On Operation Control Of A Class Of Petri Net Models With Consumable Resource By Reachability Graph

Automatic Manufacturing Systems(AMS)are manufacturing process with little or no human intervention.Such a system is prone to deadlocks due to improper allocation of resources,which causes the system to crash.In order for AMS to operate smoothly and safely,the deadlock problems in manufacturing systems must be properly addressed.As a mathematical modeling analysis tool,Petri nets are widely used in the theoretical and applied research of advanced manufacturing because of their many advantages in AMS models' construction,state analysis and deadlock control.Based on Petri nets to study the deadlock problem,the existing research has developed many analysis methods and control strategies,but no matter based on reachable graphs or structure-based analysis methods,they are mostly oriented to Petri net models with conservative token counts.For a class of AMS with a resource consumption,the corresponding subclasses of Petri nets are not considered in the existing research.However,in actual production and manufacturing,the consumption of resources is extremely common,such as the consumption of machine tools,coolants,lubricants,etc.Therefore,it is necessary to define a corresponding Petri net subclass.This paper aims to meet the control requirements in resourceconsumption manufacturing systems.Firstly,a class of Petri net subclasses with variable weights and tokens that are not conserved are defined.Then,based on the control strategy of the reachability graph,the state space partition and operation control analysis are performed on the consumable network models.Finally,an iterative control strategy for a class of consumable networks is given based on vector cover method.The main work completed in this thesis is as follows:1.Aiming to an a class of automated manufacturing systems that have consumption characteristics for resources,the work models it through Petri nets,and defines a class of Petri net models with resource consumption characteristics,which is called as consumable net.2.Based on the control needs of the consumption net,the work first distinguishs the resource consumption situations in model deadlock states.The states in which the resources are not exhausted in a model and the states in which the resources are exhausted but not meet the consumption requirement for resource utilization are chosen as the control targets.Then,according to the control targets,the concept of the first encounter control marking and the vector cover method are proposed,aiming at simplifying the number of identifiers to be considered.Finally,an algorithm for consuming various sets of identifiers in the network is given,and the codes implemented by C++ language is used as an effective tool for the operation control of large scale consumption nets.3.Based on the core idea of P-invariant controllers,this thesis conducts operational control analysis on the consumption nets.Firstly,the state coverage problem in the consumption nets and the case where the vector cover method is not applicable are illustrated,and the corresponding solution strategies are given.Then,from the perspective of linear programming,an in-depth analysis of the problems existing in the consumption net based on the reachable graph-based control strategy is described.Finally,based on an improved linear programming strategy,an iterative controller design method based on vector covering for consumable nets is proposed,and the models with different specifications are illustrated.