Soft-sensors Strategy for Estimating the Cell Biomass and the Nanobody-based Scorpion Antivenom Production: from Ml Algorithms Selection to Scale-down Bioprocess Application
Acosta-Pavas, Juan C.
Corrales, David Camilo
Alonso Villela, Susana M.
Georgakilas, Georgios
Bouhaouala-Zahard, Balkiss
Cescut, Julien
Fillaudeau, Luc
Aceves-Lara, Cesar Arturo
Pdf

How to Cite

Acosta-Pavas J.C., Corrales D.C., Alonso Villela S.M., Georgakilas G., Bouhaouala-Zahard B., Cescut J., Fillaudeau L., Aceves-Lara C.A., 2025, Soft-sensors Strategy for Estimating the Cell Biomass and the Nanobody-based Scorpion Antivenom Production: from Ml Algorithms Selection to Scale-down Bioprocess Application, Chemical Engineering Transactions, 119, 349-354.
Pdf

Abstract

The production of recombinant nanobodies (protein) in E. coli exhibits promising results for its application in serotherapy. However, the monitoring, control, and optimization of this bioprocess present limitations due to the measurements of critical variables. This study explores the potential of machine learning (ML) tools, such as soft sensors, to estimate offline variables such as protein and biomass concentration. CART Decision Tree (DT), Random Forest (RF), K-Nearest Neighbors (KNN), and Radial Basis Function (RBF) Support Vector Machine (SVM) learners were used to train and test the soft sensors. Eleven online variables were used as features. For the protein soft sensor, KNN and RBF SVM exhibit the best performance with a training and test results ?? 2 >0.99, ??????<3.578× 10 -3 , ????????<0.060, and ??????<0.052. In the biomass soft sensor, the training of all the learners showed good results ?? 2 >0.99, ??????<8.356× 10 -3 , ????????<0.091, and ??????<0.062.The soft sensors were evaluated in a scale-down process to estimate the biomass using only four variables. The results show the potential to use soft sensors for monitoring variables that are not possible to measure in real time.
Pdf