基于优化支持向量机的热耗率预测方法  

Heat rate forecasting based on optimized support vector machines

作　　者：刘超[1] 牛培峰[1] 段小龙[1] 李国强[1] 张维平[2] 陈科[1] 

机构地区：[1]燕山大学电气工程学院,河北秦皇岛066004 [2]秦皇岛职业技术学院机电工程系,河北秦皇岛066100

出　　处：《燕山大学学报》2015年第5期425-430,437,共7页Journal of Yanshan University

基　　金：国家自然科学基金资助项目(61403331;61573306);河北省自然科学基金资助项目(F2010001318)

摘　　要：针对传统的汽轮机热耗率计算模型精度较低的问题,构建了一种基于改进生物地理学优化算法优化支持向量机的热耗率预报模型。首先,提出改进的生物地理学优化算法以加强算法的优化能力,并通过4个典型的测试函数验证算法的有效性。其次,采用支持向量机建立汽轮机热耗率的预报模型,并选用径向基函数作为核函数,用改进的生物地理学优化算法优化该模型参数。最后,结合某火电站600 MW超临界汽轮机组现场数据进行热耗率预报的仿真研究,结果表明基于该算法建立的汽轮机热耗率预报模型具有较高的预报精度。To deal with the low precision of the traditional calculation model,a new optimized support vector machine（ SVM）based on improved biogeography-based optimization（ BBO） algorithm is introduced to prediction steam turbine heat rate. First,an improved BBO algorithm called B-BBO-SA is proposed to enhance the performance of original BBO algorithm by introducing a hybrid migration operator and simulated annealing. At the same time,the validity of B-BBO-SA is evaluated by four classical test functions. Then,SVM is employed to establish the prediction model of steam turbine heat rate,and the radial basis function is selected as the kernel function,in which,the B-BBO-SA is employed to serve as a method for pre-selecting SVM parameters in order to build heat rate prediction model and obtain a well-generalized model. Finally,a hybrid B-BBO-SA-SVM model is established for heat rate forecasting of a 600 MW supercritical steam turbine unit. Experimental results show that the optimized SVM model by BBBO-SA has well regression precision and generalization ability.

关 键 词：热耗率 汽轮机 生物地理学优化算法 支持向量机 

分 类 号：TK26[动力工程及工程热物理—动力机械及工程]