Lang, P.A. (Whiteshell Nuclear Research Establishment, Atomic Energy of Canacla Ltd) | Chan, T. (Whiteshell Nuclear Research Establishment, Atomic Energy of Canacla Ltd) | Davison, C.C. (Whiteshell Nuclear Research Establishment, Atomic Energy of Canacla Ltd) | Everitt, R.A. (Whiteshell Nuclear Research Establishment, Atomic Energy of Canacla Ltd) | Kozak, Everitt T. (Whiteshell Nuclear Research Establishment, Atomic Energy of Canacla Ltd) | Thornpson, P.M. (Whiteshell Nuclear Research Establishment, Atomic Energy of Canacla Ltd)
Instrumentation, installed during excavation of a shaft in granite, monitored displacements, stress changes and temperatures in the rock mass, and piezometric pressures in vertical fractures. As well, the permeability of the fractures was measured after each excavation round. The experiment demonstrated that the permeability of fractures beside the shaft decreased twofold as the shaft was excavated past the fractures. Three-dimensional finite-element modeling simulated the stress changes and displacements reasonably well, but predicted the opposite trend for permeability change. It was concluded that: a) mechanisms occurred in the fractures which are not properly understood at present, b) future experiments of this type should measure fracture displacements at the same location as the piezometric pressures and permeabilities are measured, and c) a circular shaft is preferred to a rectangular shaft for this type of experiment.
Atomic Energy of Canada Limited (AECL) is responsible for developing and demonstrating a methodology for the safe, permanent disposal of Canada's nuclear fuel wastes. Research efforts are being concentrated on geological disposal deep in stable rock formations of the Canadian Shield. As part of this research program, AECL is constructing an Underground Research Laboratory (URL) near Lac du Bonnet, Manitoba. The URL provides researchers with access to a previously undisturbed granitic rock mass to conduct a wide variety of in situ geotechnical experiments (Simmons 1986). Prior to starting construction, three years were spent characterizing the geology, hydrogeology, in situ stresses and intact rock properties at the site. Between 1984 May and 1985 March, a rectangular access shaft, 2.8 m by 4.9 m, was sunk to 255 m depth, and shaft stations were excavated at 130 m and 240 m depth. (Figure 1). Excavation was done by conventional drill and blast methods; 1.8 m bench rounds were used in the shaft. Shaft sinking was halted at depths of 15 m, 62 m, 185 m and 218 m to conduct detailed measurements of the near field hydraulic and mechanical rock-mass response to shaft excavation. Figure 1 shows the locations of the instrument arrays in the shaft and Figure 2 the layout of instrumentation at the arrays.
The URL shaft is located within the Lac du Bonnet granite batholith. Two thrust faults (referred to as fracture zones) are present between the surface and 300 m depth in the vicinity of the shaft (see Figure 1). These dip approximately 25 ° southeast. The top one (Fracture Zone #3 on Figure 1) is intersected by the shaft at 105 m depth and the second (Fracture Zone #2) lies 20 m below the shaft bottom. These fracture zones control the regional hydrogeology of the site (Davison 1984). Several splays of the Fracture Zone #2 occur in the region of the shaft and one of these intersects the shaft at 215 m depth (see Figure 1). Between the surface and the Fracture Zone #2 splay the rock is pink granite and contains a prominent vertical fracture set striking 030°; between the Fracture Zone #2 splay and the shaft bottom the rock is unfractured grey granite. The instrument arrays discussed in this paper are located in the pink granite.
ABSTRACT: A great deal of research has been directed towards the development of generalized limit equilibrium-based safety factor expressions for analyzing the stability of earth slopes. However, little attention has been paid to the implementation of advanced optimization techniques to direct the search for the critical slip surface. In order to improve the efficiency of computer-based stability analyses (and the accuracy of the solutions that these analyses yield in practice), a more advanced optimization algorithm is integrated into an existing computer program. The relative accuracy, efficiency, and reliability of the original and revised stability codes is established through extensive comparative testing involving a wide variety of realistic slope problems. The role and significance of user expertise is assessed by beginning the search at different distances from the actual critical slip surface. The new Simplex-based procedure is shown to be superior to all other search strategies currently used for computer-based slope stability analyses.
Evaluation of stability is a necessary consideration prior to any construction involving existing slopes, excavations, or man-made embankments. As such, slope stability analyses are an integral part of most geotechnical and mining projects and are therefore among the most frequent concerns of the practicing geotechnical and geological engineer.
