Compressor Fault Diagnosis System Based On Virtual Instruments

Compressor is being widely used in many kinds industries, if it breaks down during its working process, that would made product line or product machine stop and let company has a big economic loss, even it would threaten worker's life. So, it's important to research the theory of compressor's state monitor and fault diagnosis and built a perfect monitor and fault diagnosis system. In the main, this article researches screw compressor's vibration signal and built a compressor fault diagnosis system based on LABVIEW.First, this article researches the screw compressor's work theory and structure, point out some screw compressor's faults that would occur usually. Based on the research, this article analyzes the way of get the compressor's vibration and some important points that we should care about. The main analyze methods included in this article as follows:1. Using three parameters as RMS,kurtosis and margin to reflect compressor's state.2. Using wavelet energy to reflect compressor's state. Wavelet energy distributes differently under different environments, when the compressor's state is normal, it distributes uniformity, when its state is abnormal, wavelet energy distributes in a certain bound.3. Using the method of HHT to analyze vibration in the frequency field. FFT is used to analyze Linear signal, but there are many exciter, vibration signal is usual a kind of non-linear signal. HHT is a new and useful analyze method in deal with un-linear signal.4. Research two aptitude methods: AR model and support vector machine(SVM).they point out a new way for compressor's research.Combining the LABVIEW and MATLAB using the MATLAB script provided by LABVIEW,Design a compressor faults diagnosis system based on the compressor faults mechanism and research of diagnosis method. The major functions as follows: data acquisition,dynamic display,real-time monitor,signal analysis(power spectrum,correlation analysis,wavelet analysis,time-frequency analysis,HHT),state forecast based on AR model,fault diagnosis based on SVM,abnormal signal and research result playback; At the same time, the system can monitor the temperature of air suction,air exhaust,lubricating oil,cooling water and pressure of air suction,air exhaust and alarm when these parameters exceed the limited value.At last, Through the test on the text-bed, the results indicate all parts can work and diagnosis the machine's states well, it show that the system is effective and applied.