A Method for Quantifying the Dependability Attributes of Software-Based Safety Critical Instrumentation and Control Systems in Nuclear Power Plants

摘要

The lack of systematic science-based methods for quantifying the dependability attributes in software-based instrumentation and control systems in safety critical applications has shown itself to be a significant inhibitor to the expanded use of modern digital technology in the nuclear industry. Dependability attributes include reliability, safety, availability, maintainability, and security (confidentiality and integrity). Modeling the dependencies between the dependability attributes is the first step towards dependability quantification. In this research we use two methods: structured expert opinion elicitation and (hierarchical) causal mapping to extract the dependencies. A panel of fourteen international experts was identified. Each expert filled a unique questionnaire, targeted towards dependability and attributes as per his/her expertise. The questionnaires were designed in a semi-structured format. The questions were designed to elicit the attributes encompassed by dependability, the root causes of each attribute, the dependencies between attributes, and how root causes and attributes affect dependability. Then the data from the expert elicitation was analyzed and converted to fourteen hierarchical causal maps. A hierarchical causal map is divided into three levels of detail: the top layer of the causal map is called the dependence level composed of the dependability attributes and interrelationships; the middle layer is called the Event of interest (Eol) level and expresses mechanisms leading to occurrence of the main event of interest (for instance a safety critical failure) for each dependability attribute; the third layer is called Measureable Concepts level, and is composed of measures for each of the Eol contributors. Finally, a merged causal map on the dependencies between dependability attributes was developed.