Studies estimate that 156 billion tons of coal, representing 68 pct of the minable reserves in the United States, is subject to multiple-seam mining. Historically, room-and-pillar mining has dominated coal production, and this mining method has been the focus of most multiple-seam research. Advances in Longwall Mining Method have made this system more economically attractive because of its efficiency and production potential. The high productivity being achieved by longwall mining demonstrates its potential for being a substantial segment of underground coal production. Longwall mines produce more than 30 pct of all underground coal, up from 5 pct just 15 years ago. The continued growth of longwall mining without appropriate multiple-seam planning may increase the cost and risk of mining. Optimization of the mine design factors is arguably the primary means for controlling interactions between operations. The U.S. Bureau of Mines, in an effort to improve long- wall planning, is investigating multiple-seam longwall design and systems development.

Significant advances have been made in longwall design for single seams in the areas of gate road pillar design, panel layout, powered support selection, and roof control. Multiple-seam designs are less developed, but progress is being made as more longwall operators gain experience. In research, empirical and analytical in nature, the mechanics of stress transfer associated with multiple-seam longwall operations has been examined. These studies indicate that longwall interactions can extend over greater vertical distances than room-and-pillar interactions do. This may be related to the large abutment stresses produced during panel extraction, the arching stresses associated with deep, wide openings, and the caving characteristics of the strata.

Longwall mining of multiple seams has dominated European coal production for decades. In the United Kingdom, the intent is to maximize the gob and to avoid leaving pillars between panels; leaving pillars between panels is known to create stress problems in subsequent workings both above and below. Single entries are utilized in most situations. The mines also cooperate in sequencing seams and maintain a rigid phasing of panels to control interactions. These planning efforts have been successful from a ground control perspective, but problems related to cost and productivity can arise. Single-entry systems entail costly support requirements, and strict schedules for extracting panels can create production delays. Interestingly, mine planners in the United Kingdom recently have been adopting some U.S. practices. Realizing the cost benefits of retreat longwall mining, planners are attempting to incorporate this method into their multiple-seam designs. These retreat longwall designs are similar to those used in the United States, but further study is required to assess their performance. The primary concern is the different geologic settings in the United States and the United Kingdom and the behavior of overburden and interburden strata. Consequently, interactive magnitude and distance may differ although mining geometries are similar.

Eliminating interactions between operations is a difficult task because of the complex relationship between mine design and geology. The use of numerical methods for predicting interactive problems is receiving more research attention for application as a design and planning tool. Using numerical methods, combined with case study results and theoretical and statistical analyses, engineers can attempt to develop optimum mining plans for different multiple-seam conditions. MULSIM/NL is a boundary element model, developed by the USBM, for calculating stresses and displacements in tabular deposits. The model was used in this study to evaluate stress distribution and transfer for design problems that are commonly encountered in multiple-seam longwall layouts.

There are three primary design factors to consider for longwall mining multiple seams. These factors are very closely related and should be weighed equally for effective mine planning. These factors are as follows: first, the sequence or order in which the seams will be mined, which will determine the type of interaction; second, the design of the gate road pillars, which will define the magnitude of the interaction; and third, the layout of the gate roads and longwall panels, which will define the location of the interaction. Other parameters fixed by the geologic environment, such as depth and interburden thickness, will influence interaction magnitude and location and must also be considered in the design process.

In this report, the three primary design factors will be discussed in detail. Theories that describe stress transfer between multiple seams and the mechanics of interaction will be reviewed. The geological effects on mine design will also be addressed. Finally, practices for effective multiple-seam longwall design together with MULSIM/ NL model results will be presented. This information will assist longwall operators in planning safe and productive multiple-seam mines.

The 1991 Longwall Census from Coal Magazine shows there are 78 mines with longwall faces in the United States. Research by the USBM indicates that 25 mines (32 pct of the total number of mines) have mining in adjacent coalbeds, either above or below. As shown in table 1, about half of these mines report some type of interaction problem with the adjacent workings. Of the 25 mines, 16 have longwall workings in adjacent coalbeds. Layouts for gate roads and longwall panels in these mines show that four mines are superpositioning; four mines are offsetting; and the remaining eight mines slightly offset the gate roads and panels because of changes in panel widths and lengths. Nine mines have room-and-pillar workings in adjacent coalbeds. All mines attempt to lay out the panels to avoid problem areas, but a more determining factor in design is maximizing coal recovery from the property.

Seventeen of the twenty-five mines are classified as pillar load transfer interactions where the longwall operation is mining beneath an overlying mine. Two mines are superpositioning gate roads, 2 mines are offsetting gate roads, and the remaining 13 have either overlying longwall or room-and-pillar workings that can be classified as either slightly offset gate roads or randomly arranged in the case of an overlying room-and-pillar operation. A common factor in mines encountering load transfer problems is an isolated barrier pillar or a gob-solid coal boundary left in the upper mine. These features usually create stress concentrations in the gate entries and on the longwall face. Eight of the seventeen mines report experiencing ground problems because of this condition. In all but one case, the interburden between the mines is 200 ft or less and the overburden above the upper mine ranges from 800 to 2,000 ft. Three mines report severe interaction problems, but in the remaining cases, problems are moderate and manageable.

The remaining eight mines are classified as subsidence interactions where the longwall operation is mining above an underlying mine. Two mines are superpositioning the gate roads, two mines are offsetting the gate roads, and the remaining four have slightly offset gate roads or a room-and-pillar mine underlying the panels. Four of the eight mines report subsidence-related ground problems, and in all these cases the interburden between the mines is less than 200 ft and the mining height in the lower seam is 60 in or greater. Fractured strata leading to roof control problems in setup and recovery rooms and in gate entries is a common problem. These problems were usually experienced when developing across a gob-solid coal boundary in the lower mine where the subsidence trough is located.

Current trends in longwall technology to design longer and wider panels will affect the layout of mine geometries in the future. The exact superpositioning of gate roads or offsetting gate roads at the center of the overlying or underlying panel is impossible as longwall panels become wider in successive seams. Therefore, a slight offset, where the panel is positioned toward the headgate or tailgate side of the overlying and underlying panel, may become a frequent practice. As longwall panels become longer, panels and gate roads will be influenced by barriers, gob-solid coal boundaries, and other developments in overlying and underlying operations. Crossing barriers and gob-solid coal boundaries contributes to most problems, and examples of this in the field are already evident. Mine planning and coordination should be a primary consideration to minimize interactions between multiple-seam longwall layouts.

Sequencing of seams and the effects of interaction on other minable coalbeds also deserve more attention from both a health and safety and an economic standpoint. Studies indicate that interaction problems associated with subsidence may increase in the future. This is primarily true in the Appalachian Coal Region, where the better quality coals that can be longwall mined occur at lower depths. In the Appalachian Coal Region, 65 pct of the active longwall faces are located in just five different coalbeds: the Pittsburgh, Eagle, No. 2 Gas, Pocohontas No. 3, and Blue Creek. These coalbeds occur at depths ranging from 500 to 2,000 ft. As a result, a large reserve base of high-quality, low-sulfur coals is being subjected to caving from these operations. Many of the lower sulfur coalbeds are located in the Southern Appalachian Coal Region, and coal analysts speculate that the new Clean Air Act and compliance coal standards may shift future mining to these reserves. In this region, both room- and-pillar and longwall mining have contributed to coal production, but seam sequencing is still based mostly on availability and ownership. This practice has and will continue to have an effect on the minability of other coalbeds. Potentially, coalbeds that once were considered minable reserves may not be, because of interactions from other operations. To avoid higher mining costs and coal prices, mine planners should focus on adopting practices and procedures that prevent and control interactions in multiple seams.

