Towards automated hazard analysis for CPS security with application to CSTR system

Abstract

Hazard analysis to identify risks due to cyber attacks is crucial to ensure cyber-physical systems' safety and correct their behavior. This study proposes a framework to analyze cyber-physical systems (CPS) risks using system dynamics to identify the critical physical components and hazardous interactions. We divide the input space into three regions, namely safe, alarm, and hazard zones, based on some given information and the dynamics of the CPS. These regions show how by manipulation of the critical inputs system can be driven to some predefined region. Moreover, using the concept of gap metric, we refine the list of the input value of vulnerable components to update the regions. Also, as a part of hazard analysis, we study and obtain the time needed to operate the system from a safe region to a hazard region by manipulating the inputs. The proposed framework uses the physical system's model dynamics to simulate safe, alarm, and hazard mode activity. This framework's primary goal is to provide decision-makers with a framework to evaluate CPS safety measures to implement adequate safety mechanisms. Model dynamics of continuous stirred tank reactor system (CSTR) has been employed to simulate the process under different manipulation of the input signal based on the proposed framework to identify the critical inputs and hazardous scenarios. 2022