Subsidence problems due to abandoned pillar workings in coal seams F. G. Bell 7 I. A. de Bruyn

Abstract Abandoned mine works are a potential cause of ground subsidence and hence are of major concern where development or re-development is to take place. In the United Kingdom, information necessary to locate potential hazards may be available but occurs in numerous scattered locations and it may require considerable time to access the data. In recent years thematic geological maps have been produced for some of the British coal field areas. The paper describes the historical evolution of the mining system, emphasising the pillar workings which began in the sixteenth century. Methods of investigation such as downhole hammer and geophysics are briefly mentioned and the importance of zoning the land discussed. Stabilisation measures may involve occupying the workings with hydraulically-emplaced fill or cheap bulk grout. In areas of less risk, special foundation structures may be used. Resumé Les travaux miniers abandonnés sont une cause potentielle d’effondrements et doivent donc être sérieusement étudiés dans les zones où des projets de constructions sont prévus. Au Royaume-Uni les informations nécessaires pour situer les risques potentiels existent généralement, mais éparpillées dans des endroits variés, et l’accès aux données peut prendre beaucoup de temps. Au cours des dernières années, des cartes géologiques thématiques ont été publiées pour certaines parties des bassins charbonniers britanniques. L’article décrit l’évolution historique des méthodes d’exploitation, en insistant sur le système des piliers abandonn´es, qui a commencé au 16ème siècle. Les méthodes de reconnaissance telles que le marteau fond de trou et la géophysique sont brièvement évoquées et on insiste sur l’importance

Received: 19 December 1996 7 Accepted: 13 December 1997 F. G. Bell (Y) 7 I. A. de Bruyn Department of Geology and Applied Geology, University of Natal, Durban, South Africa. Fax: c27 31 260 2280 e-mail: [email protected]

de réaliser des zonages. Les mesures de confortement des sites peuvent inclure du remblayage hydraulique ou des injections peu coûteuses. Dans les zones moins exposées, on peut utiliser des fondations spéciales. Key words Coal mining 7 Pillar and stall 7 Subsidence 7 Grouting 7 Investigation 7 Hazard zonation

Introduction Urban re-development in industrial societies, together with the reduction in the number of suitable sites, has meant that areas formerly regarded as unsuitable may now be considered for development purposes. In Britain most of the large industrial centres where re-development is taking place are underlain by rocks of Coal Measure age hence the possibility of past mine workings beneath a site is a major consideration. The old workings not only present a hazard in the form of cavities beneath the surface, but the induced stresses in the bedrock may result in fractures or joints being opened. Old shallow workings close to outcrop should be expected in any urban area where exploitable beds are not covered by thick superficial deposits. Mining has been undertaken in many of the British coal fields for several centuries but the first statutory obligation to keep mining records dates only from 1850 and it was not until 1872 that the production and retention of mine plans became compulsory. However, where old mine plans do exist, they are often inaccurate and may not portray the activity at the end of the period of mining. In Britain, coal mining began on a significant scale in the thirteenth century when drifts and adits were driven in the base of quarries, from open pits or along the coal outcrop in hilly terrain. At that time the workings extended as far as natural drainage and ventilation permitted. By the fourteenth century, outcrop workings had largely given way to bell pits. The shafts of the old bell pits were usually 1.3 m in diameter and the pits rarely exceeded 12 m in depth. Extraction of the coal was around the base of the shaft Bull Eng Geol Env (1999) 57 : 225–237 7 Q Springer-Verlag

225

F. G. Bell 7 I. A. de Bruyn

Fig. 1 a Bord-and-pillar workings, Newcastle upon Tyne, 17th century. b Pillar-and-stall workings, South Wales, 17th century. c Stoopand-room workings, Scotland, 17th century. d Staffordshire square work, developed for conditions involving spontaneous combustion. Air-tight toppings could be placed in the narrow access ways to prevent spread of fire

until roof support became impractical and another shaft was sunk nearby. When a coal seam occurred more than 7 m below the surface, bell pit mining was frequently replaced by headings which radiated from a shaft into the coal seam. The pillars of coal between the headings were generally the only support to the overlying strata. It would appear that the layout 226

Bull Eng Geol Env (1999) 57 : 225–237 7 Q Springer-Verlag

of the mine was unplanned, consisting of a complex of interconnecting headings hence support pillars were irregular in shape and size. The pillar and stall method of extraction developed in the sixteenth century when underground workings were shallow but not extensive. Initially the pillars were rather haphazard in size and arrangement but with time the mining became more systematic and pillars of more uniform shape were formed by driving intersecting roadways in the seam. Also there was a tendency for the size of the stall (room) to increase. In the nineteenth century the normal stall width varied from some 1.8–4.5 m with extraction varying from 30–70% (Wardell and Wood 1965; Garrard and Taylor 1988). Several variations of the pillar and stall method were devised. Figure 1 shows four typical examples

Subsidence over pillar and stall workings

utilised in Scotland, Newcastle, Staffordshire and South Wales. However, there was no single method of working used consistently even in a particular coal field and hence now, some 100–200 years later, it is not possible to predict the layout of individual mines.

