Prediction of effluent quality in ICEAS-sequential batch reactor using feedforward artificial neural network

Area of Science:

• Environmental Engineering
• Water Treatment Technologies
• Artificial Intelligence in Environmental Science

Background:

• Municipal wastewater treatment is essential for safe discharge and reuse, especially given freshwater scarcity.
• Reclaimed water offers a sustainable solution for agricultural and industrial needs.
• Existing treatment process simulations can be complex and resource-intensive.

Purpose of the Study:

• To develop a predictive model for Jamnagar Municipal Corporation Sewage Treatment Plant (JMC-STP) effluent quality.
• To utilize a feedforward artificial neural network (FF-ANN) as an alternative to complex physical, chemical, and biological simulations.
• To predict key effluent parameters based on influent characteristics.

Main Methods:

• An investigation was conducted at the JMC-STP.
• A feedforward artificial neural network (FF-ANN) model was developed.
• The model predicts effluent quality parameters including pH, BOD, COD, TSS, TKN, AN, and TP based on influent data.

Main Results:

• The FF-ANN model demonstrated predictive capabilities for various effluent quality parameters.
• Performance metrics such as correlation coefficients (R_TRAINING, R_ALL), MAD, MSE, RMSE, and MAPE were evaluated.
• Simulation results were validated against measured values, confirming model accuracy.

Conclusions:

• The developed FF-ANN model serves as a valuable tool for optimizing wastewater treatment quality.
• The model enhances the performance and reliability of the JMC-STP.
• This approach offers an efficient alternative for monitoring and managing wastewater treatment processes.