Font Size: a A A

Diesel Engine Vibration Signal Feature Extraction Method And The Cylinder Pressure Signal Reconstruction Method

Posted on:2008-07-16Degree:MasterType:Thesis
Country:ChinaCandidate:J TangFull Text:PDF
GTID:2192360212993451Subject:Power Machinery and Engineering
Abstract/Summary:PDF Full Text Request
In recent years, the technique of diesel engine fault diagnosis, which improves the performance of diesel engines and reduces the economic loss, is quickly developed in domestic and oversea. In this paper, in order to solve practice problems, we adopt several methods to analyze the vibration signal and pressure signal on the basis of former achievements. Using these methods we obtained signal characteristic information for faults and realized diesel engine fault diagnosis.Firstly, from the starting point of vibration signals formed, the characteristics of actuation sources and their responses and the relations between them are discussed in detail. Meanwhile radiate routes of vibration signals are analyzed. The analysis is the theoretical basis for diesel engine diagnosis.Secondly, this paper implements the valve clearance, the valve leakage, the injection pressure and the injection advanced angle faults on the test bench. And the vibration signal, the cylinder pressure signal and the crack angle speed were tested and saved. According to the non-stationary feature of the vibration signal of the engine, the signal was mainly analyzed by the wavelet technique. First the signal noise was reduced using the wavelet method so that the signal-to-noise performance was improved. Then the research using wavelet packet decomposes the vibration signal and obtains the time-frequency domain characteristic parameters. Also this paper uses the time or frequency domain methods to obtain the characteristic parameters on the time or frequency domain. By the analysis of the signals, the tendency of the characteristic parameters was generalized. At the mean time, the cylinder pressure signal which is the result of the combustion and is the inversion basis from the heat energy to the mechanical energy was analyzed. In order to identify engine faults more correctly, this paper combines both the vibration and the cylinder pressure characteristic parameters.At last, this paper studies the methods that how to reconstruction the cylinder pressure signal. At present, the cylinder pressure signal was almost obtained by direct measurement. As direct measurements of the cylinder pressure were restricted by many conditions and not suitable for measurements in vehicles on the road, indirect methods have great potential value for it is economy and convenience. There are two methods for reconstruction, one is using the vibration signal, the other is using the speed signal. Accord to the non-linear feature between the speed signal, vibration signal and the in-cylinder pressure signal, a non-linear model based on BP networks is proposed for the reconstruction. By comparing these methods, the features of the methods were obtained. And the method that combining the speed signal and the vibration signal was proposed. The method makes up for the disadvantages of the former methods. At last, the reconstruction signal was valuated by the parameters. The result shows that the method performs well for reconstruction the cylinder pressure signal.
Keywords/Search Tags:diesel engine, cylinder pressure, vibration signal, wavelet analysis, BP network, fault diagnosis
PDF Full Text Request
Related items