摘要
针对传统BP神经网络训练中收敛速度较慢的缺点,将一种基于L-M算法的神经网络应用于液压泵故障诊断,并建立了基于该算法的故障诊断模型;论述了液压泵的故障特征频率,研究基于LabVIEW的频率提取与后期神经网络的处理方法.仿真结果表明:该方法和模型显著缩短了训练时间,运用神经网络方法进行液压泵故障诊断是有效的.
An improved neural network based on L-M algorithm was applied to hydraulic pump fault diagnosis against the slow convergence rate of conventional BP neural network, and the fault diagnosis model based on the algorithm was designed. Fault character frequencies of pump were discussed. The process method using LabVIEW-based frequency extraction and neural network was investigated. The simulation results indicated that this method and model could remarkably reduce the training time, and the neural network was feasible for hydraulic pump fault diagnosis.
出处
《湘潭大学自然科学学报》
CAS
CSCD
北大核心
2009年第1期148-151,共4页
Natural Science Journal of Xiangtan University
基金
湖南省教育厅资助项目(08C872)