The Study On Intelligent Fault Diagnosis System Of Diesel Engine Based On Virtual Instrument

The diesel engine has a series of performance, such as the good economy fuel, the reliable work, the good durability, the wide power scope, the slightly exhaust pollutes, for this, it is widely used not only on the heavy automobile and the project machinery, but also on the medium and small scale automobile, even more application on the passenger vehicle day by day. The development on the research of diesel engine intelligent fault diagnosis, may safeguard the production, reduce or avoid the heavy economic loss and the person injury for emergency; it also can help the serviceman to reduce the overhaul time, and improve the service quality.Artificial neural network has the unique capacity on handling the non-linearity compatible information, and it has overcame the information processing flaw by the traditional artificial intelligent method, so it is successfully used in the nerve expert system, the pattern recognize, the intelligent control, the combination to optimize, estimate and other fields. The BP neural network is the core part of front network, and it has manifested the most essence part of artificial neural network.With the great development and application of micro-electronics technology, computer technology, software technology and internet technology in the area of measurement technique and instrument, many traditional concepts of measurement instrument have been breakthroughed. Some parts of hardware in traditional instruments have been replaced by virtual instrument technique, which takes the software technique as core. The application of this technology can not only save your both valuable development time and resources, but also reduce the difficulty of design, so it brings the measurement technology to a new age.In this paper, we used the graphical programming language LabVIEW which NI Corporation develops, and called the MATLAB program by the MATLAB script node through LabVIEW, constructed a set to measure and to analyze the system of engine vibration based on the virtual instrument, which can complete the engine vibration signal data acquisition, the signal denoise, the signal characteristic extraction by wavelet packet and the intelligent fault diagnosis. In the paper we compared three kind interface driver explore methods for the general data acquisition card, finally we determined write the interface driver program of UA301 data acquisition card by calling the dynamic link library. The method, which is used easily and feasibly, runs stably and quickly. The intelligent fault diagnosis key lies in the extraction of fault characteristic, in the article signal sample is analyzed through wavelet packet theory, and the energy ratio of each sub-frequency band to occupy the total frequency band is calculated, at last 8 links of characteristic vectors are got. The fault of the sample can be discerned by trained BP network of four layers, which is structured on the basis of the characteristic vector, and the discernment obtains a good effect. It proved that extracting symptom with method of wavelet packet theory is efficiency, and the application of BP network in fault diagnosis is also feasible.