The objective of a given slope stability analysis is to identify the most probable failure mechanism (or critical slip surface) and the corresponding minimum safety factor. Although analyses by means of empirically developed field slope charts and/or stability charts may sometimes be justified, more precise analytical solutions involving a thorough program of site exploration and laboratory testing are generally recommended.
A wide variety of specific analytical methods have been developed over the years, and it is possible to subdivide these procedures into the following four classes: (1) variational methods, (2) limit analysis techniques, (3) displacement methods (finite element methods), and (4) limit equilibrium methods. A probabilistic overlay may be used with any of these approaches, if a nondeterministic analysis is desired. Each of these approaches has its advantages and limitations. However, in practice, the overwhelming majority of stability analyses are performed using limit equilibrium techniques.
A limit-equilibrium slope stability analysis is a rather straightforward problem in optimization. As such, a limit equilibrium solution consists of the following two steps: (1) the development of a merit (or objective) function -- in this case, the safety factor expression -- to serve as a scalar measure of the stability of a particular trial slip surface; and (2) the selection of a search strategy to enable the minimum of this scalar function to be accurately, efficiently, and reliably found. During the past 30 years a great deal of research has been directed towards Step (1) above. Indeed, the development and evaluation of limit equilibrium-based safety factor expressions has been one of the primary concerns of the geotechnical engineer. A variety of restricted and generalized solutions are now available, including those of Bishop, Janbu, Lowe and Karafiath, Morgenstern and Price, Spencer, Taylor, and others.
This paper proposes a new method of in situ determination of deformation modulus and Poisson's ratio of a rock mass with U.S. Department of the Interior, Bureau of Mines hydraulic borehole pressure cells. These two geomechanics constants are determined using a combination of one cylindrical cell and two pre-encapsulated flat cells installed in a single borehole. A variety of states of stress can be generated by induced loading from mining or excavation for long-term tests, and by artificial loading imposed by flatjacks for quick tests. Required instrumentation is simple and inexpensive. Test procedures are straightforward. The applicability and practicality of the technique are demonstrated by a long-term test conducted in a longwall panel at a coal mine. The investigation reveals the stress dependency of the deformation modulus and the Poisson's ratio of this coal.
The U.S. Bureau of Mines cylindrical hydraulic borehole pressure cells (CPC's) and flat hydraulic borehole pressure cells (BPC's) were initially developed in the late 1950's and the early 1960's. The CPC was used primarily for in situ determination of the modulus of rigidity of rocks (Panek, Hornsey, and Lappi, 1964), whereas the BPC was used for measuring relative change in uniaxial rock stresses. More recently, Lu (1986) developed a new technique of using combinations of one CPC and two BPC's to measure biaxial and triaxial rock stresses, both static and dynamic. The method of determining the modulus of rigidity, Gr, with a single CPC (Panek, Hornsey, and Lappi, 1964) is realistic, but the test procedure is rather tedious. Besides, the repeated use of the same cell is difficult. Hustrulid, W. and Hustrulid, A. (1975) improved the system by using a membrane and simplified the calibration and data reduction procedures. The new system is known as the Colorado School of Mines (CSM) cell. The borehole sleeve fracturing technique (Ljunggren and Stephansson, 1986) is a further development of the CSM cell technique. However, in addition to their tediousness in calibration, test procedure, and data reduction, these techniques require a mediate relation, Er = 2 Gr (l+vr) , to obtain the deformation modulus of the rock mass, Er, on the basis of an assumed value for the Poisson's ratio, vr. Since both Er and vr are stress dependent, it is impossible to study the stress dependency of Er with these techniques. A seemingly promising technique called the corejacking test has recently been proposed by Blankenship, Amadei, and Stickney (1983). This technique can provide static in situ deformation modulus and Poisson's ratio simultaneously, but the proposed test procedure remains tedious. Other existing large-scale static tests to characterize rock mass deformability include (1) the flatjack method (Loureico-Pinto, 1986), (2) the plate-bearing test, and (3) the radial jacking test (Coulson, 1979). All three techniques, however, require an assumed value for the Poisson's ratio to calculate the value of the deformation modulus, and their test operation is complicated and expensive. Again, the stress dependency of E r cannot be studied with these three techniques.
