Mine Safety and Health Administration,Pittsburgh,PA 15236,USA

Keywords:Coal bursts Underground mining Room-and-pillar Destressing

ABSTRACT In order to reach a large,untapped reserve of high-quality coal,D8 Cloverlick Mine proposed to mine a corridor nearly 600 m deep beneath the Benham Spur of Black Mountain,Kentucky’s highest peak.D8 Cloverlick Mine was extracting the Owl seam,but the corridor’s route lay approximately 20 m above century-old mine workings in the C-(Darby) seam.Adding to the concern,three serious coal bursts had recently occurred in nearby Owl seam workings.Maps of the old workings seemed to indicate that the underlying C-seam had been fully extracted.However,two of the coal bursts had occurred above areas where the C-Seam was also shown as mined out.Mine Safety and Health Administration(MSHA) Technical Support therefore investigated the records of past mining to better understand the old mine maps.Underground conditions observed in current Owl seam workings were also compared with the maps of the old C-seam workings.The study concluded that the presence of hazardous underlying remnants could not be ruled out.To mitigate the burst risk,D8 Cloverlick Mine adopted a strategy of stress probe drilling.A self-propelled coal drill was used to auger 11.5-m-long,small diameter holes in advance of mining.As each hole was drilled,the cuttings were measured to detect the presence of highly stressed coal.Ultimately the crossing was successfully completed without incident.

1.Introduction

For the past century,coal mining in Central Appalachia has powered much of the U.S.economy.Now coalfields that once seemed nearly inexhaustible are now largely mined out.In many areas,the reserves that remain are thinner and deeper than those that were mined before,and they are heavily impacted by the earlier mining above and below them.Deep cover and multiple seam interactions can cause a range of ground control problems,including an enhanced risk of coal bursts.

The storied coalfields of Harlan County,KY,typify this new reality.Remote and almost entirely isolated until early in the last century,the rich coal seams were the target of a mining boom once the railroads were pushed up the Poor Fork of the Cumberland River.Major corporations like U.S.Steel and International Harvester opened big mines and built large company towns from scratch.For decades,Kentucky was the nation’s number one coal state,and Harlan was Kentucky’s number one coal county.But today’s Harlan County miners face serious ground control hazards even to gain access to the economic reserves that are left.

Early in 2017,D8 Cloverlick Mine proposed to develop a corridor in the Owl seam beneath the Benham Spur of Black Mountain in order to reach a new reserve block of high-grade coal.Black Mountain is the highest mountain in KY,and the depth of cover above this crossing approaches 600 m.The entire route also lies approximately 21 m above nearly century-old mine workings in the C-seam (also known as the Darby seam).Adding to the concern,two severe coal bursts had recently occurred in D8 Cloverlick Mine,and a third had occurred at an adjacent Owl seam mine.

MSHA Technical Support conducted a burst risk analysis of the proposed crossing,using the approach outlined by Mark and Gauna[7].Technical Support concluded that:

The preponderance of burst risk factors evaluated for the Benham Spur crossing are rated ‘‘high”,indicating that there is an elevated risk of bursts.The most significant factor is the past history of bursts at this mine.The great depth of cover,combined with overmining possible unmapped remnant pillars with 70 feet of interburden,also point towards an elevated burst risk.These conditions are at the extreme end of the experience base in the U.S.coalfields.

Of these three factors,nothing could be done about the depth of cover or the past burst experience.On the other hand,much of the concern regarding C-seam remnant pillars was because the past mining was poorly understood.Technical Support also noted that:

Mark and Gauna state that the burst risk can actually be reduced when mining is conducted in de-stressed ground above or below an area that has been mined out[7].However,both of the bursts at D8 Cloverlick Mine occurred in areas where the available digitized maps showed that the underlying Darby Seam was fully extracted,and poor ground conditions have been encountered above other areas where the Darby was shown as mined.These experiences indicated that the available digitized Darby seam maps were inaccurate,incomplete,or incorrectly geo-referenced,and that areas portrayed as fully extracted actually contained scattered isolated remnant pillars.

The historical mine maps seemed to be the key to minimizing the risk of coal bursts caused by underlying remnant pillars.Fortunately,the original hardback maps of the C-seam workings were located at the Natural Resource Partners headquarters in Huntington,WV.High-quality scans of those hardbacks were also available(Fig.1).

At first look,however,these original maps displayed a bewildering variety of colors,letters,pillar sizes and shapes,extraction fronts,and other details(Fig.2).It was by no means obvious which pillars had been extracted,which had been left,or in some cases whether they had actually been developed at all.Adding to the confusion,very few mining dates were recorded on the map,so even the mining sequence was mystery.The investigation therefore had to begin with the history of the C-seam mining.

