Dynamic monitoring of a laboratory steel cantilever beam with induced damage

This dataset contains experimental vibration measurements collected on a laboratory cantilever steel beam monitored under operational excitation and progressive damage conditions (one weak). The beam is a slender steel rectangular section, rigidly clamped at its base and instrumented with eight uni-axis accelerometers distributed at equidistant positions along its height. The response was recorded using a synchronized multi-channel acquisition system at a sampling frequency of 2048 Hz.

The structure was excited by a servo-motor-driven linear actuator delivering periodic impulses, in order to activate a wide range of bending modes. The dataset includes acceleration time histories suitable for Operational Modal Analysis (OMA), modal parameter tracking, and vibration-based damage detection. Progressive damage scenarios were introduced by adding increasing masses at the free end of the beam, producing measurable shifts in the natural frequencies.

The data were used to validate a latent-space Operational Modal Analysis framework based on autoencoder compression and covariance-driven stochastic subspace identification (CoV-SSI). They can also be used as a benchmark dataset for system identification, modal tracking, dimensionality reduction, and structural health monitoring algorithms.