多变量预测在芳烃抽提装置甲苯塔上的自适应控制研究

Study on Adaptive Control of MultiVariable Prediction on Toluene Column of Aromatics Extraction Unit

对芳烃抽提装置甲苯塔提出了自适应预测控制方案,建立了加权最小二乘预测模型并对其在线调整。由于芳烃抽提装置中的甲苯塔存在多变量、强耦合和非线性等特点,在操作与控制方面,将自适应预测控制方案和过去的预测控制相比较,并且相应的结合在线辨识,通过仿真研究发现自适应先进控制使甲苯塔回流量和甲苯塔温差处于稳定控制,这使塔内温度分布和物料分布都得到了平衡,保证了甲苯产品质量,降低了能耗。

An adaptive predictive control scheme is proposed for toluene tower of aromatics extraction plant.A weighted least square prediction model is established and adjusted on-line.Due to the multivariable,strong coupling and nonlinear characteristics of toluene tower in aromatics extraction plant,the adaptive predictive control scheme is compared with the previous predictive control scheme in the aspect of operation and control,and the on-line identi?cation is carried out accordingly.Through the simulation study,it is found that the self-adaptive advanced control makes the return?ow of toluene tower and the temperature difference of toluene tower in stable control,which makes the temperature distribution and material distribution in the tower balance,ensures the quality of toluene products,and reduces the energy consumption.