we" llHlWllllHllHllHHIIlllllllllllllmllHlllllNllHll , LIBRARY 3 1293 10063 2730 MichiganSutc ._ University This is to certify that the thesis entitled SIMULATION 'IO EVALUATE ALTERNATIVES IN POST RICE PRODUCTION presented by Moeljarno Dj ojomartono has been accepted towards fulfillment of the requirements for _Ph._L_ degree in wall Eng. fiaszmm Major professor Date [OI/ZZi/7'CZ 0-7 639 OVERDUE FINES ARE 25¢ PER DAY~ PER ITEM this checkout from your record. laws} 000 "321. O ' ‘ ( '\ Return to book drop to remOve r l SIMULATION TO EVALUATE ALTERNATIVES IN POST RICE PRODUCTION BY Moeljarno Djojomartono A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Agricultural Engineering 1979 "r zr“ ABSTRACT SIMULATION TO EVALUATE ALTERNATIVES IN POST RICE PRODUCTION BY Moeljarno Djojomartono The problems of increasing the quality and quan— tity of rice production consists of two large parts: first, the production problems and second, the post rice production problems. The problems within the in-field post rice production operations of harvesting, threshing, handling, and drying are viewed as complex interactions and interrelations between components within the system and the components with the environment. In this disser— tation, a simulation model of post rice production opera— tions is presented as an approach to the post rice production problems. Field measurements of operation for post rice pro- duction performances and losses were carried out in Indonesia. The current post rice production research was reviewed. Field measurement data and reliable secondary data were analyzed and equations were developed to model the interactions and interrelations among post rice pro- duction operations. Moeljarno Djojomartono Environment plays an important role in the post rice production operations. Four weather related models were developed and used to capture the influence of environment on the post rice production system. Rain and no rain condition each day and the rainfall amount on rain days were simulated stochastically. The results from the weather models were used in both the field mois- ture model and the working hours model. The field mois— ture model approximates the soil moisture status and determines the in—field working condition and efficiency. The preliminary working hours model describes the avail- able working hours for each post rice production opera- tion based upon the stochastically generated rainfall during working time. Grain losses in the post rice production opera— tions were generally influenced by the timeliness of the operations. Delayed harvesting beyond the optimum date increased shattered grain losses during harvesting, threshing and handling operations. The amount of head rice as the final product from in—field post rice produc— tion process was determined by harvesting, threshing, handling and drying dates and methods. A model to account for crop maturation times was developed, and six models in discrete time form were designed to simulate the final products and losses from the post rice production opera— tions. Moeljarno Djojomartono The simulation model was used to evaluate alterna- tives in post rice production. The influence of environ- ment such as climate, the size of cultivated land and labor availability to the post rice production operations were also studied using the simulation model. The simu— lation results provided useful information to the decision makers or planners involved in the development and improve- ment of post rice production technology. The result of the evaluation can be used to identify the advantages, constraints and weakness in the varius technological alternatives. The simulation results using a given set of data and under alternatives in this research study indicated that the use of sickles and foot threading was the most appropriate method for producing high quality and quan- tity of rice. The high head rice conversion capability and lower weather dependency of a mechanical dryer sug— gested the use of mechanical dryers to increase and stablilize the annual head rice production. However, the cost of operating and owning the dryer resulted in less net annual income per hectare than useof sun drying methods. Further research in improvement and modification of model or data, and extensive use of the present model Moeljarno Djojomartono are recommended to more realistically represent and to add to one's understanding of the post rice production system. Approved by t F Z£2£2£ W r/e/fi/ /'///'-//. K ‘ 'V '/ ‘ /' If. - suing} 2 f? A“: coca. \_ Department Chairman Dedicated to my late parents ii ACKNOWLEDGMENTS I gratefully acknowledge the support of many individuals and organizations who contributed to this study. I particularly wish to express my appreciation to: Dr. Merle L. Esmay, my major professor, for his guidance, unfailing courtesy and patient encouragement during this work. Dr. Michael H. Abkin, Dr. Allan C. Rotz and Dr. Bill A.Stout, the other members of the Guidance Committee, for their helpful suggestions and continued interest. U. S. Agency for International Development for their financial support. Many staff members of the Institut Pertanian Bogor, the Sub-Directorate of Agricultural Mechanization, Department of Agriculture Republic of Indonesia and Agri— cultural Extension Service, West Java Province, for their help in collecting the field data. Finally, my wife Dati and son Michigana for their unending patience and understanding during my absence from home. TABLE OF CONTENTS LIST OF TABLES . . . . . . . . . . . LIST OF FIGURES . . . . . . . . . . . CHAPTER I. INTRODUCTION . . . . . . . . . . 1.1 Objectives . . . . . . . . II. A REVIEW OF HARVESTING, THRESHING AND DRYING OPERATIONS IN INDONESIA . . . . 2.1 Traditional Methods . . . . . 2.1.1 Panicle Harvesting (Ani-Ani Harvesting) . . . . . . 2.1.2 Sickle Harvesting . . . . 2.1.3 Foot Threshing . . . . . 2.1.4 Sun Drying . . . . . . 2.2 Mechanical Methods . . . . . . 2.2.1 Manual Dropper . . . . . 2.2.2 Mechanical Reaper . . . . 2.2.3 Manual and Mechanical Binders 2.2.4 Pedal Thresher . . . . . 2.2.5 Power Thresher . . . . 2.2.6 Small Combine- -Harvester . . 2.2.7 Mechanical Dryer . . III. FIELD STUDY . . . . . . . . . . 3.1 Methodology . . . . . 3.1.1 Harvesting Losses . . 3.1.2 Packing and Transportation Losses . . . . . . . 3.1.3 Threshing Losses . . . 3.1.4 Temporary Storage Losses . . 3.2 Results . . . . 3 2 l Harvesting Losses . . 3 2.2 Packing and Transportation Losses . . . . . . . . iv Chapter IV. VI. 3.2.3 Losses Dur 3.2.4 Temporary 3.3 Discussion . . FEASIBILITY ANALYSIS Need Analysis . ing Threshing . Storage Losses . System Identification . . . Measure of System Performance . . 4 l 4 2 4.3 Problem Statement 4 4 4 5 (Solutions) . . 4.6 Experimental Desi THE SIMULATION MODEL 5.1 Weather Model 5.1.1 Weather Da Alternative System Concepts gn . . . . ta . 5.1.2 Weather Model Development . 5.1.3 Rainfall Model . . . . 5.2 Working Hours Model . . . . . 5.3 Field Condition Model . . . . 5.3.1 Field Moisture Model . . 5.4 Plant Condition Model . . . . 5.4.1 Crop Yield Model . . . 5.4.2 Crop Maturing Model . . . 5.5 Loss Models . . 5.5.1 Shattering Losses During Harvest Operation . . 5.5.2 Packing an Losses . Losses Dur Combine- -Ha Temporary Drying Los U‘IU'lU'IU'I U'IU'IUlU'l 'O'tU'lvD-LA) U'IU'IU'I I. ooqm Overall Model . SIMULATION RESULTS AND 6.1 Two C 6.1.1 6.1.2 Harvesting 6.1.3 Changes in 6.1.4 Changes in 6.1.5 Harvesting Harvester 5.8.1 Overall Model Output . d Transportation 1ng Threshing . rvester Losses . Storage Losses ses Operation Costs and Income Model . Model Verification . . . . . DISCUSSTION . . rops per Year Schedule Ani-ani (Panicle) Harvesting with Sickles . Threshing Methods Drying Methods with a Combine- V Page 106 Chapter Page 6.2 Twelve Crops per Five Year Schedule 110 6.3 Annual Net Income per Hectare . . 112 6.4 Comparison of Alternatives . . . 119 VII. CONCLUSIONS . . . . . . . . . . 125 VIII. SUGGESTIONS FOR FURTHER RESEARCH . . . 129 BIBLIOGRAPHY . . . . . . . . . . . . 131 vi LIST OF TABLES 3.1 Grain Losses for Different Harvesting Methods and IR-38 . . 3.2 Unthreshed and Husked Grain Losses for Different Threshing Methods 6.1 Input Variable and Parameter Values Used in Simulation . . . 6.2 Parameter Values for Cost and Income Calculations . . . t Page 35 95 97 Figure 2.1 LIST OF FIGURES Manual Dropper Schematic Diagram . . . Pedal Thresher Schematic Diagram . Grain Moisture Contents and Temperature at Different Drying Times . . . . Harvesting Losses at Different Harvesting Times (IR-36) . . . . . . . . . . Handling Losses at Different Harvesting Times (IR-36) . . . . . . . . . Deteriorated Grain Losses on Temporary Storages (IR-36) . . . . . . . . . Schematic Diagram of a System . . . Schematic Diagram of All Possible Post Rice Production Operation Changes . . . Cropping' Pattern and Harvesting Sched- ules . . . . . . . . . . . . . Rainfall Model Flow Chart . . . . Cumulative Probability of Rainfall Amount in March . . . . . . . . . Rainfall Amount Model Flow Diagram . . Relationship Between Working Hours and Rainfall Amount . . . . . . . . . Soil Moisture Status Flow Chart (Part 1) . Soil Moisture Status FLow Chart (Part 2) . Continuous Delays of Different Orders (Orders = K) . . . . . . . . . . viii Page 19 22 28 34 36 4O 46 52 55 63 64 66 68 72 73 78 Figure Page 5.7 Histogram of Record and Simulated Rainfall Amount in the First 10 Days of March . . 86 5.8 Comparison of Non-rain Sequence Length from Actual Data and Simulated Results for 18 Year Period . . . . . . . . . . . 88 5.9 Overall Model Diagram (Part 1) . . . . 90 5.9 Overall Model Diagram (Part 2) . . . . 91 5.9 Overall Model Diagram (Part 3) . . . . 92 5.10 Overall Model Block Diagram . . . . . 93 6.1 Comparison of (a) Head Rice Production, (b) Broken Rice Production and (c) Shat— tered Grain Losses on Three Different Ani- ani Harvesting, Foot Threshing and Sun Drying Alternatives . . . . . . . . 100 6.2 Comparison of (a) Head Rice Production, (b) Broken Rice Production and (c) Shat- tered Grain Losses on an Alternative with Sickle Harvesting and an Alternative with Ani—ani Harvesting . . . . . . . . 102 6.3 Comparison of (a) Head Rice Production, (b) Broken Rice Production and (c) Shat- tered Grain Losses on an Alternative with Foot Threshing and an Alternative with Power Threshing . . . . . . . . . 104 6.4 Comparison of (a) Head Rice Production, (b) Broken Rice Production and (c) Shat- tered Grain Losses on Three Different Threshing Practices . . . . . . . . 105 6.5 Comparison of Annual Head Rice Production per Hectare of Three Different Dryer Capacities . . . . . . . . . . . 107 6.6 Comparison of (a) Head Rice Production, (b) Broken Rice Production and (c) Shat- tered Grain Losses on an Alternative with Sun Drying and an Alternative with Mechanical Drying . . . . . . . . . 108 ix Figure Page 6.7 Comparison of (a) Head Rice Production, (b) Broken Rice Production and (c) Shat— tered Grain Losses on Different Cropping Schedule . . . . . . . . . . 111 6.8 Comparison of (a) Head Rice Production, (b) Broken Rice Production and (c) Shat— tered Grain Losses on Different Cropping Schedules and Yield Factors . . . . 113 6.9 Comparison of Annual Net Income per Hectare on Two Different Ani-ani Harvesting, Foot Threshing and Sun Drying Alternatives . . 114 6.10 Comparison of Annual Net Income per Hectare on Three Different Threshing Practices . . 116 6.11 Comparison of Annual Net Income per Hectare on an Alternative with Sun Drying and an Alternative with Mechanical Drying . . . 117 6.12 Comparison of Annual Net Income per Hectare with Two Different Dryer Capacities and Different Dryer Price . . . . . . . 118 6.13 Comparison of Annual Net Income per Hectare for Three Different Cropping Schedules . . 120 6.14 Comparison of Annual Head Rice Production for Nine Alternatives and a Traditional Method for 1.0 Hectare Farm Size . . . . 121 6.15 Comparison of Annual Head Rice Production for Nine Alternatives and a Traditional Method for 3.0 Hectare Farm Size . . . . 122 CHAPTER I INTRODUCTION The Indonesian government has been actively under- taking BIMAS (Mass Guidance), a rice intensification pro- gram, since 1963 in order to meet the increasing demand for rice. Initially the BIMAS program was designed to deal with the production problems of rice. The BIMAS program covered 43 percent of the total dry and wet land fields in Indonesia in 1976. The introduction of the new higher yielding varie- ties, proper use of fertilizer and improved cultural prac- tices, have all contributed to increased rice yields. The average yield before the BIMAS program was 2.43 tons of rough rice per hectare (BIMAS, 1972). The average yield in 1976 was 3.66 tons per hectare, which was 51 percent above the 1963 average yield. Increases in rice produc— tion technology were not, however, followed by improved post rice production technology and loss reduction. An increasing concern about the post production losses has developed during recent years (USAID, 1978). Improved production technology provided the rice producers with increased yields, but not a comparable increase in total production. Most of the new varieties shatter more easily. Multiple cropping has shifted much harvesting to the wet season, so drying and handling losses increase. These high post production losses have tended to nullify some of the advantages of high yields. The total post rice production losses in develop- ing countries are estimated to be from 10 to 37 percent (Spurgeon, 1976). BULOG (Indonesian Board of Logistics) estimated in 1971 a total of 25 percent loss after pro— duction of rice in Indonesia. A 12 percent loss was estimated within the field operations of harvesting, handling, threshing and drying. The monetary value of post rice production losses has been more than 200 million rupiah (U.S. $320,000) annually. This was based upon a 12 percent loss of 22,732,000 tons of rough rice produced in 1974 (CBS, 1976) and a price of 75.00 rupiah per kilogram. 