Effects of systematic variation of wood adherend bending stiffness on fracture properties. Part 2. Revisiting traditional DCB analysis methods

Abstract

Part 1 of the paper evaluated how the orientation and stiffness of the layers in layered beams of constant cross section used in double cantilever beams (DCB) specimens infl uence the stiffness variation of individual adherends along their length. The behavior of bonded DCB specimens made with such materials, of which wood is a common example, are analyzed in this part 2 of the paper to determine errors associated with common data analysis methods for mode I fracture testing: simple beam theory (SBT), corrected beam theory (CBT), experimental compliance method (ECM or Berry method), and area method (AM). In particular, the fi rst three methods are described in British Standards (BS) 7991 (1991) or the American Society for Testing and Materials (ASTM) D 3433 (1999/2005), while the AM is not suggested by the mentioned standards. SBT, CBT, and ECM, although initially developed for characterizing bonds between uniform and isotropic adherends, have also been commonly applied to other materials including wood. Nevertheless, these three standardized methods may lack precision for determining the critical strain energy release rate, Ic , when applied to bonded wood adherends, due to the elastic stiffness variability that can occur along the length of the bonded beams. The AM yields more coherent results in the developed analytical procedure, although practical issues can limit its reliability with experimental results. Another physical problem that arises with the adherend stiffness variation is the onset of a slight mode II loading component that is not anticipated, nor accounted for, in the traditional data analysis methods.

Quasi‐Static Fracture Tests

“…The geometry of this problem is shown, with exaggerated proportions, in Figure 3 a DCB is shown to illustrate a typical adherend. An evaluation of the effect of grain orientation on the values of G c in pure mode I tests is proposed in [126,127], where it is shown that traditional methods for the evaluation of G c , implicitly assuming the stiffness of the beam to be constant, confuse elastic stiffness variability with G c variability, thus generating data with large scatter.…”