Nonlinear Wave Modeling And Control Design For High-purity Internal Thermally Coupled Distillation Column

The research on energy saving and control design of distillation has always been a frontier in the field of energy saving control research. Compared to the conventional distillation column, the internal thermally coupled distillation column (ITCDIC) can save as much as more than30%energy, and therefore its potential application is promising. However, the complex nonlinear dynamic characteristics, such as the strong coupling, the ill condition, the asymmetry dynamics and the inverse response, make the control design of the system difficult and become one of the main barriers for the commercialization of ITCDIC. The conventional control strategies cannot meet the tight control in ITCDIC, especially when the products reach a high purity. The complex dynamics mentioned above lead to a severe model/plant mismatch when an approximation model based on linearization or data driven is used, and the mismatch makes the model-based control strategies fail to achieve a satisfactory performance. This thesis starts from the wave theory in ITCDIC. The nonlinear wave model and the corresponding control strategies are then established to solve the challenge in ITCDIC.The main work and contributions of this thesis include:1. The wave theory is extended from conventional distillation to ITCDIC for the first time. The difference and difficulty of the extension are analyzed in detail, and then a natural wave velocity is proposed to describe the local wave characteristics in ITCDIC.2. In order to solve the mismatch between the model and the plant, a nonlinear wave model based on a shock wave velocity is presented. Then, a model predictive control based on this model is established.3. Based on the fully consideration of the wave shape variation, a instant wave velocity is proposed and the subsequent wave model highly enhances the accurary of the prediction. On the basis of the wave model, a careful investigation of the nonlinear dynamics in ITCDIC is carried out in the view of the wave theory. A generic model control is proposed based on the wave model using the instant velocity. Compared to a data-driven generic model control, the wave model based control exhibits a much better performance.4. A simplication work on the instant velocity is executed. The new simplified instant velocity leads to a wave model that is more accurate in the prediction of the product purity. Base on the simplified wave model, new generic model control strategies are presented. The self-adaptive GMC has achieved a much better performance compared to the previous controllers, and the non-adaptive GMC greatly cuts off the computation time at the cost of a very limited deterioration in performance.5. A generalized generic model control is proposed based on the simplified wave model. The newly developed cotroller not only performs well in a normal high purity as0.999, but also succeed in dealing with the very high purity control when the vapor product reach a purity of0.9999and0.99999.