Development Of Simulation Experiment System For Industrial Process Operation Optimization Based On Surrogate-Assisted Evolutionary Algorithm

Today,the biggest goal of industrial production is to obtain optimal economic benefits under the same investments.Therefore,achieving the optimal operation of industrial control and maximizing the production efficiency of the factory are the biggest requirements.This paper implements the surrogate-assisted evolutionary algorithms to solve the industrial operation optimization problems and establish a full process simulation system.It not only realizes the research on solving the industrial operation optimization problems,but also completes the verification of the optimization effect based on the surrogate-assisted evolutionary algorithms.Supported by the National Natural Science Foundation project "Data-driven industrial complex system operation optimization control and its application(61525302)",The main work of this paper is as follows:(1)Demand analysis on the development system,including functions,structures,usability and scalability.The system simulation object of industrial operation optimization problem,Tennessee Eastman(TE)chemical process model,including process flow,control objectives and mathematical description of operation optimization problems,are introduced.(2)The optimization method used in sysetem,the ensemble model assisted evolutionary algorithms(En-RVEA)are described in detail.The surrogate-assisted evolutionary algorithms solve the problem that the traditional multi-objective evolutionary algorithms are difficult to obtain the fitness value evaluation function when solving the industrial operation optimization problems.This paper innovatively applies En-RVEA to the optimization of TE process operation,and studies the optimization of industrial process operation.(3)System design.Function design:the simulation research on the optimization of TE process operation using En-RVEA algorithm is realized;Structure design:realizing the functions of the system by three parts,including En-RVEA algorithm,control algorithm and TE process model.The structure of each part and the interactions between each other are designed too.(4)System Development.Using MATLAB software to develop the En-RVEA algorithm,realizing the parameter adaptation of the TE process operation optimization problems.Developing the TE process simulation model,mainly including research and use of TE code.(5)Simulation experiment and result analysis.Running The TE process in the basic mode,the process data is acquired,the TE process operation optimization problem is mathematically analyzed,and the Radial Basis Function(RBF)neural network method is used to establish the data models between the objective functions and the decision variables,and the optimization method parameters are modified for adapting TE problem.According to the characteristics of the optimal solutions,the part of them are converted into the set values for simulation experiments,and the optimization effects are verified.The simulation study of the industrial operation optimization problem in the system is fully presented,and the optimization effects of the En-RVEA algorithm are verified.