State observers for a biological wastewater nitrogen removal process in a sequential batch reactor

Area of Science:

• Environmental biotechnology
• Biochemical engineering
• Wastewater treatment

Background:

• Biological nitrogen removal involves complex aerobic and anoxic microbial processes.
• Monitoring key state variables like microorganism and nutrient concentrations is challenging due to measurement difficulties.
• Reduced-order models offer an alternative for inferring these variables from secondary measurements (pH, redox).

Purpose of the Study:

• To investigate two reduced-order modeling approaches for biological nitrogen removal.
• To develop and compare state observers using Kalman filter structures.
• To assess the accuracy of inferred state variables against experimental data.

Main Methods:

• Developed a generic model (GM) based on the IAWQ No. 1 Model.
• Developed a specific model (SM) validated with bench-scale sequential batch reactor (SBR) data.
• Implemented Kalman filters to create a generic observer (GO) and a specific observer (SO).

Main Results:

• The generic model (GM) exhibited the poorest performance.
• The specific model (SM) showed some mismatch between model predictions and data.
• Both the generic observer (GO) and specific observer (SO) demonstrated very good agreement with experimental data.

Conclusions:

• Kalman filters enhance robustness against model errors in state variable inference.
• Reduced-order models combined with filters significantly improve process monitoring.
• This approach reduces modeling effort while ensuring adequate inference of critical process variables.