摘要
将灰色关联度分析方法应用于汽轮机性能预测BP神经网络输入层神经元的筛选。方法对样本数量、分布规律要求不高、量化结果与定性分析一致,有利于减少对技术人员经验的依赖,为汽轮机性能预测BP神经网络输入层神经元的筛选提供了科学依据。最后通过实例验证了所提出的方法的可行性。
This paper applies the analytical method of grey relational grade to screening input nodes in BP artificial neural network on performance forecasting of steam turbine. This method does not ask too much of specimen quantity and distributed law, and the quantification result consists with the qualitative analysis. It is beneficial to reducing the dependence on the experience of the technicist. This method provides scientific basis for screening input nodes in BP artificial neural network on performance forecasting of steam turbine. Feasibility on the method put forward is verified with an example in the end.
出处
《汽轮机技术》
北大核心
2010年第2期147-149,共3页
Turbine Technology
基金
河北省电力公司科研项目(KJ2008-041)
关键词
汽轮机
性能预测
BP神经网络
灰色关联度
输入层神经元
steam turbine
performance forecasting
BP artificial neural network
grey relational grade
input nodes