Potential failure of pillar and stall workings In pillar and stall workings the pillars sustain the entire weight of the overburden, which results in the pillars themselves and the rocks immediately above being subjected to additional compression. Between the pillars, the unsupported roof beds tend to sag, adding further stress to the edge of the pillars. Although the intrinsic strength of coal varies, the stability of pillars is largely controlled by a) the ratio of seam thickness to pillar width; b) depth below ground level; c) dimensions of the mined void. Madden and Hardman (1992) analysed a number of mines in South Africa and showed 1. 65% of the failures occurred at less than 70 m depth; 35% occurred at less than 40 m depth; 2. 43% of collapses had pillar widths of less than 6 m; 20% had pillar widths of less than 4 m; 3. 65% of the failures occurred where extraction exceeded 75%; 4. 60% of the pillar collapses had a pillar width to mine height ratio of less than 2; no pillar failure was recorded with a pillar width to height ratio above 4. Although stresses acting on the corners and sides of the pillars can give rise to fracturing, at depths of less than 50 m the stresses are usually too small to cause significant fracturing. Conversely, at depths greater than 120 m, the theoretical stress values would exceed the unconfined compressive strength of the strongest coals. Individual pillars in dipping seams tend to be less stable than those in horizontal strata as the overburden produces a shear force on

the pillar. The mode of failure is also influenced by the character of the roof and floor rocks. Pillars in the centre of the mined void are subjected to greater stresses than those at the periphery. Collapse in one pillar can bring about collapse in others in a chain reaction as increasing loads are placed on the adjacent pillars. Subsidence due to pillar failure in shallow mines can be greater than that which occurs in mines where total extraction has taken place as the pillars restrict the bulking of strata immediately above them. As a consequence of roof failure in the area between the pillars, they commonly become higher over time and hence their structural stability is reduced. Fallen roof material can provide confinement to pillars as well as surcharging weak floor materials which otherwise may heave. In addition to heave between the pillars, squeezes or crushes occur when the pillars are pushed into weaker horizons at floor or roof level (Piggott and Eynon 1978). Where a significant structural weakness such as a fault occurs, however, the combination of deconfinement and reduced shear strength will increase the likelihood of subsidence. Stephenson and Auguenbaugh (1978) noted that the yielding of a large number of pillars can create a shallow subsidence over a large surface area. Bruhn, Magnuson and Gray (1981) also recorded the presence of such trough or basin-like forms in the coal fields around Pittsburgh, USA. Marino and Gamble (1986) subsequently referred to such broad shallow subsidences as sags. Here the ground surface is displaced radially inwards towards the area of maximum subsidence (Fig. 2). Sag movement depends on the mine layout and the geology as well as the topographic conditions at the surface. The inward radial movement generates tangential compressive strain and circumferential tension fractures. Marino and Gamble (1986) found that the magnitude of the surface, tensile and compressive strains range from slight to severe. Sags tend to develop suddenly, the major initial movements lasting in some instances for about a week with subsequent displacements occurring over varying lengths of time. The initial movements can produce a relatively steep-sided bowl-shaped

Fig. 2 Typical sag subsidence configuration

Bull Eng Geol Env (1999) 57 : 225–237 7 Q Springer-Verlag

227

F. G. Bell 7 I. A. de Bruyn

area. The actual profile of the sag may develop more slowly and will vary considerably depending on the mine layout and the geological conditions. As a consequence, it is impossible to accurately predict the eventual profile. Carter, Jarman and Sneddon (1981) described subsidences at Bathgate near Edinburgh, attributed to failure in old mine workings at shallow depth which had a high extraction ratio. Much of the subsidence took place within a two week period and 95% had occurred within a year. The subsidence, which extended to some 0.3 m, took the form of a broad basin affecting an area of some 7 500 m 2. As further movement occurred some two years later, it was considered that the initial collapse over-stressed the adjacent updip pillars resulting in their subsequent failure. Very frequently, when a mine neared the end of its life, the pillars were robbed. The extent of the pillar robbing is believed in some cases to simulate long wall conditions, especially when the extraction ratio exceeds 85%. At moderate depths pillars have frequently been crushed and as a result the roof may have lowered such that the mined void (goaf) is almost closed. At shallow depths, however, the lower overburden pressure may be inadequate to crush the pillars hence the extent of goaf closure is much more variable. This, together with the fact that the roof may have been supported by timber props which with time may have deteriorated, means that assessment of these areas for engineering purposes is very difficult. Whenever possible, the miners avoided removal of any material other than the exploitable mineral. On occasions, however, it was clearly advantageous to produce driveways of convenient height and hence even if the roof material was strong, competent rock, they may have excavated into it. Where such over-dig occurred, the “waste” material was frequently back-stowed in the mine itself. This backstowed material would clearly reduce the height of the void and hence the potential for subsidence. Rarely, however, are areas of back-stowing recorded on old mine plans. Accurate prediction of potential subsidence requires reliFig. 4 Series of crown holes exposed at the surface, Witbank coalfield, South Africa

228

Bull Eng Geol Env (1999) 57 : 225–237 7 Q Springer-Verlag

Fig. 3 Void migration in old workings exposed by open-casting, South Staffordshire coalfield, UK

able data regarding the layout of the mine, the method of mining and the extraction ratios. Such information rarely exists. Where accurate data are available, the method outlined by Goodman, Korbay and Buchignani (1980) may be used to evaluate collapse potential. This method involves calculating the vertical stress on any pillar and hence requires assessment of the tributary area from which the load will be transferred to the individual pillar (Table 1). When a structure is to be built above old pillar and stall workings, a proportion of its load will be added to that of the overburden in the appropriate tributary area.

Void migration Where pillars remain relatively stable, the area between the pillars which has suffered a stress reduction may experi-

