基于气缸盖振动信号的柴油机故障诊断研究

在3110柴油机的1号缸上进行了气门间隙变化、断油等故障的模拟试验研究.分别在正常情况及各故障情况下测取了缸盖振动信号,对信号进行时域分段,分别用时域、频域及小波分析方法对信号进行了分析,提出了故障时频特征的提取方法,并用模糊聚类方法进行了分类分析.对正常情况(气门间隙0.3 mm)、断油故障、气门间隙过小(0.1 mm) 、气门间隙过大(0.9 mm)等4种情况的80组数据的分析结果表明,大部分数据分类效果较理想,错误率仅为6.25 %, 断油故障数据分类完全正确.对进气门间隙为0.1 mm、0.3 mm、0.9 mm的3种工况的60组数据进行了模糊聚类分析,取进气门落座信号,用绝对值均值、有效值及峰值作为特征值,分类结果完全正确.

Research on Fault Diagnosis of Diesel Engine Using Vibration Signals of the Cylinder Head

A lot of artificial diesel engine faults were simulated and investigated on first cylinder of the 3110 diesel engine.The vibration signals of cylinder head surface were obtained under the each fault conditions,such as abnormal valve clearance,fuel cut-off and gas leakage,etc.The vibration signals were partitioned,and its characteristic parameters in time domain and frequency domain were extracted based on wavelet theory and fuzzy cluster theory.The results are satisfactory through analyzing 80 vibration signals under four conditions,such as normal condition(valve clearance 0.3 mm),fuel cut-off condition,smaller valve clearance condition(0.1 mm) and larger valve clearance condition(0.9 mm).The classified results of most data are accurate,the error rate is only 6.25 %,and the classified results of data for fuel cut-off condition are very accurate.The classified results are fully accurate when analyzing the 60 vibration signals under different valve clearance condition(0.1 mm,0.3 mm,0.9 mm) by using absolute mean value,effective value and peak value of intake valve dropping signals as characteristic parameters.