Modeling of activated sludge process using artificial neuro-fuzzy-inference system (ANFIS)
Radia Maachou, Abdelouahab Lefkir, ALI Khouider, A. Bermad · Desalination and Water Treatment · 2015
The paper describes the application of a neuro-fuzzy system in order to minimize the energy consumption on controlling the nitrate production in the wastewater treatment plant by activated sludge process. Neuro-fuzzy models are based on the extraction of knowledge from data collected upstream and downstream of a treatment plant. The historical values of the observed yields associated with the energy consumed during the study period enable the prediction of the energy needed for a validation period. The energy is controlled by the excess nitrates produced, which can be a symptom of over aeration. However, the simulation data are divided into two samples (data filtered and data unfiltered of nitrate). The input parameters used in this study includes the removal yields of organic pollutants parameters and energy consumption as a decision parameter with respect to the discharge standards. The predictive power of energy shows the feasibility and robustness of the simulation approach with a filtered data.