ABSTRACT: Seismic imaging feasibility studies have been carried out in a large potash mine panel, approximately 2Km square and 1Km underground, which contains a hazardous collapse/solution structure of known dimensions. Source and receiver stations were positioned at 96 locations around the panel and tomographic images have been reconstructed for travel time data. The P wave velocity image shows a low velocity zone coinciding with the dimensions and location of the known collapse structure. Interpretation of the image is complicated by out of plane effects due to layered media. Modelling studies show how the image represents 3-D refraction artifacts superimposed on the low velocity image of the collapse structure. The complexities and problems associated with using imaging techniques in this environment are discussed and an outline for future studies is also given.
Collapse/solution structures in potash mines are hazardous and have a two fold effect on mining. If penetrated, they reduce the ore grade and can be a potential hazard to the mining operation by allowing brine/water from the overlying formations to flood the mine. Identification of these structures ahead of mining could reduce the risk of serious social and economic consequences when men are working at the face with expensive machinery. Probing ahead of the mine front using long horizontal boreholes is one method of minimising the risk of these difficulties. This method, however, does not provide information about ground conditions between boreholes. Crosshole geophysics can be utilised for this purpose and in recent years, major advances have been made using geotomographic imaging techniques. These data processing techniques allow geophysical data to be analysed so as to provide images of the rock between boreholes and inside mine structures. In principle, the approach is similar to medical CAT scanning where absorption of multiple path x-rays is used to map the density variation in the human body. In the case of geotomography, seismic (Wong et al 1983, Young et al 1987) or electromagnetic (Daily 1984, Ramirez and Daily 1985) waves which have propagated through the rock mass, are used to produce digital images which are based on velocity and/or attenuation of the signals. These images map the changing physical properties of the rock mass and allow geological interpretations to be made regarding the features imaged.
This paper describes the results of a feasibility study to image a known collapse structure. The methodology and problems associated with geotomographic imaging in layered sedimentary rocks are outlined, together with the results of model studies and the feasibility field trials.
Figure 1.(a) Plan of potash mine panel and (b) subsurface Geology.(available in full paper)
2 SEISMIC CONSIDERATIONS OF THE FIELD SITE
The potash mine panel is located at PCS Rocanville Mine, Saskatchewan, is situated 1Km below ground surface and consists of a rectangularly shaped rock mass, 1800m x 1830m in area, defined by tunnels approximately 2.5m high by 10m wide. The site, shown in Figure la, was chosen for several reasons, the most important being the presence of a large collapse structure near the panel's centre.
The elastic modulus is perhaps the single most convenient parameter for characterizing the deformability of a rock mass. This paper presents a case study of a site where the elastic modulus of the bedrock had been measured through numerous laboratory and in-situ techniques. The wealth of information acquired at the site provides an opportunity to compare the results of different test procedures and to assess some of the factors that contribute to variations in the measurements.
A major industrial development in western New York State is nearing completion. The nature and engineering characteristics of the rock formations that underlie the site have been intensively studied to a depth of almost 350 ft (107 m). The testing programs included a wide variety of both laboratory and in-situ techniques. The testing programs evolved over a period of almost 15 years as the need for new or additional data on rock mass properties became apparent and as testing procedures were refined. Although the data base was not assembled with the rigor of a research project it is reasonably comprehensive. The testing programs provided a wealth of information on the deformability of the rock mass. This paper will focus on a single parameter, the elastic (or Young's) modulus.
2 SITE DESCRIPTION
The site is located on the southern shore of Lake Ontario. It lies within the Erie-Ontario Lowlands of the Central Lowlands Physiographic Province. The soil cover consists of a relatively thin veneer of glacial tills and recent fills which have a combined thickness that seldom exceeds 12 ft (3.7 m). The bedrock surface is clean and unweathered.
The bedrock studied at the site belongs to the Oswego Sandstone and the Upper Lorraine Group, two major stratigraphic divisions of a Late Ordovician deltaic complex.
Oswego Sandstone The Oswego Sandstone is a siliceous, medium-to fine-grained, massive sandstone with minor shale and siltstone beds. The thickness of the formation increases from 35 ft (11 m) to 120 ft (37 m) across the site, a north to south distance of approximately 3000 ft (914 m). The increase occurs because of the combined effects of topographic relief and a regional dip in the bedding of approximately 0.5 degrees to the south. Upper Lorraine Group The Upper Lorraine Group is composed of two intergrading lithologic units, the Pulaski Formation and the Whetstone Gulf Formation. The Pulaski Formation has a relatively uniform thickness of approximately 100 ft (31 m). The Whetstone Gulf Formation has a thickness of at least 230 ft (70m). Both formations consist of an alternating sequence of siltstone, shale, sandstone and greywacke. The overall sequence is characterized by abrupt lithologic variations and slight gradation of one rock type into another. The bedding thickness reaches a maximum of almost 5 ft (1.5 m) near the middle of the Pulaski Formation; however, it is typically less than 1.5 ft (0.5m) and tends to decrease with depth. The distinctions between the divisions (and subdivisions) of the Upper Lorraine Group at the site are based primarily on lithologic compositions and bedding thicknesses.
