摘要
提出了一种融自组织神经网络和ＢＰ神经网络为一体的混合神经网络模型，用于电厂汽轮机系统的热耗分析。在对输入向量附加５％ 的随机量值时，输出的热耗值和原来的值是一致的，这说明ＢＰ网络具有很好的稳定性。这个特性尤其适合在线的热耗分析。图５ 表１
A hybrid artificial neural network model used to analyze the heat rate and the sensitivity of thermal parameters on it is proposed. The modeling process involving a self organizing network can rearrange the original data into several classes that are used as the training patterns of the BP artificial neural network. The stability analysis shows that the output result is still stability when 5% random value is added to the input vectors. Therefore, the BP network has its excellent robustness. Figs5, table 1 and refs 4.
出处
《动力工程》
CSCD
1999年第5期342-345,共4页
Power Engineering
关键词
电厂
汽轮机
热力系统
神经网络
性能分析
power plant steam turbine
thermal system
neural network
performance analyses