Subsidence over pillar and stall workings

ence void migration due to the spalling of the mine roof (Fig. 3). The bulking of the spalled material results in the void eventually choking itself with fallen material arresting the upward migration of the void. In areas of shallow mining, however, there may be an insufficient thickness of material for adequate bulking to take place and hence the voids may migrate through to the surface to produce a sudden crown hole collapse (Fig. 4). The factors which influence whether void migration will take place include: 1. the width of the unsupported span, 2. the thickness and dip of the seams, 3. the height of the workings, 4. the depth of overburden, 5. the nature of the cover rocks, particularly their shear strength and the geometry of the discontinuities, 6. the ground water regime. Garrard and Taylor (1988) summarised four methods which have been used to predict the collapse of roof strata above mined stalls: 1. clamped beam analysis which considers the tensile stress of the immediate roof rocks (Wardell and Eynon 1968; Hoek and Brown 1980); 2. bulking equations which consider the maximum height of collapse before a void is choked (Tincelin 1958; Price Malkin and Knill 1969; Piggott and Eynon 1978); 3. arching theories which estimate the height to which a void will occur before a stable arch develops (Terzaghi 1946; Szechy 1970); 4. coefficients based on experience and field observation which act as multipliers of either seam thickness or span width (Walton and Cobb 1984). If, as Piggott and Eynon (1978) suggest, the collapse voids adopt various geometrical forms such as conical, wedge and rectangular, different expressions must be used to calculate void migration (Fig. 5). These authors showed that for a particular width of mine opening (B) the height of collapse or migration (Dc) is a function of the original height of the mine opening (h) and the bulking factor (bf) of the overlying strata. They derived the following expressions for obtaining the height of the void migration for the three geometrical forms mentioned: 1. Conical collapse Volume of intact beds V0 p

pB 2 Dc ! 4 3

Total volume of collapse zone VcpV0 c

pB 2 !h 4

(1a) Fig. 5 Form and degree of void migration (after Piggott and Eynon, 1978). A Diagram showing notation relating to maximum (1b) height of collapse and geometry. B Postulated variation in maximum height of collapse of different modes of failure and bulking factors

But bf p

VcPV0 V0

(1c)

Hence 3h 3h or Dc p bf p 4 bf

2. Wedge collapse V0 p

(1d)

BDc1 2

(2a)

where 1plength of collapsed roadway Bull Eng Geol Env (1999) 57 : 225–237 7 Q Springer-Verlag

229

230

Bull Eng Geol Env (1999) 57 : 225–237 7 Q Springer-Verlag

Bieniawski (1967, 1970)

Salamon and Munro (1967) and Salamon (1967)

Wardell and Wood (1965)

Coates (1965)

1.1z

psi

tpp400c100

(1) tpp1320

w psi h

w 0.46 psi (Dimensions in ft) h 0.66 (wcs) 2 (2) svp1.1z psi (w) (3) Factor of safety (FS) against pillar failure strength of pillar FS p pillar loading pressure

1P

r 100 rpextraction ratio z psi sv p 1Pr

sv p

Based on tests carried out on coal pillars in South Africa, the pillars varying in height and width from 2 to 6.7 ft.

These expressions are based upon a statistical analysis of a survey of pillar dimensions in both stable (98 cases) and collapsed (27 cases) mining areas in South Africa. Summary of data Stable Collapsed 1. Depth (m) 20–200 21–192 2. Room height (m) 1.2–4.9 1.5–5.5 3. Pillar width (m) 2.7–21.3 3.4–15.9 4. Extraction ratio 38–89 49–91 5. Pillar width : height ratio 1.2–8.8 0.9–3.6 Salamon and Munro (1967) concluded that 99% of collapses occurred at safety factors lower than 1.48. Holland (1965), however, had previously suggested that a factor of 1.8 was generally neccessary and that in critically important areas it was 2.2

According to Coates this expression gives valid results where the depth of the workings is less than half their extent, as is the case with most old mine workings. When a structure is to be built over old pillared workings the additional weight on the pillars can be estimated simply by adding the weight of the appropriate part of the structure to the over-burden pressure supported by the pillar. The total load on a pillar determined in this manner is probably greater than the true load. Moreover the distribution of load over the area of the pillar is not uniform for stress is concentrated at the pillar edges. The transfer of the weight of a surface structure to residual pillars is normally calculated on the assumption that the additional weight acts vertically downwards and that there is no lateral spreading of the load. This is based on the fact that void migration causes dilation above the seam as well as joint opening so reducing or preventing such lateral distribution.

This expression is based on two assumptions, namely, that each pillar supports the column of rock over an area which is the sum of the cross sectional area of the pillar plus a part of the room area; and that the load is vertical only and uniformly distributed across the cross sectional area. Steart (1954) has pointed out, however, that the pressure is not evenly distributed, it frequently assumes a parabolic distribution with maximum pressure exerted on the central pillars of the mined out area. He further stated that the pressure gradient gradually increased with increasing area of development to a maximum given by Denkaus and that this maximum was reached when the development, if roughly circular, attained a radius equal to the depth, divided by the ratio (acb) 2/a 2 (Figure 7).

This expression was developed in relation to theories of arching and more particularly to the vertical load above a tunnel. It has been suggested that it could be used to calculate the total vertical load, including any proposed surface structure, above old mine workings.

ρ 0.5 B (1Pexp PK tan f z/B)cq exp K tan f z/B K tan f svpvertical pillar loading pressure Kpconstant representing ratio between horizontal and vertical pressure Bpbreadth of room ρ pdensity or unit weight of overburden zpdepth wpangle of internal friction qpuniform surcharge carried by soil per unit area 1 svpwz 1Pr 2abcb 2 rpextraction ratiop (acb) 2

Terzaghi (1943)

Denkaus (1962)

Based on the results of in situ tests on square pillars in Pittsburg bed. Pillars varied in width from 1 to 5.3 ft and height from 1.4 to 5.3 ft.

Pillar strength (tp)p2800!w 0.5 psi h 0.85 wpwith of pillar in inches hpheight of seam in inches

Greenwald et al. (1941)

sv p

Remarks

Expression

Author(s)

Table 1 Some expressions for determining the strength of and stress on coal pillars

F. G. Bell 7 I. A. de Bruyn

MacCourt et al. (1986) extended the work of Salamon and Munro (1967). Like Salamon and Munro, MacCourt et al. adopted tributary area theory, assuming that each pillar carried the mass of superincumbent strata immediately above it. This theory is valid when pillars have reasonably uniform geometry and are mined over an area where the width of mining exceeds the depth of seam. After a survey of recent pillar collapses in South Africa collieries, MacCourt et al. concluded that the majority occurred when the depth was less than 60 m, the pillar width was less than 6 m, the pillar width to height ratio was less than 1.75 and the extraction ratio exceeded 70%. They found that subsidence as a proportion of mined height decreased from around 0.8 at very shallow depths to between 0.1 and 0.5 at depths exceeding 100 m.

