A rockburst is defined as sudden and violent failure of rock. During the failure process, excess energy is liberated as kinetic (or seismic) energy, which causes the surrounding rock mass to vibrate. It is these vibrations that are felt by persons underground and on surface. The magnitude of a rockburst is proportional to the amplitude of the vibrations and is measured in the same way as an earthquake, using a modified Richter scale. Rockbursts are a 20th century phenomenon, the direct result of improved mining technology (especially in drilling, blasting, pumping, hoisting and ventilation), which has allowed the mines to go deeper into an ever-increasing stress environment. In Canada, the first recorded rockbursts were in the late 1920s at the gold mines in Kirkland Lake, Ont., and some of the nickel mines in Sudbury, Ont. By 1940, rockburst incidents had increased so dramatically at these two mining camps that the Ontario Mining Association appointed R. G. K. Morrison (later of McGill University) to investigate and report on the problem. Morrison’s report in 1942, in many respects, formed the basis for implementing rockburst control strategies at Ontario mines for the next 40 years. His concepts and strategies are discussed later.
In the 1960s, the gold mines at Red Lake experienced rockbursts and, in the early 1980s, so did the uranium mines at Elliot Lake. Interestingly, the mines in the Timmins area, although of the same age and depth as those in Kirkland Lake, are not rockburst- prone. During the 1980s, there has been a large increase in seismic activity in mines across Ontario.
Between 1984 and 1987 14 mines in Ontario experienced 281 large seismic events. The Quirke mine at Elliot Lake dominates the statistics and accounts for nearly half these events. (It should be noted that most of these events occurred in 1984 and 1985, and the mine has been relatively quiet since then.) Other mines with significant seismic activity are Campbell Red Lake mine at Red Lake, Ont., the Strathcona, Creighton and Copper Cliff North mines in Sudbury and the Macassa mine at Kirkland Lake. A variety of ore deposits and mining methods are used in these mines. At Red Lake and Kirkland Lake, narrow, steeply dipping vein deposits are mined by shrinkage and cut-and-fill methods. Gently dipping reef deposits at Elliot Lake are mined by room- and-pillar methods. At Sudbury, massive sulphide deposits are mined by cut-and-fill and blasthole methods.
In response to this growing rockburst problem, the Canada/Ontario/ Industry Rockburst Project was initiated in 1985. Management and funding of the project, over an initial 5-year period, is on a tripartite footing. The government of Canada, through the Canada Centre for Mineral and Energy Technology (canmet), provides the staff to operate the project. The government of Ontario, through the Ministries of Labor and Northern Development and Mines, provides funds for equipment and services. The Ontario mining industry, through Campbell Red Lake Mines, Denison Mines, Falconbridge Ltd., Inco Ltd., lac Minerals and Rio Algom, contributes its existing monitoring systems, provides field sites, assists in installation of new equipment and provides data on rockbursts at the member company mines. The rationale and objectives of the project are first to investigate the causes and mechanisms of rockbursts using existing and new seismic monitoring equipment, and then to develop techniques to alleviate or control the damage from rockbursts.
Mines in other parts of Canada have also experienced rockbursts, but to a lesser extent than those in Ontario. So-called mining-induced earthquakes have occurred above the potash mines in Saskatchewan over the past 10 years. Rockbursts have been reported at some of the gold mines in the Val d’Or area of Quebec. The lead/zinc mines in New Brunswick are rockburst-prone. In 1958, a rockburst (called a bump) at the Springhill Colliery in Nova Scotia resulted in 75 fatalities. Finally, in Newfoundland a fluorspar mine experienced rockbursts at a depth of only 150 m. Also, rockbursts have recently been recorded at the Buchans mine a few years after it was closed. In general, deep hardrock mines experience most of the rockburst problems. However, depth and hardness are not the sole criteria that one should go by, as is evident from the preceding examples.
