摘要
汽包水位预测的准确性对于锅炉火力发电和供暖设备的安全运行,起着重要的作用。针对汽包水位时滞、非线性特性,在统计学习理论(SLT)和结构风险最小化(ERM)准则基础上,建立了基于支持向量机(SVM)理论的汽包水位预测模型。采用本单位供暖锅炉的实际汽包水位数据,按照不同的汽包水位日属性和历史汽包水位数据进行了样本选择,并将该预测结果同实际汽包水位进行了比较;根据供暖规模不同,分别建立了工作日和双休日预测模型。实验结果表明,所提出的预测方法具有较高的精度。
Water-level measurement system for boiler is one of the key measuring systems in power plant.And it is important to measure and predict the drum water level.As drum water level of boiler is a time-delay,time-varying and nonlinearity system,a new mathematical model according to the support vector machines（SVM） theory was brought out,which was based on the statistical learning theory（SLT） and the empirical risk minimization（ERM） principle.Adopting the actual data from the heating boiler of author affiliat-ion,the samples were chosen according to different attributes of daily water level and historical load data.The forecasted results were compared with actual boiler water lev-el according to the scale of heating.The results show that the presented prediction method is more accurate.
出处
《机电工程》
CAS
2010年第7期29-32,共4页
Journal of Mechanical & Electrical Engineering
关键词
支持向量机
汽包水位
预测
统计学习理论
结构风险最小化
support vector machine（SVM）
drum boiler water level
prediction
statistical learning theory（SLT）
empirical risk minimiza-tion（ERM）
