基于LS-SVM与GPC算法的锅炉燃烧优化控制  

A Study on the Optimization Control of the Boiler Combustion Based on the LS-SVM and the GPC Algorithms

作　　者：张仲琪 付树强 林涛[2] 李金京 

机构地区：[1]山西鲁能河曲发电有限公司,山西忻州036500 [2]上海交通大学,上海200240

出　　处：《重庆电力高等专科学校学报》2017年第6期40-45,共6页Journal of Chongqing Electric Power College

摘　　要：基于LS-SVM算法和反馈网络建立了非线性对象全参数动态模型,实现了对非线性系统动态特性的多步预测;考虑控制的实时性,研究了将非线性LS-SVM动态模型在线线性化和构造对象实时线性CARIMA动态模型的方法;并结合GPC算法,提出了LSSVM-GPC动态优化控制策略。文中验证了LSSVM-GPC动态优化控制算法的跟踪能力和抗干扰能力,通过仿真试验给出了该算法对电站锅炉燃烧系统具有优秀的动态调节性能。This paper introduces a nonlinear dynamic model based on the LS-SVN algorithm and a feedback network,which can realize the multi-step prediction of the dynamic characteristics of the nonlinear system. Considering realtime control,it probes into online linearization of the nonlinear LS-SVM dynamic model as well as the way of the establishment of the real-time local linear dynamic model by means of the CARIMA. In combination with the GPC algorithm,it also puts forward a dynamic optimization control strategy called LSSVM-GPC,the anti-tracking and antiinterference capabilities of which have been verified in the paper. The simulation experiment has proved its good dynamic regulative performance in the combustion system of the boiler.

关 键 词：最小二乘支持向量机 动态模型 非线性 广义预测控制 优化控制 

分 类 号：TM621.2[电气工程—电力系统及自动化] TP273[自动化与计算机技术—检测技术与自动化装置]