Niole Evnek,Brent Slker,Anthony Innhione,Tim Miller

a National Institute for Occupational Safety and Health,Pittsburgh Mining Research Division,Pittsburgh,PA 15236,USA

b University of Pittsburgh,Pittsburgh,PA 15260,USA

c East Fairfield Coal Company,North Lima,OH 44452,USA

Keywords:Ground control Limestone Horizontal stress Stress control layout Crosscut Windows

ABSTRACT The Subtropolis Mine is a room-and-pillar mine extracting the Vanport limestone near Petersburg,Ohio,U.S.In February of 2018,mine management began implementing a heading re-orientation to better control the negative effects of excessive levels of horizontal stress.The conditions in the headings improved,but as expected,stress-related damage concentrated within crosscuts.The mine operator has worked to lessen the impact of the instabilities in the outby crosscuts by implementing several engineering controls.With the implementation of each control,conditions were monitored and analyzed using observational and measurement techniques including 3D LiDAR surveys.Since the heading re-orientation,several 3D LiDAR surveys have been conducted and analyzed by researchers from the National Institute for Occupational Safety and Health (NIOSH).This study examines (1) the characteristics of each 3D LiDAR survey,(2) the change in the detailed strata conditions in response to stress concentrations,and (3)the change detection techniques between 3D LiDAR surveys to assess entry stability.Ultimately,the 3D LiDAR surveys proved to be a useful tool for characterizing ground instability and assessing the effectiveness of the engineering controls used in the heading re-orientation at the Subtropolis Mine.

1.Introduction

Many underground limestone mines studied by the National Institute for Occupational Safety and Health (NIOSH) operate in limestone formations thick enough to hold rooms 4.8 to 38 m high[1].The spans of these openings are equally large averaging 13.5 m.Spans of this size require strong,massive roof members often exceeding 2 m in thickness.Mine operations leave this protective layer of the stronger limestone in the roof and floor to improve ground stability.The ground control situation becomes more complex when the mine reduces the roof and floor members,weakening the surrounding strata.Conditions are further impacted when these strata are subjected to excessive levels of horizontal stress,resulting in roof and floor instabilities that are difficult to control with primary and secondary support alone.All of these conditions are found at the Subtropolis Mine,which extracts the Vanport limestone near Petersburg,Ohio,U.S.

Stone mines operating in the Vanport limestone have been the subject of past NIOSH research.In 2012,stress damage in the Vanport limestone was documented that included numerous roof failures and offset test holes [2].A neighboring inactive underground limestone mine,the Petersburg Mine,also mined the Vanport limestone and experienced significant pillar and roof failures [3].The geologic characteristics of these mines are slightly different,with variations in moisture-sensitive floor and weak bands within the pillars,both of which have been responsible for ground instability.

Photogrammetry and LiDAR scanning have been used at both the Subtropolis Mine and Petersburg Mine to detect ground changes [4,5].The ground monitoring performed was confined to small areas of the mines because of the logistical limitations associated with taking pictures using a digital single-lens reflex(DSLR)camera or scanning with a stationary tripod-mounted LiDAR scanner.

The objective of this study is to understand the applicability of large-area,vehicle-mounted LiDAR surveying in underground limestone mining operations and to observe and understand ground instability.This application is demonstrated at the Subtropolis Mine where excessive horizontal stresses contributed to ground instabilities.Previous NIOSH research at this mine led to the decision to implement a new mining orientation to minimize stress-related damage in the headings and to concentrate damage in the crosscuts [6].Several engineering controls were implemented in the new orientation to reduce the damage in the crosscuts.

The Vanport limestone ranges from 4.9 to 6.7 m at the Subtropolis Mine.This unit thins in various locations to the south and east,which can lead to insufficient floor and roof beam thickness.The unit contains a prominent bedding plane between limestone members as well as a prominent shale band.Otherwise,the unit is massive,until high levels of horizontal stress are applied,at which point the unit begins to delaminate.Above the Vanport limestone is a weak shale,requiring the operators to leave a limestone caprock to strengthen the roof.Fig.1a shows the generalized dimensions of a typical heading.The depth of cover is approximately 57.9 m near the study area.The mining thickness is approximately 4.9 m with a heading span of 12.2 m and a crosscut width of 9.1 m.The limestone left in the floor is about 0.3 m thick,and immediately under the limestone in the floor is sandstone.The planned caprock thickness is approximately 1.2 m;however,the thickness of the caprock varies,mainly due to spalling caused by high horizontal stress conditions that are related to blasting [7].Fig.1b shows dimensions of a heading with roof instability occurring within the caprock,causing changes in the elevation of the roof line.

