Identifying Hardware Failures Systematically

Critical control systems can only be used after approval of certification authorities due to safety reasons, among other aspects. Undetected failures in such systems can be catastrophic, including the loss of human lives or huge amounts of money. The safety assessment process aims to minimize such problems. But actually it still is largely dependent on human support (engineer’s experience). To decrease this human dependency, we propose a systematic hardware-based failure identification strategy. Following common practices in industry, which use

Simulink

diagrams to design (critical) control systems, the starting point of our proposed strategy is

Simulink

diagrams. The systematic identification is performed by the model checker FDR [11]. Therefore, we translate

Simulink

diagrams into CSP

M

specifications [30]. With our strategy, engineers only need to label certain

Simulink

elements as hardware and choose specific failure names for the generic ones our strategy provides. We illustrate our work on a simple but real case study supplied by our industrial partner EMBRAER.