Improving measurement accuracy of Brazilian tensile strength of rock by digital image correlation

Since the invention of the Brazilian test about eight decades ago, discussions on the validity of indirect tensile strength measurement by the Brazilian test have always been of focus in rock fracture mechanics. According to the principles of the Brazilian test, experimental data are only valid when...

Full description

Saved in:
Bibliographic Details
Published inReview of scientific instruments Vol. 89; no. 11; p. 115107
Main Authors He, Wenhao, Chen, Keyong, Zhang, Baiyu, Dong, Kanjicai
Format Journal Article
LanguageEnglish
Published United States 01.11.2018
Online AccessGet more information

Cover

Loading…
More Information
Summary:Since the invention of the Brazilian test about eight decades ago, discussions on the validity of indirect tensile strength measurement by the Brazilian test have always been of focus in rock fracture mechanics. According to the principles of the Brazilian test, experimental data are only valid when a fracture is initiated at/near the disc-shaped planar center and propagates along the diametral loading plane. However, traditional experimental results can only show readers a final failure instead of fracture process. Meanwhile, the accuracy of analytical solutions and numerical simulations are profoundly dependent on the input of underlying assumptions. To overcome the confusion about the Brazilian test, this study proposed to combine the Brazilian test with digital image correlation. By visualizing a dynamic fracture process of fracture initiation and propagation, the validity of the Brazilian test can be verified. This study can well explain the underestimation of rock tensile strengths in traditional Brazilian tests. Additionally, the specimen should have a length to thickness ratio larger than 0.30 to minimize the compression effect near the loading contacts. Due to the intrinsic heterogeneity of the rock structure, a small deviation of fracture initiation does not make a big difference in the experimental results.
ISSN:1089-7623
DOI:10.1063/1.5065541