Bridge Structural Health Monitoring Dataset: Train-Induced Vibration Data Under Multiple Damage Scenarios

Simulated acceleration time-series dataset from a railway bridge structure under train loading with varying operational and environmental conditions. Contains vibration measurements from 18 sensor nodes (3-axis accelerometers) sampled at 256 Hz. Files: healthy.parquet: Intact structure baseline (Class 0); slightly_damaged.parquet: Baseline variations from environmental/operational factors with damage < 0.01 (Class 0); damaged_d1.parquet to damaged_d5.parquet: Five distinct damage types (Class 1). Columns: ExperimentID: Unique simulation run identifier; Velocidade: Train speed (km/h) - Environmental Operating Variable (EOV); Peso_Vagao: Wagon mass percentage - EOV; Modulo_Elasticidade: Young's modulus percentage - EOV; Irregularidade: Track profile ID - EOV; Dano_Percentual: Damage severity (0.01 to 1.0) - damaged files only; NodeID: Sensor identifier (18 nodes); Time: Timestamp (sampling rate: 256 Hz); Accel_X, Accel_Y, Accel_Z: Tri-axial acceleration. Suitable for damage detection research, anomaly detection algorithms, and structural health monitoring under environmental/operational variability. Format: Apache ParquetSource: Converted from MATLAB/HDF5 simulation data