Multimodal Signal Dataset for Fault Detection in PMSM-Driven Elevators Under Real Operating Conditions

This dataset contains synchronized time-series recordings of current and vibration signals collected from a fully operational residential elevator system equipped with a surface-mounted Permanent Magnet Synchronous Motor (PMSM). The data were acquired under healthy and faulty operating conditions, across various load levels and motion directions. The dataset supports research in predictive maintenance, signal analysis, and fault classification for PMSM-based elevator drives. All data were collected non-intrusively using industrial sensors: Hioki PW3390 power analyzer for current and THD, Triaxial accelerometer (ISO 20816-3 compliant) for vibration. The system is a 9-stop, 6-passenger MRL elevator, driven by a 5.1 kW, 12-pole PMSM motor. No signal post-processing was applied; raw sensor values are preserved for reproducibility.