Status : Verified
| Personal Name | Moreno, Anne Jadeite G. |
|---|---|
| Resource Title | Data-Driven Prediction and Characterization of Influent Wastewater Using Machine Learning |
| Date Issued | 07 July 2024 |
| Abstract | Projected water scarcity in the Philippines necessitates exploring wastewater resource recovery, but untreated wastewater discharge remains a major challenge, especially in urban areas like Metro Manila. Sewage treatment plants (STPs) play a crucial role in wastewater recovery, with Sequencing Batch Reactor (SBR) systems being particularly effective for domestic wastewater due to their flexibility, cost-effectiveness, and ability to handle varying volumes and compositions. However, the implementation of General Effluent Discharge standards by the Department of Environment and Natural Resources (DENR) has made compliance difficult for various stakeholders, including Local Government Units (LGUs) and private business owners. The stringent GES values, along with high investment and operational costs, have constrained their ability to meet the standards, highlighting the need for data-driven models to optimize treatment processes. In the present study, it evaluates the effectiveness of various machine learning models in optimizing the performance of the Sequencing Batch Reactor (SBR) system at Condominium A. The Decision Tree Classifier identified Dissolved Oxygen (DO) as the most critical parameter, with a recommended threshold of 1.67 mg/L ensuring 82% compliance with DAO 2021-19 effluent standards. Correlation analysis using the Spearman method underscored the significant relationships between DO and other key parameters, such as nitrate, phosphate, ammonia, COD, and BOD, emphasizing the importance of DO management in wastewater treatment. The study also predicts the effluent reduction in terms of Biochemical Oxygen Demand (BODeff), Chemical Oxygen Demand (CODeff), Total Suspended Solids (TSSeff), Ammonia (NH3Eff), Nitrate (NO3eff), Phosphate (PO4eff) and pH (pHeff). Among the seven machine learning models tested— Kernel Ridge Regression (KRR), K-Nearest Neighbors (KNN), Support Vector Regression (SVR), Multi-Layer Perceptron (MLP), Gradient Boosting Regression (GB |
| Degree Course | MS Environmental Engineering |
| Language | English |
| Keyword | Decision Trees Classifier, Spearman Method, Dissolved Oxygen, Machine Learning, Multi-Layer Perceptron (MLP), BOD, Sequence Batch Reactor (SBR), Wastewater |
| Material Type | Thesis/Dissertation |
Preliminary Pages
643.13 Kb
Category : F - Regular work, i.e., it has no patentable invention or creation, the author does not wish for personal publication, there is no confidential information.
Access Permission : Open Access