New strategies for reliability analysis of Programmable Logic Controllers

Programmable Logic Controllers (PLC) are widely used in industry. Reliable PLC systems are vital to many critical applications. The reliability analysis of PLC is special because the hardware component and embedded software are combined and operated in a specific manner. This paper presents a probabilistic modeling for PLC systems, which considers both the hardware and software components. Three analysis strategies are proposed for probabilistic evaluation of reliability characterization. The first method is an input-based analysis which considers the impact of errors from primary inputs. The second one is an action-based analysis. It extends the first method by considering the impact of software-based processing deviations on primary inputs. The third method is an action-traverse analysis. It computes the reliability characterization in a single topological traverse through the primary inputs during the software execution. Experimental results demonstrate the effectiveness of our approaches when compared to fault tree analysis.