Research On Key Technologies For Online Monitoring Of Valve Internal Leakage Based On Acoustic Emission

Valves play a critical role in the industrial sector by controlling the flow of fluid media,making their proper functioning essential for safe and reliable operations.However,valves often experience internal leakage due to prolonged exposure to harsh environments.Internal leakage can result in significant economic losses,environmental damage,and even casualties.Therefore,it is of great significance to develop an effective real-time monitoring system for valve internal leakage.This paper focuses on utilizing acoustic emission(AE)detection technology to construct a valve internal leakage monitoring system and addresses key aspects of monitoring valve internal leakage using AE technology.The main research contents of this paper are as follows:(1)Mechanism analysis of leakage acoustic emission signals in valves.This section examines existing valve internal leakage detection methods and highlights the feasibility and advantages of acoustic emission technology.It identifies the technical challenges of AE detection for internal valve leakage,providing guidance for subsequent work.The acoustic characteristics of valve leakage are analyzed based on fundamental principles,and relevant formulas for jet field source intensity are derived.This analysis establishes the theoretical foundation for acoustic emission detection research.(2)Development of resonant acoustic emission sensors with high sensitivity.The paper proposes a design for an acoustic emission sensor that employs PZT-5A piezoelectric ceramic as the sensitive element and alumina as the acoustic matching layer.The dimensional parameters of each component are calculated through theoretical analysis of acoustic impedance matching and resonant frequency.Finite element analysis using COMSOL verifies the theoretical analysis.The sensor is fabricated based on the design,and its sensitivity is evaluated using a dedicated test platform to generate a sensitivity curve.(3)Development of an acoustic emission monitoring system involves building the hardware system for acquiring acoustic emission signals.In this case,LabVIEW is used to develop the upper computer software for monitoring valve leakage acoustic emission(AE)signals.The software is programmed in a modular form to enable system control,data acquisition,data display,data storage,data playback,and other functionalities.To validate the system’s capabilities,damage monitoring and location verification experiments were conducted on an aluminum alloy plate.Standard lead breaking experiments were performed to simulate realistic acoustic emission sources.Signal processing techniques,specifically the SVD denoising method based on singular value decomposition and the wavelet threshold method,were combined to enhance the signal-to-noise ratio of the acquired acoustic emission signals.Through these experiments,it is observed that the designed acoustic emission acquisition system’s sensor exhibited a sensitive response to acoustic emission signals and provided a high signal-to-noise ratio.Additionally,the applied signal processing methods are proven effective.The average absolute positioning error,determined from 45 positioning tests,is found to be 4.51 mm.Overall,the developed acoustic emission monitoring system,including the hardware setup and software implementation,demonstrates the capability to capture and process acoustic emission signals with high sensitivity and improved signal quality.(4)Research on multi-classification recognition method of support vector Machine(SVM)based on multi-feature fusion of CEEMDAN reconstructed signals.An experimental system for valve internal leakage is established,and samples of different types of valve internal leakage signals and their characteristic parameters are collected.The CEEMDAN algorithm is used to filter environmental noise,and highly correlated signal features are analyzed and selected.Multi-feature fusion is applied to the reconstructed signals,and the fused characteristic vector is utilized as input for SVM-based signal recognition and classification.The achieved signal recognition rate reaches 97.5%,validating the effectiveness of the proposed method.This paper focuses on practical engineering aspects,utilizing valves as the research object.It includes the design of a high-sensitivity resonant sensor,the development of LabVIEW-based acoustic emission monitoring software,and research on a multi-classification recognition method using SVM and multi-feature fusion of CEEMDAN reconstructed signals.The successful implementation of this research allows for accurate recognition of 12 types of valve internal leakage signals,providing significant engineering application value for ensuring the safe operation of valves.