基于遗传算法优化BP神经网络的磨削力预测

BP Neural Network Based on Genetic Algorithm for Prediction of Grinding Force

为了克服传统BP神经网络的学习速率慢、容易陷入局部极小点等缺点,采用遗传算法对BP神经网络的初值空间进行遗传优化。用遗传算法来优化BP神经网络的权重和阈值,得到最佳的初始权值矩阵,并按误差前向反馈算法沿负梯度方向搜索进行网络学习的方法对磨削力进行预测。根据磨削力实验数据对网络进行训练,仿真结果表明该模型可以精确的描述砂轮速度、工件速度、磨削深度对磨削力的影响,并可以用有限的实验数据得出整个工作范围内磨削力的预测值。

In view of that BP neural network has the disadvantage of slowly leaning rate and being easily stacked into the minimal value locally,the Genetic Algorithm was utilized to optimize the initial-value space of BP neural network.The optimal initial weight-value matrix was obtained by using Genetic algorithm to optimize the weight-value and threshold of BP neural network,and the method of network learning was analyzed by using the error-forward-feedback algorithm with negative gradient searching.The network training was carried out by experiment of grinding force data,and the simulation results show that the model could accurately describe the effect of wheel＇s speed,working speed and grinding depth on grinding force.The prediction of grinding force in the working range can be obtained by using limited experiment data.