A Hybrid Neural Network Algorithm for On-Line State Inference That Accounts for Differences in Inoculum of Cephalosporium acremonium in Fed-Batch Fermentors

Abstract

One serious difficulty in modeling a fermentative process is the forecasting of the duration of the lag phase. The usual approach to model biochemical reactors relies on first-principles, unstructured mathematical models. These models are not able to take into account changes in the process response caused by different incubation times or by repeated fedbatches. To overcome this problem, we have proposed a hybrid neural network algorithm. Feedforward neural networks were used to estimate rates of cell growth, substrate consumption, and product formation from on-line measurements during cephalosporin C production. These rates were included in the mass balance equations to estimate key process variables: concentrations of cells, substrate, and product. Data from fed-batch fermentation runs in a stirred aerated bioreactor employing the microorganism Cephalosporium acremonium ATCC 48272 were used. On-line measurements strongly related to the mass and activity of the cells used. They include carbon dioxide and oxygen concentrations in the exhausted gas. Good results were obtained using this approach.

Hybrid semi-parametric modeling in process systems engineering: Past, present and future

“…The application of hybrid semi-parametric models in form of a soft-senor is very attractive for monitoring and both parallel (Lee et al, 2005) and serial (Boareto et al, 2007;Gnoth et al, 2008;Henneke et al, 2005;James et al, 2002;Jenzsch et al, 2007;Psichogios & Ungar, 1992;Schubert et al, 1994a;Silva et al, 2000Silva et al, , 2001von Stosch et al, 2011b) hybrid semi-parametric models find application. It was shown that the performance of a model in which the states and parameters were estimated by Nonlinear Programming (NLP) optimization or Extended Kalman Filter (EKF) approaches was inferior to the performance of a model in which the variable parameters were estimated using neural networks (Psichogios & Ungar, 1992), namely a hybrid semi-parametric model.…”

“…Gupta et al (1999) applied two parallel ANNs, each of which inferring a variable value, in series with three other parallel ANNs, each of which estimating a quantity that enters as an input to the mechanistic model. In Gnoth et al (2008), Silva et al (2000Silva et al ( , 2001) the prediction of one central kinetic rate (usually the specific biomass growth rate) by a first ANN, was used as an input (beside others) to another ANN, which in turn predicts another rate, e.g. the product formation rate.…”

“…for the modeling of yeast fermentations (Beluhan & Beluhan, 2000;Boareto, De Souza, Valero, & Valdman, 2007;Eslamloueyan & Setoodeh, 2011;Mazutti et al, 2010;Peres et al, 2001;Saraceno et al, 2010;Saxen & Saxen, 1996;Schubert et al, 1994aSchubert et al, , 1994b, for modeling of fungi cultivations (Chen et al, 2000;Ignova et al, 2002;Preusting et al, 1996;Silva et al, 2000Silva et al, , 2001Thibault et al, 2000;van Can et al, 1997Wang, Chen, Liu, & Pan, 2010), for modeling of bacteria cultivations (Costa, Alves, Henriques, Filho, & Lima, 1998;Gnoth et al, 2008;Henneke, Hagedorn, Budman, & Legge, 2005;Henriques et al, 1999;James et al, 2002;Jenzsch, Gnoth, Kleinschmidt, Simutis, & Luebbert, 2007;Laursen et al, 2007;Roubos et al, 2000;Thibault et al, 2000;Tholudur & Ramirez, 1999;Zuo, Cheng, Wu, & Wu, 2006;Zuo & Wu, 2000), for modeling of mammalian cell cultivations (Dors et al, , 1996Teixeira et al, 2005;Vande Wouwer et al, 2004), for modeling of insect cell cultivations (Carinhas et al, 2011), for modeling of hybridoma cell cultivations (Fu & Barford, 1995a, 1995b or for modeling the counter-ion fluxes across an ion-exchange membrane in a membrane-supported biofilm reactor …”

Artificial neural networks to infer biomass and product concentration during the production of penicillin G acylase from Bacillus megaterium

“…Syu and Tsao7 employed a NN to model the batch growth of Klebsiella oxytoca . Silva et al 37, 38 used a hybrid NN model to infer the cellular growth and product formation during production of cephalosporin C. A comparative study was performed by James et al 9 for a poli hidroxi butirate (PHB) fed‐batch fermentation process. They compared a feedforward NN, a radial basis function neural network (RBFNN), and hybrid models using either feedforward NNs or RBFNNs.…”

On-line estimation of biomass and intracellular protein for recombinant Escherichia coli cultivated in batch and fed-batch modes

“…In this study, we introduce a training strategy, which is similar to applying a neural network design for a state estimator (Linko et al, 1997;Silva et al, 2001), to determine the tunable parameters. Fig.…”