| In the production of aluminum,the regenerative aluminum smelting process is an important part for energy efficiency and product quality,which is of great significance to the efficient and economic production of aluminum.Aluminum liquid temperature is a significant variable in the aluminum smelting process,and it is costly to detect online because it requires protective temperature sensor.At the same time,the parameters of aluminum smelting production are generally set by handbooks or manual experience,resulting in non-optimal energy consumption.To address the above problems,a modeling method with mechanistic integration of multi-scale kernel technology is proposed to predict the aluminum liquid temperature.Then,a optimization problem of aluminum liquid temperature and process energy consumption is established.To obtain the optimal parameters for stabilizing the aluminum liquid temperature and saving energy at the same time,a set of improved swarm optimization algorithms is developed.This paper investigates the modelling,prediction and optimization of a regenerative aluminum smelting furnace in an industrial aluminum smelting process.The main works of the paper are listed as follows:(1)Firstly,the structure and working principle of the regenerative aluminum smelting furnace are studied,the variables related to the aluminum liquid temperature of the aluminum smelting furnace are analyzed.Then the mechanistic model of the aluminum liquid temperature is constructed by heat transfer analysis and energy conservation.To obtain the unknown functional relationship in the mechanism model,a hybrid model is developed by combining some multi-scale kernel functions with the original mechanism model.Finally,the parameter identification problem of the hybrid model is described as an optimization problem,and the optimal parameter set of the hybrid model is calculated by a swarm optimization algorithm.(2)The parametric optimization problem of hybrid models is non-convex and strongly non-linear.Traditional swarm optimization algorithms used to solve this optimization problem have the drawbacks of slow iteration and tendency to fall into local optimality.Therefore,a hybrid strategy-based sparrow search algorithm(HSSA)is proposed for solving the optimization problem.HSSA introduces chaos factors in the initialization stage of the algorithm to enrich the population diversity.In the search stage,the adaptive weights and levy flight factors are introduced to improve the accuracy and stability of the algorithm.Then,a simulated annealing mechanism is used to avoid falling into local optimum.The experimental results demonstrate that the proposed optimization algorithm has higher convergence accuracy and stronger global search capability than the original algorithm,and the hybrid model optimized by HSSA has higher accuracy than the other optimized models.(3)On the basis of the hybrid model for aluminum liquid temperature,a optimization problem of aluminum liquid temperature and process energy consumption is constructed to stabilize the aluminum liquid temperature and save energy at the same time.The optimization problem uses the hybrid model as an equation constraint with many variables and is characterized by strong non-linearity and strong coupling of variables.Therefore,the traditional gradient-based approach and the proposed HSSA algorithm do not work well.In order to further improve the performance of the algorithm,a hybrid strategybased African vultures optimization algorithm(HAVOA)is proposed.The logistic mapping enables HAVOA to have a more comprehensive global search capability before the optimization phase.To balance the search and exploitation capabilities,a weighting factor that varies with the number of iterations is introduced.The effectiveness of the method is verified by simulation experiments,and the results show that HAVOA has excellent performance and solves the optimization problem of aluminum liquid temperature and process energy consumption better than other methods. |
PDF Full Text Request