Since rockbursts are the result of a violent release of energy, an analysis of energy helps explain them. When an underground excavation is enlarged, either from mining or rock failure, the surrounding rock mass moves towards the excavation, resulting in a change in potential energy (Wt). The rock removed during enlargement also contains stored energy (Um). These two components (Wt + Um) represent the energy entering the mining operation as the result of the enlargement. This energy has to be dissipated somehow. Stresses acting on the rock that was removed are transferred to the surrounding rock mass, increasing its stored strain energy (Uc). If the excavations are internally supported by backfill, cribs and the like, then some energy is absorbed in deforming the support (Ws). Any excess energy is normally referred to as released energy (Wr). From the law of conservation of energy: Wt + Um = Uc + Ws + Wr
There are a number of ways in which energy can be released. The stored strain energy in the removed rock is obviously released. If this rock fails rather than being mined, then this energy component is consumed in the fracturing process. If the rock is removed or fails instantaneously, oscillations occur in the rock mass. Equilibrium is attained through damping, and seismic energy (Wk) is dissipated in the process. There are no alternatives, hence: Wr = Um + Wk
It is the seismic energy (Wk) that is recorded by mine microseismic systems, and it is this energy component which is responsible for the damage caused by a rockburst. From Equations 1 and 2: Wk = Wt – (Uc + Ws)
This equation indicates that, to reduce the seismic energy, the change in potential energy (Wt) has to be reduced or the energy absorbed by the support system (Ws) increased (there is no control over Uc). The former can be achieved by reducing the convergence of the rock mass and the latter by increasing the stiffness of the support (backfill).
To initiate a rockburst, part of the rock mass must be at the point of unstable equilibrium because changing stresses are driving a volume of rock to sudden failure, or a system of pillars is approaching a state of imminent collapse, or geological weakness planes are on the point of slipping. These three categories can be conveniently labelled strain, pillar and fault-slip bursts, which are familiar terminology in mining. Other conditions are that a change in stress is required to trigger a rockburst. This can be either an increase or decrease in stress depending on the type of rockburst. To initiate stress waves, an appreciable stress change must accompany the rockburst. Finally, a substantial amount of energy must be available to provide the source of the seismic energy. This reservoir can be either the stored strain energy in the surrounding rock mass or a sudden change in potential energy. Strain Bursts
Strain bursts at the edge of mine openings are caused by highstress concentrations that exceed the strength of the rock. Events can range from small slivers of rock being ejected from the walls to collapse of a complete wall as it tries to achieve a more stable shape. These types of rockbursts are normally associated with development drifts, including shafts.
Energy can be released from a number of sources. If the rock goes from a triaxial to a biaxial or uniaxial stress condition, some of the stored strain energy is released as seismic energy. Instantaneous failure of this rock will enlarge the opening and seismic energy will be released because of the elastic reactions of the rock mass. Finally, if brittle and soft rocks are present, minor slippage could occur along the contact. Pillar Bursts
Severe rockbursts, involving thousands of tonnes, have been caused by the comple
te collapse of support pillars. In some cases, the collapse of one pillar can overstress adjacent pillars and a chain-type reaction ensues. In recent times, the most significant chain reaction occurred in an old stope-and-pillar area of the Quirke mine at Elliot Lake. Over a four-year period, a zone of pillar failure gradually increased over an area of about 1,000 m on strike by 500 m on dip. The end result of this seismic activity and pillar failures was that the hangingwall could no longer span the affected area, and fracturing progressed to surface (more than 500 m) with an increase of water flow into the mine. With this fracturing of the hangingwall, rockburst incidents decreased dramatically.
