Smail Es Sellami
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Smail Es Sellami has a total of 8 co-authors. Here are the top co-authors by number of shared publications:
- Mohammed EL Hachoumi
Laboratory of Process Engineering and Environment (LGPE), Faculty of Sciences and Techniques of Mohammedia (FSTM), Hassan II University of Casablanca, Mohammedia 28806, Morocco
4 shared publications - Abdeslam Taleb
Laboratory of Process Engineering and Environment (LGPE), Faculty of Sciences and Techniques of Mohammedia (FSTM), Hassan II University of Casablanca, Mohammedia 28806, Morocco
3 shared publications - Abdelkader Es Sellami
Independent Researcher, Mohammedia 28806, Morocco
2 shared publications - Taleb Abdeslam
Laboratory of Process Engineering and Environment (LGPE), Faculty of Science and Technology of Mohammedia (FSTM), Hassan II University of Casablanca, Mohammedia 28806, Morocco
2 shared publications - Badr El Fathi
Laboratory of Process Engineering and Environment (LGPE), Faculty of Science and Technology of Mohammedia (FSTM), Hassan II University of Casablanca, Mohammedia 28806, Morocco
1 shared publication - Mohammed Karim Benhachmi
LGPE,fstm,MOHAMMEDIA,Morocco, Morocco
1 shared publication
Latest Publications
- Open Access
- 9 October 2026
- Published by MDPI AG in Preprints.org
Physics-Informed Graph Neural Networks for Leak Diagnosis Under Concurrent Leakage
Leakage diagnosis in water distribution networks is constrained by sparse instrumentation, uncertain models, and leak concurrency. We present a physics-informed spatio-temporal graph neural network evaluated end-to-end on the open Battle of the Leakage Detection and Isolation Methods (BattLeDIM) and L-Town benchmark (782 junctions, 905 pipes, 33 pressure sensors). Message passing operates on a conductance-weighted hydraulic Laplacian, trained by self-supervised sensor masking to reconstruct nodal pressure fields and latent pipe flows, embedding Hazen-Williams energy and mass-balance residuals in the loss under learned uncertainty weighting. Pressure at withheld sensors is reconstructed to 0.29 meters. A seeded ablation shows physical constraints degrade reconstruction accuracy by 0.10 meters of head while reducing nodal continuity violations by over five orders of magnitude, rendering the inferred state hydraulically admissible. Residuals feed two complementary detectors. A level-based statistic saturates under concurrent leakage; pairing it with a two-window onset contrast raises detection from 13 to 22 of 23 ground-truth leaks (95.7 %, 95 % confidence interval 79.0–99.2 %), with median detection times of 29 hours for abrupt and 74 hours for incipient events, releasing 96 % of annual leaked volume after detection. Localization transfers poorly from simulation, with a median error of 857 meters and only three leaks meeting the joint 300 meters criterion due to signal superposition under concurrency.
- 9 October 2026
- Published by MDPI AG in Preprints.org
Open AccessPhysics-Informed Graph Neural Networks for Leak Diagnosis Under Concurrent Leakage
Leakage diagnosis in water distribution networks is constrained by sparse instrumentation, uncertain models, and leak concurrency. We present a physics-informed spatio-temporal graph neural network evaluated end-to-end on the open Battle of the Leakage Detection and Isolation Methods (BattLeDIM) and L-Town benchmark (782 junctions, 905 pipes, 33 pressure sensors). Message passing operates on a conductance-weighted hydraulic Laplacian, trained by self-supervised sensor masking to reconstruct nodal pressure fields and latent pipe flows, embedding Hazen-Williams energy and mass-balance residuals in the loss under learned uncertainty weighting. Pressure at withheld sensors is reconstructed to 0.29 meters. A seeded ablation shows physical constraints degrade reconstruction accuracy by 0.10 meters of head while reducing nodal continuity violations by over five orders of magnitude, rendering the inferred state hydraulically admissible. Residuals feed two complementary detectors. A level-based statistic saturates under concurrent leakage; pairing it with a two-window onset contrast raises detection from 13 to 22 of 23 ground-truth leaks (95.7 %, 95 % confidence interval 79.0–99.2 %), with median detection times of 29 hours for abrupt and 74 hours for incipient events, releasing 96 % of annual leaked volume after detection. Localization transfers poorly from simulation, with a median error of 857 meters and only three leaks meeting the joint 300 meters criterion due to signal superposition under concurrency.
