The Method Of The Valve Leak Detection Based On Acoustic Emission

The leakage detection is one of the most important parts of the work in valve's manufacturing and operating. However, many domestic manufacturers are using an inefficient bubble method of leak detection generally. For this actuality, the acoustic emission (AE) method was researched in the article, which could respond quickly to the valve leakage rate. Through the theoretical and experimental analysis, the AE method was feasible and could make qualitative analysis quickly and easily for tiny leak on the valve, which provided rationales to realize automatic detection of the valve leakage.Firstly, various methods of leakage detection commonly used were analyzed and contrasted in their principles and characteristics in this dissertation, it was learned that the AE method was the fittest method for realizing automatic detection of the valve leakage, when the valve leakage rate was above 10^-4 Pa·m³/s. The jet's velocity field and sound field generated by valve leak were studied in theory. By FLUENT simulation analysis, it was found that the input pressure and valve port size had huge impact for velocity field, and then affected the distribution of the jet sound field.Then, a valve leakage rate measurement system was built using AE method based on the principle of valve leakage generating AE signals, which could be used to test the valve leakage higher than 10^-3 Pa·m³/s, and used an uniform air bubble method to calibrate the valve leakage in the system. After lots of experiments, it was shown that the calibration method could detect the valve leakage accurately. For the signal detection problem of susceptible to the environmental and electrical noise, the conditional circuit was designed, and the undecimated wavelet transform (UWT) was used to reduce the effect of the noise.Furthermore, a LABVIEW program was compiled to analyze and process the capturing AE signals. The characteristic quantities containing the root mean square, the fast Fourier transform (FFT) power spectrum, the FFT power density spectrum and the relative size of the amount of energy in each frequency band, could be used to characterize the valve leakage rate after the time-domain analysis, the spectral analysis and the wavelet packet analysis.Finally, after lots of experimental statistics, it was shown that the valve leakage rate obtained with the exponential function of the input pressure. As the input pressure was only one of the influencing factors of the valve leakage, it was hardly to found only one exponential function to express the relationship between them in any state of leak. Beyond that, both of the leak detection position and the input pressure could impact the strength and the spectrum distributions of the AE signals.