Significant pillar bursts have also occurred in steeply dipping, vein-type orebodies at Red Lake and Kirkland Lake. These normally occur when sill/crown pillars of shrinkage or cut- and-fill stopes reach a critical size. To understand the mechanics of pillar bursting, it is necessary to understand the concept of pillar stiffness and loading stiffness. Figure 2a) shows the simple example of a single pillar between two stopes. Figure 2c) shows the stress displacement history of the pillar. As the stress increases, the curve has a positive slope up to peak strength. After peak strength, displacement continues, but the load decreases and the slope is negative. The shape of this unloading curve depends on the type of rock. Brittle rocks have much steeper unloading curves than soft rocks.
Loading stiffness can be explained by replacing the pillar with a hydraulic jack that exerts the same load as the original pillar, as shown in Figure 2b). The loading stiffness at that location would be the unloading curve for the jack as the hydraulic pressure is released. The violence of pillar failure depends on the difference between these unloading characteristics. If the post-failure pillar unloading curve is steeper than the loading stiffness curve (that is, soft loading), as illustrated in Figure 2c), then there is a surplus of energy and failure will be sudden and violent. However, if the loading stiffness curve is steeper than the post-failure pillar curve, failure will be gradual and non-violent.
It is of interest to note the areas under the various curves since these represent the energy components. The area under the pillar curve (oab) represents the energy consumed in the fracturing process. It is made up of two components: the stored strain energy at peak strength (um) and part of the energy stored in the surrounding rock (us), or the loading system. The area between the two unloading curves represents the seismic energy (wk) that is liberated. The violence of pillar rockbursts does not come from the energy stored in the pillar but from the energy in the loading system. Fault-Slip Bursts
Slippage along a fault has long been recognized as the mechanism of an earthquake. Only recently has the same mechanism been recognized as the cause of some rockbursts in Canadian hardrock mines, especially those in Sudbury. Shear stresses act parallel to a fault or dike. Slippage is prevented so long as the clamping forces exceed the shear stress. Clamping forces are controlled by the stress acting perpendicular to the fault and the coefficient of friction on the fault. Slippage can be initiated by either an increase in the shear stress or a decrease in the perpendicular stress or coefficient of friction. Once slippage occurs, the lower dynamic coefficient of friction comes into effect, resulting in a drop in stress along the fault. Theoretical models indicate that only minor slippage and stress drops are required to initiate significant rockbursts. For instance, a rockburst of magnitude 3.0 would result from a stress drop of 3.5 mpa with an average slippage of 12 mm over a radius of 100 m.
In most cases, the damage caused by fault-slip rockbursts is minimal. There is one example of a 2.2 magnitude rockburst (at the Falconbridge mine near Sudbury) where no damage was found, although 10 to 20 mm of slippage could be observed on the fault. Normally, what damage is observed is away from the fault where the radiated seismic energy has triggered a critically loaded structure. In one case, a magnitude 3.4 rockburst caused a backfill mat to collapse in an undercut-and-fill stope some 20 m away. Alleviation and Control of Rockbursts
Two approaches can alleviate rockbursts: the strategic and tactical approach. The strategic approach is to diminish the possibility of encountering rockburst-prone ground or to reduce the severity of the rockbursts. Techniques include sequencing of extraction to minimize large energy releases or the use of backfill both to limit closure and to absorb energy otherwise liberated as seismic energy. The benefits of these techniques are only realized in the long term. The tactical approach is to accept that some rockbursting is inevitable, but to seek to limit the extent of the damage. Techniques include design of support systems that yield with the vibrations rather than snap, and destress blasting to soften the rock and control the timing of the change in potential energy. The benefits of these techniques are realized in the short term. Strategic Methods
Rockbursts are generally the result of mine planning decisions made a number of years previously. Problems are usually not encountered during initial mining but some years later when a more extensive area has been mined out. Morrison’s report in 1942 dealt with strategic methods for rockburst control. Rockbursts were attributed to the formation of “domes,” which are the fractured zones surrounding individual stopes. As stopes approach each other, the intervening pillars become increasingly stressed. If these pillars suddenly fail, the volume of the dome also suddenly increases. The rupture of the large volume of rock between two or more initial domes and the release of its accumulated energy resulted in rockbursts. The magnitude of the rockburst depended on the final size of the dome and the energy stored within it. This, in turn, was controlled by the areal extent of stoping, the depth of the working area and the physical properties of the wall rocks. Subsequently, the “doming theory” was discredited when it was shown theoretically that the size of the domes decreases with depth. However, many of the characteristics associated with domes are still relevant to rockburst control strategies. Stope span, depth and elastic properties of the wall rocks also control convergence and the change in potential energy of the surrounding rock mass. This is now recognized as being the driving force behind all rockbursts.
