基于模拟退火算法的机车齿轮箱故障诊断系统  

System Fault Diagnosis of Electric Locomotive Gear Box Based on Simulation Annealing

作　　者：范万里[1] 李刚[1] 白宇君[1] 高晓玲[1] 

机构地区：[1]兰州交通大学机电工程学院,甘肃兰州730070

出　　处：《石家庄铁道大学学报（自然科学版）》2013年第3期70-73,共4页Journal of Shijiazhuang Tiedao University(Natural Science Edition)

基　　金：青藏铁路公司科委科技研究开发计划(QZ2010-J02)

摘　　要：为了预测电力机车齿轮箱存在的潜伏性故障,设计了一种电力机车齿轮箱故障诊断系统。该系统通过VB调用Access数据库和基于模拟退火算法的BP神经网络模型,实现了对电力机车齿轮箱故障诊断的可视化操作。根据光铁谱技术原理,磨粒浓度的变化体现了齿轮箱内部的故障类型。因此,利用模拟退火的思想建立BP网络模型,收集电力机车齿轮箱的故障数据并进行归一化处理,处理后的数据作为网络的输入,故障类型的编码作为网络的目标输出。对模型进行了仿真和测试,结果表明该模型诊断电力机车齿轮箱故障准确率较高,可用于电力机车齿轮箱故障诊断系统中。The system fault diagnosis of electric locomotive gear box is designed in order to predict the potential fault of electric locomotive gear box, which realizes visualized operation of electric locomotive gear box fault diagnosis. The system calls the access database and the bp neural network model based on simulation annealing. According to the principle of RDEAES and Ferrography, the change of abrasive concentration reflects the type of gear box internal fault. Therefore, the bp network model is established to use the idea of simulation annealing, collect the fault data of electric locomotive gear box. The processed data act as the input of network and the codes of fault type act as the target output of network. The simulation and test of the model show that the model has a higher accuracy for electric locomotive gear box fault diagnosis and can be used in electric locomotive gear box fault diagnosis system.

关 键 词：电力机车 齿轮箱 BP神经网络 模拟退火算法 故障诊断 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]