Christopher Newman,David Newman

Appalachian Mining &Engineering,Inc.,Lexington,KY 40503,USA

Keywords:Bump Burst Coal Energy release rate LaModel Numerical modeling

ABSTRACT Two miners were fatally injured when a pillar bump occurred during retreat mining in a southern West Virginia coal mine.The mine was operating in the Eagle seam with overmining in the No.2 Gas and Powellton seams.A coal bump is defined as a sudden and violent failure of coal caused by the release of stored strain energy in the pillar.While significant strides have been made by academia,industry,and regulatory agencies to better understand bump conditions and mitigation techniques,coal bumps represent a long standing,highly site-specific engineering problem in which the exact failure mechanism is not clearly understood.In this case history,a cut-by-cut analysis of retreat mining operations was conducted on the 4 East Main leading up to the pillar bump event.Numerical input parameters were derived from site-specific geologic information and mine geometry for the analysis of pillar stress conditions and energy release using LaModel.An overview of stress conditions in the panel was presented including a precursor event that occurred two crosscuts inby the bump site.The methodology presented in the paper for the evaluation of the fatal bump event can be used for the identification of bump prone conditions prior to development and retreat of a mining area.

1.Introduction

On May 12,2014,a large coal pillar bump occurred while recovering pillars from the No.6 entry on the 4 East Main fatally injuring two miners at the Brody No.1 Mine located in Boon County,West Virginia.A coal pillar bump is defined at the sudden expulsion of coal from one or more pillars due to a violent release of strain energy.While the academic and industry research has identified multiple mechanisms which induce failure,it is commonly concluded that coal bumps occur due to compounding effects of mining induced stress,geology,and mining conditions.Previous case studies [1-5]indicate that coal pillar bumps may occur due to a combination of the following:(1)increased depth of cover,(2)limited interburden between active and adjacent mining,(3)multiple seam mining,(4)massive strata within the immediate and/or main roof or floor,(5) high extraction mining (retreat mining),and (6)the sequence of pillar extraction.

Pillar recovery in the 4 East Main,at the time of the accident occurred,was at an approximate depth of 281 m (925 ft),and in a multiple-seam mining condition with abandoned pillar recovery in the overlying Powellton and No.2 Gas seams.The abandoned mines are in close vertical proximity to active mining in the Eagle seam.The immediate roof of the Eagle seam is composed of a massive coarse grey sandstone within the vicinity of the bump site.Prior to this bump event,no previous incidents of coal bumps had occurred within the Eagle seam as stated within the MSHA Report of Investigation for the fatal bump event [6].

In this paper a preliminary evaluation of the stress and geological conditions within the 4 East Main during pillar recovery is presented.LaModel was employed for a cut-by-cut analysis of mining conditions in which initiated a coal pillar bump on May 9,2014 and a subsequent,fatal,coal bump one break outby on May 12,2014.Through a comparison of stress distribution,pillar stability,and energy release rate results,one is able to identify cuts within the mining cycle which may increase the propensity for the coal pillar to bump.

2.Background information

Two coal bump events occurred while recovering pillars in the 4 East Main.The first bump occurred at approximately 10:00 pm on May 9,2014 while taking the 2nd cut from pillar No.73 in the No.6 entry (Fig.1).While this was initially reported as a 2.4-m (8-ft)long rib roll within the production reports,statements from the shuttle car operator indicate that the pillar had separated from the mine roof approximately 0.15 to 0.20 m (6 to 8 in) with an observable depth of 3.0 to 4.5 m (10 to 15 ft),as stated within the WV OMHS&T Report of Investigation [7].Following the incident,the recovery of pillar Nos.72,73,and 74 was abandoned with retreat mining continuing in the subsequent break(BK 14 in Fig.1).

Fig.1.Section map of 4 East Main (after [6]).

The second bump occurred at approximately 8:45 pm on May 12,2014 while taking the 2nd cut from pillar No.78 in the No.6 entry (Fig.1).The original excavation height was 1.98 m (78 in).However,measurements taken following the bump event found a reduced mining height of 1.85 m (73 in) with 1.37 m (54 in) of loose coal filling the entry[7].Observations made during the accident investigation indicate a separation between the coal pillar and mine roof occurred that extended an undetermined depth into the pillar.No significant damage was observed within the immediate sandstone roof and no roof falls were associated with the bump event [6].

