geosciences Article
Implementation of a Fracture Modeling Strategy Based on Georadar Survey in a Large Area of Limestone Quarry Bench Mohamed Elkarmoty 1,2, * , Francesco Tinti 2 , Sara Kasmaeeyazdi 2 , Fabio Giannino 3 , Stefano Bonduà 2 and Roberto Bruno 2,† 1 2
3
* †
Department of Mining, Petroleum, and Metallurgical Engineering, Cairo University, Cairo University Street, Giza 12613, Egypt Department of Civil, Chemical, Environmental and Materials Engineering, University of Bologna, Via Terracini 28, 40131 Bologna, Italy;
[email protected] (F.T.);
[email protected] (S.K.);
[email protected] (S.B.);
[email protected] (R.B.) IDS GeoRadar s.r.l., Via Augusto Righi, 1-2, 56121 Pisa, Italy;
[email protected] Correspondence:
[email protected]; Tel.: +20-114-3888-380 Present address: Scientific Attaché, Embassy of Italy in Brazil, S.E.S. Av. das Nações Quadra 807, Lote 30, Brasilia 70420-900,
[email protected].
Received: 17 November 2018; Accepted: 10 December 2018; Published: 13 December 2018
Abstract: Rock mass fractures adversely affect the cutting of commercial-size blocks and cause rock material loss in ornamental stone quarries. In order to obtain a reliable evaluation and an optimized production of ornamental stone deposits, it is fundamental to detect fractures in a non-destructive manner identifying them through 3D deterministic modeling. In this study, a recently published fracture modeling strategy, based on Ground Penetrating Radar (GPR) survey was implemented on a large area of bench (27.0 m × 65.0 m) in a limestone quarry in Italy. The survey was done using a dual-frequency GPR system (250 MHz and 700 MHz). The objective of this work was to investigate the large-scale applicability of the mentioned fracture model for future consideration in quarrying optimization studies. Only the 700 MHz radargrams were considered for the fracture modeling, as they provided a higher resolution than the 250 MHz radargrams and a penetration depth of about 4.0 m. The bulk dielectric constant of the rock mass of the bench was estimated by averaging the velocities obtained from fitting the hyperbolic diffractions of fractures at different depths. The model showed that fractures from the same family set can have noticeable spatial variations. The results allowed us to roughly estimate the sizes of the blocks exploitable from the different rock layers of the quarry bench. Keywords: ornamental stone; fracture modeling; GPR; dielectric constant; quarrying
1. Introduction The quarrying industry produces a large amount of waste material mainly due to natural rock discontinuities characterized by different geologic forms, classes, and sizes [1,2]. Generally speaking, considering the worldwide average of ornamental stone excavation, only one third of the exploited raw material reaches the global market as a finished product, while the remaining two thirds are waste [3]. From a commercial viewpoint, ornamental stones are considered a main economic resource for many countries worldwide [4]. Therefore, since ornamental stones are natural non-renewable resources, they need to be quarried sustainably in order to reduce their environmental and economic impact. This can be carried out through a 3D detection and deterministic modeling of the fractures present in the rock mass, followed by optimized production planning aiming to optimize the coefficient of Geosciences 2018, 8, 481; doi:10.3390/geosciences8120481
www.mdpi.com/journal/geosciences
Geosciences 2018, 8, 481
Geosciences 2018, 8, x FOR PEER REVIEW
2 of 15
2 of 15
utilization [5,6]. The robustness the production strongly depends on depends the accuracy of coefficient of utilization [5,6]. Theofrobustness of theoptimization production optimization strongly on the 3D fracture modeling, which should be deterministic for this kind of application. accuracy of 3D fracture modeling, which should be deterministic for this kind of application. The the detection of of fractures in The Ground Ground Penetrating PenetratingRadar Radar(GPR) (GPR)isiswidely widelyused usedand andtrusted trustedforfor the detection fractures rock masses. Several studies dealing with GPR detection and characterization of fractures are present in rock masses. Several studies dealing with GPR detection and characterization of fractures are in the literature, addressing limestone [7,8], marble sandstone [13–15],[13–15], and granite present in the literature, addressing limestone [7,8], [9–11], marblegneiss [9–11],[12], gneiss [12], sandstone and quarries [16,17]. In addition, a numbera of publications have taken intotaken consideration the use of the GPRuse to granite quarries [16,17]. In addition, number of publications have into consideration detect and model fractures in quarried stone blocks [7,9,18–21]. The theoretical background for GPR of GPR to detect and model fractures in quarried stone blocks [7,9,18–21]. The theoretical background can be found in found [22–24]. for GPR can be in [22–24]. The most noticeable fractures from a GPR survey in The most noticeablecontributions contributionson on3D 3Ddeterministic deterministicmodeling modelingofof fractures from a GPR survey quarries have been given by [25–27]. Recently, research presented by [1] tried to improve the methods in quarries have been given by [25–27]. Recently, research presented by [1] tried to improve the proposed [25–27] by minimizing theminimizing uncertainty in fractures deterministically as methods by proposed bymainly [25–27] by mainly themodeling uncertainty in modeling fractures 3D surfaces based on a GPR survey. The modeling approach of [1] was tested and validated on a small deterministically as 3D surfaces based on a GPR survey. The modeling approach of [1] was tested area of a benchon in aa small sandstone quarry. and validated area of a bench in a sandstone quarry. The The objective objective of of this this paper paper is is to to investigate investigate the the applicability applicability and and workability workability of of the the fracture fracture modeling approach presented in [1] in a large area of a quarry bench with different of modeling approach presented in [1] in a large area of a quarry bench with different typestypes of rock rock (limestone) and geological characteristics (horizontal or semi-horizontal layers with (limestone) and geological characteristics (horizontal or semi-horizontal layers with twotwo setssets of of fractures). Moreover, the large-scale implementation presented in this paper can help the fractures). Moreover, the large-scale implementation presented in this paper can help the geogeo-engineering community to assess the applicability the fracture modeling approach in other engineering community to assess the applicability of theoffracture modeling approach in other geogeo-engineering fields, not only in quarrying. engineering fields, not only in quarrying. 2. Site Description 2. Site Description This study was carried out in a bench of a limestone quarry located in Poggio Imperiale, Foggia, This study was carried out in a bench of a limestone quarry located in Poggio Imperiale, Foggia, Italy (Figure 1). The quarried area belongs to the Apricena quarrying basin which includes the towns Italy (Figure 1). The quarried area belongs to the Apricena quarrying basin which includes the towns of Apricena, Poggio Imperiale, and Lesina. The Apricena quarrying basin is the largest mining area of Apricena, Poggio Imperiale, and Lesina. The Apricena quarrying basin is the largest mining area in the Italian Region of Apulia, covering over 24.1 km2 2 . The limestone deposit consists of a series in the Italian Region of Apulia, covering over 24.1 km . The limestone deposit consists of a series of of successions called “Fiorito Succession,” “Biancone Succession”, and “Serpeggiante Succession.” successions called “Fiorito Succession,” “Biancone Succession”, and “Serpeggiante Succession.” The The limestone deposit in the bench object of this study is made up of fine and ultra-fine-grained limestone deposit in the bench object of this study is made up of fine and ultra-fine-grained beigebeige-white micritic limestone belonging to the “Fiorito Succession” [28]. In this quarry, various white micritic limestone belonging to the “Fiorito Succession” [28]. In this quarry, various quarrying quarrying methods are used depending on the zones, ranging from the drilling and blasting method methods are used depending on the zones, ranging from the drilling and blasting method utilized in utilized in the bench under study, to the diamond wire cutting method. the bench under study, to the diamond wire cutting method.
Figure 1. Figure 1. (a) (a) A A map map of of Italy Italy showing showing the the location location of of Poggio Poggio Imperiale Imperiale (Google (Google Maps), Maps), (b) (b) A A panoramic panoramic view of the quarry; the blue star refers to the surface of the bench under study. view of the quarry; the blue star refers to the surface of the bench under study.