(1) tpp7176 w 0.459 h P0.66 kPa ρ qzf c 2 (2) sv p w2 ρ pdensity of overburden gpacceleration due to gravity zfpdepth to floor of workings cppillar centre distance tp (3) FS p sv

MacCourt et al. (1986)

b

Allows zones of minimum stability in abandoned pillar and stall workings to be located. Locations and dimensions of pillars plotted on mine plan. Factor of safety for each pillar determined and plotted on plan. Pillars which have factors of safety of less than one are considered potentially unstable and removed from plan. Their tributary areas are re-assigned to adjacent pillars and the calculation repeated. In second calculation more pillars may fail. They are removed for further reiteration of the calculation, if so required. Although method requires knowledge of exact shape of pillars, as well as strength of rocks involved, it offers practical approach to recognizing where the most potentially unstable areas exist such that appropriate measures may be taken optimally.

a

a

Fig. 7 Plan view of loading on pillars

b

USC!N-shape!N-size FS UCSpunconfined compressive strength N-shapep(0.875c0.250 w/hp) hppheight of pillar FSpfactor of safety (2) svpsvi!At ApPAw svipinitial vertical stress at roof level Atparea tributary to each pillar Appcross sectional area of pillar Awparea of pillar lost from load carrying (3) FSptp/sv

(1) tp p

2

b

Goodman et al. (1980)

1

b

Wilson (1972) suggested that a pillar was surrounded by an outer yield zone in which the stress distribution varied in linear fashion from zero at the surface to the point of failure at the pillar core. The constraint given to coal in the pillar core can increase its strength appreciably. The six expressions take no account of the strength of the coal and it has been suggested that although they may be admissible for deep working they could give underestimates of the strength of pillars at shallow depths.

(1) Wide pillars, w 1 0.003 mz ft (1.1) Square pillars svp4pz(w 2P3wmz!10 P3c3m 2z 2!10 P5) tons (1.2) Rectangular pillars sgp4pz(w)P1.5(wchmz!10 P3c 3m 2z 2!10 P5)tons (1.3) Long pillars svp4pz(wP1.5mz!10 P3)tons/ft run (2) Narrow pillars, w~0.03 mz ft (2.1) Square pillars w3 svp44p tons m (2.2) Rectangular pillars wc1 w w2 svp1333ρ 1P c m 2 3 (2.3) Long pillars w2 svp667p tons/ft run m where svpload which pillar will carry, in tons ρ paverage density of rocks in tons/ft 3 lplength of pillar in ft mpheight of roadway in ft wpwidth of pillar in ft zpdepth of overburden in ft

Wilson (1972)

Subsidence over pillar and stall workings

Bull Eng Geol Env (1999) 57 : 225–237 7 Q Springer-Verlag

231

F. G. Bell 7 I. A. de Bruyn

and VcpV0cB1H

(2b)

Hence bf p

2h 2h or Dc p 4 bf

(2c)

3. Rectangular collapse V0pB1h

(3a)

VcpV0cB1h

(3b)

Hence bf p

h h or Dc p Dc bf

(3c)

It is frequently maintained that the maximum height of void migration is directly proportional to the thickness of the mined seam and inversely proportional to the change in the volume (bulking) of the collapsed material. The height of the collapse appears to be independent of the width of the excavation, although clearly the larger the span the more likely that a collapse will occur. Although in exceptional cases the maximum height of migration may extend to ten times the height of the original roadway, it does not usually extend further than three to five times the height except at the intersection of roadways. Wider spans at these intersections give rise to higher tensile and shear forces in the roof beams thus producing conditions which are more prone to roof failure. When a competent bed thicker than 1.7 times the span width occurs in the roof rocks, the competent bed will invariably arrest the collapse. Clearly, however, the nearer the void is to the ground surface, the more likely the competent rocks will be weakened by weathering hence failure may occur even in these strata. Garrard and Taylor (1988) found the majority of void/collapse height inter-relationships were explained by variations in the width of the workings and/or the type of roof rock. Old workings with roofs of interbedded strata were generally found to be wider and to have collapsed to higher levels than workings in which the roofs were formed of sandstones, siltstones or mudstones. In addition, the collapse structures in polylithological sequences had significantly greater collapse height to width ratios and steeper failure surface angles than occurred in monolithological sequences. It would appear that in an interbedded sequence the lack of coherence in the rock unit facilitates delamination, bed separation and fracture. If a competent rock beam is to span an opening, its thickness should be at least twice the span width in order to allow arching to develop. A bed of sandstone will usually arrest a void especially if it is located some distance above the immediate roof of the mine working. In other strata, however, most voids are bridged when the span decreases to an acceptable width through corbelling rather than when a more competent bed is encountered. Chimney-type collapses can occur with abnormally high levels of migration in massive strata in which the joints diverge down232