'This loss was equivalent to 181 percent of the total Indonesian rice import of 1,509,000 tons for the year 1976/1977. Appropriate post rice production methods and tech- niques are important in helping stabilize the Indonesian economy by minimizing rice imports. Also, increased returns would accrue to the producers, farmers and workers who invest in the production systems. The traditional Indonesian post rice production operations of harvesting, handling, threshing and drying are manual, seasonal and comparatively low wage jobs. They are labor intensive operations which seem appropriate for a highly populated Indonesia. But, they are tough, tedious, in—field jobs under the hot tropical sun and often in wet soil condi— tions. Farm working conditions have improved little as compared to transportation with its introduction of one— half million motor bikes per year. The continuation of this imbalance with other fast growing industrial develop— ment in the urban areas, will bring about more emigration to the cities (Esmay, 1978). The improvement of post rice production methods and techniques should be equally as important as the rice production phase. The initial post rice production operations are harvesting, handling, threshing and drying which occur mainly in the field. The post production operations all affect the quality and quantity of rice for consumption. For example, delayed harvesting will generally increase the harvesting field losses and lower the quality of that saved. Emphasis must therefore be placed on all post rice production operations as a system. Hopefully, improved post rice production systems will provide better quantity and quality rice, minimize post production losses and improve field working condi— tions. An improved system might be brought about by adoption of selective innovations. The determination of a feasible and practical system requires knowledge of its behavior under many conditions. Some of this information can be obtained by spending substantial amounts of time and money experimenting with variations of the real world systems. Simulation modeling provides an analytical tool for extending limited experimental results and measurements to themany variations within the real world at less cost and in less time (Hillier and Lieberman, 1974). Simula- tion systems modeling of agricultural product harvesting of wheat and corn has been done by Campbell and McQuitty (1971), van Kampen (1971) and Holtman, et a1., (1973). Fridley (1974) evaluated alternate California strawberry harvest systems. Brown (1972) has done simulation model— ing on a harvest system for apples, and a simulation model for sugarcane harvest operations under stochastic conditions has been formulated by Sorensen and Gilheany (1970). This research study was designed to obtain needed field information for the verification of a post rice production model, so that simulation could be used to evaluate and identify some feasible and practical post rice production technology alternatives. The study con- centrated on the field operations of harvesting, handling, threshing and drying. In—field data collection was carried out in a rice producing area in West Java, Indonesia, during the wet and dry season of 1978. These primary data along with reliable secondary data were employed in formulating and testing the simulation model. 1.1 Objectives The general objective of this research was to identify and quantify the potential gains to be realized by the reduction of post rice production losses through improved harvesting, handling, threshing and drying tech- nology. The specific objectives were: 1. To measure field losses and machine per- formance in order to formulate and verify a representative post rice production system model. 2. To use the model to estimate post rice production losses and costs, thereby identi- fying some feasible alternatives and practi— cal combinations of rice harvesting, handling, threshing and drying technology for Indonesia. This research study was designed to provide the decision maker with reliable information for the develop- ment and improvement of post rice production technology and methods. CHAPTER II A REVIEW OF HARVESTING, THRESHING AND DRYING OPERATIONS IN INDONESIA Estimated Indonesian rice consumption per year per person in 1977 was between 123.1 and 124.6 kilograms. With a total population of l36,766,000,the estimated demand for rice was between 17,052,500 and 17,298,800 tons (CBS, 1977). The total rice production in Indonesia was 15,884,000 tons in 1976. It was produced by both tradi- tional and improved methods. The traditional methods were carried out under both dry land and wet land condi- tions. The total area of rice in 1976 was 8,368,759 hectares. This area consisted of 43 percent (3,616,089 hectares) wet land traditional, 14 percent or 1,139,342 hectares dry land traditional and 43 percent (3,613,328 hectares) under improved cultivation. Rice production included primary tillage, sowing, transplanting and weed— ing and/or spraying. The wet land traditional practices were subdivided into wet monsoon planting and dry monsoon planting. This amounted to 65 percent (2,345,617 hectares) and 35 percent (1,261,472 hectares) of the total wet land traditional area, respectively in the 1975/1976 season. The wet land improved production practices were also subdivided into wet monsoon and dry monsoon planting. Each consisted of 64 percent or 2,319,294 hectares and 36 percent or 1,294,034 hectares, respectively. The post rice production processes consist of a long and necessary series of operations to transform the mature standing field paddy into a high quality processed rice for the consumers. The post rice production systems start with the harvesting operation (cutting, bunching, laying and bundling) followed by threshing, handling and drying, milling and storage. The final operation of the in-field post rice production system is the drying opera- tion. Drying reduces the moisture content of the wet or partially wet rough rice to a dry rough rice at 13 to 14 percent wet basis for milling and/or safe storage. The harvesting time is determined by the maturity of the crop and thus depends on the planting date and cropping pattern, rice variety and weather conditions. A survey conducted by Biro Pusat Statistik (CBS) in 1977/ 1978 showed the frequency distribution of harvest time varied among provinces, but all presented a similar type of double peak distribution. The large peak was common during the wet monsoon period and the smaller during the dry monsoon period. In West Java, where the Bekasi area is located, the frequency distribution of harvest time had a high peak of 43 percent for March and April and 44 per— cent for July and August (Eriyatno, 1979). 2.1 Traditional Methods The post rice production operations were carried out by various methods, depending on size of the farm, surrounding conditions and labor availability. The tra— ditional harvesting method in most Indonesian provinces was by cutting rice panicles individually with a small knife called ani—ani. Only in two provinces, West Sumatra and North Sulawesi, was traditional rice cutting with a sickle observed (BULOG, 1977 and Eriyatno, 1979). The introduction of the high yielding varieties which shattered more easily was followed by more sickle cutting in many of the provinces. Collier, et al., (1973) stated that the use of the sickle in harvesting was logical for the new rice varieties. Traditional threshing techniques vary between regions. Beating and hand pounding are probably the most common means of threshing the indigenous varieties in many Indonesian provinces. As the new, less shatter proof high yield varieties became more popular, the manual treading and beating methods spread through the provinces. Manual treading was carried out in more than 50 percent (5 out of 9) of the provinces surveyed in 1978 (Eriyatno, 1979). Manual treading was most common in South Kaliman- tan. Smashing was most common in North and South Sulawesi. The pedal thresher was not commonly used. The practice was found. only occasionally in Java. Power threshers were not used by the individual farmers; rather, they were usually found at the mills. Drying practices depend on the cutting and harvest- ing, threshing technique used. Traditional drying was done in the sun. With the local varieties and panicle cutting, stalk paddy drying was most common. In 1978, less than 10 percent of the farmers in most provinces were doing stalk paddy drying (Eriyatno, 1979). The stalk paddy drying of the indigenous varieties was not a major operation as the paddy was commonly left standing in the field for some time after maturity for natural drying prior to harvest (Esmay, et al., 1978). Rough rice sun drying was done on various floor types. Bamboo or wood floors were most common for more than 80 percent of the rough rice sun drying in Indonesia. In West Java, East Java and North Sulawesi, concrete dry- ing floors were common (Eriyatno, 1979). The handling techniques depend on the Operations that preceed and follow it. Ani-ani cutting may be followed by either in-field or out-field threshing. The 10 harvested crop may be bundled or not, depending on the following transportation method to be used. Sickle cut- ting was usually followed by transportation with simple bamboo racks. Manual treading was generally preceeded by bag transportation after sickle cutting. Neat trans- portation of harvested stalk paddy was not necessary for manual treading to get good threshing results. The overall systems of harvesting, handling, threshing and drying varied according to the marketing system. Indonesian rice farmers traditionally marketed their surplus rice in stalk paddy bundle form (Collier, et al., 1973, and Esmay, et al., 1978). 2.1.1 Panicle Harvesting (Ani-ani Harvesting) Ani-ani is an Indonesian term for a small paddy harvesting knife, 10 centimeters long fixed crosswise on a short wooden block. The knife was used for cutting each stalk paddy (panicle) separately, approximately 20 centimeters below the panicle. Five to six panicles were cut and retained in the cutting hand, then transferred to the other hand until a large bunch of about three kilograms was accumulated. The bunches were then placed on the ground. The cut panicles were placed in a basket if the harvester brought a basket while harvesting. 11 The harvesters were mostly women. The harvesting method utilized a large number of persons for cutting and carrying the stalk paddy. Esmay, Soemangat and Eriyatno (1977) found that a skilled person was able to harvest 10 to 15 kilograms of panicle stalk paddy per hour. Collier, Wiradi and Soentoro (1973) and Ban (1970) reported that ani—ani harvesting might employ as many as 500 persons per hectare. The most a harvester could cut was 110 kilograms of stalk paddy a day. The author's study on ani—ani harvesting in the Bekasi area found a low capacity of 0.002 hectare per man hour for the new high yielding variety of IR—38. The ani—ani harvesting capacity was affected by such factors as crop height, lodging and density. Other factors of precipitation, tempertature, time of day (morn— ing or afternoon), number of harvesters, soil condition and skill of the harvester also affected harvesting rate. In general, the new high yielding varieties were short with standing "flag leave" and a dense growth form, so ani-ani harvesting was not convenient (BULOG, 1976, and Collier, et al., 1973). Esmay, et al., (1978) stated that the use of ani-ani for the easier shattering varieties may increase losses. Grist (1975) indicated that one advantage of the ani—ani method was the high degree of selection possible 12 in excluding immature grains and impurities and in harvest- ing badly lodged plants. Collier, et al., (1973) observed that the ani-ani was quite suitable for cutting the local varieties of rice that mature irregularly and whose length of stalk varies. 2.1.2 Sickle Harvesting The sickle harvesting method was first used only in islands outside of Java. Some change from ani-ani to sickle harvesting in some Java provinces has been partly due to the use of new rice varieties. There are basically two types of sickles. One has a smooth edge and the other a serrated edge. The stalks are cut about 10 to 25 centimeters above the ground with the sickle. Japanese harvesters usually cut the stalk closer to the soil surface to provide longer straw for better rope making (Stout, 1966). The sickle harvesting rate varies by harvester, experience, plant condition and the environmental condi- tion. Wet soil and hot climatic conditions reduce the speed and performance of the harvesters. Plant lodging also reduces the cutting rate. Djojomartono, Kamaruddin and Syarief (1979) reported that the average capacity of sickle harvesting was 0.019 hectare per man-hour for the IR-38 rice variety. Esmay, et al., (1978) reported a similar capacity of 0.010 hectare per man—hour. Khan 13 (1976) reported the finding of Ezaki (1963, 1969) in which the harvesting output of a skilled man in nonlodged rice was 0.010 hectare per man-hour and that of a woman was 0.006 hectare per hour. 2.1.3 Foot Threshing Foot threshing was the traditional rice threshing method in Bekasi, West Java, Indonesia. Human feet were used to tread and rub the stalk paddy against each other and on the concrete floor or bamboo mat until most of the grain kernels were separated. Stalk paddy may come from either panicle or sickle harvesting. Foot threshing was usually carried out in the yard of the farmer's house. Stalk paddy was transported from the field to the thresh— ing site on bamboo racks or in plastic bags. The rate of foot threshing was affected mainly by the condition of the stalk paddy. Very easily shattering varieties require less time for threshing. Threshing is usually done by the same men or women who do cutting. Esmay, et al., (1978) estimated that foot treading had a low capacity of 30 to 40 kilograms per man-hour. BULOG (1976) indicated that 80 kilograms per hour was the maxi— mum threshing capacity for a woman. The average threshing rate in Bekasi area was 28.41 kilograms per man—hour when done by trampling an easy shattering variety, IR—38. 14 The rubbing action of threshing can be done in various ways. In Sri Langka, threshing has been done by animals or tractors (Ilangantileke, 1978). There were two other manual threshing methods besides foot threshing: one was beating the stalk paddy on the edge of tubs, threshing board or racks; two was flail threshing (BULOG, 1976 and Eriyatno, 1979). BULOG (1976) also reported that there were two kinds of beating threshing in North Sumatra. The threshing equipment was made of bamboo. The bamboo sticks were arranged in the form of a bed, 1 by 0.4 meters for the small type and 1 by 2 meters for the large type. Mats were placed below and on three sides of the bed to prevent the grain from scattering. The threshing capacity was approximately 200 kilograms per hour for three people with the small thresher and 600 kilograms per hour for six people with the large sized thresher. A flail method was commonly used in Krawang, West Java, when the paddy condition was not suitable for ani—ani harvesting. Sickle harvested stalk paddy was often threshed on a woven bam- boo mat by beating with a piece of wood about one meter long. The capacity and resulting quality were both lower than from foot threshing. The type of wood for beating varied by area. 15 2.1.4 Sun Drying Most paddy was sun dried in Indonesia. The dry— ing was generally done in the farmer's yard or at a drying site of the cooperation or local mill. Farmers have traditionally done only limited drying when marketing the harvested indigenous varieties in bundle form. The adop— tion of improved varieties made threshing possible imme— diately after the cutting operation and the drying of rough rice was then carried out. A Post Harvest Survey of CBS in 1978 indicated that no more than 10 percent of the farmers in each province were drying their paddy in the bundle form. Two exceptions were a 20 percent finding in East Java and 73 percent in South Kalimantan. Rough rice sun drying was commonly done on a woven bamboo mat, wood layer or concrete floor. Rough rice was spread at a depth of from 2 to 10 centimeters, depending on the volume of the rough rice to be dried. A 100 square meter floor area handled approximately one ton of rough rice at 2 centimeters. Drying was usually done between seven o'clock in the morning and five in the afternoon. The rough rice was stirred regularly to bring about even drying between layers. The sun drying rate was affected by layer thick- ness, grain moisture content, floor type, amount of radia— tion, air temperature and relative humidity. Measurements 16 by BULOG (1976) in the West Java province, indicated that from 4 to 8 hours were necessary for drying rough rice from a moisture content of 17 to 19 percent down to approximately 14 percent wet basis. Sumardi and Setiawati (1978) reported that 14 hours and 4.5 hours were necessary to dry 250 kilograms of IR—34 rough rice from 28-30 per— cent and 20—22 percent, respectively, to about 14 percent moisture content. 