2.History of C-seam mining beneath D8 Cloverlick Mine

The workings in the C-seam that underlay large portions of D8 Cloverlick Mine belonged to Mine 30,which was operated by the Lynch District of the U.S.Coal and Coke Corporation,a subsidiary of U.S.Steel.Mine 30 was opened in 1920,and mining operations beneath the future D8 Cloverlick Mine reserves continued into the 1940s.The portal was located on Looney Creek in Lynch,just across the creek from Mine 31 (which today is an active tourist attraction).

Mine 30 was a very large operation for its day,designed to employ about 1000 men and produce about 1 million tons/year[8].The mine was a captive operation,and all its production was shipped to U.S.Steel’s coke plants.

Fig.1.Map showing Owl seam workings in the vicinity of Benham spur,showing burst locations,the approximate 520 m overburden contour,and the proposed route of the crossing.

Fig.2.A portion of the projected crossing(shown in pink),illustrating the variety of features shown on the hardback maps of the underlying workings in Mine 30.

Fig.3.Advancing-Robbing mining system used at Mine 30 during the 1920s,with two blocks being recovered on advance and two barrier blocks plus the 4-entry submains to be recovered on retreat.

Mine 30 employed a mining method that was apparently imported from U.S.Steel’s Connellsville,PA mines [5].Called the‘‘Advance Robbing System”(Fig.3)the method blocked out mining districts that were about 750 m wide and as much as 5 km long[8].A typical district was developed with a four-entry submain and two,two-entry haulageways driven approximately 180 m apart,thus defining four 150-m-wide blocks of coal.The two blocks on either side of the submains were initially left as barriers,the other two blocks were fully extracted on advance.Once the panel had reached its full extent,the retreat lines wheeled around to recover the barrier blocks together with the submains.Paul and Plein describe a similar ‘‘half advance-half retreat”mining system that was in use at eight large mines in southwestern PA [3].

Mining was accomplished by drilling,blasting,and then hand loading the broken coal into small rail cars.Each pair of miners would typically advance one lift per shift,approximately 4.5 m wide by 3 m deep.Machine loading was probably not introduced until the 1940s,and continuous miners not until the 1960s.

The conventional wisdom of the day emphasized as complete extraction as possible in pillar work.For example,Rice and Paul stated pillars should be drawn as soon as rooms are driven to their limit in order to have the roof break next to the solid coal,thus largely eliminating creeps and squeezes [5].Nonetheless,Paul and Plein found that the average total recovery at the ‘‘half advancehalf retreat”mines was 82%,which was typical for all the different mining systems they evaluated [3].

The Lynch mines were two of the deepest mines in the country,with cover exceeding 600 m beneath Benham Spur.George Rice,Chief Engineer for the U.S.Bureau of Mines,included them as‘‘Mine G”in his classic 1935 study‘‘Coal Bumps in the Cumberland Coalfields of VA and KY[10]”.Rice wrote that at Mine G‘‘a considerable number of bumps occurred,in which a number of men were killed and injured”.He also said that in response,U.S.Steel adjusted the method used to recover individual pillars,changing from a split-and-fender technique to an outside lift.Bryson also shows the Lynch mines on his map of the Harlan County mines that had experienced coal bursts [2].

The mine maps indicate that pillar extraction activities in Mine 30 reached the deep cover in the mid-1920s.Rice wrote that "under conditions in the Cumberland coalfield,when the cover is more than 450 m there is an increased hazard of bumps in pillar drawing,and when it is 600 m the hazard of severe bumps is acute"[10].He also said that bursts first became a serious concern in the Cumberland coalfields in 1926.Bryson noted that ‘‘small bumps occurring regularly did a great deal to impair the morale of the workmen”,and would likely have resulted in more coal being abandoned during retreat mining [2].

The fatality record at Mines 30 and 31 supports an inference that bursts became a serious problem in the mid-1920s.The Miners Memorial at the Mine 31 Portal lists the names of the miners killed in the Lynch District by year.It shows that during the early 1920s,more than 20 men were killed underground each year,likely owing to the inexperienced workforce.During 1924-1926,the average number of fatalities dropped to 7.5 per year.But the fatality rate increased again when the mining front reached the deep cover,with 12 killed in 1927 and again in 1929.

