A Modified Split Hopkinson Pressure Bar Approach for Assessing the Effect of Oscillatory Stress Fluctuations on Rock Strength

Résumé du livre

Previous experimental work has demonstrated that stress and strain rates above a lithology specific threshold are a necessary condition for rock failure to occur by fragmentation. However, this assertion is difficult to reconcile with the occurrence of pulverized rocks at distances up to 100m from principal slip zones, as the forces required to overcome the fragmentation threshold occur only within several centimeters of the fault during the propagation of an earthquake rupture. Previous researchers have invoked the accumulation of damage over multiple slip events to explain this paradox and have demonstrated multiple loading events can reduce the fragmentation threshold by a factor of two, though the experimental support of this idea has been limited to highly idealized loading pulses. Similarly, microcrack accumulation due to the rapid stress oscillations that are more representative of an earthquake rupture event may also reduce the pulverization threshold. In this work, a previously described Split Hopkinson Pressure Bar approach in which two striker bars are connected in series by an elastic spring is implemented to better mimic the loading conditions of an earthquake event. A numerical model for predicting the resulting load path for a given experimental setup is also provided. Under these more realistic compressive loading events, the stress required to overcome the pulverization threshold of Westerly Granite can be reduced to ~200MPa from ~300MPa, obtained from traditional experiments. These results suggest that the idealized load path of traditional experimental approaches overestimates the real rock strength under complex oscillatory loading. While this reduction in material strength is not in itself sufficient to account for the larger observed extent of pulverized fault zone rocks, our analyses demonstrate that macroscopic brittle failure is a non-conservative process. Therefore, when simulating prototype processes, the actual load path specific to the event of interest should be honored for the experimental results to be scalable. These results have important implications for the development of fault damage zones during the seismic cycle and the developed method may also be applicable to other geological or engineering problems that involve rapid stress cycling, such as rock blasting, and projectile impact.