Prediction Of Equipment Remaining Life And Optimization Of Maintenance Decision Under The Framework Of Industrial Internet Of Things

Accurate remaining life prediction and reasonable maintenance decisions are effective ways for manufacturing companies to reduce maintenance costs and improve production process safety.The traditional industrial manufacturing environment cannot effectively monitor the operating status of the equipment,and it is difficult to quantify the future degradation of the equipment,which makes it impossible to make reasonable maintenance decisions.Based on the Industrial Internet of Things architecture,a data-driven method is used in this thesis to predict the remaining life of equipment,and optimizes maintenance decision-making for multi-device systems based on the predicted information,so as to effectively integrate the "prediction" and "maintenance"topics in the field of predictive maintenance,and provide the decision-making basis for the equipment maintenance of manufacturing enterprises.According to the characteristics of the maintenance business under the architecture of the Industrial Internet of Things,combined with the basic maintenance process of the manufacturing enterprise,the maintenance decision-making process and the maintenance architecture based on the Industrial Internet of Things framework are analyzed,and the key technologies of the intelligent maintenance architecture are analyzed.In order to solve the problem that it is difficult to quantify the failure probability per unit time in the current remaining life model,the WTTE-RNN model is used to predict the Weibull distribution parameters of the remaining life of the equipment,so as to quantify the degradation of the equipment in unit time;and then use the combination of CNN and RNN algorithm to improve the method The parameter estimation of the life distribution is improved compared to the pure WTTE-RNN framework and the model convergence accuracy is improved.Aiming at the problem of traditional parallel multi-equipment maintenance task dispatching,an imperfect maintenance model is introduced,and a resource-constrained imperfect maintenance decision dispatching model is established from the perspective of decision risk,which can determine the equipment that need to be maintained、equipment maintenance levels、specific maintenance time and task assignment under the condition of limited maintenance personnel in each decision cycle.Then,a genetic algorithm is designed to solve the model,and verify the correctness of the model and the effectiveness of the algorithm with examples.Finally,numerical examples are used to verify the validity of the model.Aiming at the task assignment problem of series and parallel multi-device system with a fixed downtime,with the goal of minimizing maintenance costs,maximizing the probability of the system completing the next task and minimizing system downtime,a multi-target selective maintenance model of series-parallel system is established that combines task assignment with a decision model to make system decisions about the optimal duration of each rest period、the equipment that needs to perform maintenance during the rest period,and the level of maintenance performed;Then,use NSGA-Ⅲ algorithm to solve the multi-objective model.Furthermore,compared with the two-target model that only considers the minimum maintenance cost and maximizes the probability of the system completing the next task,the effectiveness of the threetarget model to shorten the system downtime is verified.Using the mainstream MVVM framework,with ASP.NET as the development platform and MySql as the back-end database,a prototype system for the health prediction and maintenance decision of the production system of the manufacturing enterprise based on B/S is developed,which provides effective support for equipment maintenance decision-making,task scheduling and spare parts management of manufacturing enterprises.