(164e) Optimization of Air Usage in Activated Sludge Processes

Wastewater treatment plants are known to utilize excessive amounts of air in activated sludge tanks in order to maintain sufficient growth of biomass and proper mixing. This approach requires intensive usage of energy for the compression of air, and this comprises up to 60% of the energy burden of an entire wastewater treatment plant (Daw et al., 2012; Spellman, 2008). Usually, the rate of aeration is kept high in activated sludge tanks so as to minimize the effects of abrupt changes in influent conditions, such as composition, temperature or volume (Owen, 1982).

In the present study, an advanced activated sludge tank model is employed to assess the feasibility of using decreased aeration rates. The activated sludge tank model is a discretized dynamic PFR model, which was calibrated and validated with the historical data from the Stickney Water Reclamation Plant in Illinois. Consisting of a train 48 CSTR tanks, the discretized model allows individual manipulation of airflow rates at each CSTR, and is an expedient conduit for performing optimization and process control studies.

Initially, constrained nonlinear optimization was performed on the model with the objective of maintaining at least 6 mg/L of dissolved oxygen and less than 0.1 mg/L of ammonia in the effluent. This was mainly done by choosing a number of pivot CSTRs along the entire activated sludge tank. Aeration rates in the pivot CSTRs were designated as optimization variables, and the aeration rates of the remaining CSTRs were calculated by linear interpolation using the neighboring pivot CSTR tanks. This allows for the utilization of different aeration profiles at different portions of the activated sludge tank. Open-loop simulations with the optimized aeration profiles obtained from this strategy have shown that the total rate of aeration can be reduced by up to 40%.

As a further step, an agent-based control framework was implemented with the objective of real-time control of the aeration rates. Using a multi-agent configuration and a specific goal for effluent quality, agents manipulated the aeration rates in each CSTR, with the objective of utilizing the minimum amount of air. Contrary to the limitations imposed by the computational burden of constrained nonlinear programming, the agents were able to decrease the aeration relatively quickly by up to 50%, and maintain it at varying influent conditions. Several dynamic scenarios, including events such as storms, were also simulated to the test the aeration profiles obtained from both approaches.

References
Daw, J., Hallett, K., DeWolfe, J., Venner, I., 2012. Energy Efficiency Strategies for Municipal Wastewater Treatment Facilities. National Renewable Energy Laboratory, Technical Report NREL/TP-7A30-53341.

Owen, W.F., 1982. Energy in wastewater treatment. Prentice-Hall, Inc., Englewood Cliffs, NJ.

Spellman, F., 2008. Handbook of water and wastewater treatment plant operations. CRC Press.