Utilization of Constraint Satisfaction Problem Algorithms in Model-Based Safety Analysis [MBSA]
Samar Elmeadawy, Akram Amin Abdellatif, Florian Holzapfel · AIAA AVIATION 2020 FORUM · 2020
Model-Based Safety Analysis (MBSA) is an introduced approach in the safety analysis field which intends to connect the physical design of a system to the safety of its operation. Both design engineers and safe engineers use a common model consisting of the physical behavior in addition to the failure behavior of a system. Such method can reduce the time and cost of the safety analysis process. A pre-implemented tool represents a new developed qualitative safety analysis tool utilizing the MBSA approach. Although the tool has shown decent results with various systems but it has some defects. The tool mainly utilized directional traversal algorithms for failure injections in the system. The tool worked correctly with a system of small number of interconnected components capturing all required failure combinations to be tested. With the increase of system components, the algorithm failed to capture redundant interconnected components but stayed looping through the system components without termination. Thus, a new systematic algorithm must be used. In this paper, a Constraint Satisfaction Problem (CSP) technique “Backtracking” is tested on some of huge complex systems. The algorithm utilizes the CSP along with inference logic to rank the interconnected components according to failure impact on the whole system. Ranking the system components reduces the operation complexity by a huge amount and prevents the brutal force of testing all failure combinations. The developed algorithm is verified by testing upon various systems as a conclusion of this work.