Since mining began at the Subtropolis Mine,excessive horizontal stress conditions have been an issue and have forced the operators to overcome ground control challenges by adopting stress control layouts.Evidence of excessive horizontal stress-related issues are apparent when examining the current mine map showing numerous layout changes implemented to minimize hazardous conditions(Fig.2).The stress control mine layout in the new study area orients headings in the direction of the principal horizontal stress and associated horizontal movement.Iannacchione et al.[6]reported the orientation of the principal horizontal stress to be N35°W.Since the new orientation has been implemented,3D LiDAR surveying of the mine has been conducted in order to assess stress-related damage between each survey date.3D LiDAR surveys can be used to measure changes in these types of challenging conditions and can help to determine high-risk strata instability areas,thereby improving worker health and safety in underground limestone mines.

2.Methods

This study presents the results of five surveys conducted by Mine Visions Systems during the following months:May 2018,August 2018,December 2018,June 2019,and November 2019(Fig.3).Each survey encompasses similar portions of the mine layout,including areas mined prior to,and after,the re-orientation of the headings within the mining front.With each subsequent survey,the entries in the new orientation and the area of the mine that have been scanned increased.The November 2019 survey data is currently being processed and only a small area of interest,discussed later,was available to analyze at the time of reporting.Each survey took approximately two to three days of scanning to complete,depending on the type of scanner used and the size of the area to be scanned.However,processing the data collected in each survey took several weeks.

Fig.1.Generalized dimensions of a typical heading at the Subtropolis Mine and dimensions of a heading with roof instability occurring within the caprock causing changes in the elevation of the roof line.

Fig.2.Subtropolis Mine map displaying mining progression and new orientation in the study area.

LiDAR data can be used to characterize variations in the shape of the mine opening and relate that change to unstable ground conditions.Stationary or static LiDAR scanning techniques have proven effective in characterizing changes in strata conditions[8].In this study,mobile or dynamic scanning techniques (continuous scanning through time) was used.While dynamic scanning techniques require special scanners,it is the processing software that is crucial for establishing point coordinates and tying together individual scans.In this study,the use of change detection methods begins to be possible at around 1-2 cm of movement.For display purposes,values larger than 2 cm are often used for visualization cutoffs.At <2 cm accuracy,fracture mapping or small-scale convergence detection,which has been recently performed in other underground limestone mines[8,9],may not be possible.However,at this site convergence >2 cm proved to be adequate in identifying unstable ground conditions and areas of fallen rock.

During the time from the first survey in May 2018,to the most recent survey in November 2019,the scanning technology that was used evolved.For the first four surveys,Mine Vision Systems used a large scanner(CRL MultiSense SL)that had to be moved around the mine via a front-end loader (Fig.4).For the May 2018 and August 2018 surveys vehicle mounted scanning was restricted to entries without obstructions,i.e.berms,equipment,etc.Therefore,beginning in the December 2018 survey,a separate handheld scanner(VLP16) was used for areas that could not be reached via vehicle.Collecting the data in this manner allowed for better access and helped to build a more comprehensive survey.However,combining these two methods increased data processing time.Both the large scanner and the handheld scanner collected data continuously while in motion,and their accuracies appear comparable.

For the November 2019 survey,the VLP16 handheld scanner was truck-mounted and used to collect the data (Fig.5).Both the large scanner(45000 points per second)and the handheld scanner(300000 points per second)collect millions of X,Y,Z data points as they move through the mine.These points are used to create a 3D LiDAR image.The collection process for the handheld scanner was straightforward.The unit is small and can be mounted to the top of a pickup truck relatively quickly.The scanner can be driven around the mine at 8 to 16 km/h.The unit was transported down all accessible entries in the study area and through several crosscuts.The range of this unit is approximately 90 m.The unit can be transferred back to a handheld status and carried into areas that a vehicle cannot access.This unit comes with a touchscreen display that automatically tracks the data in real-time as it moves through the mine.

Fig.3.Data from five 3D LiDAR surveys during May 2018,August 2018,December 2018,June 2019,and November 2019.

Fig.4.Photo of large scanner and forklift.

