The Application of Hybrid Artificial Intelligence Techniques in the Optimization of Crude Palm Oil Production

The Application of Hybrid Artificial Intelligence Techniques in the Optimization of Crude Palm Oil Production

Dr. Lily Amelia
Dosen Fakultas Teknik
Universitas Esa Unggul
The production of crude palm oil and palm kernel are greatly influenced by factors such as deterioration of the raw material, processing inefficiency as well as environmental condition. An optimisation model for the palm oil production is therefore critical in ensuring maximum revenue in production whilst minimizing palm oil and palm kernel losses and production cost. The proposed optimisation model is an integration between fuzzy expert system and multi objective programming model. Four fuzzy expert models were developed for each processing station with the aim of achieving four objectives, ie. to maximise revenue, minimise total production costs as well as minimise the total amount of palm oil and palm kernel losses. Two heuristic optimisation methods, ie. Genetic Algorithm (GA) and Direct Random Search are applied to solve this problem. The model is able to optimise the revenue, production cost as well as the total amount of palm oil and palm kernel losses. The study has also shown that Genetic Algorithm results in more optimum solutions compared to the Direct Random Search method.

Crude palm oil (CPO) and palm kernel are yielded from fresh fruit bunches of palm oil that undergo several stages of processing. After the fresh fruit bunches of palm oil are loaded into the loading ramp, the fresh fruit bunches are sterilised in a steriliser in order to separate fruits from the bunches. The sterilized fruit bunches are then transferred to a thresher using a hoisting crane where the fruits are threshed in a stripper drum and separated from the bunches. The fruits are then digested in a digester to separate the nut from the fruits. The fruits are then delivered to the screw press by means of a feed screw conveyor. Crude oil resulted from pressing is transferred into a clarification station in which the crude oil consisting of oil, water and non oil solid (NOS) is separated by using centrifugal force and heat treated at temperature 80-90 oC. The mixture of fibre and nuts from the screw press are transferred into the kernel station by a cake breaker conveyor. There after, the fibre and nuts are separated in a depericarper and entered the nut silo in order to reduce the water content. Nuts are broken in a nut cracker or a ripple mill to obtain the kernel. The mixture of kernel and shell are separated using a separating column and hydro cyclone. Finally, kernel are dried in the kernel dryer before they are stored [1].
In a palm oil production, objectives that are of interests to the decision makers are to maximise crude palm oil and kernel production, minimise palm oil and palm kernel losses during processing and to maintain production costs at the lowest possible level [2]. Nevertheless, the production of crude palm oil and palm kernel is a complex problem due to the influence of processing variables and environmental factors. Existing problems in the processing of crude palm oil and palm kernel are the lack of processing efficiencies, the properties of the raw materials that cause them to deteriorate easily, delay in processing, etc. These factors may affect the quantity and quality of oil and kernel production as well as giving impact to the production costs. Therefore, the development of an optimisation model is crucial in minimising the total amount of palm oil and palm kernel losses hence maximising revenue and minimising production costs.
The application of fuzzy logic expert system in the production process control and optimisation have been widely used since Mamdani and Assilian developed fuzzy inference fuzzy logic controller model for a steam engine [3]. Mamdani fuzzy logic expert system has been applied in areas such as production industries. Other examples of fuzzy expert control application is in hydraulic forging machine [4] and twin extruder machine [5]. Fuzzy control model for milling process optimisation was developed by Peres et al. [6]. Vagelatos et al. [7] developed fuzzy control model for optimising injection molding process. Other applications include optimisation of power distribution system operations developed by Sarfi and Solo [8], food frying process by Rywotcky [9] and fuzzy expert control for oil and gas supplies by Neuroth et al. [10]. Despite the wide application of fuzzy expert system, the technique has yet to be applied in the crude palm oil production.
Genetic Algorithm (GA) and Direct Random Search are two heuristic optimisation methods that have been widely applied in the industries. GA that was developed by John Holland in 1975 [11] has been applied in production process optimisation such as in the pulp and paper industry [12], glass furnace operation [13], fuzzy control system for hydro cyclone [14], fuzzy expert system in continuous stirring tank reactor [15], classification of welding flaw types [16], assembly process [17], semi conductor testing industry [18] and batch free radical polymerisation reactor [19]. However, GA has not been used for solving problems in the crude palm oil and palm kernel industry. Direct Random Search that was developed by Luus and Jaakola [20] has also been widely applied to solve many optimisation problems especially for chemical industries such as multi stages recycle systems [21] and phase and chemical equilibrium [22].