Economics is another factor that likely affected production at Mine 30.Steel production plummeted as the Great Depression developed after 1929,and it did not return to its previous level until the mid-1930s.Similarly,coal production in Harlan County dropped by 50% between 1929 and 1932,but had fully recovered by 1936.The Lynch mine’s fatality record broadly follows these trends.During 1932-1934,a total of just five miners were killed underground,but 23 men were killed during the following three years as production apparently increased.The end of the decade saw unionization of the Lynch and other Harlan County mines[1].Safety apparently improved significantly during the 1940s,with an average of ‘‘just”three fatalities per year.

3.C-seam maps of Mine 30

In addition to showing the extent of mining,the hardback maps also show the mining sequence.Dates were sometimes included for development mining,and a combination of colors and letters were used to show the retreat mining.From 1921 up through 1928 each letter corresponded to one quarters’ mining activity(i.e.,from ‘‘A”through ‘‘F-1”is 32 letters,representing 8 years),and afterwards one letter showed 6 months’ worth of mining.

For example,the lower portion of Fig.2 shows a portion of an angled advancing pillar extraction front tinted red and labeled U-V-W-X.The red color and these letters correspond to the four quarters of 1926.Another type of retreat mining is shown on the upper portion of the map.Labeled J-1 through M-1,the mining was conducted during 1930-1932.

The initial ‘‘advance robbing”reached Benham Spur in early 1926,and the map clearly shows that the mining method changed a couple of years later.One example of the change is shown in Fig.4.The original mining method,used until 1927 (time period Z),features long,staggered pillar lines that cross the rooms at approximately a 45° angle,and double-entry rooms driven to develop the pillars for extraction.By 1929 (time period E-1),the pillar lines are typically perpendicular to the rooms,and only single entries are driven to develop the pillars.The transition from one method to the other apparently took about a year to implement.It is likely that this change was part of the adjustment to bursting conditions that Rice mentions [10].Economics are unlikely to have triggered the change because it occurred several years before the worst years of the depression.

Fig.4.Mining method change at Mine 30.

As noted in Fig.4,A shows the original method,featuring a long,angled pillar line and two-heading room development,B is the transition period,and C displays the new method,featuring a straight pillar line and single-heading room development.

By the 1940s,no attempt was made to recover all the coal,and many large remnants were left in place.By then,Mine 30 had adopted much more conservative mining strategies in the burstprone,deep cover areas,and these likely contributed to its greatly improved safety record during the 1940s.

3.1.Georeferencing

The accuracy of the mine maps is a key concern for this investigation.One issue is how accurately the old Mine 30 maps can be georeferenced to the current D8 Cloverlick Mine map.By comparing a number of georeferencing points on the scans to their State Plane coordinates,vector displacements of their actual locations were estimated.It appears that the scans are slightly distorted,probably due to the maps’ age and to stretch during scanning.The result is that in some locations there seems to be no discrepancy in the location,in others there may be as much as 18-27 m of discrepancy.

3.2.Map accuracy and remnant pillars

The internal accuracy of the Mine 30 map is more significant.There is little reason to question the mapping of the development workings,which seem to have been quite accurately surveyed.An exception could be the unconnected ‘‘dead-end”room headings that are very common features on the map.At first glance,it appears that these rooms stop in unmined pillars.But these are actually room headings that were in the process of being developed at the time the survey was made.When the next survey conducted,the entries were already in the gob because the ‘‘pillars”shown on the map had been recovered.

This leaves the critical question unanswered:would unmined remnants that were abandoned in-between two surveys be shown on the map?In the later,1940s era mining,it is generally clear that many pillars were not recovered.Within the older retreated areas,however,recovery is generally shown as being almost total.Only a few pillars are actually marked as being‘‘unmined”.Yet when similar mining methods were used under shallow cover,in PA,only about 82% of the coal was recovered.It must have been a struggle to achieve even that much recovery beneath the far greater cover of Black Mountain,with a workforce made skittish by burst prone conditions.Rice and Paul provide a list of ways in which coal could be left in place that includes:Miners often gouge out the ribs of pillars and leave only a few feet between rooms,thus weakening the pillars so that they cannot safely be withdrawn;to prevent squeezes,many solid pillars are left,and are lost;and sometimes pillar drawing is done carelessly,and stumps containing 30 to over 100 tons are left and never recovered [5].

It seems likely none of these types of remnants would have been recorded on a survey.Instead,each survey likely shows just the boundary of the worked out area at the time of the survey,without providing any details regarding coal that was left in the gob unless it was a large,whole pillar.

