Study On Fault Diagnosis Technology For Valve-train System And The Fuel System Of Diesel Engines

As power machinery, the diesel engine is widely used in industrial production. The diesel engine plays an extremely important role in realms of the national economy and the national defense. The valve-train system and the fuel injection system are main moving part of the diesel engine, which have working environment is more severe than other moving parts,so it is the site of frequent failure. Because the two systems determine the combustion quality of the diesel engine, so the fault diagnosis of the valve-train system and the fuel injection system has important significance. At present, the fault diagnosis of the distribution structure is done through the cylinder head vibration signal and the fault diagnosis of the fuel injection system is done through the high pressure tubing pressure signal. The feature extraction and the fault identification are the main links of the fault diagnosis,which determine whether the fault diagnosis is accurate or not.In allusion to the typical faults of valve and fuel systems, the methods of feature extracting and fault diagnosis are studied in this thesis. First, the mechanisms of faults diagnosis of valve and fuel systems are analyzed, then the experimental projects are designed based on the establish of experiment platform. Second, the vibration signals of cylinder head are measured both in normal and faulty conditions. On the bases of LMD, the measured signals are decomposed into PF components, and the energy-percentage and energy-entropy of PF components are extracted as the features for faults diagnosis of valve system. Similarly,the pressure waveform of high-pressure fuel pipe is detected in different conditions of diesel engine. Based on the wavelet packet analysis, the energy factors in high-frequency and high-frequency components of pressure, as well as injection pressure, and peak pressure, and the rate of pressure rise are extracted as the characteristic parameters of faults diagnosis.Finally, the application of support vector machines is studied in the field of faults diagnosis of diesel engine. The SVM classifier is constructed based on different conditions of diesel engine, and the working conditions and fault types of diesel engine will be got according to the input characteristic parameters. The results show that it is feasible to monitor the faults of valve and fuel systems of diesel engine through the methods of feature extracting based on LMD and wavelet analysis, as well as the fault diagnosis based on support vector machines.