Fig.5.Photo of VLP16 truck-mount scanner.

Once the scan data is collected,the processing stage begins.The surveys are collected in smaller segments.Due to the large horizontal distances that are scanned,the preliminary processed data is prone to survey drift.Survey drift occurs when a scan takes place over a large distance without the ability to frequently close a loop—the larger the distance from the scan starting point to the scan finishing point,the more survey drift.Shorter scan segments may help to minimize this error during the registration process.Once the scans have been properly aligned to each other during the registration process,the data can be evaluated.

3.Results

3.1.Roof elevation contours

Once the data has been processed,the survey can be evaluated.Relative elevation contours can be made to highlight areas of damage or unevenness in the roof or floor,as was recently demonstrated at the Pleasant Gap Mine [8].For example,the June 2019 survey is shown in Fig.6 with relative roof elevations contoured 1 m above and below an approximate zero roof line.Relative elevation mapping can be complicated by mines with even gentle dips,in which case multiple regions may need to be contoured on different scales or contoured relative to a plane that follows the dip of the roof line.Since this mine is not perfectly flat,the scale was adjusted in places from a 1.0 to 3.0 m range to a 0.8 to 2.8 m range around the approximate roof line to account for the gentle dip.These ranges are meant to allow roof elevation changes to be quickly assessed,and in the case of this study,the process allowed for better viewing of roof instabilities such as fractures,cutter roof,and roof-fall areas.

The results of the type of analysis shown in Fig.6 provide a comprehensive picture of the study area and the roof instabilities the mine was experiencing.Previously mapped,stress-related damage,as observed and documented in the mine,is clearly shown in the data.Several areas of difficulty aside,the majority of scans could be aligned properly and produced a recreation of real mine roof conditions.Red areas in Fig.6 represent points locally at a higher elevation than the average roof line,areas in blue represent lower elevation,and areas in grey represent the approximate average roof line.

3.2.Change in stress conditions

One of the primary objectives of this study is to examine the change in stress-related strata damage with each survey.The November 2019 survey is the most recent,and therefore includes the largest scan area to analyze.However,this survey is still being processed and only a small portion of the data is available for analysis.The largest available survey that covers the most drivage in the mine is the June 2019 survey.This survey includes some of the primary stress-related damage areas,including several rooffall areas.

Scans within the June 2019 survey were reviewed on an individual basis to evaluate areas showing evidence of excessive levels of stress and poor roof conditions.For this survey,the study focuses on two areas within the newly re-oriented section of the mine.This first area is between crosscuts XC-2 and XC-4 and headings E15 to E21,as shown in Fig.7.This area encompasses roof-fall areas in February 2019 and March 2019.

Fig.6.June 2019 survey with roof elevations displaying roof damage such as cutter roof and roof falls (in red).

Fig.7.February 2019 and March 2019 roof-fall areas represented in the June 2019 survey with relative roof elevations.

In Fig.7,the red coloration shows the shape and extent of both the February 2019 roof fall and the March 2019 roof fall.The February 2019 roof fall initiated as a cutter-type roof instability that progressed across several entries of the mine,only to be stopped by windows which were implemented as an additional engineering control to combat the effects of excessive horizontal stress.

A window is developed by leaving an increased thickness of roof rock in the crosscuts,thus reducing the crosscut height.Windows can be full(entire crosscut length)or partial (half crosscut length)and are visible in the scan data.Further discussion of the window development and experience is provided by Evanek et al.[7].In Figs.6 and 7,the bright blue color designates locations where the roof elevation is around 1 m below the average roof line,indicating the presence of windows.

Fig.8 shows the stress damage sustained to the roof during the February 2019 roof fall relative to an approximate roof line.The approximate roof line in Fig.8 represents a zero line that is used in several other figures,where the roof above this is displayed as positive numbers and the roof below this line is displayed as negative numbers.

The March 2019 roof fall was most likely caused by stresstransfer through a nearby barrier to the east of the study area.The load transfer conditions in the barrier pillar between the old workings and the study area grew larger as the new workings reduced the size of the barrier.The conditions observed in the new workings during mining indicate that the stress was undergoing redistribution from the barrier pillar to the surrounding workings.

Fig.8.February 2019 roof-fall area with stress damage in relation to an approximate roof line.