7.25 and 4.75 hours, respectively, were needed to dry 250 kilograms of bulu (an indigenous variety) from the same initial moisture content. The high sun drying rate at Bekasi area was an average of 1.5 percentage points per hour (Djojomartono, et al., 1979). IRRI (1974) reported that the use of an unperforated clear polyethylene sheet for the drying surface resulted in a longer drying time than for a surface of woven mats or concrete floor. Technically, good weather is necessary for success— ful sun drying. Stirring frequency and good judgement as to grain moisture content are also necessary for good drying. 2.2 Mechanical Methods Institutions and experimental stations, mostly from rice producing countries, have developed new methods of agricultural operations. The International Rice 17 Research Institute (IRRI) in the Philippines, for exam- ple, has produced some semi and fully mechanical methods for tillage, planting and post production operations (IRRI, 1969). Agricultural Engineering Research Insti— tues, experimental stations and private manufactures in Japan, have developed various types of mechanical tools and machinery for agriculture. They range from simple, manually operated tools to sophisticated, large machines. IRRI and Japanese designs have been introduced into many Asian countries. The Food and Agricultural Organization of the United Nations has cooperated in several agricultural mechanization projects with the International Institute of Tropical Agriculture, Ibadan, Nigeria; The Central Rice Research Institute, Cuttack, India; L'Institute de Recherches Agronomiques Tropicales et des Cultures Virieres, Bombey and Richard-Toll, Senegal; and the Agri- cultural University of Wageningen, the Netherlands (Unonimous, 1976). The Regional South East Asian Coop- erative Post Harvest Research and Development Programme in the Philippines is also carrying out extensive research and extension programs on post harvest technology and losses. 18 2.1.1 Manual Dropper The manual dropper consists of a pair of 20-30 centimeter long serrated sickle blades facing each other and set on the end of a one to one-half meter long wooden handle. A simple wire platform was attached to the handle to bunch the cut stalk paddy temporarily (Stout, 1966 and Ezaki, 1972). Figure 2.1 shows the manual dropper sche- matic diagram. The manual dropper operator pushes the blades against rice stalks in the cutting operation. The slicing and sawing action cuts the stalks close to the soil sur- face. This action is continued until a bunch of stalk paddy is accumulated on the platform. The bunch is then laid on the ground. The Operational capacity of the dropper is not much above sickle harvesting, but the operator can cut while standing. A better job than the traditional sickle can be expected. Ezaki (1972) reported that the capacity of the manual dropper was between 0.10 to 0.16 hectares per hour. 2.2.2 Mechanical Reaper The basic principle of the walking mechanical reaper is similar to the manual dropper. A Japanese reaper uses a planetary system to rotate a circular cutting blade. The paddy stalk was then simultaneously 19 Figure 2.1.——Manua1 Dropper Schematic Diagram transported to a side platform. Nine to sixteen bunches of stalk paddy were collected on the platform, and then the large bunch was automatically laid on the ground. A 3—hp two cycle engine provided power for both the cutting operation and forward movement. The operator walked behind the machine to direct the machine in the cutting operation. The reaper capacity depends on operator skill and control. The operator controls the cutting width and forward speed. The maximum cutting width of the walking mechanical reaper was 75 centimeters. The operational judgement had to be based on the plant and soil conditions. Djojomartono, et al., (1979) reported that the average capacity of a rotary type reaper under "optimum maturity" 20 and dried soil surface was 0.046 hectare per man—hour on IR-38. A certain plant height was necessary for proper operation of the reaper. A small modification of the bunching mechanism was possible. 2.2.3 Manual and Mechanical Binders Binders cut, handle and bind the stalk paddy in one operation. Mechanical binders usually used a recipro- cating cutter bar (mower type cutter). A plastic nail type pick up device worked better than a reel for gather- ing stalks. The cutting width varied from one row (25 centimeters) to three rows (75 centimeters). The manual binder was usually a smaller one row type. A cutting device similar to the manual dropper was used. The binding device was manually operated. The attachments were all set on a two wheeled frame for ease of handling. The capacity was double that of sickle cutting and ranged from 0.020 to 0.035 hectare per hour (Ezaki, 1972). The mechanical binder was usually powered by a 3.5 to 5 horsepower engine. The mechanical binder was usually supported by two or four wide tires for mobility in wet soils. The old binder design required a prelimi- nary field cutting prior to the binder cutting. The mechanical binder capacity varied with the operator's skill and plant and soil conditions. The average capacity 21 under "optimum maturity" and dry soil conditions was 0.046 hectare per hour for a three row binder on IR-38 includ- 1‘ q ing the time for field preparation and preliminary cut— .; ting. Ezaki (1972) reported that a mechanical binder can harvest 0.06 to 0.13 hectare per hour. 2.2.4 Pedal Thresher The foot pedal drum thresher was first developed in Japan for threshing the nonshattering Japanese variety in the field or at the building site after short stalk paddy drying on bamboo racks. The foot pedal drum thresher has also been used nearly exclusively in Taiwan since the 1950's (Esmay and Wu, 1975). The foot pedal thresher consists of a threshing drum, a foot pedal and accessories. Pedaling was done at a rate of 100 times per minute to provide drum rota— tion of 400 to 450 rpm (Niko, 1948). Figure 2.2 shows a pedal thresher schematic diagram. The impacting teeth on the threshing drum separate the grainfrom the panicle while the Stalk paddy is held firmly by hand. There were one-man and two—man pedal threshers. The operational procedure for the most common two-man pedal thresher was quite simple but must be followed closely. Three or four persons were necessary for con- tinued operation. The one—man units were generally 2 TMEBHEWS HEM Figure 2.2.——Peda1 Thresher Schematic Diagram unpopular. Some foot pedal threshers were produced locally in Indonesia. BULOG (1976) and Eriyatno (1979) reported several pedal threshers in Indonesia, mainly in Java. Many of them in West Java were locally made. The pedal threshers were limited to large farmers or cooperatives. Only 26 percent, 5 percent and 2 percent of the farmers surveyed in East, Middle and West Java respectively had threshed paddy with pedal threshers. IRRI (1968) tests indicated that the performance rate of foot pedal threshers was 68 kilograms per hour for a two-man team. A similar rate of 70 kilograms per hour was found in thresher tests carried out in West Java on IR—38 (Djojomartono, et al., 1979). The pedal thresher capacity was double or triple that of beating or treading. 23 Furthermore, the pedal drum thresher was simple and light weight, so that stalk paddy threshing could be done in the field after harvesting to minimize handling losses. 2.2.5 Power Thresher There are many types of power threshers. Khan (1976) stated that threshing paddy by treading them under tractor tires has been used in some Asian countries. The method was used occasionally in Sri Langka for custom threshing. The mechanically powered threshing cylinder is more typical. Threshing cylinders can be classified as for only panicles or the through type. For panicle threshing, the stalk paddy was held by hand over the rotating drum, while the through type takes the stalk paddy through the machine. The engine-driven drum threshers have been adapted from the pedal threshers in Taiwan (Esmay and Wu, 1975). Some large old through type threshers have been used at the mill level in Indonesia. Some cooperatives and large farmers utilized small powered threshers of the panicle type. These panicle threshing type threshers could be imported or locally manufactured. Recently, one manu- facturer in Middle Java introduced locally made IRRI designed through type powered drum threshers equipped with a cleaning mechanism. 24 The result of thresher tests conducted by IRRI in 1968, indicated that the performance of IRRI drum panicle threshing power threshers varied from 126.8 to 252.1 kilogram per hour depending on the feeding rate. A small, walking combine-harvester threshing capacity was 463 kilo— grams per hour for IR—38 (Djojomartono, et al., 1979). 2.2.6 Small Combine-Harvester A rice combine-harvester does cutting, feeding, threshing, cleaning and bagging simultaneously. The panicle threshing rice combine—harvester was initially designed as a walking type combine. This panicle thresh- ing rice combine-harvester consisted of a 50 centimeter width reciprocating cutter bar (mower type), a plastic nail type pick up device, chain stalk handling and a panicle threshing type thresher. The cutter bar, pick up device and threshing units were set in a frame and sup- ported on crawler tracks to provide mobility in wet paddy fields. There were 200,000 combines in Japan in 1973 (Esmay and Wu, 1975). Several improved riding type pani— cle threshing rice combine-harvesters have been designed in Japan and adapted for wet fields in Malaysia. The com- bines, manufactured in 1975, wereequipped with a 1.5 meter cutter bar and operated at an average rate of 0.200 hec- tars per hour. The capacity was affected by soil condi— tions (Ano., 1977). 25 One walking panicle threshing type rice combine- harvester with a 50 centimeter cutter bar width provided a harvesting capacity of 0.040 hectare per hour on IR-38. During the tests, the soil condition was dry, the paddy was at its optimum maturity with 18 percent grain mois- ture content and it yielded 5.028 tons per hectare. 2.2.7 Mechanical Drying Rice is a biological material with hygroscopic qualities. It must, therefore, be dried down to 13 or 14 percent wet basis to prevent spoilage during long period storage. Sun drying systems are dependent on weather and can be a problem during wet seasons. Mechanical grain drying can be done by blowing air through layers of grain with fans. Esmay, et al., (1978) classified the moisture reduction methods into four categories: (1) ultra low temperature, (2) low temperature, (3) high temperature and (4) ultra high tem— perature. Brooker, Bakker-Arkema and Hall (1974) gave a complete discussion of cereal drying principles and systems. De Padua (1976) also discussed the rice drying principles and systems. A thermal convection solar flat-bed rice dryer was tested in Thailand by Exell and Kornsakoo (1978). A solar air heater consisted of a layer of burned rice husk and a clear plastic sheet was used. The paddy reached 26 moisture content of 14 percent wet basis from about 20 percent in two or three days while protected from rain. The flat—bed type dryer commonly used in many Asian countries was small in size with a one ton bin. The fan to blow the oil—fired heated air through the grain was driven by a three or five horsepower engine. There were 873 batch type and 7 continuous type dryers located throughout Indonesia in 1972 (Weitz- Hettelzater, 1972). They were owned by millers or coop- eratives or BULOG. The result of a BULOG and IDRC survey project in 1976 indicated that the high operational cost, the need of high operational skill and difficulties of maintenance and repair were the main problems of using the mechanical dryer in Indonesia. The mechanical dryers were only used as supplemental dryers during the rainy seasons. The batch type dryer may take from 5 to 20 hours to dry a ton of rough rice, depending on the amount of water removed from the rough rice, climatic conditions, heated air temperature, and the rate of air flow through the grain (Esmay and Wu, 1975). The new design dryer with an improved holding bin construction was developed by IRRI and is now produced locally in Indonesia. A simple divider can separate several varieties of rough rice. The loading and unloading of the rough rice is easier, 27 and the cost of the bin construction is lower with the vertical bin than with the flat-bed bin with a perforated metal floor. The average temperature and moisture content reduction of three experiments using the new high yielding variety (IR-32), on IRRI designed rice hull flat-bed dryers is shown in Figure 2.3. The results indicated that the rough rice achieved about 12 percent wet basis moisture content from about 25 percent moisture content in ten hours. 28 28 50 0 f L A A l ) ; ‘. v- _ —‘ ‘ \‘ ‘ 5 2: ENE; 4 2'” it I A H \ A A \ §E mm ‘; \ f. A, i \ l \ [A '40 E Xf" 1i 0 A 2% / \ A . ’ \“ i E; . / I 1 20‘ / O O I 1 0 I O I E II A o .30 a A./§s o g 2: A C) A MOISTURE 8 \ CDN'I'ENT \ 15« \\ \ O \ It UPPER LAYER \ 2 ‘ \ g; -20 l I it \\ --___ LOWER LAYER 0 ‘4 12 . 1 0 2 4 6 8 10 DRYING TIME , HOURS Figure 2.3.--Grain Moisture Contents and Temperatures at Different Drying Times DEGREE CENTIGRADE TEMPERATURE , CHAPTER III FIELD STUDY Rice differs from other cereal crops in that the grains are normally cooked whole rather than processed into flour. The number one objective of the post rice production operations is to transform the standing mature rice paddy crop into the highest possible quality and quantity of head rice-—good whole polished rice kernels. Many different harvesting, threshing,hand1ing and drying methods have been used in attempts to achieve the objec— tive. However, numerous varying technical, biological, physical and environmental conditions contribute to the post rice production losses of quantity and quality. Post rice production operation losses include: (1) mechanical shattering and broken kernel losses from the mechanical operations of cutting, handling, threshing and drying; (2) biological losses of over or under matur- ity, lodging; (3) rodent, bird, insect and micro—organism losses; (4) environmental losses caused by extreme tem- peratures, humidities, winds and sun light. Measurements of post rice production losses were carried out in the field using several types of tools and 29 30 machines at several crop maturity levels. The study con- centrated on harvesting, threshing, handling and tempor- ary storage Operations. The field research was carried out in a rice producing area in West Java, Indonesia in 1978, on two high yielding rice varieties, IR-36 and IR-38. The results of the field research along with avail- able reliable secondary data were employed as a base in formulating the simulation model in Chapter V. 3.1 Methodology 3.1.1 Harvesting Losses Three randomly located plots, each one hundred square meters in area were used for ani-ani and sickle harvesting on scheduled harvesting dates. Three larger plots ranging from four hundred to eight hundred square meters in area were randomly chosen for the rice reaper, rice binder and walking rice combine—harvester operations. Five one square meter plots were randomly chosen within the harvested plots to measure the grains left after the harvesting operations. These were collected and referred to as harvesting losses when converted to kilograms per hectare at 14 percent moisture content. 