[1] Hartley, C.W.S., The Oil Palm 2nd ed, Longman, New York, 1977.
[2] Eng, G.E. and M.M. Tat, “Quality Control in Fruit Processing”, JAOCS 62(2), 1985, pp. 274-281.
[3] Mamdani, E.H. and S. Assilian, “An Experiment in Linguistic Synthesis with a Fuzzy Logic Controller”, Man Machine Studies 7(1), 1975, pp 1-13.
[4] Lee, Y. H. and R. Kopp, “Application of Fuzzy Control for a Hydraulic Forging Machine”, Fuzzy Sets and Systems 118(1), 2001, pp. 99-108.
[5] Lee, S.J., C.G. Hong, T.S. Han, J.Y. Kang and Y.A. Kwon, “Application of Fuzzy Control to Start Up of Twin Screw Extruder”, Food Control 13(4-5), 2002, pp. 301-306.
[6] Peres, C.R., R.E.H. Guerra, R.H.H.A. Alique and S. Ros, “Fuzzy Model and Hierarchical Fuzzy Control Integration: An Approach for Milling Process Optimization”, Computers in Industry 39(3), 1999, pp. 199-207.
[7] Vagelatos, G.A., G.C. Rigatos and S.G. Tzafestas, “Incremental Fuzzy Supervisory Controller Design for Optimizing the Injection Molding Process”, Expert Systems with Applications 20, 2001, pp. 207-216.
[8] Sarfi, R.J. and A.M.G. Solo, “Network Radiality, Parameter and Performance Heuristics in Optimization of Power Distribution System Operations Part 2 : Fuzzification of Rule Base”, Electrical Power and Energy Systems 24(8), 2002, pp. 683-692.
[9] Rywotycki, R., “Food Frying Process Control System”, Journal of Food Engineering 59(4), 2003, pp. 339-342.
[10] Neuroth, M., P. McConnell, F. Stronach and P.Vamplew, “Improved Modeling and Control of Oil and Gas Transport Facility Operations using Artificial and Intelligence”, Knowledge Based Systems 13(2-3), 2000, pp. 81-92.
[11] Goldberg, D.E., Genetics Algorithms in Search, Optimization and Machine Learning, Addison Wesley
Publ.Co. Inc., Massachusetts, 1989.
[12] Santos, A. and A. Dourado, “Global Optimization of Energy and Production in Process Industries: A Genetic Algorithm Application”, Control Engineering Practice 7, 1999, pp. 549-554.
[13] Pina, J.M. and P.U. Lima, “A Glass Furnace Operation System using Fuzzy Modeling and Genetic Algorithms for Performance Optimization”, Engineering Applications of Artificial Intelligence 16(7-8), 2003, pp. 681-690.
[14] Karr, C.L., D.A. Stanley and B. McWhorter, “Optimization of Hydrocyclone Operation using Geno-fuzzy Algorithm”, Computer Methods in Applied Mechanics and Engineering 186(2-4), 2000, pp. 517-530.
[15] Bellarbi, K., F. Titel, W. Bourebia, and K. Benmahammed, “Design of Mamdani Fuzzy Logic Controllers with Rule Base Minimization using Genetic Algorithm”, Engineering Applications of Artificial Intelligence 18(7), 2005, pp. 875-880.
[16] Liao, T.W., “Classification of Welding Flaw Types with Fuzzy Expert Systems”, Expert System with Application 25, 2003, pp. 101-111.
[17] Tiwari, M.K. and D. Roy, “Application of an Evolutionary Fuzzy System for the Estimation of Workforce Deployment and Cross-training in an Assembly Environment”, International Journal of Production Research 40(10), 2002, pp. 4651-4674.
[18] Wang, K.J. and T.C. Hou, “Modelling and Resolving the Joint Problem of Capacity Expansion and Allocation with Multiple Resources and Limited Budget in the Semiconductor Testing Industry”, International Journal of Production Research 41(14), 2003, pp. 3217-3235.
[19] Silva, C.M. and E.C. Biscaia, “Genetic Algorithm Development for Multi-objective Optimization of Batch Free-radical Polymerization Reactors”, Computers and Chemical Engineering 27, 2003, pp. 1329-1344.
[20] Luus, R. and T.H.I. Jaakola, “Optimization by Direct Search and Systematic Reduction of the Size of Search Region”, American Institute of Chemical Engineers (AIChE) Journal 19(4), 1973, pp. 760-766.
[21] Luus, R., “Optimization of Multistage Recycle Systems by Direct Search”, Canadian Journal of Chemical Engineering 53(6), 1975, pp. 217-220.
[22] Lee, Y.P., G.P. Rangaiah, and R. Luus, “Phase and Chemical Equilibrium Calculations by Direct Search Optimization”, Computer and Chemical Engineering 23, 1999, pp. 1183-1191.
[23] Cohon,J.L., Multiobjective Programming and Planning, Academic Press, New York, 1978.
[24] Gen,M. and R. Cheng, Genetic Algorithms and Engineering Design, John Wiley and Sons, New York, 1997.

Tinggalkan Balasan

Alamat email Anda tidak akan dipublikasikan. Ruas yang wajib ditandai *