Dissertation > Industrial Technology > Energy and Power Engineering > Internal combustion engine > Diesel engine > Repair and maintenance

Study on Cylinder Pressure Recognition Method of Diesel Engine Based on Cylinder Head Vibration Signal

Author LiuShengXiRui
Tutor FeiHongZi
School Harbin Engineering University
Course Power Machinery and Engineering
Keywords diesel engine cylinder pressure recognition transfer function method RBF method local wave method
Type Master's thesis
Year 2013
Downloads 18
Quotes 0
Download Dissertation

Due to the difficulties to directly measure the pressure of the cylinder, identifyingdiesel engine cylinder pressure through other easily measured signal have a veryimportant positive significance to state detection and fault diagnosis. Here we havechosen the cylinder head vibration signal to identify the cylinder pressure. The mainresearch work of this dissertation is summarized as follows.Firstly, analysed the cylinder head vibration excitation sources, selected theidentification methods, designed the cylinder pressure identification method researchprograms based on cylinder head vibration signal, and carried out the experiments.After the analysis of cylinder pressure in the frequency domain and cylinder headvibration signal preprocessing, firstly the author established a linear system model andbased on the transfer function method to identify the cylinder pressure. Therecognition results show that, due to the cylinder head system having a nonlineartransfer characteristic, resulting in the recognition results are not very accurate basedon the linear model inevitably lost information. Therefore, the author established anonlinear system model, based on the RBF (radial basis function neural network) toidentify the cylinder pressure. The results show that, the RBF method overcomes thedefects of losing information, and gets accurate results, but its model can not beupdated in real time, so there are certain defects in adaptability and promotion.Due to the realization of these two pressure recognition methods must knowcylinder pressure in advance, therefore, the key of this study is to identify the cylinderpressure in the case of unknown pressure, which is local wave method. To overcomethe lack of local wave method recognition accuracy, the author combined waveletnoise reduction and local wave method. The results show that, although thisrecognition accuracy is not ideal, but this method does not have to know pressure inadvance. It can be said to be a fundamental solution to identify cylinder pressure, so itwill have a very good practicality and broad application prospects.

Related Dissertations
More Dissertations