An Application of Artificial Neural Network for Predicting Threshing Performance in a Flexible Threshing Device

Rice is a widely cultivated food crop worldwide, and threshing one of the most important operations combine harvesters in grain production. It complex, nonlinear, multi-parameter physical process. The flexible device has unique advantages reducing damage rate already been major concerns engineering design. Using measured test database bench, rotation speed cylinder (RS), clearance concave sieve (TC), separation (SC), feeding quantity (FQ) are used as input layer. In contrast, crushing (YP), impurity threshed material (YZ), loss (YS) output A 4-5-3-3 artificial neural network (ANN) model, with backpropagation learning algorithm, was developed to predict performance device. Next, we explored degree which inputs affect outputs. results showed that R model validation set hidden layer reached 0.980, root mean square error (RMSE) average absolute (MAE) were less than 0.139 0.153, respectively. built predicted device, regression determination coefficient R2 between prediction data experimental 0.953. revealed combined ANN method an effective approach for predicting rice. Moreover, sensitivity analysis RS, TC, SC crucial factors influencing relative importance 15.00%, 14.89%, 14.32%, FQ had least effect on performance, 11.65%. Our findings can be leveraged optimize future devices.