基于T-S模糊神经网络的Hammerstein模型预测控制

Predictive Control of Hammerstein Model Based on T-S Fuzzy Neural Network

针对化工生产过程中强非线性、大滞后、时变特点的复杂特性,提出了一种基于T-S模糊神经网络的Hammerstein模型动态矩阵预测控制方法。采用非线性控制分离策略,应用动态矩阵控制算法计算该模型动态线性部分的中间变量,作为T-S模糊神经网络的输入,进而通过T-S模糊神经网络逆映射出控制量,以实现基于T-S模糊神经网络的Hammerstein模型预测控制。pH中和过程的仿真控制实验表明,所提方法明显优于传统的PID控制方法,具有良好的设定值跟踪及抗干扰效果。

Aiming at the complex characteristics of strong non-linearity,large lag and time-varying in the chemical production process,a Hammerstein model dynamic matrix predictive control method based on T-S fuzzy neural network was put forward.By using a non-linear control separation strategy,the dynamic matrix control algorithm was used to calculate the intermediate variables of the dynamic linear part of the model,and the intermediate variables were used as the input of the T-S fuzzy neural network,then the T-S fuzzy neural network was used to reflect the control variables to complete the Hammerstein model predictive control based on T-S fuzzy neural network.Simulation control experiment results showed that that the proposed method was better than the traditional PID control.