Retrievability of nuclear waste from high-level geologic repositories is one of the performance objectives identified in 10CFR60 (Code of Federal Regulations, 1985). 10CFR60.111 states that the geologic repository operations area shall be designed to preserve the option of waste retrieval. In designing the repository operations area, rock mechanics considerations play a major role especially in evaluating the feasibility of retrieval operations. These considerations include: (1) the design thermal load and the thermomechanical response of the host rock, (2) the mode of waste emplacement, (3) the presence or absence of backfill in the entries and emplacement holes, and (4) the stability of openings during the retrieval period. This paper discusses generic considerations affecting retrievability as they relate to repository design, construction, and operation, with emphasis on regulatory and rock mechanics aspects.
The Nuclear Waste Policy Act of 1982 (NWPA) and the Code of Federal Regulations, 10CFR60, require that the nuclear waste geologic repository operations area be designed to preserve the option of retrieving the waste, should retrieval become necessary. Additionally, the Environmental Protection Agency (EPA) requires in their rule 40CFR191 that the disposal systems be selected so that removal of most of the waste is not precluded.
Retrieval is defined in 10CFR60.2 as "The act of intentionally removing radioactive waste from the underground location at which the waste had been previously emplaced for disposal". As required by the NRC rule (10CFR60.111(b)), the repository should be designed so that retrieval can start at any time up to 50 years after waste emplacement operations had been initiated (unless a different time period is approved by the Commission). This requirement makes the ability to retrieve waste before permanent closure an important criterion that must be incorporated into the repository design.
The NRC will require retrieval to protect public health and safety in the event the site, design, or operations proves to be unsuitable. Based on performance confirmation data, the NRC will determine whether the repository system or subsystem (natural or engineered) has failed or is expected to fail to meet the performance criteria. If such failure is determined at any time during the preclosure period, the NRC may direct the DOE to retrieve the waste. The DOE could retrieve waste for its own reasons without being directed by the NRC. Such activity, however, must be carried out under applicable NRC regulations governing movement of waste. It should be noted that the NRC retrievability provision in Part 60 is only intended to protect the public radiological health and safety.
An NRC decision to retrieve the waste will most likely be based on the results from the performance confirmation program. This program includes among other things: "in situ monitoring, laboratory and field testing, and in situ experiments", (10CFR60.140(c)). Performance confirmation will be initiated by DOE during site characterization and is likely to continue until permanent closure. Results from the performance confirmation program will ensure that geotechnical and other parameters used in the repository design are confirmed (see 10CFR60.141).
This paper discusses the results of vertical and horizontal drift convergence measurements taken during the mining of a respository sized drift in welded tuff in G-Tunnel on the Nevada Test Site. Results are quantified in terms of drift convergence magnitudes and rates that relate to drift stability.
Volcanic tuffs are being considered by the Department of Energy (DOg) as a medium for disposal of high-level radioactive wastes. The Nevada Nuclear Waste Storage Investigations (NNWSI) Project was established in 1977 to evaluate such disposal in geologic formations on or adjacent to the Nevada Test Site (NTS). Sandia National Laboratories (SNL), as one of the NNWSI participants, is responsible for the rock mechanics program to support the design of underground facilities of a radioactive-waste repository in tuff. Science Applications International Corporation (SAIC) has been under contract to SNL to prepare and install instrumentation for rock-mechanics experiments and to aid in the evaluations of the rock-mass responses. In preparation for the Exploratory Shaft (ES) investigations at the potential repository site at Yucca Mountain, SNL is conducting a rock mechanics program in G-Tunnel. G-Tunnel is located within Rainier Mesa on the NTS, where the welded tuffs are similar to those at Yucca Mountain (Zimmerman et al., 1984). One phase of the G-Tunnel rock- mechanics investigations is welded tuff mining evaluations, which emphasizes monitoring the behavior of welded tuff during the excavation of a repository-sized drift. The purposes of the evaluations in welded tuff have been (1) to document mining methods and drift convergence to provide repository designers with full-scale repository drift data to evaluate design concepts and numerical models and (2) to provide an opportunity for experimenters to evaluate and compare measurement techniques planned for use during ES testing programs.
