EUVIMEDEUVIMED entdecken
Crossref

Assisted Data Analysis for Nondestructive Evaluation

Eric A. Lindgren

e-Journal of Nondestructive Testing · 2026-07 · 31 (7)

DOI: 10.58286/33231 ↗

Kurz gefasst

Fachartikel in der Zeitschrift e-Journal of Nondestructive Testing, 2026-07. Verfasst von Eric A. Lindgren.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

There is increased exploration in using artificial intelligence and machine learning (AI/ML) for nondestructive evaluation (NDE) data analysis. However, data required for robust training of algorithms are limited and use of simulated data is challenged by the variability and complexity found in engineered systems. To facilitate model selection as a function of available data, this work explores the impact of data quantity and quality on the accuracy of common AI/ML algorithms. In addition, it covers methods being exploited for aviation applications that integrate heuristics and model-based techniques to assist in the analysis of NDE data. While assisting in analysis of NDE data, an additional benefit is relieving human inspectors of the tedious workload that can compromised performance. Several case studies for both metallic and polymer matrix composite structures are addressed. These include Assisted Data Analysis (ADA) algorithms that combined at least two of the three classes of algorithms to realize inspection objectives where the capability is validated by sensitivity or probability of detection studies. This approach has enabled assessments of complex and/or highly variable aerospace structures.

Zitieren

APA

Lindgren, E. A. (2026). Assisted Data Analysis for Nondestructive Evaluation. e-Journal of Nondestructive Testing, 31(7). https://doi.org/10.58286/33231

Vancouver

Lindgren EA. Assisted Data Analysis for Nondestructive Evaluation. e-Journal of Nondestructive Testing. 2026;31(7). doi:10.58286/33231

BibTeX
@article{Lindgren2026,
  author = {Lindgren, Eric A.},
  title = {Assisted Data Analysis for Nondestructive Evaluation},
  journal = {e-Journal of Nondestructive Testing},
  year = {2026},
  volume = {31},
  number = {7},
  doi = {10.58286/33231}
}
Kostenlos starten →