摘要
给出了微动齿轮的机械振动机理和故障特征,建立了三层小波神经网络,并结合遗传算法进行小波神经网络参数优化。将微动齿轮故障分为无故障、齿轮断层、齿轮面磨损脱落、齿轮面损伤,齿轮面裂痕等五种故障,通过振动试验测试故障信息,将其作为小波神经网络的训练样本,并结合遗传优化实现网络隐层节点和小波参数最佳值。仿真结果表明遗传优化的小波神经网络能够有效避免神经网络不收敛的缺点,提高学习速度,采用遗传优化神经网络进行微动齿轮故障诊断,具有较高的诊断精度和效率,可以有效应用于其他系统的故障诊断工程中。
The micro-vibration mechanism and fault characteristics of micro-gears were described,and three layer wavelet neural network was constructed and optimized with genetic algorithm.Then the faults were classified into five types,i.e.no fault,gear crack,gear face wear,tooth face attrition,tooth face crack.And the diagnosis information is tested through vibration experiment and designed as training samples of wavelet neural networks to achieve network hidden nude and optimum value of.wavelet parameter combined with genetic optimization.Simulation result shows the new method can avoid normal neural networks with convergence quality and enhance learning speed,also the higher diagnosis precision and efficiency was got,which can be effectively used in engineering diagnosis system.
出处
《机械设计与制造》
北大核心
2012年第6期176-178,共3页
Machinery Design & Manufacture
基金
河南省教育厅自然科学研究资助项目(2011B510016
2011B470005)
关键词
小波神经网络
遗传算法
故障诊断
微动齿轮
Wavelet Neural Network
Genetic Algorithm
Fault Diagnosis
Micro-Gears
