Automatic safety analysis of networked control system based on failure propagation model

Abstract

Recent developments in the modern control systems show a clear trend for increasing integration of a number of safety-related electronic systems - such as railway signalling system and x-by-wire systems - that are replacing traditional mechanical controls. These applications promise improved performance and safety through continuous high-capacity realtime data exchanges and reconfiguration in abnormal situations. However, safety analysis of networked systems is becoming extraordinarily difficult because of the complexity of interactions between components and the number of possible backup modes. The traditional analysis techniques are subjective and do not facilitate reuse, thus, making the safety analysis process time-consuming and error-prone. This paper extends the failure propagation model with a set of temporal operators, in order to demonstrate more accurate representation of practical failure behaviors of networked control systems. Computation laws of temporal operators and combinatorial operators are abstracted to transform complex temporal expressions into a standard form. In this way, the models can be solved automatically. The preliminary design of Communication Based Train Control (CBTC) system for Yizhuang Line in Beijing is used to demonstrate the approach.