Machine learning-based model construction and identification of dominant factor for simultaneous sulfide and nitrate...

Area of Science:

• Environmental Engineering
• Biotechnology
• Artificial Intelligence in Water Treatment

Background:

• Accurate water quality prediction is critical for effective simultaneous sulfide and nitrate removal (SSNR).
• Traditional models struggle with the complex microbial pathways in wastewater treatment.
• Back Propagation Neural Networks (BPNN) show promise for simulating complex biological processes.

Purpose of the Study:

• To develop and evaluate a generalized BPNN model for SSNR processes.
• To predict key water quality parameters including sulfide and nitrate removal, and byproduct formation.
• To identify critical factors influencing SSNR process performance.

Main Methods:

• Development of a generalized Back Propagation Neural Network (BPNN) model.
• Simulation and prediction of sulfide removal, nitrate removal, elemental sulfur production, and nitrogen gas production.
• Comparative analysis against traditional mathematical models.

Main Results:

• The BPNN model demonstrated strong predictive performance for effluent quality in SSNR.
• BPNN significantly outperformed traditional regression and ANOVA models.
• Crucial factors influencing process optimization and control were identified.

Conclusions:

• BPNN is a superior tool for simulating and predicting wastewater treatment processes like SSNR.
• Artificial intelligence integration enhances the efficiency of meeting water quality standards.
• The developed model provides a robust framework for SSNR process optimization.