A review of the mining conditions found that both of the bump sites occurred in deep cover(281 m or 925 ft),with the active pillar recovery section located between two gob areas,and abandoned pillar recovery in the adjacent overlying seams.The 4 East Main was developed with eight entries spaced on 26 m (87 ft) by 32-m (105-ft)pillar centers and is separated from two previously pillared panels by a 76-m(250-ft)barrier to the south and by a 54-m(180-ft)barrier to the north.Pillar recovery was conducted in close vertical proximity to the Powellton seam (24 m or 80 ft above the Eagle seam)and the No.2 Gas seam(50 m or 165 ft above the Eagle seam).Pillar recovery in the Powellton seam resulted in a 67-m(220-ft)wide remnant pillar located directly above the pillar burst site in the Eagle seam.Additionally,a gob-solid boundary in the No.2 Gas seam is located above the Eagle seam pillar burst site(Fig.2).The Eagle seam is overmined in the Williamson seam.However,given the average distance between the Eagle seam and development mining within the Williamson seam (approximately 112 m or 370 ft),interactions between the seams are highly unlikely (see Table 1).

Fig.2.Pillar bump locations in the Eagle seam relative to overmining in the Powellton seam and No.2 Gas seam (after [6]).

The immediate roof observed at the bump site was composed of a thin shale overlaid by a thick (5 m or 17 ft) medium,grey sandstone.Progressing from left to right across the No.1 Section,the immediate roof strata transitions from shale to sandstone.A review of core holes within the vicinity of the bump site similarly indicates the presence of a thick(3.0 to 7.3 m,or 10 to 24 ft),medium to coarse,grey sandstone within the immediate roof.The immediate floor consisted of 0.3 m (1 ft) of shale underlain by 1.6 m (5.5 ft) of sandstone,a thin shale parting,and an additional 3.3 m (11 ft) of sandstone.It should be noted that strong massive sandstone formations within the immediate and main roof can be associated with increased stress on the roof and pillars as observed in the field by the authors and others [8-10].

The No.4 East Main mining conditions indicated a strong potential for a coal bump given the multiple seam geometry,geologic structure,high vertical stress,stress redistribution attributable to the overlying mines,and the caving of the Eagle seam roof.Through the application of numerical modeling methods,the mining conditions within the mains were evaluated to determine bump potential with respect to mining induced stress distribution,pillar stability,and strain energy release on a cut-by-cut basis.

3.Numerical modeling

LaModel,a displacement-discontinuity boundary-element program [11],was used for the evaluation of potential pillar coal bumps through the analysis of stress distribution,pillar stability factors,and strain energy release results.The numerical model was developed on a cut-by-cut sequence mimicking the mining cycle of pillar recovery on the No.1 Section leading up to both the initial pillar bump on May 9th and the fatal pillar bump on May 12th.Site-specific mine survey data were used in developing in-seam geometry for the Eagle,No.2 Gas,and Powellton coal seams.Surface topography was similarly incorporated into the model to better simulate the overburden stress within the multiple-seam mining conditions.Given the distance between the Eagle and Williamson seams (108 m or 355 ft),it is unlikely the Williamson seam had a significant influence on the mining conditions within the Eagle seam.Therefore,the Williamson seam mining is not incorporated in the numerical model.

LaModel approximates stresses and displacements on this tabular deposit through the application of a laminated overburden model allowing for the realistic simulation of the flexibility of the stratified sedimentary overburden geology.The numerical model presented in this paper was developed utilizing the best practices for the calibration of the LaModel program for deep cover pillar retreat mining operations as outlined by Heasley et al.[12].To best represent the stress impact of the massive sandstoneimmediately overlying the Eagle seam,the lamination thickness has been defined as 30 m (100 ft) utilizing the default overburden stiffness of 20684 MPa (3000000 psi).The coal strength and gob stiffness were determined with respect to the in-program material wizards.

Table 1.eneralized geologic column for Brody No.1 Mine.

