An Application Research On The Pulverized Coal Boiler Combustion Optimization Based On The Ameliorated Flower Pollination Algorithm

The unique energy structure of our country determines the dominant position of thermal power generation in the power generation industry.More and more attention has been paid to the environmental pollution and energy waste caused by thermal power generation.How to realize high efficiency clean combustion of the boiler has become an urgent problem for thermal power plants.In recent years,researchers in boiler combustion regulation field have devoted themselves to improving the combustion methods of the pulverized coal boiler to achieve the goal of reducing nitrogen oxide emission and improving operational efficiency.However,the boiler combustion system is very complex,and it is difficult to achieve combustion optimization using traditional optimization methods.To optimize the combustion condition of the pulverized coal boiler,the research issue of this paper is how to use flower pollination algorithm and extreme learning machine to model and optimize the operation condition to reduce nitrogen oxide emission and improve thermal efficiency of the pulverized coal boiler.Firstly,an Ameliorated Flower Pollination Algorithm(AFPA)is proposed to improve search ability and simultaneously quicken the convergence speed of basic Flower Pollination Algorithm(FPA).The highlights of AFPA can be classified into the following two aspects.First,we modify the local search method of FPA.Second,the switch probability of FPA is improved.Simulation experiments based on ten sets of test benchmark problem demonstrate that the AFPA has better search performance compared with other three optimization algorithms including FPA,differential evolution algorithm and bat algorithm.Then,based on the historical operating data of the 330 MW pulverized coal boiler from a thermal power plant,Extreme Learning Machine(ELM)and AFPA are used to establish nitrogen oxide emission model and thermal efficiency model of the boiler.To verify the generalization ability of the established model,ELM and back propagation(BP)neural network are adopted to establish establish nitrogen oxide emission model and thermal efficiency model using the same data.The experimental results show that the model built by AFPA and ELM has better prediction accuracy than other two methods.Finally,according to the actual operation of the power plant,three optimization targets are proposed.The optimization targets are low emission of nitrogen oxides,high thermal efficiency and improving the thermal efficiency on the basis of reducing the emission of nitrogen oxides,respectively.Based on the established model and different optimization objective functions,AFPA is used to optimize the adjustable parameters of the pulverized coal boiler in the constraint area.The simulation experiments show that AFPA can not only optimize the nitrogen oxide emission and thermal efficiency of the pulverized coal fired boiler,but also can reduce the nitrogen oxide emissions of the pulverized coal fired boiler while improving the thermal efficiency,which may have certain reference significance for boiler combustion optimization in the power station.