Hybrid bounded error observers for uncertain bioreactor models

Area of Science:

• Bioreactor modeling
• Control systems engineering
• Observer design

Background:

• Bioreactor models often have uncertainties or partial knowledge.
• Existing observers like high-gain and asymptotic observers have limitations.
• High-gain observers offer adjustable convergence but need precise models.
• Asymptotic observers are robust to uncertainty but have fixed convergence rates.

Purpose of the Study:

• To develop novel bounded error observers for partially known bioreactor models.
• To combine the advantages of high-gain and asymptotic observers.
• To improve the convergence rate of state variable estimation in bioreactors.

Main Methods:

• Construction of hybrid bounded error observers.
• A two-step observer design approach.
• The first step utilizes high-gain observer principles for fast initial convergence.
• The second step switches to asymptotic observer principles for robust final convergence.
• Reconstruction of two state variables in the bioreactor model.

Main Results:

• The proposed hybrid observer achieves a faster convergence rate compared to classical asymptotic observers.
• The observer provides bounded errors in state variable estimation.
• Performance is demonstrated for bioreactor models with partially known dynamics.
• The observer effectively balances convergence speed and robustness to uncertainty.

Conclusions:

• Hybrid bounded error observers offer an effective solution for state estimation in partially known bioreactor systems.
• This approach enhances estimation performance by combining rapid convergence with robustness.
• The developed observer design provides a valuable tool for bioreactor control and monitoring.