摘要
将基本BP算法的改进算法(LMBP算法)利用数值分析的方法进行改进,并将其应用到矿井提升机的故障诊断系统中,改进后的算法能提高网络的训练速度;同时设计了基于ARM芯片和Vxworks操作系统的故障诊断板,并利用Borland C++ builder作为故障诊断界面的设计软件,将新集集团某矿井提升机液压站系统作为实验室仿真研究对象。结果表明,该故障诊断系统具有较好的硬件稳定性,较快的神经网络训练速度,较好的诊断结果。
LMBP algorithm was improved by using BP algorithm and applied into mine rotating machinery fault diagnosis system. The algorithm improved training speed. Designation plan of data collection board based on ARM and Vxworks operating system was designed. The interface of fault diagnosis was developed based on Borland C ＋ ＋ Builder. The hydraulic station of one rotating machinery of Xinji Company was taken as simulation research object in the lab. The results showed that the stability of hardware, the training speed of ANN and diagnosis results are good.
出处
《安徽理工大学学报（自然科学版）》
CAS
2011年第2期75-78,共4页
Journal of Anhui University of Science and Technology:Natural Science
