Robust SHM: Redundancy approach with different sensor integration levels for long life monitoring systems

Monitoring structures such as bridges or wind turbines using Structural Health Monitoring (SHM) methods allows continuous assessment of their condition and considerably faster action compared to conventional on-site inspection. Inevitably, SHM systems suffer from the same problems as the structure itself: Not only does the structure age over time, but the measurement system itself is also subject to environmental aging. To ensure robust SHM, this paper presents a concept that combines optimal sensor placement with a hardware redundancy approach to provide the most efficient, durable, and reliable SHM for structures. For this purpose, a small-scale experiment is conducted in a climate chamber equipped with various sensor types (accelerometers and strain gauges). The different levels of sensor integration allow reliable detection of sensor errors in the raw data signal. Correlation matrices are used to decide which sensor functions are most reliable at what point in time and whether sensor replacement in the SHM system is required. Furthermore, data normalisation is applied to the raw data signal to compensate for the influence of environmental conditions (ECs) such as temperature and humidity. The proposed concept is validated on the small-scale dynamically excited experiment and shows promising results to operate reliable SHM systems for several decades in the future.