The coal strength and associated yield zone material properties were calculated by the coal material wizard using the Mark-Bieniawski pillar strength equation assuming a default coal strength of 6.2 MPa(900 psi)as similarly assumed by the Analysis of Retreat Mining Pillar Stability (ARMPS),Analysis of Multiple Seam Stability (AMSS),and Analysis of Longwall Pillar Stability(ALPS) programs developed and distributed by NIOSH.The gob stiffness within the model was defined using an in-program material wizard for the No.2 Gas and Powellton seams as shown in Table 2.Given the ‘‘fresh”condition of the gob material within the Eagle seam,the calculated final gob modulus was de-rated to a modulus of 2068 MPa(300000 psi)to better simulate the uncompacted gob material resulting in increased loading on the adjacent pillars and barriers.

LaModel employs a confined core approach to modeling pillar strength.Therefore,element properties are arranged so that the weakest elements are on the pillar rib with in-situ coal strength increasing with distance into the pillar due to confinement(Fig.3).While elements along the pillar rib are subjected to uniaxial conditions,coal within the pillar is under varying degrees of triaxial confinement as is observed in the field.

4.Energy release rate calculation in LaModel

A significant amount of academic,industry,and regulatory work has been conducted to better understand the behavior and mechanism of a coal pillar bump.Historically,the energy release rate (ERR) calculation has shown a correlation to the risk and/or potential of a rock burst.The first case studies were from the deep cover hard-rock and metal mines in South Africa [13].The energy release concept was then adopted for application within coal by the U.S.Bureau of Mines for the prediction of coal pillar bumps[14,15].The ERR calculation was integrated into the LaModel program allowing users to quantify the release of gravitational potential energy from the rock mass into the mining environment in the form of heat and sound as well as dynamically in the form of a pillar coal bump [16].

For each of the six material models represented in LaModel,the total input energy,stored elastic energy,and dissipated energy aredetermined using equations developed at the U.S.Bureau of Mines(see Fig.4).For the analysis of coal bumps,the focus is on the energy released into the mining environment.The dynamic,or released energy is associated with energy changes of the material between mining steps.These energy changes can occur due to either an element’s material changing to another material,material changing to an opening,or changes from an opening to a material.These dynamic changes result in an increase in dissipated energy as well as a release of stored and kinetic energy [17].

Table 2.aterial property input parameters.

Fig.3.Example of confined pillar core approach -No.4 East Main pillar line.

The application of the ERR calculations within LaModel has been previously verified through the comparison of cut-by-cut pillar retreat analyses of the Olga mine conducted with LaModel to the result published by the U.S.Bureau of Mines[16].Additionally,Sears [16]reevaluated a pillar recovery stress analyses conducted by Newman [4]in which five cut sequences were evaluated to determine the location of the highly stressed cuts which would be more likely to cause a coal bump event.

Suddenly a frightful2 thirst seized the King, and as he saw no water near, he mounted his horse, and rode through the neighbourhood looking for a spring

5.Analysis of coal bump potential

Fig.4.Energy calculations for material models (after [17]).

For the analysis of coal bump potential using LaModel,Sears and Heasley[17]suggested a cut-by-cut analysis of retreat mining operations should be conducted while monitoring the change in energy released between mining steps.A review of the analyses conducted on the Olga mine by the U.S.Bureau of Mines as well as the case study provided by Newman [4]indicate that bumps are most likely to occur in cuts where the total energy released has spiked and following a dramatic buildup of total vertical stress.Additionally,these analyses indicate a reduction in energy release following the bump cut.Therefore,in evaluating the mining conditions in No.1 Section,LaModel results for the total vertical stress,element strain safety factor,and total energy release were monitored within a cut-by-cut sequence of retreat mining operations.In accordance with the literature,the potential for a coal bump was determined with respect to increased vertical stress and released energy as well as a reduction in pillar stability.

