Assessing the uniaxial compressive strength of extremely hard cryptocrystalline flint

Cryptocrystalline flint is an extremely hard siliceous rock that is found in chalk formations. The chalk is frequently a prefered rock type, which in recent decades is often used as a host for underground rock caverns and tunnels in Europe and North America. A reliable estimation of the uniaxial compressive strength (UCS) of the extremely strong flint, with an average UCS of about 600 MPa will provide guidance for a proper engineering design, where flint is encountered, thereby avoiding project progress delay, litigation as well as economic consequences. Conventional UCS measurement using core samples is cumbersome for flint due to the extreme strength and hardness of the rock, for which the core sample preparation process is often extremely difficult. In this study, the UCS prediction models of flints collected from the North-West Europe were developed and the validity of the developed models was investigated. A series of laboratory index tests (comprising the three-point-bending, point load, ultrasonic velocity, density, Shore hardness and Cerchar Abrasivity tests) were perfomed. The index test results were correlated with the UCS values previously determined in the laboratory using both cylindrical and cuboidal specimens to develop the UCS prediction models. Regression analysis of the UCS and the index test results was then performed to evaluate for any potential correlations that can be applied to estimate the UCS of the cryptocrystalline flint. Intensive validity and comparison studies were performed to assess the performance of the proposed UCS prediction models. This study showed that UCS of the tested flint is linearly correlated with its point load strength index, tensile strength and compressional velocity, and is parabolically correlated with its density. The present study also demonstrated that only a couple of the previously developed empirical UCS models for estimating UCS are suitable for flint, which should be used with care. Accepted version