基于铝电解过程的神经网络模型预测控制的应用研究

铝电解过程是一个复杂的非线性、时变和大时滞的工业过程体系,采用常规的控制方法,很难达到良好的控制效果。本文提出了一种基于铝电解过程的神经网络模型预测控制算法,建立了神经网络预测模型,将神经网络和模型预测控制算法相结合,实现了铝电解过程的最优控制。仿真结果表明:神经网络预测模型的输出能够很好地跟踪铝电解生产过程,其预测效果好,该预测控制方法可以使系统很快地达到稳态,具有很好的响应特性和鲁棒性。

The Research and application Of Neural Network model predictive control based on Aluminum Electrolysis Process

Aluminum electrolysis process is an industrial process system which has complicated nonlinearity ,time variation and big delay. Using regular control method is hard to attain good control result. This paper presents an neural network model predictive control algorithm based on aluminum electrolysis process, establishes the predictive model of neural network, combines the neural network and predictives control algorithm, and realizes the optimize control of aluminum electrolysis process. The simulation results show the predictive model can trace the aluminum electrolysis process. The predictive result is good. It can make the system attain stable state and has good response and robustness.