Research On Economic Modeling And Optimization For Environmental Island Of Thermal Power Units

In recent years,with the improvement of people's living standards,people's requirements for environmental protection have become higher and higher.Pollutants such as dust,NOx and SO2 emitted by coal-fired power plants have seriously affected the natural environment.In order to meet the standards for ultra-clean emissions,thermal power generators often invest a large amount of money to meet environmental protection requirements,while ignoring the energy-saving potential of thermal power units in terms of environmental protection and therefore causing great waste.So far,the research on the cost of environmental island system of thermal power units has often stayed on the individual research of each subsystem,neglecting the coupling relationship between the various subsystems and failing to analyze the environmental island economy as a whole.This paper first improves the traditional XgBoost algorithm,and compares it with the regression tree algorithm,gradient elevating algorithm and traditional XgBoost algorithm,and verifies the superiority of the improved XgBoost algorithm in model building.Then this paper analyzes the mechanism of each subsystem of the environmental island system,determines the influencing factors of each subsystem,and determines the input variables of dust emission concentration,NOx emission concentration and SO2 emission concentration model through overall analysis.The improved XgBoost algorithm was used to establish the mathematical model of dust emission concentration,NOx emission concentration and SO2 emission concentration.At the same time,the important analysis of the influencing factors and incremental analysis were used to verify the accuracy of the model.Then,based on the differential evolution algorithm,constraint processing technology and self-adaptive penalty algorithm,this paper proposes a composite differential evolution algorithm based on adaptive penalty,then studies the economy of the environmental island system of thermal power units as a whole,and proposes the environmental island economy cost model.After that,based on the dust emission model,the NOx emission concentration model and the SO2 emission concentration model obtained in the previous paragraph,the constraint conditions for the optimization problem under the premise of the emission standards are derived,and the constraints can be satisfied through a composite differential evolution algorithm based on adaptive penalty in order to optimize ajustable parameters.In the end,this paper studies how to improve the performance of the improved XgBoost algorithm.Parallel computing can effectively improve the computational efficiency of the XgBoost algorithm and making full use of the internal and external memory resources of the computer can reduce the computational overhead and increase the disk throughput.