Morrison’s main rockburst control strategy was the use of mining layouts that eliminated small remnant pillars and allowed domes to gradually expand. In practice, this meant some form of longwall configuration. The effect of weakness planes, such as faults or dikes, was explained in terms of their effect on the formation of domes and stress concentrations. The concept of slippage along these structures was not recognized. It was recommended to mine away from major weakness planes or to mine through them at a perpendicular angle rather than an acute angle. The importance of support, such as backfill, for controlling the number and severity of rockbursts was recognized, although, again, in terms of its effect on the size of domes.
It is now known that the change in potential energy of the rock mass is the driving force behind a rockburst. Potential energy is basically the pre- mining stress multiplied by the volumetric closure of a stope. We have no control over the stress, but some limited control over closure. Systematic stabilizing pillars would limit stope span and reduce the closure. However, in practice this means writing off about 20% of an orebody. This is resorted to only when the rockburst problem seems insoluble (as in South African gold mines). Backfill and, especially, stiff backfill will reduce the closure in a stope and, as indicated in Equation 3, absorb energy that otherwise would be released as seismic energy. Backfill is most beneficial in thin, t
abular deposits.
Finally, we can try to control the rate at which energy is released. Many mines can extract about 80% of the ore reserves without undue rockburst problems. It is the last 20% that causes all the problems. The last 20% is in the form of pillars that are holding back the regional closure. As these pillars are mined, large changes in potential energy occur. If longwall methods are used from the beginning of mining, then the rate of energy release is more uniform throughout the life of the mine. Tactical Methods
At some mines, below a depth of about 1,000 m, the mere fact of making a mine opening produces high stress concentrations and minor rockbursts. These problems increase in the stoping area. Although strategic methods are still important in reducing the severity of rockbursts, they will not eliminate the problem. At this point, tactical methods can be used to protect the workforce. When a rockburst occurs, a stress wave radiates out from the source. When this stress impulse hits a rigid support such as mechanical bolts, it can cause violent failure and throw the loose rock into the mine opening. This characteristic of mechnanical bolts has been observed in many mines in Ontario. Grouted rebar is also a rigid support system but, being stronger than a mechanical bolt, it takes a large stress wave to fail it. Experience in some mines indicates that friction-type support (split sets and Swellex, for example) in conjunction with wire mesh can withstand rockbursts of up to about 2.5 magnitude. The mesh contains the broken rock. For extreme conditions, the South African gold mines use a lacing support system. This consists of smooth, mild steel rebar grouted in boreholes with a “shepherd’s crook” at the collar. Wire mesh is placed against the rock and then steel cable is threaded through the “shepherd’s crook” in a diamond pattern. This type of support system has withstood a rockburst of magnitude 4.0. However, it is very expensive. Destress blasting is used in many mines in North America, in either development drifts or crown pillars of cut-and-fill stopes, in order to change the potential energy of the surrounding rock mass. This is achieved by fracturing the rock to soften it, which allows the walls to converge. However, there are potential problems with this method. One of the basic laws of rock mechanics is that stress can not be got rid of, only transferred. Consequently, destress blasting transfers stress to adjacent structures that could then burst. At some mines, rockbursts occur within minutes or hours of a production or destress blast. These mines invariably have central blasting at fixed times, with no one underground. Hence, it may be possible to choose the time of a rockburst. Mines using blasthole stoping methods tend to set off their large production blasts after the last shift on Friday, leaving the mine two days to settle down. Rockburst Prediction
