摘要
采用小波网络方法,通过对矿井提升机钢丝绳磨损度、空动时间、衬垫磨损寿命、闸瓦间隙、残压、制动盘偏摆度等关键特征参数的时间序列预测,实现了其特征参数的故障预报.由于小波网络比一般神经网络具有更多的自由度,从而使其具有更灵活有效的函数逼近能力.小波神经元的良好局部特性和多分辨率学习实现了与信号的良好匹配,使得小波网络有更强的自适应能力、更快的收敛速度和更高的预报精度.仿真和实验结果表明,预报精度满足要求.
In this paper,adopting wavelet neural network, fault forecast of characteristic parameters of mine hoist are realized by forecasting time series of key characteristic parameters of mine hoist that include abradability of steel wire rope, time of idle motion, life of pad wear away , clearance of brake shoe, remnant oil pressure and deflection degrees of brake disk. With more degrees of freedom in relation to the general neural network, wavelet neural network is of more vivid and valid ability in function approximation. With the good partial characteristic and distinguish rate learning, wavelet neural network matches commendably the signal, and has stronger self adaptation ability, more sooner convergence rate and higher forecast accuracy. Emulational and experimental results show that the forecast accuracy of wavelnet neural network may meet industrial demand.
出处
《中国矿业大学学报》
EI
CAS
CSCD
北大核心
2005年第4期528-532,共5页
Journal of China University of Mining & Technology
关键词
提升机
特征参数
小渡神经网络
故障预测
mine hoist
characteristic parameter
wavelet neural network
fault diagnosis