeds should be further investigated. 
(ii) The effects of hulled and de-hulled moringa seeds on oil yield of moringa seeds and the 
 physico-chemical properties of the expressed moringa oil should be investigated.  
(iii) The capacity of the oil expeller should be increased so that larger oil quantity can be 
 produced per unit time.          
5.3 Contributions to Knowledge 
 From the study, the physical and mechanical properties of moringa seeds in relation to 
the design of oil expeller and other moringa processing machines have been investigated and 
established.            
107 
 
 
 It was also ascertained that moisture content, heating temperature, heating time and 
applied pressure affect the yield and quality of moringa oil during expression. These processing 
conditions have to be controlled during oil expression from moringa seeds in order to obtain high 
yield and good quality oil. It was found out that at low temperatures, moringa seeds require more 
time to break down the oil cells and adjust the moisture content; while at higher temperatures, 
these are achieved at a faster rate at short durations leading to increased oil yield. However, it 
was established that extending the heating duration at higher temperatures cause substantial 
moisture loss which leads to reduction in oil yield. This research therefore developed 
mathematical model which provided optimal conditions for the moisture content, heating 
temperature, heating time and applied pressure through optimisation of the various process 
parameters. This model would serve as useful tool for moringa oil processors. The methodology 
adopted in the study could therefore be successfully employed to any process where an analysis 
of the effects and interactions of many experimental factors are required.   
 It was shown that expeller serves as a viable option for moringa oil expression for small 
to medium scale moringa oil processors due to the high extraction efficiency obtained. The 
maintenance of the expeller is cheap, the component parts are relatively cheap, available and 
affordable and the parts can be easily replaced in case of failure.  
 
 
REFERENCES 
Abdulkarim S.M., Long, K., Lai, O.M. Muhammad, S.K.S. and Ghazali, H.M. 2005. Some 
 physico-chemical properties of Moringa oleifera seed oil extracted using solvent and 
 aqueous enzymatic methods. Food Chemistry 93, 253-263. 
 
Abidakun, O.A., Koya O.A and Ajayi O.O. 2012. Effect of expression conditions on the  yield 
 of Dika Nut (Irvingia Gabonesis) oil under uniaxial compression. Proc. ICCEM,  315-
 320. 
 
Adeeko, K.A. and Ajibola, O.O. 1990. Processing factors affecting yield and quality of 
 mechanically expressed Groundnut oil. Journal of Agricultural Engineering Research, 
 45(1): 31-43. 
 
Adejumo, B.A. and Abayomi, D.A. 2012. Effect of Moisture Content on some Physical 
 Properties of Moringa oleifera Seed. IOSR Journal of Agriculture and Veterinary 
 Science, vol.1, issue 5, pp. 12-21. 
 
108 
 
 
Adejumo, B.A., Alakowe, A.T and Obi, D.E. 2013. Effect of Heat Treatment on the 
 Characteristics and Oil Yield of Moringa Oleifera Seeds. The International Journal of 
 Engineering and Science (IJES), Vol. 2, Issue 1, Pp. 232-239.  
 
Adekola, K.A. 1991. Process Optimization of Oil Expression from Coconut (cocoa 
 nucefera L.). An M.Sc. thesis, Department of Agricultural Engineering, University 
 of Ibadan, Nigeria. 
 
Ahmadi, H., Mollazade, K., Khorshidi, J., Mohtasebi, S.S. and Rajabipour, A. 2009. Some 
 Physical and Mechanical Properties of Fennel Seed (Foeniculum vulgare). 
 Journal of Agricultural Science, vol. 1, No. 1, pp 66-75. 
 
Ajao, K.R.; Ajimotokan H.A.; Olaomi, J. and Akande, H.F. 2009. Development and 
 Performance Evaluation of a Groundnut oil Expelling Machine. 
 New York Science Journal, 2009, 2(6), 76-79. 
 
Ajav, E.A and Olatunde, O.B. 2011. Mechanical Oil Expression from Groundnut (Arachis 
th nd
 hypogaea). Proceedings of the 11  International Conference and 32  Annual General 
 Meeting of The Nigerian Institution of Agricultural Engineers. Vol 32: 427-430 held 
 between October 17-20 at Ilorin, Kwara State.  
 
Ajibola, O.O., Eniyemo, S.E., Fasina, O.O. and Adeeko, K.A. 1990. Mechanical  Expression of 
 Oil from Melon Seeds. Journal of Agricultural Engineering Research, 45: 45-53. 
 
Akinhanmi, T. F. and Akintokun, P. O. 2008. Chemical Composition and Physico-
 Chemical Properties of Cashew Nut (Anacardium occidentale) Oil and Cashew Nut 
 Shell Liquid. Journal of Agricultural, Food and Environmental Sciences, 2(1): 5, pp 
 75-78. 
 
Akinoso, R. 2006. Effects of Moisture Content, Roasting Duration and Temperature on  Yield 
 and Quality of Palm Kernel (Elaeis guineensis) and Sesame (sesamum indicum) oils. A 
 Ph.D Thesis, Department of Agricultural Engineering, University of Ibadan, Nigeria. 
Akinoso, R., Igbeka, J.C. and Olayanju, T.M.A. 2006. Process Optimization of Oil Expression 
 from Sesame Seed (Sesamum indicum Linn.). Agricultural Engineering International: 
 The CIGR Ejournal. Manuscript FP 06 011. Vol. VIII, pp 1-7. 
 
Aktas, T., I. Celen and Durgut, R. 2006. Some Physical and Mechanical Properties of 
 Safflower Seed (Carthamus tinctorius L.). Journal of Agronomy, 5(4): 613-616. 
 
Anazodo, U.G.N. 1983. Force-deformation Analysis for Biomaterials in Radial Compression: 
 Maximum Contact Stress. Nigerian J. Techn., 5 (2): 1-10. 
 
Anhwange, B.A., Ajibola, V.O. and Oniye, S.J. 2004. Chemical studies of the seeds of 
 Moringa oleifera (Lam) and Detarium microcarpum (Guill and Sperr).  J. Biol. Sci., 
 4: 711-715. 
 
Anjou, K. 1972. Manufacture of rapeseed oil and meal. In: Rapeseed: cultivation,  composition, 
 processing and utilization (edited by L. Appelqvist and R. Ohlson). Elsevier Publishing 
 Co., New York. 
 
109 
 
 
Anwar, F. and Bhanger, M.I. 2003. Analytical characterization of Moringa oleifera seed oil 
 grown in temperate regions of Pakistan. J. Agric. Food Chem., 51: 6558-6563. 
 
Anwar, F. and Rashid, U. 2007. Physico-Chemical Characteristics of Moringa oleifera Seeds 
 and Seed Oil from a Wild Provenance of Pakistan. Pak. J. Bot 39(5): 1443-1453.   
 
Anwar, F., Latif, S., Ashraf, M. and Gilani, A.H. 2007. Moringa oleifera: A food plant with 
 multiple medicinal uses. Phytother Res., 21 (1): 17-25. 
 
Anwar, F., Zafar, S.N. and Rashid, U. 2006. Characterization of Moringa oleifera seed oil 
 from drought and irrigated regions of Punjab, Pakistan. Grasas Y Aceites, 57 (2), 60-
 168. 
 
Arafa, G.K. 2007. Some Physical and Mechanical Properties of Sesame Seeds concerning 
 the selection of Separation Unit. Misr. J. Ag. Eng., 24(2): 415-429.  
 
Asoegwu, S., Ohanyere, S., Kanu, O. and Iwueke, C. 2006. Physical Properties of  African Oil 
  Bean Seed (Pentaclethra macrophylla)” Agricultural Engineering International: The 
 CIGR  Ejournal Manuscript FP 05 006. Vol. VIII. August, 2006. Pp 1-16. 
 
Aviara, N.A., Ehiabhi, S.E., Ajibola, O.O., Oni, S.A., Power, P.P., Abbas, T. and Onuh, O.A, 
2011. Effects of moisture content and temperature on the specific heat of soya bean, 
 Moringa oleifera seed and Mucuna flagellipes nut. Int J Agric & Biol Eng, vol 4, No 
 1, pp  87-92. 
 
Aviara, N.A., Power, P.P. and Abbas, T. 2012. Moisture-dependent physical properties of 
 Moringa oleifera seed relevant in bulk handling and mechanical processing. Industrial 
 Crops and Products, vol. 42, Pp. 96–104. 
 
Awolu, O.O., Obafaye, R.O. and Ayodele, B.S. 2013. Optimization of Solvent Extraction 
 of Oil from Neem (Azadirachta indica) and its Characterizations. Journal of Scientific 
 Research & Reports 2(1): 304-314. 
Ayotunde,  E.O., Fagbenro, O.A. and Adebanyo, O.T. 2011. Toxicity of aqueous extract  of 
 Moringa seed powder to Nile tilapia (Oreochromis niloticus) fingerlings, International 
 Research Journal of Agricultural Science, vol. 1, no. 4, p. 142-150. 
 
Balke, D. T. 2006. The production of higher value food ingredients from white mustard via 
 aqueous extraction. Ph.D. Thesis. Department of Chernical Engineering and Applied 
 Chemistry, University of Toronto, Canada.  
 
Bamgboye, A.I and Adejumo, O.I. 2011a. Effects of processing parameters of Roselle seed on its 
 oil yield. International Journal of Agricultural and Biological Engineering, Vol. 4, No.1, 
 82-86. 
 
Bamgboye, A.I. and Adejumo, O.I. 2011b. Mechanical properties of Roselle (Hibiscus 
 sabdariffa L) seeds. Agricultural Science Research Journals Vol. 1(8) pp. 178-183. 
 
Bamgboye, A.I. and Adebayo, S.E. 2012. Seed moisture dependent on Physical and 
 Mechanical properties of Jatropha curcas. Journal of Agricultural Technology, Vol. 
 8(1): 13-26.   
110 
 
 
 
Bargale, P.C. 1997. Mechanical oil expression from selected oilseeds under uniaxial 
 compression. Ph.D. thesis, University of Saskatchewan, Canada. 
 
Blake, J.A. 1982. Evaluation of an On-farm Press. In: Vegetable oil fuels. ASAE  Publication 
 No. 482, St. Joseph, MI 49055, 247 –251. 
 
Bongiriwar, D.R., Pordwal-Desai, S.R. and Sreenivasan A. 1977. Studies on the defatting of 
 Peanuts and Soybeans for Ready–to-eat snack items. Indian Food Packer,  31(3): 61 –76. 
 
Box, G.E.P. and Wilson, K.B. 1951. On the experimental attainment of optimum  conditions. J.  
 R. Statist. Soc., 13, 1–38. 
 
Box, G.E.P. and Norman, R.D. 1987. Empirical Model-Building and Response Surfaces;  Wiley: 
 Hoboken, NJ, USA; p. 424.   
 
Box, G.E.P., Hunter, W.G. and Hunter, J.S. 1978. Statistics for Experiments. John Wiley  and 
 Sons Inc. New York, pp. 335-375. 
 
Brett, C.J. 2005. Clinical  Perspectives  on  the  Health  Effects  of  Moringa  oleifera: A 
 Promising Adjunct  for  Balanced  Nutrition  and  Better Health. KOS Health 
 Publications, pp 1 -5. 
 
Caviedes, J. 1996. Aqueous processing of rapeseed (canola). M.Sc. Thesis. Department of 
 Chernical Engineering and Applied Chemistry, University of Toronto, Canada. 
 
Cecoco 1988. Cecoco oil expelling equipment for small scale and rural industrial purposes 
 leaflet. NSAE Proc. 1992, vol. 16 pp 102-110. 
 
Chukwu, O. and Sunmonu, M.O. 2010. Determination of Selected Engineering Properties 
 of Cowpea (Vigna unguiculata) Related to Design of Processing Machines. 
 International Journal of Engineering and Technology, 2 (6): 373 – 377. 
Compaore, W.R., Nikirma, P.A., Bassolr, H.I.N., Savadogo, A., Mouecoucou, J., Hounhouigan, 
D.J. and Traore, S.A. 2011. Chemical Composition and Antioxidative Properties of Seeds of 
 Moringa oleifera and Pulpss of Parkia biglobosa and Adansonia digitata commonly used 
 in Food Fortification in Burkina Faso. Current Research Journal of Biological Sciences 
 3(1): 64-72. 
 
Corrêa, P.C., SchwanzdaSilva, F., Jaren, C. and Afonso, P.C. 2007. Physical and  mechanical 
 properties in rice processing. Journal of Food Engineering, 79, 137-142. 
 
Cuesta, C., Sanchez-Muniz, F. J. and Hernandez, I. 1991. Evaluation of non-polar methyl 
 esters by column and gas chromatography for the assessment of used frying oils. J. 
 Am. Chem.Soc. 1991, 68, 443-445. 
 
Cvjetko, M., Jokic, S., Lepojevic, Z., Vidovic, S., Maric, B. and Redovnikovic, I.R. 2012. 
 Optimization of the Supercritical CO 2 Extraction of Oil from Rapeseed Using 
 Response Surface Methodology. Food Technol. Biotechnol. 50 (2) 208-215. 
 
111 
 
 
Dahot, M.U. 1988. Vitamin contents of flowers and seeds of Moringa oleifera. Pak. J. 
 Biochem., 21: 21-24. 
 
Darvishi, H. 2012. Moisture-Dependent Physical and Mechanical Properties of White 
 Sesame Seed. American-Eurasian J. Agric. & Environ. Sci., 12(2): 198-203. 
 
Davies, R.M. 2010. Engineering Properties of Three Varieties of Melon Seeds as Potentials 
 for Development of Melon Processing Machines. Advance Journal of Food Science 
 and Technology 2(1): 63-66. 
 
Dedio, N. and Dorrell, D.G. 1997. Factors affecting the pressure extraction of oil from flax 
 seed. J. Am. Oil Chem. Soc.; 54 (9): 313-315. 
 
Delaveau P. and Boiteau, P. 1980. Plantes Medicinales et Phytotherapie; 14, 1, pp 29-33. 
 
Doerr, B. 2005. Moringa water treatment. Moringa- An Echo Technical Note. North Fort  Myers, 
 USA.   
 
Dogra, P.D., Singh, B. P. and Tandon, S. 1975. Current Science 44, 1, pp. 31. 
 
Dutta, S.K.; Nema V.K. and Bhardwaj R.K. 1988. Physical properties of grain. J. of 
 Agr.Eng. Res. 39, 259-268. 
 
Fahey, J.W 2005. Treatment of Malnutrition with Moringa oleifera. J Sci.; 13: 29-50.  
 
Farsaie, A. and Singh, M.S. 1983. On farm sunflower oil production. ASAE paper No. NAR 83-
 307, ASAE, St. Joseph, MI 49085. 
 
Fasina, O.O. and Ajibola, O.O. 1989. Mechanical Expression of Oil from Conophornut            
 (Tetracarpidium conophorum). Journal of Agricultural Engineering Research, 44: 
 275-287. 
 
Fasina, O.O. and Ajibola, O.O. 1990. Development of Equation for the yield of Oil 
 expressed from Conophornut. Journal of Agricultural Engineering Research, 46: 45-
 53. 
 
Foidl, N., Makkar, H.P.S. and Becker, K. 2001. Potential of Moringa oleifera in agriculture and 
 industry. Potential of Moringa products development, pp. 20. 
 
Folkard, G., Sutherland, J. and Shaw, R. 2004. Moringa oleifera. Water and Environmental 
 Health at London and Loughborough, Pp. 109-112. 
 
Folkard, J.P., Travis, V.E., J.P. Sutherland and Holmes, R.O.H. 1993. Innovative water and 
 wastewater treatment for developing countries. J. Ind. Wat. Res. Soc.; pp 29-32. 
  
Garnayak, D.K., R.C. Pradhan, S.N. Nalk and Bhatnagar, N. 2008. Moisture-dependent physical 
 properties of Jatropha seed. Indust. Crops Prod., 27:127-129. 
 
Gate 1979. Oil Presses: An Introduction. GATE/GTZ Publications, Eschborn, Germany.  Gobin, 
 A. M. L., Uguru, M. I. and Deckers, J. 2001. Castor. In Crop production in 
112 
 
 
 Tropical Africa. (R. H. Raemaebers  ed). Directorate General for International Co-
 operation (DGIC), Brussels.  Pp. 725-732. 
 
Ghasemi Varnamkhasti, M., Mobli, H., Jafari, A., Rafiee, S., Heidary Soltanabadi, M., and 
 Kheiralipour, K. 2007. Some Engineering Properties of Paddy (var. Sazandegi). 
 International Journal of Agricultural and Biology, 5: 763-766. 
 
Goja, A.M. 2013. Physico-chemical properties of oil produced from Moringa oleifera, 
 Jatropha curcas and Carthamus tinctorius L. seeds. International Journal of Advanced 
 Research, Volume 1, Issue 4, 181-187.  
 
Gumble, L. O. and Maina, C. 1990. Friction coefficient of cereal seeds on various  surfaces. 
 Agricultural Mechanization in Asia, Africa and Latin America, 21(4), 61-64. 
 
Gupta, R.K. and Das, S.K. 2000. Fracture resistance of Sunflower seed and kernel to 
 compressive loading. Journal of Food Engineering, vol.46, pp.1-8. 
 
Hall, A.S.; Holowenko, A.R. and Laughlin, H.G. 1980. Schaum‟s Outline Theory and 
 Problem of Machine Design. McGraw- Hill Book Company, London. 
 
Hammonds, T.W., Harris, R.V. and Head, S.W. 1991. The influence of moisture content  on the 
 extraction of oil from fresh grated Coconut. Tropical Science; 31: 73 –81. 
 
Hamzat, K.O. and Clarke, B. 1993. Prediction of oil yield from Groundnuts using the 
 concept of Quasi-equilibrium oil yield. Journal of Agricultural Engineering Research, 
 55: 79- 87. 
 
Hartwell, J.L. 1995. Plants used against Cancer: A Survey. Lloydia, pp. 30-34. 
 
Ibrahim, A. and Onwualu, A.P. 2005. Technologies for extraction of oil from oil-bearing 
 agricultural products: A review. Journal of Agricultural Engineering and Technology, 
 13: 58-70. 
Ibrahim, S.S., Ismail, M., Samuel, G., Kamel, E. and Azhari, T. 1974. Beniseed, a 
 Potential Oil Source. Agric. Res. Rev., 52: 47-50. 
 
Idouraine, A.,  Kohlhepp, E.A. and Weber, C.W. 1996. Nutrient Constituents from Eight  Lines 
 of Naked Seed Squash (Cucurbita pepo  L.). J. Agric. Food Chem. 44: 721–724. 
 
