摘要
为了准确地进行燃煤机组负荷预测,引入支持向量机(SVM)方法建立了锅炉炉膛多层火焰图像与机组负荷之间的复杂关系模型.将该方法应用于某660 MW燃煤锅炉机组中,用训练后的SVM模型进行负荷预测,并与BP神经网络模型预测结果进行比较.结果表明:采用SVM方法预测机组负荷,模型能够辨识出火焰辐射图像与机组负荷之间的复杂关系,实现对负荷的准确预测;SVM模型预测精度比BP网络模型高,SVM模型具有预测精度高、泛化能力强等优点,且模型训练时间较短.
In order to accurately forecast the load of a coal-fired power unit,a new algorithm based on Support Vector Machine（SVM） method is presented,which establishes a model to reflect the complicated relations between the load and the furnace flame Images of the coal-fired unit.The trained SVM model has been applied to a 660MW coal-fired unit,of which the results are compared with that of BP neural network model.Results show that with SVM model,accurate load prediction can be realized,since the model is able to identify complicated relations between the load and the furnace flame Images.Compared with BP model,SVM model has a virtue of higher forecasting accuracy,stronger generalization performance and shorter training time.
出处
《动力工程学报》
CAS
CSCD
北大核心
2011年第8期619-623,643,共6页
Journal of Chinese Society of Power Engineering
基金
华北电力大学青年教师科研基金资助项目(200911032)
华北电力大学中央高校基本科研业务费专项资金资助项目(11QG72)
关键词
能源与动力工程
负荷预测
支持向量机
火焰图像
energy and power engineering
load forecasting
support vector machine
flame image
