Development of a Weak But Brittle Analog Sediment Rock for Experimental Study

ABSTRACT Laboratory-tested results in rock mechanics have played a crucial role in understanding rock behavior and failure mechanisms. This paper outlines the development of an analog rock, similar to sandstone, for an experimental study that yields a low unconfined compressive strength yet exhibits brittle failure characteristics. The F-75 Ottawa sand, Type I/II commercial grade Portland cement, air entrainment admixture, and water constitute the raw mixtures of the developed analog rock. The result is a reproducible homogeneous and isotropic rock. The effect of the air entrainment admixture and freezing-thawing cycle on the analog rock are investigated. This study amplifies the notable macroscopic fracturing after a limited post-yield strain. The unconfined compressive strength has been bounded at around 1.5 MPa, and the material continues to exhibit brittle behavior at a range of confinement. A set of Brazilian indirect tensile tests was used to assess the tensile strength, including the brittle failure analysis governed by the ratio between the unconfined compressive strength and tensile strength. In addition, the material physical properties were obtained, along with the thin-section-based porosity analysis. The synthetic sandstone is developed for a scale-model experimental study of brittle failure in tunnels using a true-triaxial apparatus that include spalling, dog-earing, and rockburst. INTRODUCTION Brittle failure in the underground excavation can be sudden and catastrophic, resulting in large cracks and fractures such as spalling and rockburst. This stress-induced failure is typically encountered in sparsely fractured or massive rocks at deep underground excavations (Martin et al., 1997). Therefore, understanding the nature of undesirable rock mass brittle failure is essential to provide mitigation plans for the associated risks. One approach to enhance the understanding of behavior and failure mechanisms of rocks is through laboratory testing. However, it is challenging to collect natural, homogeneous, and undisturbed rock specimens of varying sizes from the field. Additionally, the natural specimen may introduce alterations to the variables that could differentiate the investigation plan in terms of reproducibility (Johnston & Choi, 1986). Hence, developing a repeatable analog rock mix can be considered with a flexibility range of controlled properties. In this case, the prepared analog rock specimens must have the characteristics of the investigated phenomenon (i.e., brittle behavior).