基于遗传算法和支持向量机的低NO_x燃烧优化  

Support Vector Machine and Genetic Algorithms to Optimize Combustion for Low NO_x Emission

作　　者：王春林[1] 周昊[1] 李国能[1] 凌忠钱[1] 岑可法[1] 

机构地区：[1]能源清洁利用国家重点实验室(浙江大学),浙江省杭州市310027

出　　处：《中国电机工程学报》2007年第11期40-44,共5页Proceedings of the CSEE

基　　金：国家自然科学基金项目(60534030;50576081)。~~

摘　　要：大型四角切圆电站锅炉NOx排放是造成环境污染的重要因素,也是电厂关心的重要问题。影响燃煤锅炉NOx排放量的因素众多而且复杂。对锅炉NOx排放特性进行建模预测,并结合优化算法实现燃烧优化是降低锅炉NOx排放的有效方法。文中应用支持向量机算法建立了大型四角切圆燃烧锅炉NOx排放特性模型,接合遗传算法,利用NOx排放的热态实炉试验数据对模型进行了校验,对锅炉运行参数进行了优化。结果表明,通过遗传算法的寻优,NOx排放量有比较明显的降低。支持向量机与遗传算法相结合与其它方法相比具有泛化能力好,计算速度快等优点,是锅炉NOx排放控制的有效工具。NOx emission is a main factor that has great impacts on the environment. It was affected by many factors and complicated. Building a model to predict NOx emission is a good way to optimize the coal combustion and reduce NOx emission. A support vector machine （SVM） model predicting the NOx emission of a high capacity boiler was developed and verified. Good predicting performance was achieved with the proper learning parameters choosing by genetic algorithms. Low NOx emissions were achieved by combing genetic algorithms and SVM model to optimize operating parameters. The modeling results show that the combination of support vector machine and genetic algorithms has good ability to optimize combustion, it has good generalization ability and higher calculation speed comparing with other approaches.

关 键 词：锅炉 燃烧 NOx 支持向量机 遗传算法 

分 类 号：TK223[动力工程及工程热物理—动力机械及工程]