Modern day monitoring and control challenges outlined on an industrial-scale benchmark fermentation process

Goldrick, Stephen; Duran-Villalobos, Carlos A.; Jankauskas, Karolis; Lovett, David; Farid, Suzanne S.; Lennox, Barry

doi:10.1016/j.compchemeng.2019.05.037

Cited by 38 publications

(21 citation statements)

References 37 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Other solutions involve the use of Digital Twins to generate unlimited simulated data sets. This data can be used to develop and evaluate ML algorithms for process optimisation and speed up the readiness of these algorithms to implement once experimental data becomes available [ 41 ]. Other challenges have been raised by Mowbrey et al about the data produced in the biopharmaceutical sector.…”

Section: Resultsmentioning

confidence: 99%

A decade in review: use of data analytics within the biopharmaceutical sector

Banner

Alosert

Spencer

et al. 2021

Current Opinion in Chemical Engineering

Self Cite

View full text Add to dashboard Cite

Highlights Data analytics has increasing significantly in recent years in the biopharma sector. No clear trend observed between algorithm utilisation and data size. PLS was found to be most applied algorithm within the biopharmaceutical sector. Majority of the data analytics applications are focused on USP operations. Data analytics will play a key role as the sector transitions towards Industry 4.0.

show abstract

Section: Resultsmentioning

confidence: 99%

A decade in review: use of data analytics within the biopharmaceutical sector

Banner

Alosert

Spencer

et al. 2021

Current Opinion in Chemical Engineering

Self Cite

View full text Add to dashboard Cite

show abstract

“…IndPenSim can simulate modified designs of the reactor and inputs parameters. There are automatically controlled variables such as medium pH regulated by a feedback proportional integral derivative (PID) controller and manually controlled variables such as substrate flow rate controlled by operator recipe . Each batch by default is set to run for 230 h. Importantly for the study, the simulator allows for the simulations of 8 process faults that are shown in Table .…”

Section: Methodsmentioning

confidence: 99%

“…It was developed using a historical batch database of a 100,000 L fermentation of Penicillium chrysogenum to accurately simulate the process under different conditions. 18 By considering multiple available modes, delays in assay measurements, simulated manual operation of feeding strategies, inaccurate sensor readings, and random deviations across the penicillin development, the simulator aims to accurately generate data for data analytics studies. 18 Penicillin is typically modeled in two sequential phases: the growth phase and the secondary metabolic product production phase.…”

Section: Methodsmentioning

confidence: 99%

Design and Optimization of a Penicillin Fed-Batch Reactor Based on a Deep Learning Fault Detection and Diagnostic Model

Hematillake

Freethy

McGivern

et al. 2022

Ind. Eng. Chem. Res.

View full text Add to dashboard Cite

The application of a supervised deep convolutional autoencoder was tested against partial least-squares-discriminant analysis (PLS-DA) for fault detection and diagnosis in a penicillin fed-batch process. In silico data was generated with a comprehensive simulator (IndPenSim) of an industrial-scale penicillin fed-batch simulator of operation under normal batch conditions and 8 fault batch conditions. A composite face-centered design response surface was applied to optimize key bioreactor design parameters based on a profit function that was directly dependent on fault detection results. The application of PLS-DA and the NN modeling resulted in an average fault detection rate (FDR) across all faults and process parameter configurations of 72.5% and 95.9%, respectively. In classifying complex fault conditions, the deep learning model greatly surpassed the PLS-DA model, and this improvement translated into a 25.0% increase in realized profit with the supervised deep convolutional autoencoder when compared to PLS-DA monitoring.

show abstract

“…It has fault generation capability built-in so it can be used for testing fault detection algorithms and for process fault identification. (Goldrick et al, 2019; This paper uses a modified version of the simulator. Measurement noise and random variations were removed to move towards deterministic model.…”

Section: Simulation Modelmentioning

confidence: 99%

Adaptation framework for an industrial digital twin

Koistinen¹,

Ohenoja²,

Tomperi³

et al. 2021

Linköping Electronic Conference Proceedings

View full text Add to dashboard Cite

Digital twins for performance-oriented applications in industrial environments require systematic model maintenance. Model adaptation requires efficient optimization tools and continuous evaluation of measurement quality. The adaptation and model performance evaluation are based on the modeling error, making the adaptation prone also to the measurement errors. In this paper, a framework for combining model adaptation and measurement quality assurance are discussed. Two examples with simulated industrialscale biopharmaceutical penicillin fermentation are presented to illustrate the usability of the framework.

show abstract

Modern day monitoring and control challenges outlined on an industrial-scale benchmark fermentation process

Cited by 38 publications

References 37 publications

A decade in review: use of data analytics within the biopharmaceutical sector

A decade in review: use of data analytics within the biopharmaceutical sector

Design and Optimization of a Penicillin Fed-Batch Reactor Based on a Deep Learning Fault Detection and Diagnostic Model

Adaptation framework for an industrial digital twin

Contact Info

Product

Resources

About