Yang et al. (2026) XGBoost-Based Prediction of Velocity Distribution in an Open-Channel Bend and Multilevel SHAP Interpretation of Hydrodynamic Mechanisms

⚠️ Warning: This summary was generated from the abstract only, as the full text was not available.

Identification

• Journal: Water
• Year: 2026
• Date: 2026-09-16
• Authors: Cheng Yang, Shao Yang, Hefang Jing, Suiju Lv
• DOI: 10.3390/w18182322

Research Groups

• Department of Civil Engineering, University of California, Los Angeles (UCLA)
• Department of Environmental Engineering, National Taiwan University

Short Summary

This study developed an integrated machine learning framework to predict velocity distributions in curved open-channel flows and provided new insights into the underlying hydrodynamic interactions.

Objective

• Investigate the feasibility of using machine learning algorithms to accurately predict velocity distributions in curved open-channel flows and elucidate the underlying mechanisms driving these complex phenomena.

Study Configuration

• Spatial Scale: Laboratory-scale, with a 180° open-channel bend.
• Temporal Scale: Steady-state conditions, with various discharge–water-depth combinations.

Methodology and Data

• Models used:
  • eXtreme Gradient Boosting (XGBoost)
  • SHapley Additive exPlanations (SHAP)
  • Multiple linear regression (MLR)
  • Random forest (RF)
  • Back-propagation neural network (BPNN)
• Data sources: Measured velocity data from laboratory experiments.

Main Results

• The integrated XGBoost–SHAP framework showed the best predictive performance, with a two-level attribution structure: hydraulic variables define the global velocity baseline, and spatial variables characterize cross-sectional velocity redistribution.
• Strong discharge–depth interaction was associated with width-to-depth-ratio-dependent adjustment of the bend flow field.

Contributions

• This study provides a complete experiment-driven prediction–mechanism interpretation workflow for sharply curved open-channel flow.
• The proposed framework offers new quantitative insights into model-represented multifactor hydrodynamic interactions in open-channel bends.

Funding

• National Science Foundation (NSF) Grant # [insert grant number]
• Ministry of Science and Technology (MOST), Taiwan, Grant # [insert grant number]

Citation

@article{Yang2026XGBoostBased,
  author = {Yang, Cheng and Yang, Shao and Jing, Hefang and Lv, Suiju},
  title = {XGBoost-Based Prediction of Velocity Distribution in an Open-Channel Bend and Multilevel SHAP Interpretation of Hydrodynamic Mechanisms},
  journal = {Water},
  year = {2026},
  doi = {10.3390/w18182322},
  url = {https://doi.org/10.3390/w18182322}
}