Islau, M., Marks, B. and Bakker-Arkema, F. 2002. Optimization of commercial ear-corn dryers. 
 Agricultural Engineering International: The CIGR Journal of Scientific Research and 
 Development. Manuscript FP 04 007. Vol VI.   
 
Jazie, A.A., Pramanik, H., Sinha, A.S.K. and Jazie, A.A. 2013. Egg Shell as eco-friendly 
 Catalyst for Transesterification of Rapeseed Oil: Optimization or Biodiesel Production. 
 Special Issue of International Journal of Sustainable Development and Green Economics 
 (IJSDGE), ISSN No.: 2315-4721, V-2, I-1, 2, pp 27-32. 
 
Kagwacie, O.C. and Anozie, N.A. 1995. Effect of Processing Conditions on Solvent 
 Extraction of oil from Rubber seeds. Journal of Agricultural Technology, 3(1): 31-40. 
 
113 
 
 
Kalogo, Y., Rosillon, F.,Hammes, F. and Verstraete, W. 2000. Effect of a water extract of 
 Moringa oleifera seeds on the hydrolytic microbial species diversity of a UASB reactor 
 treating domestic wastewater. Letters in Appl. Microbio., 31: 259-264. 
 
Karaj, S. and Müller, J. 2010. Determination of Physical, Mechanical and Chemical 
 Properties of Seeds and Kernels of Jatropha curcas L. Industrial Crops and Products. 
 Elseveir Journal, pp. 1-10. 
 
Kemper, T. G. 2005. Oil extraction. In F. Shahidi (Ed.) Bailey’s Industrial Oil and Fat 
 Product. Vol. 5, 6th edn., p. 1. John Wiley & Sons, New York. 
 
Khan, L.M. and Hanna, M.A. 1983. Expression of Oil from Oilseeds: A review. Journal  of 
 Agricultural Engineering Research, 54: 215-229. 
 
Khan, L.M. and Hanna, M.A. 1984. Expression of Soybean Oil. Transactions of the ASAE, 27 
 (91): 190-194. 
 
Khurmi, R.S. and Gupta, J.K. 2004. “Theory of Machines”, Eurasia Publishing house, New 
 Delhi, India. 
 
Lawson, H. 1995. In Food Oils and Fats: Technology, Utilization and Nutrition; Chapman 
 and Hall: New York. 
 
Lowell, J.F. 1999. Moringa oleifera: Natural nutrition for the tropics, Dakar, Senegal Church 
 World Service. 
 
Maduako, J.N. and Faborode, M.O. 1990.Some Physical Properties of Cocoa Pods in 
 relation to primary processing. Ife Journal of Technology, 2: 1-7. 
 
Manuwa, S.I. and Muhammad, H.A. 2010. Moisture Content and Compression Axis Effects on 
 Mechanical Properties of Shea Kernels. Journal of Food Technology, 8(3), 89-94. 
 
Maroyi, A. 2007. The Utilization of Moringa oleifera in Zimbabwe: A Sustainable 
 Livelihood Approach. J Bot Sci.  23:172-185.  
 
Metha, A. and Aggrawal, B. 2008. Antiashmatic activity of Moringa oleifera Lam. Indian 
 Journal of Pharmacology; 40:28‐31.  
 
Mohammed, A.S., Lai, O.M., Muhammad, S.K.S., Long, K. and Ghazali, H.M. 2003. 
 Moringa oleifera, Potentially a New Source of oleic acid-type for Malaysia, In 
 Investing in Innovation, Bioscience and Biotechnology, vol 3, pp 137-140. 
 
Mohsenin, N.N. 1970. Physical Properties of Plant and Animal Materials. 1st edition; 
 Gordon and Breach Science Publishers, New York. 
 
Mohsenin, N. N. 1980. Physical Properties of Plant and Animal Materials. 2nd edition; 
 Gordon and Breach Science Publishers, New York. 
 
Mohsenin, N.N. 1986. Physical properties of Plant and Animal Materials. 2nd edition; 
 Gordon & Breach Science Publishers: New York. 
114 
 
 
 
Montgomery, D.C 2005. Design and Analysis of Experiments: Response Surface Method  and 
 Designs; John Wiley and Sons, Inc.: Hoboken, NJ, USA. 
 
Mubeen, A. 1998. Machine Design. Khanna Publishers, 2-B Nath Market, Naisarak Delhi 
 – 11 00 06: 698-730. 
 
Mwthiga, G. and Moriasi, L.A. 2007. Study of yield characteristics during mechanical oil 
 extraction of preheated and ground soybeans. Journal of Applied Sciences Research, 
 3(10), 1146-1151. 
 
Narayana, S. K.V., King, P., Gopinadh, R. and Lakshmi, D.K.N. 2012. Application of 
 Artificial Neural Networks and Statistical Methods in Coconut Oil Processing. 
 International Journal of Advanced Computer and Mathematical Sciences, Vol 3, Issue 
 2, pp 209-214.  
 
National Research Council 2006. Moringa. Lost Crops of Africa: Volume II: Vegetables.  Lost 
 Crops of Africa. 2. National Academies Press. Pp. 23-40. 

Ngoddy, P.O. and Ihekoronye, A.I. 1985. Integrated Food Science and Technology for  the 
 Tropics. Macmillan Education Limited, London. Pp. 23-28. 
Nzikou, J.M., L. Matos, J.E. Moussounga, C.B. Ndangui, A. Kimbonguila, Th. Silou, M. Linder 
and Desobry, S. 2009. Characterization of Moringa oleifera Seed Oil Variety Congo-Brazzaville. 
 Journal of Food Technology. Volume 7, Issue 3, Pp. 59-65. 
Ogunsina, B.S., Indira, T.N., Bhatnagar, A.S., Radha, C., Sukumar, D. and Gopalakrishna, A.G.  
 2011. Quality characteristics and stability of Moringa oleifera seed oil of Indian origin. 
 Journal of Food Science and Technology, 5 (3): 503-510. 
Ohlson, J.S.R. 1976. Processing effects on oil quality. Journal of American Oil Chemists 
 Society, 53 (6): 299-301. 
Ojiako, E.N. and Okeke, C.C. 2013. Determination of antioxidant of Moringa oleifera seed oil 
 and its use in the production of a body cream. Asian Journal of Plant Science and 
 Research, 3(3):1-4. 
 
Olajide, J.O. 2000. Process Optimization and Modelling of Oil Expression from Groundnut and 
 Sheanut Kernels. A Ph.D thesis, Department of Agricultural Engineering, University of 
 Ibadan, Nigeria. 
 
Olaniyan, A.M. 2010. Effect of extraction conditions on the yield and quality of oil from  castor 
 bean. Journal of Cereals and Oilseeds.Vol. 1(2), pp. 24-33.  
 
Olaomi, J. 2008. Design and Construction of a Motorized Groundnut oil Expelling 
 Machine. B.Eng Thesis, Department of Mechanical Engineering, University of Ilorin, 
 Nigeria. 
 
Olayanju, T.M.A. 2002. Design, Fabrication and Evaluation of a Beniseed (Sesanum 
 indicum L.) Oil Expeller. A Ph.D thesis, Department of Agricultural Engineering, 
 University of Ibadan, Nigeria.  
115 
 
 
 
Oliver-Bever, B. 1986. Medicinal Plants in Tropical West Africa. Cambridge University Press, 
 Cambridge.  
 
Olson, M.E. and Carlquist, S. 2001. Nutritional Quality of Moringa oleifera leaves. Bot J 
 Linnean Soc.; 135(4): 315-348 
 
Onwualu, A. P., Akubuo, C.O. and Ahaneku, I.E. 2006. Fundamentals of  Engineering in  
 Agriculture. (1st ed.). Lagos: Immaculate Publications Ltd. 
 
Orhevba, B. A., Sunmonu, M. O. and Iwunze, H.I. 2013. Extraction and Characterization  of 
 Moringa oleifera Seed Oil. Research and Reviews: Journal of Food and Dairy 
 Technology.Vol. 1, Issue 1, Pp. 22-27.    
 
Ozumba I.C. and Obiakor, S.I. 2011. Fracture resistance of palm kernel seed to compressive 
 loading. Journal of Stored Products and Postharvest Research, Vol. 2(13), pp. 248-253. 
 
Pauwels, E.K. 2011. The Protective Effect of the Mediterranean diet. Focus on cancer and 
 cardiovascular risk. Med princ pract, 20(2): 103-111.  
 
Pearson, D. 1981. The Chemical Analysis of Food. Churchill; pp. 580-581. 
 
Peleg, M. and Bargly, E.B. 1982. Physical Properties of Foods. AVI Publishing Co. Inc.  West 
 Port Connecticut.  
 
Pominski, J., Pearce, H.M. and Spadero, J.J. 1970. Partially defatted Peanuts-factors 
 affecting oil removal during pressing. Food Technology, 1970; 24(6): 92-94. 
 
Practical Action 2008. Small Scale Oilseed Processsing. A Practical Action Technical Brief. 
 
Price, M.L. 2007. The Moringa Tree. ECHO Technical Notes, USA. Pp. 1-19. 
 
Raymond, H.M. and Douglas, M.C. 2002. Response Surface Methodology Process and 
 Product Optimization Using Designed Experiments, 2nd ed.; John Wiley & Sons,  Inc.: 
 Hoboken, NJ, USA.  
 
Razavi, S.M.A., S. Yeganehzad and Sadeghi, A. 2009. Moisture Dependent Physical 
 Properties of Canola Seeds. J. Agric. Sci. Technol.,Vol. 11: 309-322. 
 
Reddy, T.S. and Bohle, N.G. 1993. Mechnical expression of oil from Mustard Seeds. Journal of 
 Agricultural Mechanization, vol. 24, No 3, pp 42-46.  
 
Rosenthal, A., Pyle, D. L. and Niranjan, K. 1996. Aqueous and enzymatic processes for edible 
 oil extraction. Enzyme Microbial Technology, 19: 402–420. 
 
Santos A.F., Argolo A.C., Coelho L.C. and Paiva, P.M. 2005. Detection of water soluble  lectin 
 and antioxidant component from Moringa oleifera seeds. Water Res. 39(6):975-980. 
 
Seifi, M.R. and Alimardani, R. 2010. Moisture-Dependent Physical Properties of  Sunflower 
 Seed (SHF8190). Journal of Modern Applied Science, Vol. 4, No. 7, 135- 143.  
116 
 
 
 
Seifi, M., R., Alimardani, R., Akram, A. and Asakereh, A. 2010. Moisture-Dependent 
 Physical Properties of Safflower (Goldasht). Advance Journal of Food Science and 
 Technology, 2(6): 340-345. 
 
Shafiee, S., Hashemi, J. and Kheiralipul, K. 2009. Moisture-Dependent Physical Properties of 
 Castorbean Seed. World Applied Sciences Journal, 7(10): 1216-1221. 
 
Shridhar, B.S., Beena, K.V., Anita, M.V. and Paramjeet, K.B. 2010. Optimization and 
 Characterization of Castor Seed Oil, Leonardo Journal of Sciences, Issue 17, p. 59-70.  
 
Siddhuraju, P. and Becker, K. 2003. Antioxidant properties of various solvent extracts of  total 
 phenolic constituents from three different agro-climatic origins of drumstick tree 
 (Moringa oleifera Lam.). J. Agri. Food Chem., 15, 2144-2155. 
 
Singh, M.S., Farsaie, A., Stewant, L.E. and Douglas, L.W. 1984. Development of  Mathematical 
 Models to predict sunflower oil expression. Transactions of American Society of 
 Agricultural Engineers, 27 (4): 1190-1194. 
 
Sitkei,  G. 1986.  Mechanics of  Agricultural Materials. Budapest: Akademiai Kiado. 
 
Sivakumaran K., Goodrum, W.J. and Bradley, A.R. 1985. Expeller Optimization of Peanut 
 Oil Production. Transactions of the American Society of Agricultural Engineers, 28 (1): 
 316-320. 
 
Sivala, K., Bhole, N.G. and Mukherjee, R.K. 1992. Effect of moisture on rice bran oil 
 expression. Journal of Agricultural Engineering Research, 50: 81-91. 
 
Southwell, K.H. and Harris, R.V. 1992. Extraction of oil from oilseeds using the hot water 
 flotation method. Tropical Science, 82: 251 –256. 
 
Southwell, K.H., Harris, R.V. and Swetman, A.A. 1990. Extraction and refining of oil 
 obtained from dried Avocado fruit using a small expeller. Tropical Science, 30: 121 -
 131. 
 
Shridhar, B.S, Beena, K.V., Anita, M.V. and Paramjeet, K.B. 2010. Optimization and 
 Characterization of Castor Seed Oil. Leonardo Journal of Sciences, Issue 17, p. 59-70.  
 
Srikanta-Rao, P.V. 1978. A search for appropriate technology for the village oil industry.  
 ATDA Publications, Lucknow, India.   
 
Stevenson, D.G., Eller, F. J., Wang, L., Jane, J. L., Wang, T. and Inglett, G.E. 2007. Oil and 
 Tocopherol Content and Composition of Pumpkin Seed Oil in 12 Cultivars. J. Agric. 
 Food Chem. 55: 4005-4013. 
 
Suganya, C.; Soottawat, B. and  Nisit S. 2000. Separation and Quality of Fish Oil from 
 Precooked and Non-Precooked Tuna Heads. Food Chemistry; 69:289–294. 
 
Takeoka, G.R., Gerhard H. F. and Dao, L.T. 1997. Effect of Heating on the Characteristics and 
 Chemical Composition of Selected Frying Oils and Fats. J. Agric.  Food Chem. 1997, 45, 
117 
 
 
 3244−3249. 
 
Tarighi, J., Mohtasebi, S.S. and Seyfi, M. 2011. Some Mechanical and Physical Properties of 
 Corn Seed (Var. DCC 370). African Journal of Agricultural Research, Vol. 6(16), pp, 
 3691-3699. 
 
Taser, O.F., Altuntas, E. and Ozgoz, E. 2005. Physical properties of Hungarian and 
 common Vetch seeds. J. Appl. Sci. 5, 323–326. 
 
Tavakoli H., Rajabipour A. and Mohtasebi, S.S. 2010. Moisture-Dependent Some 
 Engineering Properties of Soybean Grains. Agricultural Engineering International: 
 the CIGR Ejournal. Manuscript 1110. Vol. XI. February, 2009. 
 
Tunde-Akintunde, T.Y. 2000. Predictive models for evaluating the effective of some 
 processing parameters on yield and quality of some soybean products. A Ph.D thesis, 
 Department of Agricultural Engineering, University of Ibadan, Nigeria. 
 
Tunde-Akintunde, T. Y., Akintunde, B. O. and Igbeka, J.C. 2001. Effects of Processing  Factors 
 on Yield and Quality of Mechanically Expressed Soybean Oil. Journal of Engineering 
 Technology, 1(9): 39-45.    
 
Uzama, D., Thomas, S.A., Orishadipe, A.T., and Clement, O.A. 2011. The Development  of a  
 Blend of Moringa Oleifera Oil with Diesel for Diesel Engines. Journal of Emerging 
 Trends in Engineering and Applied Sciences (JETEAS) 2(6): 999-1001. 
 
Varma, R.K. 1998. Design of an Oil Expeller. In Processing and Storage of Oilseeds and 
 Products for Food Uses. Course Manual for Summer School, Central Institute of 
 Agricultural Engineering, Bhopal M.P. India. Pp. 378-388. 
 
Ward, J. A. 1976. Processing high oil content seeds in continuous screw press. Journal of 
 American Oil Chemist Society, 53: 261 –264. 
 
Ward, J. A. 1982. Pre-pressing of oil from rapeseed and sunflower. Journal of American Oil 
 Chemist’s Society. 61: 1358-1361. 
 
Weiss, E. A. 1983. Oilseed Crops. Longman Group Limited, London. 
 
Weiss, E.A. 2000. Oilseed crop. 2nd Edition Blackwell Longman Group Ltd. USA. 131- 164.   
 
Williams, S.W. and Rathod, K.L. 1983. A case study of expeller production of soybean  flour in 
 India. College of Agriculture, University of Illinois at Urbana-Champagne. Int. Agric. 
 Pub. INSTOY Series No. 3.  
 