Bull Eng Geol Env (1999) 57 : 225–237 7 Q Springer-Verlag

wards. Walton and Cobb (1984) mentioned that three times the width of the stall appeared to be an upper limit to the extent of most void migrations in abandoned coal mines, although 1.5 times the span is another frequently quoted factor. The height of collapse was shown by Garrard and Taylor (1988) to be proportional to its width, which itself was related to both the type of roof rock and the thickness of the worked coal seam. Sandstone overlaid the widest workings whereas the narrowest workings had a mudstone roof. Nonetheless, voids with sandstone roofs collapsed to higher levels than those with mudstone or siltstone roofs. Garrard and Taylor also suggested that where the span of the working is known or can be estimated, the collapse height could be obtained by multiplying this by 2.68 in the workings they investigated. However, in view of the difficulty of obtaining stall dimensions in abandoned workings, the height of the void migration is normally assessed from the thickness of the mineral extracted and the difference between the density of the roof material in situ and when collapsed. Few of the workings investigated by Garrard and Taylor (1988) had completely bulked. From their observations these authors questioned the relationship derived from bulking theory and suggested that the maximum height of collapse was likely to be equal to 9.8 times the seam thickness (see also Piggott and Eynon 1978). More commonly, however, the height of collapse is three to six times the seam thickness as recorded by Statham, Golightly and Treharne (1987) in the South Wales Coalfield. Exceptions occur however and void migrations in excess of twenty times the seam thickness have been recorded, particularly in dipping seams where large quantities of water can re-distribute the fallen material. Statham et al. (1987) noted that in South Wales the upper limit of the migration ratio in gently dipping seams (~57) was 8 compared with about 18 in steeply dipping seams ( 1 127). The re-distribution of collapsed material may lead to the formation of “supervoids” and their migration to rock head then results in large scale subsidence at the ground surface. Under such circumstances, simple analysis using the bulking factor proves inadequate (Carter 1985). It should be noted that the depth of superficial deposits should be included in the thickness of the overburden and that where weak superficial deposits exist, they may “flow” into the void hence reducing the depth of the surface depression but frequently increasing its diameter.

Investigation in subsidence areas The location of sub-surface voids due to mineral extraction is of prime importance when re-developing areas previously undermined. Attempts should be made to determine the number and depth of mined horizons, the extraction ratio, the pattern of the layout and the condition of the old pillar and stall workings. The sequence and type of roof rock may provide some indication as to whether void migration

Subsidence over pillar and stall workings

has taken place and hence its likely extent. Wherever possible, it is important to establish the state of the old workings, noting whether they are open, partially collapsed or collapsed and the extent of fracturing or bed separation. Each investigation should be designed specifically for the particular coal field and project. An appropriate desk study and reconnaissance survey should be undertaken. Sources of information in Britain are listed in Table 2 but caution is advised when examining old maps and mine plans as these are likely incomplete and inaccurate. Whilst the use of aerial photography and remote sensing imagery may facilitate the detection of subsidence features in rural areas, they are generally of little value when the mined area has been developed previously. Over the past 30 years, attempts have been made to develop geophysical methods for the location and delineation of abandoned mine workings. A variety of surface traversing techniques are available which provide readings at close intervals, facilitating the location of shallow voids. The selection of the most appropriate technique necessitates consideration of four parameters: penetration, resolution, signal to noise ratio and contrast in physical properties. From careful analysis of the geophysical trace, it may be possible to distinguish man-made cavities from natural cavities as the former are frequently created using blasting and/or machinery, resulting in fractures which may extend up to two diameters away from the cavity. Whilst some authors consider it is possible to detect a cavity at a depth of less than twice its diameter, McCann, Jackson and Culshaw (1987) concluded that where a cavity was deeper than its own diameter, it was unlikely to produce a measurable anomaly. For further information on geophysical investigations, reference should be made to such literature as McCann et al. (1997). The location of old mine workings is generally undertaken by exploratory drilling using a downhole hammer technique (air or water flush). This is a relatively quick and inexpensive method of investigation and as the drill rods are able to free-fall, any void encountered is readily identified. The drilling is normally carried out on a grid pattern such that where appropriate, infilling holes can be undertaken at closer spacings. One of the principal objectives when investigating old mine workings is to consider their condition as well as their extent. As a consequence, it is advisable to undertake some rotary coring which permits an examination of the strata rather than simply the dust/water returns obtained from the downhole hammer holes. Although it is valuable to have piezometers installed, these only indicate the ground water levels while the use of water absorption tests may assist in assessing potential grout requirements. Whilst it may be possible to enter some shafts and/or old galleries, extreme care should be taken not only regarding the stability of the strata but also the possible presence of toxic gases which are generally not visible. The use of downhole cameras or closed circuit television may be preferable to man entry. Where the mines are flooded, rotating ultrasonic scanners may be used within the voids. Scanning can be carried out horizontally and/or at an angle as

required to maximise the information obtained regarding the cavity.

Hazard zonation Mine hazard assessment is usually undertaken on a site basis. Regional assessment, although less frequent, can provide an overview of the problems involved such that planners, for example, can determine the most difficult areas for development and avoid imposing unnecessarily rigorous conditions in areas where such restrictions are not warranted. One of the first attempts at hazard zoning in an area of old mine working was made by Price (1971). After a detailed investigation at a site at Airdrie, Scotland, he was able to propose “safe” and “unsafe” zones. In the safe zones the cover was regarded as sufficiently thick to preclude a subsidence hazard, ie some 10 m of rock or 15 m of till. In these areas normal foundations could be used for two-storey dwellings. On the boundaries between the safe and unsafe zones, the dwellings were constructed with reinforced foundations, or rafts as an added precaution against unforeseen problems. Development was prohibited in the unsafe area. In this way, Price produced what we would now consider to be a thematic mining information plan. In recent years, such thematic maps have been made for many areas of the United Kingdom. Early thematic maps produced by the British Geological Survey distinguished between areas where the mining was within 30 m of the surface and where it was below this depth (McMillan and Brown 1987). This 30 m depth is based on limited information and therefore subject to interpretation. It assumes that bulking factors of 10 to 20% will affect the strata involved in void migration. However, where roadways much higher than the seam thickness have been excavated, such generalisations are not applicable. Further, Carter (1985) noted that old workings at depths exceeding 50 m have collapsed, affecting the ground surface. In some areas the thematic maps only record workings where mine plans are available hence only the more recent mining is included. Whilst such positive information is clearly advantageous, the lack of any indication of potentially undermined areas may lead to a false sense of security or conversely, the blighting of areas which may not warrant it. The recognition of multiple seam working led to the requirement that areas of shallow working (~30 m below rock head) should be identified in terms of seams worked. In these areas separate maps were prepared depicting: 1. total known mining, 2. current mining, 3. known mining within 30 m of rock head, 4. location of shafts and drill holes which encountered shallow workings, 5. mining for minerals other than coal and ironstone. In the Fife area of Scotland, known and inferred shallow mine workings are differentiated on the same map. Further, an indication of the area in which mining might be Bull Eng Geol Env (1999) 57 : 225–237 7 Q Springer-Verlag

