Safety Evaluation of Hybrid Systems
Leila Boucerredj, Nasr Eddine Debbache · 2018
This paper proposes a new complete approach based on qualitative and quantitative analysis to the safety evaluation of hybrid systems, the goal of qualitative analysis is to extract all minimal feared scenarios that lead a system in a failure state, directly from the proof trees of linear logic to establish the causality between undesirable events and normal operations. The proof trees of linear logic contain events that are the result of event in the scenario, but not strictly necessary for the final production of the critical feared state. The size of the proof tree is proportional to the number of firing transitions in the sequent provable. For this raison we develop a qualitative analysis approach based on the concept of minimal cutsets of the fault tree applied to the proof trees of linear logic. The aim is reducing the size of the proof trees of linear logic and generates a minimum number of feared states. Then we complete our study by quantitative analysis to the safety evaluation of hybrid systems based on the results of qualitative analysis; the modeling formalism adopted of this class of systems is based on the hybrid automata model.