INTRODUCTION
Atomic Energy of Canada Limited (AECL) is currently investigating excavation damage in highly stressed granitic rock at the Underground Research Laboratory (URL) as part of the Canadian program in nuclear waste disposal. The Mine-by experiment, which is part of this effort, is a large-scale excavation response test that involves monitoring rock displacements and microseismicity in an instrumented volume of rock at the 420 level of the URL as a 3.5-mdiameter, horizontal tunnel is excavated through it. The tunnel is oriented parallel to 02 to maximize differential stresses in cross-section. A stress analysis indicates compressive stresses of 165 MPa in the roof and floor of the tunnel and tensile sidewall stresses of-15 MPa [1]. Prominent notches developed in the roof and floor of the tunnel during excavation, where considerable microseismic activity was observed [2]. This paper summarizes acoustic emission (AE) and ultrasonic velocity results from a portion of the Mine-by tunnel sidewall where peak tensile stresses are expected.
EXPERIMENTAL METHODS
The AE array was installed along the northwest wall of the tunnel, adjacent to the working face, after the tunnel had attained a length of 22 m. Four parallel, 1.3-m-long, 76-mm-diameter boreholes provided access to the interior of the tunnel wall. The boreholes were arranged in a diamond pattern and were inclined at 11 ø so that two boreholes aligned with the o?-o 2 plane and two with the o2-o s plane. The AE array enclosed a rectangular prism of rock about 0.5 m s in volume. Five Panametrics V103 1.0 MHz compressional transducers were spring-loaded against the rock surface in each borehole. Hemispherical brass caps protected the transducer faces and provided smooth contact surfaces with the rock. Velocity surveys were made before and after each AE monitoring period. The received signals were amplified 60 dB and passed through 10 kHz high-pass filters. Waveforms were sampled every 0.1 gsec and recorded on removable hard disks with four Nicolet 440 10 MHz four-channel digital oscilloscopes. Data acquisition was software controlled with an IBM PC via a GPIB bus. Most AE acquisition procedures were identical to those used in the velocity surveys. Four transducers in the borehole nearest the tunnel face served as triggers for the other 12 channels. AE monitoring proceeded continuously, except for intervals when extensive work was carried out at the tunnel face. The monitoring began several days after Round 17 was completed and continued for 3.5 weeks, during which two 0.5-m tunnel extensions (Rounds 18 and 19) were made and a third (Round 20) begun.
P-wave arrivals were manually picked for each velocity survey. Straight-line raypaths were assumed in calculating all velocities. An ellipsoid was statistically fit to the velocity data and used for AE source location. Errors in velocities due to uncertainties in transducer location are estimated to be ñ40 m/s. P-wave arrivals, and particularly clear S-wave arrivals, were manually picked for over I300 AE events. AE locations were calculated using an iterative steepest-descent fitting program. Travel-times from 3 velocity surveys were input into the source location program to assess location accuracies. Location errors averaged 44 mm for the 12 central transducers and 59 mm for the entire array. Event locations outside of the array are very approximate at best. Focal mechanisms were determined for events that located within 0.75 m of the array centre and had at least eight clear polarity picks. Nodal planes were fit to the shear solutions assuming a double-couple mechanism, and pressure and tension axes were determined.
