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 Auto Regressive Model and Neural Network

作　　者：黄泉水[1] 江国和[2] 肖建昆[2] 

机构地区：[1]中国船舶科学研究中心,无锡210482 [2]江苏科技大学机械与动力工程学院,镇江212003

出　　处：《振动．测试与诊断》2009年第3期362-365,共4页Journal of Vibration,Measurement & Diagnosis

摘　　要：建立了一种基于AR与RBF神经网络结合的诊断模型,模拟柴油机气阀漏气、气门间隙异常等故障,采用NI公司PCI-4472采集卡在LabVIEW7.1平台上开发了柴油机缸盖振动信号采集分析与诊断系统。首先,对利用该系统采集的缸盖振动信号样本建立AR模型并进行AR谱估计,从中提取5个特征参数,然后利用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,便于故障在线监测与诊断系统的开发。A fault diagnosis method based on the auto regressive(AR) model and the neural networks was put forward. Through the simulation of the faults such as the gas leakage and the abnormal lash of a diesel engine,a diagnosis system for the cylinder head where the faults occured frequently was developed by using the N1-4472 signal acquistion module on the LabVIEW platform. An AR model was established using the vibration signal samples from the cylinder head,and the AR spectra were estimated and five feature parameters were extracted. Then,a radial-based-function(RBF) network was used to distinguish the fault type. Results show that the mehtod is effective for developing the on-line monitoring and diagnosis system of the diesel engine.

关 键 词：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]