Open Access
Subscription or Fee Access
A Roadmap to Account for Potential Uncertainties in Non-Destructive Testing during Structural Health Monitoring of Composites
Abstract
In the context of structural health monitoring (SHM) applications, this article first presents an example of damage detection uncertainty in a typical fiberglass woven thermoplastic plate structure with pre-embedded disbonds of various sizes and locations. A Taguchi experimental design has been implemented to statistically assess the significance of uncertainty in the performed ultrasonic NDT tests. Next, a Gaussian Process (GP)-based SHM framework has been suggested to account for such possible uncertainties in training datasets of SHM applications, and hence arriving at more robust damage systems.