基于混沌神经网络动态联想记忆的电机故障诊断

Rotor Fault Diagnosis of Induction Motors Based of a Dynamic Associative Memory of Chaotic Neural Network

采用一种由混论神经元构成的神经网络，通过研究其非线性动力学特性、混炖吸引子轨迹以及对初始条件的敏感性，实现混炖神经网络的动态联想记忆。在此基础上，应用混沌神经网络对异步电动机鼠笼转子断条故障进行动态记忆和恢复。研究结果表明，基于混炖神经网络的故障诊断有助于故障模式的记忆和重视。

An associative neural network is used with chaotic neural models interconnected through aconventional auto-associative matrix of synaptic weights. Dynamic associative memory and essentialcharacteristics of chaotic neural network is dealt with: nonperiodic chaos, chaotic attractors and sensitivity tostarting condition. In the paper faults of three phase induction motors with broken bare is diagnsed usingdrpamic associative memory of chaotic neural network. Diagnose result suggest that the chaotic neutal networkis beneficial to dynamic memory retrieval and faults identification. And chaotic neural network has faulttolerance.