3.1.2 Packing and Trans- portatiOn Losses The harvested stalk paddy were packed and trans— ported using bamboo racks, plastic bags or bamboo baskets 31 to the threshing location, 600 meters away from the field. Plastic sheets were laid under the bamboo rack while pack- ing. The bamboo rack was covered by a plastic sheet dur- ing transportation. The amount of shattered grains on the sheet were counted and converted to kilograms per hectare to provide packing and transportation losses. 3.1.3 Threshing Losses Stalk paddy harvested by ani—ani and sickle were then threshed using manual foot threshing or a pedal thresher. The stalk paddy harvested using mechanical harvesters were threshed using a power thresher from the walking rice combine—harvester. The grain which was blown off, husked or left unthreshed in the straw after the threshing operation was referred to as threshing losses. These were collected and converted to kilograms per hectare. The threshing opera- tions were done only at one maturity level, assumed as the optimum maturity level judged by the farmers. 3.1.4 Temporary Storage Losses A portion of the harvested stalk paddy and threshed rough rice were used for temporary storage study. Harvested stalk paddy and threshed rough rice were left fresh and undried inside a storage room for five days, to imitate the delay of threshing and drying operation. 32 Spoiled grain was examined every day and judged from the outside appearance as to losses of stalk paddy and rough rice from temporary storage. 3.2 Results 3.2.1 Harvesting Losses Regression equations were fitted to field data collected on the harvesting operation for the IR-36 rice variety. Losses due to ani-ani and sickle harvesting operations were used as dependent variables against the number of days before or after the "optimum maturity" date judged by the farmers as the independent varible. Equation 3.1 illustrates the predictive equation fitted to the field data of ani-ani harvesting. PHL = 57.7 + 0.16*ND 3.1 Equation 3.1 had a r value of 0.93. Equation 3.2, with a r value of 0.94 represents the prediction equation of sickle harvesting losses. SHL = 52.14 + 1.37*ND 3.2 In equations 3.1 and 3.2, PHL = the predicted ani—ani harvesting losses in kg/ha 33 SHL = the predicted sickle harvesting losses in kg/ha ND = the independent harvesting time variable, number of days before or after the opti- mum harvesting time. Regression curves for ani—ani and sickle harvest— ing operations varying with the number of days before and after the optimum harvesting date are illustrated in Figure 3.1. The mechanical rice reaper, rice binder and rice combine-harvester could not be operated properly on the IR—36 rice variety, because of the plant's short height to the canopy of 77.5 centimeter. The IR—38 rice variety with an average height of 20 centimeter more than the IR—36, was better adapted to the Japanese mechanical rice reaper, rice binder and walking rice combine—harvester. The grain shattering losses after harvesting IR-38 rice by ani-ani, sickle, rice reaper, rice binder and rice combine-harvester are shown in Table 3.1. The measurements were made at the "Optimum maturity" level, with a grain moisture content of 18 percent and dry soil condition. 3.2.2 Packing and Transpor- tation Losses The packing and transportation losses with bamboo racks for IR-36 varied with the harvesting date, before ‘1 (I ' Eff—1 34 A ANIHMH HMAESTHTE 0 SICKLE HARVESTING 70.0 0 § \. . / c x /’ . / U; / m / u: / § / 60.0 // " w. E PHL: 57.7+O.l6*NP_____.———// 0 U) ‘ ”/k / é // A // E / o / / / ,F/ SHL: 52.14+l.37*ND / / 1| 50.0 x /‘ / / l 0 / / L / / -4 -2 O 2 4 6 8 10 HARVESTING TIME, DAYS BEFORE OR AFTER OPTIMUM DATE Figure 3.l.—-Harvesting Losses at Different Harvesting Times (IR—36) 35 TABLE 3.l.--Grain Losses for Different Harvesting Methods and IR-38 Losses Harvesting Methods Kg per ha Percent Ani-ani Harvesting 158.8 Sickle Harvesting 135.1 Mechanical Rice Reaper 111.6 Mechanical Rice Binder 100.5 Rice Combine Harvester 121.1 2.4 and after the "optimum maturity" as depicted in Figure 3.2. The total amount of packing and transportation losses referred as to handling losses was used as a dependent variable against the independent harvest date variable in obtaining the fitted prediction equation: BRTL = 7.82 + 0.54*ND 3.3 where: BRTL = the predicted handling losses for 600 meters, kg/ha ND = the number of days before or after the optimum harvesting time. The handling losses equation had a high coeffi— cient of determination of 0.98. Packing and transporta- tion of stalk paddy with plastic bags or bamboo baskets 36 BRTD= 7.82+0.54*ND ’//’/ ES. QTQTEPORT ioss / Co 90 PA ll”, , / I / I), ,I I I I IIIG' I/,// I ,I , III/IN l// I III] I [I ,I I’ll/I I I/’/,I ,I I l / / I ll, / I I I I / I / I” I I I I [I l]/’ Ill/I’llll I” //’I I/ ,I’ , ,I //I I // I’ I I I/,I /’/’/ I/,I / I I I] I I / I I I ,/A 12.0 10.0 0 C I 6 Wm? co m\4:\0& .mmmeH gag/E: 0 0 -2 EM“ESTD«STIME,ERXS BETTE ORZEYTR.OPEU§E4I$TE -4 -6 3.2.--Hand1ing Losses for Different Harvesting (IR—36) Times Figure 37 were assumed to avoid the shattering losses, and measure- ments were not made on these cases. 3.2.3 Losses During Threshing The amount and percentage of unthreshed grain and husked grain losses after threshing operations on IR-36 and IR-38 rice varieties are shown in Table 3.2. The measurements were done on stalk paddy harvested at optimum harvesting time. A power thresher could not be used to thresh the IR—36 stalk paddy properly due to the short stems . TABLE 3.2.-—Unthreshed and Husked Grain Losses for Differ- ent Threshing Methods IR—36 Variety IR-38 Variety Threshing Methods Kg per ha Percent Kg per ha Percent Unthreshed Grain Foot Threshing 17.5 0.6 22.9 Pedal Threshing 143.2 4.8 29.2 Power Threshing * * 0.0 0.0 Husked Grain Foot Threshing 2.6 0.1 0.0 0.0 Pedal Threshing 6.8 0.2 0.0 Power Threshing * * 6.3 0. 38 3.2.4 Temporary Storage Losses The percentage of deteriorated grain losses after stalk paddy and rough rice temporary storages were used as the dependent variable against the independent variable of the number of temporary storage days. Regression equations were used to fit the data obtained. Equation 3.4 represented the model used to fit the predictive equation to the collected data on stalk paddy temporary storage. SPTSL = —2.43 + 8.47*NDTS 3.4 where: SPTSL = the predictive percentage of deteri- orated grain losses, percent. NDTS = the independent number of temporary storage days variable, day. had a r (coefficient of determination) value of 0.97. Equation 3.5 is the predictive equation for rough rice temporary storage losses, RRTSL = 5.5 + l4.73*LN(NDTS) 3.5 where: RRTSL = the predictive percentage of deteri— orated grain losses, percent 39 NDTS = the independent number of temporary storage day variable, day The coefficient of determination for equation 3.5 was 0.92. The curves of the predictive equations represent- ing the variations in percentage of spoiled grain losses to the independent number of temporary storage day varia- ble NDTS, are illustrated in Figure 3.3. 3.3 Discussion Shattering losses after harvesting as well as after handling operations showed an increase after optimum harvesting time and a decrease before the optimum harvest— ing date. Rice paddy plants continue their biological process until maturity. The grain becomes more suscepti- ble to shattering at harvesting and handling operations as it matures and drys. At optimum harvesting time, IR—36 hadlower shattering losses after ani—ani and sickle harvesting than IR—38. This difference in shattering losses reflected the behavior of varietal characteris- tics. The percentage of unthreshed and husked grains varied with the methods of threshing. Table 3.2 illus- trated the influence of the threshing method on the per- centage of unthreshed and husked grains. Foot threshing 40 50.0 0 STALK PADDY TEMPORARY STORAGE m 0 ROUGH RICE TEMPORARY STORAGE c/ E 40.0 E m 33‘ SPTSL: -2.43+8.47*NDT U) m 30.0 S g g 20.0 6 E s 10.0 0 // RRI‘SL= 5.5+l4.73*LN(NDTS) 0.0 O l 2 3 4 5 TEMPORARY STORAGE TIME, DAYS Figure 3.3.—-Deteriorated Grain Losses in Temporary Storages (IR-36) 41 with a rubbing action for threshing the grain from the straw showed a low tendency to produce husked grains. The high speed rotating drum teeth from the power thresher, on the other hand, caused husked grain losses. The unthreshed grain losses may result from improper feeding of stalk paddy. Deteriorated grain losses due to delay in thresh— ing of stalk paddy or delay in drying of rough rice were seen to increase with the increase in the number of days in temporary storage. High moisture grain will ferment, germinate and generally deteriorate in quality, if it is not stored under proper conditions. CHAPTER IV FEASIBILITY ANALYSIS The systems approach consists of five major phases (Manetsch and Park, 1976): (1) Feasibility Evaluation, (2) Abstract Modeling, (3) Implementation Design, (4) Implementation and (5) Operation. The feasibility evaluation generates a set of viable alternative solutions capable of satisfying the needs of the people for whom a development program is being designed. The first step of a feasibility evalua- tion is the needs analysis. Based upon the current situa- tion, an analysis of the needs was made to determine the real need (Asimow, 1962). Having established the real needs, the next step is system identification. This step includes the system boundary, output desired to meet the real need, input necessary to accomplish desired outputs, inputs from environment, undesired outputs resulting from the inputs, and system parameters. After the real needs have been stated and the system has been identified, an explicit statement of the 42 43 problem can be made. Included in the problem formulation is a statement of the criteria for evaluating alterna- tive solutions. Generation of a broad range of alternative solu- tions follows the problem formulation. These alternative solutions should then be analyzed to determine those that do not seem to be feasible with current knowledge. The set of feasible solutions were then studied in more detail in an effort to predict which systems have good potential for reduction to practice. A simulation was developed to aid in this study. 4.1 Needs Analysis The traditional post rice production techniques in Indonesia have evolved through the years as a result of climatic, economic, social and cultural factors. Evo- lution often provides an appropriate technique for the prevailing conditions. Thus, the traditional methods and techniques were probably most appropriate for the indigenous varieties. But once the balance of traditional operations is disturbed, changes must follow. The BIMAS program has introduced new high yield— ing varieties which, along with other cultural practices, increased the rice yields. Besides the high yielding Characteristics, the new grain varieties have a much shorter growing season and are more susceptible to 44 shattering losses than the indigenous varieties. With the proper mix of resource inputs, the new grain varieties increase the opportunity for multiple cropping. The harvesting season then, often has to be handled in the wet season. Traditional sundrying is more difficult if not impossible and increases quality losses. The increased quantity of rice requires more harvesting capac- ity than the traditional method to harvest the paddy dur- ing its optimum harvesting time and reduce excessive losses. As described in Chapter III, shattering losses of new rice varieties increased significantly beyond its optimum maturity. An imbalance between rice production and post production factors was created by the focus only on rice production practices through the BIMAS program. New varieties may justify some new techniques or methods. Alternative post rice production systems may be appro- priate for the new higher yielding varieties to reduce excessive losses and thus, to save the increased rice yields for marketing and consumption. Esmay (1977) stated that the limited effort pre— viously directed towards improving post production opera— tions has been scattered piecemeal and ineffectively. Isolated attempts have been made from time to time in different countries to improve specific tools, machines 45 or methods without consideration of the total post pro- duction phase. In other cases attempts have been made to import modern technology directly from the industrialized countries. In most cases, appropriate attention has not been given to minimizing losses with appropriate technol- ogy that is labor intensive, simple and locally made. The evidence indicated a need for an improved post rice production system for Indonesia to minimize losses and increase total rice production. 