Predicting coagulation-flocculation process for turbidity removal from water using graphene oxide: a comparative...

Area of Science:

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

Background:

• Turbidity removal is crucial for water purification.
• Graphene oxide (GO) shows promise as an adsorbent for water contaminants.
• Data-driven modeling can optimize water treatment processes.

Purpose of the Study:

• To model and predict turbidity removal efficiency using graphene oxide (GO).
• To compare the performance of artificial intelligence (AI) models: artificial neural network (ANN), support vector regression (SVR), and adaptive neuro-fuzzy inference system (ANFIS).
• To evaluate AI models against response surface methodology (RSM).

Main Methods:

• Developed AI models (ANN, SVR, ANFIS) using pH, GO dosage, and initial turbidity as inputs.
• Selected input variables using the partial mutual information (PIM) algorithm.
• Assessed model accuracy using statistical metrics: MSE, RMSE, MAE, and R².

Main Results:

• ANN demonstrated the highest accuracy with R² of 0.949 for validation data and minimal errors.
• ANN accurately predicted 76.1% of data points with <10% relative error.
• SVR showed the weakest performance, while ANFIS and RSM had comparable results.

Conclusions:

• Artificial neural network (ANN) is the most effective technique for modeling turbidity removal by GO.
• Response surface methodology (RSM) is valuable for understanding parameter interactions before ANN modeling.
• AI, particularly ANN, offers a powerful approach for optimizing water treatment processes.