The development of composite materials for structural components necessitates the need for methods to evaluate and characterize their damage state after experiencing various loading conditions. A composite differs from traditional structural materials, such as a metallic alloy, by consisting of a heterogeneous combination of constituent materials, often in layers, to improve the overall strength of the laminate. Delaminations are one common damage mechanism concern with composites. When adhesion decreases between adjacent plies, a void can form in the interior of a laminate causing a localized, weakened region. Assessing the damage state of a material, especially when unobservable from the surface of a material, provides insight into its health and operational ability. Nondestructive evaluation (NDE) is one method for characterizing the damage state of a material without causing an alteration to any component of a structure.&nbsp;This work aimed to analyze data from an embedded magnetostrictive particle NDE method with artificial neural networks (ANNs) to detect delaminations embedded in experimental CFRP beams.&nbsp;&nbsp;Particles of the magnetostrictive material Terfenol-D were embedded in CFRP beams during their fabrication. Delaminations were intentionally created using a piece of tape placed at the midplane of the CFRP beams. Voltage readings were captured along the length of the beams using an actuating-sensing coil apparatus; the sensing coil picked up the magnetic flux from the embedded Terfneol-D particles and converted the magnetic energy to electrical energy according to Faraday's law of induction. The experimental voltage data was preprocessed to eliminate any redundant features. A total of five features were found applicable for this two-class problem of detecting the presence of an embedded delamination. ANNs have shown favor in making damage predictions, especially with NDE techniques, because of their ability to model nonlinear relationships and correlate features internally without external manipulation from a user. Multi-layer perceptrons (MLPs) were trained with a subset of the experimental dataset using two approaches: a single 90-10 training-testing partition and a k-fold cross-validation scheme. The mean accuracy of the MLPs was found to be approximately 71%. While MLPs were found to predict the presence of embedded delaminations, their performance was tied to their architecture and activation functions. The results demonstrate the potential for the implementation of ANNs to indicate delaminations using a noncontact, internally contained sensing method.

A Magnetostrictive Particle Delamination Detection Method in a Composite Material Using Artificial Neural Networks

Paper Proceedings: https://asmedigitalcollection.asme.org/SMASIS/SMASIS2021/volume/85499

ASME 2021 Smart Materials, Adaptive Structures and Intelligent Systems (SMASIS2021)

Supporting the need among researchers worldwide for additional online and multimedia content, ASME now offers the ASME Conference Video Collection. This platform will collect and provide long-term access to conference presentation videos, encompassing all technology areas within ASME's technical conference and events program. About 30,000-40,000 engineers across all mechanical engineering and related disciplines participate in ASME conferences each year. This includes authors of approximately 6,000 to 8,000 technical papers published by ASME, the majority of whom also submit video presentations for our virtual conferences. The video collection will be available as an add-on to existing product packages or as a separate subscription.