Research On Boiler Combustion Optimization Based On Improved Multi Objective PSO Algorithm

With the rapid development of modern economy,to ameliorate the environment issues and to save energy have aroused wide concerns.To further improve the boiler efficiency and reduce the emissions of pollutants are of great significance for coal-fired power plant.Moreover,the optimization of boiler combustion takes the advantages of low cost and high controllability,meaning that conducting the research on the combustion optimization is necessary.Boiler combustion is a complicated process with a strong coupling and high delay.Considering the easily changeable electrical load and various kinds of coals that boilers burn,the present study aims to develop an ideal model for boiler combustion system which can use different types of coal under different loads,while remaining the best energy efficiency simultaneously.In this paper,based on the combustion characteristics of the boiler,least squares support vector machine(lssvm)was applied for regression modeling,and the optimized multiobjective particle swarm optimization(mopso)algorithm was thereafter employed for multi-objective purpose.In the case of modeling,the least squares support vector regression machine(lssvr),coupled with thermal operational data,was used to build an offline model.The obtained data was used to test this model,which was then compared to that of support vector regression machine.The results show that the lssvr model is more accurate in predicting and better in general fitting.In terms of the optimization of the operation parameters,this paper adopted an modified multi-objective pso algorithm,which uses chaotic inertia weight.the global optimal value was obtained according to the ratio of contemporary crowded distance in the external particles to its Euclidean distance.Particles in the external files were determined by stereo space grid,they were then combined with particle crowding distance to update the space,and thus obtaining the best data for evenly distributed Pareto.On the basis of the above algorithm,the result shows that the validity of the modified algorithm is very well after the optimization of secondary air through multi-objective approach.