Laboratory investigation of scale effects in bursting failure of quasi-brittle materials

Understanding the scale effect in the behavior of solid materials is important when small-scale models are used to predict the behavior of real structures. Tremendous efforts have been devoted to investigating the scale effect in the strength of quasi-brittle materials. However, limited studies have focused on the whether there is a scale effect in bursting failure, which is a typical failure pattern of quasi-brittle structures. Here, we conduct a series of unconfined compression tests on coal samples using a specialized designed loading machine to produce bursting failure in the laboratory. We use cubic coal samples with a large range of sizes from 50 to 300 mm to investigate the scale effect in bursting failure of quasi-brittle materials. We use acoustic emission technology to monitor microcreaking within the samples and a high-speed camera to capture the bursting failure process. We demonstrate that there is a strong scale effect in bursting failure of quasi-brittle materials in terms of mechanical properties including compressive strength and deformability, strain energy density, bursting severity, and precursor time. We found that ejection of localized pieces from the skin of the structure is a precursor of catastrophic bursting failure of quasi-brittle materials. We conclude that the relationship between the precursor time (T) and structure size (M) of bursting failure is logarithmic and can be described as logT = a + bM where a and b are constants. This relationship provides a guideline when models obtained from laboratory tests are used to predict strength and behavior at the intermediate (mines and rock slopes) and large scales (earthquakes).