基于进化神经网络的齿轮可靠性预测

Evolutionary neural network based prediction of gears reliability

人工神经网络模型中广泛应用的是BP(BackPropagation)模型 ,针对BP算法存在收敛速度慢 ,容易陷入局部最小点的缺陷 ,本文用遗传算法 (GeneticAlgorithm :GA)训练神经网络 (ArtificialNeuralNetwork :ANN) ,取代了一些传统的学习算法 ,设计了GA +BP学习算法 用遗传算法和神经网络相结合的方法求解了齿轮弯曲疲劳寿命的预测问题 仿真结果表明 ,组合GA与BP可以克服单纯使用BP易陷入局部极小等问题 ,取得了较为满意的效果 。

Among various artificial neural network models presented, the Back Propagation(BP) model is most widely used. To improve the convergence speed and get rid of the problem of easily falling into local minimum point during the training of BP model, this paper proposes a new learning algorithm which add GA to the training of ANN. The problem for predicting bending fatigue life of gears is solved by using the method combining BP with GA. The simulation results show that the combination of GA and BP can effectively overcome the problem of easily falling into local minimum point and give higher accuracy of predictions.