Research On PDET-DCFT Method For Risk-Informed Safety Analysis Of NPP

The current safety analysis methods widely conducted in nuclear power plants(NPP)include two main types,Deterministic Safety Analysis(DSA)and Probabilistic Safety Analysis(PSA).Due to the interactions among random failures of equipment,operating states of systems,and operator actions occurred in NPP,either of the two methods for evaluating nuclear safety is insufficient or unnecessarily conservative.Thus,the nuclear industry has been investigating how to scientifically find out the contributors and laws of nuclear accidents,accurately recognize the safety boundaries,guide the design and safety review of nuclear energy systems,and finally develop safe and economic nuclear energy applications.In terms of basic principles of reliability,the operating state and failure probability of equipment vary with the dynamic operating conditions of NPP.New actions would be triggered by the changes and perturbations of operating parameters such as system pressure and temperature.In turn,the responses of those parameters during NPP operation would be caused by the state transitions of equipment such as switching between operating and standby units,random failures.In recent years,the newly developed Risk-informed Safety Margin Characterization(RISMC)technology has explored the coupling of DSA and PSA,in order to study the contributors to core damage,the accident evolution with their possibilities and consequences,and realistically evaluate safety margin and accident risk.As a newly explored technology of safety analysis,the main idea of RISMC is the coupling analysis of dynamic probabilistic risk analysis(DPRA)and system dynamic simulation.Dynamic Event Tree(DET)is one of the effective methods.However,for nuclear power plant accident analysis with dynamic characteristics,the RISMC analysis with dynamic event tree still needs to further solve the following problems:(1)The interactions between system process parameters and the reliability of equipment has not been fully considered;(2)There are many accident sequences that simulate all possible state combinations of equipment,and it is difficult to treat the dependencies of system;(3)Analyzing the system and equipment reliability under dynamic operating conditions;(4)Reasonably balancing the modeling precision,accuracy,and calculation amount in the coupling analysis of DET and system dynamic simulation.The more refined the simulation model is,the more realistic the results are,but the calculation amount would become unbearable.The lack of precision in DET model affects the accuracy of riskinformed safety margin.In this thesis,a new RISMC framework is proposed that combines Progression-dependent Dynamic Event Tree(PDET)and Dynamic Condition Fault tree(DCFT)to quantify the riskinformed safety margin indicators,so as to closely couple DSA with PSA.Based on this framework,this paper focuses on the following research to simulate the evolution of accidents:(1)The Progression-dependent Dynamic Event Tree method is proposed to make it suitable for complex system RISMC analysis.In the current DET methodology,the random state,state duration,and environmental conditions of equipment are usually simplified when quantifying the probabilities of accident sequences,so the real dynamic conditions can’t be fully reflected.For this reason,a conditional probability model for PDET branch is proposed related to equipment state,state duration,and process variables,so that it can reflect the influence of equipment operation history and environmental conditions in the pre-order branches.At the same time,the PDET risk analysis method is proposed to estimate the probabilities of accident sequence.(2)Aiming at the problem of equipment reliability analysis under dynamic operating conditions,the construction of Dynamic Conditional Fault Tree(DCFT)is proposed in this thesis,which replaces part of the PDET model.The DCFT hierarchical modeling and updating method is studied.To divide the dynamic progression of accident,the concept of Minimum Mission Phase(MMP)is introduced.Taking the system configuration and process parameters of an MMP as the DCFT conditions,the DCFT module reconstruction,the logical structure updating,time-dependent probability updating,and equipment reliability parameter updating related to process parameters are studied.In addition,in the case of component random failure already occurred,a hybrid treatment method is addressed for common cause failure,in which both independent failure factors and common cause factors are considered comprehensively,resulting in formulas for calculating CCF event probabilities.(3)In order to balance the precision of modeling,the accuracy of risk quantification,and the amount of calculation,the RISMC evaluation method integrating PDET with DCFT is proposed.PDET modeling is performed for the system-level failures,while the subsystem/equipment level/failure mode level failures are represented by DCFT.Then,by studying the coupling relationship between dynamic probabilistic risk assessment and system simulation,the mapping rules of equipment state transition and functional characteristic parameters are established.The accident risk quantification method combined with PDETDCFT is clarified.Finally,in order to demonstrate and verify the feasibility of the above method,the margin analysis of Large Break Loss of Coolant Accident on the main coolant pipeline of NPP is taken as an example.In the future,the PDET-DCFT method could provide new theoretical support for design and operational decisions,such as evaluating the risk-informed safety margin under various accident scenarios,quantifying the impact of proposed design changes on safety margin,verifying the success criteria of safety systems,and suggesting revisions of nuclear safety regulations.The above research in this thesis shows that PDET-DCFT method for the risk-informed safety analysis could make up for the shortcomings of current safety analysis methods.It has the following characteristics:Firstly,the PDET-DCFT model fully considers the complex coupling relationship between equipment state,state duration,and time responses of system process parameters,which is consistent with the dynamic characteristics of NPP accident progression.This model could automatically generate new branches and update the branch probability,so that realistically evaluate the risk level of accident.Secondly,the PDET-DCFT method can support nuclear power plants to carry out RISMC analysis,simulating the impact of timing order of events on NPP safety and risk level.It realistically evaluates the risk-informed safety margin under various operating conditions of nuclear power plants.Besides,it also helps to avoid excessive sequences arising from the problem of branching explosion as much as possible.Third,the integration of PDET with DCFT can not only represent the possible combinations of component state transitions in a complex system,but also have the ability to deal with simultaneous failures of multiple components and common cause failures,which makes up for the deficiency of RISMC analysis using dynamic event tree alone.It is beneficial to ensure accurately evaluating risk-informed safety margin.Last,the PDET-DCFT method could effectively identify the evolutionary paths of accidents,eliminate unnecessary conservatism and personnel judgments in the existing safety analysis.