摘要
利用支持向量回归算法,建立了汽轮机热耗率计算模型。介绍了支持向量回归算法的原理,对算法中的参数选择进行了探讨。对某300MW机组汽轮机热耗率计算进行了建模,并与RBF神经网络回归模型进行了比较。结果表明:基于支持向量回归算法的模型具有较强的泛化能力,适于在线应用。通过对输入参数添加随机扰动量分析表明,该模型比传统计算模型具有更好的稳定性,能更准确地计算汽轮机热耗率。
A calculation model of steam turbine heat rate is founded by using Support Vector Regression （SVR）. The algorithm of SVR is represented and parameter selection of the algorithm is discussed. Calculation model of heat rate for a 300 MW steam turbine is built, which is compared with the model of hear rate based on RBF. it is indicated that the model based on SVR has more forcible generalization ability and can be applied on-line. The analyzing through adding a random distrubance quantity to input parameter indicates that this model has better stability than conventional method and can calculate the heat rate of steam turbine more accurately. It is an effective method for modeling the operating performance of the steam turbine. Figs 5, table 1 and refs 8.
出处
《动力工程》
EI
CSCD
北大核心
2007年第1期19-23,49,共6页
Power Engineering
关键词
动力机械工程
汽轮机
热耗率
支持向量机
回归分析
power and mechanical engineering
steam turbine
heat rate
support vector machine
regression analysis
