Development of Preventive Maintenance System Arising in Advanced Eddy Current Testing Using Network Tomography

This paper is concerned with a cyber-infrastructure arising in eddy current based nondestructive evaluation system for nuclear power plants. The system consists of an inspection module, a computational module and a communication interface module which can visualize useful information through interaction between the inspection module and the computational module. The design and performance of computational and communication interface modules are shown. Introduction: Preventive maintenance has become recent critical issues in Japanese power plants. Various kinds of stress corrosion cracks have been found out at shroud and re-circulating tubes of BWR plants since 2002 in Japan. Especially, interest has grown for detecting a crack on the surface of SUS304 components. From the fact that natural defects like stress corrosion crack (SCC) cannot be detectable by visual inspection, it is highly required to develop a feasible computational method for identifying and characterizing such defects. Eddy current testing (ECT) has advantages on the detection of surface and subsurface cracks [1][2]. Our major concern in this paper is to design a structural monitoring system based on the advanced ECT for nuclear power plants. To this end, we are developing a new remote diagnosis system using network communication. Figure 1 illustrates the overall configuration of a remote non-destructive system considered here. The system strategy is composed of three modules, inspection module, computational module, and communication interface module. In the inspection module, we develop the advanced ECT sensor equipped with information socket that makes it possible to communicate information among three modules. The computational module is to develop ECT simulator using a server system provided with high performance computational environments. The module is able to provide appropriate simulation data using the forward and inverse analysis for in-service inspection of real plants. The communication interface is worked as information board between human experts and the proposed system. In this paper, we describe the overall configurations of the proposed system, demonstrate how the system works, and explain current stage of our efforts. Fig. 1 Overall configuration of a remote diagnosis system Inspection Module: Figure 1 depicts a prototype of the multi coil ECT sensor used in the inspection module. The inner and outer diameters of each coil are 1 [mm] and 2.4 [mm], respectively. The height of each coil is 2[mm] with 520 turns. The sensor is composed of 35 coils. Fig. 2 Multi coil ECT sensor with 35 channels As illustrated in Fig.3, there are two typical scanning patterns according to the combination of exciting and detecting coil patterns. The so-called T-scan using 32 channels allows us for detecting horizontal crack, while U-scan using 16 channels provides us sensitive detection for vertical cracks.