A model predictive optimal control system for the practical automatic start-up of anaerobic digesters

Area of Science:

• Environmental Engineering
• Biotechnology
• Process Control

Background:

• Optimal start-up of anaerobic digesters (AD) is challenging and costly, particularly regarding process monitoring.
• Existing methods often lack efficiency and cost-effectiveness in achieving stable operation.

Purpose of the Study:

• To develop and evaluate a non-linear model predictive control (NMPC) system for the optimal and cost-effective automatic start-up of AD treating soluble, non-recalcitrant substrates.
• To identify a minimum set of low-cost control variables (CVs) for effective process monitoring and control during start-up.

Main Methods:

• Development of a non-linear model predictive control (NMPC) system with two configurations.
• Utilized a simplified tailored AD model as the prediction model within the NMPC system.
• Evaluated the NMPC system using a virtual high-rate AD reactor simulation based on the Anaerobic Digestion Model No. 1 (ADM1).

Main Results:

• The NMPC system successfully achieved optimal start-up, reaching target methane production rates significantly faster (18-39 days) than manual strategies (70-75 days).
• Demonstrated superior set-point tracking error performance for key control variables.
• Identified two distinct, effective NMPC feeding strategies leading to successful reactor start-up without destabilization.

Conclusions:

• NMPC offers a robust and efficient solution for the automatic start-up of anaerobic digesters.
• The formulation of the NMPC objective function is critical for achieving successful and robust start-up.
• This approach reduces start-up duration and enhances process stability, offering significant cost and operational benefits.