RADAR I\/Q DATA FOR MACHINERY IMBALANCE DETECTION

This dataset contains continuous-wave (CW) radar recordings of motor vibration collected to support the detection and characterization of rotational imbalance. A total of 1,802 independent experiments were performed at a sensor-to-target distance of 30 cm; each experiment comprises 30 seconds of baseband in-phase (I) and quadrature (Q) samples recorded at 10 kHz (300,000 samples per channel). During each recording, motor speed and load were held constant; experimental conditions spanned speeds from 500 to 1500 rpm and loads from 0 to 3 Nm. Controlled rotational imbalance was introduced by coupling a disk to the motor and attaching discrete masses, producing four labeled classes: baseline (0 g), light imbalance (10 g), medium imbalance (20 g), and severe imbalance (30 g). Data are organized in a ZIP archive containing four class-specific folders. The dataset is intended for supervised classification, regression of imbalance severity, spectral\/time\u2011frequency analysis, demodulation, and feature\u2011extraction method development, and evaluation of noncontact sensing approaches for rotor health monitoring.Radar frequency: 24 GHz (CW mode)Sampling frequency: 10 kHzMeasurement duration: 30 secondsSignal length: 300,000 samples (I and Q channels)Motor speed: 500-1500 rpm (11 equally spaced levels)Load torque: 0-3 Nm (6 equally spaced levels)Dataset Structure:Total number of experiments: 1802 (# of experiments per classes  {462, 464, 413, 463})Data formats: CSVRaw I\/Q signalsThis dataset can be utilized for mass imbalance fault diagnosis, machine learning algorithm development, and signal processing method validation in rotating machinery condition monitoring applications.