Longwall Mining Design Factors

Seam Sequencing

The sequence in which the seams will be mined is the first fundamental decision confronting mine planners. In the past, coalbeds were mined in no particular order with regard to controlling interactions and reducing ground control problems. Seam sequencing was based mostly on economics, availability, and ownership. Unfortunately, this practice still continues today in many instances. For long-wall mining, there are four possible mining sequences involving multiple seams :

1. Seams are longwall mined in a descending order, with mining completed in the upper seams before any mining is initiated in the lower seams (fig. 1).
2. Seams are longwall mined in an ascending order, with mining completed in the lower seams before any mining is initiated in the upper seams (fig. 2).
3. Seams are longwall mined simultaneously, and mine plans may or may not be coordinated with one another (fig. 3).
4. Seams are longwall mined randomly (fig. 4).

The mining sequence determines the type of interaction the longwall developments will experience. Discussed later are the primary types of interaction, and the supporting theories that describe the mechanism of interaction, with special reference to geological considerations.

Interaction for Descending Order

A descending order of extraction is considered the most preferable practice for optimum control over seam interaction. Seams sequenced in this order are impacted by stress transferred from overlying gate road pillars, gob- solid coal boundaries, and isolated barriers. Referred to as pillar load transfer, this interaction can usually be predicted with reasonable accuracy, and design changes can be implemented to minimize damage to underlying operations. The mechanics of stress transfer between workings has been analyzed extensively through case study documentation and the use of mathematical and photoelastic models. Two theories have been developed to explain interactions due to load transfer from overlying workings: “pressure bulb” theory and “arching” theory.

Pressure Bulb Theory

In pressure bulb theory, the pillar is assumed to be the major structural element in the transfer of stress. The pressure bulbs are best represented as contour lines of stress as shown in Figure 5. In this simplified example developed by Peng, the pillars are superpositioned and the weight of the vertical load is equally shared by neighboring pillars using the tributary area method. The highest stress occurs near the top and bottom of the pillar, decreasing vertically to zero influence at a distance approximately four times the pillar width. The vertical

stress experienced by the pillars is equal to the sum of the contour lines, but because of pillar superpositioning, this analysis is also extended to include additional stress from neighboring pillars. This analysis was developed for perfectly elastic, isotropic, and homogeneous materials. This approach can be used to estimate the load transfer to underlying pillars when gate roads are superimposed.

Haycocks used photoelastic models to further investigate the pressure bulb concept under anisotropic conditions representative of various geological conditions. The models simulated stress transfer and dissipation as a function of three major variables: pillar geometry and loading, interburden layering (stratification), and interburden elastic modulus. The effects of depth and pillar geometries were studied using pillars of varying widths loaded with stress profiles ranging from uniform to several types of peak-trough profiles. Interburden stratigraphy was simulated using layered materials of varied thickness and elastic properties. This research showed that:

1. The distribution of stress on the pillar (stress profile) affects both the distance and the magnitude of the load transfer.
2. Peak-trough stress profiles dissipate stress with less influence than uniform stress profiles do.
3. High modulus layering of the interburden tends to inhibit pressure bulb formation, while low modulus layering increases the vertical and horizontal distances stress can be transferred from overlying pillars.

Haycocks concluded that pressure bulb theory is useful in analyzing pillar load transfer when a “passive” interaction occurs. This condition occurs when gate road pillars are superpositioned and the lower seam pillars are sufficiently large to prevent them from yielding.

Su studied pillar load transfer mechanisms with the aid of finite element modeling. This work led to the following conclusions:

1. Pillar shape influences the transfer of stress: A rectangular pillar will transfer less load, and also, the interactive distance will be less than would be the case for a square pillar of equal load-bearing capacity.
2. The elastic modulus of the coal pillar has a negligible effect on the transfer of load. In situ horizontal stresses also have a negligible effect on the downward transfer of stress.
3. Strata inclination will not distort the pressure bulb contours below a large pillar. However, as pillar size decreases the contours will be increasingly distorted under the same strata inclination.
4. The absolute values of the pressure bulb contours are proportional to overburden, and interactions will be-come more severe as depth increases.

The important geologic factors that influence pressure bulb formation include depth and the thickness and stratigraphy of the interburden. The significance of depth in load transfer mechanics is apparent, as stress increases with depth. The physical characteristics of the interburden are equally important, as the magnitude and distance of load transfer is largely dependent upon the thickness, stratification, and degree of fracturing. In general, interaction potential between seams decreases as the interburden increases. The relative stiffness of the individual strata that comprise the interburden and their ability to deform under load is a function of their elastic modulus. Strata that have a high elastic modulus, such as sandstone, tend to dampen stress transfer but are more prone to shear failure. Strata that have a low elastic modulus, such as shale, transfer load and bend more readily. The number of individual beds that characterize the interburden can also influence the magnitude and distance of stress transfer. A high degree of strata layering transfers load in greater magnitude and distance than thick massive strata do. Fracture zones created by adverse in situ conditions weaken the strata and, therefore, will lower the elastic modulus and increase interactive potential.

Although this research was directed at room-and-pillar workings, it may provide insight into the pressure bulb mechanisms for longwall gate roads. Essentially, gate roads are representative of narrow room-and-pillar developments. Pillar width for gate road developments rarely exceeds 100 ft. Since the vertical influence of the pressure bulb is directly related to pillar width, load transfer by pressure bulb mechanics will occur mostly in seams of close proximity. Analyses of room-and-pillar case studies conducted by Haycocks suggest that pressure bulb interactions may be limited to 110 ft of vertical distance. Field studies show that longwall interactions can extend over much greater distances, suggesting that some other load transfer mechanism may be responsible. In European investigations, researchers have found that interburden distances for which interaction effects are no longer encountered ranged from 210 ft to as much as 750 ft. U.S. studies have shown similar interactive distances involving interactions between room-and-pillar and longwall operations. Two mine design factors may be responsible for these large interactive distances. First, long gob-solid coal boundaries were left in the overlying mine. These boundaries, where gob meets solid coal or a pillar line, are known carriers of high abutment stresses. Therefore, the pressure bulbs associated with these features can be transferred over much greater vertical distances. Second, the longwall panels were relatively narrow in width in relation to their depth. Interactive distance associated with these types of openings may be increased because of arching stresses.

Arching Theory in Mining

Theories of arching have dominated rock mechanics research for decades; the concept was first introduced into mining literature as early as 1885. Pressure arch theory may be more applicable than pressure bulb theory in explaining the large interactive distances associated with longwalls. Arching theory assumes the mine opening is the major factor in the transfer of stress. Stress transfer is the result of the pressure arch that can form around a mine opening upon excavation.

Dinsdale presented the first significant work on this theory, He showed that the arch is usually elliptical in shape and that it exists both above and below the mine opening. As shown in figure 6, it consists of “intradosal ground” (tensile zone) enveloped by the “extradosal ground” (compressive zone). Large abutment pillars or barriers support the extradosal ground; the pressure created is known as the abutment pressure. Transfer of stress to developments mining beneath the abutment, can occur through the pressure bulb effect, The abutment is a high-stress zone usually occurring where the gob and solid coal meet. Unlike the uniformly loaded pillars in the example in figure 5, the abutment stress profile is characterized by peaks and troughs, with the peak occurring at the pillar edge. Since the load is not uniformly distributed, the pressure bulb becomes distorted and can transfer stress to greater depths than in the uniformly loaded case.

Dinsdale refers to the formation of “minor” and “major” pressure arches in underground excavations. Minor pressure arches can form independently from pillar to pillar as shown in figure 7 provided the strength of the pillar in situ exceeds that of the abutment pressure. If the pillars yield or fail because of excessive pressure, their load is transferred to neighboring barriers or abutment pillars and a major pressure arch will form as shown in figure 8. The magnitude of the abutment pressure and the shape and height of the arch are dependent upon the depth, the opening width, and the physical nature of the strata. Dinsdale noted that a relationship exists between panel width and depth. He theorized that if the mine opening became too wide, the arch would no longer span the opening and the extradosal ground would fail, leading to subsidence on the surface.

