A Novel Supervisory Control Algorithm to Avoid Deadlock in a Manufacturing System Based on Petri Net in Presence of Resource Failure

It is well established that resource failure, including robots and machines, in a manufacturing system can result in deadlocks. This issue not only hampers the system's performance but can also inflict significant damage on the manufacturing process. In this paper, we present a new algorithm developed through modeling of a manufacturing system using Petri net that ensures the liveness of the net in the event of such a failure. To detect possible failures, we first design a recovery subnet that is integrated into the resource. Next, we analyze the effects of failures on each state of the network to identify forbidden states. Finally, we propose an algorithm that optimally adds control places and establishes new constant vectors within the network, enabling effective management of remaining resources across different parts of the net. The proposed algorithm has been implemented in a system featuring three manufacturing lines, demonstrating its error-free operation while ensuring key properties such as boundedness, liveness, and performance continuity within the net.