Shape optimization of a cruciform geometry for biaxial testing of polymers

Abstract

a b s t r a c tThe presented literature review of cruciform shapes used for biaxial characterization of materials indicates that the majority of shapes can be divided into two large groups when the following selection criteria are taken into consideration: (i) the shape of the outer boundaries and (ii) the load capacity needed to achieve failure in the biaxial region. Manipulation of the outer shape boundaries appears to be essential to bundle the applied loads to the central zone where failure is intended to be built up. For each group, one particular cruciform design is reported whereby the outer boundaries are based on a single curved shape. Although the use of discontinuous double radii edges should be avoided according to earlier reports [1,2], it is shown here through the construction of an optimization algorithm, that the use of a single curve for the outer boundaries leads to strains in the arms that are strongly dependent on these single curved edges. Numerical simulations based on the finite element method as well as experiments performed on polymeric test pieces in combination with DIC measurements, show good agreement on this matter and demonstrate this sensitivity very clearly.

Effect of through-thickness compression on in-plane tensile strength of glass/epoxy composites: Experimental study

“…A study by Smits et al [8] suggested a new cruciform type of a specimen for obtaining reliable biaxial failure data for composites based on finite-element simulations of such specimens combined with respective experiments. Subsequently, shape optimization of these specimens for biaxial testing was investigated in some works [9,10]. A generic optimization process based on the finite-element approach was developed to provide geometry modifications that could fulfil predefined requirements for a successful biaxial test.…”

Basic studies in biaxial tensile tests

“…This can be performed either by finite element simulations, where a constitutive model has to be assumed, or by experimental testing (or a combination of both). Shape optimization using finite elements might either be done by trial and error contour definitions, or by applying numerical tools from optimization …”

“…This can be performed either by finite element simulations, where a constitutive model has to be assumed, or by experimental testing (or a combination of both). Shape optimization using finite elements might either be done by trial and error contour definitions, [27,30,34] or by applying numerical tools from optimization. [20,36,37] Since we are interested in talking about homogeneity of the strain distribution in a certain region, we have to introduce a measure.…”

“…Recently, Lamkanfi et al investigated the influence of the geometrical discontinuities due to the production of the milled zone and the fillet corners on the strain distribution within laminate composites. Later on, in Lamkanfi et al, the study of shape optimization of polymer cruciform specimens was presented with the objective of generating damage in the middle area. By reviewing recent experimental literature, they made distinction between the cases where the load trajectories are left unchanged, that is, high loads (up to 1500 kN per arm) are applied, and the cases where more moderated loads (100 kN per arm) are applied and the load trajectories are manipulated by changing the curvature of the corner between the arms.…”

Mechanical Behavior of a Fuselage Stiffened Carbon-Epoxy Panel under Debonding Load

“…Concerning the composite structure, this one has to be created case by case, in a representative scale, in order to regroup the specific singularity details of interest, such as stiffener-skin interfaces. For the multi-axial modular testing machine, this must have the means to apply mechanical, thermal and/or electrical loads very similar to those found on the real composite structures [5,6]. One of the many advantages to build an in-house testing machine is the capability to design it with the specific needs of the research.…”