Model Predictive Control Algorithm And System For Boiler Combution Efficiency In Power Station

The boiler operation in power station is affected by many factors.For one thing,the combustion efficiency as the important indicator of reaction fuel burning degree,it directly influences the production efficiency and economic benefits of power station.For anther thing,the Nitrogen Oxides(NOx)produced by combustion are the main source of the gas pollution.Therefore,it is the hot issues of boiler research to optimal control of the productive process of the station boiler to achieve high efficiency and low emission.For the problems in the control of power plant boilers,the data driving model of boiler combustion parameters(boiler combustion efficiency,NOx emission and boiler load)were constructed with 660 MW power plant boiler as the research object,the model predictive control algorithm was designed and carried out experimental research,the specific studies as follows:Firstly,based on the basic principle of least squares support vector machine(LS-SVM),a multiple working conditions data-driven modeling method(DDMMF)with feature selection was proposed to establish a prediction model for key parameters of boiler combustion,including: 1)The classification regression tree(CART)algorithm was employed to analyze the actual production data,and the variables with the importance greater than 5% were selected as the model input.2)The Kth nearest neighbor(KNN)classifier was used to classify the samples to distinguish the production data belong to different working conditions;3)A data-driven model based on LSSVM of differential evolution algorithm(DE)was designed,and DE dynamic was adopted to optimize LSSVM parameters to construct a boiler critical parameter prediction model under different conditions;4)The prediction model was dynamically corrected for further improvement of the prediction accuracy.Secondly,based on boiler key parameter model,the boiler combustion efficiency rolling optimization model was constructed by considering load constraints,pollutant emission constraints and boundary constraints,and the optimal controller parameters were calculated via an optimization model solved by the DE algorithm.Finally,the boiler combustion efficiency predictive control system was established to realize the user management,production data management,formulation of control strategy,model management and graphical display of model prediction control results,and assist the operators to effectively adjust the important parameters of the boiler.The experimental results based on the actual production data demonstrate: 1)The predictive model can accurately predict the boiler key parameters and meet the demands of boiler combustion process control and optimizing;2)The predictive control algorithm of model can effectively control the boiler combustion efficiency,the average error of simulation is less than 3%.