An Efficient Method to Assess Resilience and Robustness Properties of a Class of Cyber-Physical Production Systems

Abstract

Widely available real-time data from the sensors of IoT infrastructure enables and increases the adoption and use of cyber-physical production systems (CPPS) to provide enterprise-wide status information to promptly respond to business opportunities through real-time monitoring, supervision and control of resources and activities in production systems. In CPPS, the failures of resources are uncertainties that are inevitable and unexpected. The failures of resources usually lead to chaos on the shop floor, delayed production activities and overdue orders. This calls for the development of an effective method to deal with failures in CPPS. An effective method to assess the impacts of failures on performance and create an alternative plan to mitigate the impacts is important. Robustness, which refers to the ability to tolerate perturbations, and resilience, which refers to the capability to recover from perturbations, are two concepts to evaluate the influence of resource failures on CPPS. In this study, we developed a method to evaluate the influence of resource failures on CPPS based on the concepts of robustness and resilience. We modeled CPPS by a class of discrete timed Petri nets. A model of CPPS consists of asymmetrically decomposed models of tasks. The dynamics of tasks can be represented by spatial-temporal networks (STN) with a similar but asymmetrical structure. A joint spatial-temporal networks (JSTN) model constructed based on the fusion of the asymmetrical STNs is used to develop an efficient algorithm to optimize performance. We characterized robustness and resilience as properties of CPPS with respect to the failures of resources. We analyzed the complexity of the proposed method and conducted experiments to illustrate the scalability and efficiency of the proposed method.