Advanced

Open-Loop Optimal Control of Chromatographic Separation Processes

Sellberg, Anton LU (2018)
Abstract
Chromatography is an important separation process for several industries due to its ability to separate components with similar physical properties, which are difficult to separate otherwise. This method may also give high product purity, which is necessary for production of biopharmaceuticals and fine chemicals. However, conventional feedback control systems cannot be utilized to control most chromatographic separations due to the long time delays and complex system dynamics.

This work presents an open-loop optimal control framework that has been used to control chromatography processes. The framework is model-based and uses a simultaneous modelling and optimization approach. Thus, the model and the controls are fully... (More)
Chromatography is an important separation process for several industries due to its ability to separate components with similar physical properties, which are difficult to separate otherwise. This method may also give high product purity, which is necessary for production of biopharmaceuticals and fine chemicals. However, conventional feedback control systems cannot be utilized to control most chromatographic separations due to the long time delays and complex system dynamics.

This work presents an open-loop optimal control framework that has been used to control chromatography processes. The framework is model-based and uses a simultaneous modelling and optimization approach. Thus, the model and the controls are fully discretized and system is modelled and optimized simultaneously by solving a large scale optimization problem.

To improve the performance of the chromatographic separation, a zero-order hold elution trajectory was investigated as an alternative to step and linear gradients. The zero-order hold trajectory has been implemented in three different separation cases. First, a ternary batch separation was investigated. It was found that the zero-order hold trajectory resulted in an operating point that is more robust than the optimal operating point for step and linear gradient elution, leading to better overall performance of the process when subjected to disturbances. Secondly, the zero-order hold elution trajectory was applied to a multi-column counter-current solvent gradient purification process consisting of two columns. A single column analog model was developed to facilitate the implementation of the elution trajectories and the cyclic semi-continuous operation of the two columns was introduced as a constraint in the optimization to ensure cyclic steady state at the optimal operating point. Thirdly, the zero-order hold framework was extended to include the volumetric flow rates of the process. This enabled multi-objective optimization of the loading phase in capture chromatography. The variable flow rate loading scheme was 39\% more productive and resulted in 15\% higher process yield compared to the equally weighted single flow rate scheme.

To facilitate implementation of the zero-order hold trajectories and control of commercial chromatography systems a software interface was developed. The software interface is written in the general purpose programing language Python and allows bi-direction communication between the open-loop optimal control framework and the available sensors and actuators. (Less)
Please use this url to cite or link to this publication:
author
supervisor
opponent
  • Professor Kawajiri, Yoshiaki, Nagoya University, Nagoya, Japan
organization
publishing date
type
Thesis
publication status
published
subject
keywords
Chromatography, Modelling, Optimal Control, Parameter Estimation
pages
229 pages
publisher
Department of Chemical Engineering, Lund University
defense location
lecture hall K:B, Kemicentrum, Naturvetarvägen 14, Lund University, Faculty of Engineering LTH, Lund
defense date
2018-05-04 13:15
ISBN
978-91-7422-582-2
978-91-7422-583-9
language
English
LU publication?
yes
id
d7118c99-02a3-45ea-a689-132e0fc2537b
date added to LUP
2018-04-09 11:33:51
date last changed
2018-05-29 09:39:18
@phdthesis{d7118c99-02a3-45ea-a689-132e0fc2537b,
  abstract     = {Chromatography is an important separation process for several industries due to its ability to separate components with similar physical properties, which are difficult to separate otherwise. This method may also give high product purity, which is necessary for production of biopharmaceuticals and fine chemicals. However, conventional feedback control systems cannot be utilized to control most chromatographic separations due to the long time delays and complex system dynamics.  <br/><br/>This work presents an open-loop optimal control framework that has been used to control chromatography processes. The framework is model-based and uses a simultaneous modelling and optimization approach. Thus, the model and the controls are fully discretized and system is modelled and optimized simultaneously by solving a large scale optimization problem. <br/><br/>To improve the performance of the chromatographic separation, a zero-order hold elution trajectory was investigated as an alternative to step and linear gradients. The zero-order hold trajectory has been implemented in three different separation cases. First, a ternary batch separation was investigated. It was found that the zero-order hold trajectory resulted in an operating point that is more robust than the optimal operating point for step and linear gradient elution, leading to better overall performance of the process when subjected to disturbances. Secondly, the zero-order hold elution trajectory was applied to a multi-column counter-current solvent gradient purification process consisting of two columns. A single column analog model was developed to facilitate the implementation of the elution trajectories and the cyclic semi-continuous operation of the two columns was introduced as a constraint in the optimization to ensure cyclic steady state at the optimal operating point. Thirdly, the zero-order hold framework was extended to include the volumetric flow rates of the process. This enabled multi-objective optimization of the loading phase in capture chromatography. The variable flow rate loading scheme was 39\% more productive and resulted in 15\% higher process yield compared to the equally weighted single flow rate scheme.<br/><br/>To facilitate implementation of the zero-order hold trajectories and control of commercial chromatography systems a software interface was developed. The software interface is written in the general purpose programing language Python and allows bi-direction communication between the open-loop optimal control framework and the available sensors and actuators.},
  author       = {Sellberg, Anton},
  isbn         = {978-91-7422-582-2},
  keyword      = {Chromatography,Modelling,Optimal Control,Parameter Estimation},
  language     = {eng},
  month        = {05},
  pages        = {229},
  publisher    = {Department of Chemical Engineering, Lund University},
  school       = {Lund University},
  title        = {Open-Loop Optimal Control of Chromatographic Separation Processes},
  year         = {2018},
}