SMB色谱分离过程简化模型辨识

Reduced Model Identification of SMB Chromatographic Separation Process

在线性等温线条件下针对葡萄糖和果糖两组分SMB色谱分离过程,依据机理模型设计了恰当的仿真实验,基于MIMO-ARX模型类对其进行了简化模型辨识。根据模型验证结果,分析了不同样本数据和模型结构参数对模型精度的影响。获得的简化模型对机理模型的动态行为具有较好的逼近能力．对基于模型的控制器设计具有一定的指导意义。

MIMO-ARX model is used to model glucose and fructose simulated moving bed （SMB） chromatographic separation process under the linear isotherm. A proper simulation experiment based on the rigorous model is designed and the MIMO-ARX model is identified. According to model validation results, the influence on identification precision is analyzed by examining different sample data and parameters for effecting model structures. Good approximation ability of the reduced model for the rigorous model dynamic behavior is achieved. It is helpful to the model based controller design. Finally, limitations of the reduced model are presented and some prospects are also given.