Research On Electricity Consumption Prediction Model And Electricity Consumption Optimal Decision Optimization Algorithm Of Cement Raw Material Grinding System

The cement raw material grinding system is the initial system in the cement process industry,it has an important influence on the subsequent production quality.Electricity consumption is a key index to measure the energy consumption of cement raw material pulverizing system,the accurate electricity consumption prediction can guide the decision variables’ optimization strategy.However,due to the complex industrial process,the production equipment and related variables involved in the cement raw grinding system are numerous.What’s more,there are dynamic characteristics such as coupling and time-ductility between the variables,the conventional modeling method can’t carry out accurate mathematical analysis of the cement raw grinding system.In this paper,a Multi-Layer Recurrent Highway Networks(ML-RHN)is proposed to model the electricity consumption prediction of cement raw grinding system.Based on the model,an improved Whale Optimization Algorithm based on Logistic-Tent mapping & Gaussian random walk(LT-GWOA)to design the decision variables’ optimization strategy.The optimal electricity consumption value satisfying the fineness constraint of raw material is used to solve the corresponding decision variable value,which can guide the resource scheduling in the production process.Specific research contents are as follows:In view of the problems of numerous variables and complex relationships in the cement raw grinding system,this paper analyzes the process mechanism of the cement raw grinding system,selects the system variables that have the main impact on electricity consumption in the real energy data management database as the model’s initial input,and further verifies the variables rationality through correlation analysis.Aiming at the influence of nonlinear characteristics such as time delay and coupling of cement raw grinding system data on the prediction accuracy,this paper uses the sliding window to integrate the timing information of input data and construct the input data matrix of the model.Based on the time memory characteristics of recurrent highway network,time correlation information among variables is extracted to eliminate the influence of delay on prediction accuracy.To solve the problem of strong coupling between system variables,a multi-layer network structure is designed,and a layered attention mechanism is added to ensure that the model can pay attention to the characteristics of different variables.The electricity consumption prediction accuracy of cement raw material grinding system is improved effectively by using the above scheme.Aiming at the problem that the decision variable optimization strategy of cement raw powder grinding system relies excessively on manual experience to reduce the accuracy of the strategy,this paper proposes a LT-GWOA algorithm to design the decision variable optimization strategy,which effectively improves the scientific setting method of decision variables.Finally,using the real database data of cement production enterprises,the above electricity consumption prediction model and the optimization algorithm of decision variables were carried out several control experiments.The results show that the error of the predicted value is small and the improved optimization algorithm can greatly improve the optimization range of power consumption,the above model and strategy can complete the accurate prediction of electricity consumption and scientific optimization of decision variables,which is helpful to improve the economic and ecological benefits of cement production.