Research On Data-driven Optimization Of Energy-saving Scheduling Method For Workshop

This paper focuses on the energy consumption analysis and energy-saving model establishment of the manufacturing workshop,as well as the multi-objective intelligent algorithm for solving the problem.It solves the problems of data storage and sharing between heterogeneous databases in manufacturing workshop,high cost of adaptive value calculation of intelligent algorithm and so on.It has a certain theoretical reference and engineering application significance for the production energy saving and improving efficiency of the enterprise workshop.Firstly,this paper makes a lot of literature research on manufacturing data storage and sharing method,agent model method and multi-objective intelligent algorithm of workshop energy saving.Then,aiming at the problem of data storage and sharing between heterogeneous databases in manufacturing data,this paper analyzes the characteristics of database storage,sharing technology and extract transform load（ETL）technology,and puts forward the synchronous storage strategy of different database data in manufacturing data.This paper proposes to measure the efficiency of data storage and sharing between heterogeneous databases by using synchronous processing time and CPU utilization.Then,in order to explore the energy-saving potential of the manufacturing workshop,an improved multiple objectives multi verse optimizer（IMOMVO）algorithm is proposed based on the"shutdown-restart"strategy is used to solve the multi-objective hybrid pipeline workshop energy-saving model based on the mixed integer nonlinear programming model,which aims to minimize the maximum completion time and total energy consumption.By setting up different weighting coefficients,three workshop production scenarios are established.At last,in order to solve the problem of high cost of energy-saving target,i.e.fitness evaluation,in the process of intelligent algorithm optimization,a feature vector extraction method based on matrix coding mechanism is proposed,and the introduction of kernel function is conducive to the solution of energy-saving target by extreme learning machine（ELM）.This paper analyzes the computational complexity of the imomvo algorithm which needs to build the agent model,establishes the agent model based on ELM,and designs the data-driven optimization algorithm framework of workshop energy saving.The average synchronization processing time between SQL2SQL,SQL2My SQL and SQL2Mongo DB is 2.07ms,2.60ms and 0.74ms respectively.The average CPU utilization is 15.7%,57.9%and 3.2%respectively.The proposed IMOMVO algorithm is compared with the improved genetic simulated annealing algorithm（IGSA）.The results verify the effectiveness and superiority of the proposed model and algorithm.El M algorithm and error back propagation（BP）algorithm are used to verify the prediction performance and compare the calculation time.The experimental results show that the goodness of fit of ELM algorithm is 0.97381,and the prediction performance is better than that of BP algorithm.The average calculation time of a single fitness value is 5.4×10^-4seconds,only 18.5%of the real solution.It shows that ELM has a good effect on the prediction of energy saving target in workshop.