Modeling And Optimization Of Boiler Combustion Based On Improved Gray Wolf Algorithm

The combustion of coal in coal-fired power station boilers will emit acidic oxides and cause pollution to the environment.The reduction of environmental pollution on the basis of ensuring the combustion efficiency of the boiler is of great significance to my country’s energy conservation and emission reduction.Due to the complexity of the combustion process of coal-fired boilers and the coupling of parameters,traditional methods are difficult to establish an effective and accurate boiler combustion prediction model.This paper takes a 660MW ultra-supercritical once-through boiler as the research object,and uses BP neural network and support vector machine to model and predict the NOx emission and boiler thermal efficiency of the boiler.Among them,the boiler combustion characteristic model established by the BP neural network has a maximum relative error of 2.3%and an average relative error of 0.95%when predicting the NOx emission concentration.The maximum relative error of predicting the boiler thermal efficiency has reached 1.2%,and the average relative error The error is 0.74%.The maximum relative error of NOx emission concentration prediction using the boiler combustion characteristic model established by support vector machine is 1.3%,the average relative error is 0.3%,the maximum relative error of boiler thermal efficiency prediction is 1.84%,and the average relative error is 0.65%.The above simulation prediction results show that the two modeling methods have good prediction accuracy and generalization ability,which are in line with the expected results of the experiment.In addition,this paper also analyzes the influence of different coal mill combinations with different secondary air distribution modes(pagoda air distribution,inverted pagoda air distribution,equal air distribution,waist drum air distribution)on the combustion characteristics of the boiler.Experimental results It shows that the NOx emission concentration reaches the lowest 172.1mg/m3 when the ABCDE coal pulverizer is combined with the inverted pagoda air distribution,and the boiler thermal efficiency reaches the highest 94.58%when the ABDEF coal pulverizer is combined with the equal air distribution.In this paper,the basic gray wolf algorithm,particle swarm algorithm and improved gray wolf algorithm are used to optimize the support vector machine model.First,the performance of the three algorithms is compared and analyzed.The calculation results show that the improved gray wolf algorithm has the fastest iteration speed and smaller relative error.The optimized training sample NOx emission concentration average relative error is 0.19%,and the boiler thermal efficiency average relative error It is 0.49%.The iteration speed of the particle swarm algorithm is the slowest,the maximum relative error of the optimized NOx emission concentration is 0.22%,and the average relative error of the boiler thermal efficiency is 0.54%.The comparison shows that the improved gray wolf algorithm has better performance and more accurate prediction.Using the improved gray wolf algorithm to optimize the support vector machine parameters can further improve its generalization ability and prediction accuracy.Finally,a multi-objective combustion optimization study is carried out on coalfired boilers.Based on the above-established support vector machine model boiler combustion characteristics prediction model,the minimum NOx emission concentration and the maximum boiler thermal efficiency are used as the objective function,and the NSGA-Ⅲ algorithm is used respectively.And NSGA-Ⅱ algorithm for combustion multi-objective optimization.The experimental results show that the Pareto front obtained by using the NSGA-Ⅲ algorithm has better convergence and distribution.After the optimization of the NSGA-Ⅲ algorithm,the NOx emission concentration range is 155mg/m3-215mg/m3,compared with the NOx emission in the historical operating data The concentration range is 212mg/m3-302mg/m3,and the reduction ratio reaches 26.9%.At the same time,the optimized boiler thermal efficiency range is 93.5%-94.4%,which is 0.8%higher than the historical maximum efficiency of the boiler operation before optimization.The optimization results have certain reference significance for the economic and environmental protection operation of coal-fired power station boilers.