Dissertation
Data-driven approach for damage detection leveraging the digital thread
Doctor of Philosophy (Ph.D.), Drexel University
Jun 2024
DOI:
https://doi.org/10.17918/00010457
Abstract
Damage incubation, initiation and growth in materials and structures represents a significant engineering challenge related to their safe use, maintenance and remaining useful life in engineering applications. Common damage sources include cracks, delaminations and corrosion, among others, the detection of which has long been the objective of applying Nondestructive evaluation (NDE) methods. Sensing and data processing of common NDE are currently plagued by the challenge of handling large volumes of data, the processing of which often also requires offline analysis and expert knowledge. This creates a significant challenge when attempting to assess the structural integrity of complex engineering structures in operation, fast and reliably. In this context, this research proposes to leverage recent advances in digital processing, as well as developments in the use Artificial Intelligence (AI) to develop automated and data-driven ways to perform damage detection. To achieve this goal, the research presented introduces a Digital Thread between sensors and decision making, leveraging an Internet of Things architecture coupled with a data fusion approach which uses both deep learning and information entropy models. To demonstrate the potential of this methodology, two representative cases of damage detection involving laboratory cases of crack monitoring in a metal alloy subjected to mechanical testing and of defect detection while performing polymer 3D-printing are presented. These cases were chosen as representative examples in which a data-driven approach for damage detection could be used not only for diagnostics but also for prognostics and feedback control. The development of this approach considers the need for scalability, customization and performance in near real-time. As such, requirements, system development as well as performance assessment are presented.
Metrics
105 File views/ downloads
34 Record Views
Details
- Title
- Data-driven approach for damage detection leveraging the digital thread
- Creators
- Sarah Malik
- Contributors
- Antonios Kontsos (Advisor)Ahmad Raeisi Najafi (Advisor)
- Awarding Institution
- Drexel University
- Degree Awarded
- Doctor of Philosophy (Ph.D.)
- Publisher
- Drexel University; Philadelphia, Pennsylvania
- Number of pages
- xx, 319 pages
- Resource Type
- Dissertation
- Language
- English
- Academic Unit
- College of Engineering (1970-2026); Mechanical Engineering (and Mechanics) (1970-2026); Drexel University
- Other Identifier
- 991021890212104721