Linearization of the activated sludge model ASM1 for fast and reliable predictions

Area of Science:

• Environmental Engineering
• Biochemical Engineering
• Wastewater Treatment

Background:

• Activated sludge processes are crucial for wastewater treatment, particularly for nitrogen removal.
• The Activated Sludge Model No. 1 (ASM1) is a standard but complex model for simulating these processes.
• Increasingly stringent effluent standards necessitate efficient and accurate modeling tools.

Purpose of the Study:

• To develop a reduced-complexity version of the ASM1.
• To maintain high predictive value and biological interpretability of state variables.
• To significantly decrease computation time for practical applications.

Main Methods:

• Generated input/output data from the full ASM1 model.
• Linearized nonlinear kinetic terms of ASM1 and rewrote the model in state-space format.
• Identified parameters using generated data and employed Multi-Model interpolation to cover the operational range.

Main Results:

• Achieved a significantly less complex model compared to the full ASM1.
• The reduced model retains biological interpretability of all state variables.
• Demonstrated a substantial reduction in computation time while maintaining predictive accuracy.

Conclusions:

• The proposed reduced ASM1 is a valuable tool for risk assessment in wastewater treatment.
• The model's efficiency makes it suitable for on-line Model Predictive Control (MPC) strategies.
• This simplification facilitates the evaluation of treatment plants under stricter environmental regulations.