Key points are not available for this paper at this time.
Abstract Modern aerospace systems depend on dense sensor networks that continuously interact with Structural Health Monitoring (SHM) systems to support predictive maintenance and flight safety. However, the reliability of sensor data is often taken for granted, and existing SHM frameworks primarily focus on detecting structural damage without assessing the health of the sensing layer itself. This limitation can lead to ambiguous interpretations, where structural degradations and sensor faults produce similar anomalies in the measured responses, resulting in false alarms or missed detections. To overcome this challenge, this paper introduces a sensing digital twin that extends the concept of structural digitalization to the sensing infrastructure. The proposed framework establishes a physics-informed, self-consistent virtual twin of the sensing network designed to monitor, correct, and quantify the reliability of measurements in real time. The twin is assembled by integrating reduced-order models of the structural dynamics, optimal sensor placement, and realistic sensor-fault mechanisms into an augmented temporal dataset. Three core modules are trained and embedded within the twin: a physics-aware classifier that discriminates between structural and sensing degradations, a correction module that reconstructs signals affected by sensors faults through manifold projections, and a reliability module that estimates the probabilistic level of trust associated with each sensor. The capabilities of the sensing digital twin are demonstrated on a composite wing panel undergoing progressive damage and sensor degradations. Results show that the framework achieves in real-time high classification accuracy, effective signal correction, and complete recovery of sensing reliability, supporting physically coherent and reliability-aware data for SHM under representative aerospace operating conditions.
Fiore et al. (Mon,) studied this question.