Sparse structural health monitoring using signature-informed hybrid modeling – SPARSE-SHM

Structural Health Monitoring (SHM) enables load estimation and damage detection from in-situ sensor measurements. However, its widespread use is hindered by the required number of sensors and the cost of the electronic devices for data collection and processing. The SPARSE-SHM project tackles these challenges by reducing both sensors density (spatial sparsity) and data volume (temporal sparsity). It relies on a novel concept of signature-informed hybrid modeling. The idea is to build a reduced order model from the system equations which captures just enough spatial and temporal information to identify the parameters of interest, with uncertainty quantification. Dedicated real-time data assimilation techniques are used to compensate for the limitations of the model, enabling reliable monitoring with minimal sensors count and optimal time sampling. By combining physics and data, this hybrid approach paves the way to sparse, cost-effective, and robust monitoring solutions for complex systems.