Related Experiment Video
Updated: Oct 14, 2025
11:53
Spatial Multiobjective Optimization of Agricultural Conservation Practices using a SWAT Model and an Evolutionary Algorithm
Published on: December 9, 2012
13.1K
A new insight for real-time wastewater quality prediction using hybridized kernel-based extreme learning machines
Javad Alavi1, Ahmed A Ewees2, Sepideh Ansari1
1Department of Environmental Sciences and Engineering, Kheradgarayan Motahar Institute of Higher Education, Mashhad, Iran.
Environmental Science and Pollution Research International
|November 6, 2021
Summary
Predicting chemical oxygen demand (COD) in wastewater is crucial. Hybrid machine learning models, particularly the KELM-salp swarm algorithm, accurately forecast real-time COD, improving treatment plant management.
Area of Science:
- Environmental Engineering
- Machine Learning
- Wastewater Treatment
Background:
- Accurate prediction of inlet chemical oxygen demand (COD) is essential for effective wastewater treatment plant operation and management.
- The non-stationary and complex nature of inlet COD values presents a significant challenge for reliable forecasting.
Purpose of the Study:
- To develop and evaluate novel machine learning models for the accurate prediction of real-time inlet chemical oxygen demand (COD).
- To investigate the efficacy of hybridizing kernel-based extreme learning machines (KELMs) with intelligent optimization algorithms for COD prediction.
- To assess the impact of incorporating time-series learning and consumer behavior data on prediction accuracy.
Main Methods:
- Hybridization of Kernel-based Extreme Learning Machines (KELMs) with intelligent optimization algorithms, including the Salp Swarm Algorithm (SSA).
- Integration of time-series learning and consumer behavior patterns derived from water-use data (hourly/daily) as supplementary inputs.
- Systematic comparison of various hybrid KELM model configurations and input combinations.
Main Results:
- The optimal model configuration utilized up to 2-day lag values of COD along with other wastewater properties.
- The KELM-salp swarm algorithm (SSA) model demonstrated superior performance compared to other hybrid models.
- The best-performing KELM-SSA model achieved a minimum root mean square error (RMSE) of 0.058 and a mean absolute error (MAE) of 0.044.
Conclusions:
- Hybrid KELM models, particularly when optimized with the SSA, offer a reliable approach for real-time COD prediction in wastewater treatment.
- The integration of historical COD data and water-use patterns significantly enhances prediction accuracy.
- The developed models provide a valuable tool for improving the planning and management of wastewater treatment processes.
