Advances in Engineering Advances in Engineering features breaking research judged by Advances in Engineering advisory team to be of key importance in the Engineering field. Papers are selected from over 10,000 published each week from most peer reviewed journals.
Moritz von Stosch, Rui Oliveira, Joana Peres, Sebastião Feyo de Azevedo
Journal of Process Control, Volume 22, Issue 7, August 2012
Abstract
A general hybrid semi-parametric process control framework is proposed in this study. The motivation was the integration of different levels of knowledge systems into a general hybrid semi-parametric control structure, of which the general linear controller or the PID controller are, for instance, particular cases.
Several hybrid semi-parametric control structure variants and tuning methods are benchmarked in relation to a simulated bioprocess control problem, namely closed-loop control of the biomass concentration through manipulation of the substrate feeding rate, coupled with the closed-loop control of the dissolved oxygen concentration through the stirring velocity. The results demonstrate that (i) due to the hybrid approach the control loop can be closed without any additional identification experiments; (ii) the incorporation of different types of knowledge can enhance the controller performance, when compared to structures without a priori knowledge; (iii) knowledge incorporation seems to facilitate the tuning of the controller; and (iv) the control action can be analyzed in relation to structural information incorporated into the hybrid controller.
A general hybrid semi-parametric process control framework
Moritz von Stoscha, Rui Oliveirab, Joana Peresa, Sebastião Feyo de Azevedoa
a LEPAE, Departamento de Engenharia Quimica, Faculdade de Engenharia, Universidade do Porto, Rua Dr. Roberto Frias s/n, 4200-465 Porto, Portugal
b REQUIMTE, Departamento de Quimica, Faculdade de Ciências e Tecnologia, Universidade Nova de Lisboa, 2829-516 Caparica, Portugal
A general hybrid semi-parametric process control framework is proposed in this study, that seeks to overcome the structural limitations of PID controllers by using Artificial Neural Networks (ANNs), while limiting the typical shortcomings of ANNs through the integration of other available process knowledge. This control structure is general and customizable to particular applications, the General Linear Controllers or the classical PID controller constituting particular cases of the proposed structure.
Controller design and tuning is carried out at the aid of hybrid semi-parametric process models. Such models provide excellent prediction capabilities at low development costs wherefore they are ideal for the tuning of the controller parameters with respect to desired specifications. In addition, hybrid process models enable efficient nonlinear control designs, such as Model Reference Control (MRC) or Generic Model Control (GMC).
In this study two parameter identification methods are proposed. The first, having as feature that the parameters of the process model and of the controller are simultaneously identified from process data. The second, permitting the tuning of the controller parameters beyond the characteristics that can be learned from the process data.
A simulated control problem is taken as case study, where the control of biomass through the manipulation of the substrate feeding rate and the control of dissolved oxygen concentration through the manipulation of the stirring velocity are considered. The simulated process employed is characterized by highly non-linear kinetics and different time scales of the inherent dynamics, which can result in bang-bang behavior. Furthermore the biomass and the dissolved oxygen control loops are coupled. Several possible combinations of hybrid control structures were investigated for this case study and the following points can be highlighted:
i) Without any additional identification experiments the control loop can be closed, thanks to the hybrid approach;
ii) The controller performance is enhanced through the incorporation of different types of knowledge, when compared to structures without a priori knowledge;
iii) Controller tuning seems to be facilitated by knowledge incorporation; and
iv) The control action becomes interpretable in relation to structural information incorporated into the hybrid controller.
Go to Journal
