基于改进ART2神经网络的汽轮机故障诊断

针对反向传播神经网络(BPNN)学习收敛速度慢、易陷入局部极小值等问题,探讨了采用ART2神经网络的汽轮机组故障诊断方法。根据ART2网络学习与记忆的特点,对模型进行改进,解决了原始ART2网络权值学习的随机偏移问题,有效地过滤了噪声。仿真结果表明:该诊断方法快速、准确且易于工程实现。

Based on Improving ART2 Network Diagnosing Steam Turbine Set Failure

Point to some problems of back propagation neural networks(BPNN)like slow convergence of learning and liability of dropping into local minima,ART2 neural network,a new way of fault diagnosis of steam turbine units is suggested. According to the learning and remembering feature of ART2,noise and random excursion were improved by the new model.Simulation results show that the proposed method is featured by swiftness,accuracy and easy to practical application.