Extraction Of Valve Clearance Fault Signal From A High Power Density Diesel Engine

As a common power device,diesel engine plays an important role in the national economy and people's daily life.Its structure is complex and its working status is changeable.According to statistics,in the cause of diesel engine failure in a variety of reasons,the proportion of the fault of the valve is higher.Based on the actual demand of diesel engine fault diagnosis,this paper chooses the valve clearance that affects the combustion of diesel engine as the research object,analyzes the fault diagnosis method of the valve clearance of diesel engine by analyzing the cylinder head vibration as the breakthrough point.The high power diesel engine contains a lot of motion mechanism and its structure is compact and complex.During the working process,the vibration excitation source is many,the frequency distribution is wide,the working time is long,and the load is high.If the direct measurement method is used to get the vibration signal,it is a typical non-stationary signal.At present,most of the vibration signal feature extraction methods for internal combustion engines are suitable for simple fault diagnosis of some single cylinder and less excitation sources.But the vibration signal processing and extraction effect of high power diesel engine is not very ideal.Aiming at the signal of large power diesel engine is a typical nonlinear and non-stationary signals when it works,this paper designed and developed high-power diesel mechanical vibration signal collecting device,using Empirical Mode Decomposition(EMD)and Polymeric Empirical Mode Decomposition(EEMD)method to preprocess the valve clearance fault signal of high-power diesel engine.Then,a variety of information entropy is used to extract the feature.The experiment shows that the original signal of the polymerization mode decomposition can get more effective characteristic parameters.In the end,a support vector machine,which is very suitable for small sample fault identification,is used to identify the fault mode.The experimental scheme designed in this paper is conducive to the acquisition and analysis of vibration signals of the internal combustion engine.The correlation coefficient obtained through the aggregation of empirical mode decomposition has strong distinguishability and can effectively distinguish the real IMF component from the IMF component containing noise,and the real component is reconstructed to reduce the noise.Analysis of the vibration signal of valve clearance in diesel engine shows that: EEMD has a very good effect on solving non-linear and non-stationary problems,and after EEMD decomposition,the IMF has highlighted the different local characteristic information of the original signal,and through the analysis of the IMF,it extracts the characteristic variable that can describe the fault information,which is the the energy ratio can effectively describe the characteristics of different vibration faults,combined with the multiple information entropy feature,can accurately identify and diagnose the valve clearance of high-power diesel engine,and provide a new method for feature extraction and fault diagnosis of other complex faults of internal combustion engine.