5.1.Analysis of coal bump potential for May 9,2014 event

The 4 East Main was initially modeled as partially retreated with the fourth lift taken from pillar No.70 in the No.4 entry.For each subsequent cut,the total vertical stress,element strain safety factor,pillar stress safety factor,and total energy release were analyzed and compared to determine if the given cut indicated a potential to bump.While the total vertical stress and pillar safety factor results were used in evaluating the distribution and redistribution of stress on the pillar line,the energy release rate provides an indication of bump potential.As suggested by Sears and Heasley [17],the larger the increase in released energy between steps,the more likely the energy will be dynamically released as a coal pillar bump.A review of the total energy release per step indicates a maximum energy release (approximately 4335680 Nm or 3200000 ft-lbs) occurred due to the extraction of the first lift from pillar No.72 from No.5 entry (Fig.5) indicating the potential for a pillar bump.As retreat mining progresses across the section,the total energy release decreases which is then followed by a sharp increase in energy release following the extraction of the second lift from pillar No.73 in No.6 entry indicating the potential for a pillar bump.

Total energy release results indicate an initial peak in released energy occurred following the extraction of the 1st lift from pillar No.72 within the No.5 entry(Step 6)indicating a potential bump.In reviewing the total vertical stress results for Step 6 and the preceding Step 5 of the numerical model did not indicate a significant rise in pillar stress prior to the potential bump event nor was there as significant redistribution of stress on the pillar line following extraction.Similarly,a review of the pillar safety factors indicates that the pillar core has not yielded (Fig.6).While total energy released results indicated a potential bump event,review of stress distribution and pillar safety factor results do not correlate to a potential bump event.Therefore,it would be unlikely that a bump would occur during the extraction of the 1st lift from pillar No.72.

Fig.5.Total energy release results for bump occurring on May 9,2014.

The subsequent spike in total energy released occurs in Step 18 with the extraction of the first lift from pillar No.73 within the No.6 entry indicating a potential bump.A review of the total vertical stress results for Steps 17 and 18 do not indicate a significant rise in pillar stress prior to the potential bump event nor was there a significant redistribution of stress following pillar extraction.A review of the pillar safety factors was similarly conducted and results indicate that the pillar core has not yielded (Fig.7).While total energy released results indicated a potential bump event,review of stress distribution and pillar safety factor results do not correlate to a potential bump event.Therefore,it would be unlikely that a bump would occur during the extraction of the first lift from pillar No.73.

Fig.7.LaModel results for the 1st lift from pillar No.73 from No.6 entry.

The final spike in total energy released was observed in Step 20 with the extraction of the second lift from pillar No.73 within the No.6 entry indicating a potential bump.In reviewing the total vertical stress results for Step 20 and the preceding Step 19,there was a significant redistribution of stress following the extraction of the second lift.Furthermore,pillar safety factor results indicate that the pillar core has completely yielded (Fig.8).Given an increase in the total energy released,a significant stress redistribution on the pillar line,and a yielded pillar core,the extraction of the second lift from pillar No.73 within the No.6 entry has a high potential to result in a pillar bump.

5.2.Analysis of coal bump potential for May 12,2014 event

The May 12,2014 bump event was similarly analyzed using the methodology previously conducted for the May 9,2014 bump event.A review of the total energy release per step indicates a maximum energy release (approximately 5,690,580 Nm or 4,200,000 ft-lbs)occurred due to the extraction of the first lift from pillar No.78 in No.5 entry(Fig.9)indicating the potential for a pillar bump.As retreat mining progresses across the section,the energy release decreases followed by a sharp increase in energy release after the extraction of the first lift from pillar No.79 in the No.6 entry indicating the potential for a pillar bump.

Total energy release results indicate an initial peak in released energy occurred following the extraction of the first lift from pillar No.78 within the No.5 entry(Step 5)indicating a potential bump.In reviewing the total vertical stress results for Step 5 and the preceding Step 4 of the numerical model,results did not indicate a significant redistribution in stress following extraction.Similarly,a review of pillar safety factor results indicates that the pillar core has not yielded (Fig.10).Therefore,it would be unlikely that a bump would occur during the extraction of the first lift from pillar No.78 within the No.5 entry.

Fig.8.LaModel results for the 2nd lift from pillar No.73 from No.6 entry.

Fig.9.Total energy release results for bump occurring on May 12,2014.

Fig.10.LaModel results for the 1st lift from pillar No.78 from No.5 entry.