 
 
 
118 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
APPENDIX A 
 
PHYSICAL PROPERTIES OF MORINGA SEEDS 
 
Table A1. Moisture Content of Moringa Seeds 
 
 Sample A Sample B Sample C 
Weight of Container (g) 24.4 23.6 24.1 
Weight of Container + Sample before drying (g) 74.6 73.7 74.2 
Weight of Container + Sample after drying (g) 70.6 70.0 70.4 
Loss in Weight (g) 3.5 3.7 3.8 
Moisture Content (% wet basis) 6.97 7.39 7.58 
 
6.97+7.39+7.58
Average Moisture Content of Moringa Seeds =  
3
 
Average Moisture Content of Moringa Seeds = 7.31% wet basis 
 
 
119 
 
 
Table A2. Linear Dimensions of Moringa Seeds 
 
2
S/NO L (mm) W (mm) T (mm) Da (mm) Dg (mm) S (mm ) Φ Ra 
1 9.31 8.23 6.95 8.163 8.105 206.37 0.871 0.884 
2 10.02 9.10 5.72 8.280 8.049 203.53 0.803 0.908 
3 9.29 8.59 7.82 8.567 8.646 229.44 0.931 0.925 
4 9.78 8.84 7.14 8.587 8.515 227.78 0.871 0.904 
5 9.28 8.32 7.06 8.220 8.169 209.65 0.880 0.897 
6 10.30 9.16 7.06 8.840 8.733 239.59 0.848 0.889 
7 9.14 8.90 7.51 8.517 8.485 226.18 0.928 0.974 
8 10.68 7.85 6.99 8.507 8.368 219.99 0.784 0.735 
9 8.84 7.68 6.61 7.710 7.656 184.14 0.866 0.869 
10 9.08 7.95 6.84 7.957 7.904 196.27 0.870 0.876 
11 7.95 7.55 6.63 7.377 7.355 169.95 0.925 0.950 
12 8.59 8.27 6.68 7.847 7.800 191.13 0.908 0.963 
13 9.19 7.79 7.49 8.157 8.124 207.34 0.884 0.848 
14 7.86 7.15 6.80 7.270 7.257 165.45 0.923 0.910 
15 8.22 7.72 7.11 7.683 7.670 184.82 0.933 0.939 
16 7.85 7.33 6.96 7.380 7.371 170.69 0.939 0.934 
17 8.30 8.16 7.69 8.050 8.046 203.38 0.969 0.983 
18 8.85 8.38 7.55 8.260 8.242 213.41 0.931 0.947 
19 8.55 8.43 7.75 8.243 8.236 213.10 0.963 0.986 
20 8.52 8.28 7.25 8.017 7.997 200.91 0.939 0.972 
21 9.36 7.23 6.88 7.823 7.751 188.74 0.828 0.772 
22 10.42 9.91 6.56 8.963 8.782 242.29 0.843 0.951 
23 9.21 8.19 7.95 8.450 8.433 223.42 0.916 0.889 
24 8.25 8.03 6.61 7.630 7.594 181.17 0.920 0.973 
25 8.46 8.07 7.42 7.983 7.972 199.66 0.942 0.954 
 
 
Table A2 (Cont’d) 
 
2 2 2
S/NO L (mm) W (mm) T (mm) Da (mm ) Dg (mm ) S (mm ) Ra Φ 
26 9.31 8.60 6.31 8.073 7.965 199.31 0.856 0.924 
27 8.47 7.62 7.36 7.817 7.803 191.28 0.921 0.900 
28 8.97 8.68 7.64 8.430 8.410 222.20 0.938 0.968 
29 8.39 8.02 7.22 7.877 7.861 194.14 0.937 0.956 
30 9.50 8.04 7.58 8.373 8.335 218.25 0.877 0.846 
31 9.51 8.67 6.91 8.363 8.290 215.90 0.872 0.912 
32 8.74 8.52 6.74 8.000 7.947 198.41 0.909 0.975 
33 9.65 9.23 6.65 8.510 8.398 221.57 0.870 0.956 
34 9.09 8.58 7.02 8.230 8.181 210.26 0.900 0.944 
35 8.45 8.25 7.66 8.120 8.113 206.78 0.960 0.976 
36 9.28 8.50 7.82 8.533 8.513 227.67 0.917 0.916 
37 9.26 8.96 6.84 8.353 8.279 215.33 0.894 0.968 
38 8.82 8.57 7.24 8.210 8.180 210.21 0.927 0.972 
39 8.49 8.08 7.30 7.957 7.941 198.11 0.935 0.952 
40 8.95 8.66 7.58 8.397 8.375 220.35 0.936 0.968 
41 8.68 8.27 7.37 8.107 8.088 205.51 0.932 0.953 
120 
 
 
42 8.86 8.66 6.59 8.037 7.967 199.41 0.899 0.977 
43 8.83 7.89 6.54 7.753 7.695 186.02 0.871 0.894 
44 9.82 8.81 6.34 8.323 8.186 210.52 0.834 0.897 
45 8.55 8.34 7.41 8.100 8.084 205.31 0.945 0.975 
46 9.67 8.81 6.39 8.290 8.165 209.44 0.844 0.911 
47 8.63 8.40 7.72 8.250 8.241 213.36 0.955 0.973 
48 8.12 7.58 6.95 7.550 7.535 178.37 0.928 0.933 
49 8.28 7.80 6.44 7.507 7.465 175.07 0.902 0.942 
50 8.96 8.63 7.19 8.260 8.223 212.43 0.918 0.963 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Table A2 (Cont’d) 
 
2 2 2
S/NO L (mm) W (mm) T (mm) Da (mm ) Dg (mm ) S (mm ) Ra Φ 
51 8.88 8.27 7.97 8.373 8.365 219.83 0.942 0.931 
52 9.72 8.73 6.59 8.347 8.239 213.25 0.848 0.898 
53 8.60 7.81 7.35 7.920 7.903 196.22 0.919 0.908 
54 9.24 7.28 6.99 7.837 7.776 189.96 0.842 0.788 
55 9.12 8.25 7.73 8.367 8.347 218.88 0.915 0.905 
56 8.51 8.23 7.13 7.957 7.934 197.76 0.932 0.967 
57 8.66 8.22 7.21 8.030 8.006 201.36 0.924 0.949 
58 9.03 8.84 7.30 8.390 8.353 219.20 0.925 0.979 
59 9.60 8.68 7.59 8.623 8.584 231.49 0.894 0.904 
60 10.17 9.33 6.55 8.683 8.534 228.80 0.839 0.917 
61 8.63 8.23 4.91 7.257 7.039 155.66 0.816 0.954 
62 7.39 6.91 5.60 6.633 6.588 136.35 0.891 0.935 
63 7.61 7.44 4.12 6.390 6.156 119.05 0.809 0.978 
64 7.00 6.71 6.07 6.593 6.582 136.10 0.940 0.957 
65 8.86 8.58 6.73 8.057 7.998 200.96 0.903 0.968 
121 
 
 
66 8.88 8.60 4.84 7.440 7.177 161.82 0.808 0.968 
67 7.68 6.83 5.46 6.657 6.592 136.52 0.858 0.889 
68 7.10 6.77 5.77 6.547 6.521 133.59 0.918 0.954 
69 10.15 9.69 4.48 8.107 7.609 181.89 0.750 0.955 
70 8.63 8.40 4.17 7.067 6.711 141.49 0.778 0.973 
71 6.58 6.06 5.82 6.153 6.145 118.63 0.934 0.921 
72 7.15 6.71 4.77 6.210 6.117 117.55 0.856 0.938 
73 7.64 7.07 4.98 6.563 6.455 130.90 0.845 0.925 
74 7.99 7.69 5.46 7.047 6.948 151.66 0.870 0.962 
75 7.46 7.04 6.95 7.150 7.147 160.47 0.958 0.944 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Table A2 (Cont’d) 
 
2 2 2
S/NO L (mm) W (mm) T (mm) Da (mm ) Dg (mm ) S (mm ) Ra Φ 
76 8.19 7.08 4.37 6.547 6.328 125.80 0.773 0.864 
77 6.90 6.56 4.94 6.133 6.070 115.75 0.880 0.951 
78 7.45 7.16 4.63 6.413 6.274 123.66 0.842 0.961 
79 7.29 6.62 5.40 6.437 6.388 128.20 0.876 0.908 
80 7.75 7.07 4.61 6.477 6.321 125.52 0.816 0.912 
81 7.47 7.00 5.31 6.593 6.524 133.71 0.873 0.937 
82 8.26 7.60 4.51 6.790 6.566 135.44 0.795 0.920 
83 7.10 6.08 5.36 6.180 6.139 118.40 0.865 0.856 
84 7.07 6.24 5.89 6.400 6.381 127.92 0.903 0.883 
85 7.24 6.97 6.40 6.870 6.861 147.89 0.948 0.963 
86 8.11 7.43 6.50 7.347 7.317 168.20 0.902 0.916 
87 6.44 6.05 5.73 6.073 6.066 115.60 0.942 0.939 
88 6.58 6.36 4.50 5.813 5.732 103.22 0.871 0.967 
89 7.07 6.62 4.11 5.933 5.773 104.70 0.817 0.936 
90 7.09 6.58 5.54 6.403 6.370 127.48 0.898 0.928 
122 
 
 
91 7.01 6.35 4.09 5.817 5.668 100.93 0.809 0.906 
92 6.75 6.27 5.97 6.330 6.322 125.56 0.937 0.929 
93 7.87 6.92 4.46 6.417 6.239 122.29 0.793 0.879 
94 7.64 5.67 4.92 6.077 5.973 112.08 0.782 0.742 
95 7.70 7.54 7.12 7.453 7.449 174.32 0.967 0.979 
96 7.07 6.88 6.27 6.740 6.731 142.33 0.952 0.973 
97 8.07 7.93 7.32 7.773 7.766 189.47 0.962 0.983 
98 6.98 6.53 4.52 6.010 5.906 109.58 0.846 0.936 
99 7.27 5.98 5.23 6.160 6.104 117.05 0.840 0.823 
100 9.49 8.78 7.10 8.457 8.395 221.41 0.885 0.925 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Table A3. True Density of Moringa Seeds 
 
3 -3
S/No Mass (g) Volume (cm ) True Density (gcm ) 
1 10.20 14.00 0.729 
2 10.00 12.00 0.833 
3 20.10 20.00 1.005 
4 20.20 20.00 1.010 
5 30.20 30.00 1.007 
6 30.00 28.00 1.071 
7 40.20 40.00 1.005 
8 40.10 40.00 1.003 
9 50.00 50.00 1.000 
10 50.10 48.00 1.044 
 
Table A4. Bulk Density of Moringa Seeds 
 
3 -3
S/No Mass (g) Volume (cm ) Bulk Density (gcm ) 
1 163.70 260.00 0.630 
123 
 
 
2 145.10 230.00 0.631 
3 173.60 275.00 0.631 
4 141.80 205.00 0.692 
5 142.60 207.00 0.689 
6 162.00 250.00 0.648 
7 157.10 228.00 0.689 
8 175.80 266.00 0.661 
9 138.70 203.00 0.683 
10 148.80 225.00 0.661 
 
 
Table A5. One Thousand Seed Weight of Moringa Seeds 
 
S/No Measured Value (g) W1000 
1 23.8 238.00 
2 24.0 240.00 
3 24.6 246.00 
4 23.4 234.00 
5 24.1 241.00 
6 23.7 237.00 
7 24.0 240.00 
8 23.8 238.00 
9 23.9 239.00 
10 23.9 239.00 
 
 
 
 
 
 
 
 
Table A6. Dynamic Angle of Repose of Moringa Seeds 
 
S/No h (mm) D (mm) θ 
1 3.5 17.5 21.8 
2 3.6 18.5 21.3 
3 4.0 20.7 21.1 
4 4.0 20.3 21.5 
5 3.4 18.0 20.7 
6 3.8 18.2 22.7 
7 3.6 19.0 20.8 
8 3.8 19.2 21.6 
9 3.6 19.3 20.4 
10 3.8 18.3 22.6 
 
 
 
Table A7. Static Coefficient of Friction of Moringa Seeds on different Surfaces 
 
124 
 
 
S/No Glass Stainless Mild Steel Galvanized Rubber Wood 
Steel Steel 
o o o o o o
 μ  μ  μ  μ  μ  μ 
1 50 1.192 48 1.111 52 1.280 45 1.000 67 2.356 58 1.600 
2 43 0.932 49 1.150 51 1.235 52 1.280 67 2.356 54 1.376 
3 45 1.000 49 1.150 58 1.600 45 1.000 62 1.881 53 1.327 
4 45 1.000 48 1.111 56 1.483 55 1.428 62 1.881 60 1.732 
5 47 1.072 46 1.036 53 1.327 46 1.036 63 1.963 55 1.428 
6 45 1.000 48 1.111 53 1.327 51 1.235 69 2.605 63 1.963 
7 44 0.966 47 1.072 54 1.376 51 1.235 67 2.356 62 1.881 
8 47 1.072 48 1.111 52 1.280 55 1.428 65 2.145 63 1.963 
9 45 1.000 49 1.150 55 1.428 55 1.428 68 2.475 55 1.428 
10 46 1.036 48 1.111 55 1.428 53 1.327 63 1.963 54 1.376 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
APPENDIX B 
 
MECHANICAL PROPERTIES OF MORINGA SEEDS 
 
Table B1. Mechanical Properties of Moringa Seeds at Peak 
 
S/No Force at Peak Deformation at Peak Stress at Peak Energy to Peak 
2
(N) (mm) (N/mm )  (N.m) 
1 53.800 5.0660 44.800 0.1268 
2 60.400 5.2500 50.300 0.1371 
3 57.900 5.1770 48.300 0.1043 
4 59.700 5.0790 49.800 0.1564 
5 66.900 5.1110 55.800 0.1662 
6 65.600 5.1270 54.700 0.1419 
7 53.700 5.0810 44.700 0.1317 
8 51.000 5.0730 42.500 0.1121 
9 58.600 4.8500 48.800 0.1347 
10 63.500 5.1340 52.900 0.1395 
125 
 
 
11 54.300 5.0980 45.300 0.1185 
12 59.300 5.0620 49.400 0.1708 
13 56. 300 5.3540 46.900 0.1348 
14 48. 900 5.0760 40.800 0.1056 
15 59. 700 5.0350 49.800 0.1309 
16 50.400 5.0950 42.000 0.1188 
17 67.000 5.0650 55.800 0.1393 
18 62.700 5.1330 52.300 0.1523 
19 63.100 5.0610 52.600 0.1264 
20 57.900 5.0610 48.300 0.1384 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Table B2. Mechanical Properties of Moringa Seeds at Break 
 
S/No Force at Break Deformation at Break Stress at Break Energy to Break 
2
(N) (mm) (N/mm )  (N.m) 
1 53.700 5.1910 44.800 0.1335 
2 60.400 5.2500 50.300 0.1371 
3 57.500 5.2270 47.900 0.1072 
4 59.300 5.0920 49.400 0.1572 
5 66.900 5.1110 55.800 0.1662 
6 65.600 5.1270 54.700 0.1419 
7 53.700 5.0810 44.800 0.1317 
8 51.000 5.0730 42.500 0.1121 
9 57.700 5.0650 48.100 0.1471 
10 63.500 5.1340 52.900 0.1395 
11 54.300 5.0980 45.300 0.1185 
12 59.300 5.0620 49.400 0.1708 
13 56.300 5.3540 46.900 0.1348 
14 48.900 5.0760 40.800 0.1056 
126 
 
 
15 59.300 5.0530 49.400 0.1319 
16 50.400 5.0950 42.000 0.1188 
17 67.000 5.0650 55.800 0.1393 
18 62.600 5.1510 52.200 0.1534 
19 63.100 5.0610 52.600 0.1264 
20 57.900 5.1150 48.300 0.1415 
 
 
Table B3. Mechanical Properties of Moringa Seeds at Yield 
 
S/No Force at Yield   Stress at Yield Energy to Yield Young Modulus 
2 2
(N) (N/mm )  (N.m) (N/mm ) 
1 37.800 31.500 0.0425 116.94 
2 40.000 33.300 0.0685 81.52 
3 33.600 28.000 0.0415 106.12 
4 46.600 38.800 0.0595 127.46 
5 32.800 27.300 0.0297 147.95 
6 39.000 32.500 0.0530 97.37 
7 46.700 38.900 0.0873 83.02 
8 41.000 34.100 0.0717 87.32 
9 43.700 36.400 0.0645 88.75 
10 43.000 35.800 0.0662 108.78 
11 42.500 35.400 0.0752 90.59 
12 54.300 45.200 0.1215 98.66 
13 25.600 21.300 0.0316 77.13 
14 19.800 16.500 0.0205 66.65 
15 42.400 35.300 0.0677 116.25 
16 39.600 33.000 0.0630 78.34 
17 34.400 28.600 0.0439 88.62 
18 37.000 30.800 0.0465 92.58 
19 40.200 33.500 0.0594 97.34 
20 39.700 33.100 0.0536 101.82 
APPENDIX C 
 
TOTAL OIL CONTENT OF MORINGA SEEDS BY SOXHLET METHOD 
 
Table C1. Total Oil Content of Moringa Seeds  
 
 Sample A Sample B Sample C 
Weight of filter paper (g) 0.54 0.52 0.52 
Weight of sample 10.00 10.00 10.00 
Weight of filter paper + sample before extraction (g) 10.54 10.52 10.52 
Weight of filter paper + sample after extraction (g) 6.81 6.89 6.82 
Loss in Weight (g) 3.73 3.63 3.70 
Oil Content (%) 35.4 34.5 35.2 
 
35.4+34.5+35.2
Average Oil Content of Moringa Seeds =  
3
 
127 
 
 
Average Oil Content of Moringa Seeds = 35.03%  
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
APPENDIX D 
 
AMOUNT OF WATER ADDED TO MORINGA SAMPLES 
 
Amount of water to be added to moringa samples to raise the moisture content from 7.31% to 8% 
100−7.31
(  - 1) x 1000 = 7.5 ml 
100−8.00
 
Amount of water to be added to moringa samples to raise the moisture content from 7.31% to 9% 
100−7.31
(  - 1) x 1000 = 18.6 ml 
100−9.00
 
Amount of water to be added to moringa samples to raise the moisture content from 7.31% to 
10% 
128 
 
 
100−7.31
(  - 1) x 1000 = 29.9 ml 
100−10.00
 
Amount of water to be added to moringa samples to raise the moisture content from 7.31% to 
11% 
100−7.31
(  - 1) x 1000 = 41.5 ml 
100−11.00
 
Amount of water to be added to moringa samples to raise the moisture content from 7.31% to 
12% 
100−7.31
(  - 1) x 1000 = 53.3 ml 
100−12.00
 
 
 
 
 
 
 
 
 
 
 
 
 
APPENDIX E 
 
OIL YIELD OF MORINGA AT DIFFERENT PROCESSING CONDITIONS 
 
Table E1. Yield of Moringa Oil at different Processing Conditions (Experiment 1) 
S/No MC HT Ht AP Oil Yield 
1 12.00 70.00 25.00 15.00 24.78
2 14.00 60.00 20.00 20.00 22.51
3 14.00 80.00 30.00 10.00 26.81
4 10.00 80.00 30.00 10.00 18.87
5 12.00 70.00 35.00 15.00 19.04
6 12.00 70.00 25.00 15.00 22.98
7 14.00 80.00 20.00 10.00 25.41
8 16.00 70.00 25.00 15.00 18.51
129 
 
 
9 14.00 80.00 20.00 20.00 27.73
10 10.00 60.00 20.00 20.00 16.97
11 12.00 70.00 25.00 25.00 23.82
12 14.00 60.00 30.00 20.00 23.23
13 10.00 80.00 30.00 20.00 21.63
14 10.00 60.00 30.00 20.00 17.01
15 8.00 70.00 25.00 15.00 13.82
16 14.00 60.00 20.00 10.00 20.55
17 12.00 70.00 25.00 15.00 24.84
18 12.00 70.00 25.00 5.00 10.84
19 14.00 80.00 30.00 20.00 29.10
20 12.00 70.00 15.00 15.00 13.33
21 12.00 70.00 25.00 15.00 25.18
22 12.00 70.00 25.00 15.00 24.73
23 14.00 60.00 30.00 10.00 21.25
24 10.00 60.00 20.00 10.00 15.88
25 12.00 70.00 25.00 15.00 23.24
26 12.00 50.00 25.00 15.00 18.85
27 10.00 80.00 20.00 20.00 16.77
28 12.00 90.00 25.00 15.00 20.65
29 10.00 80.00 20.00 10.00 16.76
30 10.00 60.00 30.00 10.00 15.41