- Open Access
Drift-Aware Digital-Twin Readiness for Two-Stage Reverse Osmosis: Plant-Scale Time-Aware Machine Learning and Operating-Regime Validation
Reverse osmosis (RO) digital twins require predictive models whose accuracy remains credible as plant operating conditions change. We evaluated the readiness of predictive models for a full-scale, two-stage RO plant using operational spreadsheets and temporally structured validation. An audit of 70 legacy worksheets identified heterogeneous layouts. A normalized source provided 476 train-day records from 17 operating months between June 2022 and November 2023, while 55 records from April and May 2024 formed an independent later-period holdout. Four supervised regression algorithms (Ridge Regression, Random Forest, Extra Trees, and Gradient Boosting) were evaluated for second-stage permeate flow and conductivity. Under a random split, Random Forest achieved R2 values of 0.876 and 0.833, with root mean square errors of 8.48 m3/h and 63.26 µS/cm, respectively. Chronological validation reduced R2 to 0.042 and 0.118. On the later holdout, R2 fell to −1.173 and −0.918, with corresponding root mean square errors of 11.14 m3/h and 218.65 µS/cm. Rolling-origin tests revealed substantial variability among successive periods. Distributional diagnostics identified changes in feed pressure, first-stage permeate conductivity, second-stage flow, and second-stage conductivity; the available data could not establish their physical causes. These findings show that performance under random splitting is insufficient to establish readiness for future use. A membrane-process digital twin requires traceable data, chronological evaluation, operating-regime surveillance, controlled model updates, and human-supervised advisory operation.
Open AccessDrift-Aware Digital-Twin Readiness for Two-Stage Reverse Osmosis: Plant-Scale Time-Aware Machine Learning and Operating-Regime Validation
Reverse osmosis (RO) digital twins require predictive models whose accuracy remains credible as plant operating conditions change. We evaluated the readiness of predictive models for a full-scale, two-stage RO plant using operational spreadsheets and temporally structured validation. An audit of 70 legacy worksheets identified heterogeneous layouts. A normalized source provided 476 train-day records from 17 operating months between June 2022 and November 2023, while 55 records from April and May 2024 formed an independent later-period holdout. Four supervised regression algorithms (Ridge Regression, Random Forest, Extra Trees, and Gradient Boosting) were evaluated for second-stage permeate flow and conductivity. Under a random split, Random Forest achieved R2 values of 0.876 and 0.833, with root mean square errors of 8.48 m3/h and 63.26 µS/cm, respectively. Chronological validation reduced R2 to 0.042 and 0.118. On the later holdout, R2 fell to −1.173 and −0.918, with corresponding root mean square errors of 11.14 m3/h and 218.65 µS/cm. Rolling-origin tests revealed substantial variability among successive periods. Distributional diagnostics identified changes in feed pressure, first-stage permeate conductivity, second-stage flow, and second-stage conductivity; the available data could not establish their physical causes. These findings show that performance under random splitting is insufficient to establish readiness for future use. A membrane-process digital twin requires traceable data, chronological evaluation, operating-regime surveillance, controlled model updates, and human-supervised advisory operation.