Haycocks studied the effects of minor and major arching through the use of finite element and stress-vector plots. This research led to an improved understanding of arching mechanisms under multiple-seam conditions. The models showed that when openings are narrow and in close proximity, minor pressure arches in adjacent seams

can interact, resulting In abnormally high lateral and abutment pressures. For very wide openings, such as those created by longwall mining, major pressure arch formation is likely to create points of excessive pressure in seams above or below. Minor pressure arches are more applicable in describing interactions between narrow entries in close proximity, whereas major pressure arches describe

the larger interactive distances associated with wide, deep openings.

In empirical studies, researchers have attempted to describe the height and width of the major pressure arch in actual mining conditions. The National Coal Board estimated the width of the arch as 15 pct of the seam depth plus 60 ft. Holland theorized that the arch was elliptical in shape and the maximum height was approximately twice its width. Peng estimated the height of the arch from 30 to 50 times the seam thickness. Haycocks states that caving and arching are two closely related strata movements and that it is the caving and sagging of the gob that creates the major pressure arch and ultimately its influence on overlying and underlying seams. He gives several formulas for determining the width and height of the arch. Haycocks also found that the arches can have various shapes, from a dome to a modified parabolic shape, depending on geology. The primary considerations are rock type, the degree of bedding in the strata, and the presence of natural fractures and joints. In strong, monolithic strata with little jointing, the arch formation may be low or the strata may bridge for narrow panel widths. In weak, highly bedded and jointed strata, the arch will have difficulty supporting itself regardless of panel width. In actual mining practices, most geologic conditions fall somewhere in between, as strata are composed of interbedded units of shale and sandstone of varying thickness and strength. The height and shape of the arch can vary greatly under these conditions and should be studied on a case-by-case basis with consideration for the ratio of seam depth to longwall panel width.

The formation of a major pressure arch in longwall excavations rests on two basic assumptions. First, the ratio of the seam depth to longwall panel width must be at some critical value so that the arch can support itself. Second, the gate road pillars must be of sufficient in situ strength to support the abutment pressure. Theoretically, if both of these conditions are met, arches can form from panel to panel as illustrated in figure 9. The arch is more likely to form for subcritical panel widths. In subcritical panels, some subsidence usually occurs, but less than that occurring in a critical panel. Equations for determining subcritical widths for longwall panels are given by Mark. If the abutment pressure exceeds the pillar strength, the arch will fail, but another arch may form if the excavation is sufficiently deep.

Studies show that arching stresses can either hinder or benefit mining in overlying and underlying seams. The extradosal ground forms the zone of high compressive stress that can cause ground control problems in the roof, floor, and pillars. Other studies show that arching may be beneficial to mining. The intradosal ground or tension zone is actually a destressed region in relation to the surrounding strata, and conceivably the stress encountered in this zone may actually be less than that created by the cover load.

Interaction for Ascending Order

Coalbeds that are longwall mined in an ascending order can experience interactions resulting from subsidence. This sequence of extraction causes the overlying strata to fail and is characterized by different zones of movement. The extent of each zone is a function of the subsidence process and the geologic composition of the overlying strata. Subsidence can produce two types of ground failure in overlying coalbeds. In the first type, the strata bend in response to subsidence, leading to the formation of a trough. The second type, known as interseam shearing, occurs when subsidence produces highly inclined shear or shear-tensile failures, displacing the coalbed to lower elevations. Design and geologic factors will be further addressed as they relate to these two types of subsidence failure.

Subsidence Trough

Longwall mining usually leads to uniform and predict-able subsidence as documented by surface measurements

A surface profile for supercritical and critical panels consists of a subsidence trough, the outer limits defined by the angle of draw, and an area of maximum subsidence. As shown in figure 10, the subsidence trough is identified by distinct zones of tension and compression. An inflection point, typically located directly superjacent to the ribline, distinguishes the two zones on the surface: the tension zone over the solid coal and the compression zone over the mined-out panel. Recent subsidence measurements over longwall panels in the Appalachian Coalfields indicate that the inflection point actually develops over the mined-out panel. This implies that the tension zone, for subsurface subsidence, would also be located over the mined-out panel. Field studies of room-and-pillar operations affected by subsidence support this assumption as the most severe roof conditions have been shown to occur over the lower mine gob immediately upon crossing the ribline. The factors that influence the exact location of the tension zone require further study and are likely related to caving behavior and bending of the individual beds.

The magnitude of roof problems in the subsidence trough is dependent upon the extent of the tension zone. This zone is associated with the formation of fractures and opening of joints, which lead to poor roof conditions. Increased flows of methane and water are frequently encountered as fractures provide pathways for their migration. The compression zone has been observed to cause only minor ground problems related to pillar instability, mostly rib spalling. Beyond the subsidence trough lies the zone of maximum subsidence. In this zone, the extraction efficiency of longwalls allows the ground to subside uniformly, usually resulting in improved ground conditions.

The amount of damage subsidence can cause to an overlying coalbed is largely dependent upon five factors: (1) the lower seam mining height, (2) the distance between the coalbeds or interburden thickness, (3) the angle of draw and the caving angle, (4) time or the age of the workings, and (5) the geologic characteristics of the interburden. Some studies have documented the effects of subsidence on overlying operations, and only recently has meaningful field measurement of subsurface movement and strata response to subsidence been undertaken in the United States. However, the magnitude and extent of subsidence damage coalbeds will experience can be assessed qualitatively using the factors listed above and other relevant information.

The first two factors are related and can be used to determine if overlying coalbeds are influenced by caving, fracturing, or sagging as shown in figure 11. Each zone can be identified by the failure characteristics of the strata, but the regions are not well defined and may vary in vertical extent.

Studies involving full extraction mining show that coalbeds lying within 20 times the lower seam mining height are most susceptible to subsidence-related damage, as they usually fall within the caving zone. The caving zone can actually be further divided into various regions of strata failure. The complete caving zone is a region of severely disturbed strata, and the failure is best described as highly fragmented to platey blocks. Studies indicate that this zone is generally three to six times the mining height depending on the expansion ratio or bulking factor of the rock. Coalbeds located within this distance may not be minable. The region from 6 to 12 times the lower seam mining height is an area of partial cave. The completely caved rock below offers some support to the overlying beds, yet failure may still be severe enough to cause the strata to fracture into large blocks. The strata within the subsidence trough may have a significant degree of bending, leading to intense fracturing, or may be displaced as a result of shear stresses. Ground conditions in the zone of maximum subsidence may even be less than ideal depending on the caving characteristics of the strata. Severe to moderate roof control problems should be anticipated when developing in the partial caving zone. A distance of 12 to 20 times the lower seam mining height defines the upper limits of the caving zone. In this region, strata may separate along bedding planes and fracture or joints may open, but the individual beds tend to remain intact. Also, displacements due to shearing are less likely to occur. These conditions become less intense progressively higher in the zone. Ground control problems may be severe in the subsidence trough, but ground conditions should improve in the zone of maximum subsidence.

Coalbeds lying within 20 and 50 times the lower seam mining height usually fall in the zone of fracturing. The bending of the strata is not as abrupt, and although fractures exist they are less pronounced. Ground control problems will mainly be encountered in the subsidence

trough, but generally the coalbed should be less difficult to mine. Damage to overlying coalbeds is minimal at distances exceeding 50 times the lower seam mining height as this is the zone of sagging. In this zone, bending is gradual and distributed over a larger horizontal distance. The strata throughout the subsided area tend to sag uniformly and mining should encounter virtually no problems.