There are two aspects to rockburst prediction: location and time. In some cases, it is possible to identify rockburst-prone areas of a mine based on microseismic activity or computer models that give stress distributions. Prediction of time is much more elusive. Earthquakes have been taking place over a much longer time than have rockbursts, with still no adequate method of prediction. In some cases, a build-up in microseismic activity precedes pillar bursts. However, it very much depends on the time scale examined. Only minutes or hours before a burst, trends usually do not appear. However, the microseismic record over the preceding months could indicate a gradual build-up in activity. This, in fact, did happen at one mine that experienced rockbursts. Rockburst Monitoring
Three types of seismic monitoring systems are being deployed to monitor a large range of seismic events. The Eastern Canada Seismic Network is operated by the Geophysics Division of the Geological Survey of Canada. Seismometers are located across Ontario, Quebec and the Maritimes to monitor naturally occurring earthquakes. These sensors also record the larger rockbursts, and magnitude values can be calculated. In Ontario, the existing seismic network has been augmented by additional seismograph stations at Sudbury, Elliot Lake, Red Lake and Kirkland Lake. These stations primarily provide magnitude values for rockbursts down to possibly 1.0. A secondary purpose is to have permanent records of the larger seismic events.
The most sophisticated new network is in the Sudbury Basin. Three seismograph stations are located on the north, south and west rim of the Basin. This allows triangulation for locating hitherto unlocatable rockbursts. The signals for each seismometer are digitized at the sensor and are continuously transmitted over dedicated phone lines to Science North. Here, the signals are recorded on drum recorders, which are on public display. Also, the signals are continuously transmitted over a data line to the Geophysics Division in Ottawa, where the data are analysed and the magnitude values calculated. Macroseismic monitoring systems are designed to capture the complete waveforms of the larger seismic events. Strong-motion triaxial geophones are installed 500 m to 1,000 m away from active mining so as not to saturate the sensors. Five geophones per system are installed on surface, underground or both. Analysis of wave forms should provide additional information on the mechanism of rockbursts. The direction of first motion should indicate the type of rockburst. For instance, pillar bursts are implosions and the first motion is inward, whereas a blast is an explosion and the motion is outward. For fault-slip bursts, opposite sides of the fault move in opposite directions. Peak particle velocity is a measure of the damage caused by a rockburst similar to that in blasting. It is used to design support systems so that they can withstand certain levels of rockbursting. Integration of the waveforms gives the seismic energy liberated in a burst, which is fundamental in understanding the energy balance. Analysis of the signal frequencies has been used in seismology to define the main rockburst parameters, such as radius, slippage and stress drop.
Five microseismic systems will be installed. The first two are already operating at Quirke and at Falconbridge’s Strathcona mine. Other systems will be installed at Inco’s Creighton mine, the Campbell Red Lake mine and Lac Mineral’s Macassa mine. Microseismic monitoring systems are owned and operated by the mining companies. Currently, 14 systems, all made by Electrolab, have been installed in Ontario mines. Additional systems are located in a potash mine in Saskatchewan and a lead/zinc mine in New Brunswick. These systems use up to 64 extremely sensitive geophones installed underground. Their prime function is to locate the source of a seismic event within a few seconds of its occurrence. An indication of the relative magnitude is also obtained.
The mining companies have put a lot of effort into software development for these systems. Digitized mine plans and sections are produced showing location of events. In many cases, these are available to production personnel before a shift. These systems have significantly influenced mine planning. Dr David G. F. Hedley is Senior Research Scientist, Elliot Lake Laboratory, CANMET, Energy, Mines and Resources Canada. This paper was presented earlier this year at a seminar in Val d’Or, Que. — 30 —
Be the first to comment on "A ROCKBURST PRIMER"