【摘要】：燃煤電站鍋爐不僅消耗大量的煤炭資源,也排放出了大量的大氣污染物。在提高鍋爐效率降低煤耗和控制污染物排放來緩解高耗能和重污染的突出問題上,基于智能算法的鍋爐燃燒優(yōu)化方法相比傳統(tǒng)方法有其獨特優(yōu)勢,可以為電站鍋爐高效率低污染運行提供有效的指導。通過分析300MW鍋爐和1000MW鍋爐的運行數(shù)據(jù),利用支持向量回歸機建立了NOx排放SVM模型和鍋爐熱效率SVM模型,比較了基于RBF核函數(shù)和Sigmoid核函數(shù)的模型性能；分別采用自適應遺傳算法(GA)和果蠅優(yōu)化算法(FOA)對模型參數(shù)進行優(yōu)化,并比較了各自模型的泛化能力和預測精度性能,結(jié)果表明FOA優(yōu)化的SVM模型更具有優(yōu)勢。以建立好的鍋爐NOx排放和鍋爐熱效率的SVM模型為基礎,結(jié)合自適應遺傳算法,在不考慮對效率影響的情況下,單一對NOx排放進行優(yōu)化時,NOx排放水平在優(yōu)化后能夠得到大幅降低,但是與此同時也會造成鍋爐熱效率的下降；在不考慮對NOx排放影響的情況下,單一對鍋爐熱效率進行優(yōu)化時,鍋爐熱效率在優(yōu)化后能有較大的增幅,但是同時會導致NOx排放量有所上升。文中以1000MW機組鍋爐為對象,在建立NOx生成量和鍋爐熱效率SVM模型的基礎上提出了一種運用改進的多目標遺傳算法(改進型NSGA-Ⅱ)進行多目標燃燒優(yōu)化的方法,同時考慮鍋爐熱效率和NOx生成這兩個目標,對鍋爐運行參數(shù)進行優(yōu)化,得出了由一系列可行解組成的Pareto最優(yōu)解集,其中有很多可行解同時滿足鍋爐效率的提高和NOx生成的降低這兩個目標,為機組運行人員提供參考,達到高效率且低NOx生成的鍋爐燃燒優(yōu)化目的。
[Abstract]:The coal-fired utility boiler not only consumes a lot of coal resources, but also emits a lot of air pollutants. In order to improve boiler efficiency, reduce coal consumption and control pollutant emissions to alleviate the problem of high energy consumption and heavy pollution, the combustion optimization method based on intelligent algorithm has its unique advantages compared with traditional methods. It can provide effective guidance for high efficiency and low pollution operation of utility boiler. By analyzing the operation data of 300MW boiler and 1000MW boiler, the SVM model of NOx emission and the SVM model of boiler thermal efficiency are established by using support vector regression machine. The performance of the model based on RBF kernel function and Sigmoid kernel function is compared. Adaptive genetic algorithm (GA) and Drosophila Optimization algorithm (FOA) are used to optimize the model parameters, and the generalization ability and prediction accuracy of each model are compared. The results show that the SVM model optimized by FOA has more advantages. Based on the established SVM model of boiler NOx emission and boiler thermal efficiency, combined with adaptive genetic algorithm, without considering the effect on efficiency, the level of NOx emission can be greatly reduced when single optimization of NOx emissions is carried out. But at the same time, the boiler thermal efficiency will decrease. Without considering the influence of NOx emission, the boiler thermal efficiency can increase greatly when the boiler thermal efficiency is optimized. But it will also lead to an increase in NOx emissions. Taking the boiler of 1000MW unit as an object, based on the establishment of SVM model of NOx production and boiler heat efficiency, a method of multi-objective combustion optimization using improved multi-objective genetic algorithm (NSGA- 鈪，

本文編號：2199802