基于LS-SVM的汽轮机组加热器运行经济性评价方法  

Least square support vector machine based economic evaluation method for heaters in steam turbines

作　　者：李勇[1] 刘哲[1] 

机构地区：[1]东北电力大学能源与动力工程学院,吉林吉林132012

出　　处：《热力发电》2014年第2期9-11,15,共4页Thermal Power Generation

基　　金：吉林省科技发展计划项目(20110409)

摘　　要：提出了采用最小二乘支持向量机(LS-SVM)及选择径向基(RBF)核函数模型计算汽轮机组低压加热器端差应达值与相应实测值之间的偏差对加热器性能进行评价的方法。以某300MW机组6号低压加热器的正常运行状态为例,验证了采用LSSVM模型计算加热器端差应达值的可行性,并应用该计算模型对6号低压加热器在故障状态下的经济性进行计算分析。结果表明,该加热器上、下端差的实测值与应达值之间的均方差分别为0.536 3、5.055 3℃,均远大于其正常状态时的相应均方差值(0.005 3、0.007 8℃),说明该加热器发生了故障。经检查,证明其故障为该机组6、7号低压加热器之间的旁路门发生内漏及6号低压加热器水位偏低。表明该计算模型可应用于低压加热器运行热经济性评价,并可实现对加热器故障的早期发现。The least squares support vector machine （LS-SVM） and RBF kernel function model were applied to calculate the target value of terminal temperature difference of low-pressure heat- ers ,and performance evaluation of the heaters was conducted according to the difference between the target value and measured value. Taking normal operation state of the No. 6 low-pressure heater in a 300 MW unit turbine as the example,the feasibility of using the LS-SVM model to cal- culate the target value of terminal temperature difference of the low-pressure heater was verified. Finally,economical efficiency of the heater at fault state was evaluated by the LS-SVM model. The results showed that, the mean square deviations between the measured and target value of up- per and bottom terminal temperature difference of the heater were 0. 536 3 ℃ and 5. 055 3 ℃ ,re- spectively,both far larger than that at the normal operation state （0. 005 3 ℃ and 0. 007 8 ℃）. This indicated fault has happened on the heater. The inspection validated that, leakage occurred on bypass door between the No. 6 and No. 7 low-pressure heater, and water level in the No. 6 low- pressure heater was too low. Therefore, the LS-SVM model was preferred to evaluate the econom- ic efficiency of the low-pressure heater＇s performance,and it also can discover the heater fault at an early time.

关 键 词：LS-SVM 300 MW机组 低压加热器 端差 应达值 

分 类 号：TK267[动力工程及工程热物理—动力机械及工程]