An Implementation of DFM for Reliability Modeling and Analyzing of AP1000 FWCS

Abstract

Both in the process of designing instrumentation and control (I&C) systems of nuclear power plants and replacement of aging and obsolete ones, digital technology are more and more popular. However, traditional static methodology for probabilistic risk assessment (PRA) may not yield satisfactory results when applied to digital systems. As a result, some dynamic methodologies are employed in the analysis for I&C systems of nuclear power plants. The dynamic flowgraph methodology (DFM) is one of the favorable methodologies which brings out most positive features and least negative features. The I&C system of AP1000 is a distributed digital I&C system, but its PRAs are based on the event tree (ET)/fault tree (FT) approach. The results are admitted to be enlarged to some extent. Also, under some circumstances, ET/FT approach cannot account for the coupling between triggered or stochastic logical events. Regarding these disadvantages, DFM is introduced in modeling of Feed-water Control System (FWCS) of AP1000. A basic system of FWCS is constructed, which contains the main components and control logic of FWCS. Furthermore, failure mode and effect analysis (FMEA) of the FWCS system is carried out in the way of analyzing each component individually. After these efforts, a DFM model of the basic system is built utilizing the DFM language. After the DFM model has been produced, the application of deductive analysis shows different initial conditions or events that can cause the FWCS system to fail in different modes.