Contact: E-mail
Project title: Dynamic model based process performance monitoring for life sciences
Host organization: USTRATH
Supervisors: Dr. Alison Nordon (USTRATH), Prof. Julian Morris (USTRATH), Prof. Krist Gernaey (DTU)
Duration: 36 months
Objectives:
- Develop dynamic multivariate statistical process performance monitoring tools based on (i) Canonical Variate Analysis state space models and (ii) Gaussian Process Regression models for nonlinear processes. The ESR will work with DTU (Prof. Krist Gernaey) on application of the monitoring methods for selected fermentation processes. The work will include (a) the extension of CVA approaches to included process data-driven confidence bounds allowing enhanced confidence in bio-process malfunction diagnosis rather than just ‘detection’, (b) the development of a Gaussian Process Regression approach for nonlinear processes.
- Demonstrate the effectiveness of the developed tools on fermentation processes. Particular care will be taken to be cognizant of ICH GAMP 5 and CFR21 Part 11 operational compliance in the software developments.
Expected results:
- New dynamic statistical bio-process performance monitoring methodologies and algorithms.
- Demonstration of enhanced process malfunction diagnosis particularly for subtle and incipient process disturbances.
- Demonstration of the tools developed on fermentation process for production of industrial enzymes.
Planned secondments:
- DTU, M18, 2 months, on-site on industrial fermentation processes
- BAYER, M36, 2 months, exposure to fermentation technology and processes