【作者】 翟永杰； 乔弘； 李海丽； 韩璞； 刘柱云；

【Author】 ZHAI Yong-jie,QIAO Hong,LI Hai-li,HAN Pu,LIU Zhu-yun 1.School of Control Science and Engineering,North China Electric Power University,Baoding,Hebei 071003,China 2.School of Energy and Power Engineering,North China Electric Power University,Beijing 102206,China 3.Guodian Nanjing Automation Co.,LTD,Nanjing 210003,China

【机构】 华北电力大学控制科学与工程学院； 华北电力大学能源与动力工程学院； 国电南京自动化股份有限公司；

【摘要】 影响火电厂经济运行的一个重要因素是许多重要热工技术参数和经济参数难以在线实时测量,为解决热工参数软测量建模问题,对粒子群优化算法与序列最小优化算法进行了结合研究,提出了粒子群SMO算法。该方法首先从训练数据的性能信息中挖掘SMO学习机参数,以此指导粒子群的速度和范围设置,然后将改进SMO算法同PSO算法结合进行双层结构参数寻优,有效减少人为因素和数据噪声的影响,使算法建模具有较高精度。结合实际工艺,将该方法应用于氧量软测量建模。仿真结果表明,该模型能够准确预测氧量变化,具有较高的精度和良好的应用前景。更多还原

【Abstract】 An important factor impacting the economic operation of power plant is many important technical parameters and economic parameters are difficult to be real-time measured online.To solve the problem of soft measurement modeling of thermal parameters,we research the PSO algorithm and sequential minimal optimization algorithm and propose SMO algorithm based on PSO.Firstly,the method mines SMO learning machine parameters from the performance information of training data to determine the pace and scope settings of PSO parameters;then combines the improved SMO algorithm with PSO algorithm to optimize double-layer structure parameters.It can reduce the impact of noise data and human factors effectively and make the modeling algorithm have higher accuracy.With the actual technology,the method has been applied to soft measurement modeling of oxygen.Simulation results show that the model can predict oxygen changes accurately and it has high accuracy and good prospects.更多还原