The in situ elastic moduli of rock are measured holographically by a borehole instrument in marble and oil shale. A point force induced displacement field is recorded with double exposure holography. Data consists of a fringe pattern on the viewed hologram. The elastic moduli are determined by modeling of the fringe pattern. Only Young's modulus is fairly constrained by the present data reduction. For an applied force of 520 ± 50 Newtons, a marble was found to have a Young's modulus of 56 ± 8 GPa, near an ultrasonically determined value of 48 ± 6 GPa. In an oil shale, an applied force of 295 ± 10 Newtons indicates a Young's modulus of 25 ± 1GPa.
Accurate knowledge of the in situ elastic moduli is required by most conventional stress measurement techniques (.McGarr and Gay, 1978). The holographic in situ stressmeter, which records a micron scale stress-relief dis- placement field (Bass, Schmitt, and Ahrens, 1986), requires values of Young's modulus and Poisson's ratio for a volume of rock with dimensions 10 cm square by 1 cm deep adjacent to s. borehole wall surface. Other stress measurement techniques: the C.S.I.R. "doorstopper" cell, (Leemah, 1969), the C.S.I.R. triaxial strain cell, (Leemah and Hayes, 1966), and the direct strain-gauge technique, (Swolfs, Handin, and Pratt, 1974) would benefit from small scale elastic moduli magnitudes. With this primary motivation in mind, double exposure holography is applied to the measurement of the in situ elastic moduli. In situ elastic moduli are difficult to measure due to problems in reproducing the original rock properties such as the state of stress, fluid saturation, jointing, and microfractures (Cheng and Johnson, 1981; Jaeger, 1979). Conventional in situ testing methods: compression tests, plate bearing tests, fiat jack tests, dilatometer tests, the "petite seismique" test, and the borehole jack tests (Bieniawski, 1978; Jaeger, 1979), measure the rock elastic moduli on varying length scales, force levels, and strain magnitudes. The rock modulus is dependent upon the sample scale and the density of discontinuities (Bieniawski and Van Heerdeen, 1975). For example, the N-X borehole jack method applies a uni-directional force of up to 703 KNewtons via two 20.3 cm long platens to the walls of a 7.62 cm diameter borehole (Goodman, Van, and Heuze, 1970). Changes in the hydraulic fluid pressure and the borehole wall diameter are measured; typical displacements are on the order of 0.2 into (Meyer and McVey, 1974) resulting in strain magnitudes near 1000 µstrains. The test is affected by the tensional properties of the rock surrounding the borehole (Heuze and Amadei, 1985). In a plate bearing test, compressional forces of up to 11.8 MNewtons are applied to areas on the order of I m2(Jaeger, 1979). Induced displacements are 0.25 to 2.5 mm and are measured with dial gauges; strain levels encountered are also approximately 1000 µStrains. The holographic test, described in this paper, is on a much smaller scale. The test, similar in principle to plate bearing tests, requires application of a normal point force to the borehole wall of 50 Newtons to 5 KNewtons.
In situ experiments at the Department of Energy's Nevada Test Site (NTS) have extended the technology of propellant-based fracturing to include liquid- filled, uncased boreholes. Previous work had developed the technology for liquid-free, uncased wellbores. Test results show that fracturing behavior is considerably more complex in liquid- filled boreholes than in the liquid-free case. Compared to liquid-free boreholes, liquid-filled boreholes have faster pressure risetimes and lower peak pressures, large hydraulic-type fractures propagating several meters beyond the end of the test zone, numerous oscillations in the pressure history, and variable types of fracturing within a single test region. Multiple fracturing is initiated in the wellbore of a fluid-filled borehole when the proper propellant mixture is used but, unlike fluid-free wellbores where up to eight fractures propagate, only one to three hydraulic- and/or shear-type fractures propagate. Hydrodynamic effects involving liquid compression during borehole pressurization appear to be responsible for the differing phenomena observed in fluid- filled boreholes.
INTRODUCTION AND BACKGROUND
In many gas and oil wells where production is dependent upon the wellbore intersecting natural fractures, stimulation is required after drilling to obtain commercially economic production. Traditionally, this has been done with high explosives or with hydraulic or foam fracturing. High explosives, however, can produce a crushed, compacted region in the vicinity of the wellbore that can effectively seal off production. Hydraulic or foam fracturing avoids the crushing but produces only a single fracture across the wellbore. Because this fracture is usually parallel to existing natural fractures, communication with the formation may not increase significantly. Tailored-pulse fracturing techniques have been developed for liquid-free wellbores that avoid crushing yet produce up to eight major fractures radiating from the wellbore. One such technique, Sandia's High Energy Gas Fracturing (Cuderman,1984a; Cuderman and Northrop, 1984; Cuderman, 1984b), uses a full wellbore charge of propellant tailored to produce the optimum borehole pressurization. These techniques enhance production be- cause of the increased number of fractures available to intersect natural fractures that otherwise would not communicate with the wellbore. Experiments in liquid-free boreholes included in situ experiments to determine pressure pulse and fracture behavior as a function of borehole diameter and propellant mixture; these tests were conducted in G-Tunnel at the Department of Energy's NTS, where the fracture patterns were directly observable by mining back to uncover the test bed. Experiments in Devonian shale at Rowan County, Kentucky, and Meigs County, Ohio, proved the technology developed at NTS in ash-fall tuff was equally applicable in a different rock type. From this work, the technology for propellant- based fracturing in liquid-free boreholes emerged. Data from more than 30 field tests demonstrated that the pressure risetime, t , is the most critical factor in determining the type of fracturing obtained, and that the principal in situ stresses control fracture orientation. Figure 1 is a typical pressure pulse obtained during multiple fracturing in uncased, liquid-free wellbores. An initial pressure rise is normally observed, after which the pressure drops and then again increases until the propellant burn is complete.