233

F. G. Bell 7 I. A. de Bruyn

Table 2 Sources of information in Britain (after Stratham et al. 1987) Source

Type of information

Completeness and reliability

British Coal Abandoned Mine Plans Department Opencast Executive

Abandoned mine plans, working mine plans, open-cast site files, subsidence incidences, local knowledge

Usually quite reliable but incomplete; degree of incompleteness declines near to outcrop. Best information if from open-cast records

Mineral Valuer

Enquiries for site developments, subsidence incidences; mineral leases

Reliable and complete within itself but mainly site specific

British Geological Survey

Geological maps and memoirs, field slips and note books, drill hole files (much information duplicated from local authority holdings). Represents a summary and distillation of mining information to procedure a geological interpretation

Complete but reliability largely determined by information held by British Coal. Generally good for study area

Ordnance Survey

Various editions of OS maps giving historical background, location of mines etc.

Complete and reliable

Aerial photographs

Subsidence features visible, often at several different dates

Complete but reliability depends on interpretation

Local authorities, land authorities, land reclamation bodies

Local knowledge of problems and subsidence incidences, site investigation reports, mining reports, drill/grout records

Complete and reliable but usually very site specific

County archives

Historical background, mine plans, leases etc.

Incomplete but usually reliable

Specialist contractors and consultants

Site investigation and mining reports, drill/grout records. Local and general knowledge. Much information already available from local authorities

Site specific

Remote archives

Historical background, leases etc.

Incomplete but usually reliable

Statutory untertakings

Records of problems

Not applicable

General contractors and consultants

Site investigations, mining reports, drill/grout records. Many of documents available from local authorities

Complete an reliable but very site specific

Others

Variable – generally historical background, anecdotes

Often incomplete and unreliable

expected can be obtained by plotting all drill holes which encountered colliery spoil outside areas of working shown on abandonment plans. In an assessment of the degree of risk due to mining incidences related to abandoned mine workings in South Wales, Statham et al (1987) analysed 388 events and found: 1. 64% occurred in open land, posing no threat to persons or property; 2. 21% occurred when people were nearby or threatened property; 3. 15% caused damage to highways, buildings or other property but only one event resulted in minor injury. In the South Wales Coalfield, such an assessment would imply a low level of hazard when considering the stability of old workings. Assuming a typical incident affects an area of 5 m 2 the probability of any event occurring on any 25 m 2 plot is in the order of 10 –7 per year. Even if the number of subsidence incidences which have remained undiscovered increased above the total figure by a factor of three, the overall risk would still be low. Statham et al. (1987) recorded that over 90% of the incidences occurred within 100 m of the outcrop of the coal seam. The authors produced a Development Advice Map for South Wales 234

Bull Eng Geol Env (1999) 57 : 225–237 7 Q Springer-Verlag

which showed two zones within the outcrop area of the worked seams. The zones were based on the migration ratio (thickness of rock cover divided by extraction thickness). Migration ratio values of six and ten were expected to contain 90 and 100% respectively of relevant subsidence incidences. Such maps can provide valuable information to engineers regarding both the feasibility of construction and the likely scale of ground investigation required at a specific site. Culshaw, Bell and Cripps (1988) noted the value of three types of maps – those showing: 1. the extent of made ground, including infilled open cast workings and expanses of mine or open pit waste; 2. the location of known shafts, adits and other mine entrances; 3. the known extent of worked and/or unworked mineral at various depths. The scale of the maps would clearly vary depending on the local requirement and the amount of information available. In Britain they are generally between 1 : 10 000 and 1 : 50 000. In the upper Forth Estuary Gostelow and Browne (1988) attempted to zone the ground underlain by old mine workings in terms of its suitability for different types of foundations.

Subsidence over pillar and stall workings

With the production of thematic maps, it is essential that the users understand how the maps were compiled and the value of the data on which they are based. The fact that areas are indicated as undermined should not necessarily lead to planning blight. Too frequently, users of thematic maps treat them as accurate representations without appreciating the scant information on which the maps were produced. Generally engineering problems in areas of past mining only occur if the development is not properly planned, designed and constructed with reference to the state of the undermined ground. Whilst the engineer should examine thematic maps when they are available, they should never replace site investigations.

Remedial measures Healy and Head (1984) discussed the remedial methods for the treatment of areas of old mine workings. The method to be used should always be evaluated from the specific site conditions, taking due cognisance of the economic implications of alternative methods. One of the most difficult assessments to make is related to the possible effect of progressive deterioration of old mine workings and the potential subsidence risk. The placement of any new structure must be carefully considered to ensure that the ground is not adversely affected by the additional load and/or that the ground is sufficiently stabilised that it will not suffer distress during the anticipated life of the structure. In some instances it may be advisable to move the structure onto more stable ground. For example, following a site investigation for a hotel near Newcastle upon Tyne, UK, it was found that part of the site was underlain by old pillar and stall workings and that coal might be extracted from beneath the site within the next few years. It was recommended that the hotel complex should be re-located and the design of the building altered. A typical example of the re-design of a school and town layout is given in Fig. 6. Following an appreciation of the details of the underground workings, it was possible to re-develop the site in a much safer manner. When sites in the vicinity of an undermined area are assessed it is normal to take precautions within an angle of draw of 257 (from the vertical) as subsidence could occur within this zone as a result of lateral movement. Stabilisation may involve: 1. bulk excavation and re-compaction if the worked seam is very shallow; 2. piled foundations – although there is a danger that the pile may bear on strata which itself has been undermined and hence could become unstable; 3. reinforced beamed rafts to spread the load for low rise building; 4. socketed rafts to provide a jacking facility to re-align the structure should distortion occur; 5. steel mesh reinforcement or geonets for highway support;