The third factor, the angle of draw and the caving angle, can be used to roughly estimate the extent of the subsidence trough within the overlying coalbed. The angle of draw defines the outer limit of subsidence and projects out over the solid coal from the ribline as shown in figure 12. Accurate calculation of the angle of draw has been the focus of longwall subsidence research over the last decade. Most studies on the subject conclude that the appropriate angle for eastern U.S. applications ranges between 15° and 25° for flat-lying coalbeds. The caving angle is not explicitly defined in rock mechanics literature, but observers have referred to it as a “negative angle of draw.” This implies that it projects out over the gob from the ribline as shown in figure 12.

Studies of surface subsidence and caving behavior may provide some insight into a clear definition and value of the caving angle. King used this angle to estimate the side abutment load for longwall pillars. He assumed that this angle defined a failure plane where the weight of the overburden was evenly distributed between the pillars and the gob. King referred to the angle as a “shear angle,” and he estimated its value at 25°. Mark took a similar approach in developing a method for determining longwall pillar loads; he referred to the angle as the “abutment angle.” He estimated it at approximately 10° to 23° based on back-calculation from pillar stress measurements conducted in three coalbeds. Singh referred to the angle as the “angle of fracture” and used physical models to study its behavior and formation in different mediums. He noted the shear-tensile failure in its development and estimated its range from 5° to 15°. Peng implicitly called it a caving angle and based on observation estimated its range from 15° to 35° with an average of 25° applicable for most conditions. Different geological conditions may be the determining factor for this broad range of values. Obviously, caving mechanics warrants further study to better delineate appropriate caving angles based on stratigraphy.

From purely a geometric perspective, the angle of draw and the caving angle can be used to determine the extent of the subsidence trough. Figure 12 illustrates a simplified example of this zone’s influence using an angle of draw of 25° and a caving angle of 25°. Strata bending progresses gradually, beginning at the angle of draw, and becomes more pronounced superjacent to the ribline. Upon crossing this boundary, fracturing and poor roof conditions due to tensile stress are most apparent. Proceeding farther, the inflection point denotes the end of the tension zone and the start of the compression zone. Beyond this, the strata begin to settle uniformly in the zone of maximum subsidence. The closer the overlying coalbed is to the mined-out panel, the lesser the extent of the subsidence trough, but the greater the bending in the strata. As interburden increases, so does the horizontal extent of the subsidence trough, but the bending is less severe and it is distributed over a larger distance. Since the degree of fracturing is directly related to bending, the potential for ground problems decreases as interburden increases.

The fourth factor, the effect of time, should also be taken into account when assessing potential interactive problems due to subsidence. Sufficient time delay in mining seams damaged by subsidence is an important factor, as delay allows for caving and compaction of the gob, which lessens damage to overlying gate entries upon development. Studies involving subsidence measurements over longwall panels have shown that additional surface subsidence may occur over a mined-out panel as an adjacent panel is being mined, King referred to this as a “triggered subsidence.” This concept can also be applied to multiple seams when mining above a mined-out panel. Redistribution of stress may cause additional subsidence and ground problems in the upper mine.

Zhou classified subsidence interactions into three possible categories based on a time factor: (1) Lower seam mining is currently active with upper seam mining, an active condition; (2) mining in the lower seam is complete but the ground is still in the process of settling, both an active and passive condition; and (3) subsidence is complete and the ground has settled reaching a new state of equilibrium, a passive condition. A method to quantify this time variable in terms of upper seam stability was attempted through a rating method, which also considered other factors such as interburden thickness and percent sandstone, lower seam mining height, and percent extraction. This method was empirically developed from room- and-pillar and longwall workings for categories 2 and 3 above. Interestingly, little if any field documentation is available for active conditions, category 1. Since the effects of subsidence on overlying active mines, in terms of time, are almost immediate, the active condition has the potential of being the most damaging of the three.

The fifth factor, stratigraphy, can influence the other four factors that determine subsidence damage. Factors of primary importance are the number and thickness of the individual beds, the relative stiffness of each stratum, the presence of natural fractures or joint systems, and the cohesion between individual beds. Strata response to bending stress differs depending on the properties of the individual beds that form the entire rock mass. The elastic modulus indicates the stiffness of the rock and its ability to deform. In general, shales have a low elastic modulus, and the beds tend to bend and fracture as an

entire unit. Sandstones, with a high elastic modulus, are stiffer formations, and the beds tend to shear and displace. The degree of fracturing or natural jointing of the rock also affects its elastic modulus. Fractures lower the strength and the elastic modulus of the rock. Shales usually have more pronounced fracture systems than the more competent sandstones. Cohesion between the individual beds is also an important consideration for it determines if slippage will occur along bed interfaces, leading to differential movement and bed separation. A significant number of field studies that detail subsidence characteristics in relation to stratigraphy are forthcoming. Such information will improve damage assessment and help in the formation of mining plans to control interactive problems.

Interseam Shearing

Interseam shearing can be very damaging to overlying coalbeds because the strata tend to shear and displace as shown in figure 13. This interaction is more likely to occur when both operations are “active” as described by Zhou. Field studies have documented this type of failure in room-and-pillar operations. Holland observed that in 38 cases of coalbeds affected by subsidence, about 75 pct had a shear-tensile failure. The greatest potential for shearing occurs in coalbeds lying within 12 times the extracted seam height or in the partial caving zone. Haycocks reports that shearing is most likely to occur when interburden is less than 33 ft, but opening width is critical and under severe circumstances shearing may extend through to the surface, displacing large areas of coal.

Stratigraphy is the most important factor in defining the shear potential of the overlying strata. Studies indicate that physical properties of the strata, namely high elastic modulus, are a prominent feature in shear failure. Therefore, shearing occurs more frequently in sandstones and similar rock types. Stemple noted in his study

that the shear angle differs depending on the strata type. The angle of failure has been observed to range from nearly vertical for thick sandstone units to 25° for highly bedded strata, and it appears to coincide with the shear or caving angle discussed previously. This tendency for hard strata to resist bending and then shear can also be observed at the longwall face, as competent sandstone roofs may “hang up” before abruptly fracturing at the supports. Fractures and joint systems inherent to the strata are also critical factors in interseam shearing. Vertical joints have few cohesive properties and, therefore, provide planes for shear failure. Faults, which are offsets in the coalbed, may complicate or contribute to shear failure.

Simultaneous and Random Orders of Extraction

Extraction sequences that are simultaneous or random can lead to various interactive problems related to Zhou’s first two categories described previously. Simultaneous mining implies that both mines will be worked at the same time in the same area so that “active” interactions may occur between the two operations. Although the simultaneous longwall mining of multiple seams has not been documented in the United States, attempts to longwall mine thick seams simultaneously in Europe may provide some insight into design strategies. Wilson cites foreign experiences with multilift thick-seam mining and presents design considerations. Probably the most important factor in simultaneous extraction of two seams is co-ordination. Gate road development and longwall panel extraction in both mines should be planned to control interactive problems. In most cases, a descending order of extraction is preferable practice. Gate road development and panel extraction in the upper mine should be kept in advance of the lower mine so the damaging effects of subsidence are avoided. Simultaneous extraction can lead to obvious production and scheduling problems. For instance, downtime on the longwall face of the upper mine will delay mining of the lower face until these problems are corrected. Since simultaneous mining is such an economically risky endeavor, long-term mine planning and rigid scheduling of gate road development and panel extraction are essential.

