| As a major thermal power source of of industrial, vehicles, ships and mining, diesel engines diagnosis technology become the focus of research in recent years. So its diagnosis technology was more mature and widely. A design of on-line testing integration of host computer and console computer is proposed in this paper.And the fault diagnosis of diesel engine is achieved by algorithm of LMD-SVM.This design of diagnostic system’s console computer based on S3C6410 processor of ARM11, its vibration sensor is ADXL345, its operating system is Linux. This devise can store vibration data, and it also can tarnsmit the data to console computer by its serial ports.This design of diagnostic system’s host computer mainly used the tool of MATLAB. An user Interface which is designed by MATLAB GUI is used to control and display the dynamic vibration data. The algorithm of LMD-SVM is designed by script program, was used to pattern classification and fault diagnosis.The working principle of LMD-SVM algorithm is that divide the original signal into several PF components, and solve every approximate entropy of PF component, extract several approximate entropy which can be the feature vector, put the training samples into the support vector machines to be trained.Then we put the testing samples into SVM to be sorted by mode, achieved fault diagnosis last.Finally, R6105 AZLD diesel engines as main research object, and selected 6 conditions. The online collection of signal was achieved by building a diesel engine test platform, debugging host and console computer. There are 180 group of test samples and its passed rate is 91.67%.The experiment present that this mothod is useful in engineering. |