Fig. 6 A Proposed layout of structures before site investigation. B Re-appraisal of site layout to avoid problems from old mine workings (after Price, Malkin and Knill 1969)

6. mass infill, either by hydraulic methods in an open system or grouting in a closed system. When mines can be entered there is much to be said for the hydraulic placement of bulk fill material. It is important that wherever possible the fill reaches the mine roof and it may be necessary to construct barriers to restrict the area being filled. In most cases, however, the disused workings are infilled with a cementitious grout emplaced through drill holes on a grid system. Whilst it is essential that the material is pumpable, it should not spread so easily that it extends beyond the perimeter of the area to be treated. Most grouts include bulk materials such as fly ash and sand mixes with a sufficient cement content to provide a strength in the order of 1 N/mm 2. Particularly where there is a significant dip to the seam, it may be necessary to create a perimeter barrier using a thicker grout and/or Bull Eng Geol Env (1999) 57 : 225–237 7 Q Springer-Verlag

235

F. G. Bell 7 I. A. de Bruyn

Table 3 came prominent in the United Kingdom in the sixteenth Depth and thickness of seams at the site of Tyne Bridge, Gates- century. In these workings, pillars were left in place to suphead port the roof rocks. There was no legal obligation to proSeam

Average depth below g.l. (m)

Seam thickness (m)

Nature of workings

High Main Top Main Bottom Main Main Maudlin

16.8 41.2 71.7 126.5 209.0

1.67 1.70 0.66 0.45–0.91 1.07

Pillar and stall Presumed old workings No record of workings Thicker part worked Presumed old workings

gravel/sand emplaced through larger diameter holes around the down-dip periphery of the site. The adequacy of the infilling can be assessed by the quantities of grout injected, the use of water absorption tests and/or downhole cameras (Segatto and Heinz 1992). An example of a treated area is the A1 trunk road approach to the Tyne Bridge at Gateshead in northern England. The bridge is a pre-stressed concrete structure with support piers founded on sandstone, each pier carrying some 2000 tonnes. Generally the sandstone was recorded at 0.6–4.9 m depth but it was not present in some areas where it was believed quarrying had taken place in the past. Below the sandstone, five coal seams were found (Table 3). Following the site investigation, it was considered that the risk of surface subsidence due to void migration from the High Main seam was low but that the stress on the pillars and the seat earth beneath the seam could result in some subsidence damage. In view of the relatively high risk structure involved, it was decided to pressure grout the workings in the High Main seam although the seams below were not treated. A perimeter wall some 1 675 m in length was formed to contain the grout. This barrier was formed of approximately 2 300 m 3 of grout, the mix being one part cement, 2.5 parts fly ash, 10.5 parts sand and 0.1 part bentonite. Infilling within the perimeter wall took 2 850 m 3 of grout using a mix of one part cement, 5 parts fly ash and 18 parts sand. The grout holes were set out at approximately 3 m intervals and provided a 18 m wide treated area beneath each pier. After completion of the infill grouting, the discontinuities in the overlying sandstone were grouted with a 1 : 1 mix of cement and fly ash at a pressure of 275 kPa. The entire thickness of the Main Seam was not worked and it was calculated that for the 0.9 m worked seam the total grout occupied some 40% of the seam volume. The adequacy of the grouting was determined by injecting water in the last holes. As full returns resulted it was assumed that the grouting exercise had been satisfactory.

duce and deposit mine plans prior to 1872 and hence most of the old mines were uncharted and their location can be only be found by appropriate site investigation. Prior to any construction on Coal Measures strata, therefore, it is essential to undertake an appropriate field investigation, the design of which may well benefit from a thorough desk study and/or other preliminary work such as geophysics. Once the nature and extent of the old workings has been determined, the ground can be zoned in terms of the hazard they represent. The depth below ground surface has frequently been used as the basis for hazard zoning, with 30 m below the ground surface or rock head often being taken as the boundary between safe and questionable ground. Special foundations such as rafts and piles have been used in areas of shallow workings but in general ground treatment is undertaken involving the infilling of the voids with a grout formed of cement:PFA:sand.