An objective of the welded tuff mining evaluations has been to document the behavior of the rock mass during the excavation of a repository-sized drift having cross-section dimensions of 6.1 m wide, and 4 m high. As part of the evaluations, drift convergence measurement stations, located near unmined faces, were instrumented and measurements were taken while the drift was advanced in rounds, nominally 2 to 3 m long. Tape extensometer (TE) and multiple-point borehole extensometer (MPBX) measurements were used to provide vertical and horizontal displacement data useful for describing drift convergence phenomena. In this paper, the drift convergence measurements, quantified by magnitudes and rates, are described and analyzed. The results are directed toward evaluating drift convergence phenomena in terms of drift stabilities. Instrumentation and test-method evaluations are beyond the intended scope of the paper.
Drift convergence measurements can be useful in predicting long-term stabilities of drifts. The two quantities of magnitude and rate synthesize effects of geological processes, rock properties, in situ stresses, and groundwater pressure, although the latter is not a factor in G-Tunnel or at Yucca Mountain. For the work described here, the support systems were designed using empirical rock mass classification systems (Barton et al., 1974 and Bieniawski, 1976). Once openings are mined then observational techniques using measurements aid in evaluating rock behavior.
Engineering properties for soft rocks with zero to low Rock Quality Designation (RQD) were evaluated and established for design of large drilled piers with diameters varying from 1.8 to 3.6 m in order to support transmission structures. Numerous field pressure meter and rock anchor pullout test results, and empirical correlations were used.
Extensive geotechnical investigation for a transmission facility project in New York has recently been performed. Complex geological conditions within the investigation area (Figure 1) were encountered. 'Rock cores obtained from the boring program indicate that the top 6 m rocks at most of the locations are weathered and have Rock Quality Designation (RQD) below 20 percent. The RQD is based on a modified core recovery procedure, which is based indirectly on the number of fractures and the amount of softening or alteration in the rock mass as observed in the rock cores from a drill hole. Instead of counting the fractures, an indirect measure is obtained by dividing the total length of hard and sound core pieces which are 10.2 cm or greater in length by the total length of the core run. At those locations, the rocks are commonly found to be shale, siltstone, claystone, and sometimes interbedded with thin layers of sandstone or limestone. Large drilled piers (I.8 to 3.6 m diameter) with rock sockets and rock anchors were designed to support the structures to resist heavy wind moments, uplift and lateral forces. In order to adequately design the foundations embedded in rocks and having low RQD, numerous pressure meter tests were performed at different elevations in each selected borehole along the transmission route. Many rock anchor pull-out tests were also performed to test bond strengths among anchor, epoxy grout and rocks. This paper will detail the interpretation of the test results, the analysis, and the basis of the establishment of the engineering properties of the soft rocks for the design of the transmission foundations within a complex geological region.
2 CEOLOCIC SETTINGS
The area of southeastern New York State shares many of the formational and structural features that characterize the geology of eastern North America. The rock succession includes all of the common kinds and types of sedimentation. Plutonic and dike rocks are also present in the eastern region. Glacial deposits of Pleistocene-age sand, gravel, clay and till mantle much of the investigated area (Newland, 1933). The rocks in the north end of the area, which are of the Ordovician, Silurian and Devonian age, consist of shale, siltstone, sandstone, limestone and dolostone. The Catskill Mountains in the middle of the area rise about 610 m above the adjacent parts of the Allegheny Plateau. The mountains consist of Medial Devonian rock of shale, siltstone and sandstone. The rocks at the south end are slightly metamorphosed and recrystallized, and are of Cambrian-Ordovician age. They are also located at the Northern edge of the Hudson Highlands. They consist of limestone, folded and faulted shales, slate, sandstone and mudstone. Subsurface conditions within this investigated area have been obtained through an extensive geotechnical investigation program (G/C, 1984, 1986a and b).
The next phase of the Nevada Nuclear Waste Storage Investigations (NNWSI) Project is the characterization of Yucca Mountain, a possible site for a high-level nuclear-waste repository. Site characterization will include experimental programs in many technical fields, including rock mechanics. Because the data gathered during site characterization will be used to design a repository and assess its performance with respect to regulatory requirements, the planning for a program in experimental rock mechanics is necessarily complex. The planning approach includes performance allocation, an interactive process by which the data needs for design and performance assessment of the repository are used to determine the direction and magnitude of the experimental programs.
