Intelligent modeling of material separation in combine harvester’s thresher by ANN

ABSTRACT: In order to minimize combine harvester loss, it is necessary that product process such as cutting, conveying, threshing, separating and etc., should be optimized. Threshing is one of these processes, which has more effect on combine performance. Maximum separation of threshed grains from concave is desired, because it will decrease the load of straw walker and other separation units which lead to reduction of straw walker losses. In order to evaluate effecting parameters on separation efficiency, experiments were conducted in 4×3×3 factorial pattern with Randomized Blocks design. Independent variables in these experiments were, stem height, feed rate, threshing clearance ratio and rotational velocity of threshing cylinder. To offer an intelligent model to forecasting grains separation, Neurosolution was used. Results showed a network with 7 neurons in hidden layer had minimum MSE with R2=0.81. Furthermore, results showed that the amount of grain separation had dependent to threshing clearance, speed of threshing cylinder, stem height and feed rate, respectively. The amount of grain separation has increased with reduction in stem height, feed rate, threshing clearance ratio and speed up of cylinder.