【摘要】 近年来,神经网络在故障诊断领域中应用广泛,针对传统BP神经网络的不足,利用基于遗传机制的神经网络进行齿轮的故障诊断。先采用遗传算法对神经网络的权值进行全局优化,然后引入动量因子和自学习速率改善传统BP算法的收敛速度慢缺点,最后结合MATLAB编程语言进行齿轮故障诊断,结果与一般BP网络的诊断结果进行比较,证明该方法的正确性和有效性。更多还原

【Abstract】 Recently, neural network is widely used in Fault Diagnosis field. Gears Fault Diagnosis based on genetic neural network is used to avoid the defect of the tradition BP neural networks. Genetic Algorithm (GA) is used to optimize the initial weights in global view, and then improve the network problems of slow convergence speed by momentum factor and self study speed. At last use MAT-LAB language in the gear fault diagnosis. The result was compared by the generally network which testify that method correctness and validity.更多还原