Table E2. Yield of Moringa Oil at different Processing Conditions (Experiment 2) 
S/No MC HT Ht AP Oil Yield 
1 12.00 70.00 25.00 15.00 23.73
2 14.00 60.00 20.00 20.00 23.82
3 14.00 80.00 30.00 10.00 25.11
4 10.00 80.00 30.00 10.00 18.41
5 12.00 70.00 35.00 15.00 17.73
6 12.00 70.00 25.00 15.00 25.45
7 14.00 80.00 20.00 10.00 25.86
8 16.00 70.00 25.00 15.00 16.33
9 14.00 80.00 20.00 20.00 28.21
10 10.00 60.00 20.00 20.00 15.96 
11 12.00 70.00 25.00 25.00 24.77
12 14.00 60.00 30.00 20.00 25.05
13 10.00 80.00 30.00 20.00 20.45
130 
 
 
14 10.00 60.00 30.00 20.00 16.76
15 8.00 70.00 25.00 15.00 13.45
16 14.00 60.00 20.00 10.00 18.20
17 12.00 70.00 25.00 15.00 23.63
18 12.00 70.00 25.00 5.00 12.63
19 14.00 80.00 30.00 20.00 27.86
20 12.00 70.00 15.00 15.00 14.97
21 12.00 70.00 25.00 15.00 22.85
22 12.00 70.00 25.00 15.00 24.89
23 14.00 60.00 30.00 10.00 19.44
24 10.00 60.00 20.00 10.00 14.73
25 12.00 70.00 25.00 15.00 25.10
26 12.00 50.00 25.00 15.00 16.97
27 10.00 80.00 20.00 20.00 16.82
28 12.00 90.00 25.00 15.00 20.97
29 10.00 80.00 20.00 10.00 15.89
30 10.00 60.00 30.00 10.00 13.77
 
 
 
 
 
 
 
Table E3. Yield of Moringa Oil at different Processing Conditions (Experiment 3) 
S/No MC HT Ht AP Oil Yield 
1 12.00 70.00 25.00 15.00 24.51
2 14.00 60.00 20.00 20.00 23.44
3 14.00 80.00 30.00 10.00 24.73
4 10.00 80.00 30.00 10.00 19.37
5 12.00 70.00 35.00 15.00 17.93
6 12.00 70.00 25.00 15.00 23.83
7 14.00 80.00 20.00 10.00 26.57
8 16.00 70.00 25.00 15.00 16.98
9 14.00 80.00 20.00 20.00 27.44
10 10.00 60.00 20.00 20.00 16.03
11 12.00 70.00 25.00 25.00 22.40
12 14.00 60.00 30.00 20.00 25.37
131 
 
 
13 10.00 80.00 30.00 20.00 19.96
14 10.00 60.00 30.00 20.00 17.84
15 8.00 70.00 25.00 15.00 13.76
16 14.00 60.00 20.00 10.00 19.92
17 12.00 70.00 25.00 15.00 23.78
18 12.00 70.00 25.00 5.00 10.78
19 14.00 80.00 30.00 20.00 28.77
20 12.00 70.00 15.00 15.00 14.38
21 12.00 70.00 25.00 15.00 24.57
22 12.00 70.00 25.00 15.00 23.11
23 14.00 60.00 30.00 10.00 19.33
24 10.00 60.00 20.00 10.00 14.11
25 12.00 70.00 25.00 15.00 24.56
26 12.00 50.00 25.00 15.00 16.55
27 10.00 80.00 20.00 20.00 17.41
28 12.00 90.00 25.00 15.00 19.41
29 10.00 80.00 20.00 10.00 15.65
30 10.00 60.00 30.00 10.00 14.20

 
 
 
 
 
 
APPENDIX F 
 
MATERIAL BALANCE IN TERMS OF OIL AND CAKE OUTPUT 
 
Table F1. Material Balance in terms of Oil and Cake Output (Experiment 1) 
S/No MC HT Ht AP Oil Output Cake Output (g) Total Output (g) Efficiency (%) 
1 10.00 70.00 25.00 15.00 247.8 706.8 954.6 95.46 
2 11.00 60.00 20.00 20.00 225.1 691.4 916.5 91.65 
3 11.00 80.00 30.00 10.00 254.1 686.2 940.3 94.03 
4 9.00 80.00 30.00 10.00 188.7 751.4 940.1 94.01 
5 10.00 70.00 35.00 15.00 190.4 747.7 938.1 93.81 
6 10.00 70.00 25.00 15.00 229.8 697.8 927.6 92.76 
7 11.00 80.00 20.00 10.00 268.1 677.6 945.7 94.57 
8 12.00 70.00 25.00 15.00 185.1 753.6 938.7 93.87 
132 
 
 
9 11.00 80.00 20.00 20.00 277.3 646.4 923.7 92.37 
10 9.00 60.00 20.00 20.00 169.7 794.0 963.7 96.37 
11 10.00 70.00 25.00 25.00 238.2 687.6 925.8 92.58 
12 11.00 60.00 30.00 20.00 232.3 701.1 933.4 93.34 
13 9.00 80.00 30.00 20.00 216.3 706.0 922.3 92.23 
14 9.00 60.00 30.00 20.00 170.1 776.5 946.6 94.66 
15 8.00 70.00 25.00 15.00 138.2 801.2 939.4 93.94 
16 11.00 60.00 20.00 10.00 212.5 740.7 953.2 95.32 
17 10.00 70.00 25.00 15.00 248.4 696.0 944.4 94.44 
18 10.00 70.00 25.00 5.00 108.4 843.2 951.6 95.16 
19 11.00 80.00 30.00 20.00 291.0 675.3 966.3 96.63 
20 10.00 70.00 15.00 15.00 133.3 778.7 912.0 91.20 
21 10.00 70.00 25.00 15.00 251.8 698.8 950.6 95.06 
22 10.00 70.00 25.00 15.00 247.3 695.9 943.2 94.32 
23 11.00 60.00 30.00 10.00 205.5 716.6 922.1 92.21 
24 9.00 60.00 20.00 10.00 158.8 756.9 915.7 91.57 
25 10.00 70.00 25.00 15.00 232.4 696.5 928.9 92.89 
26 10.00 50.00 25.00 15.00 188.5 743.2 931.7 93.17 
27 9.00 80.00 20.00 20.00 167.7 799.6 967.3 96.73 
28 10.00 90.00 25.00 15.00 206.5 750.9 957.4 95.74 
29 9.00 80.00 20.00 10.00 167.6 765.8 933.4 93.34 
30 9.00 60.00 30.00 10.00 154.1 801.6 955.7 95.57 
 
 
 
Table F2. Material Balance in terms of Oil and Cake Output (Experiment 2) 
S/No MC HT Ht AP Oil Output Cake Output (g) Total Output (g) Efficiency (%) 
1 10.00 70.00 25.00 15.00 237.3 700.9 938.2 93.82 
2 11.00 60.00 20.00 20.00 238.2 704.8 943.0 94.30 
3 11.00 80.00 30.00 10.00 258.6 669.6 928.2 92.82 
4 9.00 80.00 30.00 10.00 184.1 760.9 945.0 94.50 
5 10.00 70.00 35.00 15.00 177.3 764.9 942.2 94.22 
6 10.00 70.00 25.00 15.00 254.5 664.0 918.5 91.85 
7 11.00 80.00 20.00 10.00 251.1 687.7 938.8 93.88 
8 12.00 70.00 25.00 15.00 163.3 771.3 934.6 93.46 
9 11.00 80.00 20.00 20.00 282.1 662.3 944.4 94.44 
10 9.00 60.00 20.00 20.00 159.6 763.5 923.1 92.31 
11 10.00 70.00 25.00 25.00 247.7 712.6 960.3 96.03 
12 11.00 60.00 30.00 20.00 250.5 668.6 919.1 91.91 
133 
 
 
13 9.00 80.00 30.00 20.00 204.5 746.7 951.2 95.12 
14 9.00 60.00 30.00 20.00 167.6 778.4 946.0 94.60 
15 8.00 70.00 25.00 15.00 134.5 786.2 920.7 92.07 
16 11.00 60.00 20.00 10.00 194.4 725.3 919.7 91.97 
17 10.00 70.00 25.00 15.00 236.3 692.2 928.5 92.85 
18 10.00 70.00 25.00 5.00 126.3 806.7 933.0 93.30 
19 11.00 80.00 30.00 20.00 278.6 662.0 940.6 94.06 
20 10.00 70.00 15.00 15.00 149.7 813.6 963.3 96.33 
21 10.00 70.00 25.00 15.00 228.5 737.6 966.1 96.61 
22 10.00 70.00 25.00 15.00 248.9 702.9 951.8 95.18 
23 11.00 60.00 30.00 10.00 182.0 786.1 968.1 96.81 
24 9.00 60.00 20.00 10.00 147.3 785.9 933.2 93.32 
25 10.00 70.00 25.00 15.00 251.0 674.3 925.3 92.53 
26 10.00 50.00 25.00 15.00 169.7 786.8 956.5 95.65 
27 9.00 80.00 20.00 20.00 168.2 759.1 927.3 92.73 
28 10.00 90.00 25.00 15.00 209.7 726.8 936.5 93.65 
29 9.00 80.00 20.00 10.00 158.9 763.3 922.2 92.22 
30 9.00 60.00 30.00 10.00 137.7 801.7 939.4 93.94 

 
 
 
 
 
Table F3. Material Balance in terms of Oil and Cake Output (Experiment 3) 
S/No MC HT Ht AP Oil Output Cake Output (g) Total Output (g) Efficiency (%) 
1 10.00 70.00 25.00 15.00 245.1 703.7 948.8 94.88 
2 11.00 60.00 20.00 20.00 234.4 688.4 922.8 92.28 
3 11.00 80.00 30.00 10.00 247.3 675.3 922.6 92.26 
4 9.00 80.00 30.00 10.00 193.7 754.7 948.4 94.84 
5 10.00 70.00 35.00 15.00 179.3 758.2 937.5 93.57 
6 10.00 70.00 25.00 15.00 238.3 719.0 957.3 95.73 
7 11.00 80.00 20.00 10.00 265.7 667.7 933.4 93.34 
8 12.00 70.00 25.00 15.00 169.8 773.6 943.4 94.34 
9 11.00 80.00 20.00 20.00 274.4 649.2 923.6 92.36 
10 9.00 60.00 20.00 20.00 160.3 794.8 955.1 95.51 
11 10.00 70.00 25.00 25.00 224.0 717.3 941.3 94.13 
12 11.00 60.00 30.00 20.00 253.7 666.4 920.1 92.01 
13 9.00 80.00 30.00 20.00 199.6 725.0 924.6 92.46 
134 
 
 
14 9.00 60.00 30.00 20.00 178.4 760.3 938.7 93.87 
15 8.00 70.00 25.00 15.00 137.6 797.4 935.0 93.50 
16 11.00 60.00 20.00 10.00 199.2 733.5 932.7 93.27 
17 10.00 70.00 25.00 15.00 237.8 701.7 939.5 93.95 
18 10.00 70.00 25.00 5.00 107.8 815.0 922.8 92.28 
19 11.00 80.00 30.00 20.00 287.7 648.1 935.8 93.58 
20 10.00 70.00 15.00 15.00 143.8 806.3 950.1 95.01 
21 10.00 70.00 25.00 15.00 245.7 701.6 947.3 94.73 
22 10.00 70.00 25.00 15.00 231.1 687.7 918.8 91.88 
23 11.00 60.00 30.00 10.00 193.3 736.4 929.7 92.97 
24 9.00 60.00 20.00 10.00 141.1 796.0 937.1 93.71 
25 10.00 70.00 25.00 15.00 245.6 700.0 945.6 94.56 
26 10.00 50.00 25.00 15.00 165.5 750.5 916.0 91.60 
27 9.00 80.00 20.00 20.00 174.1 769.4 943.5 94.35 
28 10.00 90.00 25.00 15.00 194.1 746.5 940.6 94.06 
29 9.00 80.00 20.00 10.00 156.5 775.2 931.7 93.17 
30 9.00 60.00 30.00 10.00 142.0 782.4 924.4 92.44 








APPENDIX G 
 
MODELLING OF OIL YIELD 
Design Summary 
 
Study Type  Response Surface  Experiments 30 
Initial Design  Central Composite  Blocks No Blocks 
Design Model  Quadratic 
 
 Response       Name       Units        Obs Minimum Maximum Trans Model 
 Y1                  Yield           %             30 11.42 28.58 None Quadratic 
 Factor        Name         Units Type            Low Actual     High Actual      Low Coded          High Coded 
 A                 Moisture Content     % wb Numeric          9.00                  11.00                    -1.000 1.000 
o
 B                 Heating Temp.         C Numeric          60.00                80.00                    -1.000 1.000 
 C                 Heating Time         mins Numeric          20.00                30.00                    -1.000 1.000 
 D                 Applied Pressure      MPa Numeric          10.00                20.00                    -1.000 1.000 
 
Design Matrix Evaluation for Response Surface Quadratic Model 
4 Factors: A, B, C, D  
135 
 
 
Degrees of Freedom for Evaluation 
Model          14 
Residuals    15 
Lack 0f Fit 10 
Pure Error  5 
Corr Total    29 
     Power at 5 % alpha level for effect of 
Term  StdErr** VIF Ri-Squared 1/2 Std. Dev. 1 Std. Dev. 2 Std. Dev. 
    A                        0.20                    1.00                    0.0000                20.9 %                63.0 %               99.5 % 
  B 0.20 1.00 0.0000 20.9 % 63.0 % 99.5 % 
  C 0.20 1.00 0.0000 20.9 % 63.0 % 99.5 % 
  D 0.20 1.00 0.0000 20.9 % 63.0 % 99.5 % 
  A2 0.19 1.05 0.0476 68.7 % 99.8 % 99.9 % 
  B2 0.19 1.05 0.0476 68.7 % 99.8 % 99.9 % 
  C2 0.19 1.05 0.0476 68.7 % 99.8 % 99.9 % 
  D2 0.19 1.05 0.0476 68.7 % 99.8 % 99.9 % 
  AB 0.25 1.00 0.0000 15.5 % 46.5 % 96.2 % 
  AC 0.25 1.00 0.0000 15.5 % 46.5 % 96.2 % 
  AD 0.25 1.00 0.0000 15.5 % 46.5 % 96.2 % 
  BC 0.25 1.00 0.0000 15.5 % 46.5 % 96.2 % 
  BD 0.25 1.00 0.0000 15.5 % 46.5 % 96.2 % 
  CD 0.25 1.00 0.0000 15.5 % 46.5 % 96.2 % 
 **Basis Std. Dev. = 1.0 
Measures Derived From the (X'X)-1 Matrix 
Std             Leverage                  Point Type 
1 0.5833  Fact 
2 0.5833  Fact 
3 0.5833  Fact 
4 0.5833  Fact 
5 0.5833  Fact 
6 0.5833  Fact 
7 0.5833  Fact 
8 0.5833  Fact 
9 0.5833  Fact 
10 0.5833  Fact 
11 0.5833  Fact 
12 0.5833  Fact 
13 0.5833  Fact 
14 0.5833  Fact 
15 0.5833  Fact 
136 
 
 
16 0.5833  Fact 
17 0.5833  Axial 
18 0.5833  Axial 
19 0.5833  Axial 
20 0.5833  Axial 
21 0.5833  Axial 
22 0.5833  Axial 
23 0.5833  Axial 
24 0.5833  Axial 
25 0.1667  Center 
26 0.1667  Center 
27 0.1667  Center 
28 0.1667  Center 
29 0.1667  Center 
30 0.1667  Center 
Average = 0.5000 
 
          Maximum Prediction Variance (at a design point) =  0.583 
          Average Prediction Variance  =  0.500 
          Condition Number of Coefficient Matrix = 1.667 
          G Efficiency (calculated from the design points) =   85.7 %  
          Scaled D-optimality Criterion =  1.388 
      
Correlation Matrix of Regression Coefficients 
   Intercept  A  B  C   D  A2  B2 
Intercept 1.000 
A  -0.000 1.000 
B  -0.000 -0.000 1.000 
C  -0.000 -0.000 -0.000 1.000 
D  -0.000 -0.000 -0.000 -0.000  1.000 
A2  -0.535 -0.000 -0.000 -0.000  -0.000 1.000 
B2  -0.535 -0.000 -0.000 -0.000  -0.000 0.143 1.000        
C2  -0.535 -0.000 -0.000 -0.000  -0.000 0.143 0.143 
D2  -0.535 -0.000 -0.000 -0.000  -0.000 0.143 0.143 
AB  -0.000 -0.000 -0.000 -0.000  -0.000 -0.000 -0.000 
AC  -0.000 -0.000 -0.000 -0.000  -0.000 -0.000 -0.000 
AD  -0.000 -0.000 -0.000 -0.000  -0.000 -0.000 -0.000 
BC  -0.000 -0.000 -0.000 -0.000  -0.000 -0.000 -0.000 
BD  -0.000 -0.000 -0.000 -0.000  -0.000 -0.000 -0.000 
137 
 
 
CD  -0.000 -0.000 -0.000 -0.000  -0.000 -0.000 -0.000 
 
C2                   D2                AB                  AC                   AD                    BC               BD                      CD 
C2 1.000        
D2 0.143              1.000 
AB -0.000             -0.000              1.000 
AC -0.000             -0.000              -0.000         1.000 
AD -0.000             -0.000              -0.000         -0.000      1.000 
BC -0.000             -0.000              -0.000         -0.000      -0.000 1.000 
BD -0.000             -0.000              -0.000         -0.000      -0.000 -0.000 1.000 
CD -0.000             -0.000              -0.000         -0.000      -0.000 -0.000                 -0.000             1.000 
  
 
 
 
 
 
 
 
 
 
 
Correlation Matrix of Factors [Pearson's r] 
 
  A  B  C  D   A2  B2  C2 
A  1.000 
B  0.000 1.000 
C  0.000 0.000 1.000 
D  0.000 0.000 0.000 1.000 
A2  0.000 0.000 0.000 0.000  1.000 
B2  0.000 0.000 0.000 0.000  -0.111 1.000 
C2  0.000 0.000 0.000 0.000  -0.111 -0.111 1.000 
D2  0.000 0.000 0.000 0.000  -0.111 -0.111 -0.111 
AB  0.000 0.000 0.000 0.000  0.000 0.000 0.000 
AC  0.000 0.000 0.000 0.000  0.000 0.000 0.000 
AD  0.000 0.000 0.000 0.000  0.000 0.000 0.000 
BC  0.000 0.000 0.000 0.000  0.000 0.000 0.000 
BD  0.000 0.000 0.000 0.000  0.000 0.000 0.000 
CD  0.000 0.000 0.000 0.000  0.000 0.000 0.000 
138 
 
