Study And Implementation Of Collaborative Modeling Key Technology In Probability Safety Analysis

Nuclear energy is an effective type of energy which can solve the energy problem of the world today. Safety is the primary subject of utilizing nuclear energy. Effective and efficient safety analysis methods can help to improve the safety of nuclear power plants. Probability Safety Analysis (PSA), as one of the main safety analysis approaches applied in nuclear industry, played a significant role in nuclear safety analysis.Models including Fault Trees and Event Trees are inputs of PSA. Correct PSA results can’t be achieved without appropriate and detailed PSA models.As one of the representative large and complex systems, the nuclear power plant PSA models can be very large and complex. Efficient and effective PSA modeling methods are needed to handle the building, managing and processing of nuclear power plant PSA models. On the premise of full investigation, the existing PSA programs and tools have several disadvantages:low model saving and loading efficiency, the model data managing mode isn’t direct enough for the users, the model file forms of one program can’t be used on another program, poor model data and logic compatibility, unable to support collaborative modeling of different programs, most of the exsiting PSA programs still use standalone architecture which is used by the first generation of PSA programs, more workload for staffs and much time cost when handling large and complex PSA models. Collaborative modeling technology, such as collaborative using of multiple PSA programs and multi-user distributed collaborative modeling, can solve these problems.Aiming at solving the problems introduced above, this paper plans to study a set of innovative PSA collaborative modeling methods. Based on the methods, several software modules were designed and developed. In order to satisfy the need of practical PSA works, the study of this paper has made a systematic research on key methods and technology of PSA collaborative modeling based on the joint point of collaborative and PSA modeling technologies. Several aspects are included: efficient PSA model strorage technology, cross-platform collaborative modeling, PSA model projects collaborative using, new PSA software architecture deisgn and multiple clients PSA collaborative modeling based on network. On the basis of PSA program platform developed by the Institute of Nuclear Energy Safety Technology, Chinese Academy of Sciences·FDS Team. The main contributions and works of this paper are:(1) a new multiple PSA model file cross-platform collaborative using method was proposed. On the basis of model unified model storage mechanism which can save the logic and data of PSA models separately. By applying the method, the program can identify and analyze the model files’codes automatically and rebuild the fault tree and event tree model logics quickly. The PSA models represented by the model files can be completely rebuilt. This method improves the versatility and compatibility of PSA model files and solves problems of cross-platform collaborative modeling. (2) A new collaborative modeling method for standalone PSA programs was proposed. In order to solve the problem of data collision in collaborative using of multiple PSA projects, conflicting data processing mechanism of PSA model projects and complex tree graph logical structure merging algorithm are developed. The data and logic validity and integrity of the merged projects are ensured. This method can solve the problems in model updating and sub models’collaborative modeling. (3) A new multiple users and computers collaborative modeling method was proposed. This method combined the advantage of C/S architecture with traditional standalone PSA softwares innovatively. Multiplex users and roles management mechanism and PSA modeling work decomposition and controlling algorithm were developed so that the multiple users and computers’paralled collaborative modeling can be applied in PSA modeling to solve the problems in traditional PSA modeling such as low paralled level, long redundancy time and low overall efficiency.Moreover, the methods and algorithms introduced above are implemented in software modules. Models are the most important basis of PSA, and effective and creative PSA collaborative modeling methods can improve the efficiency of PSA modeling, ensure the correctness of models, promote the application of PSA methods and programs in nuclear power plants. The PSA collaborative modeling methods introduced in this paper are implemented in complete PSA models of several Nuclear Power Plants. The models are accurate and reliable. The time spent on modeling and processing is shortened significantly compared to that by traditional PSA programs. This work laid a good foundation for PSA analysis and calculation.