Abstract
Quantitative characterization of impact damage in polymer matrix composite panels is desired to inform the initial conditions of damage evolution models for subsequent mechanical loading. Previous work from the authors focused on predicting X-ray Computed Tomography data from Ultrasonic Testing data to produce a 3D representation of the damage. The predicted damage with this approach contained artifacts near matrix cracks, the tips of delaminations, and damage occluded by the topmost delaminations. In this work, a data fusion approach is developed to segment the damage using both sets of data directly. This ‘data fusion’ approach involves the training of a classifier using both UT and XCT data as inputs and predicting damage/no damage as the output. Details of the model and data processing are described, along with the resulting segmentation.
How to Cite:
Sparkman, D. M., Welter, J. T. & Uchic, M. D., (2019) “Data fusion approach for ultrasonic and X-ray computed tomography data”, Review of Progress in Quantitative Nondestructive Evaluation .
Downloads:
Download PDF