Andrade Neto A S, Secchi A R, Souza M B, Barreto A G
PEQ/COPPE, Universidade Federal do Rio de Janeiro, Rio de Janeiro, Brazil.
PEQ/COPPE, Universidade Federal do Rio de Janeiro, Rio de Janeiro, Brazil.
J Chromatogr A. 2016 Oct 28;1470:42-49. doi: 10.1016/j.chroma.2016.09.070. Epub 2016 Sep 30.
An adaptive nonlinear model predictive control of a simulated moving bed unit for the enantioseparation of praziquantel is presented. A first principle model was applied at the proposed purity control scheme. The main concern about this kind of model in a control framework is in regard to the computational effort to solve it; however, a fast enough solution was achieved. In order to evaluate the controller's performance, several cases were simulated, including external pumps and switching valve malfunctions. The problem of plant-model mismatch was also investigated, and for that reason a parameter estimation step was introduced in the control strategy. In every studied scenario, the controller was able to maintain the purity levels at their set points, which were set to 99% and 98.6% for extract and raffinate, respectively. Additionally, fast responses and smooth actuation were achieved.
本文提出了一种用于吡喹酮对映体分离的模拟移动床装置的自适应非线性模型预测控制方法。在所提出的纯度控制方案中应用了第一性原理模型。在控制框架中，对于这类模型的主要关注点在于求解它所需的计算量；然而，还是获得了足够快的求解结果。为了评估控制器的性能，模拟了几种情况，包括外部泵和切换阀故障。还研究了装置 - 模型失配问题，因此在控制策略中引入了参数估计步骤。在每个研究场景中，控制器都能够将纯度水平维持在其设定点，提取物和萃余物的设定点分别设定为99%和98.6%。此外，还实现了快速响应和平滑驱动。