However,the mining sequence provides some clues as to where it would have been very difficult to extract all the coal.For example,where two pillar lines are shown in Fig.4 to have converged from different directions it seems unlikely that all the pillars could have been successfully recovered.The map occasionally shows some pillars that are shown to have been extracted before others that are located inby,and in those cases it seems likely that the inby pillars were abandoned.Pillars within the haulage ways would have been the last to be extracted,and may have also been left unmined in some instances.

It also might be expected that once bursts began to occur,and then when the miners were adjusting to the new mining method that was adopted in response,that more coal might have been left than was usual.But without more detail from original sources,educated guesses must be made.Therefore,the next stage of the project was to conduct underground mapping in the Owl seam to see if patterns of roof and rib conditions observed there might provide clues about the C-seam undermining.

3.3.Underground mapping

An investigation of rib conditions,floor heave,roof distress,and other possible evidence of stress or subsidence was conducted in some of the accessible,deep cover areas D8 Cloverlick Mine.Then the mapped underground conditions were compared with the Cseam maps.The goal was to determine which features of C-seam mining shown on the underlays might correlate with signs of multiple seam interaction in the Owl seam.

Some of the areas that were mapped underground had clearly been impacted by undermining.Fig.1 shows that five Owl seam panels either stopped short or temporarily dropped some entries in the deep cover,central portion of the mine.Mining in the Cseam beneath this area began in the mid-1920s with very regular pillar recovery using the original method.In 1929,a second pillar line was established using the new mining method.At the southern edge of the entire C-seam mining district,there is a 6-to 9-m-wide barrier which separates this pillar recovery from another district that was extracted several years later.The area includes at least four haulageways.

Some of the most severe conditions in these panels were above a C-seam haulageway that that was the boundary between one pillar line that advanced from west to east and another that advanced from east to west.The 0-right Owl seam panel(Fig.1)stopped just at the point where the underlying pillar lines must have passed each other (Fig.5).Rice wrote that ‘‘when two adjacent pillar retreat lines,moving in opposing directions,pass each other”a‘‘special hazard”is created [10].Therefore,it seems likely that the pillars in the haulageway were never recovered.Several of the other haulageways were also associated with distress in the overlying works.

As shown in Fig.5,A is the heavy rib sloughing and poor roof conditions above a two-entry haulageway that separated two retreat panels moving in opposite directions;and B illustrated the poor ground conditions above a pillar marked ‘‘Lost”.

In several other panels,the greatest damage appears to be above a haulageway just at the point where the C-seam extraction began using the new mining method.Several other highly stressed areas lie above retreat mining that was conducted in 1930.In contrast,very little rib sloughage was mapped above mining conducted very slowly between 1931 and 1936,even though it contained another haulageway.Similarly,good roof and rib conditions were observed in a nearby submains,even though two pillar lines converge on the haulageway underneath it.

Fig.5.Severe ground conditions mapped in the 0-right panel of D8 Cloverlick Mine.

3.4.C-seam mining beneath coal burst events

Two coal bursts had occurred in D8 Cloverlick Mine prior to the proposed crossing,and a third had occurred in an adjacent mine.The site of the first one at D8 Cloverlick Mine is shown in Fig.6.Only six pillars had been recovered when the burst occurred.Two features are shown on Fig.6 that may have contributed to the burst.The first are the series of unconnected,apparently‘‘dead-end”rooms almost directly beneath the burst site.These rooms seem to outline a large C-seam pillar that might have been left unmined,but we now believe that the survey ‘‘snapshot”caught the pillars in mid-development and that they were subsequently recovered.On the other hand,the intersecting haulageways directly beneath the burst pillar are probably more significant.After the burst,‘‘subsidence shoulders”were mapped underground that closely follow the orientation of one of these haulageways.These subsidence shoulders are fairly subtle features,however,and they were not picked up by the pre-mining geotechnical survey.

Bursts during development are extremely unusual,but the second burst at D8 Cloverlick Mine occurred within an Owl seam pillar that had not even been fully outlined(Fig.7)[9].In the C-seam this area was accessed by a two-entry haulageway,with pillar extraction during the second half of 1929 using the straight line mining method.However,while consistently poor rib conditions were observed on the left side of the section in Mine X prior to the burst,they did not correspond to any particular feature of the underlying mining.

Fig.6.Retreat burst at D8 Cloverlick Mine,where the outline of Owl seam workings is shown in black,the pillared area is shaded;and the red star is the location of the burst.

Fig.7.Development burst at Mine X,where the red star shows that the partially developed pillar where the burst occurred was directly above a two-entry haulageway.

The burst at the adjacent mine occurred during retreat mining,and resulted in the permanent disability of a shuttle car operator.The burst occurred where the cover was only 360 m.Even more unusual,there was no C-seam undermining in the vicinity of the burst.This incident underlines the unpredictability of coal bursts,as well as their destructive power.

