基于AR模型和神经网络的柴油机故障诊断  

Fault Diagnosis of Diesel Engine Based on AR Model and Neural Networks

作　　者：黄泉水[1] 江国和[1] 肖建昆[1] 

机构地区：[1]江苏科技大学机械与动力工程学院,江苏镇江212003

出　　处：《噪声与振动控制》2008年第3期60-63,91,共5页Noise and Vibration Control

基　　金：江苏科技大学人才资助项目(35020701)资助;企业协作技术攻关课题(校编2006JSB005)

摘　　要：基于虚拟仪器技术,采用NI公司PCI-4472采集卡及LabVIEW7.1开发柴油机缸盖振动信号采集分析系统。模拟柴油机气阀漏气、气门间隙异常等故障,并对利用该系统采集的缸盖振动信号样本进行AR谱估计,从中提取5个特征参数,最后利用RBF神经网络进行故障模式识别。结果表明:该诊断方法具有较高的精度,结合LabVIEW与Matlab的平台,便于故障在线监测与诊断系统的开发。Based on virtual instrument technology, the vibration signals acquisition and analysis system for diesel engine cylinder head was developed by LabVIEW7.1 and PCI-4472 Datacard. By simulating gas leakage and abnormal lash of diesel engine, the vibration signals samples was measured and the AR spectrum was computed by this system. Five parameters were extracted from the AR spectrum. Finally, the RBF neural networks was used to carry out fault mode identification. The results show that the method is high precision and it is effective to develop the on-line motoring and diagnosis system by Lab-VIEW and Matlab.

关 键 词：振动与波 柴油机 AR模型 故障诊断 RBF神经网络 LABVIEW 

分 类 号：TK421[动力工程及工程热物理—动力机械及工程] TP183[自动化与计算机技术—控制理论与控制工程]