Random seam sequences should be avoided because they only serve to complicate the design process. In the Eastern Coalfields, longwall operators may frequently encounter room-and-pillar operations in adjacent seams. In many instances, interactions are inevitable, but improvements in ground stability can usually be achieved through careful planning and fundamental ground control methods. An operator should first assess the interactive potential of the workings, then adjust mine plans accordingly. This could include changing gate road pillar design and longwall panel orientation and size or increasing support. Primary factors to consider include the interburden distance, depth, and geometry of the workings. Several case studies document the effects of room-and-pillar workings on longwall operations. These studies showed that ground problems are usually encountered when mining over or under gob-solid coal boundaries or large isolated barrier pillars. In one case, abandoned room-and-pillar workings were encountered in seams both above and below the active longwall. Interactions were likely, but difficult to predict because of the complex geometry of the workings as shown in figure 14. The decision to orient the panels in this manner was based largely on maximizing the reserve potential of the property. Although problems were encountered, mostly related to subsidence, the operator successfully mined several panels using this layout. In another case, a longwall panel encountered overlying room-and-pillar workings as shown in figure 15. Interactions on the longwall were mainly experienced as tailgate problems, creating moderate ground instability. In both these cases, the operators were able to incorporate previous knowledge in the design process and planned accordingly when room-and-pillar workings were encountered.

Gate Road Pillar Design

The primary function of the gate road pillars is to maintain the stability of the headgate and tailgate entries as the longwall panels are mined. These entries facilitate the movement of workers, coal, and equipment as well as provide ventilation and escape in emergencies. Considerable progress has been made in longwall pillar design in recent years because of significant and noteworthy research. It is beyond the scope of this section to detail the findings of this work; however, the principal design concepts and their implications for multiple-seam interaction will be discussed.

There are two basic approaches to longwall pillar design. The yield pillar approach uses small pillars, which are designed to fail in a controlled manner and transfer their load to nearby abutments by means of a pressure arch. The conventional approach uses large pillars, sometimes referred to as “abutment or squat pillars,” which are designed to carry all of the loads to which they are subjected as the longwall panels are extracted.

Yield Pillar Approach

Holland was the first to propose the concept of yielding pillars for ground control purposes. His work mainly focused on yield pillar applications for room-and-pillar developments and the formation of a pressure arch for improved entry stability. Mark presents the most recent advances in yield pillar theory, design, and application for longwall gate roads. He specifies three fundamental assumptions for yield pillar design. First, the pillars must deform so the main roof bridges, transferring load and creating a pressure arch. Second, there must be a solid abutment nearby of sufficient load-bearing capacity, such as a longwall panel or abutment pillar, to accept this transferred load. Third, the pillar must fail in a controlled manner but maintain enough strength to support the weight of the rock within the pressure arch. Figure 8 illustrates these three assumptions. Several examples of yield pillar successes and failures in U.S. longwall mines are cited in his work. A common denominator in successful applications of yield pillars is using increased support density in entries, usually in the form of wood cribs. Mark gives some preliminary formulas for determining the width of yield pillars, but concludes that design is based mostly on experimentation and that no real quantitative guidelines exist.

European experience has shown that gate road designs that minimize pillar width are the most favorable for controlling multiple-seam interactions. Leaving large pillars between panels creates stress concentrations in seams both above and below. To prevent this, the gob is maximized by utilizing single-entry gate roads and advancing longwalls. This mining method is not permissible in U.S. mines, so other strategies must be adopted. Two- or three-entry gate road systems that utilize yield pillars that tend to “crush-out” when both panels are mined may reduce interactions. In the case of pillar load transfer, the pillars no longer carry the abutment load, which is distributed into the nearby gob, and abutment loads are less likely to become concentrated in lower operations. In the case of subsidence, the subsidence profile at the pillar edges is reduced; therefore, interactions due to strata bending or shearing are lessened.

Maleki investigated the use of yield pillars in longwall gate entries for controlling interactions between two seams at a western mine. Using both computer-aided design and underground trials, he concluded that no three-entry system that utilized yield pillars or a combination of yield pillars and conventional pillars would satisfactorily stabilize the gate roads in both seams. A yield and conventional pillar design resulted in the most stable conditions for the upper mine gate roads, but this design created interaction problems in the lower mine. Layouts using two yield pillars resulted in the most stable conditions for the lower mine, but the least stable for the upper mine. Maleki concluded that a two-entry system that utilized one yield pillar may provide stability for both mines.

Forrest also studied the use of yield pillars for controlling multiple-seam longwall interactions. Using photo-elastic models, he showed that yield pillars can reduce

shear stress in the roof and floor of the lower seam better than conventional pillar designs can. However, he noted that narrow pillars that do not completely yield and that maintain a strong confined core may increase the interaction in the lower seam. Generally, the research showed that yield pillars can result in significant reductions of stress in the lower mine gate roads and improved pillar and entry stability.

Although the yield pillar approach seems to be the optimum method for controlling longwall interactions, designing a gate road system that utilizes all yielding pillars is a complex problem. One major reason is that the post-yield and stress transfer mechanics of yield pillar systems are not yet fully understood. Until significant advances are made in the field under single-seam conditions, this concept will have limited application for multiple seams. A more practical approach utilizes conventional pillar designs with consideration given to interactions between seams.

Conventional Longwall Mining Approach

The conventional approach for gate road design uses large pillars that are designed to accept all the abutment load to which they are subjected. Proven methods for designing conventional gate road pillars have been presented by several researchers. Carr developed a design method that has been extremely successful for longwalls operating in the Warrior Coal Basin of Alabama. Choi was the first to develop a method specifically for U.S. conditions. Hsuing considered the properties of the roof and floor in his designs. Analysis of Longwall Pillar Stability, or ALPS, was originally developed by Mark and Bieniawski at Pennsylvania State University and was recently refined by the USBM. Each method has its own attributes, and most longwall mines in the United States utilize one of these methods to design gate roads. Interactions between longwall operations that use conventional pillar designs depend primarily on the depth, the interburden thickness and composition, the panel width, and the pillar size.

The magnitude of the side abutment load on the gate road pillars after the longwall panel(s) are mined determines the intensity of the interaction in lower operations. The side abutment load is represented as the wedges of strata defined by the abutment angle as described by Mark and shown in figure 16. This side abutment can be transferred to lower operations when coalbeds are mined in descending order. If gate roads are superpositioned, the lower mine pillars must be designed to accept this additional load on development. If the gate roads are offset, load from overlying pillars is then transferred to the longwall panel and face supports and they must contend with the additional load. Depth is the principal factor in determining the magnitude of the abutment load. The panel width is a secondary factor. Supercritical and critical panel widths produce slightly more abutment load for a given depth then subcritical panels do. Pillar size does not affect the value of the abutment load, but larger pillars may serve to distribute this load over a wider area.

When seams are mined in ascending order and subsidence is the consideration, the total width of the gate roads defines the resulting subsidence profile in overlying coalbeds. The wider the gate roads, the greater the distance between the subsidence troughs of two adjacent mined-out panels. As discussed previously, factors related to interburden thickness and stratigraphy, mining height, the angle of draw, and the caving angle determine the extent of damage an overlying coalbed will experience.

This applies to supercritical and critical panel widths where subsidence is experienced on the surface. Subcritical panels may cause interactions due to arching if the gate road pillars are sufficiently large to support the arch. In this case, interactions in underlying or overlying operations may experience high stress in the compressive zone of the arch as depicted in figure 9.

Several field studies have documented the effectiveness of conventional gate road pillar designs where room-and- pillar workings were encountered in adjacent seams. These studies showed that properly designed pillars, where consideration is given to the transfer of stress, performed satisfactorily under multiple-seam conditions. The ALPS method was used to assess pillar strength and stability factors at these sites. The studies found that when stability factors were less than 1.1 after tailgate loading, interactions were more damaging to pillars and entries. The research recommends that stability factors range between 1.3 and 1.5 after tailgate loading to maintain stability.

Other Design Considerations

Other design considerations to reduce the influence of the gate road pillars can include a combination design of yield and conventional pillars or pillar recovery techniques. Yield pillars that flank one or both sides of a larger conventional pillar are known as yield-abutment design. This design may prove useful in controlling interactions if the yield pillars crush-out, thereby reducing the relative width of the gate road.