- Open Access
Operational Stability and Treatment Efficiency of a Reverse Osmosis Plant Treating Highly Mineralized Water: Results of a Four-Year Follow-Up Study
This study evaluates the long-term operational performance of a full-scale reverse osmosis desalination plant treating highly mineralized water under real industrial operating conditions. Unlike conventional laboratory or pilot-scale investigations, the analysis is based on four years of operational data (2020–2024), providing a representative assessment of membrane system behavior under variable environmental and hydraulic conditions. Statistical analyses including descriptive statistics, time-series analysis, Pearson correlation, and inter-annual comparisons were performed using Python-based data-processing tools. An Operational Stability Index was also introduced to quantify process consistency over time. The results demonstrate stable desalination performance, with salt rejection exceeding 90%. Operational monitoring indicated effective pretreatment conditions, low SDI values, and controlled membrane fouling under long-term operation. Differential pressure evolution and cleaning-in-place performance further confirmed the effectiveness of the implemented fouling mitigation strategy. Membrane autopsy investigations revealed the coexistence of biological fouling, silica scaling, and localized oxidative degradation, highlighting the importance of integrated pretreatment and operational control. The findings demonstrate the value of long-term industrial monitoring for understanding membrane performance, fouling behavior, and operational stability in full-scale RO systems. This approach supports data-driven operational optimization and may contribute to the future development of predictive monitoring and intelligent desalination-management frameworks as a Digital Twin.
Open AccessOperational Stability and Treatment Efficiency of a Reverse Osmosis Plant Treating Highly Mineralized Water: Results of a Four-Year Follow-Up Study
This study evaluates the long-term operational performance of a full-scale reverse osmosis desalination plant treating highly mineralized water under real industrial operating conditions. Unlike conventional laboratory or pilot-scale investigations, the analysis is based on four years of operational data (2020–2024), providing a representative assessment of membrane system behavior under variable environmental and hydraulic conditions. Statistical analyses including descriptive statistics, time-series analysis, Pearson correlation, and inter-annual comparisons were performed using Python-based data-processing tools. An Operational Stability Index was also introduced to quantify process consistency over time. The results demonstrate stable desalination performance, with salt rejection exceeding 90%. Operational monitoring indicated effective pretreatment conditions, low SDI values, and controlled membrane fouling under long-term operation. Differential pressure evolution and cleaning-in-place performance further confirmed the effectiveness of the implemented fouling mitigation strategy. Membrane autopsy investigations revealed the coexistence of biological fouling, silica scaling, and localized oxidative degradation, highlighting the importance of integrated pretreatment and operational control. The findings demonstrate the value of long-term industrial monitoring for understanding membrane performance, fouling behavior, and operational stability in full-scale RO systems. This approach supports data-driven operational optimization and may contribute to the future development of predictive monitoring and intelligent desalination-management frameworks as a Digital Twin.
- 10 December 2025
- Published by Trans Tech Publications, Ltd. in Materials Science Forum
- Vol. 1172, 85-98
Optimization of the Osmosis Process for the Potabilization of Wadi Oum Er-Rbia Water: Analysis of Hydraulic and Quality Parameters
The main objective of this study is to optimize the reverse osmosis process in order to ensure the potabilization of water from the Oued Oum Er-Rbia, by determining the most influential parameters. To the best of our knowledge, this is the first study to apply daily PCA-based monitoring on Oued Oum Er-Rbia’s raw water to optimize membrane operation under Moroccan field conditions.To better understand the interactions between quality and hydraulic parameters influencing membrane performance, data were collected from Oued Oum Er-Rbia over multiple seasons. The parameters monitored included turbidity, salinity, temperature, and Silt Density Index (SDI), all known to affect fouling and pretreatment requirements.
- 10 December 2025
- Published by Trans Tech Publications, Ltd. in Materials Science Forum
- Vol. 1172, 85-98
Optimization of the Osmosis Process for the Potabilization of Wadi Oum Er-Rbia Water: Analysis of Hydraulic and Quality Parameters
The main objective of this study is to optimize the reverse osmosis process in order to ensure the potabilization of water from the Oued Oum Er-Rbia, by determining the most influential parameters. To the best of our knowledge, this is the first study to apply daily PCA-based monitoring on Oued Oum Er-Rbia’s raw water to optimize membrane operation under Moroccan field conditions.To better understand the interactions between quality and hydraulic parameters influencing membrane performance, data were collected from Oued Oum Er-Rbia over multiple seasons. The parameters monitored included turbidity, salinity, temperature, and Silt Density Index (SDI), all known to affect fouling and pretreatment requirements.