NDE data is often discarded the moment the test is done, eliminating the ability to learn how a part evolves over time. Even when data is saved, it lacks interconnectivity, limiting how easily it can be integrated into a holistic representation of the part. We show how to improve the way NDE data is accessed, searched, and managed through customizable transformations and representations. We illustrate how fully digital process records can be used to generate the interconnections that are the key to automated search. We propose a future system where the data is automatically registered with physical geometry, and Bayesian inference is used to obtain best estimates of the part condition. The system will incorporate the history of motivations, rationale and methods of data acquisition to derive the meaning and intent of a high level query and give the best response. Combining otherwise-disconnected data sets provides the diverse information needed to answer high-level questions.

Related Publications