Pillar recovery techniques involve extracting one or several of the gate road pillars with the longwall panel. The benefits of this technique are obvious, from both a productivity and a multiple-seam standpoint. Successful demonstration of pillar recovery for single-seam mining has been documented in several case studies. These studies show that pillar recovery is a feasible technique, but the benefits may be outweighed by the cost of ground support. Excessive support costs may be incurred, because entries and crosscuts must be supported in advance with materials the longwall shear is capable of cutting, such as piers constructed of cement-flyash mixtures.

Orientation of Gate Roads and Panels

The two basic design approaches to gate road and panel orientation are offsetting and superpositioning. The decision to offset or superimpose gate roads when mining multiple seams has been a topic of much concern in longwall design. Model studies have shown that superpositioning of gate roads produces the most adverse stress conditions on the gate road pillars and that in most cases offsetting pillars is preferable practice. The major advantage of offsetting is that the gate roads can be developed in a relatively destressed area, over or under the mine gob. The disadvantage is that stress concentrations due to pillar load transfer or subsidence may create problems on the longwall face.

In superpositioning, the longwall face has the advantage of operating in the destressed area but the gate road pillars are subjected to potential stress concentrations. Since the stability of the gate roads is of primary importance, superpositioning is rarely practiced, but it may be required under special circumstances. Geologic factors, such as faults or sandstone channels within the coalbed, may render a block of coal unminable, thereby shifting gate roads to a superpositioned arrangement. Superpositioning may also be required near the perimeter of a coal property to maximize recovery. Whether adopted as part of the regular mining plan or as a contingency, superpositioned gate road pillars must be properly designed to contend with interactions.

Offset Gate Roads

Offsetting positions the gate roads at the centerline of the longwall panel in the previous seam as shown in figures 17 and 18. Gate road development typically benefits from this arrangement. Accepted practice assumes that the load at the center of the gob returns to the normal cover load after longwall panel mining as shown in figure 19. Under these conditions, designing pillars at depths for a single-seam case should be sufficient. Recent field investigations, by the USBM, of gob behavior at a longwall in Virginia indicate that the stress at the center of the gob remained below that of the normal cover load. These studies show that load-carrying capacity of the gob depends upon the depth, panel width, and the composition of the overlying strata. These factors influence the caving characteristics of the gob and the load it will carry as it compacts and consolidates. This research may prove that offsetting is more advantageous to gate road development than previously thought.

For descending extraction (fig. 17), the interaction on the longwall face is contingent upon the abutment load the overlying gate road pillars carry. A certain percentage of this load will be transferred to lower operations depending upon the distance between the seams and the interburden composition. The strata behind the face have little load- bearing capacity immediately after caving, so the supports and longwall face must contend with the bulk of the transferred stress. Support instability problems can occur directly beneath the overlying pillars if the horizontal and vertical components of stress associated with the pressure bulb effect become excessive. Ground control problems can arise in the roof, floor, or longwall face because of vertical and shear stresses. Soft roof strata

can lead to falls between the supports and longwall face. Hard roof strata may cause coal bumps or less violent spalling on the longwall face. Soft floor strata create floor heave and difficulties in advancing supports.

For ascending extraction (fig. 18), the degree of subsidence damage in the roof, floor, and coal will determine the difficulty of mining. Damage assessment, discussed previously, is related to the lower seam mining height, caving characteristics, and stratigraphy. The front abutment stress is a major factor contributing to potential ground instability as the longwall panel is mined. Field studies show that if subsidence causes fracturing problems, the front abutment may create further instabilities in the roof, leading to difficulties in advancing supports.

Superimposed Gate Roads

Superimposing positions the gate roads in successive seams in a columnized arrangement as shown in figures 20 and 21. This allows the longwall panel to operate within the destressed region of the gob, but pillars must withstand the effects of interaction.

For descending extraction (fig. 20), transferred stress originates from the abutment load the overlying gate road pillars carry as shown in figure 22. Designing pillars in the

lower mine to withstand this additional load is the primary concern. There are two basic options to consider for conventional pillar designs. First, the two mines are planned individually and the upper mine pillars are sized initially for a single-seam case. The lower mine pillars are then sized to accept the additional stress that will be transferred. In most instances, the pillars in the lower mine will be slightly larger in width. If the center entries are superimposed, the wider lower mine pillars would cause the lower mine gate roads to be slightly offset in relation to the upper mine, thereby shortening the panel width as shown in figure 23. In the second option, the pillars can be designed in both mines simultaneously so similar sizes are achieved in both operations as shown in figure 24. This requires deliberate mine planning and a committed effort to longwall mine both seams, knowing that pillars in the upper mine may be slightly overdesigned to achieve properly designed pillars in the lower mine.

Currently, there are no proven methods for estimating the quantity of stress transferred and safe pillar sizes for superpositioned gate roads. Empirical studies that document interactions involving superpositioned gate roads and recommendations for safe pillars sizes are lacking, but designs based on numerical methods have provided insight into the problem. The USBM has recently developed a method for estimating stress transfer and ensuring safe pillar sizes specifically for superimposed gate roads. The USBM’s MULSIM/NL model, a boundary element computer program, was used to analyze the transfer mechanics between gate road pillars. Analysis of Longwall Pillar Stability (ALPS), an empirically based method for designing longwall gate road pillars, was used to calibrate model input parameters. The attributes of the MULSIM/NL model and ALPS were combined to develop a method for estimating lower mine pillar stress and to recommend limits for safe pillar designs when super-positioning longwall gate roads. The report (IC 9305) is available from the USBM upon request.

Superpositioned gate roads for an ascending order of extraction are shown in figure 21. As illustrated in this figure, if similar-sized pillars are used in both mines the outer entries in the upper mine will be positioned over the lower mine gob or in the tension zone. Ground conditions in these entries may be severe, particularly when they are located within the caving and fracture zone of the lower mine. There are two options to consider to avoid this situation. First, pillar width in the upper mine can be reduced so outer entries will be located above the lower mine pillars as shown in figure 25. This arrangement is usually suitable for maintaining these entries, but in extreme cases of fracturing it may be necessary to locate entries farther over the underlying gate roads. A conservative method for estimating entry location is to determine the outer limits of the subsidence trough based on the angle of draw. Figure 26 illustrates entry location assuming an angle of draw ranging between 15° and 25°.

Once a pillar size is selected, proper consideration must then be given to the pillar’s strength to ensure its stability during the longwall loading cycle. In the second option, the pillars in the upper mine can be designed sufficiently large so that entries are developed away from the effects of the tension zone. The inner limits of this zone and the location of entries can be estimated by using the caving angle. Figure 27 gives an example assuming an angle ranging from 15° to 25°. There are two obvious drawbacks to this arrangement related to recovery. First, the pillars are substantially overdesigned for their depth, and second, the panel width is reduced considerably.

Extraction of two seams that were longwall mined in ascending order using superpositioned gate roads is documented in a USBM study. In this case, outer entries for the upper mine gate roads were located over lower mine pillars as in figure 25. This arrangement proved to be successful with over 20 longwall panels mined. Although ground problems were encountered, they were moderate and controllable, occurring mostly in setup and recovery room entries that crossed over the lower mine gob as in one study area (fig. 28).

Staggered Gate Roads

There may be special design considerations for gate road layouts when three or more seams are mined in a descending order. If gate roads are superpositioned, pillar designs must consider the transfer of stress and, therefore, pillars must become increasingly larger in successive seams, leading to the loss of reserves. Offsetting gate roads lessens pillar support requirements with alternating seams being superpositioned. But even in this situation, pillars still must be larger to account for the transfer of stress. An alternative to offsetting and superpositioning is a staggered arrangement. As shown in figure 29, gate roads in successive seams are offset slightly so they are always developed beneath the gob and away from the influence of overlying pillars. This method has the following advantages. First, the benefits of developing beneath the gob are maintained and a buildup of stresses in gate roads is avoided. Second, recovery is improved as pillars can be designed at depths as if in single-seam situations and do not become excessively large to account for the multiple-seam situation. Third, a more balanced subsidence profile is produced where consideration for surface strains is a design constraint. This method has been successfully practiced in coalfields in the United Kingdom to increase percent extraction and avoid overlying gate road pillars in three-seam scenarios.

