Multi-objective Dynamic Robust Operation Optimization Method Based On Time Series For Continuous Annealing Process

The purpose of process operation optimization is to provide the optimal parameter setting values for the production process,which has a significant impact on improving product quality,reducing energy consumption and increasing production efficiency.There are many uncertainties in the production process of continuous annealing.In actual production,these uncertainties will have a significant impact on the safety and economy of production.However,the traditional robust operation optimization only sets a fixed degree of robustness according to the fluctuation of the production process parameters.Although the influence of the uncertainties is much reduced,the economy of production is sacrificed.For this problem,this dissertation starts with the actual production process investigation and proposes a robust multi-objective optimization approach based on time series prediction.Firstly,a multi-objective dynamic robust operation optimization model for the continuous production process is established,and subsequently the real-time prediction of the fluctuation of production process parameters is carried out by time series forecasting model based on LSSVM,so as to realize the dynamic adjustment of the robustness.Finally,the improved NSGA-? algorithm based on the selection of polar coordinates is used to solve the operation optimization model,so as to further improve the economic benefits of the production process under the premise of ensuring the production safety.Specific research includes:(1)Aiming at the actual situation of production process,the unit production capacity,the unit energy consumption and the strip product quality are selected as optimization objectives.The model variables and constraints are determined by analyzing the parameters related to the target,and the dynamic uncertainty of production process is further analyzed.Based on the analysis of the traditional static robust operation,a multi-objective dynamic robust operation optimization model for the continuous production process is established.(2)With respect to the dynamic uncertainties in the production process,a time series prediction model of process parameters fluctuation is set up based on Least Squares Support Vector Machines(LSSVM),and the optimization of model parameters based on improved differential evolution algorithm method is proposed.The time series prediction model is used to predict the fluctuation range of the parameters of the next optimization period,which can be used to realize the dynamic adjustment of operation optimization robustness.(3)Aiming at the multi-objective dynamic optimization model of continuous operation,an improved NSGA-II algorithm based on polar coordinate selection is designed to solve the model.This algorithm mainly improves the traditional NSGA-II algorithm from the aspects of diversity and convergence of population.That is,choosing the parent solutions to carry out the crossover and mutation by the mechanism of polar selection to accelerate the convergence of the algorithm at the premise of diversity maintenance.(4)The algorithm proposed in this dissertation is applied to the optimization of multi-objective dynamic robust operation optimization model,and compared with the traditional static robust operation optimization,and the computational results illustrate the efficiency of the proposed multi-objective dynamic robust operation optimization based on time series prediction.