ABSTRACT The increasing reliance on dense sensing networks in advanced aircraft structures places stringent requirements on the validity and interpretability of measured data used for Structural Health Monitoring (SHM). Conventional SHM approaches assume nominal sensor behavior, an assumption that becomes fragile when sensor degradation and structural damage generate similar response patterns in composite structures [1-2]. This work introduces a real-time digital twin of the sensing network grounded in scientific machine learning, in which onboard sensor measurements are assimilated to learn a physics-informed self-consistent representation of the sensing layer. Within this framework, the learning process enables autonomous discrimination between structural anomalies and sensing faults, reconstruction of sensor-corrupted signals, and continuous probabilistic estimation of sensor reliability. The digital twin integrates a reduced-order representation of the coupled structure-sensor system, an optimal sensor placement strategy, and machine learning modules that leverage physics informed schemes to (i) attribute the source of anomalous measurements, (ii) recover corrupted observations through projection onto physically admissible manifolds, and (iii) assign each sensor a continuously updated reliability score expressed as a probability-calibrated trust metric. The approach is demonstrated on a carbon fiber UAV wing panel subjected to concurrent structural damage and sensor degradation, where results show accurate fault discrimination, restoration of physically consistent strain responses representative of the true structural state, and real-time tracking of individual sensor reliability. REFERENCES [1] Rocha H., Semprimoschnig C., Nunes J.P., Sensors for process and structural health monitoring of aerospace composites: A review, Engineering Structures, Vol.237, 2021. [2] Ariyaratnam S., Ermacora M., Di Fiore F., Mainini L., Virtual sensing reliability characterization for next-generation green aircraft resilience, AIAA Aviation Forum and Ascend, pp.3342, 2025.