Petri Net-Based Supervisory Control Of Automated Manufacturing Systems With Failure-prone Resources

One of the driving factors of a nation’s economic growth is its productivity level.Manufacturing automation is viewed as the critical and core vehicle for enhance production technologies such that fewer resources,such as raw materials,equipment and labor,can be utilized to produce more high-quality outputs.Technically,contemporary automated manufacturing systems(AMSs)are in essence resource allocation systems in which their sub-processes or components interact with each other to perform various tasks concurrently while competing for limited reusable manufacturing resources.Optimal distribution and sharing of these limited resources are desired in AMSs.In this sense,it is necessary to optimally configure and share the limited resources in an automated manufacturing system.The optimality of allocating resources could be hampered if there is no effective and robust control strategies in place to deal with situations that may arise in a system such as deadlocks and blockages.Recent years have seen extensive studies on deadlock problems,along with the substantial achievements made in this regard.However,a vast majority of them are based on the premise that the resources or system’s components are all reliable.Contrary to this notion of absolute systems reliability,in reality,failures do occur due to aging,breaking down of tools,wear and tear,faults in some resource’s components or due to failure of sensors.A circular wait can be created due to failure of resources or other components,which could stall the production of other part types that do not require the services of the failed resources or components.Based on the Petri net theory,this research presents deadlock control policies for certain classes of automated manufacturing systems with failure-prone resources,which can handle both deadlock and blockage propagation in the presence or absence of resource failures.The main contributions made in this work are summarized as follows:1.A robust deadlock control policy for automated manufacturing systems with multiple unreliable resources that are not configured by central buffers is proposed.The policy permits legal states as many as possible and ensures that parts not requiring unreliable resources can be automatically processed without human intervention if one or multiple unreliable resources break down.The policy development is based on a modified neighborhood policy,namely a single route neighborhood,which handles the allocation of failure-prone resources in a system.To guarantee deadlock-free operations in the remaining parts of the system,monitors are designed for emptiable strictly minimal siphons.2.A Petri net-based adaptive deadlock control policy that does not require extra buffers is reported.If an unreliable resource fails,three classes of buffer spaces can be temporarily used to store parts that require the failed resources in their impending processing routes.The proposed deadlock adaptive control comprises of control places and switch controllers.If an unreliable resource fails,the switch controllers are activated to move the part types that require the failed resource in their subsequent processing stages into borrowed buffer spaces.After a failed resource is recovered,the system returns to its normal operating mode.The part types in the borrowed buffer spaces will be returned to their last processing stages before the occurrence of the failure and continued along their processing route.3.In discrete event systems such as automated manufacturing systems,sensors are normally used to observe occurrences of events in order to control the systems and prevent them from entering undesired states such as deadlocks.Events that are observable under normal conditions may become unobservable if sensors or any element that detects their occurrences fails.In Petri net models,events are modelled by transitions and their occurrences by transition firings.Therefore,a transition may have two operational modes:the observable mode under normal conditions and the unobservable mode under failure conditions.To this end,an adaptive deadlock control that prevents deadlocks in the systems with a bimodal transition is proposed.The adaptive supervisory control is composed of control places designed to make all the siphons in the net systems max-controlled,and bimodal transition controllers designed to remark siphons that may become under-marked if bimodal transitions are operating in an unobservable mode.The supervisory control policies proposed in this work,unlike most of the policies in the literature,are robust to failure of components or resources in a system.The policies can guarantee continued production of part types that do not require a failed resource.Furthermore,the policies ensure that deadlocks/blockages do not occur if a failed resource is repaired and put back to operation.