3.5.Owl seam mining experience and proposed crossing

In summary,the underground mapping conducted and the past burst sites pointed to several features of the C-seam mining that are associated with high stress in the Owl seam.Interactions were found to occur above haulageways and the above the boundaries where two pillar lines meet.In addition,many of the distressed areas and the two burst sites overlie areas that were mined during the period 1929-1931.Few problems in the Owl seam were observed when overmining C-seam workings mined before or after the 1929-1931 era.

The proposed Benham Spur crossing was projected to be almost entirely above worked out C-seam mining.However,the C-seam workings contain several of the features that were identified above,including three pillar line boundaries and two haulageways.In addition,about 300 m of the crossing was to be underlain by 1929-1931 workings that are contemporaneous with the mining beneath the two burst sites.Another potential risk is that the geo-referencing might not be entirely accurate.An error of 18 m or so could place the leftmost entry of the corridor directly above another two-entry haulageway.Finally,there is the risk that a burst-prone condition could be created without any indication on the C-seam underlay map.

3.6.Probe drilling and destressing

The NIOSH in 2010 Report to Congress on the hazards of deep cover pillar recovery discussed a number of techniques that have been employed around the world to reduce the risk of bursts [4].Probe drilling,combined with de-stressing,is the one that seemed most relevant to development mining.

Probe drilling is based on the principle that the closer highly stressed coal is to the face or rib,the greater the risk that mining will produce a burst.Probe drilling is used to determine the location of the highly stressed coal.Small (typically 54 mm diameter)auger holes are drilled in a pattern around the entry,and the cuttings are measured as they are removed from the hole.If the coal is yielded and de-stressed,the volume of the cuttings will nearly equal the volume of the hole (after accounting for the swell).If highly stressed coal is encountered,however,the volume of cuttings will greatly exceed the theoretical volume of the hole.In addition,various dynamic effects,such as audible knocking and jamming of the drill rod,may be encountered.

In German mines,probe hole drilling is considered to be the only reliable technique for establishing when a risk of rock burst is present[6,11].If high stress is encountered too close to the face,then de-stressing operations must be conducted before mining can continue.

Fig.8.Stress probe drilling at D8 Cloverlick Mine during the Benham Spur crossing.

Fig.9.Coal drill used to auger the stress probe drill holes.

The crossing plan for D8 Cloverlick Mine called for probe drilling in all entries when depth of cover is over 480 m and C-seam undermining was present.The probe drilling consisted of three holes being drilled prior to each cut of each entry,one as close as practical to the face at a 45° angle into the left and right ribs,and one straight ahead in the face (Fig.8a).The hole pattern was designed such that mining never advanced closer than 6 m to a location that had not been tested for stress.

Drill cuttings were captured and measured.Each 3 lineal m of 54 mm diameter hole produced approximately 12 L of cuttings.Up to 24 L could be produced without triggering additional actions.If more than 24 L of cuttings were produced,all work in the affected place would be stopped pending more detailed evaluations.The drillers also watched for other indicators of excessive stress including excessive knocking/thumping,jamming of the drill rods,floor vibrations,etc.The measurements and observations were recorded on drill log sheets for each probe drill hole(Fig.8b).Holes were also painted and marked on the rib,and records kept and made available in the mine office.All the holes were drilled using a standard self-propelled coal drill (Fig.9).

Ultimately,the crossing was completed without incident.No high stress zones were identified,thus there was no need for dropping entries or for de-stressing.

4.Conclusions

Multiple seam mining under deep cover can create a serious risk of coal bursts.In this instance,a history of past bursts,including one on development,added to the concern.

It was hoped that by deciphering the 90-year-old map of the underlying workings it would be possible to identify a low-risk path for the projected corridor.An extensive effort,employing a number of primary sources as well as underground mapping,was largely successful in interpreting the map.Unfortunately,the presence of highly stressed remnants could still not be ruled out.

Since mine planning could not reduce the burst risk to a comfortable level,engineering controls were necessary.Stress probe drilling is considered the international gold standard for identifying zones of highly stressed,burst prone coal.The holes were drilled in advance of every cut,so that the mining machine was never closer than 6 m from untested coal.

This case history is the first application of stress probe drilling in the U.S.during development mining,though the technique has a long history in Germany and elsewhere.Simple and low tech,it is also effective.While it reduced the corridor’s advance rate its other costs were minimal.Its successful application at D8 Cloverlick Mine may prove useful to future mines facing similar circumstances.