 
 
  D2  AB  AC  AD  BC  BD  CD 
D2  1.000 
AB  0.000 1.000 
AC  0.000 0.000 1.000 
AD  0.000 0.000 0.000 1.000 
BC  0.000 0.000 0.000 0.000  1.000 
BD  0.000 0.000 0.000 0.000  0.000 1.000 
CD  0.000 0.000 0.000 0.000  0.000 0.000 1.000 
 
 
 
 
 
 
 
 
 
 
 
 
 
Sequential Model Sum of Squares 
                   Sum of                     Mean                         
Source          Squares                         DF                   Square                    F Value               Prob > F 
Mean 12497.86 1 12497.86 
Linear  337.16 4 84.29 7.58 0.0004 Suggested 
2FI  14.06 6 2.34 0.17 0.9821 
Quadratic 121.80 4 30.45 3.22 0.0429 Suggested 
Cubic  71.36 8 8.92 0.88 0.5712 Aliased 
Residual 70.63 7 10.09 
Total  13112.86 30 437.10 
   
Lack of Fit Test      
                     Sum of                       Mean                 
Source  Squares DF                     Square F Value Prob > F 
Linear  277.78 20 13.89 1212.33 < 0.0001 Suggested 
2FI  263.73 14 18.84 1644.27 < 0.0001 
Quadratic 141.93 10 14.19 1238.86 < 0.0001 Suggested 
Cubic  70.57 2 35.28 3079.77 < 0.0001 Aliased 
139 
 
 
Pure Error 0.057 5 0.011 
 
 Model Summary Statistics 
  Std.  Adjusted Predicted 
 Source Dev. R-Squared R-Squared R-Squared PRESS 
 Linear 3.33 0.5482 0.4759 0.3634 391.50 Suggested 
    2FI  3.73 0.5711 0.3453 0.2246 476.85 
   Quadratic 3.08 0.7691 0.5536 -0.3294 817.61 Suggested 
    Cubic  3.18 0.8852 0.5242 -15.5233 10161.83 Aliased 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ANOVA for Response Surface Quadratic Model 
 
Analysis of variance table [Partial sum of squares] 
 
                 Sum of                Mean    
Source        Squares                DF            Square             F Value      Prob > F 
 
Model         473.01                   14               33.79               3.57             0.0099  significant 
A                186.26 1                 186.26             19.68 0.0005 
B         54.06 1                 54.06          5.71 0.0304 
C 14.48 1                 14.48 1.53 0.2352 
D 82.36 1                 82.36 8.70 0.0099 
A2 68.22 1                 68.22 7.21 0.0170 
B2 14.25 1                 14.25 1.51 0.2387 
C2 52.87 1                 52.87 5.59 0.0320 
D2 30.87 1                 30.87 3.26 0.0911 
AB 7.18 1 7.18 0.76 0.3974 
AC 0.62 1 0.62                 0.065 0.8021 
AD 2.46 1 2.46 0.26 0.6173 
BC 1.90 1 1.90 0.20 0.6602 
140 
 
 
BD 1.45 1 1.45 0.15 0.7008 
CD 0.44 1 0.44                 0.046 0.8330 
Residual       141.99                  15 9.47 
Lack of Fit   141.93                  10               14.19               1238.86       < 0.0001                      significant 
Pure Error    0.057 5                 0.011 
Cor Total     615.00                   29 
Std. Dev.     3.08                     R-Squared     0.7691 
Mean           20.41  Adj R-Squared 0.5536 
C.V.            15.07  Pred R-Squared -0.3294 
PRESS        817.61                Adeq Precision 6.952 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Coefficient  Standard                                                                95% CI          95% CI 
Factor Estimate DF Error Low High VIF 
Intercept 24.21 1 1.26                 21.53 26.89 
A-A 2.79 1 0.63 1.45 4.12 1.00 
B-B 1.50 1 0.63 0.16 2.84 1.00 
C-C 0.78 1 0.63 -0.56 2.12 1.00 
D-D 1.85 1 0.63 0.51 3.19 1.00 
A2 -1.58 1 0.59 -2.83 -0.32 1.05 
B2 -0.72 1 0.59 -1.97 0.53 1.05 
C2 -1.39 1 0.59 -2.64 -0.14 1.05 
D2 -1.06 1 0.59 -2.31 0.19 1.05 
AB 0.67 1 0.77 -0.97 2.31 1.00 
AC -0.20 1 0.77 -1.84 1.44 1.00 
AD 0.39 1 0.77 -1.25 2.03 1.00 
BC 0.35 1 0.77 -1.29 1.98 1.00 
BD -0.30 1 0.77 -1.94 1.34 1.00 
CD 0.37 1 0.77 -1.47 1.80 1.00 
   
Linear Model for Oil Yield      
Final Equation in Terms of Coded Factors: 
Y = +20.41 + 2.79Mc + 1.50HT + 0.78Ht + 1.85Ap   
Final Equation in Terms of Actual Factors: 
141 
 
 
Y = -27.39433 + 2.78583Mc + 0.15008HT + 0.15533Ht + 0.37050Ap   
[Std. Dev. = 3.33, R-Squared = 0.5482, Mean = 20.41, Adj R-Squared = 0.4759 
C.V. =  16.33, Pred R-Squared = 0.3634, PRESS = 391.50, Adeq Precision = 10.163]. 
2FI Model for Oil Yield  
Final Equation in Terms of Coded Factors: 
Y = +20.41 + 2.79Mc + 1.50HT + 0.78Ht + 1.85Ap + 0.67McHT - 0.20McHt + 0.39McAp + 
0.35HTHt - 0.30HTAp + 0.17HtAp  
Final Equation in Terms of Actual Factors: 
Y = +29.69192 - 2.10042Mc - 0.60204HT - 0.034167Ht - 0.15775Ap + 0.067000McHT - 
0.039250McHt + 0.078500McAp + 0.006900HTHt – 0.006025HTAp + 0.006600HtAp  
[Std. Dev. =  3.73, R-Squared = 0.5711, Mean = 20.41, Adj R-Squared = 0.3453 
C.V. = 18.26, Pred R-Squared = 0.2246, PRESS = 476.85, Adeq Precision = 6.130]. 
Quadratic Model for Oil Yield   
Final Equation in Terms of Coded Factors 
2 2 2 2
Y = +24.21 + 2.79Mc + 1.50HT + 0.78Ht + 1.85Ap - 1.58Mc  - 0.72HT  - 1.39Ht  - 1.06Ap  + 
0.67McHT - 0.20McHt + 0.39McAp + 0.35HTHt - 0.30HTAp + 0.17HtAp 
 
Final Equation in Terms of Actual Factors 
2
Y = -203.79542 + 29.44125Mc + 0.40712HT + 2.74250Ht + 1.11525Ap – 1.57708 Mc  – 
2 2 2
0.007208HT  – 0.055533Ht  – 0.042433Ap  + 0.067000McHT – 0.039250McHt + 0.078500McAp 
+ 0.006900HTHt – 0.006025HTAp + 0.006600HtAp 
[Std. Dev. = 3.08, R-Squared = 0.7691, Mean = 20.41, Adj R-Squared = 0.5536, 
C.V. = 15.07, Pred R-Squared = -0.3294, PRESS = 817.61, Adeq Precision = 6.952]. 
Cubic Model for Oil Yield 
Final Equation in Terms of Coded Factors: 
2 2 2 2
Y = +24.21 + 4.67Mc + 2.28HT + 0.55Ht + 0.64Ap - 1.58Mc  - 0.72HT  - 1.39Ht  - 1.06Ap  + 
3 3
0.67McHT - 0.20McHt + 0.39McAp + 0.35HTHt - 0.30HTAp + 0.17HtAp - 0.94Mc  - 0.39HT  + 
3 3 
0.11Ht  + 0.60Ap - 0.41McHTHt - 0.12 McHTAp - 0.11McHtAp - 0.058HTHtAp 
Final Equation in Terms of Actual Factors: 
Not available for aliased models. 
[Std. Dev. = 3.18, R-Squared = 0.8852, Mean = 20.41, Adj R-Squared = 0.5242, 
C.V. = 15.56, Pred R-Squared = -15.5233, PRESS = 10161.83, Adeq Precision = 5.080].  
Where, 
Y = Oil yield (%) 
142 
 
 
Mc = Moisture content (%wb)  
o
HT = Heating temperature ( C) 
Ht = Heating time (mins) 
Ap = Applied pressure (MPa) 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Diagnostics Case Statistics 
 
Standard  Actual         Predicted         Student      Cook's Outlier                Run  
Order       Value           Value                  Residual      Leverage  Residual       Distance        t                    Order 
1 14.91 13.62 1.29 0.583  0.649 0.039 0.636 24 
2 19.36 17.46 1.90 0.583  0.957 0.085 0.954 16 
3 16.10 15.19 0.91 0.583  0.456 0.019 0.444 29 
4 25.33 21.71 3.62 0.583  1.821 0.309 1.993 7 
5 14.46 14.55 -0.086 0.583 -0.043 0.000 -0.042 30 
6 20.20 17.60 2.60 0.583  1.309 0.160 1.344 23 
7 18.88 17.50 1.38 0.583  0.695 0.045 0.682 4 
8 26.16 23.23 2.93 0.583  1.473 0.203 1.539 3 
9 16.32 16.81 -0.49 0.583 -0.248 0.006 -0.240 10 
10 23.26 22.22 1.04 0.583  0.523 0.025 0.510 2 
11 17.00 17.18 -0.18 0.583 -0.092 0.001 -0.089 27 
12 27.79 25.27 2.52 0.583  1.268 0.150 1.297 9 
13 17.20 18.40 -1.20 0.583 -0.604 0.034 -0.590 14 
14 24.55 23.02 1.53 0.583  0.769 0.055 0.758 12 
15 20.68 20.15 0.53 0.583  0.268 0.007 0.259 13 
16 28.58 27.45 1.13 0.583  0.568 0.030 0.555 19 
17 13.68 12.33 1.35 0.583  0.681 0.043 0.668 15 
18 17.27 23.47 -6.20 0.583 -3.123 0.910 -5.100 * 8 
19 17.46 18.32 -0.86 0.583 -0.435 0.018 -0.423 26 
20 20.34 24.33 -3.99 0.583 -2.007 0.376 -2.268 28 
21 14.23 17.10 -2.87 0.583 -1.446 0.195 -1.506 20 
22 18.23 20.21 -1.98 0.583 -0.996 0.093 -0.996 5 
23 11.42 16.26 -4.84 0.583 -2.437 0.554 -3.029 18 
24 23.66 23.67 -0.010 0.583 -0.005 0.000 -0.005 11 
25 24.30 24.21 0.092 0.167  0.033 0.000 0.032 25 
26 24.09 24.21 -0.12 0.167 -0.042 0.000 -0.041 6 
143 
 
 
27 24.34 24.21 0.13 0.167  0.047 0.000 0.045 1 
28 24.08 24.21 -0.13 0.167 -0.046 0.000 -0.044 17 
29 24.24 24.21 0.032 0.167  0.011 0.000 0.011 22 
30 24.20 24.21 -8.333E-003 0.167 -0.003 0.000 -0.003 21 
 
* Case(s) with |Outlier T| > 3.50   
  
 
 
  











 
 
 
 
APPENDIX H 
 
STATISTICAL ANALYSIS RESULTS 
 
Tests of Between-Subjects Effects 
Dependent Variable: Yield 
Source Type III Sum of Df Mean Square F Sig. 
Squares 
a
Corrected Model 614.945  24 25.623 2236.489 .000 
Intercept 4360.995 1 4360.995 380651.270 .000 
MC 311.447 3 103.816 9061.608 .000 
HT 101.561 4 25.390 2216.197 .000 
TH 110.887 3 36.962 3226.275 .000 
AP 163.243 3 54.414 4749.580 .000 
MC * HT 3.566 1 3.566 311.243 .000 
MC * TH .990 1 .990 86.415 .000 
MC * AP 1.346 1 1.346 117.451 .000 
HT * TH .717 1 .717 62.570 .001 
HT * AP 1.243 1 1.243 108.515 .000 
TH * AP .197 1 .197 17.155 .009 
MC * HT * TH 1.361 1 1.361 118.817 .000 
MC * HT * AP .123 1 .123 10.694 .022 
MC * TH * AP .110 1 .110 9.641 .027 
HT * TH * AP .030 1 .030 2.620 .166 
MC * HT * TH * 
.000 0 . . . 
AP 
144 
 
 
Error .057 5 .011   
Total 13112.862 30    
Corrected Total 615.002 29    
a. R Squared = 1.000 (Adjusted R Squared = .999) 
 
Estimated Marginal Means 
Grand Mean 
Dependent Variable: Yield 
Mean Std. Error 95% Confidence Interval 
Lower Bound Upper Bound 
a
19.651  .021 19.597 19.705 
 
 
 
 
 
 
 
Univariate Analysis of Variance 
Between-Subjects Factors 
 Value Label N 
8.00 Lowest 1 
9.00 Lower 8 
Moisture Cont 10.00 Medium 12 
11.00 Higher 8 
12.00 Highest 1 
50.00 Lowest 1 
60.00 Lower 7 
Heating Temp 70.00 Medium 13 
80.00 Higher 8 
90.00 Highest 1 
15.00 Lowest 1 
20.00 Lower 8 
Heating Time 25.00 Medium 12 
30.00 Higher 8 
35.00 Highest 1 
5.00 Lowest 1 
10.00 Lower 8 
App. Pressure 15.00 Medium 12 
20.00 Higher 8 
25.00 Highest 1 
 
145 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Tests of Between-Subjects Effects 
Dependent Variable: FFA 
Source Type III Sum of df Mean Square F Sig. 
Squares 
a
Corrected Model 77.550  24 3.231 137.911 .000 
Intercept 378.570 1 378.570 16157.475 .000 
MC 1.194 3 .398 16.986 .005 
HT 69.816 4 17.454 744.943 .000 
TH 3.438 3 1.146 48.912 .000 
AP .710 3 .237 10.094 .015 
MC * HT .002 1 .002 .086 .781 
MC * TH .008 1 .008 .333 .589 
MC * AP .002 1 .002 .086 .781 
HT * TH .250 1 .250 10.670 .022 
HT * AP .004 1 .004 .180 .689 
TH * AP 4.444E-005 1 4.444E-005 .002 .967 
MC * HT * TH .005 1 .005 .193 .679 
MC * HT * AP .000 1 .000 .005 .947 
MC * TH * AP .002 1 .002 .090 .776 
HT * TH * AP .005 1 .005 .193 .679 
MC * HT * TH * 
.000 0 . . . 
AP 
Error .117 5 .023   
Total 892.303 30    
Corrected Total 77.667 29    
146 
 
 
a. R Squared = .998 (Adjusted R Squared = .991) 
 
 
 
Estimated Marginal Means 
Grand Mean 
Dependent Variable: FFA 
Mean Std. Error 95% Confidence Interval 
Lower Bound Upper Bound 
a
5.156  .030 5.079 5.234 
 
 
 
 
 
 
 
 
Univariate Analysis of Variance 
Between-Subjects Factors 
 Value Label N 
8.00 Lowest 1 
9.00 Lower 8 
Moisture Cont 10.00 Medium 12 
11.00 Higher 8 
12.00 Highest 1 
50.00 Lowest 1 
60.00 Lower 7 
Heating Temp 70.00 Medium 13 
80.00 Higher 8 
90.00 Highest 1 
15.00 Lowest 1 
20.00 Lower 8 
Heating Time 25.00 Medium 12 
30.00 Higher 8 
35.00 Highest 1 
5.00 Lowest 1 
10.00 Lower 8 
App. Pressure 15.00 Medium 12 
20.00 Higher 8 
25.00 Highest 1 
 
147 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Tests of Between-Subjects Effects 
Dependent Variable: Colour 
Source Type III Sum of df Mean Square F Sig. 
Squares 
a
Corrected Model 6.755  24 .281 105.542 .000 
Intercept 597.183 1 597.183 223943.497 .000 
MC .102 3 .034 12.779 .009 
HT 6.145 4 1.536 576.080 .000 
TH .337 3 .112 42.090 .001 
AP .009 3 .003 1.121 .424 
MC * HT .000 1 .000 .104 .760 
MC * TH .000 1 .000 .104 .760 
MC * AP .000 1 .000 .104 .760 
HT * TH .014 1 .014 5.104 .073 
HT * AP .000 1 .000 .104 .760 
TH * AP .000 1 .000 .104 .760 
MC * HT * TH .005 1 .005 1.875 .229 
MC * HT * AP .001 1 .001 .469 .524 
MC * TH * AP .001 1 .001 .469 .524 
HT * TH * AP .005 1 .005 1.875 .229 
MC * HT * TH * 
.000 0 . . . 
AP 
Error .013 5 .003   
Total 1305.660 30    
Corrected Total 6.768 29    
148 
 
 
a. R Squared = .998 (Adjusted R Squared = .989) 
 
Estimated Marginal Means 
Grand Mean 
Dependent Variable: Colour 
Mean Std. Error 95% Confidence Interval 
Lower Bound Upper Bound 
a
6.583  .010 6.557 6.609 
 
 
 
 
 
 
 
 
 
 
 
Univariate Analysis of Variance 
Between-Subjects Factors 
 Value Label N 
8.00 Lowest 1 
9.00 Lower 8 
Moisture Cont 10.00 Medium 12 
11.00 Higher 8 
12.00 Highest 1 
50.00 Lowest 1 
60.00 Lower 7 
Heating Temp 70.00 Medium 13 
80.00 Higher 8 
90.00 Highest 1 
15.00 Lowest 1 
20.00 Lower 8 
Heating Time 25.00 Medium 12 
30.00 Higher 8 
35.00 Highest 1 
5.00 Lowest 1 
10.00 Lower 8 
App. Pressure 15.00 Medium 12 
20.00 Higher 8 
25.00 Highest 1 
 
149 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Tests of Between-Subjects Effects 
Dependent Variable: Impurity 
Source Type III Sum of df Mean Square F Sig. 
Squares 
a
Corrected Model 2.117  24 .088 367.465 .000 
Intercept 100.960 1 100.960 420665.954 .000 
MC .038 3 .013 52.369 .000 
HT 1.837 4 .459 1913.795 .000 
TH .137 3 .046 190.055 .000 
AP .008 3 .003 11.163 .012 
MC * HT 2.500E-005 1 2.500E-005 .104 .760 
MC * TH .002 1 .002 8.438 .034 
MC * AP 6.944E-005 1 6.944E-005 .289 .614 
HT * TH .005 1 .005 20.417 .006 
HT * AP .000 1 .000 1.157 .331 
TH * AP 2.778E-006 1 2.778E-006 .012 .919 
MC * HT * TH .001 1 .001 5.208 .071 
MC * HT * AP .000 1 .000 .833 .403 
MC * TH * AP .000 1 .000 1.875 .229 
HT * TH * AP 1.250E-005 1 1.250E-005 .052 .829 
MC * HT * TH * 
.000 0 . . . 
AP 
Error .001 5 .000   
Total 219.201 30    
Corrected Total 2.118 29    
150 
 
 
a. R Squared = .999 (Adjusted R Squared = .997) 
 
Estimated Marginal Means 
Grand Mean 
Dependent Variable: Impurity 
Mean Std. Error 95% Confidence Interval 
Lower Bound Upper Bound 
a
2.682  .003 2.674 2.690 
a. Based on modified population marginal mean. 
 
