Deadlock detection and avoidance in FMS with shared resources and parallel machines using Petri net

Abstract

Deadlock is a critical case in flexible manufacturing system that results from concurrency and resource sharing for processing different types of parts. It leads to less efficiency and fewer parts to be produced. Previous attempts to solve this issue focused on putting constraints on process flow of a simple system to avoid deadlock. Yet, this way reduces the performance of system and hence the number of final products. Besides, it does not ensure its capability in complex systems. Thus, the objective of this dissertation is to model and simulate complex flexible manufacturing system of industrial company that has shared resource and parallel machines using timed colored petri net. It also aims to analyse the system in terms of utilization in each stage and detect the deadlock where high utilization is found. Another objective is to avoid the deadlocks that have effect on the daily production of the system. The manufacturing system was modelled and simulated using CPN tool. Then, the analysis of the simulation in CPN was performed. It showed that deadlock exists in wire straightening with 88.55%, injection trolley 3 with 86.392%, spinning machine 1_2 with 90.611% and spinning machine 1_3 with 82.311% utilization. Four boilers and mould revolution were also determined as deadlocks having 87.75% and 98.295% utilization respectively. Deadlock avoidance was then conducted through testing six plans in the manufacturing system to improve the daily production of the company. It was concluded that the best one is by adding another resource in wire straightening, injection trolley 3 and mould revolution stages. The simulation in CPN showed that this plan can achieve 22.047% increase in production. The utilization in the three deadlocks became in between 55% to 60%. This method for detecting and avoiding deadlock was validated in Delmia Quest and the simulation of improvement showed that the suggested plan will improve the current production by 22.134% increase.