论文部分内容阅读
采用核电厂安全审查大纲技术的分支NUREG-0800 BTP7-14分别建立基于贝叶斯(Bayes)网络的阶段评估模型以及综合评估模型。在阶段评估模型中,确立8个阶段,通过13个一级指标、74个二级指标、326个三级指标来完成对软件阶段性的实时评估。选用Hugin贝叶斯网络分析工具,针对测试对象展开预测推理及敏感性分析。经过测试后得到该软件在生命周期不同阶段对标准的符合程度,经综合评估模型推理,可得该软件在标准层面的可靠性指标是98.84%。经敏感性分析,可以定性地发现软件在生存周期中存在的薄弱环节,为评估核安全级数字化仪控系统的可靠性和安全性奠定基础。
The NUREG-0800 BTP7-14, which is a branch of the nuclear power plant safety review outline, establishes a phase assessment model based on the Bayesian network and a comprehensive evaluation model. In the stage evaluation model, eight phases are established, and real-time evaluation of the software stage is completed through 13 first-level indicators, 74 second-level indicators and 326 third-level indicators. Hugin Bayesian network analysis tool is selected to predict the reasoning and sensitivity analysis of the test object. After testing, the software is found to meet the standards at different stages of the life cycle. Based on the comprehensive evaluation model, the reliability index of the software at the standard level is 98.84%. Sensitivity analysis can qualitatively reveal weaknesses in software life cycle and lay a foundation for assessing the reliability and safety of nuclear safety grade digital instrument control system.