No abstract

Cyclic loading affects the electrical conductivity and fluid permeability of kaolinite clay‐bearing sandstone. The effective confining pressure on two sandstones was repeatedly raised and lowered between 3.5 and 31.5 MPa. Permeabilities dropped by 30% for Berea sandstone and 90% for Muddy J sandstone after three cycles; however, total pore volume always returned to its initial value after each cycle. Water salinity had little effect on these results. Electrical conductivity showed no change when a very conductive pore fluid, 1 M KCl water, was used, however the conductivity showed the same decreases as permeability when demineralized water was the pore fluid. These results show that clay particle rearrangement in the pores is probably responsible for the permeability changes.

We measured the densities, total and microcrack porosities, and ultrasonic velocities of a number of core samples from an 856‐m‐deep core hole near the Banco Bonito vent at Valles Caldera, New Mexico. Reductions in porosity with depth define a zone from about 600 m down where hydrothermal mineralization and recrystallization have been most active. This zone is also reflected in a large decrease in the anisotropy of acoustic velocities. Stress orientation estimates based on microcrack orientations at the 812‐m depth as determined by differential strain curve analysis, as well as anelastic strain recovery measurements on a sample from 472‐m depth, show a horizontal E‐W minimum compression direction and a maximum compression inclined about 30° from vertical.

No abstract

The Los Alamos National Laboratory is developing a program in large‐volume testing of rocks. One part of this program is the measurement of size effects on the constitutive properties of rock. The other part is the development of models to explain and predict these effects. We use a unique 44‐MN hydraulic press to load large specimens to failure, while measuring stress‐strain behavior, failure strength, and acoustic emission. In this paper, results of stress‐strain measurements taken on a single sample of "Texas Pink" granite are reported. We took two sets of data with gauge lengths of 0.3 and 10.0 cm. We also discuss existing models of size effect on strength and conclude that a stochastic model is needed to replace existing empirical deter ministic ones. Estimates of initial crack abundance together with our simple model of failure explain the variation of strength with size reported by some investigators.