Low- to Mid-Rise Reinforced Concrete Buildings Dataset: Structural Layouts and Nonlinear Lateral Response for Surrogate Modeling
Description
This dataset contains nonlinear lateral capacity data for 8,747 procedurally generated reinforced concrete buildings, analyzed by nonlinear pushover analysis in both principal directions, resulting in 17,494 samples. The dataset includes: - Building layouts: structural layouts of all buildings, together with key structural information including geometry, member sections, materials, and applied loads. - Tabular representation: global structural features and corresponding pushover targets. - Full-Structure representation: complete structural graphs with node, connectivity, and feature data. - Representative-Floor representation: representative-floor graphs combined with global building attributes. For each representation, the same four output quantities are provided: - Maximum spectral displacement (Sdmax) - Spectral acceleration (Sa) capacity curve, represented by 25 equally spaced points - First-mode participation factor (PF1) - First-mode effective modal mass coefficient (α1) These outputs provide the nonlinear lateral capacity and first-mode modal properties required for subsequent capacity-spectrum-based performance evaluation using ATC-40. The dataset also includes training, validation, and test splits, together with the feature scalers used for each machine-learning representation. Full details on the data structures, input features, output variables, and their respective ranges/bounds are provided in the accompanying dataset_summary.pdf.
Files
Institutions
- Polytechnic School of AlgiersAlgiers, Algiers