4.2 System Identification Figure 4.1 illustrates a system schematically. The post rice production system included the field opera- tions involved in transforming a standing rice crop into a storable commodity for market and consumption. The operations involved were cutting, bunching, transporting, threshing and drying. The system constraints were: a. One new high yeilding variety was used b. One geographical region was considered. The total simulated area was assumed to be within the one geographical region and with the same climatic conditions c. The total acreage was assumed to be well irrigated and with good drainage. A good dry soil condition was always assumed for the day before the first day of harvesting 46 Emummm m mo Emummflo oeuwfiwaomll.a.v musmflm if V L Empm>m p2m§mozm_ Sahm>m 47 d. The one main objective in the post rice production operation was to achieve maximum production and minimum losses Employment and financial evaluations were imposed when feasible, and practical alternative systems in reduc- ing losses were evaluated. Based upon the needs and identification of the system, the following were considered: Desired Outputs: outputs, inputs, and constraints a. Minimize quantity and quality losses as measured in individual system (kg/ha b. Increase rice total tonnage or kg/ha) c. Improve grain operations and/or in the overall or %) to the consumer as measured in and/or yield per crop area (kg/yr quality as measured by percentage of head, broken and cracked kernels and foreign material (%) d. Minimize operation costs and improve the pro- ducers profit position (rupiah/system/hr or rupiah/system/yr) Environmental Input Variables: a. Climatic and weather variables b. Prices of post rice production tools and machines 48 c. Government policies, i.e., rice price and grading d. Wage and interest rate e. Fuel and lubrication prices Controllable Inputs: a. Operations rate (ha/hr) b. Working hours (hr/day) c. Post rice production systems or combinations of various types of harvesting, handling, threshing and drying. Uncontrollable Inputs: a. Total acreage (ha/season) b. Production practices; cropping pattern, rice variety, planting date and harvest date (number of harvest seasons/yr) c. Crop condition at harvest time, e.g., moisture content and lodging d. Credit availability e. Rural technical level System Design Parameters: a. Performance characteristics of post production technologies and machines b. Labor input (number of people per operation) c. Planning horizon (yrs) All of the outputs and inputs listed were con— sidered to be a part of the system under study; however, 49 some were not included in the simulation model. Justifi- cation for omitting these factors (i.e., credit avail- ability and rural technical level) was based upon the lack of available data to include a meaningful considera- tion, and that for the purpose of this study, an inordi— nate amount of time would have been required to develop the needed data. 4.3 Problem Statement Evaluate traditional and alternative methods for post rice production in Indonesia under alternative crop— ping patterns and farm sizes, to: a. Minimize qualitative and quantitative losses b. Increase consumable grain production in quan— titative and qualitative terms c. Maintain labor utilization at an acceptable level d. Improve financial feasibility e. Provide technical and cultural feasibility 4.4 Measure of System Performance The main objective of improving the post rice pro- duction operations was to maximize the quantity and qual- ity of rice production by the minimization of losses. The performance of alternative post rice production methods were compared to the traditional methods. 50 Negative incremental performance in terms of production was not desired. Similarly, positive incremental per- formance in terms of losses was not desired. Accumulation of some desired outputs was consid— ered necessary to evaluate a system which might show effectiveness in the long run. Evaluations of perform— ance for a post production system where machines were involved were made at five year periods. The cost—benefit ratio for incremental production through alternative post rice production operation methods was compared to the performance of traditional operations under given conditions. 4.5 Alternatives of System Concepts (Solutions) The types of changes which might be made in post rice production techniques to bring about the desired outputs were classified as follows: a. Changes in harvest practices only b. Changes in threshing practices only c. Changes in drying practices only d. Changes in harvesting and threshing prac- tices e. Changes in harvesting and drying practices f. Changes in threshing and drying practices g. Changes in harvesting, threshing and drying practices 51 The traditional post rice production method in the Bekasi area, West Java Indonesia was: -—Harvesting: panicle harvesting with ani-ani, unbundled short stalk paddy --Handling: bamboo basket handling --Threshing: manual foot threshing (treading threshing) of wet stalk paddy —-Drying: Sun drying Cleaning was usually done after drying. The sun drying operation was to dry rough rice to 14 percent moisture content wet basis. The possible changes of post rice production operation are shown schematically in Figure 4.2. The walking rice combine-harvester was also considered as one of the post rice production alternatives. There are 23 possible post rice production opera- tion combinations resulting from the possible changes of harvesting, handling, threshing or drying operations. However, only twelve combinations, including traditional post rice production operation, were studied using the simulation model. These twelve combinations are shown on page 53. 4.6 Experimental Design The twelve combinations of post rice production Operation alternatives were then studied in more detail 52 COADOSUOHm moam #mom wanflmmom Had mo Emummflo ofiumEonomll.N.v wusmflm are mzozczcoo are .3 8x: 58:8 Eom 02;”:— . mmmcmnv mcoflwmummo Emoimcaczsoo BE 55o mzozczcoo is an ax: 850:8 Eom ©2235! ozEmmEE l @2250 .5585 also; 328:: Even. oz_>~_a 3mm ezmma 8.03. 08:13 ozEmmEE . I em=:_m saga SE28: e885 =5:me Exam Em: 62.33.: . A oz :50 A 53 .mQOADMHmmo cofluospoum woflu “mom HMGOflpHUdHB ¥ Amman Hmwmm>HMQlwanEoo o0 NH cam Hmwmm>umnlmcflnaoo mo Ha “whuo Hwaom mxomm OHMOAm QBEMmmm ca Howuo Hmpom II maxoflw oemmzmm m “chum uoom mxomm maxoflm oemmmmm w cam Hw3om mxomm OHROHm mazmmmm n sum Hmpom II waxoflm mammzmm m cam uoom mxowm waxoflw memmmmm m Momma Hm3om pwxmmm wHOHsmm Dazmmmm w Hmwuo uoom poxmmm waoflcmm 08mmmmm m csm Hm3om uwxmmm 0H0flzmm mBzmmmm N cum Doom uwxwmm QHOHsmm «memmmmm H maflmnm mcfinmoune mcflapcmm mcflumm>umm @000 .oz mcoflumuflnfioo soflumummo 54 to predict which systems were most practical and had the greatest potential for reduction of losses. The four con- straints described in system identification were applied in all simulations. The variables that reflect the changes in farming practices are: a. i. ii. Multiple cropping schedule: two crops per year twelve crops per five years. It was assumed, in this cropping schedule, the harvesting seasons were in March and August in the first year, in January, June and November in the second year, in April and September in the third year, in February, July and December in the fourth year and in May and October in the fifth year. The following five years would have the same harvesting schedule The harvesting schedule for both cropping patterns is illustrated in Figure 4.3. b. Seven different farm sizes: i. ii. iii. iv. 1 hectare per crop 2 hectares per crop 3 hectares per crop 4 hectares per crop 55 me5©wsom mcflwmm>umm paw mcumwumm mcflmmouollm.w wusmam GZHBmfléfiflmu flHHHHHH % mqbommom m¢mw mmm mmOmU 03H 56 v. 5 hectare per crop vi. 7.5 hectare per crop vii. 10 hectare per crop The shorter growing season of the higher yielding varieties and their ability to mature with less regard to day length increased the opportunity for multiple cropping. Esmay (1973) stated that an equivalent of 50 kilograms of rough rice per hectare is lost each day the land is idle with 120—day rice capable of yielding five tons per hectare. One of the BIMAS program objectives for high production was to maximize the number of crops per year. Harvesting was then necessary during two or three seasons per year. The farm size determination was based upon the average size of the individual farmers and the financial condition of cooperative organizations. Individual farm sizes ranged from one to five hectares. Larger farms, 7.5 hectares and 10 hectares were hypothesized as at the cooperative level (BUUD). CHAPTER V THE SIMULATION MODEL A computer simulation model was formulated to simulate the various combinations of post rice production operations. The simulation approach was chosen because of the system complexity and the stochastic aspect of the post rice production operations (Holtman, et al., 1970). Simulation models may be formulated in either the discrete or continuous time form. The choice between the continuous or discrete time model depend upon: (1) the level of detail necessary to answer relevant questions, (2) the frequency of events or the flow rate of objects relative to the mimimum time interval, and (3) the cost of programming and operating the models (Manetsch and Park, 1974). The outputs should provide all seasonal, annual and planning period information. After considering all of the input information, the discrete time form was selected with a simulation time increment of one day. Submodels and components of the post rice production model are described in the following sections. 57 58 5.1 Weather Model The weather or climatic condition is described by numerous parameters such as: rainfall amount and duration, barometric pressure, dry-bulb and wet-bulb temperatures, evaporation, solar radiation, wind speed and direction. The rainfall, temperatures, radiation and evaporation parameters are critical to harvesting, handl- ing, threshing and drying operations. Rain, nonrain days and the rainfall amount determine the work or no work status of the in—field post rice production operation of harvesting, in—field threshing and sun drying. Tempera- ture, radiation and evaporation directly affect the sun drying operation (i.e., the drying period), and the soil moisture status which in turn affects the capability of various in-field post production operations. Temperature, radiation and evaporation parameters, however, were not included in this study. The judgement for omitting these factors was based upon the lack of available data to appropriately account for the effects of these factors on operational performances. The output of weather--the rainfall occurrences and rainfall amount-—were used as an input to the in-field post rice production operations in the overall model. The weather component of the "environment" for the post rice production activities can be obtained in two ways: 59 One is by using historical weather records. This method provides an exact replication of historical occurrences, but in a series of limited length and requires a large number of cards and long computation time. Two is a method that generates daily weather factors for larger time series than available from historical records. This second method uses historical records for the base of probability occurrence, and then stochastically generates the parameters. The method is thus called the stochastic weather simulator. 5.1.1 Weather Data Fifty year weather records for rainfall, tempera- tures (mean, maximum and minimum), pressure, sunshine hours, relative humidity and wind (velocity and direction) were available from 52 weather stations in Java. Smith (1973) felt that data before 1939-40 were probably more reliable than afterwards. The effect of the Second World War and subsequent political disturbances in Indonesia resulted in the loss of considerable meteorogical data. There were 20 meteorogical stations in West Java, but none was located in the Bekasi area. The closest station to the Bekasi area was in Jakarta. It was located at an altitude of seven meters, which was similar to the Bekasi area. Eighteen years (1959—1976) of Jakarta 60 weather data were then used to stochastically generate weather factors representative of the Bekasi area. 5.1.2 Weather Model Development Considerable research has been reported on weather parameters. Some reports included evaluations for only one typical location, while others were for many loca— tions and several weather parameters. Sorensen (1967) constructed a generalized weather model to simulate regions with gradient extremes of annual rainfall from five to fifty inches for any given year. Iowa rainfall probabilities were evaluated and used in an agricultural production analysis by Link (1968). Jones (1970) developed a weather simulation model to use the Markov Chain method to simulate daily rainfall, tempera— ture and evaporation for Mississippi. A spatial corre- lation between the simulated weather in the Manapla and Victorias areas in the Philippines was maintained with the weather simulation model developed by Panol (1973). In Panol's study, one day lag auto-correlation among the weather variables was also used. Dumont and Boyce (1974) described methods of simulating five weather variables for any location for which data was available. Oldeman (1975) evaluated the climatic condition and constructed an agro-climatic map of Java. 61 5.1.3 Rainfall Model A day was defined as wet in this study if 0.1 millimeter or more rainfall occurred during that day. A probability of rainfall was described as a probabilistic process. It was assumed that the conditional probability of rainfall at the ith day was determined by the weather sequence until the (i-l) days. Let: FDi = event that was sequentially dry for i days DD. = event that (i+l) day was dry RDi = event that was sequentially wet for i days WDi+l = event that (i+l) day was wet P(DDi+l/FDi) = conditional probability that (i+l) day was dry, given i sequential dry days P(WDi+l/RDi) = conditional probability that (i+l) day was wet, given i sequential wet days. FDirlDDi+ 1 was the event that (i+l) days were sequen— tially dry and therefore, 0 = , FDi DDi+l FDi+l 5 l RDir')WDi+l was the event that (i+l) days were sequen— tially wet and therefore, 62 From equation 5.1 and 5.2, the conditional probability P(DDi+l/FDi) and P(WDi+l/RDi) are: r) P(FDi DDi+ P(FD. ) 1) _ _ 1+1 P(DDi+l/FDi) - p(FDi) - P(FDiT 5.3 P RD.r\W . P . P(WD /RD ) = ( l D1+l) = (RD1+1) 5 4 i+l i P(RDi) P(RDi) ' P(FDi+l), P(FDi), P(RDi+l) and P(RDi) were obtained from historical records of daily rainfall. The conditional probability, then, was calculated by using equations 5.3 and 5.4. The probability of a day being wet was deter- mined by P(DDi+l/FDi), P(WDi+l/RDi) and a random number generated using program RN = RANF(—l). The process is depicted in Figure 5.1. The rainfall distribution during a given wet day was described by a method with two random numbers. The total rainfall for each ten day period was divided into ten categories. A cumulative probability distribution was made for each ten day period from the eighteen year weather records. Figure 5.2 shows the cumulative proba- bility distribution of three ten day periods in March. The first random number determines the rainfall category 63 IDAY=IDAY+1 DRY INCREMENT OF SEQUENTIAL DRY DAY I=I+l IDAY=IDAY+1 DRY Figure 5.l.——Rainfa11 Model Flow Chart INCREMENT OF SEQUENIIAL WET DAY I=I+l 100‘ E80 [L] D4 5 6 o E E 2 ‘3‘ 40. E ! U o 64 _ ._‘___ , —.:_"_. _..—_:;‘—_--.=-—O ../"" / 1‘ / / €;/ Ill 0"_0 FIRST 10 DAYS OF MAKE-I ._-... SECOND 10 DAYS OF MARCH A—-—A THIRD 10 DAYS OF MARCH 20.0 40.0 60.0 80.0 100.0 RAINF‘ALL AMOUNT, MID/ETERS Figure 5.2.--Cumulative Probability of Rainfall Amount in March 65 of the distribution, and the second one defines the spe- cific rainfall amount. The process is shown in Figure 5.3. 