以某330MW煤粉汽包锅炉为测试对象,以支持向量回归机为基础,利用最小二乘支持向量机和一种新的寻优算法———万有引力搜索算法进行了综合建模及参数优化.结果表明:建立的模型较好地实现了对电厂锅炉NOx质量浓度排放量的预测;与遗传算法、蜂群算法和粒子群算法相比,万有引力搜索算法能更好地找到未知优化参数,使得所建模型具有更高的预测能力和泛化能力,从而有效控制燃煤电厂NOx的排放量. Taking a 330 MW pulverized coal-fired drum boiler as the test object and the support vector regression machine as the foundation, the integrated modeling and parameter optimization are carried out by using the least square support vector machine and a new optimization algorithm - gravitation search algorithm. The results show that the established model can better predict the NOx emission of power plant boilers. Compared with genetic algorithm, bee colony algorithm and particle swarm optimization algorithm, the gravitation search algorithm can better find unknown optimization parameters, Model has a higher predictive ability and generalization ability, so as to effectively control the NOx emissions from coal-fired power plants.