Research On Key Technologies Of Manufacturing Execution System For Large-Scale Petroleum Equipment

Large-scale petroleum equipment manufacturing enterprises have a low degree of informatization in the workshop,unable to obtain the status data of the machine tools in the workshop in time,and the quality management of the products is still in the stage of using paper documents,which makes it difficult to carry out efficient quality traceability.At the same time,because of its multi-variety and small batch characteristics,it is faced with great difficulties in production scheduling.In view of the above phenomena,machine tool status monitoring algorithm,product quality traceability algorithm,and production scheduling algorithm based on deep reinforcement learning that integrate processing equipment status and product processing quality are proposed.Equipment status data and product quality data are used in production scheduling,with the goal of making the production plan more in line with the actual situation of the enterprise.Finally,a manufacturing execution system(Manufacturing Execution System,MES)containing the above-mentioned modules is developed and applied to large-scale petroleum equipment manufacturing enterprises.The main research contents of this paper are as follows:(1)The workshop machine tool information model is established and the machine tool condition monitoring algorithm is researched to realize the condition monitoring of the machine tool in this paper.First,the workshop machine tool information model is constructed to represent the state of the machine tool.Then the machine tool status monitoring algorithm is studied to complete the data collection of the machine tool status.The collected machine tool status data can be used to monitor the status of the workshop machine tool.At the same time,the machine tool status data can be used with the scheduling algorithm to enhance the rationality of the algorithm.(2)In this paper,the product quality traceability model,product quality data collection algorithm,product knowledge graph and the quality traceability algorithm based on the knowledge graph are studied to achieve product quality traceability.First,the production process of products in the enterprise is divided into many processing nodes and a product quality traceability model based on processing nodes is established.The barcode-based product quality data collection algorithm is researched to obtain product quality data.Product quality data and other processing node data are used to construct a product knowledge graph.A product quality traceability algorithm based on the knowledge graph is proposed to achieve product quality traceability.At the same time,product quality information is fed back to the scheduling algorithm so that the scheduling algorithm can adjust the processing plan in time according to the current product processing progress.(3)A multi-objective optimization model including processing time,processing cost,variance of equipment task volume,and sustainability indicators is constructed,and a multi-objective optimization algorithm for production scheduling based on deep reinforcement learning is studied to achieve production scheduling that meets the requirements in this paper.Through the mathematical description of the enterprise production scheduling problem,the constraint conditions are determined,and a multi-objective optimization model for processing time,processing cost,equipment task variance and sustainability indicators as the optimization goals is constructed.Disjunction graph and multi-agent deep deterministic policy gradient algorithm are combined in order to study the multi-objective optimization algorithm of production scheduling based on deep reinforcement learning that integrates the state of the machine tool and the quality of the product.The reliability of the algorithm has been verified based on a variety of data sets.(4)Manufacturing execution system including machine tool status monitoring,product quality traceability and production scheduling modules is developed.Manufacturing execution system containing the above three core modules is developed and applied to large-scale petroleum equipment manufacturing enterprises.