5.2 Working Hours Model The maximum working hours per day for in—field post rice production operations was assumed as eight hours. The drying operation using a mechanical dryer can be over eight hours, and its maximum can be set by chang- ing the value of a parameter from the overall model. Threshing wet paddy directly after the cutting operation was a common threshing practice in the Bekasi area, West Java, especially when the foot threshing method was used. The persons who were doing the threshing opera— tion were the persons who cut the paddy. Only two hours or less of available working hours per day were assigned to the harvesting operation in the working hours model. If more than two working hours were available on a par— ticular day, and it was not a combine harvesting method, the available working day was divided into two similar hours. One half of the available working hour was assigned to harvesting and handling operations and another to the threshing operation. The available working hours per day were defined by the rainfall amount on a particular day, between six RANDOM NUMBER RN=RANF (-1) FIND THE RAINFAIL CATAGORY (ONE TO TEN) RANmM NUMBER RN=RANF(- CONVERT CATAGORY TO RAINFML ANDUN'I‘ RAINFALL AIVDUNT Figure 5.3.--Rainfall Amount Model Flow Diagram 67 o'clock in the morning and six o'clock in the evening and the type of in-field operation. The relationship of the working hours and the rain— fall amount was assumed as shown in Figure 5.4. The available working hours for the combine—harvester on rain days were assumed to be one—half that of other operations. 5.3 Field Condition Model The field condition which greatly affects post rice production operations was described by the top soil moisture status. Wet soil or standing water reduces the tractability of harvesting machines. The soil condition also affects traditional or manual harvesting operations. Standing water or wet soil is critical for sickle harvest- ing because: (1) the cut grain will draw moisture from the wet soil; (2) the mobility of the harvester is reduced under wet soil condition whether using a sickle or ani- ani. Observation during the course of this study in the Bekasi area showed that harvesters prefer doing the cut- ting job on dry soil rather than on wet soil. There was a tendency for the harvesters to leave panicle uncut on wet and badly lodged paddy fields. Many soil moisture studies have been done for areas where complex machines are used. The moisture dis— tribution in the soil surface layer as well as soil type 68 8 U: E Q N 6‘ 5 TI: (5 4‘ E E 21 0 f '7 0.0 5.0 10.0 15.0 RAHTNEI. AMIEE, MILLDflflERS Figure 5.4.—-Re1ationship Between Working Hours and Rainfall Amount 69 determines the machine's tractability. Shaw (1963), and Baier and Robertson (1966) developed a soil moisture budget model which considered soil type, evaporation, evapotranspiration, rainfall and run off. Shaw (1965) developed and tested a work day model which transformed climatic observations into work day sequences. Frisby (1970) described a technique for predicting the number of good days available for primary tillage in the spring and fall for soil types in Central Missouri. Tulu, et al., (1974) developed models to analyze timeliness costs for corn production. Research data on soil moisture in Indonesia was too limited for a meaningful analysis. It was therefore decided not to develop a model of the effect of climatic conditions and soil moisture status on in—field post rice production operations. Instead, a simpler model was developed to estimate a working day base upon sequential rainfall occurrences for several days. This model was considered adequate for predicting the field condition resulting from excessive rainfall and its effects on the performance of mobile type machines (i.e., combine- harvester). 5.3.l Field Moisture Model The soil moisture model estimated the soil mois— ture status on particular nonrain days, preceeded by wet days. The model assumptions were: 70 l. A non—rain day for a well-drained field was considered as a working day when less than 10.0 millimeters of rainfall on the previous day. 2. A rain day with a precipitation of less than 10.0 millimeters was combined with the capability of evaporating 50 percent of the rainfall in excess of 10.0 milli- meters on the previous day. 3. A nonrain day was defined as a work day for the combine-harvester when there was less than or equal to 10.0 rainfall in excess on the previous day with less or equal to 10.0 millimeters precipitation. Up to 20.0 millimeters of rainfall may be in excess on a work day before it affects the manual harvesting rate. More rain in excess of 20.0 millimeter would reduce the harvesting rate by 20 percent. 4. One nonrain day is necessary to evaporate the excess rainfall resulting from 10.0 millimeter on the previous day. The model was used to check if a nonrain day was a work day based upon the soil moisture status. The model based 71 its decisions on rainfall and climatic conditions of up to three days before the particular day of concern. The model flow chart is shown by Figure 5.5. 5.4 Plant Condition Model Robertson and Weille (1973) reported that rice yields varied from farm to farm and from country to coun- try. The lowest yields were about 1400 to 2000 kilograms per hectare in the developing countries where traditional production practices were used, and the high yields ranged from 4500 to 5300 kilograms per hectare with high yield- ing varieties and optimum crop inputs and management. The yield variation was a result of interactions and interrelations of many factors such as weather, variety, fertility of soil, farm management practices and preva- lence of diseases and insects. Among these factors, the weather parameters were the most significant contributing factors. Owen (1971) reviewedstudies on the effects of temperature on the growth and development of rice. He concluded that adverse temperature, acting either alone or through the interaction of day length and other fac- tors from sowing to floral initiation, largely determine the maximum potential yield. Rainfall amount and dis- tribution were also found critical for rainfed rice 72 VDRKINGIKXHESDDDEL ERFLA(2) =R-EL-5-(3’LmM2) .___J \J Figure 5.5.—-Soil Moisture Status Flow Chart (Part 1) 73 NO WEE GEF.=.8 Figure 5.5.——Soi1 Moisture Status Flow Chart (Part 2) 74 production. IRRI (1969) found that grain yields were reduced by inadequate moisture during the early stage of crop growth. A reduction in grain yield was not found when the reproductive and ripening stages of the paddy growth were preceded by continuous flooding. IRRI (1967) also found that solar radiation during the last 30 to 45 days before harvest was highly correlated with grain yields. Murata (1975) simulated the rice yield from climatic factors, solar radiation, sunshine hours and temperature. Nitrogen application was another example. IRRI (1970) data indicated that the nitrogen absorbed at flowering was closely related to the grain yield of improved varieties. Three models were necessary to adequately reflect the plant conditions during harvest. The first model determined the potential yield on the season. The second model took into account the yielding or maturing time. The second model did not allow all of the paddy to enter the maturity stage at the same time. The planting sched- ule along with weather factors (i.e., temperature) influ— ence the maturity time (IRRI, 1975). The third model consisted of a crop moisture reduction model. Three types of moisture reduction take place in the rice crop. First, the grain loses moisture in the field while attached to the paddy plant; second, the stalk paddy loses moisture 75 after harvest; and third, grain loses moisture after it is threshed. Many researchers have worked on these prob- lems. Philips and O'Callaghan (1974) modified the wheat moisture content vs. days since ear emergence, from Geslin and Jordan's data, to identify the wheat moisture content when harvested. The relation between climatological variables and field moisture of corn was modeled by Schmidt and Hallauner (1966). Ano. (1976) used the days after 50 percent heading as the characteristic which determines the stage of maturity rather than the moisture content. The grain moisture varies within the day and between days, especially during the wet season. A yield model and a rice maturing time model were developed for this study. From Anonymous (1976), the optimum harvesting date was used as the determinant for losses, rather than as a function of moisture content. 5.4.1 Crop Yield Model The yield model was built to describe the rela— tion between the effect of the tillage season's weather condition and accomplishment of the harvesting operation in the previous season to the next season yield. Mathe— mathically, the yield of the next season was described as: = * OYt+l AOY YLDCt+l 5.5 76 YLDCt+l = f(SHDAR, AREA, STTLC, TTILD) 5.6 where, OY = potential yield at season (t+1), ton per hectare AOY = average potential yield, ton per hectare YLDC = yield coefficient for season (t+l) SHDAR = harvested area at previous season (t) STTLC = standard tillage capacity, for particu- lar method, ha/hr TTILD = total tillage working day for previous season (t), days. Available tillage working days were determined by the total tillage working days and the weather conditions during the period. The area prepared for planting was determined by the standard tillage capacity and available tillage working hours. The yield coefficient was assumed a function of the fraction of the area harvested and the fraction of the area prepared for planting. The yield model, however, was used only for determining the yield of the second and the following seasons with twelve crops per five years of production practice. 5.4.2 Crop Maturing Model The realty that all of the rice fields in a par- ticular area would not mature at the same time was accounted for by utilizing the SEASONALIZATION Subroutine 77 (Manetsch, 1977). The total yield in a given area was indicated by OPY. It was distributed over the maturity time span between TSM (time of starting maturity) and TFM (time of final maturity) as shown in Figure 5.6. The determination of yield was accomplished by using the DELAY Subroutine of the FORDYN simulation language (Llewellyn, 1965). The portion of mature crops entered as input into the overall post production model, in terms of optimum mature crops. It was specified in the overall model, that about 60 percent of the yield matured on the first day of the harvesting period. When the harvesting operation had to be done either in advance or at delay time, the field (potential) yield of that particular area was not at its optimum level. The in—field yield reduction, due to immature kernels or pre-harvest shattering losses, was calculted by using a modified formula for IR-8 (Bhole, et al., 1970), given as, (equation 5.7) YFLF _ l 0 _ 0.32 + 0.428*(OPD+i) —(0.007*(OPD+i)**2 " ' 0.32 + 0.428*OPD — 0.007*OPD**2 where: YFLF = yield loss coefficient on day i OPD = optimum day of harvest (= 30.57), day i = specific number of days, before or after OPD. AM u mHmUHOV mumpuo wcwumMMHo mo whoamo msossfluCOUI|.m.m wusmflm A as? was 2%: 8. m2? ammo mo 8.2m v P >55 34 SA mm; 84 Ed 86 mm.0 III], r , \ -\) \\ o.o u/x zit/:1! \ \ \\ \ ”HUI/(I’ll, \ \\ \\ \\ / //u// *1 1; \ \ // N.l\ Hus // CA 8 / / \\ 7 / / \\ / I /II\A / 2 us £4 / \ / \ / \\ o.m / \ mm um / \\ m.m 79 Equation 5.7 was derived from field yield vs. grain mois- ture content data collected by Bhole, et al., 1970. The conversion from moisture content to number of days after 50 percent heading used the predictive equation fitted to the moisture content vs. days after 50 percent heading collected by Anonymous (1976). 5.5 Loss Models Losses under various conditions were simulated with predictive equations fitted to the field data. 5.5.1 Shattering Losses During Harvest Operation Less shattering losses occurred for most varie- ties before and during the optimum maturity period. Grain harvested before complete maturity increases losses of immature kernels as modelled in Section 5.4.2. Similarly, grain harvested after the optimum maturity date increases shattering losses. The amount of shattering losses after harvest was influenced by the harvesting method. The predictive equations 3.1 and 3.2 fitted to ani—ani har— vesting and sickle harvesting were used to simulate the shattering losses. The cracked grain loss due to ani-ani or sickle harvesting was set equal zero in the simulation model. 80 5.5.2 Packing and Transpor- tation Losses Transportation with bamboo baskets or plastic bags can avoid most transportation shattering losses. The model then, BBTL = 0.0 5.8 The predictive equation 3.3 was used to model the handling shattering losses with bamboo racks. 5.5.3 Losses During Threshing The losses during stalk paddy threshing can be unthreshed grain, husked grain or cracked grain losses. The amount of unthreshed, husked and cracked grains were influenced by the crop maturity level, varietal character- istics and the threshing methods. The unthreshed and husked losses models in this study assumed that unthreshed and husked losses were con— stant over maturity stages. The cracked grain losses model used in the overall model was derived from the equa- tions used by Ilangantileke (1978) and Anonymous (1976), given as: a. Foot threshing: i. Unthreshed grain losses, FTULP = 0.005 5.9 81 ii. Husked grain losses, FTHLP = 0.0 5.10 iii. Cracked grain losses, FTCLP = 0.0 5.11 b. Pedal threshing: i. Unthreshed grain losses, PULP = 0.006 5.12 ii. Husked grain losses, PHLP = 0.0 5.13 iii. Cracked grain losses, PCLP(1) ‘ 10.106 -o.3773*opp +0.00724*opp**2 — 1.0 5.14 c. Power threshing: i. Unthreshed grain losses, PWULP = 0.0 5.15 ii. Husked grain losses, PWHLP = 0.001 5.16 iii. Cracked grain losses, 86.702 -5.493*(PPD+i)+0.09277*(PPD+i)**2 PWCLP = 86.702 -5.493*PPD +0.09277*PPD**2 _ 1.0 5.17 All of the values of threshing losses are fractions of the threshed grain amount. Thus, the threshing losses were in kilograms. OPP in equation 5.14 is optimum harvesting (26.06) days after 50 percent heading. PPD in _lO.106-0.3773*(OPP+i) +0.00724*(OPP+i)**2 82 equation 5.17 is also optimum harvesting time, with the value of 29.61 days after 50 percent heading. 5.5.4 Combine—Harvester Losses The grain shattering losses after harvesting IR-38 with a 50 centimeter cutting width walking rice combine- harvester were 121.1 kg/ha at the optimum harvesting time. Because the data for shattered grain losses with the combine-harvester for different harvesting dates were not available, the model for estimating the shattering loss was derived from the data collected at optimum harvesting time and the data from sickle harvesting. The equation was: CHL = 46.6 + 1.37*NDD 5.18 where: CHL = combine-harvester shattering losses, kg/ha NDD = number of days after or before the optimum harvesting date, days The threshing operation within harvesting with a combine-harvester would also result in cracked grain losses. Equation 5.17 was used to predict the cracked grain losses due to harvesting with a walking rice combine—harvester. 83 5.5.5 Temporary Storage Losses The predictive equations 3.4 and 3.5 fitted to stalk paddy temporary storage data and to rough rice temporary storage data were used to model the losses which occurred in these threshing or drying operation delays. The model should be used for the number of temporary stor- age days more than or equal to one day. The use of a zero day temporary storage would create negative losses. 