Research On Health Evaluation And Preventive Maintenance Of Multi-equipment Production Line

With the continuous development of modern manufacturing,manufacturing technology is gradually developing in the direction of automation,intelligence and integration.The equipment scale of the multi-equipment production system continues to expand,the degree of automation continues to improve,and the production efficiency of the production system has been greatly improved,which also puts forward higher requirements for the reliability of the production system.In a multi-equipment production system,there are correlations between equipment.If a certain equipment fails,the related equipment will also be shut down.Therefore,in order to reduce maintenance costs,preventive maintenance should be performed on equipment whose health condition has degraded to a certain extent during each maintenance process.At the same time,preventive maintenance of the production line needs to be based on the health status information of the equipment in the production line and the overall health of the production line.This thesis was supported by the 2020 major science and technology project of Sichuan Province: "Development and Application Demonstration of a Digital Intelligent Processing Complete Production Line for Large-Size Aluminum Alloy Parts"(Grant No.2020ZDZX0024).The large-size aluminum alloy parts production line acted as the research object.In view of the preventive maintenance of the multi-equipment hybrid production line,this thesis was carried out on the single-equipment health evaluation,the multi-equipment system health evaluation and the preventive maintenance strategy of the multi-equipment hybrid production line.The specific content is as follows:First of all,in order to obtain the health index of a single device,a method for dividing the health status of a single device based on a variational autoencoder and an unsupervised clustering algorithm was proposed.The equipment health index was obtained by measuring the deviation level between the current state of the equipment and the equipment health status based on the Mahalanobis distance.Then,after obtaining the health index of a single device,in order to calculate the system health index of the production line,a system health index construction method based on granular computing was proposed.In this thesis,a directed graph was introduced,and the node importance evaluation method was used to reflect the influence of the production line structure.The data redundancy was divided by the importance of the data,and the data with different redundancy were fused with different fusion processing to construct the system health index.Finally,the accuracy of the method was proved by setting up a comparative verification experiment.Finally,a preventive opportunity maintenance method for the production line based on the equipment health index was proposed.When the system health index or the equipment health index reaches the preventive maintenance threshold,the production line will be shut down for maintenance,and all equipment that meets the opportunistic maintenance conditions will be maintained.This thesis took the lowest total maintenance cost of the production line system in a limited time as the goal,and established the model with the evolution law of the equipment health index as the constraint,and solved it through genetic algorithm to obtain the best maintenance threshold combination.The experiment showed this method is better than periodic maintenance in terms of total maintenance cost,maintenance times,and equipment utilization.