SDOF SHM benchmark

This dataset contains three years of numerically generated Structural Health Monitoring (SHM) data, considering a single sensor setup on a fixed–fixed steel beam benchmark. It includes hourly acceleration records (3-minute signals at 100 Hz), mid-span deflection values, input profiles (temperature, loads, ambient vibration), and labels for all simulated conditions. The data captures the combined effects of Environmental and Operational Variability (EOVs), multiple fast-varying damage scenarios, slow-varying damage due to corrosion, and eight classes of sensor faults or malfunctions. The benchmark is divided into five sub-datasets covering undamaged behaviour, abrupt stiffness loss, progressive deterioration at different rates, signals contaminated a posteriori with sensor faults and malfunctions, and combined conditions. All simulations were produced using a state-space formulation of an SDOF model for dynamic response and an Euler-Bernoulli beam model for static response, with stochastic live load modelling and temperature-dependent material properties. The repository also provides the full code (Jupyter notebooks) used to generate the data, allowing users to reproduce, modify, or extend the benchmark. This dataset is intended as an open reference for testing and validating SHM methods under realistic interacting effects, including temperature variations, operational changes, structural degradation, and measurement anomalies.