摘要
齿轮故障诊断的基本方法是利用振动信号,通过提取特征信息,比较各次测量中的差异进行诊断。根据BP神经网络诊断方法,运用虚拟样机技术建立齿轮模型,模拟各种各种故障,提取振动信号,然后给出BP神经网络训练的特征频率,使其对数据进行分析学习。当输入待诊断样本后,BP网络对数据进行对比分析得出诊断结果。经过仿真检验表明该方法对减速箱齿轮副的故障诊断有效,对其他旋转机械的故障诊断和维修保养也具有一定的实用价值。
The gear fault diagnosis method is based on the vibration signal,through extracting the feature information,comparing the measurements difference diagnose.Based on the BP neural network diagnosis method using virtual prototype technique and gear model is built,simulating various fault,the vibration signals are extracted;and then the training of BP neural network characteristic frequency is given,the data is analyzed and studied.When the diagnostic samples are input,BP network is compared and analyzed for data,and the diagnosis results are got.Through the simulation test,it shows that the method is effective for gear fault diagnosis in the reduction gear box,and also has a certain practical value for other rotating machinery fault diagnosis,repair and maintenance.
出处
《机械研究与应用》
2012年第6期170-172,174,共4页
Mechanical Research & Application
关键词
BP神经网络
时域平均分析法
齿轮副故障诊断
仿真分析
BP neural networks
analysis method of time domain average
fault diagnosis of geers
simulation analysis
