Neural Network Approach To Automated Condition Classification Of A Check Valve By Acoustic Emission Signals

This paper presents new techniques under development for monitoring the health and vibration of the active components in nuclear power plants. The purpose of this study is to develop an automated system for condition classification of a check valve one of the components being used extensively in a safety system of a nuclear power plant. Acoustic emission testing for a check valve under controlled flow loop conditions was performed to detect and evaluate disc movement for valve failure such as wear and leakage due to foreign object interference in a check valve. It is clearly demonstrated that the evaluation of different types of failure types such as disc wear and check valve leakage were successful by systematically analyzing the characteristics of various AE parameters. It is also shown that the leak size can be determined with an artificial neural network.