分形在柴油机燃油系故障诊断中的应用
Fault Diagnosis of Diesel Engine's Fuel System Based on Fractals
-
摘要: 通过研究柴油机高压油管在喷油过程中,产生的振动信号中包含的丰富的燃油系工况信息,发现这一信号具有分形特征,即统计意义上的自相似性,类似一种1/f噪声.利用分形理论可以将其丰富的信息浓缩于信号的功率谱指数或容量维数中,并通过分析它的功率谱指数对燃油系各种工说的敏感性,进行燃油系的不解体诊断,实例诊断表明,它是一种较好的故障诊断方法。Abstract: The vibrations of diesel engines high pressure lines during injection time contains much information of the fuel system condition,the research shows that the signal has fractal characteristics, i. e. self-similar stochastic structure within the data. It belongs to 1/f noise family. The information of the signal can be concentrated in one parameter, Hurst Exponent or Power Spectrum Density (PSD) exponent. By analyzing the sensitivity of the PSD exponent in variant conditions of the fuel system, a non-intrusive fault diagnosis method to diesel engines is supplied.