5.5.6 Drying Losses Bhole, et al., (1970) found that mechanical drying gave different head yield from sun drying. The percentage of head yield would be the same if the rough rice were harvested after the optimum harvesting date when sun- cracked grain had developed while the paddy was left standing in the field. The model for predicting cracked rice grain losses after drying with sun drying or mechani— cal drying methods was derived from Bhole, et al., (1970) data. The conversion from moisture content to the number of days after 50 percent heading was based upon the data collected by Anonymous (1976). The predictive model for cracked grain losses with sun drying time was, (1 84 -28.13 +4.13*(OSD+i-7) - OiSB* (OSD+i-7) **2 -28.13 +4.13*(ODS) -0.058*OSD**2 5.19 SCLP(i) =l.O - where: SCLP = the percentage of cracked grain losses with sun drying on day i, percent OSD = optimum harvesting date (35.8), day i = number of days after or before the optimum harvesting date, days. The predictive model for cracked grain losses from mechanical drying was given as: -l.0 +4.lO*(DSD+i+1) -0.07* (DSD+i+l)**2 -l.O+4.lO*DSD - 0.07*DSD**2 DCLP(i) =1.0 - 5.20 where, DCLP = the percentage of cracked grain losses with a mechanical dryer on day i, percent DSD = optimum harvesting date (29.29), day i = number of days after or before the optimum harvesting date, days 5.6 Operation Costs and Income Model The total operation costs over the planning period were calculated by summing all of the fixed and variable costs associated with post rice production operations. The gross return over the planning period was calculated by adding the gross return from head rice production and the gross return from production of rice other than head 85 rice. The income, before subtracting the rice production costs, was assumed as the gross return minus the total post production costs. The average annual net income was computed by multiplying the net income by a capital recovery factor, (1 +r)n - 1 r(1 +r)n CRF= where, CRF = capital recovery factor r = rate of return per time period n = planning period or number of specific time periods. 5.7 Model Verification The weather model and rainfall model were verified using actual weather and rainfall data in March and Aug— ust. The distribution of rainfall from the simulation and the actual records in the first ten day period of March for ten years are shown in Figure 5.7. It is shown in Figure 5.7 that the simulation results agree very closely with the actual data. Another important consideration in comparing simu- lated rainfall data with actual rainfall data is the fre- quency of occurrence of wet days, or the distribution of the number of consecutive nonrain days. The results of 86 100 < 80 . —--- RECORD DATA E .._-. 511mm RESULTS E 60 . E H .4 H g 40 < £2 E C— 20 - O —. r---: 0.0 20.0 40.0 60.0 80.0 100.0 RAINFALL AMOUNT, INEILDJIETERS Figure 5.7.--Histogram of Record and Simulated Rainfall Amount in the First 10 Days of March 87 silumlated nonrain days for 18 years agree very closely with the actual data for the same 18 year period, as sum— marized in Figure 5.8. Since the field moisture model and the working hours model were basically preliminary models based upon assumptions, no attempt was made to verify these models. No verification were made for the loss models which were predictive equations; however, the logical and theoreti- cal consistencies of all these models within the overall model are discussed in the results (Chapter VII). 5.8 The Overall Model A specific high yielding variety with a known opti- mum maturity date and known shattering loss grown on par- ticular land was used as an input to the model. The crop maturing model determined the portion of mature crop on a particular day of the total known area. The portion of mature crop entered available optimum mature crop at the day of harvesting. The harvesting was initiated if it was a work day (nonrain) for that particular post rice pro- duction operation combination. After the available area was harvested, the next operation, such as threshing or drying, followed if there was rice crop (stalk paddy or rough rice) available and the operational condition per— mitted. Otherwise, there were operational delays which caused post rice production losses. The total simulation 88 ”i 100. i\ DATA , SIMULATED °\‘ , o lST 10 DAYS OF MARCH * \\ o 2ND 10 DAYS OF MARCH ————— 80‘ \\ A 3RD 10 DAYS OF MARCH —.—.— g3 E 60. 8 m [321 g 40. 20' 0 LENGTH OF NON-RAIN DAY, DAY Figure 5.8.--Comparison of Nonrain Sequence Lengths from Actual Data and Simulated Results for 18 Year Period 89 model continued until all operations were completed or until the maximum post rice production period was achieved. The initial condition of the farm and the post rice pro- duction operation were defined in the beginning of each run. Two different cropping patterns, seven different farm sizes, a planning period and a number of simulations could be controlled by changing the data set. The over— all model also provided the means and standard deviation for the average annual net income from several simulations. Figure 5.9 shows the overall model's flow chart, and Figure 5.10 shows the overall block diagram. 5.8.1 Overall Model Outputs The outputs from each post rice production opera— tion combination were: 1. The total grain production (quality and quantity) 2. The total operation costs and net income; seasonal and annual 3. The total grain losses These outputs were used as a basis for a comparison analy— sis of various alternatives in post rice production. 9O READINPUT ._PW L... Mfr... j CALL, WEATHER, SOIL NDISTURE & MDRKING HOURS NO HARVESTING OPERATION l HARVESTING & HANDLING OPERATIONS l DELAYED HARVESTING & HANDLING CALCULATE LOSSES HARVESTING & HANDLING IDSSES mRKING HOURS L (5 Figure 5.9.-—Overa11 Model Flow Diagram (Part 1) CADIHATE SEEK RMIH SKXEGBILBSES (ZECUDMHETHRE§HNG ICBSES & WORKINGIKXEE L CAUJEATE RIKH KKE STORAGE LOSSES CAICULATE DRYING LOSSEss. WORKING HOURS Figure 5.9.——Overa11 Model Flow Diagram (Part 2) 92 AND INCOME Figure 5.9.——Overall Model Flow Diagram (Part 3) 93 EmgmMHo Moon H0002 HHmuw>OII.OH.m wusvflm : 1Q OZHMEQ >452 .E 02:35 2022550 Cmdfifiwnfiz: T Uzi—muz— F wfimfl L|\l_ , 8595 52 6:8 ll||I||||In .E OZHIMEEF HJII OZ Tram mm: Fz E: E :5 3.5 is ,5 525 1.0 2.0 3.0 4.0 5.0 7.5 10.0 Cb) FARM SIZE ( ha ) é“ - A 8 J _ rMelve crops per five year schedule '5 S : Two crops per year schedule : Alb 7 e ----- . ' Uicgt g a a 1 - g: a tag. 1.0 2.0 3.0 4.0 5.0 7.5 10.0 (c) FARM SIZE ( ha ) Figure 6.7.-—Comparison of (a) Head Rice Production, (b) Broken Rice Production and (c) Shat- tered Grain Losses on Different Cropping Schedules mac‘- —- x — ,‘ 7 112 Figure 6.8 depicts the effects of cropping sched- ules and the yield factors on alternative farms with sickle harvesting, bamboo rack handling, foot threshing and mechanical drying. The use of the yield factor in the model, as might be expected, affected the increment in head rice production. 6.3 Annual Net Income Per Hectare Figure 6.9 illustrated the difference in annual net income from using 25 person and 50 person crew sizes in harvesting with ani—ani. The annual net income was defined as the amount of money received by the producers after subtracting the annual post production cost from the annual gross return of selling the head rice and broken rice. The high head rice production of 2329 kilo- grams per hectare for 3.0 hectare farm size and 50 persons workingjxiharvesting, affected the high annual net income per hectare of 196,674 rupiah for the same farm size. The maximum annual net income, both for 50 person harvest— ing and 25 person harvesting teams were for the 2.0 hectare farm size. The annual net income per hectare was affected by the threshing method used. The lower net income per hectare of using a power thresher at any size of farm between 1.0 to 5.0 hectare may be caused by the high expense of owning and operating the power thresher. The m~ — ----- . . ll3 — Twelve crops per five year schedule :2 Twelve crops per five year schedule and é? yield factor :2 g 4 g Two crops per year schedule E E3” 2 a 32 3 ‘ C a-I- E _ 2 Q a E is 1 3.0 4.0 5.0 7.5 10.0 (a) FARM SIZE ( ha ) :4 52° - rIWelve crops per five year schedule U E:::jfmerwacnxm pa:fiwayearsxhahfleamd r—s O 4 II . f‘-’ 3 yield factor fig? "4 Egggg'nm3cropsgerjwfiu'schamfle 3 D 2 Es s .. E3; 2 - a s E a 1 - 3.0 4.0 5.0 7.5 10.0 Cb) FARM SIZE ( ha ) 52 o — ‘IWelve crops per five year schedule g 8 4 - [:3 Twelve crops per five year schedule and E H yield factor t:£.. - _ Egggg Emacropsgxx'yearsrhahfle a 553 3 ° :8: z - Z .— 3 $5 s. 35:: 1 - .W 3.0 4.0 5.0 7.5 10.0 (c) FARM SIZE ( ha ) Figure 6.8.—-Comparison of (a) Head Rice Production, (b) Broken Rice Production and (c) Shat- tered Grain Losses on Different Cropping Schedules and Yield Factors 114 50 persons in ani-ani harvesting and foot threshing E:::::]25 persons in ani—ani harvesting and foot threshing N O O 1000 Rupiah H O O ANNUALNEI‘INCDMEPERHECI‘ARE 1.0 2.0 3.0 4.0 5.0 PAH-i SIZE ( ha ) Figure 6.9.-—Comparison of Annual Net Income per Hectare on Two Different Ani-ani Harvesting, Foot Threshing and Sun Drying Alternatives 115 lower head rice production from power threshing when compared to the pedal thresher or foot thresher also resulted in a low annual net income per hectare. Figure 6.10 illustrates the effect of using the power thresher on the annual net income per hectare. The use of a mechanical dryer affected the annual net income as illustrated in Figure 6.11. The reason for a low annual net income per hectare for the post rice production operation with a mechanical dryer was probably due to higher mechanical dryer operation costs than sun drying. The use of a walking rice combine-harvester resulted in similar total head rice and broken rice pro- duction to the traditional harvesting method with 25 person crew size, but less head rice production due to the use of a mechanical thresher while threshing the stalk paddy. Moreover, the given high initial price of a combine-harvester (5,000,000 rupiah) caused an expensive combine operation expense and thus, resulted in negative annual net income per hectare. Until a low cost combine- harvester or custom combining is available such that low combine operation costs and positive annual net income can be achieved, technology in combination with the use of a combine—harvester is not considered practical. Increasing 20 percent of drying capacity resulted in 6 percent higher annual net income per hectare for a 116 wooepomnm mcflsmmuce ucmeMMHQ mouse co wnmwowm mom wEoocH umz Hm52c< Mo cOmHHmmEOUII.OH.m muson A an o mam E5 or. manhup HMOHQMQUQB can mcflcmmmcu Mmaom .mcflumo>nmc waxoflmmmmmmm mung 3359182 new 328% amend imfinmmfiifi 2%an mfig Hogog can. mfimwflfl uoom imfiumwg 0.3ng .OOH .oom umdnu 0001 EHVJDEH 213d EMOONI JEN ’IVDNNV 117 mcflmna HMOHcmcomE cuflz m>aumcmwwafl cm can manhuo 25m cuflz o>flpMCwaad cm co wymuomm Mom wEoocH pmz Hmsccfl mo COmflHmmEOUII.HH.o wusmflm h an o Hum cam/E o.m 0.? o.m o.N wo.H IIBIdnH 00m EfiflflflififldfimbdfliflfllTflth .oom mcflhfiuHmOHQqfiHEHUGmHHflfimefiVubOM gaflumw?flw~flcnlflsmnHHHHu @536 Sam 98 mfimmg boom .mfiumwg flawlflsfl|_ 118 5.0 hectare farm size if the same dryer price was used. A 5 percent increase in annual net income per hectare could be expected by increasing both the dryer capacity and the dryer price by 20 percent and operating on a farm size of 5.0 hectares. The change of annual net income per hectare due to the increasing capacity and price of the dryer is summarized in Figure 6.12. _ 0.150 ton/hr and a price of 120,000 Rupiah 0.125 ton/hr and a price of 100,000 Rupiah 85200' E5 “2 Ea g3 E 100- g 1.0 2.0 3.0 4.0 5.0 7.5 Efldi SIZE ( ha ) Figure 6.12.-—Comparison of Annual Income per Hectare with Two Different Dryer Capacities and Different Dryer Price 119 Figure 6.13 illustrates the effects of different cropping schedules and the yield factor. A higher number of crops per year increased the head rice and broken rice production; however, the two crops per year schedule was seen to give the highest annual net income per hectare. 6.4 Comparison of Alternatives The simulated head rice production for nine selected technology combinations on a two crops per year schedule are compared in Figure 6.14 for a 1.0 hectare farm size and in Figure 6.15 for a 3.0 hectare farm size. As it has been discussed in previous sections separately, the simulation results showed an increase in head rice productions in all cases when a mechanical dryer was compared to sun drying. The use of power threshing in a technology combina— tion resulted in lower head rice production than using the foot threshing method. Harvesting by a rice combine— harvester, in which a mechanical thresher is used while threshing the stalk paddy, also caused a decrease in head rice production. The change from foot threshing to pedal threshing in combination with sickle harvesting and mechanical drying slightly reduced (1/2 percent) the head rice production. The head rice production of all combinations with ani—ani harvesting did not differ from the head rice 120 I'll. fimflwacnxm pa:fiwayem:schahfle TWelvecncpsgxm'fivejwfim'scemfle and yield factor C: We creps per year N O O 1000 Rupiah AMfimLIET HKIMEIERIECEME 1.0 2.0 3.0 4.0 5.0 FARM SIZE ( ha ) Figure 6.13.——Comparison of Annual Net Income per Hectare for Three Different Cropping Schedules 121 fimbpnmmmmsnsm flnwaumimmmo armpits” fits-1.1mm WWW bumsamnwd'bunsamms Wmmm sumsarqn wpad ‘btrnmm WIS F'_ M W pup bangs-“1': :00; 'bmsanmx ems WWW burqsam: mod 'arpsam 313st l REMPHMBFR Wm :00; ‘btrnsam ems L bmlzpmmeqawwe Museum-ice; WIRE-TIN Alternatives and a Traditional Method for 1.0 Hectare Farm Size Exrgkxp uns pus bumsazqn .ravnd ‘bunsaueq Tue-3mg Burlap“ I hmsamzauu flnmnwmqrmhnw m As I. bx OOOI NOLLDDCIOHd 33m (NEH ’IVDNNV Figure 6.14.—-Comparison of Annual Head Rice Production from Nine 122 Snap m1: pus ' bmsazu'eq 31:;qu Earp/11pm” gamma WWW bL—gqsammcd'btrpsmms 6mm Tampa: pus amsamifififi’flmwmnmamms WW1”? snmmnnnag'fixumuaxflxfls 531519135111? W300; 'fitrpsaueqnxe.m muss-re mm'wpsamms Wm” snmmnnnmu‘flnfiammifixfls bu'pup mm pus mm 3003 ’W FIE-Zn!!! bupup us we bumsam Janod ‘bmsaam rue-m NOILDDGOHd EDIE GVHH TVDNNV Figure 6.15.