Consistent Identification of CFRP Viscoelastic Models from Creep to Dynamic Loadings

Abstract

Rate‐dependent models require creep or mechanical tests at various strain rates in order to be identified and validated. Different geometries coexist for creep and static tests (normative geometry) and for dynamic tests. Therefore, due to geometrical sample considerations, experimental results could be inconsistent for identification or validation procedures, inducing, for example, differences on the shear modulus only due to the change of geometry. The objective of this work is to present an improved sample geometry that allows to obtain consistent mechanical tests results at various strain rates highlighting the rate dependencies of laminates. In particular, a complete mechanical validation of the sample geometry for dynamic tests is successfully performed in order to avoid inconsistency. Results of static and dynamic tests on the validated geometry are analysed, and the rate dependency of the elastic properties of the UD T700GC/M21 mesoscopic ply is highlighted on a wide strain rate range (10−3 to 102 s−1). Finally, the identification of a non‐linear viscoelastic model is performed on dynamic and creep tests results in order to obtain a representative model for dynamic, static and creep loadings, and to demonstrate the importance of introducing the improved geometry for the dynamic tests.

Crush responses of composite cylinder under quasi-static and dynamic loading

“…Berthe et al [29] conducted high rate coupon testing on T700GC/M21 material which showed the shear modulus increased by 40% when the loading rate was increased from 10 -3 −1 to 88 −1 but no discernible change in the transverse modulus. This suggests that the ply configuration is the cause of the apparent rate independence observed in this study.…”

A consistent experimental protocol for the strain rate characterization of thermoplastic fabrics

“…In Berthe's work, the same discrepancies in the nonlinear domain were observed between Type A and B specimens. No inconsistencies were observed in the viscoelastic range, however, which was the main focus of the authors' studies.…”

“…All dimensions are summarized in Table . Concerning the viscoelastic part of the response, Type B specimens are able to ensure the representativity of Type A specimens with a free length set to only twice the width and a further reduced section . Section reduction is however not considered in this study.…”

“…Further, no normalized experimental procedures exist for high strain rates, which leads to inconsistencies when different materials are compared . Most studies reveal an increase of elastic properties with strain rate, both for carbon fiber and glass fiber composites and both for the fibers' (0° and 90°) and the off‐axis (45°) directions. Conversely, the moduli in the fibers' directions were found strain rate insensitive in Harding and Welsh and Taniguchi et al,, while the shear modulus tends to decrease with strain rate in Shokrieh and Omidi and Brown et al Almost all studies on the strain rate sensitivity of failure properties reveal an increase of failure stresses for the fibers' and the shear orientations, but contradictory trends are also noticed regarding ultimate strains.…”

“…Some of these inconsistencies are directly related to the use of dedicated reduced specimens in dynamic studies . Shock wave propagation period, load line inertia, and resonance effects are classical difficulties encountered in dynamic experiments using servo‐hydraulic machines .…”