Server Failure Detection for Robustness Analysis in Automated Manufacturing Systems With Arbitrary Resource Failures

Abstract

In the paradigm of Petri nets (PNs), this paper focuses on the maximally permissive robustness analysis for automated manufacturing systems (AMSs) with arbitrary resource failures through server failure detection. First, unreliable system of sequential systems with shared resources (U-S4Rs) are constructed by adding detection subnets to the original PNs for each unreliable place. Based on different failed statuses of unreliable resources, the reachable marking set of a U-S4R is partitioned into one safe and multiple failed subsets, where there exists a one-to-one relationship between all markings in the safe subset and those of the original system. Second, subnet systems corresponding to part types are categorized to provide the definitions of the robust, strongly robust, and weakly robust markings. Third, we provide algorithms to compute the sets of strongly connected, weakly connected, and dangerous markings in each failed zone. When resources are in a failed status ${\mathcal {X}}_{f}$ , we establish an ${\mathcal {X}}_{f}$ -failure-dependent marking set. The robustness of all markings within that set can be evaluated by examining the connectedness of the markings derived from those in that set. Moreover, for all markings not in the ${\mathcal {X}}_{f}$ -failure-dependent marking set but in the safe subset, we provide conditions further to determine the robustness of these markings. Finally, examples are presented to illustrate the proposed methods and show the advantage over the existing ones.Note to Practitioners—In practical manufacturing scenarios, it is crucial to analyze and control automated manufacturing systems (AMSs) to ensure uninterrupted production in the case of resource failures. Failure to handle these failures effectively can result in unexpected blockages, reduced throughput, and compromised product quality. Most existing works have been dedicated to robust control policies in order to guarantee the robustness. However, there are still some states that do not satisfy the robust control policies despite their robustness. This paper specifically investigates the maximally permissive robustness analysis of AMSs with arbitrary resource failures using Petri nets. Some significant results are proposed to determine the robustness of all markings for the original system when the unreliable resources are in different failed statuses. Therefore, based on the results of the maximally permissive robustness analysis, researchers can design optimal robust controllers to enhance the flexibility of manufacturing processes considerably.