Detection of crack onset in double cleavage drilled specimens of plaster under compression by digital image correlation e Theoretical predictions based on a coupled criterion R. Romani a , M. Bornert b , D. Leguillon a, * , R. Le Roy b, c , K. Sab b a Institut Jean Le Rond d'Alembert, Sorbonne Universit es, Universit e P. et M. Curie, CNRS UMR 7190, 4 Place Jussieu, 75005 Paris, France b Laboratoire Navier, Universit e Paris-Est, CNRS UMR 8205, ENPC, IFSTTAR, 77455 Marne-la-Vall ee, France c Ecole Nationale Sup erieure d'Architecture Paris-Malaquais (ENSAPM), 14 Rue Bonaparte, 75272 Paris, France article info Article history: Received 19 March 2014 Accepted 9 December 2014 Available online 17 December 2014 Keywords: Fracture Drilled specimens Image correlation abstract Geomaterials such as rocks and concrete are brittle or quasi-brittle materials. Tensile tests carried out to observe the initial phases of crack nucleation are difficult to achieve because of the unstable nature of the tests. Instead, compression tests on drilled specimens offer a greater stability. When subjected to a compressive loading, two opposite cracks take place and grow from the cavity, parallel to the load. This crack nucleation is experimentally studied in rectangular drilled specimens of plaster with a centred cylindrical hole which size is assumed to be small with respect to the dimensions of the specimen. The results are compared to a theoretical prediction of the crack onset derived from the coupled criterion of Leguillon. Due to the difficulty of determining the crack initiation directly by the naked eye, 2D Digital Image Correlation is used. The nucleation event is determined by analysing the history of deformations at some points where the crack is expected to start. The predictions are proving to be in good agreement with the experimental results. © 2014 Elsevier Masson SAS. All rights reserved. 1. Introduction In civil engineering and rock mechanics, compression tests like the Brazilian test (see Li and Wong (2013) for a review) or the double cleavage drilled compression test (DCDC) (Sammis and Ashby, 1986; He et al., 1995; Fett et al., 2005, 2009; Plaisted et al., 2006; Plaisted and Nemat-Nasser, 2007; Wong et al., 2006) are often preferred to tensile ones because they offer a greater stability. Under a compression load, a tension crack (mode I) initiates in the direction of the compression loading, then it grows gradually and stably as the loading increases without resulting in the complete failure of the specimen. However, it should be noted that in all the above mentioned papers, it is the growth of a crack from a pre-cut that is studied, not the initiation of the crack itself which is the subject of the present analysis. The aim of this paper is to study the crack initiation in quasi- brittle materials using drilled specimens. Plaster is chosen as a model of brittle geomaterial, it is cheap and easy to handle. Moreover some authors pointed out that it can also be a model for industrial ceramics (Vekinis et al., 1993; Meille et al., 2003). The prediction of crack initiation cannot be carried out using classical brittle fracture criteria because they lead to a paradox. The Griffith criterion based on energy is unable to predict new crack nucleation and the maximum tensile stress condition often results in unrealistic conclusions. To solve the paradox, one of the authors proposed a coupled criterion, which involves two conditions that must be satisfied simultaneously: one based on energy and the other on stress (Leguillon, 2002). The energy condition derives directly from an energy balance between un-cracked and cracked states. As a consequence of this balance, it is derived that the crack jumps a given length. The stress condition states that the tensile stress must be greater than the tensile strength all along the ex- pected crack path (jump). The verification of this theory is based here on compression tests carried out on drilled specimens made of plaster. The crack onset is experimentally determined using Digital Image Correlation (DIC) (Peters and Ranson, 1982; Sutton et al., 1983; Chu and Peters, 1985). Our goal is not to definitively validate the theory but only to provide a positive additional element of appreciation for this cri- terion. In particular, we have tried to show experimentally that the * Corresponding author. E-mail address: dominique.leguillon@upmc.fr (D. Leguillon). Contents lists available at ScienceDirect European Journal of Mechanics A/Solids journal homepage: www.elsevier.com/locate/ejmsol http://dx.doi.org/10.1016/j.euromechsol.2014.12.002 0997-7538/© 2014 Elsevier Masson SAS. All rights reserved. European Journal of Mechanics A/Solids 51 (2015) 172e182