Mixed mode fracture testing of adhesively bonded wood specimens using a dual actuator load frame

Abstract

An experimental evaluation of mixed mode fracture tests conducted on adhesively bonded wood specimens using a dual actuator load frame is presented. This unit allows the fracture mode mixity to be easily varied during testing of a given specimen, providing improved consistency, accuracy, and ease of testing over a range of loading modes. Double cantilever beam (DCB) type specimens made of southern yellow pine (Pinus spp.) wood substrates bonded with a commercially available one part polyurethane adhesive were tested over a wide range of mode mixities from pure mode I to pure mode II. The critical strain energy release rate (SERR) values were calculated from the measured load, displacement, and crack length data, in combination with material properties and specimen geometric parameters, and compared on a versus fracture envelope plot. Mean quasi-static fracture energy values were calculated to be 390 J m-2 and 420 J m-2 for mode I and mode II fracture, respectively. For various mixed mode phase angles, the critical SERR values were partitioned into mode I and mode II components. In mixed mode loading conditions the cracks were typically driven along the interface, which resulted in lower total fracture energy values when compared with those measured under pure mode I loading conditions. A drop in measured fracture energy of approximately 45% was observed with mode mixity phase angles as small as 16°, implying that engineering designs based on results from the popular mode I DCB test could be nonconservative in some situations. Fracture surfaces obtained at different mode mixities are also discussed. An improved understanding of fracture behavior of adhesively bonded wood joints under mixed mode loading through generation of fracture envelopes could lead to improved designs of bonded wood structures.

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

“…SBT and CBT assume that the elastic characteristics of the adherend and adhesive materials are independent of the crack length. The accuracy of this assumption is deemed to be confi rmed by the coeffi cient of correlation, R 2 , of the CBT data fi t. The R 2 value obtained experimentally for CBT and ECM is often very near unity, even in materials with intrinsic variability such as wood (Gagliano and Frazier 2001 ;Nicoli et al 2009 ;Singh et al 2010 ), masking any possible effects of adherend variability. Nevertheless, Ic data for bonded wood specimens are often characterized by larger scatter (Triboulot et al 1984 ;Singh et al 2010 ;Nicoli et al 2012a ) than seen for uniform adherends, raising a question about errors introduced by methods recommended in the standards that were developed for uniform properties.…”

Quasi‐Static Fracture Tests

“…Experience with southern yellow pine [110] and yellow poplar [109,119] has indicatively shown that typical values of G c for mode I tests are in the order of few to several hundreds Joules per square meter, whereas for other structural materials such as bonded steel and aluminum, values in the order of thousands Joules per square meter are not unusual. Without going too much into the detail of the quantitative results of fracture tests in different material systems, a general pattern that one can usually see for bonded wood specimens is that G c data are characterized by data scatter that is commonly higher than what is obtained in homogeneous materials also when failure occurs within the adhesive layer.…”

“…Given the influence of a large number of factors and the fact that wood is a natural material, it is commonly found that fracture measurements taken on wood and bonded wood are characterized by data spread that is usually larger than the corresponding values obtained when testing more uniform adherends [107][108][109][110]. In particular, Triboulot et al [111] indicated with experimental and numerical studies that the high variability in wood properties requires researchers to deal with extensive statistical analysis when critical fracture toughness values are to be evaluated.…”

Effects of systematic variation of wood adherend bending stiffness on fracture properties: Part 1. Influence of grain angle

“…Fracture measurements taken on wood and bonded wood joints are characterized by data spread that is usually larger than what is obtained when testing more uniform adherends (Ebewele et al 1979(Ebewele et al , 1980Triboulot et al 1984 ;Gagliano and Frazier 2001 ;Conrad et al 2003 ;Singh et al 2010 ). This fi nding has often been associated with random variations of properties that are diffi cult to detect in a natural material, but there are also morphological aspects that can possibly be accounted for.…”