一种基于NARX神经网络的非线性多步预测模型
首发时间:2013-01-02
摘要:神经网络是一种非常优秀的非线性建模工具,在工程实际中得到了广泛的应用。根据网络结构中是否存在反馈回路,神经网络可分为动态神经网络和静态神经网络。本文针对工业过程中具有强非线性特点的被控对象,提出了一种基于动态神经网络的非线性多步预测模型;设计了非线性预测模型的结构辨识方法,并对模型的预测性能进行了仿真验证。仿真结果表明,该预测模型能够较好地预测非线性系统的未来输出信息,为非线性多步预测控制器的设计和应用打下了良好的基础。
关键词: 模型预测控制 预测模型 递归神经网络 NARX神经网络 模型辨识
For information in English, please click here
A Multi-Step Nonlinear Predictive Model Based on NARX Neural Network
Abstract:Neural Network is a kind of execllent nonlinear modeling tools. According to the structure, neural network can be divided into two categories which are dynamic neural network and static neural network. This paper proposed a nonlinear multi-step predictive model based on dynamic neural network for the systems which have strong nonlinearity, designed a special model identification method, and provided simulation experiment. The simulation results show that the nonlinear predictive model can give relatively good predictive output information of the nonlinear system, and can be used in multi-step predictive controller.
Keywords: Model Predictive Control Predictive Model Recurrent Neural Network NARX Neural Network Model Identification
论文图表:
引用
No.****
同行评议
共计0人参与
勘误表
一种基于NARX神经网络的非线性多步预测模型
评论
全部评论0/1000