As a part of the activities that have resulted from the Nuclear Waste Policy Act of 1982, Yucca Mountain, Nevada is being studied as a possible site for a repository for high-level nuclear waste. Partly on the basis of explorations of the site, the U.S. Department of Energy (DOE) recently (May 1986) selected Yucca Mountain as one of three candidate sites to be fully characterized. The experiments and tests conducted during site characterization will provide the data required for the design and analysis of the repository and for the assessment of the performance of the repository with respect to the licensing requirements in the federal regulations.
Although the interpretation of design and performance-assessment data needs in rock mechanics is necessarily complex (see section 2), some primary licensing concerns are evident from the regulations. The repository must be designed so that openings (shafts, ramps, drifts, and waste-emplacement boreholes) can be constructed and maintained using readily available technology. Because the option to retrieve the waste must be maintained up to closure of the repository, the expected stability of openings must be consistent with the planned methods and period of retrieval (over 80 years). During the post-closure period, the rock-mechanics concerns are in two areas: 1) the behavior of the rock around waste-emplacement boreholes as it affects containment of waste in waste packages and 2) the creation of preferential pathways (e.g., continuous large fracture systems) for water infiltration or the release of radionuclides.
Studies of the rock characteristics for a Yucca Mountain repository have been ongoing for several years (e.g., Tillerson and Nimick (1984)). The present repository plans provide for underground storage of the waste in a unit of fractured, densely-welded volcanic tuff above the water table. Samples of tuffs from outcrops and exploratory drillholes have been studied in the laboratory to obtain preliminary estimates of mechanical and thermal properties. Because an under- ground facility has not been available at Yucca Mountain, in situ testing has been conducted in G-Tunnel on the Nevada Test Site, about 40 miles from Yucca Mountain (see Zimmermann and Finley (1987) for a recent summary). The testing location within G-Tunnel contains welded and nonwelded tuff beds that, based on laboratory tests on small samples, are mechanically and thermally similar to those at Yucca Mountain.
Analysed on the strata displacement measured data collected from c311ieries, it can be found that when the hanging section of main roof is caving, a part of the fixed section in the abutment place would be appeared in rebound condition. It is clear that this manifestation may be used as an index to predict the variation of the roof pressure in longwall coal mining in order to explain this manifestation, a model as a semi-infinite long beam to be clamped on a Winkler elastic foundation has been put forward in this paper. Besides, the position of the fractured surface and the rebound district of the main roof also had been discussed in this paper.
In coal mining, especially in longwall mining, the fracture of the main roof would affect the distribution of the abutment pressure ahead of the longwall face and following the fracture of main roof a series of phenomena would be happened around face area. For instance, the convergence in the working area and the load acting on the support would be increased and sometimes some kinetic phenomena may be happened. Because of this, the mining engineers and researchers in China always paid special attention to this problem.
2 MECHANICAL MODEL OF MAIN ROOF BEFORE AND AFTER ITS FRACTURE
In according of the relationship between length of the longwall face, thickness of the main roof and its span of first weighting the main roof hanging over the caved area can be suggested as a "plate" fixed (or simpled) to the boundary pillar and coal face. By the experience in the laboratory, the fracture form of these "plate" can be described in Fig. 1. Fig. la shows the fracture form when the span of first weight (L) is less than the length of coal face (1). Fig.1b indicates that both of length is approximately same. Fig.1c shows when L is much longer than 1.
It is particularly interesting to notice the process of the plate collapse. The first crack is occurred on the top of the plate above the middle of the longer supported line and then the crack is extended along the longer line. In the meantime the cracks will be appeared on the top along the shorter boundary line. At the end the boundary cracks would be connected each other and formed as an ellipse shown in Fig. 1. Following this, the plate will collapse as a "X" form.
Fig. 1 (available in full paper)
Fig. 2 (available in full paper)
After further removal of the coal face, the "plate" will collapse periodically and its form is indicated in Fig.2.
According to the measured data in situ the behaviour of the overlying strata across the middle of the face line can be represented in Fig. 3. Clearly, this case is just for the condition Fig.1a.
Fig. 3 (available in full paper)
From Fig.3 a mechanical model as a semi-infinite long beam to be clamped on a Winkler elastic base can be suggested as illustrated in Fig. 4.