Developing a generic relation for predicting sediment pick-up rate using symbolic soft computing techniques

Area of Science:

• Fluid mechanics and sediment transport.
• Application of computational intelligence in environmental engineering.

Background:

• Sediment pick-up rate is crucial for understanding sediment transport in open channels.
• Traditional experimental and numerical methods have limitations.
• Soft computing approaches for sediment pick-up rate prediction are underexplored.

Purpose of the Study:

• To develop and evaluate soft computing models for predicting sediment pick-up rate.
• To establish a dimensionless relationship for sediment pick-up rate prediction.
• To compare the performance of genetic programming (GP), grammatical evolution (GE), and gradient boosting machine (GBM) algorithms.

Main Methods:

• Utilized two experimental datasets for training and validation.
• Employed genetic programming (GP), grammatical evolution (GE), and gradient boosting machine (GBM).
• Input variables included dimensionless Froude number, particle diameter, and depth-averaged turbulent kinetic energy.
• Performance was assessed using statistical indices, visual plots, and uncertainty quantification (Tsallis and Renyi entropies).

Main Results:

• Developed three distinct mathematical expressions for sediment pick-up rate prediction.
• Grammatical evolution (GE) yielded the most accurate predictions.
• Performance evaluation confirmed GE's superiority over GP and GBM.

Conclusions:

• Soft computing, particularly GE, offers a powerful and accurate approach for predicting sediment pick-up rates.
• The developed dimensionless models provide a valuable tool for sediment transport analysis.
• Further research can explore other soft computing techniques and a wider range of flow conditions.