Model Based Verification of Safety-Critical Systems

Abstract

Dependability analysis like reliability, safety, performability etc. of safety-critical systems (SCS) have been modeled using various modeling techniques such as unified modeling language (UML), fault tree, failure mode effect analysis, and reliability block diagrams (RBDs). These techniques are capable to model all the system requirements, and the developed replica is implicitly accepted by all the stakeholders. These techniques demonstrate the static properties of a system and fail while capturing the dynamic behavior. Dynamic reliability block diagrams (DRBDs), which are extension to RBDs provide a framework for modeling the dynamic behavior of SCS. However, the analysis of a DRBD model in order to locate and identify the critical aspects of reliability and safety such as nonliveness, deadlock, design errors, or faulty state, is not trivial when done manually. This paper presents a novel approach for model based verification for digital feedwater control system (DFWCS) of a nuclear power plant (NPP) by developing its formal model using DRBD and then analyzing it using colored Petri nets for full proof design. The techniques to improve the faulty design are also proposed. Finally this model is proved to be bounded and deadlock-free.