|
Effluent quality prediction of one of the urban wastewater treatment plants using machine learning algorithms
|
Amin Bagheri    , Mohsen Sadani , Mohammadreza Karimi * |
| MPH student School of Public Health and Safety, Shahid Beheshti University of Medical Sciences, Tehran, Iran |
|
|
Abstract: (58 Views) |
Background: Mathematical and statistical simulators can significantly reduce the management costs of wastewater treatment systems. This research aimed to predict the effluent quality of an urban wastewater treatment plant in Tehran using machine learning algorithms from 2017 to 2021.
Methods: This descriptive-analytical study utilized monitoring data from the influent and effluent of the wastewater treatment plant, which were prepared for analysis. In the second stage, the data were refined, processed, and converted into dummy variables to facilitate entry into data mining algorithms. The Artificial Neural Network (ANN) algorithm and the M5 tree model were then evaluated to identify the best model for predicting the concentration of Chemical Oxygen Demand (COD) in the effluent. In this process, 70% of the data were allocated for training and 30% for validation using Python software. The best model was selected based on regression analysis, comparing the R² and RMSE indices.
Results: The findings indicated that the ANN model, with a coefficient of determination (R²) of 0.72, outperformed the M5 model, which had an R² of 0.68, in predicting the output COD concentration—an indicator of the treatment plant's efficiency. Additionally, regression analysis revealed that BOD₅ and TSS exhibited the highest correlation with CODout.
Conclusion: The results of the ANN and M5 models were within an acceptable range based on statistical indicators, demonstrating their potential for effectively estimating data in wastewater treatment plants.
|
|
| Keywords: Machine learning algorithms, artificial neural network, M5 model tree, regression, chemical oxygen demand |
|
|
Full-Text [PDF 914 kb]
(30 Downloads)
|
Type of Study: Research |
Subject:
Special
Received: 2024/08/28 | Accepted: 2024/09/29 | Published: 2024/11/13
|
|
|
|
|
|
|
| Send email to the article author |
|
|
|
| Add your comments about this article |
|
|
|