Mulsim/NL Modeling

MULSIM/NL is a boundary element model, developed by the USBM, for calculating stresses and displacements in tabular deposits. The model provides the capability to analyze many coal raining situations and to determine the effects of the three-dimensional stress redistribution caused by mining in either single or multiple seams. The model was used to examine the average stress distribution across a longwall face for the following design arrangements: first, a longwall panel mined above and below a gob-solid coal boundary as shown in figure 30, and second, a longwall panel mined below isolated gate roads positioned at various locations across the longwall face as shown in figure 31.

To establish stress trends related to the above design arrangements, certain design and geologic criteria were fixed to limit the number of influence variables. The longwall panel and pillar dimensions used in this study were chosen to generally represent current conditions in the field. The 1992 Longwall Census from Coal Magazine was used as a basis for selection. The critical dimension for the longwall panel, as used in the model, is the panel width, since the model was used to examine average stress across the longwall face. There were 93 operating longwall faces in 1991, with the average width being approximately 715 ft. Based on this figure, a panel width of 700 ft was selected for analysis in the model.

The census also shows that 46 pct of the operators use a three-entry gate road, 38 pct use a four-entry gate road, and the remaining 16 pct use other configurations or were unknown. Based on these figures and to narrow the extent of the investigation, a three-entry gate road using 60- by 80-ft pillars was selected for study.

The fixed geologic criteria included the depth, the interburden thickness and physical characteristics, coalbed thickness and physical characteristics, in situ stresses, and geologic discontinuities. Of these, depth and interburden thickness are the critical factors influencing stress transfer. Statistical analysis of case study information and photo- elastic model results show that interactive probability is most sensitive to the relationship between these two parameters. For this analysis, the depth to the longwall panel was kept constant at 800 ft as was the interburden thickness at 50 ft. The depth was selected as the average operating depth from longwalls in the 1992 Longwall Census. Also, this depth is the ALPS design limit for a three-entry gate road with 60- by 80-ft pillars in a 6-ft coalbed. An interburden of 50 ft was selected to represent a close-seam interaction. Under this condition, the magnitude of stress transfer to the longwall face would be maximized, thereby making it easier to distinguish stress trends.

The next most important parameters are those of the coalbeds, which include their thicknesses, coal strength, and dip. In situ stresses, such as a high horizontal stress, may affect overall stress magnitude, but the transfer of stress is more directly related to the normal cover load, which is a function of depth. Geologic discontinuities, such as clay veins, are more likely to cause localized instability in the workings than to have an effect on the overall transfer of stress.

In relation to this study, the MULSIM/NL model has two geologic shortcomings. First, geologic discontinuities cannot be represented in the model. Second, individual strata that characterize the overburden and interburden cannot be represented, so a generic modulus is chosen to depict the overall lithology. Assuming that the overburden and interburden is one homogeneous, isotopic material makes the strata reactions stiffer than they really are. Therefore, the elastic modulus of the material is lowered in order to more closely approximate a stratified rock mass. Physical properties of strata in the MULSIM/ NL model are represented by the elastic modulus and Poisson’s Ratio. The following linear elastic modulus values for the overburden and interburden, coal, and gob are given below. A Poisson’s Ratio of 0.25 was assumed for all cases.

The coalbeds were assumed to be 72 in thick. This figure was chosen because it represents an average operating height from the 1992 Longwall Census. The coalbeds were also assumed to be flat-lying deposits with no dip. The assumed density of the overburden was 162.5 lb/ft.

Crossing A Gob-Solid Coal Boundary

A longwall panel mined above or below a gob-solid coal boundary occurs frequently in the Eastern Coalfields when abandoned room-and-pillar or longwall workings are en¬countered in adjacent seams. The MULSIM/NL model was used to evaluate two design considerations related to the average vertical stress across the longwall face for this situation. First, the model considers the direction of mining in relation to the gob-solid coal boundary to evaluate whether it would be more advantageous for the panel to be mined from the solid to the gob side or the gob to the

solid side of the boundary. Second, the model was used to determine the angle of approach, or the angle at which the longwall face should be oriented to the overlying boundary to minimize stress across the face. Various model runs were made with the longwall face oriented at 0°, 15°, 30°, 45°, and 60° to the boundary. The center of the panel served as the reference point to the boundary, as the longwall face was mined from 250 ft approaching to 250 ft past this reference point as shown in figure 32. The panel was mined in 50-ft increments through this distance and the average stress across the face was calculated at each increment. This produced 11 stress values, which were then normalized to a single-seam situation. The single-seam runs used the same input parameters mentioned above, but without the mining in the upper or lower seam. Normalizing the data in this manner simplified the analysis and comparison of trends without using actual stress values.

The first design arrangement examined is the longwall being mined below the boundary from the solid to the gob side as shown in figure 30A. The results of this analysis are shown in figure 33A. As shown in the graph, the normalized average stress across the longwall face was above the single-seam datum until the face was approximately 50 to 100 ft under the gob. At this point, the stress decreased to less than the single-seam datum for all angles of approach. The lowest stress was experienced at the 0° angle and was approximately 0.45 times the single-seam stress.

The second design arrangement examined is the longwall being mined below the boundary from the gob to the solid side of the boundary as depicted in figure 30B. The results of this analysis are shown in figure 33B. As shown by the graph, the normalized average stress across the longwall face remained below the single-seam datum when advancing toward the boundary for all angles of approach.

Immediately upon crossing the boundary, the stress began to increase at different rates and exceeded the single-seam datum at various distances past the boundary depending on the angle of approach.

The last design arrangements examined are the longwall panel being mined above boundary as shown in figures 30C and 30D, respectively. The results of these two situations are shown in figures 33C and 33D. The trend of normalized average stress across the longwall face followed a similar trend as mining below the boundary (figs. 33A and 33B). In comparing these two sets of graphs (mining above versus mining below the boundary), there appears to be little difference in the average face stress produced. One should be cognizant of the fact that this analysis only considers stress and does not take into account the strata displacements and fracturing that subsidence will create. In other words, although the stresses may be nearly the same when mining above or below the boundary, the quality of the strata for the subsidence case may be more of a determining factor in assessing the conditions on the longwall face.

To address the first design consideration, the data suggest that it is better to mine the longwall panel from the gob to the solid side of the boundary rather than from the solid to the gob side. This relationship is best shown by comparing figures 33A and 33B using the 0° angle graphs. In the solid to gob case (fig. 33A), the maximum stress is experienced under the solid approximately 75 ft from the boundary and is over 1.6 times the single-seam stress. The minimum stress is experienced well under the gob, 250 ft past the boundary, and is approximately 0.45 times the single-seam stress. In comparison, the gob to solid case (fig. 33B) shows that the maximum stress occurs about 50 ft past the boundary under the solid coal and is approximately 1.25 times the single-seam stress. The minimum stress occurs at 100 ft approaching the boundary and is approximately 0.30 times the single-seam stress. A similar trend of minimum and maximum stresses also applies for the other angles of approach. The reason for this difference is that in the solid to gob case, the vertical stress becomes concentrated and is supported by a decreasing area as the longwall panel is mined. This is illustrated in figure 33A. The average stress across the face reaches a peak from 50 ft approaching to directly under the boundary, then decreases abruptly at about 50 to 75 ft under the gob.

