An impact source localization technique for a nuclear power plant by using sensors of different types.

In a nuclear power plant, a loose part monitoring system (LPMS) provides information on the location and the mass of a loosened or detached metal impacted onto the inner surface of the primary pressure boundary. Typically, accelerometers are mounted on the surface of a reactor vessel to localize the impact location caused by the impact of metallic substances on the reactor system. However, in some cases, the number of accelerometers is not sufficient to estimate the impact location precisely. In such a case, one of useful methods is to utilize other types of sensor that can measure the vibration of the reactor structure. For example, acoustic emission (AE) sensors are installed on the reactor structure to detect leakage or cracks on the primary pressure boundary. However, accelerometers and AE sensors have a different frequency range. The frequency of interest of AE sensors is higher than that of accelerometers. In this paper, we propose a method of impact source localization by using both accelerometer signals and AE signals, simultaneously. The main concept of impact location estimation is based on the arrival time difference of the impact stress wave between different sensor locations. However, it is difficult to find the arrival time difference between sensors, because the primary frequency ranges of accelerometers and AE sensors are different. To overcome the problem, we used phase delays of an envelope of impact signals. This is because the impact signals from the accelerometer and the AE sensor are similar in the whole shape (envelope). To verify the proposed method, we have performed experiments for a reactor mock-up model and a real nuclear power plant. The experimental results demonstrate that we can enhance the reliability and precision of the impact source localization. Therefore, if the proposed method is applied to a nuclear power plant, we can obtain the effect of additional installed sensors.

Crown Copyright © 2010. Published by Elsevier Ltd. All rights reserved.