During 1982 and 1983, it became increasingly clear that North River No. I Mine would have to buy a set of replacement longwall shields. Despite extensive efforts to make improvements, the shields that were in place at that time were incapable of holding the roof. North River Energy had limited experience in specifying and acquiring longwall roof supports, but a great deal of experience with our conditions and needs. We there- fore compiled a shield evaluation method that uniquely responded to our situation. The resulting system emphasized teamwork among operating, maintenance and engineering personnel. The charge from management was threefold: select shields to support maximum expected loads, optimize face operations, and pay attention to detail. Final selection was made objectively, based on all available information. We chose a 5516 KN (620 tons)support with a minimum yield load density of 760 KN/m2 (7.9 tons/ft2). This paper reviews our process of specification, design, prototype testing, and fabrication, and offers some relevant suggestions.
Volumes of material have been written about longwall mining and roof control. Little, however, is directed toward helping the mining company choose and test the roof support that optimizes both roof control and operational efficiency. This is probably because every mine is unique in its combination of roof and operating conditions. The process of successfully choosing roof supports is dependent on roof and equipment analysis. Both of these come largely from the mining company engineers and the roof support vendors. There is no standard method of comparing analyses and proposals, so great care must be exercised in choosing one roof support over another. Such critical parameters as canopy ratio, support capacity, and roof support density can be calculated by different methods. For accurate comparisons, one method must be chosen to recalculate these and other parameters from proposal drawing measurements. When design and specification are complete, the prototype shields must be carefully tested for compatibility and durability. Finally, once the prototype shields have passed every test, the series fabrications must be manufactured to the required standards.
The first question is: How much support is necessary to hold the roof? We used every available resource including books, articles, government agencies, other mines, consultants, and vendors, as well as the "laboratory" of our own longwall. To calculate necessary support capacity, we determined the weight of the roof block that would rest on the support (Figure 1). The width of that block is easy to figure. It is the distance from the center of one shield to the center of the next. The length and height are a little more difficult. The length of the roof block equals the length of the shield canopy, plus the distance from the tip of the shield to the face, plus the roof rock overhang behind the shield. Generally the rear over- hang is not broken off square with the face. In our case, many observations behind the operating longwall indicated that the rear of the roof block usually breaks at about a 35 degree angle.
Single-row bench blasts at a reduced scale in the field were used to study the effects of delay time on fragmentation. Nine shots with a spacing of 21 in. and a burden of 15 in.(S/B : 1.4) and nine shots with a spacing of 30 in. and a burden of 15 in. (S/B : 2.0) were completely screened to determine particle size distributions. Delay intervals between holes ranged from 0 to 45 msec, corresponding to effective delay intervals ranging from 0 to 36 msec per foot of burden. Each shot was instrument- ed with strain and pressure gages to measure in situ blast dynamics and to evaluate interactions between blastholes. Finer overall fragmentation was produced for shots with delay intervals between 1 and 17 msec per foot of burden. Only for shots within this optimum delay range was it observed that the strains induced by stress waves constructively interacted with strains induced by gas pressure from an earlier detonated hole. Coarsest fragmentation resulted at delay intervals less than 1 msec/ft, where stress waves from each hole were observed to interact destructively, and at delay intervals greater than 24 msec/ft, where no interaction between holes was observed, indicating a condition that can be considered as the firing of ingle hole shots independently.
Manufacturers have recently improved the precision of pyrotechnic caps, and benefits of improved fragmentation have been reported (1, 2). The Bureau of Mines is currently conducting research on blast induced fragmentation of rock. Tests thus far have been performed in the laboratory and at reduced scale in the field and have mostly been concerned with the effect of delay time on fragmentation and the interaction between shot-holes. Although the research program will eventually include full-scale blasts, initial testing in the laboratory provided an effective means for establishing a methodology of controlled experimentation. The tests at reduced scale in the field per- mitted the screening of the entire muckpile to develop fragmentation assessment techniques and results to optimize the expensive full-scale field tests. This report discusses the reduced-scale field tests and results. The reduced-scale field tests were conducted at the University of Missouri's Experimental Mine in Rolla. This site was chosen for its accessibility and geology and for the cooperation available from the University. Furthermore, the results of previous research conducted at the mine on blast design and fragmentation were reported in several theses (3-5). Although these studies investigated various design factors affecting fragmentation, such as coupling, initiation sequence, primer location, and air gap, there were tests that provided a comparison to the Bureau test results.
The 45-in. bench of dolomite in which the experiment was conducted is part of the Jefferson City Formation. The rock is of irregular grain size, 10% calcite, and thick-bedded with a specific gravity of 2.65 and longitudinal and shear velocities of 14,800 and 8,100 fps, respectively (5). Velocities of the dolomite were measured in situ using accelerometers located behind the blastholes and determined to be 14,700 fps (longitudinal) and 8,100 fps (shear).
The need to develop accurate predictive hydraulic fracturing simulators arises from the consideration of vertical containment for the fracture. Fracture height growth from a pay zone into adjacent layers has been shown (see Cleary ) to depend on three sets of parameters for the different strata: