基于组合式改进BP神经网络的矿用空压机故障诊断  

Fault Diagnosis of the Air Compressor for Mining Based on Modular Combined Improved BP Neural Network

作　　者：张强[1,2,3] 张书会[1] 张楠 

机构地区：[1]辽宁工程技术大学机械工程学院,辽宁阜新123000 [2]工业装备结构分析国家重点实验室,大连116023 [3]中国煤矿机械装备有限责任公司 [4]沈阳煤业(集团)有限责任公司,沈阳110122

出　　处：《机械设计与研究》2013年第1期110-113,共4页Machine Design And Research

基　　金：辽宁省教育厅科研计划资助项目(L2011050);中国煤炭工业协会科学技术研究指导性计划资助项目(MTKJ2010-310);工业装备结构分析国家重点实验室开放基金资助项目(GZ1107)(GZ0818)

摘　　要：矿用空压机是煤矿现代化安全生产的重要组成设备,针对目前矿用空压机经常出现的故障,收集了矿用空压机故障征兆和其对应的故障类型,将故障样本数据和模糊神经网络相结合,并根据改进BP神经网络确定网络的输入和输出向量,对矿用空压机进行组合式故障诊断,诊断结果与实际情况比较吻合。运用MATLAB实现神经网络故障诊断仿真,仿真结果表明诊断误差较小,输出向量与实际故障矩阵结果接近。The coal mine air compressor is an important equipment for safe production. Aiming at the faults occurring currently and frequiently in air compressors, so their symptoms and corresponding fault types were collecting. Combined fault sample data with fuzzy neural network, the network input and output vectors were determined according to the inlproved BP network, so the fault of air compressor is diagnosed. The diagnosis is consistent with the reality. Using the MATLAB to simulate neural network fault diagnosis and the simulation results show that the diagnosis error is small, and the output vector and the actual fault matrix results are approximate.

关 键 词：矿用空压机 改进BP算法 故障诊断 神经网络 

分 类 号：U664.21[交通运输工程—船舶及航道工程]