원자력 발전소 매설 배관 누수 탐지를 위한 센서 네트워크

초록

It is important to apply real-time remote sensing system for premature pipeline leakage detection of nuclear power plant because of economic damage and human injury due to leakage. There are two typical methods to detect pipeline leakage. One is the physical method using acoustic rod and the other is cross-correlation method by receiving vibration signals from sensors at both ends. In the former method, it needs a specialist with expertise in listening and is hard to process in real-time. The latter method is greatly vulnerable to noise signals. There are a lot of efforts to solve the limitation. However, most of the methods proposed so far are not suitable for practical systems because those require additional filtering algorithms. The filtering algorithms are difficult to process in real time and thus not suitable for nuclear power plants. In order to solve the above problems, we propose an intelligent sensor network system based on the leakage detection with improved cross-correlation function. For the performance evaluation, we implement the testbed and simulation. The field test and simulation results show that the ad-hoc structure is superior than the linear structure when the grid size is less than 5, whereas the linear structure is more advantageous when the grid size is greater than 5.

키워드