Data from this study also indicates the difference in stressrelated damage between the previous heading orientation and the current orientation.Fig.6 displays an area of decreased stress effects (reduction in rapid roof elevation changes,shown in red)that occurred once the re-orientation was complete and two mining fronts joined together.The area of decreased stress related failures was also confirmed by the operator.The process of the two mining fronts joining together can be seen in Fig.9.In September 2018,immediately following the August 2018 survey,the mine was in the beginning stages of re-orientation in two separate developments around XC-37.Prior to re-orientation,crosscuts were driven east-west and headings were driven north-south.Following the re-orientation,headings were driven in the northwest direction.Several of the headings driven prior to the reorientation experienced significant stress damage,especially XC-37.Therefore,the headings driven in the new orientation had to be accessed around XC-37,creating two separate mining fronts.The development of two separate fronts caused stresses to be redirected towards XC-37,which is the most northern crosscut in the previous orientation (east-west).During the December 2018 survey,the two mining fronts became more distinctive and the concentration of stress narrowed in XC-37 and reached its peak.By June 2019,the stresses were redirected around the singular mining front and the stress effects were reduced accordingly.Fig.6 shows an overall decrease in stress effects once the new orientation was well established to the north-east of crosscut XC-7.This decrease in stress effects was also confirmed by the mine operators.

3.3.Change analysis of crosscut XC-37

During the development of the two separate mining fronts,XC-37 sustained increased stress damage as a result of the horizontal stress being redirected around the mining front and focused towards the crosscut.After the two mining fronts merged into one singular mining front,which can be seen in the June 2019 survey,the horizontal stress appeared to have subsided in this crosscut.This study also focuses on evaluating the damage experienced in XC-37 during the redistribution of stress.Fig.10 displays the damage to the roof in crosscut XC-37 by using the same relative elevation contour analysis described earlier to create Figs.6 and 7.However,Fig.10 also includes a surface that is created through triangulation and is intended to improve visualization of the instabilities in the roof.Fig.10 examines XC-37 and contains examples of roof instabilities in the form of cutter or guttering roof.Conversely XC-36,also shown in Fig.10,has relatively uniform roof elevations indicating stable roof conditions.

Fig.10 is useful for identifying the damage caused by excessive horizontal stress;however,it does not show how the damage progressed over time.Therefore,change detection was used to track this damage to capture the magnitude and extent of rock displacements.

Displacements such as roof falls or pillar spalling may be very easy to detect even with poor accuracy,precision,and alignment because of the magnitude of movement.However,the same is not true for roof convergence and floor heave.Multiple surveys were used to detect change in XC-37.Unfortunately,due to limited access,the June 2019 survey did not include XC-37,and the August 2018 survey only included half of the crosscut.Therefore,the analysis includes the full extent of XC-37 for the May 2018,December 2018,and November 2019 surveys,as well as half of XC-37 for the August 2018 survey.

The overall change is measured using the point cloud from the first set of scans(May 2018)and last set of scans(November 2019).Fig.11 shows the overall movement in the roof that XC-37 experienced during this time period.Negative values indicate movement into the opening or sag and postive values indicate rock removal,both representative of roof instabilities.Conversely,Fig.12 shows the overall movement in the floor in XC-37 during this time.Both Figs.11 and 12 demonstrate the changes occurring in XC-37 during this period where the horizontal stress is believed to have been elevated.In Figs.11 and 12,the positive numbers represent movement into the entry and the negative numbers represent material loss.

Fig.9.Increases in stress concentrations at XC-37 as two mining fronts converge.

Fig.10.The crosscut XC-37 area represented in the June 2019 survey triangulation with roof elevations and contour lines.

Fig.11.Overall change detection in the roof of XC-37 from May 2018 to November 2019.

Fig.12.Overall change detection in the floor of XC-37 from May 2018 to November 2019.

At the Subtropolis Mine,the orientation of the principal horizontal stress is N35°W.The most significant stress-related damage is expected to occur perpendicular to the principal orientation of the horizontal stress field.However,XC-37 was mined in an east-west orientation prior to the adoption of the stress control layout.Therefore,damage was expected to concentrate at the far west corner of the entry and propagate at an angle across the crosscut and perpendicular to the principal horizontal stress.Fig.13 displays change detection in the roof from May 2018 to December 2018.In Fig.13,the amount of roof sag(green and blue)and the amount of roof loss or damage (red and yellow) have increased.The roof sag is represented in the legends as positive numbers,or material moving in from the original surface into the entry.The roof loss or damage is represented in negative numbers,or material taken away from the original surface.Fig.13 shows significant damage at the left side of the face.This damage appears to propagate across the crosscut to the other rib at an angle perpendicular to the principal horizontal stress,as anticipated.Once at the far rib,the damage may propagate along that rib until the damage begins to concentrate at the far corner again.In Fig.13,this process repeats until the damage reaches a set of three large concrete cribs.The damage concentrates in front of the cribs and then is forced to move around the cribs.Once past the cribs,the damage propagates again from the cribs to the rib at an angle perpendicular to the principal horizontal stress.

