一种煤矿带式输送机故障诊断方法

摘要

In view of problems of timeliness and reliability of fault diagnosis for coal mine belt conveyor are seriously affected by various fault types and the mutual influence of symptoms,a fault diagnosis method of coal mine belt conveyor was put forward.The method adopts fault diagnosis technologies combining with rough set and neural network,uses rough set attribute reduction algorithm to optimize input fault symptoms set,and obtains the optimal reduction set.The reduced minimum condition attribute set was input into BP neural network to train in a reasonable manner,and diagnosis decision rules was obtained through continuous learning and optimization.The reduced samples of the corresponding test symptoms set attribute were input into the trained network to diagnose fault,so as to identify corresponding fault.The simulation results show that the method can fully remove redundant information,speed up network training,and improve fault diagnosis accuracy of belt conveyor.%针对煤矿带式输送机故障种类繁多且各征兆存在交叉,严重影响故障诊断的时效性和可靠性的问题,提出了一种煤矿带式输送机故障诊断方法.该方法采用粗糙集与神经网络相结合的故障诊断技术,通过粗糙集属性约简算法优化输入的故障征兆集,得到最优约简集;将约简后的最小条件属性集输入BP神经网络进行合理训练,经过不断学习优化,最终得到诊断决策规则;将约简的相应测试征兆属性样本输入训练好的网络进行故障诊断,判别出相应故障.仿真结果表明,该方法能够充分删除冗余信息,加快网络训练速度,提高带式输送机故障诊断精度.