RESEARCH ON THE RELIABILITY OF DIGITAL INSTRUMENTATION AND CONTROL SYSTEM OF NUCLEAR POWER PLANT BASED ON DYNAMIC FLOWGRAPH METHODOLOGY

摘要

With the development of information and computer technology, the Digital Instrumentation and Control (I&C) System has been widely used in nuclear power plants, which leads the tendency of NPPS' construction and rebuilding on digital I&C system. As an approximate approach, conventional fault tree approach has been used quite often in the analysis of nuclear power plants' Probability Safety Assessment (PSA), which combine with system components' failure modes in order to modeling the digital system's failure. However, for the reason that conventional fault tree approach has a great disadvantage on analyzing the reliability of digital I&C system, which may not be able to fully describe the dynamic behavior of digital I&C system with significant hardware/software/human action process interaction, multi-failure modes and logic loops, it cannot carry on effective modeling and evaluation of digital I&C system. Therefore it is necessary to establish some dynamic approaches to modeling digital I&C system. As a new probability safety analysis method, Dynamic Flowgraph Methodology (DFM) can model the relationship between time sequence and system variables because of its dynamic property. Therefore, DFM can be used to analyze the impact of software failure, hardware failure and external environment, which are closely related to the reliability of the whole system. In the first place, this paper introduces the theoretical basis, model elements and the modeling procedures of DFM and demonstrates how Dynamic Flowgraph Methodology (DFM) can be applied to Reactor Protection System with interactions between hardware/software and physical properties of a controlled process. Meanwhile, in this case, DFM and fault tree methodologies are both used to conduct the PSA for the same top event by calculating the probability of it and finding out the prime implicants of DFM and minimal cutsets of conventional fault tree. During the process of analysis, we mainly evaluate the reliability of reactor trip function of Reactor Protection System (RPS) by using DFM and conventional fault tree approach and mainly focus on modeling the four-way-redundant voting logic and the reactor trip breaker logic. Finally, through the comparison of this two methods and model results, it is concluded that there is a distinct advantage of DFM over conventional fault tree approach by using multi-logic to fully display the fault mode and utilizing decision table to describe the interaction between software and hardware. In general, conclusion can be drawn that, as a dynamic approach, Dynamic Flowgraph Methodology could be more accuracy and effective than conventional fault tree approach in analysis, ensuring the reliability and safety of the whole digital I&C system.