The above pattern of damage is something that has been observed in several other instances at the Subtropolis Mine.However,the 3D LiDAR surveys allow for a more detailed analysis of this type of damage propagation.This technology also provides a unique tool when mapping change detection over different time periods.For instance,Fig.14a displays the change in the roof of XC-37 from May 2018 to August 2018,whereas Fig.14b displays the change in the roof at XC-37 from May 2018 to December 2018.The change in roof conditions from Fig.14a to Fig.14b can be seen clearly.In Fig.14b,the amount of roof sag(green and blue)and the amount of roof loss or damage (red and yellow) have increased.As stated previously,the roof sag is represented in the legends as positive numbers,or material moving in from the original surface into the entry.The roof loss or damage is represented in negative numbers,or material taken away from the original surface.In August 2018,mining was divided into two fronts and was just beginning to concentrate stresses towards XC-37 (Fig.9).By December 2018,the two mining fronts were closing in on each other and stress concentrations towards XC-37 were assumed to be at their highest.The 3D LiDAR survey images in Fig.14a and b show evidence of additional roof instabilities as the two mining fronts converged around XC-37.Similarly,Fig.15a and b show the change elevation within the floor of XC-37.Fig.15a represents the change detection from May 2018 to August 2018,whereas Fig.15b displays the changes from May 2018 to December 2018.Between August 2018 and December 2018,XC-37 was cleared of debris on the floor several times.Excess material on the floor in this time period was piled next to the cribs and can be seen in Fig.15b.This material is shown in purple and blue and represent positive numbers.It is also clear from Fig.15b that there is an increase in floor heave over this time period,which is represented in green.

4.Discussion and conclusions

Although roof conditions have been the primary concern at the Subtropolis Mine,floor heave has occurred in other portions of the mine.The orientation of the heaved floor rock indicates it is also related to excessive horizontal stress.In crosscut XC-37,the data suggests floor heave occurred at similar times as the sag in the roof.It may be that in this case,floor heave was a precursor to the roof sag and may help to better predict roof damage,as was believed to be the case by Slaker et al.[5].The ability to use 3D LiDAR surveys to map floor heave and track its progression may assist in the prediction of poor ground conditions.

Fig.13.Change detection in the roof of XC-37 from May 2018 to December 2018 (refer to Fig.3 for the mine layout during these two scan dates).

Fig.14.Change detection in the roof of XC-37.

Fig.15.Change detection in the floor of XC-37.

Despite the widespread roof damage,the ribs in this mine have primarily remained in good condition.However,in XC-37,slight movement along the shale band was observed on the northern rib.This movement appears to be related to differential horizontal movement along a contact plane within the rib.This movement was mapped in the mine but was not discernible in the survey data.The XC-37 was observed by the mine operators as having some of the most challenging conditions in the mine,and as a result was sometimes inaccessible.

The 3D LiDAR surveys displayed the ability to map relative roof elevation contours.This data provides a more detailed view of roof damage in February 2019 and March 2019 roof fall areas.The scans provided insight to barriers,such as windows and large pillars,that halted the progression of the roof falls.The survey data also shows an increase in roof damage and roof sag in XC-37 during periods of elevated concentrations of horizontal stress.

The 3D LiDAR surveys were used to analyze the change in strata conditions in response to variations in mine layouts.This allowed for the assessment of different engineering controls used to mitigate excessive horizontal stress conditions.The 3D LiDAR surveys helped to evaluate the application of the stress control mine layout in reducing strata instabilities.The 3D LiDAR surveys prove to be useful for performing wide-area damage assessments,which may improve the safety conditions in underground limestone operations.

Disclaimer

The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the National Institute for Occupational Safety and Health,Centers for Disease Control and Prevention.Mention of any company or product does not constitute endorsement by NIOSH.