Desalination, Год журнала: 2023, Номер 568, С. 117016 - 117016
Опубликована: Сен. 27, 2023
Язык: Английский
Desalination, Год журнала: 2023, Номер 568, С. 117016 - 117016
Опубликована: Сен. 27, 2023
Язык: Английский
Journal of Water Process Engineering, Год журнала: 2024, Номер 65, С. 105839 - 105839
Опубликована: Июль 30, 2024
Язык: Английский
Процитировано
4Desalination and Water Treatment, Год журнала: 2024, Номер 320, С. 100683 - 100683
Опубликована: Авг. 3, 2024
The integration of renewable energy sources with multi-energy systems present challenges and opportunities to enhance sustainability. Among these, solar stills have emerged as a solution for water desalination. With the advent expert system technologies, avenues are opened improving operational efficiency distillers. This paper presents an innovative approach utilizing correlation analysis, ReliefF feature selection, k-Nearest Neighbor (kNN) algorithm forecasting cumulative distillate output double slope still. analysis is based on 6-cases-based dataset, which includes variations in relative different operational-environmental conditions. Key features that significantly impact overall performance were identified manage distiller productivity. findings reveal maximum was 1610 ML/m2.day due incorporating reflective materials phase change (PCM) enhancing distillation rates. kNN model evaluated its R2, RMSE, CVRMSE, best models achieving scores 0.995, 0.0033, 0.1666, respectively. These metrics underscore effectiveness proposed machine learning predicting output, thereby enabling informed management processes. Combining technologies computational intelligence holds significant promise sustainable environmental management, study presented.
Язык: Английский
Процитировано
4Case Studies in Thermal Engineering, Год журнала: 2024, Номер 61, С. 104924 - 104924
Опубликована: Авг. 6, 2024
_Predicting solar energy is essential for efficient power system planning and the successful integration of renewable sources. This study aims to develop a framework evaluating various machine learning models feature selection strategies prediction. The research applies six models, i.e., linear regression (LR), random forest (RF), neural networks (NN), K-nearest neighbor (KNN), gradient boosting (GB), AdaBoost, datasets from 2019 2021 collected at Abiod Sid Cheikh station in southern Algeria. Various statistical indicators, including R2, RMSE, MAE, Adj-R, were analyzed assess model performance. analysis revealed that R2 values ranged 0.591 0.996 kW/m2, RMSE 0.510 1.78 MAE 0.357 0.856 kW/m2 across different models. KNN NN showed significant errors, while GB RF demonstrated strong accuracies (RMS = 0.9). AdaBoost LR excelled real-time short-term predictions, exhibiting an RMS 0.99. offers comprehensive evaluation method selecting most suitable findings can assist planners engineers choosing appropriate accurate prediction, thereby enhancing efficiency systems. Improved prediction contribute more reliable into grids, supporting transition cleaner sources reducing environmental impacts.
Язык: Английский
Процитировано
4Materials Today Chemistry, Год журнала: 2025, Номер 45, С. 102616 - 102616
Опубликована: Март 5, 2025
Язык: Английский
Процитировано
0Journal of Thermal Analysis and Calorimetry, Год журнала: 2025, Номер unknown
Опубликована: Апрель 28, 2025
Язык: Английский
Процитировано
0Separation and Purification Technology, Год журнала: 2025, Номер unknown, С. 133627 - 133627
Опубликована: Май 1, 2025
Язык: Английский
Процитировано
0Desalination, Год журнала: 2023, Номер 568, С. 117016 - 117016
Опубликована: Сен. 27, 2023
Язык: Английский
Процитировано
2