摘要
以求解汽轮机变工况运行时的最优主蒸汽压力为目标,首先根据汽轮机运行数据,利用回声状态网络(ESNs)建立热耗率预测模型,然后利用改进的磷虾群优化(I-KH)算法的全局搜索能力,在可行的主蒸汽压力范围内对所建预测模型进行主蒸汽压力寻优,并将优化后的汽轮机滑压运行曲线与厂家设计压力曲线进行对比.结果表明:优化后机组各个负荷下对应的热耗率均有所下降;按照优化后的汽轮机滑压运行曲线运行可有效降低机组热耗率.
To obtain the optimal main steam pressure of a steam turbine under variable working conditions, a heat rate prediction model was established using operating data o{ the turbine by echo state networks （ESNs）, with which the optimal initial steam pressure was searched in the range of permitted pressure by taking use of the global searching ability of improved krill herd algorithm （I-KH）, and the optimized tur- bine sliding pressure operation curve was compared with the design curve. Results show that the unit heat rate reduces at various loads after optimization, which could be effectively lowered if the turbine runs ac- cording to the optimized sliding pressure operation curve.
出处
《动力工程学报》
CAS
CSCD
北大核心
2015年第9期709-714,共6页
Journal of Chinese Society of Power Engineering
基金
国家自然科学基金资助项目(60774028)
河北省自然科学基金资助项目(F2010001318)
关键词
汽轮机
热耗率
回声状态网络
磷虾群优化算法
最优初压
steam turbine
heat rate
echo state network
krill herd algorithm
optimal initial steam pres-sure
