UML-based safety analysis of distributed automation systems

HAZOP (hazard and operability) studies are carried out to analyse complex automated systems, especially large and distributed automated systems. The aim is to systematically assess the automated system regarding possibly negative effects of deviations from standard operation on safety and performance. Today, HAZOP studies require significant manual effort and tedious work of several costly experts. The authors of this paper propose a knowledge-based approach to support the HAZOP analysis and to reduce the required manual effort. The main ideas are (1) to incorporate knowledge about typical problems in automation systems, in combination with their causes and their effects, in a rule base, and (2) to apply this rule base by means of a rule engine on the description of the automated system under consideration. This yields a list of possible dangers regarding safety risks and performance reductions. These results can be used by the automation experts to improve the system's design. Within this paper, the general approach is presented, and an example application is dealt with where the system design is given in the form of a UML class diagram, and the HAZOP study is focused on hazards caused by faulty communication within the distributed system.