Application of artificial neural network in prediction of the combine harvester performance

作者：Gundoshmian Tarahom Mesri^*; Ghassemzadeh Hamid Reza; Abdollahpour Shamsollah; Navid Hossein

来源：Journal of Food Agriculture and Environment, 2010, 8(2): 721-724.

• Neural networks
• combine performance
• modeling

Harvesting is one of the most critical operations in grain production. The purpose of grain harvesting is to recover grains from the field and separate them from the rest of the crop material in a timely manner with minimum grain loss while maintaining highest grain quality. A grain combine performs many functional processes. These are gathering and cutting (or in case of windrows, picking up), threshing, separation, and cleaning. Combine performance is affected by design factors, operating conditions and crop properties. Any means to increase the productivity and efficiency of the agricultural combine harvester has immediate benefits for the producer. Until recently, no reliable models of complex physical processes and no techniques to handle design of such systems in nonlinear, multi parameter and multiple objective conditions were available. In this study, a three layer perceptron neural network, with back propagation (BP) training algorithm, was developed for modeling of the combine performance. The optimum structure of neural network was determined by a trial and error method and different structures were tried. The model investigates the influence of the wheat yield, crop variety, crop moisture content, crop height, height of cut, threshing drum speed, concave clearance, fan speed, chaffer opening and lower sieve opening on the combine performance.

• 出版日期2010-4