Study On Multi-objective Optimization On Economic Operation And NO_x Emissions For Coal-fired Boiler Based On Intelligent Algorithm

Intelligent algorithm was applied to multi-objective optimization of boiler combustion is one of the important measures for energy conservation and emissions reduction, in this paper, based on a variety of intelligent algorithm, multi-objective optimization was carried out on the boiler combustion. First of all, taking a 660 MW coal-fired boiler as the research object, using BP artificial neural network to establish the boiler combustion prediction model in the case of a 660 MW load, with 23 boiler combustion operation parameters as the input, the boiler thermal efficiency and NO_x emissions for the output. The range of model training error is-1.1?10-9 and 1.33?10-9,the absolute value of NO_x emission calibration sample average error is 3.0398%, the absolute value of the boiler thermal efficiency calibration sample average error is0.1129%, the prediction model has higher accuracy and good generalization. Based on the prediction models of the combustion, boiler combustion optimization model is set up using genetic algorithm optimization algorithm, among them, the weight coefficient method is used to transformed the multi-objective optimization problem into a single objective optimization problem, and different weight ratio(0.1-0.9 ?0.2-0.8?0.3-0.7?0.4-0.6?0.5-0.5?0.6-0.4?0.7-0.3?0.8-0.2) were optimized respectively. The optimizing results under different weight ratio is different, as the weight ratio of NO_x emissions and the thermal efficiency of the boiler from 1-9gradually increase to 8-2, the value of NO_x emissions by 176 mg/m3 decline gradually to 111 mg/m3, the boiler combustion heat loss gradually rise from4.24%( The boiler thermal efficiency is 95.76%) to 6.05%( The boiler thermal efficiency is 93.95%). All of its solution set form the Pareto solution set, and presents the concave Pareto frontier. The fourth chapter introduced the multi-objective evolutionary algorithm based on decomposition to the boiler combustion multi-objective optimization, Chebyshev decomposition strategy is used for boiler combustion optimization, the optimized NO_x emissions in the range of 112 mg/m3~183mg/m3, boiler heat loss is 4.3% ~ 5.8%, compared with the traditional genetic algorithm, the optimization effect is a bit poor, and each have advantages anddisadvantages of both methods.In this paper, using the other 330 MW boiler as the research object at the same time, first of all, pay attention to the boiler full load operation condition, improve the original genetic algorithm, and can get an improved BP- GA optimization model, in an acceptable range for the boiler heat efficiency, optimization for NO_x emission.The boiler thermal efficiency constraints was 93.5%, 93%, 92.5%, 93%, the optimized NO_x emissions down to 380 mg/m3,350 mg/m3,320mg/m3,310 mg/m3 respectively, the optimization effect is obvious. According to the boiler, based on the running data under the 300 MW and 270 MW load, vector evaluation of genetic algorithm optimization model is established, the condition of high quality solution set of 300 MW boiler thermal efficiency and NO_x emissions is 92.93% ~ 92.93% and367 mg/m3~413 mg/m3 respectively, the condition of high quality solution set of270 MW boiler thermal efficiency and NO_x emissions is 92.26%~93.56% and 360mg/m3~416 mg/m3 respectively, both optimal value set is good.