Laboratory Dataset on Local Scour Around Complex RC Piers with Pile Groups

This dataset contains 336 clear-water experimental cases of local scour depth around complex reinforced concrete (RC) bridge piers supported by pile caps and pile groups. The data were compiled from multiple peer-reviewed laboratory studies and reflect a broad range of hydraulic, geometric, and sediment conditions representative of real-world bridge foundations. Each entry includes detailed parameters describing:
- Hydraulic conditions: flow velocity, water depth, flow duration, Froude number
- Sediment properties: median grain diameter, geometric standard deviation
- Pier column geometry: width, length, shape
- Pile cap configuration: width, length, height, shape, submergence level
- Pile group arrangement: number of piles, spacing, diameter, orientation relative to flow
- Measured scour depth: maximum equilibrium scour depth
The dataset also includes a set of derived dimensionless parameters based on Buckingham π-theorem, which were used in the machine learning analysis in the corresponding publication. These dimensionless variables enable model generalization across different scales and experimental setups.
This dataset supports the paper titled: “Optimizable Machine Learning Models for Accurate Prediction of Flood-Induced Local Scour Around Complex Reinforced-Concrete Piers”
(submitted to Proceedings of the Institution of Civil Engineers - Water Management, 2025)