基于先验知识的长短记忆RBF网络结构

Structure of RBF with long and short term memory based on field knowledge

提出了一种基于先验知识的RBF-LSTM(RBF with Long and Short Term Memory)网络结构。该网络将专业背景知识引入到神经网络的结构构造中,提出了具有长短期记忆功能的网络结构。同时引入了剪枝理论,使网络具有更精简的结构。将这种网络结构应用于热工过程中过热气温动态特性建模,仿真结果表明该神经网络模型具有较高的建模精度以及泛化能力。

Structure of PBF-IASTM （RBF with Long and Short Term Memory） based on field knowledge was proposed. This net usesd the field background knowledge to supervise the train, and the net had the memory of long term and short term. The net had a compact network structure because of pruning theory. The network was applied to build the dynamic model of the superheated steam temperature in thermal process. Simulate results show that the neural model has high approximation accuracy and good generalization ability.