INTRODUCTION
The examination of the effect of scale on measured in situ stress data has been one focus of an experimental program conducted by AECL Research at the Underground Research Laboratory (URL) [ 1]. The stress determination methods used in the program include conventional overcoring, shaft convergence measurements, under-excavation tests, and the bored raise rosette overcoring experiment presented in this technical note. The purpose of the bored raise rosette overcoring experiment was to ascertain whether or not the method can be used to determine in situ rock stresses, and to examine the possible effect of scale on the stress results obtained by the overcore technique.
Bored raise rosette overcoring was previously conducted by Brady et al. [2] at the Mount Isa Mine in Australia. Brady et al. conducted 14 sets of overcore stress determination experiments along the length of a 1.8-m-diameter vertical excavation. Each station consisted of four strain rosettes located around the circumference of the raise, and each rosette included five gauges, formed by a pair of measurement pins. Brady et al. [2] concluded that the bored raise overcoring method successfully determined the field stresses in the rock.
A similar experiment was conducted in the ventilation raise at the URL. The experiment consisted of overcoring four resistance strain gauge rosettes located on the surface of a bored ventilation raise employing a procedure similar to that of Brady et al. [2] (Figure 1). The URL experiment also included CSIR (Council for Scientific and Industrial Research) overcore stress measurements conducted in boreholes extending radially away from the ventilation raise and considered the effects of elastic anisotropy on the stress results. The large-scale (bored raise rosettes) and small-scale (CSIR overcore) tests were intended to determine whether or not a scale effect exists for overcore tests at the URL.
The measured strains during bored raise overcoring are related to the in situ stresses in the same manner as for CSIR overcore tests [3]. In both CSIR and bored raise overcoring the rock is assumed to behave elastically, and therefore the stresses around a cylindrical opening can be defined theoretically using continuum mechanics. In order to obtain the six unknown components of the in situ stress tensor, six independent observations are required. However, in either the bored raise or the CSIR overcoring, more than six strain observations are recorded; hence, a least squares approach is implemented to obtain representative values for the magnitudes and directions of the three principal stresses.
The Underground Research Laboratory is located within the Lac du Bonnet batholith approximately 100 km northeast of Winnipeg, Manitoba. The batholith is representative of granitic intrusions in the Precambrian Canadian Shield. At the test location, the rock mass is predominantly massive and unfractured, and the rock is described as a homogeneous to slightly gneissic, medium-grained, grey granite.
The experiment layout is illustrated in Figure 1. The rosettes were positioned on the wall of the ventilation raise in two paris at right angles to one another (VRI 1 and 2 formed one pair and VRI 3 and 4 were the other). Positioning the rosettes at right angles allowed a reference check on the rosettes' circumferential strains. According to elastic theory, the sum of the circumferential strains from VRI 1 and VRI 2 should equal the sum of similar gauge readings from VRI 3 and VRI 4. Each bored raise rosette was individually oyercored using a 300-mmdiameter bit. Temperature-controlled drilling water was supplied and the drill was instrumented to monitor drill speed, torque, thrust and penetration depth.