 
 
 
 
 
 
 
 
 
Between-Subjects Factors 
 Value Label N 
8.00 Lowest 1 
9.00 Lower 8 
Moisture Cont 10.00 Medium 12 
11.00 Higher 8 
12.00 Highest 1 
50.00 Lowest 1 
60.00 Lower 7 
Heating Temp 70.00 Medium 13 
80.00 Higher 8 
90.00 Highest 1 
15.00 Lowest 1 
20.00 Lower 8 
Heating Time 25.00 Medium 12 
30.00 Higher 8 
35.00 Highest 1 
5.00 Lowest 1 
10.00 Lower 8 
App. Pressure 15.00 Medium 12 
20.00 Higher 8 
25.00 Highest 1 
 
151 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
a
Multivariate Tests  
Effect Value F Hypothesis df Error df Sig. 
b
Pillai's Trace 1.000 634411.992  3.000 3.000 .000 
b
Wilks' Lambda .000 634411.992  3.000 3.000 .000 
Intercept b
Hotelling's Trace 634411.992 634411.992  3.000 3.000 .000 
b
Roy's Largest Root 634411.992 634411.992  3.000 3.000 .000 
Pillai's Trace 1.783 2.440 9.000 15.000 .061 
Wilks' Lambda .001 11.504 9.000 7.452 .002 
MC 
Hotelling's Trace 154.565 28.623 9.000 5.000 .001 
c
Roy's Largest Root 150.856 251.427  3.000 5.000 .000 
Pillai's Trace 2.269 3.878 12.000 15.000 .008 
Wilks' Lambda .000 54.150 12.000 8.229 .000 
HT 
Hotelling's Trace 3225.408 447.973 12.000 5.000 .000 
c
Roy's Largest Root 3202.698 4003.373  4.000 5.000 .000 
Pillai's Trace 2.166 4.330 9.000 15.000 .006 
Wilks' Lambda .000 30.751 9.000 7.452 .000 
TH 
Hotelling's Trace 212.549 39.361 9.000 5.000 .000 
c
Roy's Largest Root 182.729 304.548  3.000 5.000 .000 
Pillai's Trace 1.692 2.157 9.000 15.000 .090 
Wilks' Lambda .018 3.526 9.000 7.452 .050 
AP 
Hotelling's Trace 19.450 3.602 9.000 5.000 .086 
c
Roy's Largest Root 17.627 29.378  3.000 5.000 .001 
b
Pillai's Trace .269 .369  3.000 3.000 .783 
MC * HT b
Wilks' Lambda .731 .369  3.000 3.000 .783 
152 
 
 
b
Hotelling's Trace .369 .369  3.000 3.000 .783 
b
Roy's Largest Root .369 .369  3.000 3.000 .783 
b
Pillai's Trace .830 4.867  3.000 3.000 .113 
b
Wilks' Lambda .170 4.867  3.000 3.000 .113 
MC * TH b
Hotelling's Trace 4.867 4.867  3.000 3.000 .113 
b
Roy's Largest Root 4.867 4.867  3.000 3.000 .113 
b
Pillai's Trace .103 .115  3.000 3.000 .946 
b
Wilks' Lambda .897 .115  3.000 3.000 .946 
MC * AP b
Hotelling's Trace .115 .115  3.000 3.000 .946 
b
Roy's Largest Root .115 .115  3.000 3.000 .946 
b
Pillai's Trace .904 9.424  3.000 3.000 .049 
b
Wilks' Lambda .096 9.424  3.000 3.000 .049 
HT * TH b
Hotelling's Trace 9.424 9.424  3.000 3.000 .049 
b
Roy's Largest Root 9.424 9.424  3.000 3.000 .049 
b
Pillai's Trace .461 .854  3.000 3.000 .550 
b
Wilks' Lambda .539 .854  3.000 3.000 .550 
HT * AP b
Hotelling's Trace .854 .854  3.000 3.000 .550 
b
Roy's Largest Root .854 .854  3.000 3.000 .550 
b
Pillai's Trace .045 .047  3.000 3.000 .984 
b
Wilks' Lambda .955 .047  3.000 3.000 .984 
TH * AP b
Hotelling's Trace .047 .047  3.000 3.000 .984 
b
Roy's Largest Root .047 .047  3.000 3.000 .984 
b
Pillai's Trace .890 8.070  3.000 3.000 .060 
b
Wilks' Lambda .110 8.070  3.000 3.000 .060 
MC * HT * TH b
Hotelling's Trace 8.070 8.070  3.000 3.000 .060 
b
Roy's Largest Root 8.070 8.070  3.000 3.000 .060 
b
Pillai's Trace .577 1.365  3.000 3.000 .402 
b
Wilks' Lambda .423 1.365  3.000 3.000 .402 
MC * HT * AP b
Hotelling's Trace 1.365 1.365  3.000 3.000 .402 
b
Roy's Largest Root 1.365 1.365  3.000 3.000 .402 
b
Pillai's Trace .669 2.025  3.000 3.000 .289 
b
Wilks' Lambda .331 2.025  3.000 3.000 .289 
MC * TH * AP b
Hotelling's Trace 2.025 2.025  3.000 3.000 .289 
b
Roy's Largest Root 2.025 2.025  3.000 3.000 .289 
b
Pillai's Trace .589 1.432  3.000 3.000 .388 
b
Wilks' Lambda .411 1.432  3.000 3.000 .388 
HT * TH * AP b
Hotelling's Trace 1.432 1.432  3.000 3.000 .388 
b
Roy's Largest Root 1.432 1.432  3.000 3.000 .388 
b
Pillai's Trace .000 .  .000 .000 . 
b
MC * HT * TH Wilks' Lambda 1.000 .  .000 4.000 . 
* AP bHotelling's Trace .000 .  .000 2.000 . 
b
Roy's Largest Root .000 .000  3.000 2.000 1.000 
153 
 
 
a. Design: Intercept + MC + HT + TH + AP + MC * HT + MC * TH + MC * AP + HT * TH + HT * AP + 
TH * AP + MC * HT * TH + MC * HT * AP + MC * TH * AP + HT * TH * AP + MC * HT * TH * AP 
b. Exact statistic 
c. The statistic is an upper bound on F that yields a lower bound on the significance level. 
 
 
 
 
 
 
 
 
 
 
 
 
Tests of Between-Subjects Effects 
Source Dependent Variable Type III Sum of df Mean Square F Sig. 
Squares 
a
FFA 77.550  24 3.231 137.911 .000 
Corrected b
Colour 6.755  24 .281 105.542 .000 
Model 
c
Impurity 2.117  24 .088 367.465 .000 
FFA 378.570 1 378.570 16157.475 .000 
Intercept Colour 597.183 1 597.183 223943.497 .000 
Impurity 100.960 1 100.960 420665.954 .000 
FFA 1.194 3 .398 16.986 .005 
MC Colour .102 3 .034 12.779 .009 
Impurity .038 3 .013 52.369 .000 
FFA 69.816 4 17.454 744.943 .000 
HT Colour 6.145 4 1.536 576.080 .000 
Impurity 1.837 4 .459 1913.795 .000 
FFA 3.438 3 1.146 48.912 .000 
TH Colour .337 3 .112 42.090 .001 
Impurity .137 3 .046 190.055 .000 
FFA .710 3 .237 10.094 .015 
AP Colour .009 3 .003 1.121 .424 
Impurity .008 3 .003 11.163 .012 
FFA .002 1 .002 .086 .781 
MC * HT Colour .000 1 .000 .104 .760 
Impurity 2.500E-005 1 2.500E-005 .104 .760 
154 
 
 
FFA .008 1 .008 .333 .589 
MC * TH Colour .000 1 .000 .104 .760 
Impurity .002 1 .002 8.438 .034 
FFA .002 1 .002 .086 .781 
MC * AP Colour .000 1 .000 .104 .760 
Impurity 6.944E-005 1 6.944E-005 .289 .614 
FFA .250 1 .250 10.670 .022 
HT * TH Colour .014 1 .014 5.104 .073 
Impurity .005 1 .005 20.417 .006 
FFA .004 1 .004 .180 .689 
HT * AP Colour .000 1 .000 .104 .760 
Impurity .000 1 .000 1.157 .331 
FFA 4.444E-005 1 4.444E-005 .002 .967 
TH * AP Colour .000 1 .000 .104 .760 
Impurity 2.778E-006 1 2.778E-006 .012 .919 
FFA .005 1 .005 .193 .679 
MC * HT * 
Colour .005 1 .005 1.875 .229 
TH 
Impurity .001 1 .001 5.208 .071 
FFA .000 1 .000 .005 .947 
MC * HT * 
Colour .001 1 .001 .469 .524 
AP 
Impurity .000 1 .000 .833 .403 
FFA .002 1 .002 .090 .776 
MC * TH * 
Colour .001 1 .001 .469 .524 
AP 
Impurity .000 1 .000 1.875 .229 
FFA .005 1 .005 .193 .679 
HT * TH * AP Colour .005 1 .005 1.875 .229 
Impurity 1.250E-005 1 1.250E-005 .052 .829 
FFA .000 0 . . . 
MC * HT * 
Colour .000 0 . . . 
TH * AP 
Impurity .000 0 . . . 
FFA .117 5 .023   
Error Colour .013 5 .003   
Impurity .001 5 .000   
FFA 892.303 30    
Total Colour 1305.660 30    
Impurity 219.201 30    
FFA 77.667 29    
Corrected 
Colour 6.768 29    
Total 
Impurity 2.118 29    
a. R Squared = .998 (Adjusted R Squared = .991) 
b. R Squared = .998 (Adjusted R Squared = .989) 
c. R Squared = .999 (Adjusted R Squared = .997) 

155 
 
 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
APPENDIX I 
 
MODELLING OF FREE FATTY ACID 
 
Design Summary 
 
Study Type Response Surface  Experiments 30 
Initial Design Central Composite  Blocks No Blocks 
Design Model Linear 
 
 Response       Name       Units              Obs   Minimum Maximum Trans Model 
 Y1                   FFA        mg/KOH/g      30   2.42  7.40 None Linear 
 
 Factor        Name         Units Type            Low Actual     High Actual      Low Coded          High Coded 
 A                 Moisture Content     % wb Numeric          9.00                  11.00                    -1.000 1.000 
o
 B                 Heating Temp.         C Numeric          60.00                80.00                    -1.000 1.000 
 C                 Heating Time         mins Numeric          20.00                30.00                    -1.000 1.000 
 D                 Applied Pressure      MPa Numeric          10.00                20.00                    -1.000 1.000 
 
Design Matrix Evaluation for Response Surface Linear Model 
 
4 Factors: A, B, C, D  
Degrees of Freedom for Evaluation 
Model          4 
Residuals    25 
156 
 
 
Lack 0f Fit 20 
Pure Error  5 
Corr Total    29 
 
   Power at 5 % alpha level for effect of 
Term  StdErr** VIF Ri-Squared 1/2 Std. Dev. 1 Std. Dev. 2 Std. Dev. 
    A                        0.20                    1.00                    0.0000                21.8 %                65.3 %               99.7 % 
  B 0.20 1.00 0.0000 21.8 % 65.3 % 99.7 % 
  C 0.20 1.00 0.0000 21.8 % 65.3 % 99.7 % 
  D 0.20 1.00 0.0000 21.8 % 65.3 % 99.7 % 
 **Basis Std. Dev. = 1.0 
 
 
 
 
 
 
 
 
Measures Derived From the (X'X)-1 Matrix 
Std             Leverage                  Point Type 
1 0.2000  Fact 
2 0.2000  Fact 
3 0.2000  Fact 
4 0.2000  Fact 
5 0.2000  Fact 
6 0.2000  Fact 
7 0.2000  Fact 
8 0.2000  Fact 
9 0.2000  Fact 
10 0.2000  Fact 
11 0.2000  Fact 
12 0.2000  Fact 
13 0.2000  Fact 
14 0.2000  Fact 
15 0.2000  Fact 
16 0.2000  Fact 
17 0.2000  Axial 
18 0.2000  Axial 
19 0.2000  Axial 
20 0.2000  Axial 
21 0.2000  Axial 
157 
 
 
22 0.2000  Axial 
23 0.2000  Axial 
24 0.2000  Axial 
25 0.0333  Center 
26 0.0333  Center 
27 0.0333  Center 
28 0.0333  Center 
29 0.0333  Center 
30 0.0333  Center 
Average = 0.1667 
 
          Maximum Prediction Variance (at a design point) =  0.200 
          Average Prediction Variance  =  0.167 
          Condition Number of Coefficient Matrix = 1.000 
          G Efficiency (calculated from the design points) =   83.3 %  
          Scaled D-optimality Criterion =  1.195 
 
       
Correlation Matrix of Regression Coefficients 
   Intercept  A  B  C   D   
Intercept 1.000 
A  -0.000 1.000 
B  -0.000 -0.000 1.000 
C  -0.000 -0.000 -0.000 1.000 
D  -0.000 -0.000 -0.000 -0.000  1.000 
 
Correlation Matrix of Factors [Pearson's r] 
  A  B  C  D    
A  1.000 
B  0.000 1.000 
C  0.000 0.000 1.000 
D  0.000 0.000 0.000 1.000 
 
Sequential Model Sum of Squares 
                   Sum of                     Mean                         
Source          Squares                         DF                   Square                    F Value               Prob > F 
Mean 792.90 1 792.90 
Linear  67.63 4 16.91 59.71 < 0.0001 Suggested 
2FI  1.56 6 0.26 0.89 0.5202 
Quadratic 1.20 4 0.30 1.04 0.4213  
Cubic  3.48 8 0.44 3.61 0.539 Aliased 
158 
 
 
Residual 0.84 7 0.12 
Total  867.61 30 28.92 
   
Lack of Fit Tests      
                     Sum of                       Mean                 
Source  Squares DF                     Square F Value Prob > F  
Linear  6.96 20 0.35 14.86 0.0036 Suggested 
2FI  5.41 14 0.39 16.48 0.0030 
Quadratic 4.21 10 0.42 17.97 0.0026 
Cubic  0.73 2 0.36 15.52 0.0072 Aliased 
Pure Error 0.12 5                        0.023 
 
 
 Model Summary Statistics 
  Std.  Adjusted Predicted 
 Source Dev. R-Squared R-Squared R-Squared PRESS 
 Linear 0.53 0.9052 0.8901 0.8574 10.65 Suggested 
    2FI  0.54 0.9261 0.8872 0.7488 18.77 
    Quadratic 0.54 0.9421 0.8880 0.6731 24.42 
    Cubic  0.35 0.9887 0.9532 -0.4037 104.87 Aliased 
 
ANOVA for Response Surface Linear Model 
 
Analysis of variance table [Partial sum of squares] 
 
                 Sum of                Mean    
Source        Squares                DF            Square             F Value      Prob > F 
 
Model          67.63                    4                 16.91               3.57           < 0.0001  significant 
A               0.26        1               0.26              19.68 0.3486 
B                  61.47 1                 61.47          5.71 < 0.0001 
C 4.74 1                 4.74 1.53 0.0004 
D 1.16 1                 1.16 4.09 0.0540 
Residual       7.08                   25 0.28 
Lack of Fit   6.96                   20                 0.35               14.86        0.0036                      significant 
Pure Error    0.12 5                 0.023 
Cor Total     74.71                   29 
 
Std. Dev.     0.53                     R-Squared     0.9052 
Mean           5.41  Adj R-Squared 0.8901 
C.V.            10.35  Pred R-Squared 0.8574 
PRESS        10.65                Adeq Precision 29.468 
 
 
 
Coefficient  Standard                                                                95% CI          95% CI 
Factor Estimate DF Error Low High VIF 
Intercept 5.14 1 0.097                 4.94 5.34 
159 
 
 
A-A 0.10 1 0.11 -0.12 0.33 1.00 
B-B 1.60 1 0.11 -1.82 -1.38 1.00 
C-C 0.44 1 0.11 -0.67 -0.22 1.00 
D-D 0.22 1 0.11 -0.44                 4.124E-003 1.00 
   
Linear Model for Free Fatty Acid      
Final Equation in Terms of Coded Factors: 
FFA = +5.14 + 0.10Mc - 1.60HT - 0.44Ht - 0.22Ap 
Final Equation in Terms of Actual Factors: 
FFA = +18.18808 + 0.10375Mc - 0.16004HT - 0.088917Ht - 0.043917Ap 
[Std. Dev. = 0.53, R-Squared = 0.9052, Mean = 5.14, Adj R-Squared = 0.8901 
C.V. =  10.35, Pred R-Squared = 0.8574, PRESS = 10.65, Adeq Precision = 29.468]. 
2FI Model for Free Fatty Acid  
Final Equation in Terms of Coded Factors: 
FFA = +5.14 + 0.10Mc - 1.60HT - 0.44Ht – 0.22Ap + 0.16McHT - 0.15McHt - 0.14McAp - 
0.01HTHt + 0.11HTAp - 0.13HtAp  
Final Equation in Terms of Actual Factors: 
FFA = +17.57746 + 0.16813Mc - 0.34592HT + 0.30208Ht + 0.21033Ap + 0. 015687McHT - 
0.029625McHt - 0.028125McAp - 0.000213HTHt + 0.002288HTAp - 0.005325HtAp  
[Std. Dev. =  0.54, R-Squared = 0.9261, Mean = 5.14, Adj R-Squared = 0.8872 
C.V. = 10.49, Pred R-Squared = 0.7488, PRESS = 18.77, Adeq Precision = 19.608]. 
Quadratic Model for Free Fatty Acid   
Final Equation in Terms of Coded Factors 
2 2 2 2
FFA = +5.49 + 0.10Mc - 1.60HT - 0.44Ht - 0.22Ap - 0.077Mc  - 0.19HT  - 0.097Ht  – 0.067Ap  + 
0.16McHT - 0.15McHt - 0.14McAp - 0.011HTHt + 0.11HTAp - 0.13HtAp 
Final Equation in Terms of Actual Factors 
2
FFA = -2.07323 + 1.69937Mc - 0.079479HT + 0.49521Ht + 0.29021Ap - 0.076562Mc  - 
2 2 2
0.001903HT  - 0.003863Ht  - 0.002663Ap  + 0.015687McHT - 0.029625McHt - 0.028125McAp - 
0.000213HTHt + 0.002288HTAp - 0.005325HtAp 
[Std. Dev. = 0.54, R-Squared = 0.9421, Mean = 5.14, Adj R-Squared = 0.8880, 
C.V. = 10.45, Pred R-Squared = 0.6731, PRESS = 24.42, Adeq Precision = 16.854]. 
Cubic Model for Free Fatty Acid  
Final Equation in Terms of Coded Factors: 
160 
 
 
2 2 2 2
FFA = +5.49 – 0.15Mc - 1.96HT - 0.33Ht - 0.15Ap – 0.077Mc  - 0.19HT  - 0.097Ht  – 0.067Ap  + 
3 3
0.16McHT - 0.15McHt - 0.14McAp - 0.011HTHt + 0.11HTAp - 0.13HtAp + 0.12Mc  + 0.18HT  - 
3 3 
0.059Ht  - 0.034Ap + 0.099McHTHt + 0.13 McHTAp - 0.12McHtAp + 0.15HTHtAp 
Final Equation in Terms of Actual Factors: 
Not available for aliased models. 
[Std. Dev. = 0.35, R-Squared = 0.9887, Mean = 5.14, Adj R-Squared = 0.9532, 
C.V. = 6.76, Pred R-Squared = -0.4037, PRESS = 104.87, Adeq Precision = 17.130]. 
 