-Comparison of Annual Head Rice Production from Nine Alternatives and a Traditional Method for 3.0 Hectare Farm Size 123 production of all combinations involving sickle harvest- ing for a 1.0 hectare farm size (Figure 6.14). When the farm size increases to 3.0 hectares, however, the tech- nology combinations with the use of ani—ani harvesting resulted in incomplete harvesting and delays in harvest— ing; thus, the head rice productions were lower than all combinations with sickle harvesting (Figure 6.15). Comparison of simulated head rice production of various technology combinations in post rice production indicated that the use of sickle harvesting with 25 per— sons, foot threshing and sun drying was the most appro— priate technology combination for the farm size up to 3.0 hectares. The change from sun drying to a mechanical dryer within the technological combination of sickle harvesting and foot threshing increased the head rice production by 38 percent for the 3.0 hectar farm size. The owning and operating expenses of a mechanical dryer, however, would also increase, and thus the annual net income per hectare decreased. For the 3.0 hectare farm size, the use of a mechanical dryer in combination with sickle harvesting and foot threshing resulted in a 6 percent lower annual net income per hectare than if sun drying was used. An effort to reduce mechanical drying expenses through custom dry— ing, cooperative drying practices, subsidy or credit by a 124 government program could increase the head rice produc— tion for the country and offset the decrease in the annual net income. Simulation output shows only a slight decrease in head rice (1/2 percent), therefore there is some basis for recommendation to improve the working conditions in threshing practices. The use of the pedal thresher instead of manual treading method will make the threshing job more convenient and attractive. CHAPTER VII CONCLUSIONS The conclusions of this research study are: 1. The field measurements on post rice production operations and losses indicated: a. Any delay in the harvesting operation increased shattered grain losses. Shattered grain losses for sickle harvesting of the IR—36 rice variety was about 50 kilograms per hectare when harvested at optimum matur- ity, and increased to 68 kilograms per hectare when har— vesting was delayed nine days. b. Bamboo rack handling losses increased when stalk paddy was harvested after optimum maturity. For IR-36, the transportation loss was 11.5 kilograms per hectare when harvested six days after the optimum harvest- ing date. The handling losses were about 8 kilograms when the harvesting was at the optimum maturity. c. Improper temporary storage losses for wet rough rice and wet stalk paddy of the IR—36 rice variety was 25 percent and 42 percent respectively by the fifth day. 125 126 2. The system simulation of post rice production operations illustrated the influence of environment on the system as well as interactions within the system. As verified by post rice production data, the system simu— lation provided an evaluation of the post rice production system for the selection of post rice production tech- nology alternatives. The simulation indicated: a. The low traditional harvesting capacity of 0.050 hectare per hour with 25 person crew in ani-ani as the bottleneck of the overall system for farms over 1.0 hectare in size. The final rice production was determined by the harvesting capacity as affected by delays and incomplete harvesting while the sun dryer capacity of 0.125 ton per hour was never fully utilized. If harvest- ing labor were abundant, the simulation with an assumed 100 harvesters indicated that the sun drying capacity of 0.125 ton per hour was the limiting factor in the post production Operation for a 3.0 hectare farm size. b. The interacting effects between the impact of the power threshing drum teeth and the delay in harvest- ing caused more broken rice kernels. The use of power drum thresher in combination with 25 persons in ani—ani harvesting, therefore increased the broken rice production for farm sizes over 1.0 hectare because of the delay in harvest. 127 c. The change from sun drying to a mechanical dryer within the technological combination of sickle harvesting and foot threshing increased the head rice pro- duction by 38 percent for 3.0 hectare farm size. The owning and operating expenses of a mechanical dryer; how— ever, would increase and thus lower the annual net income per hectare by 6 percent. The direct application of the simulation model was in the selection of technology alternatives for the post rice production operations. The simulation indicated that 25 persons for sickle harvesting, bamboo rack hand- ling, foot threshing and sun drying was the most appro- priate technology combination based upon head rice production and income for farm sizes up to 3.0 hectares. Further information that can be drawn from this simulation study are: a. The overall post production model illustrates the influence of the weather on the system, the inter- actions of the rice paddy plants with the operations, the interactions between the post rice production operations (i.e., the effect of harvesting delay on the threshing and drying operations), and finally, the overall effect on rice production, grain losses and income. The compre- hensive illustration of the system can contribute sub- stantially to a better understanding of the behavior of the post rice production system. 128 b. The post rice production model with twenty simulations for each technology alternative provided pre- liminary results when dealing with weather variability and associated risk. The simulation showed a larger vari— ation of rice production and a stronger dependency of harvesting operations on environmental condition when a combine-harvester was used compared to manual (ani-ani or sickle) harvesting. The simulation also indicated more independence of mechanical drying operations from the weather variability than sun drying. CHAPTER VIII SUGGESTIONS FOR FURTHER RESEARCH 1. Model modification needs further research. Particularly, the field moisture model and the working hour model need further development and verification. Whether the field moisture model adequately and realisti- cally represents the real soil moisture status needs further verification. The working hours model also needs further testing. The two basically preliminary models reveal important functions in determining the comparative performance of technology alternatives, but need further work. 2. A greater data base on post rice production is necessary to increase the reliability of the simula- tion results. Data which are necessary to improve the model performance are: a. Data related to lodging. Maturity level of the paddy was the only parameter used to predict the losses in the present model. Interacting effects of weather and varietal characteristics, such as, lodging may become a sensitive loss factor. 129 130 b. Weather variables other than rainfall and the effect on drying operations. The present model indi- cated the variations in risk and performance of sun drying and mechanical drying systems by using rainfall as the main parameter. Further study may indicate an importance for other weather variables (e.g., ambient temperature or humidity). 3. Further application of the present model is needed. Suggestions for application are: a. Other locations or regions. b. Identify the optimum size of machines or tools in one technology package. c. More rice varieties. d. Develop more sensitivity tests to: ——indicate logical or theoretical inconsis- tencies in the model, --add to user's understanding of and insights into the model and corresponding real system, --provide a direct way to test alternatives, ——suggest data collection priorities by indicating those parameters which are most sensitive to the performance of the model. BIBLIOGRAPHY 131 BIBLIOGRAPHY Anonymous, 1976. Mechanization of Rice Production. India- Nigeria-Senegal. An International Coordinated Research Project: 1970-1976. Published by Arrangement with the Food and Agriculture Organi- zation of the United Nations and the Agricultural University of Wageningen, the Netherlands. Anonymous, 1977. Research on Rice Mechanization in the Tropic, Agricultural Machinery Research Institute, Japan. Araullo, E. V., D. B. De Padua and Michael Graham, 1976. Rice: Postharvest Technology. International Development Research Center, Ottawa, Canada. Asimow, Morris, 1962. Introduction to Design. Prentice Hall, Inc., Englewood Cliffs, N.J. Baier, W. and Geo. W. Robertson, 1966. A New Versatile Soil Moisture Budget, Can. J. Plant Sci. (46). 299-315. Ban, Toshizo, 1970. "Present Status and Problems of Rice Processing Procedures from Harvesting through Milling in South East Asia--A Case Study in Indonesia." Rice Production in South East Asia, Japan. Bhole, N. G., S. Lal, V. V. Rama Roa and J. Wimberly, 1970. RPEC, India. Publication 701, Indian Institute of Technology, Kharagpur, India. BIMAS (Badan Pengendali Bimas), 1972. Evaluation on Bimas Program. Unpublished. Brooker, D. B., F. W. Bakker-Arkema and C. W. Hall, 1974. Drying Cereal Grains. Avi Publishing Co. Inc., Westport, Connecticut. Brown, Galen K., 1972. Mechanical Harvest System Simula- tion and Design Criteria for Processing Apples. Unpublished Ph.D. Thesis, Michigan State Uni- versity. 132 133 BULOG (Badan Urusan Logistik), 1976. Rice Postharvest Technology, Progress Report for International Development Research Center, Ottawa, Canada. Unpublished. Campbell, W. D. and J. B. McQuitty, 1972. An Evaluation of Harvest Simulation as an Aid to Decision Making, Canadian Agricultural Engineering. 14 (l). 11:14. CBS (Brio Pusat Statistik), 1976. Statistical Yearbook of Indonesia, Jakarta, Indonesia. CBS (Brio Pusat Statistik) and BIMAS, 1977. Analysis and Statistics of Rice Production Development Pro- gram. Unpublished. Collier, William L., Gunawan Wiradi and Soentoro, 1973. Recent Changes in Rice Harvesting Methods. Agricultural Development Council. New York. Djojomartono, Moeljarno, Kamaruddin Abdullah and Rizal Syarief ST., 1979. In—field Post Rice Produc— tion Losses on Farm in West Java. A Paper Pre— sented at the Grains Post Harvest Workshop, BULOG and S. E. Asia COOperative Post Harvest Research and Development Programme, Jakarta, Indonesia. Dumont, A. G. and D. S. Boyce, 1974. The Probabilistic Simulation of Weather Variables. Journal of Agricultural Engineering Research (1974) 19. 131:145. Eriyatno, 1979. System Modeling on Rice Milling Technol- ogy in Indonesia. Unpublished Ph.D. Thesis, Michigan State University. Esmay, M. L. and Ming—wu Wu, 1975. Adoption of the Drum Thresher for New Rice Varieties in Asia, Agri- cultural Mechanization in Asia, Spring, 1975. 15:19. Esmay, M. L., Soemangat, Eriyatno and A. Phillips, 1977. Draft of Manual on Post Rice Production in the Tropic. East West Center, Honolulu, Hawaii. (In printing) Exell, R. H. B. and Sommai Kornsankoo, 1978. A Low Cost Solar Rice Dryer. Asian Institute of Technology, Bangkok, Thailand. 134 Ezaki, H., 1972. The Improvement of Rice Harvesting Mechanization in Japan. Meeting of Experts on the Mechanization of Rice Production and Pro— cessing, Paramaribo, Surinam, 27 Sept. - 2 Oct., 1971. Fridley, R. B., 1973. Simulation of Strawberry Production in California to Evaluate Socio-economic Implica— tions of Alternate Harvest System. Unpublished Ph.D. Thesis, Michigan State University. Frisby, James C., 1970. Estimating Good Working Days Available for Tillage in Central Missouri, Trans- action of the ASAE, 1970. 641:643. Hillier, F. S. and G. J. Lieberman, 1974. Operation Research. Holden Day Inc., San Francisco, Cal. Grist, D. H., 1975. Rice. Fifth Edition, Longmans, London. Holtman, J. B., L. K. Pickett, D. L. Armstrong and L. J. Connor, 1973. A Systematic Approach on Simulating Corn Production Systems. Transaction of the ASAE l6 (1). 19:23. Ilangantileke, Sarath G., 1978. Post Rice Production Sys- tems Analysis. Unpublished Ph.D. Thesis, Michigan State University. IRRI (The International Rice Research Institute), 1967. The International Rice Research Institute Annual Report for 1967, Los Banos, Laguna, Philippines 149:188. , 1968. The International Rice Research Institute Annual Report for 1968, Los Banos, Laguna, Philippines. 240:305. , 1969. The International Rice Research Institute Annual Report for 1969, Los Banos, Laguna, Philip- pines. 109:130 and 190:220. , 1970. The International Rice Research Institute Annual Report for 1970, Los Banos, Laguna, Philip- pines. 123:156. , 1974. The International Rice Research Institute Annual Report for 1974, Los Banos, Laguna, Philip— pines. 349:360. 135 IRRI, 1975. The International Rice Research Institute Annual Report for 1975, Los Banos, Laguna, Philippines. 300:321. Jones, James W., Rex F. Colwick and E. Dale Threadgill, 1972. A Simulated Environmental Model of Tempera— ture, Evaporation, Rainfall and Soil Moisture, ASAE Paper 70—404. van Kampen, J. H., 1971. "Farm Machinery Selection and Weather Uncertainty." System Analysis in Agri- cultural Management, John Wiley & Sons Austra- lania, Pty., Ltd. Khan, Amir U., 1976. "Harvesting and Threshing Equipment and Operations." Rice Post Harvest Technology. I Eds. IRDC, Ottawa, Canada. Link, D. A., 1968. Research Needs for Farm Machinery Scheduling. Computers and Farm Machinery Manage— ment Seminar Proceedings. ASAE Publication PROC— 468. 28:32. Llewllyn, Robert W., 1965. FORDYN, An Industrial Dynamics Simulator. Raleigh, North Carolina, Privately Printed. Manetsch, Thomas J. and Gerald L. Park, 1976. System Analysis and Simulation with Application to Econ- omic and Social Systems. Parts I and II. Depart— ment of Electrical Engineering and System Science, Michigan State University, East Lansing, Michigan. Murata, Yoshio, 1975. Estimation and Simulation of Rice Yield from Climatic Factors. Agricultural Meteo— rology, 15(1975). 117-131. Niko, Sadaichi, 1948. Introduction to Agricultural Mach— ineries (in Japanese), Shinnokisha, Japan. Oldeman, L. R., 1975. An Agro-climatic Map of Java. Cen— tral Research Institute for Agriculture, Bogor, Indonesia. Owen, P. C., 1971. The Effects of Temperature on the Growth and Development of Rice. Field Crop Abstracts, Vol. 24, No. l. Panol, F. Y., 1972. Modeling Yield and Weather for Sugar- cane Production Simulation. Unpublished Ph.D. Thesis, Michigan State University. 136 Robertson, G. W. and G. A. De Weille, 1973. Rice and Weather, Reprinted from the FWMO BULLETIN," January 1973. Schmidt, J. L. and A. R. Hallauner, 1966. Estimating Harvest Date of Corn in the Field. Crop Science, Vol. 6. 227—231. Shaw, R. H., 1963. Estimation of Soil Moisture under Corn. Agricultural and Home Economics Experiment Sta— tion Research Bulletin, 520. Iowa State Uni- versity of Science and Technology, Ames, Iowa. , 1965. Estimation of Field Working Days in Spring from Meteorological Data. Iowa State Journal of Science 39(4). 393:402. Smith, G. W., 1973. A Study of the Agrometeorology of Indonesia. Report for the Land Capability Appraisal Project, Food and Agricultural Organi- zation of the United Nations, Bogor, Indonesia. Sorensen, Eric E., 1967. Regional Rain Model's with Space and Time—correlated Structure. Management Science, Vol. 14, No. 3. 239:249. , and J. F. Gilheany, 1970. A Simulation Model for Harvest Operations under Stochastic Conditions. Management Science, Vol. 16, No. 8. B—429zB—565. Spurgeon, David, 1976. Hidden Harvest, International Development Research Center, Ottawa, Canada. Stout, B. A., 1966. Equipment for Rice Production. Farm Power and Machinery Specialist Agricultural Engin- eering Branch Land and Water Development Division, Food and Agriculture Organization of the United Nations, Rome. Sumardi and Jetty Setiawati, 1978. The Effect of Harvest— ing Times on Milling Rice Grade (in Indonesian). Central Research Institute for Agriculture, Indonesia. Tulu, M. Y., J. B. Holtman, R. B. Fridley and S. D. Parson, 1974. Timeliness Costs and Available Working Days Shelled Corn. Transaction of the ASAE 17(4). 789:800. Weitz—Hettelsater, 1972. Economics and Engineering Study: Rice Storage, Handling and Marketing in Indonesia. Kansas City, Missouri. TRTE UNIV. LIBRARIES ‘I H 1 WI \| 3100632730 MICHIGAN 5 “WWW“ 3129