浙江电力

2019, v.38;No.283(11) 103-108

[打印本页] [关闭]
本期目录(Current Issue) | 过刊浏览(Archive) | 高级检索(Advanced Search)

基于最小二乘支持向量机的汽轮机低压缸排汽焓计算
Calculation of Low-pressure Cylinder Exhaust Enthalpy Based on Least Squares Support Vector Machine

杨斌,柳琦,张芹,高原,雷鸣,余鹏,何皓,刘真全
YANG Bin,LIU Qi,ZHANG Qin,GAO Yuan,LEI Ming,YU Peng,HE Hao,LIU Zhenquan

摘要(Abstract):

为了在线计算汽轮发电机组的经济性,基于LSSVM(最小二乘支持向量机)建立了一种汽轮机低压缸排汽焓在线计算模型。首先分析汽轮机低压缸排汽焓影响因素,确定LSSVM模型的输入变量与输出变量,采集历史数据,数据预处理后剔除明显坏点,再对各参数进行归一化,将其转化为无量纲量,最后将归一化处理后的数据用于LSSVM模型的训练,再用性能试验的数据对模型进行验证,得到基于LSSVM的汽轮机低压缸排汽焓计算模型。结果表明:基于LSSVM的汽轮机低压缸排汽焓计算模型能够有效预测低压缸排汽焓,误差范围在1%以内,低压缸排汽焓的预测值比试验值平均小约5 kJ/kg。低压缸排汽焓的预测值与试验值保持着相同的变化规律。
For online calculation of the economy of steam turbine units, a calculation model for low-pressure cylinder exhaust enthalpy based on LSSVM(least squares support vector machine) is established. Firstly, the influencing factors of low-pressure cylinder exhaust enthalpy are analyzed to determine the input variations and output variations of LSSVM, collect historical data and weed out the bad points after data pre-processing;secondly, the parameters are normalized and converted into dimensionless variables, and then the data after normalization is used for LSSVM model training. Finally, the model is verified by the data of the performance test to obtain a calculation model low-pressure cylinder exhaust enthalpy based on LSSVM. The results show that the calculation model for low-pressure cylinder exhaust enthalpy based on least squares support vector machine can effectively predict low-pressure cylinder exhaust enthalpy, and the error range is within 1%; the predicted value of low-pressure cylinder exhaust enthalpy is 5 kJ/kg smaller than the measured average value;the predicted value and the test value of low-pressure cylinder exhaust enthalpy share the same change law.

关键词(KeyWords): 汽轮机;经济性;排汽焓;最小二乘支持向量机
steam turbine;economy;exhaust enthalpy;LSSVM

Abstract:

Keywords:

基金项目(Foundation):

作者(Author): 杨斌,柳琦,张芹,高原,雷鸣,余鹏,何皓,刘真全
YANG Bin,LIU Qi,ZHANG Qin,GAO Yuan,LEI Ming,YU Peng,HE Hao,LIU Zhenquan

DOI: 10.19585/j.zjdl.201911017

参考文献(References):

扩展功能
本文信息
服务与反馈
本文关键词相关文章
本文作者相关文章
中国知网
分享