Nuclear safety is one of the key issues for a nuclear power plant (NPP). Digital instrumentation and control (I&C) systems have been employed gradually in the newly-built and upgraded NPPs, while the reliability of software brings great challenges to the Probability Risk Assessment (PRA) of NPPs. Software testing is regarded as one of the most important methods to guarantee the quality of safety software. The testing data can then be adopted to assess the coding quality by reliability modelling. As the variety of digital I&C systems, software modelling methods corresponding to particular I&C system, as well as a model which is suitable in all situations, are both expected.
The Reactor Protection System (RPS) in High Temperature Gas-Cooled Reactor - Pebble bed Module (HTR-PM) is the first digital RPS designed and to be operated commercially in China. As the designer, we also took part in the software testing work of this digital I&C system. In this paper, we gave a comprehensive introduction to the software testing and reliability modelling research of RPS in HTR-PM, including the objective, tools, methods, testing strategy, organizational structure, and the implementation phases.
During the testing experience of safety software of RPS in HTR-PM, we collected the software abnormal reports which could be employed for the reliability analysis to evaluate the quality of the safety software. We introduced the data mining and reliability modelling research according to the abnormal reports. Different characteristics of faults could be used for software reliability modelling, such as software version, fault severity, test stage, submission date, debugging data, and so on. In the end we introduced a software modelling method based on severity analysis of the abnormal reports.
The work we showed in this paper can contribute to improve the process of testing and reliability analysis for other digital I&C systems in NPPs.