6.Discussions

The potential for a coal bump event was determined with respect to increased energy release,vertical stress redistribution,and the stability of the pillar core during a cut-by-cut analysis of the pillar recovery on No.1 Section.While a review of total energy release results identified five potential bump events,the recovery of the second lift from pillar No.73 within the No.6 entry and the recovery of the first lift from pillar No.79 within the No.6 entry were identified as bump events with high potential given the redistribution of stress on the pillar line and the stability of the pillar core in relation to an increase in total energy release.

Numerical modeling results were compared to site observations detailed within the MSHA investigative report as well as site photographs.As previously shown in Fig.1,the initial pillar bump occurred during the recovery of the second lift from the No.73 pillar within the No.6 entry as indicated by LaModel results.The second bump event occurred during the recovery of the second lift from the No.78 pillar within the No.6 entry.However,LaModel results indicate a high bump potential during the recovery of the previous cut;the first lift from the No.79 pillar.Furthermore,in comparing underground condition mapping conducted by MSHA investigators to the element strain safety factor results as determined by LaModel,one finds that the results mimic the rib instability observed in pillar Nos.79 and 80 as well as the rib sloughage denoted on the outby pillar row (Fig.12).

Fig.11.LaModel results for the 1st lift from pillar No.79 from No.6 entry.

Fig.12.Element strain safety factor results for the analysis of rib condition.

7.Summary and conclusions

A preliminary numerical model was developed for the evaluation of coal pillar bump potential within the 4 East Main of the Brody No.1 Mine.On May 9,2014 pillar No.73 bumped during pillar recovery injuring one miner.A second bump occurred on May 12,2014 while recovering pillars Nos.78 and 79 fatally injuring both the continuous miner and mobile roof shield operators.Prior to these events,no pillar bumps had occurred within the Eagle seam.

In reviewing mine mapping and geological information,the mining conditions on the 4 East Main resembled those conditions associated with previous coal bumps in the Appalachian coalfields.The 4 East Main was located in an area of deep cover,had multipleseam stress interaction (barrier pillar in the overlying Powellton seam and a gob-solid boundary in the overlying No.2 Gas seam),and a massive stratum located within the immediate roof and floor.A review of energy release results identified the location(s),or cut(s),which have an increased potential to initiate a coal pillar bump in the 4 East Main.A total of five potential bump events were identified within the LaModel results.These potential bumps were further analyzed in regard to the stress redistribution on the pillar line as well as the stability of the pillar core to further distinguish potential bumps from actual bump events.From the LaModel results,the two bump events that occurred at the Brody No.1 Mine were both categorized as having a high potential to initiate a bump event.Additionally,LaModel results mimicked site-specific condition mapping as conducted by MSHA investigators.

The methodology detailed within this paper provides a numerical technique for the evaluation of coal pillar bump potential through the comparison on stress distribution,pillar stability,and energy release results.While the numerical approach correctly identified the potential to bump,the exact locations of the coal pillar bump events were not precisely identified with LaModel,predicting the May 12,2014 bump event a lift prior to the location of the actual event within the cut sequence.To better locate the release of energy to the exact cut within the mining sequence,the numerical model should be updated with additional sitespecific material properties and behaviors to more accurately replicate the mining conditions.

The analysis of the bump events at the Brody No.1 Mine finds that while the ERR calculation provides an indication of bump potential,analysis and comparison to stress distribution and pillar safety factor results allow for better classification of the actual bump events.Through the application of LaModel,panel and pillar layouts can be better designed to alleviate pillar bump potentials as well as evaluate the pillar bump mitigation techniques such as leaving two rows of pillars between the mining hazard and the pillar line or the more aggressive practice of utilizing ‘‘bump cuts”within the row outby the pillar line [4].

Acknowledgements

The authors would like to acknowledge Mr.Eric Legg and Mr.Gary Hensley who were fatally injured during the coal pillar bump that occurred on May 12,2014 on the No.1 Section in the 4 East Main of Brody No.1 Mine.May the continued analysis of the Brody No.1 Mine bring the industry closer to understanding the coal pillar bump failure mechanism such that further loss of life can be prevented.