Research On Energy-Efficient Scheduling Software System Of Electro-Fused Magnesium Furnaces

Electro-fused magnesium furnace is a metallurgical equipment for producing advanced refractory materials,which has the characteristics of high energy consumption and high material consumption.Enterprises usually adopt the production method of multiple electro-fused magnesium furnaces in parallel and batch smelting.During the smelting process,the working conditions change dynamically,and the electric energy demand of a single furnace fluctuates sharply.The total electric energy demand of the furnaces is limited by the contract value of the power supply department.At present,enterprises are still using a single electric control policy,does not consider the difference between the maximum demand constraint furnaces and furnaces' operation,with direct operational experience of employees of electro-fused magnesium furnaces opening power.It is not only easy to damage smelting equipment.Moreover,the power efficiency is reduced.To address the problem,this thesis designed and developed an energy-efficient scheduling software for electro-fused magnesium furnaces.The software not only has basic energy management functions,but also realizes energy-efficient scheduling functions.That is under the premise of meeting the maximum demand constraint,the electric load of the electro-fused magnesium furnaces is scheduled from the two perspectives of time and space to improve the overall power efficiency of the furnaces.The software not only reduces the number of power outages during the smelting process,but also reduces the electrical load peak of the furnaces.The research work mainly includes the following aspects:Firstly,the thesis analyzes the demand for energy-efficient scheduling software for electrofused magnesium furnaces.The thesis analyzes the energy consumption and energy saving potential of electro-fused magnesium furnaces.In view of the characteristics of changes in working conditions and fluctuations in electrical energy demand,basic energy management functions including production planning,operation monitoring,abnormal alarms,statistical analysis,etc.,as well as energy-efficient scheduling functions are proposed.The energy-efficient scheduling function is to schedule the limited electric energy resources between the group furnaces and at different slip times,so as to maximize the overall power efficiency of the group furnaces under the maximum demand constraints.Secondly,the thesis studies the energy-efficient scheduling approach of electro-fused magnesium furnaces.The thesis presents the overall strategy framework for energy-efficient scheduling of electro-fused magnesium furnaces,that is rolling prediction of the dynamic changes of electro-fused magnesium furnaces' operating conditions and electric energy demand,and judging whether to schedule group furnaces electric energy demand according to the demand forecast results.The rolling prediction and scheduling strategy improves the reliability of prediction and scheduling.In order to improve the real-time performance of electric energy scheduling,an energy-efficient dispatching decision-making approach based on rule-based reasoning and dynamic adjustment strategy is designed.This approach combines domain knowledge and expert experience to give a better power scheduling decision result,and through the online dynamic adjustment of the current setting value approach,to ensure that the scheduling result is implemented.The experimental results show that the approach is more reliable.Thirdly,the thesis designs an approach of predicting working conditions based on the reasoning of confidence rules and an approach of predicting power demand based on LSTM,and implements them in the energy efficiency scheduling software for magnesia smelting furnaces.The thesis analyzes the data of smelting conditions,extracts relevant characteristic attributes,and establishes a confidence rule base for predicting conditions.Tests with actual production data show that the accuracy of the predicted working conditions is over 85%.The electric energy demand is affected by the working conditions and the fluctuation of the electrode,and the fluctuation is strong,so the LSTM model is used to predict the electric energy demand.By adjusting the number of neurons and other parameters to compare and select the optimal parameters,the MAE test error result is1638.47 k W,the RMSE test error result is 2225.64 k W,and the MAPE test error result is 3.032%.Fourthly,the thesis designs and implements the energy-efficient scheduling software for electro-fused magnesium furnaces.The software adopts modular design and B/S architecture.The functional modules are divided into user management,energy management and energy-efficient scheduling.Among them,the energy-efficient scheduling module realizes several parts of working condition prediction,electric energy demand prediction,group furnaces electric energy dispatching and dynamic adjustment of single furnace current setting value.The software adopts C#programming language and.NET development framework to realize the back-end of energyefficient scheduling software.The software adopts SVG,HTML,CSS,JQuery,Highcharts and other technologies to complete front-end development.After testing and verification,the software has shown good interaction effects.