Through-Thickness Damage Timeline of Fiber Composites under Dynamic Loading

The share of composite materials in major civilian structures has been steadily increasing since the 1950s. Their excellent stiffness/strength-to-weight ratio makes them ideal for applications where weight saving and increased maneuverability are essential. During the past decades, material researchers, structural engineers and manufacturers have been bridging the gap between composite structures and traditional metallic structures.

Their appeal is now permeating into the military sector, they are not only seen as a weight-saving opportunity for the main structural components but also as a vector to replace passive defense systems such as armor plates. Such plates tend to be very thick, and hence heavy, since they constitute the first line of defense against near-field blasts, ballistic and fragmentation attacks in active conflict areas.

The aforementioned threats subject the target to strain variations that occur at a substantially higher rate than quasistatic loads such as fight maneuvering loads. Quasi-static loads induce strain-rates between 10-3 and 10-2 s-1. Such low strain-rates correspond to the loading capacities of traditional servo-hydraulic machines making the  characterization of both metallic and composite materials straightforward for said strain-rates, even for out-of-plane or interlaminar properties.

Moving up in the dynamic domain, high strain-rates can be generated in various manners. High-speed servohydraulic machines are able to generate intermediate strain-rates between 10-2 and 1 s-1. In order to reach higher strain-rate levels, other known experimental methods are the Charpy machine, impact loading, the shock tube, small explosive charges or the Split Hopkinson Pressure Bar.

These experimental methods are often used to characterize material properties on a coupon level, but are usually hard to scale-up when trying to perform tests at a sub-component or component level. In such cases, full-scale blast tests with real-life explosive charges are applied, where strain-rates can reach up to 104 s-1.

Performing tests on full-scale composite panels with realistic explosive charges is a non-trivial task. A secluded location, far from civilian activity, needs to be secured and personnel trained in explosive handling is usually required. What's more, because of the amplitude of the blast, these locations tend to be open-air facilities, where the explosive and the specimens are subjected to changing weather conditions, which can greatly aect the performance of plastic explosives, and hence be a nuisance for test repeatability.

Additionally, measurements are difficult to perform during blast tests. Any kind of instrumentation device risks being damaged during a blast. The explosion cloud can even interfere with the imaging equipment, since it may engulf the test specimen obstructing the video recording during the initial phases.

Therefore, only postmortem inspections can be performed on the composite plates. The through-thickness damage timeline generated by the blast is concealed within the thickness of the panels. Consequently, the panels need to be cut so that the extension of the delaminations can be measured. However, the experimentalist remains oblivious to when and where delaminations initiated and how they propagated.

This work introduces a new experimental setup, the Narrow Beam Impact Test (NBIT), which aims to generate similar loading levels and damage modes as those seen during near- field blast tests but under controlled laboratory conditions. The NBIT consists of a narrow composite beam that is impacted by a soft polyethylene impactor while its thickness is exposed to two high-speed cameras.

By doing so, the through-thickness damage timeline is exposed to the observer in real time. The rest part of the thesis describes the NBIT experimental setup, including mishaps and ideas that were not adopted in the nal design. Several impactor materials were tested and high density polyethylene was selected due to its tendency to generate a smaller impact  ash while not disintegrating during the impact.

The second part describes the experimental results of NBITs performed on composite beams made of E-Glass/Polyester. Damage modes within the impact energy range [50 - 1050] J were thoroughly described and three aspects were analyzed in order to establish the full through-thickness damage timeline: delamination onset time, delamination onset location and delamination propagation.

It was measured that the first delamination onset occurred at an average of 20 s after impact. It was concluded that the delaminations onset time remains constant with respect to the impact energy. Delamination onset was consistently located at the rearmost third of the specimen's thickness and was seen to migrate backwards with increasing impact energy.

The delamination propagation velocity was also measured and three propagation zones were identied with a rapid first propagation speed of ca. 1700 m.s-1. Finally, the backside deformed shape is extracted using image tracking software for later analysis. Additionally, the deformed shape measurements allowed for discussing the dynamic eect on the longitudinal Young's modulus and the in-plane compressive strength.

An increase of 27% and 63% was measured respectively. The fourth part focuses on the numerical modeling of the experiment. Part four describes a Finite Element Model of the NBIT built using the explicit  commercial code LS-DYNA. The eects of the material modeling of the impactor are described along with two contact based delamination models.

A parametric study on interlaminar properties was carried out and it was found that the delamination propagation was shear dominated. Thanks to the experimental deformed shape extracted in part two, the Interlaminar Shear Strength and mode II critical energy release rate were seen to increase by 79% and 10% respectively.

The fifth and final part introduces an analytical model based on 1-D elastic wave propagation mechanics in order to understand the through-thickness stress state of the beams tested in part two. The analytical model was able to predict the delamination onset and location for the first delamination mode.

Finally, Appendix A discusses a tangent of this work that arose while testing dierent impactor materials: the dynamic testing of polymers. The sum of the work presented in this thesis provides an insight into the timeline of through-thickness damage generated under dynamic loading in thick monolithic composites.

Such improved understanding paves the way for designers of defense systems in order to tailor the architecture of lighter armor plates made of composite materials .