Additive manufacturing is of great importance to many NASA missions due to its versatility and fast fabrication capability. In particular the fabrication of Inconel steel as well as plastic parts are relevant to the production of aerospace components. These parts often can contain voids and intra-granular impurities that lead to critical flaws that reduce performance in terms of strength and durability. These critical flaws can be volumetric or fracture-like in nature and are difficult to detect in the finished part. The goal of the proposed research is to develop a non-destructive method to not only detect and quantify these defects, but also to correlate the results to the expected strength and fatigue life of the part. To accomplish this, we will utilize complementary non-destructive evaluation (NDE) modalities and machine learning (ML) to document critical flaws. The analysis of scanned parts will be independently performed, and the data will be correlated to the NDE observations. The data from a large number of additively manufactured aerospace components will be used to create a training set for machine learning to provide a high-speed inspection process. The Phase I program will demonstrate feasibility of our novel method whereas the Phase II program will logically extend Phase I research to implement a practical system.
The evaluation technique along with the machine learning algorithm we propose is a viable method for rapid inspection of additively manufactured aerospace components post production. The development of this technology will not only address immediate needs relevant to NASA missions, but will also have substantial impact on non-destructive inspection in general.
Additive manufacturing is a rapidly growing fabrication method which is applicable to many industrial applications. This novel evaluation process to be developed under this program will have immediate benefits to other original equipment manufacturers (OEMs) by removing the need to inspect components by destructive means.