Prediction and Analysis of Forged Workpiece’s Precision in Hydraulic Press Based on BP Neural Network and Genetic Algorithm
Workpiece’s precision is an important indicator of hydraulic press. In order to accurately predict the accuracy of the part, a method that combined the genetic algorithm and neural network is put out. Design of orthogonal experiment (DOE) is used to determine the input samples of neural network training and testing samples. The output samples are obtained by finite element analysed method (FEA). Through optimizing weights and thresholds of BP neural network using genetic algorithms, prediction model of workpiece’s precision is established. The established predict model overcomes the shortcomings of slow to convergence and easy to fall into the local minimum point of BP neural network model . By comparing the neural network forecast result with FEA ‘s results, it can be seen that the established prediction model has good fitting and generalization ability. So the model can be used to predict the workpiece’s precision.
Zhengyi Jiang and Chunliang Zhang
Y. C. Li et al., "Prediction and Analysis of Forged Workpiece’s Precision in Hydraulic Press Based on BP Neural Network and Genetic Algorithm", Advanced Materials Research, Vols. 97-101, pp. 2598-2602, 2010