| Summary: | The observed fractal nature of both fault length distributions and earthquake magnitude-frequency distributions suggests that there may be a direct relationship between the structure of active fault systems and the resulting seismicity. In much previous theoretical work, a positive correlation between the exponent D from the fracture length distribution and the seismic or AE b-value has been inferred from a simple dislocation model of the seismic source. The main aim of this project was to test this relation quantitatively, for tensile fracture, using results from controlled laboratory experiments. Field work on natural tensile fracture systems was also carried out to test that the results from the artificial fracture in the laboratory were relevant to the natural fractures that were being simulated. First, a series of double torsion tensile tests on crystalline rock, carried out at the Fracture Mechanics Laboratory, University College London, is described. A program of tests was carried out on two rock types, granodiorite and granite of different grain size, to test the effects of rock type and specimen size on the laboratory results. The main controlling variables, used during the tests, that affect the results, are the presence or absence of fluid, at ambient temperature and pressure, in the crack tip environment, and stress intensity (and hence crack velocity). Microseismic acoustic emissions (AE) were monitored during subcritical crack growth under controlled conditions of constant stress intensity, K<SUB>I</SUB>, and quantitative structural analyses of the resulting fracture patterns were carried out on the same specimens. AE microseismicity is found to reflect different aspects of microcrack propagation in the laboratory. AE b-values range form 1.0 to 2.4 and vary depending upon fluid presence at the crack tip. The greater the fluid-rock interaction, the higher the relative proportion of smaller seismic sources (thus producing a higher b-value).
|
|---|
