The Application Research On Acoustic Emission Technique For Fault Detection Of Valve Leakage

Valve is a sort of indispensable equipment used for controlling fluid flow in power plant. The efficiency of thermal power units would be greatly reduced, or even there is a direct threat to the safety operation of staff and units, as soon as an important valve leaks. At present, all the leakage detection methods of power plant valve have certain limitations in on-time and accurate diagnose of valve operation fault. And thus, there is an important significance and broad application prospect to research the real-time quantitative detection and diagnosis of the power plant's valve leakage fault.The detection method of valve leakage fault based on acoustic emission technology is mainly studied by combining theoretical analysis and experimental research in this paper, and then a type detection system of valve leakage fault is designed and developed.First, the mechanism, characteristics, and detection method of acoustic emission signals generated from valve leakage are discussed on the basis of previous studies. And a new calculation method of the valve leakage rate based on the features of acoustic emission signals is obtained, which provides the important theoretical foundation for quantitative diagnosis of valve leakage faults.Secondly, a simulation experiment platform of valve leakage fault and a detection system used to detect acoustic emission signals generated from valve leakage are set-up, and then the quantitative relationship between acoustic emission characteristic parameters and internal valve leakage rate is studied under different operating conditions with experiment method. The change of AE parameters, counts, RMS, energy, amplitude and dominant frequent component, with the variation of valve sizes, valve types, inlet pressure levels and size of leak, are obtained after the experimental data analysis and process based on Generalized Least squares method. Thus, the best AE parameters, counts, RMS and energy, which can characterize the internal fluid leakage through valve are found and the rules for valve leakage failure diagnosis based on acoustic emission detection are built. Simultaneously, the quantitative relationship between valve leakage rate and the characteristics of acoustic emission signal, RMS, is researched, and the calculation formulas of valve leakage rate of the three types of valve tested in lab are obtained, which put the quantitative diagnostic theory of valve leakage fault into practice.Finally, in this thesis, on the base of theoretical and experiment research, a type detection system of valve leakage fault is designed and developed using LabVIEW as developing platform. The specific testing index of AE characteristics are set up in the system according to the acoustic emission signal trait of valve leakage fault, with which the noise can be filtered out adaptively and the quantitative diagnosis of valve leakage rate and grades alarm are realized. It is proved that the quantitative detection performance of the detection system of valve leakage fault is quite well, with sensitive reaction and less than±16% detection error, by applying the system in lab and power plants. Compared to the PCI-2 AE testing system, the system developed by author of this thesis is more simple, visual and targeted.