Study On The Key Algorithms Of Fault Tree And Event Tree Modeling And Analysis

Researches on the reliability analysis and safety assessment of system are significant in the progress of nuclear industries and national defense of our country. Fault tree and event tree modeling and analysis technologies are recognized as the important means of system reliability analysis and safety assessment.The improvement of key technologies and algorithm related to the process of fault tree/event tree modeling and analysis, such as analysis boundary conditions, not-logic managing and common-cause failures, can not only decrease the mistakes and time consuming in the modeling process but also increase the calculation speed and precision in the analysis process. This dissertation dedicated in the researching and resolving these questions:(1) A linked fault tree construction algorithm for event tree analysis based on improved De Morgan algorithm and boundary conditions of analysis was developed. In the light of the different accuracy treatments of not-logic in the system model, the timesaving of analysis is 8%～33% compared to traditional method. Its correctness and efficiency was approved by multiple international benchmarks tests.(2) A fault tree transfer pages reusing algorithm based on multiple switching events sets was proposed. The values conflictions problem of the system model basic elements in event tree analysis with multi boundary conditions sets was solved through the multiple nesting disposing of effective switching events set for fault tree transfer page. The calculation speeds were 7%～43% faster than international similar program Risk Spectrum, which was validated by multiple international benchmarks and Qinshan nuclear power plant’s real models.(3) A new method called Common-Cause Breakdown Structure (CCBS) based on Multi-Common-Cause Failures (MCCF)was proposed to manage the asymmetrical Common-Cause Failures (CCF), which improve the precision of the asymmetrical CCF analysis. Independent failure splitting hypothesis and symmetrical impacts hypothesis of CCF were elaborately abstracted from the traditional symmetrical assumptions while not only much weakened but also very reasonable and true in most cases, which make the CCF events treated by CCBS can be handled by general CCF models without limitation from the components types or the CCF models. The correctness and validity of CCBS was approved by comparing to conservative and optimistic methods.On the basis of these researches, the related key algorithms were implemented and tested in the fault tree and event tree modeling and analysis modules of the reliability and probabilistic safety assessment program RiskA developed by the Institute of Nuclear Energy Safety Technology, Chinese Academic of Science/FDS Team. The results of test and verification on the international benchmarks and nuclear power plants, such as Qinshan, proved the effectiveness and correctness of these improvements. The development of these key technologies and algorithms contribute to the higher efficiency and accuracy in the modeling and analysis of large complex system and promote the development of fault tree and event tree modeling and analysis technologies and their further application in the nuclear energy utilization and national defense of our country.