Develop mathematical models
Formulate nonlinear dynamic models of continuous bioreactors with and without biomass recirculation.
Education / Ph.D. Research
Ph.D. Research Project · 2017–2021
Nonlinear Model Predictive Control of the Bioethanol Production Process by Continuous Fermentation
Research in chemical and biochemical reaction engineering, nonlinear process analysis and advanced process control.
The Challenge
Stable operation of a strongly nonlinear bioprocess.
Continuous bioethanol fermentation can provide efficient and continuous production, but the process exhibits complex nonlinear behaviour. Depending on operating conditions, the bioreactor may develop multiple steady states, metabolic transitions, oscillations or unstable operating regions.
The central challenge of this research was to understand these behaviours and develop a control strategy capable of guiding the process toward selected operating points while respecting process limitations.
Research Objectives
From process understanding to advanced control.
Analyse nonlinear behaviour
Identify steady states, stability boundaries, bifurcations and changes in dominant metabolic pathways.
Create process maps
Translate the nonlinear process analysis into practical maps showing stable and undesirable operating regions.
Design predictive control
Develop nonlinear model predictive controllers with integrated process constraints.
Research Workflow
A model-based approach to process control.
01
Continuous fermentation process
→
02
Nonlinear mathematical model
→
03
Steady-state and stability analysis
→
04
Process maps and constraints
→
05
Nonlinear predictive control
Methods
Tools and engineering methods
Key Outcomes
Main research results
- Mathematical models for two continuous bioreactor configurations.
- Identification of stable, unstable and oscillatory process regions.
- Process maps combining product concentration, metabolic behaviour and operating constraints.
- Numerical libraries describing the process boundaries.
- One- and two-dimensional nonlinear predictive controllers integrated with process constraints.
- Comparison of classical PI control with nonlinear model predictive control.
Why It Matters
The methods extend beyond bioethanol fermentation.
Although bioethanol production was the application studied in this project, the underlying approach is much broader. Mathematical modelling, nonlinear analysis, process maps, constraint handling and model predictive control can also be applied to other chemical and biochemical processes.
This research created the foundation for my later work in process simulation, digital engineering and model-based process development.
Scientific Output
Four peer-reviewed publications
The research was documented through a series of publications covering nonlinear process analysis, dynamic bifurcations, process maps and model-based control.