Évènements

16 Dic
16/12/2020 14:00

Sciences & Société

Soutenance de thèse : Changyi XU

Operational dependability model generation

Doctorant : Changyi XU

Laboratoire INSA : AMPERE (UMR5005)

Ecole doctorale : ED160 : Électronique, Électrotechnique et Automatique

Assessing the complex industrial system to be on dependable service is what the engineers and researchers have long been aiming for. Recent advanced researches in the Model-based safety assessment, especially the Structre Analysis and Component Modeling, provide the practicable methodologies to assess the dependability, yet a lack  of the framework which is able to assess both the structure and the various behaviors of the components in one uniformed model retains them to achieve the excellent assessment. Moreover, as the system’s operations are not considerable in the models, the service in the aspect of operational dependability is not able to be assessed both in quality and in quantity. Although several existing assessment tools have already show their potential to model the various behaviors in the form of n-state models or consider the operations as repair priority to be event sequence in the model, fusing ‘structure’, ‘various behaviors’ and ‘operations’ is still a challenge, highlighting a need for one viable framework that bridge the gap among them both by quality or quantity. In this research, a formal model generation approach is studied to bridge this gap, which is able to assess the system operatinal dependability by considering the system structure, various behaviors, and operations. Here, the composition of the component models is introduced in order to generate a global model of the system, the total breakdown states are identified according to the resulted failure expression for the purpose to fully consider the system’s structure, and the operational dependability is further realized by quality by applying the trajectory specifications, while by quantity by developing a cost evaluating technology termed Capacity Calculation Fault Tree. In the end, a demonstration of a miniplant system illustrates the wide potential of this research for guaranteeing the dependable service of complex industrial systems.