Where, 
Mc = Moisture content, %wb  
o
HT = Heating temperature, C 
Ht = Heating time, mins 
Ap = Applied pressure, MPa 
FFA = Free Fatty Acid, mg/KOH/g 
 
 
 
 
Diagnostics Case Statistics 
 
Standard  Actual         Predicted         Student      Cook's Outlier                Run  
Order       Value           Value                  Residual      Leverage  Residual       Distance        t                    Order 
1 7.06 7.30 -0.24 0.200    -0.508 0.013 -0.500 14 
2 7.21 7.51 -0.30 0.200          -0.629 0.020 -0.621 18 
3 3.45 4.10 -0.65 0.200          -1.368 0.094 -1.393 10 
4 3.88 4.31 -0.43 0.200          -0.900 0.041 -0.897 12 
5 6.84 6.41 0.43 0.200            0.898 0.040 0.894 15 
6 7.05 6.62 0.43 0.200            0.903 0.041 0.900 23 
7 2.76 3.21 -0.45 0.200          -0.949 0.045 -0.947 16 
8 2.93 3.42 -0.49 0.200          -1.028 0.053 -1.029 19 
9 7.03 6.86 0.17 0.200            0.352 0.006 0.345 26 
10 7.16 7.07 0.090 0.200            0.189 0.002 0.185 20 
11 3.34 3.66 -0.32 0.200           -0.676 0.023 -0.669 28 
12 3.65 3.87 -0.22 0.200           -0.461 0.011 -0.453 24 
13 6.74 5.97 0.77 0.200            1.610 0.130 1.667 21 
14 4.83 6.18 -1.35 0.200            -2.839 0.403 -3.378 30 
15 2.64 2.77 -0.13 0.200            -0.279 0.004 -0.274 4 
16 2.82 2.98 -0.16 0.200            -0.337 0.006 -0.330 5 
17 4.66 4.93 -0.27 0.200            -0.575 0.017 -0.567 8 
18 6.07 5.35 0.72 0.200             1.516 0.115 1.559 22 
19 7.40 8.34 -0.94 0.200            -1.979 0.196 -2.111 7 
20 2.42 1.94 0.48 0.200             1.008 0.051 1.009 1 
21 6.41 6.03 0.38 0.200             0.798 0.032 0.792 11 
22 4.16 4.25 -0.092 0.200            -0.193 0.002 -0.189 17 
23 5.98 5.58 0.40 0.200             0.840 0.035 0.835 2 
24 4.83 4.70 0.13 0.200             0.269 0.004 0.264 9 
25 5.36 5.14 0.22 0.033             0.419 0.001 0.412 29 
26 5.67 5.14 0.53 0.033             1.011 0.007 1.012 13 
161 
 
 
27 5.55 5.14 0.41 0.033             0.782 0.004 0.775 27 
28 5.27 5.14 0.13 0.033             0.247 0.000 0.242 6 
29 5.61 5.14 0.47 0.033             0.896 0.006 0.893 3 
30 5.45 5.14 0.31 0.033             0.591 0.002 0.583 25 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
APPENDIX J 
 
MODELLING OF COLOUR INTENSITY 
 
Design Summary 
 
Study Type Response Surface  Experiments 30 
Initial Design Central Composite  Blocks No Blocks 
Design Model  2FI 
 
 Response       Name       Units              Obs   Minimum Maximum Trans Model 
 Y2                  Colour      LU                  30   5.70  7.60 None 2FI 
 
 Factor        Name         Units Type            Low Actual     High Actual      Low Coded          High Coded 
 A                 Moisture Content     % wb Numeric          9.00                  11.00                    -1.000 1.000 
o
 B                 Heating Temp.         C Numeric          60.00                80.00                    -1.000 1.000 
 C                 Heating Time         mins Numeric          20.00                30.00                    -1.000 1.000 
 D                 Applied Pressure      MPa Numeric          10.00                20.00                    -1.000 1.000 
 
 
Design Matrix Evaluation for Response Surface 2FI Model 
      
162 
 
 
4 Factors: A, B, C, D  
Degrees of Freedom for Evaluation 
Model          10 
Residuals    19 
Lack 0f Fit 14 
Pure Error  5 
Corr Total    29 
     Power at 5 % alpha level for effect of 
Term  StdErr** VIF Ri-Squared 1/2 Std. Dev. 1 Std. Dev. 2 Std. Dev. 
    A                        0.20                    1.00                    0.0000                21.4 %                64.2 %               99.6 % 
  B 0.20 1.00 0.0000 21.4 % 64.2 % 99.6 % 
  C 0.20 1.00 0.0000 21.4 % 64.2 % 99.6 % 
  D 0.20 1.00 0.0000 21.4 % 64.2 % 99.6 % 
  AB 0.25 1.00 0.0000 15.8 % 46.5 % 96.6 % 
  AC 0.25 1.00 0.0000 15.8 % 47.6 % 96.6 % 
  AD 0.25 1.00 0.0000 15.8 % 47.6 % 96.6 % 
  BC 0.25 1.00 0.0000 15.8 % 47.6 % 96.6 % 
  BD 0.25 1.00 0.0000 15.8 % 47.6 % 96.6 % 
  CD 0.25 1.00 0.0000 15.8 % 47.6 % 96.6 % 
**Basis Std. Dev. = 1.0 
 
  
 
  
Measures Derived From the (X'X)-1 Matrix 
Std             Leverage                  Point Type 
1 0.5750  Fact 
2 0.5750  Fact 
3 0.5750  Fact 
4 0.5750  Fact 
5 0.5750  Fact 
6 0.5750  Fact 
7 0.5750  Fact 
8 0.5750  Fact 
9 0.5750  Fact 
10 0.5750  Fact 
11 0.5750  Fact 
12 0.5750  Fact 
13 0.5750  Fact 
14 0.5750  Fact 
15 0.5750  Fact 
16 0.5750  Fact 
17 0.2000  Axial 
18 0. 2000  Axial 
163 
 
 
19 0. 2000  Axial 
20 0. 2000  Axial 
21 0. 2000  Axial 
22 0. 2000  Axial 
23 0. 2000  Axial 
24 0. 2000  Axial 
25 0.0333  Center 
26 0.0333  Center 
27 0.0333  Center 
28 0.0333  Center 
29 0.0333  Center 
30 0.0333  Center 
Average = 0.3667 
          Maximum Prediction Variance (at a design point) =  0.575 
          Average Prediction Variance  =  0.367 
          Condition Number of Coefficient Matrix = 1.000 
          G Efficiency (calculated from the design points) =   63.8 %  
          Scaled D-optimality Criterion =  1.528 
      
 
Correlation Matrix of Regression Coefficients 
   Intercept  A  B  C   D AB AC   
Intercept 1.000 
A  -0.000 1.000 
B  -0.000 -0.000 1.000 
C  -0.000 -0.000 -0.000 1.000 
D  -0.000 -0.000 -0.000 -0.000  1.000   
AB  -0.000 -0.000 -0.000 -0.000  -0.000 1.000 
AC  -0.000 -0.000 -0.000 -0.000  -0.000 -0.000 1.000 
AD  -0.000 -0.000 -0.000 -0.000  -0.000 -0.000 -0.000 
BC  -0.000 -0.000 -0.000 -0.000  -0.000 -0.000 -0.000 
BD  -0.000 -0.000 -0.000 -0.000  -0.000 -0.000 -0.000 
CD  -0.000 -0.000 -0.000 -0.000  -0.000 -0.000 -0.000 
 
  AD                     BC                         BD                      CD 
AD  1.000                                             
BC  -0.000   1.000                                
BD  -0.000     -0.000               1.000                  
CD  -0.000      -0.000               -0.000         1.000                                       
  
164 
 
 
  
Correlation Matrix of Factors [Pearson's r] 
 
  A  B  C  D   AB AC   AD 
A  1.000 
B  0.000 1.000 
C  0.000 0.000 1.000 
D  0.000 0.000 0.000 1.000 
AB  0.000 0.000 0.000 0.000  1.000   
AC  0.000 0.000 0.000 0.000  0.000 1.000  
AD  0.000 0.000 0.000 0.000  0.000 0.000 1.000 
BC  0.000 0.000 0.000 0.000  0.000 0.000 0.000 
BD  0.000 0.000 0.000 0.000  0.000 0.000 0.000 
CD  0.000 0.000 0.000 0.000  0.000 0.000 0.000 
 
  BC  BD  CD   
BC  1.000 
BD  0.000 1.000 
CD  0.000 0.000 1.000 
  
Sequential Model Sum of Squares 
                   Sum of                     Mean                         
Source          Squares                         DF                   Square                    F Value               Prob > F 
Mean 1304.16 1 1304.16 
Linear  6.48 4 1.62 285.06 < 0.0001 Suggested 
2FI  0.065 6 0.011 2.67 0.0472 Suggested 
Quadratic 0.020 4               5.083E-003 1.35 0.2989  
Cubic  0.033 8               4.167E-003 1.25 0.3908 Aliased 
Residual 0.023 7               3.333E-003 
Total  1310.78 30 43.69 
  
Lack of Fit Tests      
                     Sum of                       Mean                 
Source  Squares DF                     Square F Value Prob > F 
Linear  0.13 20                  6.433E-003 2.41 0.1669 Suggested 
2FI  0.064 14                  4.548E-003 1.71 0.2895 Suggested 
Quadratic 0.043 10                  4.333E-003 1.63 0.3086    
Cubic  0.010 2                  5.000E-003 1.88 0.2468 Aliased 
Pure Error 0.013 5                  2.667E-003 
 
165 
 
 
 Model Summary Statistics 
  Std.  Adjusted Predicted 
 Source Dev. R-Squared R-Squared R-Squared PRESS 
 Linear 0.075 0.9785 0.9751 0.9678 0.21 Suggested 
    2FI  0.064 0.9884 0.9822 0.9614 0.26 Suggested 
   Quadratic 0.061 0.9914 0.9834 0.9594 0.27  
    Cubic  0.058 0.9965 0.9854 0.7795 1.46 Aliased 
 
 
 
 
 
 
 
 
 
 
 
 
 
ANOVA for Response Surface 2FI Model 
 
Analysis of variance table [Partial sum of squares] 
 
                 Sum of                Mean    
Source        Squares                DF            Square             F Value      Prob > F 
 
Model          6.54                      10               0.65               161.42          < 0.0001  significant 
A      0.17 1                 0.17              41.13            < 0.0001 
B         5.80 1                 5.80               1431.58        < 0.0001 
C 0.48 1                 0.48                118.85         < 0.0001 
D 0.027 1                 0.027 6.58 0.0189 
AB            1.000E-002 1            1.000E-002 2.47 0.1327 
AC            1.000E-002 1            1.000E-002 2.47              0.1327 
AD 0.022 1 0.022 5.55 0.0293 
BC            2.500E-003 1            2.500E-003 0.62 0.4419 
BD            1.000E-002 1            1.000E-002 2.47 0.1327 
CD            1.000E-002 1            1.000E-002           2.47 0.1327 
Residual       0.077                   19           4.053E-003 
Lack of Fit   0.064                   14           4.548E-003               1.71        0.2895                      not significant 
Pure Error    0.013 5            2.667E-003 
Cor Total     6.62                    29 
166 
 
 
 
Std. Dev.     0.064                     R-Squared     0.9884 
Mean           6.59  Adj R-Squared 0.9822 
C.V.            0.97  Pred R-Squared 0.9614 
PRESS        0.26                Adeq Precision 51.018 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Coefficient  Standard                                                                95% CI          95% CI 
Factor Estimate DF Error Low High VIF 
Intercept 6.59 1 0.012                6.57 6.62 
A-A 0.083 1 0.013               0.056 0.11 1.00 
B-B 0.49 1 0.013 0.46 0.52 1.00 
C-C 0.14 1 0.013 0.11 0.17 1.00 
D-D 0.033 1 0.013             6.135E-003 0.061 1.00 
AB -0.025 1 0.016 -0.058              8.311E-003 1.00 
AC 0.025 1 0.016             -8.311E-003 0.058 1.00 
AD 0.037 1 0.016              4.189E-003 0.071 1.00 
BC 0.012 1 0.016 -0.021 0.046 1.00 
BD -0.025 1 0.016 -0.058              8.311E-003 1.00 
CD 0.025 1 0.016             -8.311E-003 0.058 1.00 
   
Linear Model for Colour Intensity      
Final Equation in Terms of Coded Factors: 
CI = +6.59 + 0.083Mc + 0.49HT + 0.14Ht + 0.033Ap 
Final Equation in Terms of Actual Factors: 
CI = +1.51000 + 0.083333Mc + 0.0491674HT + 0.028333Ht + 0.006667Ap 
[Std. Dev. = 0.075, R-Squared = 0.9785, Mean = 6.59, Adj R-Squared = 0.9751 
C.V. =  1.14, Pred R-Squared = 0.9678, PRESS = 0.21, Adeq Precision = 63.919]. 
2FI Model for Colour Intensity  
167 
 
 
Final Equation in Terms of Coded Factors: 
CI = +6.59 + 0.083Mc + 0.49HT + 0.14Ht + 0.033Ap - 0.025McHT + 0.025McHt + 0.037McAp + 
0.012HTHt - 0.025HTAp + 0.025HtAp  
Final Equation in Terms of Actual Factors: 
CI = +2.42250 + 0.020833Mc + 0.075417HT - 0.054167Ht - 0.058333Ap - 0.002500McHT + 
0.005000McHt + 0.007500McAp + 0.000250HTHt - 0.000500HTAp + 0.001000HtAp  
[Std. Dev. =  0.064, R-Squared = 0.9884, Mean = 6.59, Adj R-Squared = 0.9822 
C.V. = 0.97, Pred R-Squared = 0.9614, PRESS = 0.26, Adeq Precision = 51.018]. 
Quadratic Model for Colour Intensity   
Final Equation in Terms of Coded Factors 
2 2 2
CI = +6.57 + 0.083Mc + 0.49HT + 0.14Ht + 0.033Ap + 0.002Mc  + 0.027HT  + 0.002Ht  + 
2
0.002Ap  - 0.025McHT + 0.025McHt + 0.037McAp + 0.012HTHt - 0.025HTAp + 0.025HtAp  
Final Equation in Terms of Actual Factors 
2
CI = +4.00208 - 0.020833Mc + 0.037500HT - 0.058333Ht - 0.060833Ap + 0.002083Mc  + 
2 2 2
0.000271HT  + 0.000083Ht  + 0.000083Ap  - 0.002500McHT + 0.005000McHt + 0.007500McAp 
+ 0.000250HTHt - 0.000500HTAp + 0.001000HtAp  
[Std. Dev. = 0.061, R-Squared = 0.9914, Mean = 6.59, Adj R-Squared = 0.9834, 
C.V. = 0.93, Pred R-Squared = 0.9594, PRESS = 0.27, Adeq Precision = 45.251]. 
Cubic Model for Colour Intensity  
Final Equation in Terms of Coded Factors: 
2 2 2
CI = +6.57 + 0.092Mc + 0.51HT + 0.16Ht + 0.042Ap + 0.002Mc  + 0.027HT  + 0.002Ht  + 
2
0.002Ap  - 0.025McHT + 0.025McHt + 0.037McAp + 0.012HTHt - 0.025HTAp + 0.025HtAp + 
3 3 3 3 
0.004Mc  - 0.008HT  - 0.008Ht  - 0.004Ap - 0.012McHTHt - 0.025 McHTAp + 0.025McHtAp - 
0.012HTHtAp 
Final Equation in Terms of Actual Factors: 
Not available for aliased models. 
[Std. Dev. = 0.058, R-Squared = 0.9965, Mean = 6.59, Adj R-Squared = 0.9854, 
C.V. = 0.88, Pred R-Squared = 0.7795, PRESS = 1.46, Adeq Precision = 37.585].  
 
