Identification and optimal control of fructo-oligosaccharide production

Schorsch, Julien; Kinnaert, Michel; Fekih-Salem, Radhouane; Dewasme, Laurent; Castro, C. C.; Wouwer, Alain Vande

doi:10.1016/j.ifacol.2018.09.345

Cited by 3 publications

(7 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…1‐kestose was observed to be the major product among all different types of obtained FOS, as well as in our research. Furthermore, Schorsch et al 17 used the enzyme from Aureobasidium pullulans and the reaction mechanism was similar to our research. The FOS yield was 50%.…”

Section: Resultssupporting

confidence: 70%

Micro‐kinetic model of fructo‐oligosaccharide synthesis for prebiotic products

Pravilović

Janković

Veljković³

et al. 2023

AIChE Journal

View full text Add to dashboard Cite

A new micro‐kinetic model of the enzyme‐catalyzed synthesis of fructo‐oligosaccharides (FOS) was developed. A commercial enzyme mixture Pectinex® Ultra SP‐L derived from Aspergillus aculeatus was used. A variety of initial enzyme concentrations (1–5 vol%) and sucrose concentrations (400–600 g/L) were experimentally investigated and included in kinetic modeling. Several variations of kinetic mechanisms and corresponding models have been examined. A hybrid genetic algorithm was used to predict the kinetic parameters simultaneously for all experimental data. The best fitting model has been adopted, and with an average error of 13.34%, it describes the experimental data very well. The influence of initial concentrations on the conversion of sucrose and production of FOS is being carefully investigated. It was shown that the initial sucrose concentration significantly affects the highest level of FOS concentration, but the enzyme concentration controls the time at which maximum is reached as well as the rate of FOS decomposition.

show abstract

Section: Resultssupporting

confidence: 70%

Micro‐kinetic model of fructo‐oligosaccharide synthesis for prebiotic products

Pravilović

Janković

Veljković³

et al. 2023

AIChE Journal

View full text Add to dashboard Cite

show abstract

“…The parameter identification of model (27)(28)(29) with the 2B-2FB data configuration (see Table 5) show that all parameter uncertainties are very significantly reduced. The FIM is regular (see Table 2) with rank(FIM) = rank(S) = −numerical rank(S) = dim ( FIM ) = 23.…”

Section: Numerical Resultsmentioning

confidence: 99%

“…(i) and (iii) can be relatively well understood since the presence of multiplicative terms containing parameters may reinforce parameter dependency while (ii) and (iv) can be interpreted as a consequence of the cascade form of model (27-29) considering rates 3 and 4, as well as 5-7. The interpretation of model (27)(28)(29) would therefore be based on the following chemical/biological assumptions:…”

Section: Numerical Resultsmentioning

confidence: 99%

“…To investigate the impact of the parameter uncertainties on the model prediction, additional computations can be achieved, including the estimation of the covariance of the model prediction error, or alternatively, Monte Carlo analysis which considers all scenarios generated from the (27)(28)(29) represented as a heatmap. Parameters are numbered from 1 to 23 following the sequence of the first column of Vmt GF2 − 0.789 k 4 Vmh GF2 − 0.95 k 12 k 10 − 0.725 k 12 k 5 − 0.868 k 13 Vmt GF3 − 0.814 In the case of batch B1, Fig.…”

Section: Numerical Resultsmentioning

confidence: 99%

“…This bottom-up approach has to be distinguished from the top-down approach presented in [27], where a dynamic model is developed from scratch, starting from the experimental database, and applying principal component analysis and parameter estimation to develop a data-driven model. Without deflowering the content of this article and the present one, it can nevertheless be concluded that the two approaches lead to models whose predictive capacity is similar.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Sensitivity analysis and reduction of a dynamic model of a bioproduction of fructo-oligosaccharides

Fekih-Salem

Dewasme

Castro

et al. 2019

Bioprocess Biosyst Eng

Self Cite

View full text Add to dashboard Cite

Starting from a relatively detailed model of a bioprocess producing fructo-oligosaccharides, a set of experimental data collected in batch and fed-batch experiments is exploited to estimate the unknown model parameters. The original model includes the growth of the fungus Aureobasidium pullulans which produces the enzymes responsible for the hydrolysis and transfructosylation reactions, and as such contains 25 kinetic parameters and 16 pseudo-stoichiometric coefficients, which are not uniquely identifiable with the data at hand. The aim of this study is, therefore, to show how sensitivity analysis and quantitative indicators based on the Fisher information matrix can be used to reduce the detailed model to a practically identifiable model. Parametric sensitivity analysis can indeed be used to progressively simplify the model to a representation involving 15 kinetic parameters and 8 pseudo-stoichiometric coefficients. The reduced model provides satisfactory prediction and can be convincingly cross validated.

show abstract