Study On Fault Diagnosis Of Fuel Injection System In Diesel Engines Based On Visual Instrument

Diesel engines are presently one of the most prevalent-applied dynamic facilities and their oil-heating systems are further significant in the engines as integrity on account that the operating condition directly affects the normal running of diesel engines with determination of their running performances. In other words, the dysfunction of fuel injection systems would apparently deteriorate the diesel engine performances. Approximately one third of the diesel engine malfunctions are brought about by the failures of oil-heating systems, which would probably lead to the fuel-combusting deterioration of diesel engines, the reduction of dynamic, economical and unfailing performances and the increase of harmful emissions. Therefore, the timely diagnosis for the malfunction of the oil-heating systems is of great importance. However, the compressed fluid as operation media in the diesel-heating systems makes greatly high running pressure, various system configurations and complicated relations between diverse interactive parameters. Accordingly, the conventional malfunction diagnosis method of reduplicated testing the diesel engines on the testing-stands requires much manpower and material input, prolongs the research cycles and couldn't meet the demands of modern production with high efficiency.hi this paper, it is introduced the technologies of signal acquisition, signal wavelet analysis, waveform recognition, neural networks, visual instruments all applied in the malfunction diagnosis of fuel injection systems. Based on the study and analysis of the principles, characteristics and application examples of these techniques, the malfunction diagnosis testing system for fuel injection systems on visual instruments was put forward and applied with desirable effects.The principal study contents and conclusions are as follows:(1) According to the function, structure, operation process and mechanism of the diesel injection systems, the method of detecting the pressure waveforms in high-pressure fuel tubes was studied to diagnose the failure of diesel injection systems. The results show that the pressure waveforms in high-pressure fuel tubes contain most operation condition information and could recognize sensitively various malfunctioninformation.(2) With the introduction of visual instrument technology in the intelligent diagnosis for the malfunctions of the diesel injection system, the failure-diagnosis software platform was developed in Lab VIEW using the waveform signals acquired in the high-pressure tubes by the data-acquisition board. The connection of the visual data-acquisition board with computers was either through USB ports or PCI boards and thus the signal acquisition could realized via multipath. Moreover, changing or adding the acquiring information was convenient by the method of changing the sensors and the program settings. In addition, LabVIEW is prevalently used as a type of programming language as a result of its characteristics of visualization and modularization. The wavelet analysis and BP neural network were processed in MATLAB called by LabVIEW.(3) The dispersion wavelet analysis was applied to disintegrate the pressure waveform in high-pressure oil tubes and one approximation waveform component a3 at the low frequency and three others dl, d2, d3 at the high frequency was consequently decomposed. The characteristic parameters in time domain were extracted by the recognition of a3 and the low-frequency energy factor and the high-frequency energy factor were disintegrated by calculation of a3 and d3. The results indicate the wavelet analysis is an effective method for processing the pressure waveform in the high-pressure oil tubes maintaining characteristics in both time and frequency domains;(4) With reference to the compressed pressure characteristics of high-pressure tubes, the theory was put forward to recognize the waveforms of low-frequency approximation signals disintegrated by wavelet analysis and select 8 parameters, i.e. the response pressure of oil ejectors, the maximum injection pressure, the oil pressure rise rate, the aftereffect peak value, the fuel injection holding angle, the affereffect wave width, the low-frequency energy factor and the high-frequency energy factor. These parameters all effectively reflected the operation condition of fuel injection.(5) The BP neutral network is a type of large-scale parallelism nonlinear...