Where, 
Mc = Moisture content, %wb  
o
HT = Heating temperature, C 
Ht = Heating time, mins 
Ap = Applied pressure, MPa 
168 
 
 
CI = Colour Intensity, LU 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Diagnostics Case Statistics 
 
Standard  Actual         Predicted         Student      Cook's Outlier               Run  
Order       Value           Value                  Residual      Leverage  Residual       Distance        t                    Order 
1 5.90 5.89 6.667E-003 0.575 0.161 0.003 0.156 14 
2 6.00 5.98 0.015 0.575 0.361 0.016 0.353 18 
3 6.90 6.95 -0.052 0.575 -1.245 0.191 -1.264 10 
4 7.00 6.94 0.057 0.575 1.365 0.229 1.399 12 
5 6.10 6.05 0.048 0.575 1.165 0.167 1.176 15 
6 6.20 6.24 -0.043 0.575 -1.044 0.134 -1.047 23 
7 7.20 7.16 0.040 0.575 0.964 0.114 0.962 16 
8 7.30 7.25 0.048 0.575 1.165 0.167 1.176 19 
9 5.90 5.89 0.015 0.575 0.361 0.016 0.353 26 
10 6.10 6.13 -0.027 0.575 -0.643 0.051 -0.632 20 
11 6.90 6.84 0.057 0.575 1.365 0.229 1.399 28 
12 7.00 6.98 0.015 0.575 0.361 0.016 0.353 24 
13 6.10 6.14 -0.043 0.575 -1.044 0.134 -1.047 21 
14 6.60 6.48 0.12 0.575 2.771 0.944 3.494 30 
15 7.20 7.15 0.048 0.575 1.165 0.167 1.176 4 
16 7.40 7.39 6.667E-003 0.575 0.161 0.003 0.156 5 
17 6.40 6.43 -0.027 0.200 -0.468 0.005 -0.458 8 
18 6.70 6.76 -0.060 0.200 -1.054 0.025 -1.057 22 
19 5.70 5.61 0.090 0.200 1.581 0.057 1.651 7 
20 7.60 7.58 0.023 0.200 0.410 0.004 0.401 1 
21 6.30 6.31 -1.000E-002 0.200 -0.176 0.001 -0.171 11 
22 6.80 6.88 -0.077 0.200 -1.346 0.041 -1.378 17 
23 6.50 6.53 -0.027 0.200 -0.468 0.005 -0.458 2 
24 6.60 6.66 -0.060 0.200 -1.054 0.025 -1.057 9 
25 6.60 6.59 6.667E-003 0.033 0.107 0.000 0.104 29 
26 6.60 6.59 6.667E-003 0.033 0.107 0.000 0.104 13 
27 6.60 6.59 6.667E-003 0.033 0.107 0.000 0.104 27 
28 6.50 6.59 -0.093 0.033 -1.491 0.007 -1.545 6 
169 
 
 
29 6.60 6.59 6.667E-003 0.033 0.107 0.000 0.104 3 
30 6.50 6.59 -0.093 0.033 -1.491 0.007 -1.545 25 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 APPENDIX K 
 
MODELLING OF OIL IMPURITY 
 
Design Summary 
 
Study Type Response Surface  Experiments 30 
Initial Design Central Composite  Blocks No Blocks 
Design Model  Linear 
 
 Response       Name       Units              Obs   Minimum Maximum Trans Model 
 Y3                  Impurity      %                  30   2.12  3.20 None Linear 
 
 Factor        Name         Units Type            Low Actual     High Actual      Low Coded          High Coded 
 A                 Moisture Content     % wb Numeric          9.00                  11.00                    -1.000 1.000 
o
 B                 Heating Temp.         C Numeric          60.00                80.00                    -1.000 1.000 
 C                 Heating Time         mins Numeric          20.00                30.00                    -1.000 1.000 
 D                 Applied Pressure      MPa Numeric          10.00                20.00                    -1.000 1.000 
 
 
Design Matrix Evaluation for Response Surface Linear Model 
 
4 Factors: A, B, C, D  
170 
 
 
Degrees of Freedom for Evaluation 
Model          4 
Residuals    25 
Lack 0f Fit 20 
Pure Error  5 
Corr Total    29 
 
Power at 5 % alpha level for effect of 
Term  StdErr** VIF Ri-Squared 1/2 Std. Dev. 1 Std. Dev. 2 Std. Dev. 
    A                        0.20                    1.00                    0.0000                21.8 %                65.3 %               99.7 % 
  B 0.20 1.00 0.0000 21.8 % 65.3 % 99.7 % 
  C 0.20 1.00 0.0000 21.8 % 65.3 % 99.7 % 
  D 0.20 1.00 0.0000 21.8 % 65.3 % 99.7 % 
 **Basis Std. Dev. = 1.0 
 
               
 
 
 
 
 
 
Measures Derived From the (X'X)-1 Matrix 
Std             Leverage                  Point Type 
1  0.2000      Fact 
2                  0.2000   Fact 
3                  0.2000   Fact 
4                  0.2000   Fact 
5                  0.2000   Fact 
6                  0.2000   Fact 
7                  0.2000   Fact 
8                  0.2000   Fact 
9                  0.2000   Fact 
10                 0.2000   Fact 
11                 0.2000   Fact 
12                 0.2000   Fact 
13                 0.2000   Fact 
14                 0.2000   Fact 
15                 0.2000   Fact 
16                 0.2000   Fact 
17                 0.2000   Axial 
18                 0.2000   Axial 
19                 0.2000   Axial 
171 
 
 
20                 0.2000   Axial 
21                 0.2000   Axial 
22                 0.2000   Axial 
23                 0.2000   Axial 
24                 0.2000   Axial 
25                 0.0333   Center 
26                 0.0333   Center 
27                 0.0333   Center 
28                 0.0333   Center 
29                 0.0333   Center 
30                 0.0333   Center 
Average = 0.1667 
 
          Maximum Prediction Variance (at a design point) =  0.200 
          Average Prediction Variance  =  0.167 
          Condition Number of Coefficient Matrix = 1.000 
          G Efficiency (calculated from the design points) =   83.3 %  
          Scaled D-optimality Criterion =  1.195 
 
Correlation Matrix of Regression Coefficients 
   Intercept  A  B  C   D   
Intercept 1.000 
A  -0.000 1.000 
B  -0.000 -0.000 1.000 
C  -0.000 -0.000 -0.000 1.000 
D  -0.000 -0.000 -0.000 -0.000  1.000 
 
Correlation Matrix of Factors [Pearson's r] 
  A  B  C  D    
A  1.000 
B  0.000 1.000 
C  0.000 0.000 1.000 
D  0.000 0.000 0.000 1.000 
 
  
     
Sequential Model Sum of Squares 
                   Sum of                     Mean                         
Source          Squares                         DF                   Square                    F Value               Prob > F 
Mean 215.85 1 215.85 
Linear  1.98 4 0.49 126.50 < 0.0001 Suggested 
2FI  0.024 6                  3.942E-003 1.01 0.4478 
172 
 
 
Quadratic 0.022 4                  5.595E-003 1.62 0.2205  
Cubic  0.032 8                  3.950E-003 1.37 0.3450 Aliased 
Residual 0.020 7                  2.880E-003 
Total  217.92 30 7.26 
 
Lack of Fit Tests      
                     Sum of                       Mean                 
Source  Squares DF                     Square F Value Prob > F  
Linear  0.097 20                  4.829E-003 20.12 0.0017 Suggested 
2FI  0.073 14                  5.210E-003 21.71 0.0016 
Quadratic 0.051 10                  5.056E-003 21.07 0.0018 
Cubic  0.019 2                  9.479E-003 39.50 0.0009 Aliased 
Pure Error       1.200E-003 5                  2.400E-003 
 
Model Summary Statistics 
  Std.  Adjusted Predicted 
 Source Dev. R-Squared R-Squared R-Squared PRESS 
 Linear 0.063 0.9529 0.9454 0.9300 0.15 Suggested 
    2FI  0.062 0.9643 0.9455 0.8717 0.27 
    Quadratic 0.059 0.9751 0.9518 0.8590 0.29 
    Cubic  0.054 0.9903 0.9598 -0.3153 2.73 Aliased 
 
ANOVA for Response Surface Linear Model 
 
Analysis of variance table [Partial sum of squares] 
 
                 Sum of                Mean    
Source        Squares                DF            Square             F Value      Prob > F 
 
Model          1.98                      4                 0.49               126.50           < 0.0001  significant 
A                 0.027        1                 0.027              6.82 0.0150 
B                  1.74 1                 1.74          444.54             < 0.0001 
C 0.19 1                 0.19 49.70             < 0.0001 
D                  0.019 1                 0.019 4.93 0.0358 
Residual       0.098                  25               3.911E-003 
Lack of Fit   0.097                  20               4.829E-003        20.12        0.0017                          significant 
Pure Error    1.200E-003         5                2.400E-004 
Cor Total      2.08                   29 
Std. Dev.     0.063                     R-Squared     0.9529 
Mean           2.68  Adj R-Squared 0.9454 
C.V.            2.33  Pred R-Squared 0.9300 
PRESS        0.15                Adeq Precision 42.168 
 
 
 
Coefficient  Standard                                                                95% CI          95% CI 
173 
 
 
Factor Estimate DF Error Low High VIF 
Intercept 2.68 1     0.011                 2.66 2.71 
A-A 0.033 1 0.013           7.041E-003 0.060 1.00 
B-B -0.27 1 0.013 -0.30 -0.24 1.00 
C-C -0.090 1 0.013 -0.12 -0.064 1.00 
D-D -0.028 1 0.013               -0.055            -2.041E-003 1.00 
   
Linear Model for Oil Impurity      
Final Equation in Terms of Coded Factors: 
CL = +2.68 + 0.033Mc - 0.27HT - 0.090Ht - 0.028Ap 
Final Equation in Terms of Actual Factors: 
CL = +4.76817 + 0.033333Mc - 0.026917HT - 0.018000Ht - 0.005667Ap 
[Std. Dev. = 0.063, R-Squared = 0.9529, Mean = 2.68, Adj R-Squared = 0.9454 
C.V. =  2.33, Pred R-Squared = 0.9300, PRESS = 0.15, Adeq Precision = 42.168]. 
2FI Model for Oil Impurity  
Final Equation in Terms of Coded Factors: 
OI = +2.68 + 0.033Mc + 0.27HT - 0.090Ht - 0.028Ap + 0.014McHT - 0.028McHt - 0.015McAp - 
0.005HTHt + 0.012HTAp - 0.011HtAp  
 
Final Equation in Terms of Actual Factors: 
OI = +3.82442 + 0.11958Mc - 0.041917HT + 0.050750Ht + 0.018083Ap + 0.001375McHT - 
0.005500McHt - 0.003000McAp - 0.000100HTHt + 0.000250HTAp - 0.000450HtAp  
[Std. Dev. =  0.062, R-Squared = 0.9643, Mean = 2.68, Adj R-Squared = 0.9455 
C.V. = 2.33, Pred R-Squared = 0.8717, PRESS = 0.27, Adeq Precision = 28.465].  
Quadratic Model for Oil Impurity   
Final Equation in Terms of Coded Factors 
2 2 2 2
OI = +2.73 + 0.033Mc - 0.27HT - 0.090Ht - 0.028Ap - 0.010Mc  - 0.026HT  - 0.011Ht  - 0.013Ap  
+ 0.014McHT - 0.028McHt - 0.015McAp - 0.005HTHt + 0.012HTAp - 0.011HtAp 
Final Equation in Terms of Actual Factors 
2
OI = +1.25708 + 0.311250Mc - 0.005750HT + 0.072417Ht + 0.034083Ap - 0.009583Mc  - 
2 2 2
0.000258HT  - 0.000433Ht  - 0.000533Ap  + 0.001375McHT - 0.005500McHt – 0.003000McAp - 
0.000100HTHt + 0.000250HTAp - 0.000450HtAp 
[Std. Dev. = 0.059, R-Squared = 0.9751, Mean = 2.68, Adj R-Squared = 0.9518, 
C.V. = 2.19, Pred R-Squared = 0.8590, PRESS = 0.29, Adeq Precision = 25.921]. 
Cubic Model for Oil Impurity 
Final Equation in Terms of Coded Factors: 
174 
 
 
2 2 2 2
OI = +2.73 + 0.024Mc - 0.27HT - 0.12Ht - 0.032Ap - 0.010Mc  - 0.026HT  - 0.011Ht  - 0.013Ap  + 
3
0.014McHT - 0.028McHt - 0.015McAp - 0.005HTHt + 0.012HTAp - 0.011HtAp + 0.005Mc  - 
3 3 3 
0.001HT  + 0.017Ht  + 0.002Ap + 0.023McHTHt + 0.015 McHTAp - 0.011McHtAp + 0.011HTHtAp 
Final Equation in Terms of Actual Factors: 
Not available for aliased models. 
[Std. Dev. = 0.054, R-Squared = 0.9903, Mean = 2.68, Adj R-Squared = 0.9598, 
C.V. = 2.00, Pred R-Squared = -0.3153, PRESS = 2.73, Adeq Precision = 22.985].  
 
Where, 
Mc = Moisture content, %wb  
o
HT = Heating temperature, C 
Ht = Heating time, mins 
Ap = Applied pressure, MPa 
OI = Oil Impurity, % 
 
 
 
Diagnostics Case Statistics 
 
Standard  Actual         Predicted         Student      Cook's Outlier                Run  
Order       Value           Value                  Residual      Leverage  Residual       Distance        t                    Order 
1 2.98 3.04 -0.056 0.200 -1.010 0.051 -1.010 14 
2 3.11 3.10 6.833E-003 0.200  0.122 0.001 0.120 18 
3 2.46 2.50 -0.038 0.200 -0.682 0.023 -0.675 10 
4 2.57 2.56 5.167E-003 0.200  0.092 0.000 0.091 12 
5 2.86 2.86 3.500E-003 0.200  0.063 0.000 0.061 15 
6 2.91 2.92 -0.013 0.200 -0.235 0.003 -0.231 23 
7 2.26 2.32 -0.058 0.200 -1.040 0.054 -1.042 16 
8 2.32 2.38 -0.065 0.200 -1.159 0.067 -1.167 19 
9 2.94 2.98 -0.040 0.200 -0.712 0.025 -0.705 26 
10 3.07 3.05 0.024 0.200  0.420 0.009 0.413 20 
11 2.44 2.44 -1.500E-003 0.200 -0.027 0.000 -0.026 28 
12 2.52 2.51 0.012 0.200  0.212 0.002 0.207 24 
13 2.85 2.80 0.050 0.200  0.897 0.040 0.893 21 
14 2.66 2.87 -0.21 0.200 -3.692 0.681 -5.363 * 30 
15 2.21 2.26 -0.051 0.200 -0.921 0.042 -0.918 4 
16 2.30 2.33 -0.028 0.200 -0.504 0.013 -0.496 5 
17 2.64 2.62 0.024 0.200  0.435 0.009 0.428 8 
18 2.81 2.75 0.061 0.200  1.090 0.059 1.095 22 
19 3.20 3.22 -0.021 0.200 -0.369 0.007 -0.363 7 
20 2.12 2.14 -0.024 0.200 -0.429 0.009 -0.422 1 
21 2.83 2.86 -0.032 0.200 -0.578 0.017 -0.570 11 
22 2.61 2.50 0.11 0.200  1.925 0.185 2.043 17 
23 2.76 2.74 0.021 0.200  0.375 0.007 0.369 2 
24 2.66 2.63 0.034 0.200  0.614 0.019 0.606 9 
25 2.72 2.68 0.038 0.033  0.613 0.003 0.605 29 
175 
 
 
26 2.72 2.68 0.038 0.033  0.613 0.003 0.605 13 
27 2.71 2.68 0.028 0.033  0.450 0.001 0.443 27 
28 2.74 2.68 0.058 0.033  0.938 0.006 0.935 6 
29 2.74 2.68 0.058 0.033  0.938 0.006 0.935 3 
30 2.75 2.68 0.068 0.033  1.100 0.008 1.105 25 
 
* Case(s) with |Outlier T| > 3.50   
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
APPENDIX L 
 
AUTOCAD DRAWINGS OF THE MORINGA EXPELLER AND THE COMPONENT PARTS 
 
176 
 
 
 
Figure L1. The Hopper 
  
 
Figure L2. The Electric Motor  
 
177 
 
 
 
 
Figure L3. The Roasting Chamber  
 
Figure L4. The Oil Barrel 
 
178 
 
 
 
Figure L5. The Wormshaft 
 
 
Figure L6. The Barrel Oil Cover 
179 
 
 
 
Figure L7. The Oil Trough 
  
 
Figure L8. The Temperature Regulator  
180 
 
 
 
Figure L9. Belt and Pulley Arrangement  
 
 
Figure L10. The Bevel Gears  
181 
 
 
 
 
Figure L11. The Stirrer  

  
Figure L12. The Electric Heater   
182 
 
 
 




 
Figure L13. The Pillow Bearings   
 
 
 
 
Figure L14. The Machine Base  
183 
 
 


 
Figure L15. The Front View 
 
184 
 
 
 
 
Figure L16. The Right End View 
185 
 
 
 
Figure L17. The Left End View 
186 
 
 
 
Figure L18. The Plan View 
 
 
187 
 
 
 
Figure L19. The Schematic View 
 
188 
 
 
 
 
Figure L20. The SchematicExploded View 
 
PART DESCRIPTION PART DESCRIPTION 
A Temperature Regulator N Bolt and Nut 
B Hopper O Hopper Passage 
C Wormshaft  P Belts 
D Oil Barrel Q Electric Motor Stand 
E Worms R Electric Motor 
F Hopper Base S Driving Pulley 
G Mechanism Link T Electric Heater 
H Pillow Bearing U Roasting Drum 
I Cake Trough V Roasting Drum Stand 
J Barrel Oil Cover W Stirrer Blade 
K Oil Trough X Bevel Gears 
L Machine Frame Y Stirrer  
M Driven Pulley Z Sliding Gate 
 
 
189 
 
 
 
 
 
 
Figure L21. The Wireframe View 
 
 
 
 
190 
 
 

 
Figure L22. The Main View 
191 
 
 
 
Figure L23. Multiple Views 











192 
 
 


Figure L24. The Exploded View 

PART DESCRIPTION PART DESCRIPTION 
A Temperature Regulator N Bolt and Nut 
B Hopper O Hopper Passage 
C Wormshaft  P Belts 
D Oil Barrel Q Electric Motor Stand 
E Worms R Electric Motor 
F Hopper Base S Driving Pulley 
G Mechanism Link T Electric Heater 
H Pillow Bearing U Roasting Drum 
I Cake Trough V Roasting Drum Stand 
J Barrel Oil Cover W Stirrer Blade 
K Oil Trough X Bevel Gears 
L Machine Frame Y Stirrer  
M Driven Pulley Z Sliding Gate 



193