Design And Implementation Of Acoustic Emission Leak Detection Analysis System For Sealed Pressure Vessel

With the continuous improvement of production technology in the world,the application of sealed pressure vessels has reached a new height.During the use process,due to cracking,collision,friction,corrosion and other conditions,it is easy to cause container leakage,which will seriously impact on people’s economic and life safety.Acoustic emission is a physical phenomenon due to the rapid release of instantaneous elastic waves by a material.The mechanical vibration signal caused by leakage of a sealed pressure vessel is an acoustic emission signal.Aiming at the problem of safety accidents caused by leakage of sealed pressure vessels,combined with domestic and foreign acoustic emission leak detection methods,a set of sealed emission vessel acoustic emission leak detection and leak identification system was designed.Mainly include the following research contents:Researched the generation mechanism of the leaked acoustic emission signal,established a sealed pressure vessel acoustic emission signal detection model,and comprehensively analyzed the relevant characteristics of the signal before,immediately,and after leakage.The overall scheme design of the system is realized,and key technologies such as characteristic parameter analysis,leak detection,feature extraction,and leak classification identification of leaked acoustic emission signals are mainly studied.Aiming at the problem of real-time leak detection of sealed pressure vessels,we developed a multi-channel,high-sampling rate acoustic emission leak detection software,which realized the modular design of the software functions,including user management,system settings,high-speed collection and storage,waveform display,characteristic parameter display,3D energy column display,real-time spectrum analysis,real-time leak alarm and other functional modules.The function of each module is independent,which can reduce the coupling relationship between the modules and improve the applicability and portability of the system.Aiming at the classification and identification of acoustic emission signal leaks in sealed pressure vessels,a software algorithm for acoustic emission signal EMD feature extraction and SVM classification identification is proposed.The leaked acoustic emission signal is a high-noise,weak signal,non-linear,non-stationary signal.Using EMD to separately decompose the acoustic emission signals of various leak sizes into multiple IMF components,combined with characteristic parameter analysis and HHT Characteristic distribution and internal connection of ringing count,amplitude,energy and HHT marginal spectrum frequency of different leaky signals.The weak acoustic emission signal feature extraction under strong background noise is realized,and the feature vectors of multiple sets of different leak sample signal data are trained by SVM to obtain the acoustic emission signal SVM prediction model.Finally,taking the sealed pressure vessel as the research object,an experimental platform of acoustic emission leak detection and analysis was built for experimental verification.Run the software to collect and store the acoustic emission signal in real time.After the system analyzes the signal characteristic parameters,it performs real-time data display and leak alarm,and tests the historical data query and export functions.Experimental verification of EMD feature extraction and SVM classification recognition of multiple leaky acoustic emission signals.The experimental results proved the practicality and effectiveness of the acoustic emission signal leak classification and identification method.