基于WLS-SVM-SFS模型的电站锅炉燃烧优化  

Combustion Optimization for Utility Boilers Based on WLS-SVM-SFS Model

作　　者：牛玉广[1] 沙超[1] 孟庆伟[1] 谷佳琪 靳鑫[1] 

机构地区：[1]华北电力大学新能源电力系统国家重点实验室,北京102206

出　　处：《热力发电》2014年第5期7-12,共6页Thermal Power Generation

基　　金：国家科技支撑计划项目:大型电站燃煤锅炉燃烧在线优化节能减排技术(2011BAA04B01)

摘　　要：NOx排放量和锅炉效率模型是电站锅炉燃烧优化的基础。采用抗干扰能力更强的加权最小二乘支持向量机(WLS-SVM)建立了NOx排放模型。将序列前向选择(SFS)与WLS-SVM相结合建立了锅炉效率模型,在不影响模型精度前提下去除了模型中的冗余成分,精简了模型结构,提高了模型计算速度。采用遗传算法,以所建模型为基础,提出了一种可兼顾锅炉效率和NOx排放量的优化燃烧方案。实际应用结果表明,该优化方案使锅炉效率平均提高0.386%,NOx排放量平均降低99.147mg/m3。NOx emission and boiler efficiency model is the foundation of boiler combustion optimi-zation.The weighted least squares support vector machine （WLS-SVM）was employed to estab-lish the NOx emission model.By combining the WLS-SVM with sequential forward selection （SFS）,the boiler efficiency model was built.Moreover,on the premise of not affecting the model accuracy,the redundancy compositions in the model was reduced,which simplifies the model structure and enhances the calculation speed.On the basis of the established models,an optimiza-tion method which can effectively balance the boiler efficiency and NOx emissions was put for-ward,by applying the genetic algorithm.The results show that this optimization program im-proved the boiler efficiency and NOx emission by 0.386% and 99.147 mg/m3 on average,respec-tively.

关 键 词：NOx 排放量 锅炉效率 燃烧优化 

分 类 号：TK223[动力工程及工程热物理—动力机械及工程]