自反馈RBF网络在高炉热状态模型预测中的应用

在RBF神经网络的输入上加入了自反馈的神经元,提出了自反馈RBF神经网络,使网络对过去时态的数据具有了记忆能力,对该网络进行了稳定性分析后,采用层叠的自反馈神经元以增加网络的动态处理能力,并设计了自反馈RBF的在线训练算法,通过对混沌时序数据的仿真实验证明该算法的有效性。在此研究基础上,建立了高炉的热状态预测控制模型,预报铁水中硅的含量以达到判断高炉热状态的目的,实验表明该模型提高了高炉热状态的预报精度。

Application of self-feedback RBF NN in prediction model for heat state of blast furnace

Self-feedback neural unit is added into the inputs of RBF NN and the self-feedback RBF which had the capacity of memorizing the past-time temporal data is presented.After analyzing stability of the network,the stacks of self-feedback neural are adopted to increase the dynamic handling capacity of network.And the online training algorithm is designed for self-feedback RBF.The simulation results from chaotic time series data indicates the proficiency of the algorithm.Based on these studies,prediction model of the heat state of blast furnace is builded,and used to forecast the content of silicon in molten iron to judge the blast furnace heat state.The experimental results show that this model to improve the forecasting accuracy of the blast furnace heat state.