References

Bieniawski ZT (1967) An analysis of the results from underground tests aimed at determining the in situ strength of coal pillars. CSIR Rep MEG 569, Pretoria Bieniawski ZT (1970) In situ large scale testing of coal. In: Proc Conf In Situ Testing of Rocks and Soils, Brit Geotech Soc, pp 67–74 Bruhn RW, Magnuson MO, Gray RE (1981) Subsidence over abandoned mines in the Pittsburgh Coalbed. In: Geddes JD (ed) Proc 2nd Int Conf on Ground Movements and Structures, Cardiff. Pentech Press, London, pp 142–156 Carter PG (1985) Case histories which break the rules. In: Ford MC, Topping BHV, Whittington HW (eds) Mineworkings 1984. Proc Int Conf Construction. In: Areas of abandoned mineworkings, Edinburgh. Engineering Technics Press, Edinburgh, pp 20–29 Carter P, Jarman D, Sneddon M (1981) Mining subsidence in Bathgate, a town study. In: Geddes JD (ed) Proc 2nd Int Conf Ground movements and structures, Cardiff. Pentech Press, London, pp 101–124 Coates DF (1965) Pillar load. 1. Literature survey and new hypothesis. Dept Mining & Tech Surv, Ottawa, Mines Branch, Res Rep R168 Culshaw MG, Bell FG, Cripps JC (1988) Thematic geological mapping as an aid to engineering hazard avoidance in areas of abandoned mineworkings. In: Forde MC (ed) Mineworkings 1988. Proc 2nd Int Conf Construction in areas of abandoned mineworkings, Edinburgh. Engineering Technics Press, Edinburgh, pp 69–76 Denkaus HG (1962) Critical review of the present state of scientific knowledge related to the strength of mine pillars. J S Afr Inst Min Metall 63 : 59–75 Garrard GEG, Taylor RK (1988) Collapse mechanisms of shallow coal mine workings from field measurements. In: Bell FG, Culshaw MG, Cripps JC, Lovell MA (eds) Engineering geology of underground movements (Eng Geol Spec Pub No 5). Geological Society, London, pp 181-192 Goodman RE, Korbay S, Buchignani A (1980) Evaluation of Following mining from quarry floors and/or hill slopes and collapse potential over abandoned room and pillar mines. Bull the development of bell pits, pillar and stall mining beAEG 17 : 27–37

Conclusions

236

Bull Eng Geol Env (1999) 57 : 225–237 7 Q Springer-Verlag

Subsidence over pillar and stall workings

Gostelow TP, Browne MAE (1988) Engineering geology of the upper Forth Estuary. Rep Brit Geol Surv 16 Greenwald HP, Howarth HC, Hartman I (1941) Experiments on the strength of small pillars of coal in the Pittsburg bed. US Bur Mines, Tech Pap 605 (1939), Rep Invest 3575 Healy PR, Head JM (1984) Construction over abandoned mine workings. CIRIA Spec Pub 32, London Hoek EL, Brown ET (1980) Underground excavations in rock. Inst Mining & Metallurgy, London MacCourt L, Madden BJ, Schüman EHR (1986) Case studies of surface subsidence over collapsed bord and pillar workings in South Africa. In: Proc Symp Effect of Underground Mining on Surface, Sandton. SANGORM, pp 25–32 Madden BJ, Hardman DR (1992) Long-term stability of bord and pillar workings. In: Proc Symp Construction over mined areas, Pretoria. S Afr Inst Civ Eng, Yeoville, pp 37–51 Marino G, Gamble W (1986) Mine subsidence damage from room and pillar mining in Illinois. Int J Mi Geol Eng 4 : 129–150 McCann DM, Eddleston M, Fenning PJ, Reeves GM (1997) Modern geophysics in engineering geology (Geol Soc Eng Geol Spec Pub). Geological Society, London McCann DM, Jackson PD, Culshaw MG (1987) The use of geophysical surveying methods in the detection of natural cavities and mine shafts. Q J Eng Geol 20 : 59–73 McMillan AA, Brown MAE (1987) The use and abuse of thematic information maps. In: Culshaw MG, Bell FG, Cripps JC (eds) Planning and engineering geology (Eng Geol Spec Pub 4). Geological Society, London, pp 237–246 Pigott RJ, Eynon P (1978) Ground movements arising from the presence of shallow abandoned mine workings. In: Geddes JD (ed) Proc 1st Conf Large Ground Movements and Structures, Cardiff. Pentech Press, London, pp 749–780 Price DG (1971) Engineering geology in the urban environment. Q J Eng Geol 4 : 191–208 Price DG, Malkin AB, Knill JL (1969) Foundations of multistorey blocks on Coal Measures with special reference to old mine workings. Q J Eng Geol 1 : 271–322

Salamon MDG (1967) A method of designing bord and pillar workings. J S Afr Inst Min Metall 68 : 68–78 Salamon MDG, Munro AH (1967) A study of the strength of coal pillars. J S Afr Inst Min Metall 68 : 53–67 Segatto P, Heinz WF (1992) Backfilling of coal mines. In: Proc Symp Construction over mined areas, Pretoria, S Afr Inst Civ Eng, Yeoville, pp 237–241 Statham I, Golightly C, Treharne G (1987) The thematic mapping of the abandoned mining hazard – a pilot study for the South Wales Coalfield. In: Culshaw MG, Bell FG, Cripps JC (eds) Planning and engineering geology (Eng Geol Spec Pub No 4). Geological Society, London, pp 255–268 Steart FA (1954) Strength and stability of pillars in coal mines. J Chem Met Min Soc S Afr 54 : 309–325 Stephenson RW, Augenbaugh NB (1978) Analysis and prediction of ground subsidence due to coal mine entry collapse. In: Geddes JD (ed) Proc 1st Conf Large Ground Movements and Structures, Cardiff. Pentech Press, London, pp 100–118 Szechy K (1970) The art of tunnelling. Akademias Kiado, Budapest Terzaghi K (1943) Theoretical soil mechanics. Wiley, New York Terzaghi K (1946) Introduction to tunnel geology. In: Proctor RV, White T (eds) Rock tunnelling with steel supports. Commercial Shearing & Stamping Co., Youngstown, Ohio, pp 17–99 Tincelin E (1958) Pression et deformations de terrain dans les mines de fer de lorraine. Jouve Editeurs, Paris Walton G, Cobb AE (1984) Mining subsidence. In: Attewell PB, Taylor RK (eds) Ground movements and their effects on structures. University Press, London, 216–242 Wardell K, Eynon P (1968) Structural concept of strata control and mine design. Trans Inst Min Metall A 77 : 125–150 Wardell K, Wood JC (1965) Ground instability arising from old shallow mine workings. Proc Midlands Soil Mech Fnds Soc 7 : 5–30 Wilson AH (1972) An hypothesis concerning pillar stability. Min Eng 131 : 409–417

Bull Eng Geol Env (1999) 57 : 225–237 7 Q Springer-Verlag

237