ARANEA · ARANEA: Physics-Enhanced AI for Soil–Pile Interaction Monitoring

Ensuring the resilience of critical infrastructure—especially offshore wind turbines and bridges—requires next-generation Structural Health Monitoring (SHM) that delivers real-time, interpretable, and uncertainty-robust diagnostics. Deterministic and stochastic model-updating have advanced the field, but gaps remain in scalability, treatment of non-linear dynamics, and application to soil–pile interaction (SPI). SHM supports safe operation, early damage detection, and condition-based maintenance, reducing risk, extending service life, and lowering costs. This project advances vibration-based SHM by fusing acceleration and strain in a multi-modal framework and applying it to SPI, explicitly handling uncertainty and non-linearity. It goes beyond current practice to provide predictive, uncertainty-aware asset management aligned with the EU Green Deal and Europe’s digital transition.ARANEA combines physics-based modelling and data-driven learning in a hybrid model-updating framework. It fuses acceleration and strain with physics-enhanced AI for SPI in offshore foundations. The method integrates output-only system identification with quantified uncertainty, reduced-order digital twins of the structure–soil interface for real-time monitoring, and inversion of soil parameters via two routes: physics-centred evolutionary optimisation with geotechnical priors, and physics-informed machine learning (PINNs, physics-constrained ODEs, INNs). Validated on synthetic, laboratory, and field datasets, the framework embeds explainability and uncertainty throughout to deliver diagnostics that are auditable and trustworthy for operators. Grounded in substructure formulations such as PISA-type models, ARANEA provides reliable, real-time decision support aiming to reduce inspections, extends asset life, and advances Europe’s sustainability and digitalisation goals.