Similar experiences have been documented in the United Kingdom in mining beneath gob-solid coal boundaries in coalfields. In one case, a longwall face approached the boundary from the solid to the gob side of the boundary. As the face advanced closer to the over-lying boundary, the ground conditions deteriorated, because a high vertical stress was being supported by a diminishing area of the panel and pillars. Once past the boundary and under the gob, face conditions improved. When the same face en-countered the next boundary and advanced from the gob to the solid, the vertical stress encountered was lessened.

The second design consideration concerns the angle of approach, or the proper angle to orient the longwall panel to the boundary to minimize stress across the face. When examining figure 33B, one would conclude that a 60° angle would be the best choice, but this is not completely true. Since we are examining average stress across the face, the face for a 60° angle is always influenced by the overlying gob, as shown in figure 32, and this reduces the average stress. One must realize that peak stresses occur along the face directly beneath the boundary, and these peak stresses

may be a primary factor in selecting an approach angle. Figures 34A and 34B graph the normalized peak stress for approaching the boundary from the solid to the gob and from the gob to the solid, respectively. As shown in the graphs, the normalized peak stresses when mining from the solid to the gob are much higher than when mining from the gob to the solid. This is a similar trend to that of the normalized average stress as shown in figures 33A and 33B. But when comparing the normalized average stress (figs. 33A and 33B) with the normalized peak stress (figs. 34A and 34B), the trends are significantly different. In general, the normalized average stress decreases as the angle of approach increases; however, the normalized peak stress increases as the angle of approach increases. For the 0° case, the graphs are similar, as the normalized average and peak stresses coincide with each other. As the angle increases, the graphs begin to vary, and the 60° angle of approach best illustrates the difference. In the case of mining from gob to solid, the normalized average stress (fig. 33B) for the 60° angle remains below the single-seam datum because the influence of the gob lowers the average stress across the face. Conversely, the normalized peak stress (fig. 34B) for the 60° angle is much higher, ranging from 1.2 to 1.4 over the single-seam datum. The peak stress will transverse the face and occur directly beneath the boundary. Based on the mining conditions,

the operator must decide how to distribute the stress across the face: Either encounter the peak stress all at once, the 0° case, or distribute the peak stress across the face, as in the 60° case. The best approach may be to limit the angle to under 30°. This would lessen the influence of the peak stress across the face and permit the face to advance beyond the boundary as quickly as possible.

A final design consideration concerns the placement of the headgate and tailgate entries when the boundary is crossed at an angle. In the above analysis, the tailgate was placed under the overlying gob as depicted in figure 32. Reversing the headgate and tailgate entries had little affect on the average stress across the longwall face but did increase the stress on the tailgate side of the panel considerably, Since the tailgate usually experiences the most adverse conditions, it is recommended that the tailgate be placed under the overlying gob to reduce stress.

Isolated Gate Roads above Panel

There are two fundamental approaches when seams are extracted in descending order. First, the gate roads can be superpositioned in the two seams as shown in figure 20; in this case, the primary consideration is the proper design of the lower seam gate road pillars to accept the resulting load transfer. Second, the gate roads can be offset along the longwall panel as shown in figure 17; in this case, the primary design consideration is the magnitude of the resultant stress concentration on the longwall face. The first design consideration has been recently addressed in USBM IC 9305, in which procedures were established for properly designing lower seam gate road pillars. The second design arrangement, offset gate roads, will be the focus of this discussion.

Offsetting gate roads is the preferable practice because gate roads in the lower seam can be positioned beneath the overlying gob in a destress zone as shown in figure 17. The center of the longwall panel must then contend with the transfer of stress from the overlying gate roads. However, there may be instances in which positioning the overlying isolated gate roads at the center of the underlying panel is not possible. For instance, the panel width in each seam may differ, as the current trend for longwalls is increasingly wider panels. This situation precludes the exact alignment of the overlying gate roads at the center of the underlying panel when a series of adjacent panels are to be mined. As a result, the overlying isolated gate roads may be positioned at various locations across the panel width, either at center or toward the tailgate or headgate sides as shown in figures 31A through 31C. The MULSIM/NL model was used to evaluate these design layouts. The input parameters described previously were used in the analysis.

Figures 35A through 35C correspond to figures 31A through 31C and show the calculated stress profiles across the width of the longwall panel for each location of the gate roads. The stress in the graphs has been normalized to a single-seam situation. The profiles show a predictable trend of high stress concentration across the longwall face directly beneath the overlying gate roads. The normalized peak stress for the headgate and tailgate locations is

approximately 1.8 times the single-seam stress. For the center of the panel location, normalized peak stress is almost 2.0 times that of the single seam. The area of the panel positioned under gob will experience less stress than in the single-seam case. Unlike the gob-solid coal boundary case, in which the boundary is approached and passed, the stress profiles shown will affect the panel for its entire length.

Several design factors should be considered when positioning the overlying gate roads. First, since the ground conditions on the tailgate entries are usually more severe, positioning the gate roads toward the tailgate side will increase stress in these entries. Therefore, when possible overlying gate roads should be kept toward the headgate side to improve tailgate conditions. Second, positioning the overlying gate roads at the center of the panel is the best practice even though it produced slightly higher normalized peak stress. The panel has the ability to distribute stress more evenly and over a larger area. This is a better situation than concentrating the stress at the panel edges or on the gate roads, as would be the case with the other arrangements. Third, the longwall support will also be influenced by this load transfer. Improved conditions should result if the center supports, rather than the supports at the panel edge, contend with the additional stress.

Summary

Experience has shown that mining in one seam can affect subsequent operations in seams both above and below the one being mined. In the past, coalbeds in the United States were mined in no particular order with regard to controlling interactions between operations and reducing associated ground control problems. The primary factors influencing seam sequencing were related to economics, availability, and ownership. Unfortunately, this practice still continues today in many instances. Ground problems associated with multiple-seam mining can cause significant damage to the mine structure, resulting in escalated mine costs, reduced safety for employees, and the loss of minable reserves. Ground problems are usually compounded by poor mine planning and a lack of knowledge of the factors that contribute to multiple-seam interactions. Improvements in mine planning and design that would control or eliminate multiple-seam interactions present a major challenge to the mining industry.

Longwall mining, because of its extraction efficiency and increased productivity, is rapidly gaining a larger portion of underground coal production. As the number of longwall operations increases and seams become mined out, the likelihood of encountering workings in adjacent seams will increase as well. The successful practice of designing a multiple-seam longwall layout depends primarily on the operator’s intrinsic knowledge of local geology and strata behavior. However, models can be used to provide insight into relative stress transfer and distribution in multiple seams. The USBM’s MULSIM/NL model was used for this purpose. Since very little empirical information exists to assist operators with multiple-seam longwall design, numerical models provide a tool for initial design. The model expedites design analysis and provides insight into the relative stress distribution and transfer that occurs under various design conditions. The analysis of different designs using numerical models has considerable potential in helping operators find solutions to complex multiple- seam interactive problems.

The purpose of this report is to provide mine operators with practical information for planning and designing multiple-seam longwall mines. It provides design strategies that should be useful when workings in over- or underlying seams are encountered. The report has focused on three fundamental design principles for planning safe and productive multiple-seam longwall mines: (1) the sequencing of seams; (2) the design of gate road pillars; and (3) the layout of the gate roads and panels. A descending order of extraction is preferable, because the severe ground problems associated with subsidence should be avoided in most cases. Two basic approaches are available for gate road pillar design: yield pillars and conventional pillars. For the most part, yield pillars require- further study to assess their performance under multiple-seam conditions. However, there are several conventional pillar design approaches available to the operator that have been used successfully in the field. Finally, when orienting the gate roads and panels, either a superpositioned or offset arrangement can be used, depending on mining conditions. When gob-solid coal areas are encountered, it is best to approach the boundary from the gob to the solid side and keep the approach angle under 30°. Using these fundamental principles and through careful planning and site-specific experience, potential problem areas can be delineated and design changes implemented to control interactions.