基于改进果蝇算法的电站锅炉燃烧建模与优化

发布时间：2018-05-18 15:58

本文选题：电站锅炉 + 支持向量机　；参考：《哈尔滨理工大学》2017年硕士论文

【摘要】：电站锅炉不仅需要大量的煤炭资源,也排放出许多的大气污染物。在处理锅炉效率和控制污染物这个问题上,本文基于改进果蝇算法优化的锅炉燃烧方法,相比原始果蝇算法具有明显的特性,可以为电站锅炉高效率、低污染运行提供可靠的指导。为达到提高锅炉热效率同时减少NOx排放的目标,提出改进果蝇算法(MFOA)优化支持向量机(SVM)的锅炉建模方法。针对果蝇(FOA)算法寻优精度低、收敛速度慢的问题,本文使用三维搜索及自适应变步长的策略改进果蝇算法,并完成对SVM中的惩罚因子C、核参数g和不敏感损失系数ε,这三个参数的寻优,使支持向量机对锅炉燃烧系统模型预测更加准确。另外根据不同时间段的样本数据来检验MFOA-SVM模型的预测能力,并且与未改进果蝇算法优化支持向量机(FOA-SVM)的预测效果进行比较,仿真结果表明,改进的果蝇算法具有较强的参数寻优能力,能够有效提高SVM模型的准确性,进而提高SVM模型寻优精度;与FOA-SVM模型对比,MFOA-SVM模型寻优的精度更高,收敛速度更快,可以更准确地预测锅炉热效率和NO_x排放。最后以所建燃烧模型为基础,使用MFOA算法对锅炉进行单目标和多目标优化,优化结果表明,文中所提出的燃烧优化方案可以有效提高锅炉效率和降低NO_x排放量。
[Abstract]:Utility boilers not only need a large amount of coal resources, but also emit a lot of air pollutants. On the problem of boiler efficiency and pollutant control, this paper based on the improved Drosophila algorithm optimization boiler combustion method, compared with the original Drosophila algorithm has obvious characteristics, and can be used for the utility boiler high efficiency. Low pollution operation provides reliable guidance. In order to improve boiler thermal efficiency and reduce NOx emission, a boiler modeling method based on improved Drosophila algorithm (MFOAA) and optimized support vector machine (SVM) was proposed. In order to solve the problem of low precision and slow convergence rate of FOAA algorithm, this paper improves the algorithm of Drosophila by using three dimensional search and adaptive variable step size strategy. The penalty factor C, kernel parameter g and insensitive loss coefficient 蔚 in SVM are optimized, which makes the prediction of boiler combustion system model more accurate by support vector machine (SVM). In addition, the prediction ability of MFOA-SVM model is tested according to the sample data of different time periods, and compared with the prediction effect of unimproved Drosophila algorithm optimization support vector machine (FOA-SVM). The simulation results show that, The improved Drosophila algorithm has strong parameter optimization ability, which can effectively improve the accuracy of SVM model and improve the accuracy of SVM model, compared with FOA-SVM model, MFOA-SVM model has higher accuracy and faster convergence speed. Boiler thermal efficiency and NO_x emissions can be predicted more accurately. Finally, based on the established combustion model, the single-objective and multi-objective optimization of boiler is carried out by using MFOA algorithm. The optimization results show that the proposed combustion optimization scheme can effectively improve boiler efficiency and reduce NO_x emissions.
【学位授予单位】：哈尔滨理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP18;TM621.2

【相似文献】