基于LS-SVM的供热机组调峰能力预测方法研究  

Study on Prediction Method of Heat Supply Unit's Peak-load Regulation Capacity Based on LS-SVM Model

作　　者：杨海生 向鹏 王兴国 李浩 朱霄珣[3] YANG Hai-sheng;XIANG Peng;WANG Xing-guo;LI Hao;ZHU Xiao-xun(State Grid Hebei Energy Technologies Service Company Limited,Shijiazhuang 050090, China;School of Energy and Power Engineering,North China Electric Power University,Baoding 071003, China;State Grid Hebei Electric Power Company Limited Research Institute,Shijiazhuang 050022, China)

机构地区：[1]国网河北能源技术服务有限公司,石家庄050090 [2]国网河北省电力有限公司电力科学研究院,石家庄050022 [3]华北电力大学能源动力与机械工程学院,保定071003

出　　处：《汽轮机技术》2019年第1期63-66,共4页Turbine Technology

基　　金：国家科技支撑计划项目(2014BAA06B01)

摘　　要：针对传统研究供热机组调峰能力方法存在的工作量大、结果偏离实际运行工况等问题,提出一种基于最小二乘支持向量机(LS-SVM)预测机组调峰能力的方法。对机组运行数据预处理后,选择部分数据建立训练样本集训练LS-SVM模型,然后利用训练出的模型得到供热机组在不同热负荷下的调峰范围,并以某供热机组在供热期的运行数据为例,得到了该供热机组的调峰能力,结果表明,基于最小二乘支持向量机得出的调峰能力区间比较准确,并且与传统方法相比在实际运行中可复现性更好。Aiming at the problems such as large workload and deviating from actual operating conditions in the traditional research methods of heat supply unit, a method based on least squares support vector machine (LS-SVM) is proposed. After preprocessing the unit's operation data, part of the data is selected to establish a training sample set and train the LS-SVM model. Then use the trained model to obtain the peak-load regulation capacity range of the heat supply unit under different heat loads. Taking the operation data of a heat supply unit during the heating period as an example, the peaking capacity of the heat supply unit was obtained. The results show that the peak-load regulation capability range based on the least squares support vector machine is accurate, and is more repeatable than the traditional method in actual operation.

关 键 词：供热机组 调峰能力 最小二乘支持向量机 运行数据 预测 

分 类 号：TK212[动力工程及工程热物理—动力机械及工程]