On-line optimization of four-zone simulated moving bed chromatography using an Equilibrium-Dispersion Model: II. Experimental validation

Lee, Ju Weon; Kienle, Achim; Seidel‐Morgenstern, Andreas

doi:10.1016/j.ces.2020.115808

Cited by 10 publications

(3 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…An extensive experimental validation of the on-line optimization concept is presented in a second part considering the separation of two bicalutamide enantiomers as a case study [21].…”

Section: Discussionmentioning

confidence: 99%

On-line optimization of four-zone simulated moving bed chromatography using an equilibrium-dispersion model: I. Simulation study

Lee

Kienle

Seidel‐Morgenstern

2020

Chemical Engineering Science

Self Cite

View full text Add to dashboard Cite

“…An extensive experimental validation of the on-line optimization concept is presented in a second part considering the separation of two bicalutamide enantiomers as a case study [21].…”

Section: Discussionmentioning

confidence: 99%

On-line optimization of four-zone simulated moving bed chromatography using an equilibrium-dispersion model: I. Simulation study

Lee

Kienle

Seidel‐Morgenstern

2020

Chemical Engineering Science

Self Cite

View full text Add to dashboard Cite

“…TMB parameters will be estimated globally by a Software-In-the-Loop (SIL) scheme solved by Particle Swarm Optimization (PSO) in Matlab connected to the SMB phenomenological models in gProms, each isotherm at a time. Lee; Kienle; Seidel-Morgenstern (2020a, 2020b) and Lee; Seidel-Morgenstern (2019) applied an equilibrium dispersion cell model with nonlinear adsorption isotherms based on online optimization that does not need to solve the PDAE, but nonlinear equations “cell-by-cell”. − On the other hand, here, parameters will be estimated globally by the SIL scheme solved by orthogonal colocation in 20 intervals, using second-order polynomials with two collocation points in a 1 Tb Ryzen 7 laptop. This procedure is performed four times, one for each ξ k .…”

Section: Methodsmentioning

confidence: 99%

Strategies for Simulated Moving Bed Model Parameter Estimation Based on Minimal System Minimal Knowledge: Adsorption Isotherm Equation Screening and Estimability Analysis

Santos,

Rebello,

Prudente

et al. 2023

Ind. Eng. Chem. Res.

View full text Add to dashboard Cite

The simulated moving bed (SMB) is a widely recognized technique for the resolution of several high-value-added compounds, and its mathematical modeling plays a crucial role in predicting its behavior. However, modeling can be challenging as it needs parameter estimation and topology definition, especially for such complex systems. SMB modeling might incorporate different adsorption isotherm equations (AIEs), which might better represent the behavior of a certain mixture to be separated. This work proposes two methodologies for SMB modeling and AIE selection relying on minimal system knowledge that can provide a representative model by a global parameter estimation instead of the traditional "Step by Step" approach. The first methodology proposes an architecture for global parameter estimation that encompasses the screening of the most suitable AIE in the lower level and the definition of the minimum number of experiments necessary for model identification in the upper level. In the second methodology, an extended generic isotherm, a combination of AIE, is formulated; and an estimability analysis is carried out to determine the estimable parameters. Both methodologies apply Latin Hypercube Sampling (LHS) for the design of experiments and emulation of a real unit by incorporating noise in the experimental data under a Gaussian error distribution. Moreover, uncertainty evaluation and data reconciliation are performed to enhance the robustness, reliability, and precision of the proposed approaches. The methodologies are validated by considering experimental data from the literature, apart from the LHS-generated samples, and are subject to cross-validations. Finally, the two methodologies are compared, with the first methodology demonstrating a superior performance for the case study. This work contributes to the advancement of SMB modeling, providing valuable tools for parameter estimation and AIE selection.

show abstract

“…This is especially challenging in the optimization process where the simulations of SMB processes under various operation conditions must be repeated hundreds and even thousands of times. Therefore, the optimization method based on mechanistic model is not suitable for online control purpose because the online optimization results must be obtained rapidly so that the operating conditions can be adjusted quickly enough 25‐27 . To this end, substituting the mechanistic model with machine learning model may be a solution because machine learning can reduce the simulation time greatly.…”

Section: Introductionmentioning

confidence: 99%

Using a machine learning model for the optimal design of simulated moving bed processes and its application to separate rebaudioside A and stevioside

Yao

Zheng

Chen

et al. 2021

J of Chemical Tech & Biotech

View full text Add to dashboard Cite

BACKGROUND Solving a mechanistic model of simulated moving bed (SMB) is extremely time consuming, therefore, it is inefficient for online control using the optimization method based on such a mechanistic model. To this end, the machine learning model can be regarded as a potential substitution to accelerate the optimization process. RESULTS Several machine learning algorithms were applied to predict the product purities under various operation conditions using two SMB processes as case studies: i.e., a sugar separation of rebaudioside A and stevioside, as well as an enantioseparation of 1,1′‐bi‐2‐naphthol racemate. The results indicate that the random forest (RF) model and the deep neural network (DNN) model provide a satisfactory accuracy with a mean absolute error (MAE) lower than 0.19% (RF) and 0.08% (DNN), respectively. During the optimization process to maximize the feed flowrate under specific purity requirements, among the two selected models, DNN model showed a better generalization ability than RF model and gave a feed flowrate 10% higher than the highest value in the training dataset, which was consistent with the result obtained by using the mechanistic model. The optimized operation conditions for sugar separation were verified experimentally and the achieved purities of rebaudioside A and stevioside were 99.2% and 98.8%, respectively. CONCLUSION The DNN model was successfully used to substitute the mechanistic model to reach a rapid optimization of SMB processes with an improved efficiency of 103 ~ 104 times. © 2021 Society of Chemical Industry (SCI).

show abstract

On-line optimization of four-zone simulated moving bed chromatography using an Equilibrium-Dispersion Model: II. Experimental validation

Cited by 10 publications

References 16 publications

On-line optimization of four-zone simulated moving bed chromatography using an equilibrium-dispersion model: I. Simulation study

On-line optimization of four-zone simulated moving bed chromatography using an equilibrium-dispersion model: I. Simulation study

Strategies for Simulated Moving Bed Model Parameter Estimation Based on Minimal System Minimal Knowledge: Adsorption Isotherm Equation Screening and Estimability Analysis

Using a machine learning model for the optimal design of simulated moving bed processes and its application to separate rebaudioside A and stevioside

Contact Info

Product

Resources

About