Conformance checking for programmable logic controller programs and specifications

Abstract

Verification of industrial control systems' software is an important task, as the cost of failure in these systems is typically high. Formal verification methods can complement the currently used testing techniques, especially if requirements are formally specified. Behavioural specifications can be used to perform conformance checking against the implementation. However, the typical conformance relations are often more sensitive to differences than the controlled processes in case of many control systems, resulting in counterexamples during verification that are considered as false positives in practice. To overcome this issue, we introduce conformance relations adapted to control systems based on programmable logic controllers (PLCs) with different levels of permissibility. The relations can be selected by the control engineers, depending on the required compliance levels. Defining the new relations and a model checking-based method to check them makes conformance checking a powerful tool for the verification of industrial control systems.