Implementation of Acoustic Emission technique in early detection of control valve seat leakage

Control valves are important components in control systems, which play a crucial role in ensuring plants run efficiently. The effect of valve failure is not only major production loss but also high maintenance cost. In a plant, a common technique used for valve maintenance is based on the valve condition replacement at fixed time interval. Most of the time the replacement only happens after the valve has obvious failure and during the plant shutdown. Common case of failure in valves is the internal valve seat leakage. The valve leak tightness test is conducted to detect the occurrence of internal leak in valve, however it is costly and time consuming to isolate the valve from the site and dismantling the valve. Significant money can be saved if leak in the valve can be detected early and if the test can be conducted online without having to isolate the equipment from the site. Therefore, this paper aims to study the effectiveness of Acoustic Emission technique in detecting early leakage in control valve on-site with fully closed position. The paper discusses a method based on different types of statistical analysis parameters such as variance, standard deviation, max amplitude on frequency domain analysis of AE signals to distinguish between healthy and unhealthy control valves. A real time AE measurement system is developed and tested. The acquired AE signatures are processed and analyzed using MATLAB. © 2014 IEEE.