The rock mass of the bench was characterized by two sub-vertical sets of fractures, identified as Set 1 and Set 2 (Figure 2). Set 1 refers to the fractures propagating parallel to the bench face under study (strike direction to North-West), whereas Set 2 refers to the fractures propagating perpendicularly to the bench face under study (strike direction to South-West). As evident in Figure
Geosciences 2018, 8, x FOR PEER REVIEW
3 of 15
2, the spacing of the fractures in Set 1 was larger than that in Set 2, allowing the quarried blocks in Geosciences 2018, 481 a marketable size. The aperture size of the out-cropping fractures ranged 3from of 15 this bench to 8,have Geosciences 2018, 8, x FOR PEER REVIEW 3 of 15 about 0.2 cm to 3.0 cm. The rock mass layering could not be observed visually on the bench face, 2, the spacing the fractures 1 was larger than that ineffect Set 2, (black allowing the white quarried blocks in whose height wasof 2.30 m, dueintoSetthe chemical blasting and stripes) on the The rock mass of the bench was characterized by two sub-vertical sets of fractures, identified as this bench to have a marketable size. The aperture size of the out-cropping fractures ranged from appearance of the bench face (Figure 2) and to the thin-aperture bedding planes between the layers, Set 1 about and Set (Figure 2).cm. SetThe 1 refers to thelayering fractures propagating parallelvisually to the bench under study 0.22the cm to 3.0view rock mass could not be observed theface bench face, as shown in side of the bench (Figure 3). However, the layers in thisonquarry were clearly (strike direction North-West), Set 2 refers to the fractures perpendicularly whose heighttowas m, duewhereas to the chemical blasting effect (black propagating and white stripes) on the stratified parallel to the2.30 bench surface. to theappearance bench face study to South-West). As evident Figure the 2, the spacing of under the bench face(strike (Figuredirection 2) and to the thin-aperture bedding planesinbetween layers, The bench surface was reasonably flat, and thus suitable for the GPR survey activities. The rock shown in in theSet side view larger of the bench (Figure 3). However, thethe layers in this quarry clearly of theasfractures 1 was than that in Set 2, allowing quarried blocks were in this bench to massstratified of the bench was wet since the GPR survey was carried out 8 h after the end of 10-h of parallel to slightly the bench surface. have a marketable size. The aperture size of the out-cropping fractures ranged from about 0.2 cm to light rain.The However, the surface of the benchand was dry due toforthe the area after the rain. bench surface was reasonably thus suitable thesun GPRdrying surveyface, activities. The rock 3.0 cm. The rock mass layering could notflat, be observed visually on the bench whose height was The surveying was carried out on a day characterized by a temperature of 14 °C and a relative mass of the bench was slightly wet since the GPR survey was carried out 8 h after the end of 10-h of 2.30 m, due to the chemical blasting effect (black and white stripes) on the appearance of the bench face humidity of 71%. light rain. However, the surface of the bench was dry due to the sun drying the area after the rain. (Figure 2) and to the thin-aperture bedding planes between the layers, as shown in the side view of the The surveying was carried out on a day characterized by a temperature of 14 °C and a relative bench (Figure 3). However, the layers in this quarry were clearly stratified parallel to the bench surface. humidity of 71%.
Figure A view of the studied bench(under (under the the arrow) thethe twotwo setssets of out-cropping Figure 2. A of studied bench arrow)highlighting highlighting sets of out-cropping out-cropping Figure 2. A2.view view of the the studied bench (under the arrow) highlighting the two of fractures in two large normal rock faces. The orientation of the rock face, where Set 2 is highlighted fractures in two two large large normal normalrock rockfaces. faces.The Theorientation orientationofofthe therock rockface, face, where 2 highlighted is highlighted where SetSet 2 is in infigure, the figure, is parallel to the studiedbench benchface. face. in is parallel to the studied thethe figure, is parallel to the studied bench face.
Figure 3. A left-side view of the bench face where horizontal layers were observed visually (named a, b, c, and d).
Figure 3. A A left-side left-side view view of of the the bench bench face face where where horizontal horizontal layers layers were were observed observed visually visually (named (named a, a, b, c, and d).
The bench surface was reasonably flat, and thus suitable for the GPR survey activities. The rock mass of the bench was slightly wet since the GPR survey was carried out 8 h after the end of 10-h of light rain. However, the surface of the bench was dry due to the sun drying the area after the rain. The surveying was carried out on a day characterized by a temperature of 14 ◦ C and a relative humidity of 71%.
Geosciences 2018, 8, 481
3. TheGeosciences GPR Survey 2018, 8, x Design FOR PEER REVIEW Geosciences 2018, 8, x FOR PEER REVIEW
4 of 15
4 of 15 4 of 15
A three-dimensional rectangular GPR survey grid was designed with four vertices named A, B, C, 3. The GPR Survey Design 3. The GPR Survey Design and D, as shown in Figure 4. The A–B vertex was parallel and 3.0 m away from the bench face, since A three-dimensional rectangular GPR survey grid was designed with fourbearing vertices named A, B, side most of the neglected area within theseGPR 3.0 survey m wasgrid fullwas of stone debris. of the A, A–B A three-dimensional rectangular designed withThe four vertices named B, C, and D, as shown in Figure 4. The A–B vertex was parallel and 3.0 m away from the bench face, ◦ was 345.0 North. The dimensions of thewas rectangular grid measured through a 1.0 cm C, and from D, as the shown in Figure 4. The A–B vertex parallel and 3.0were m away from the bench face, since most of the neglected area within these 3.0 m was full of stone debris. The bearing of the A–B sincetape. most An of the neglected area within 3.0 m was stone debris. bearing of the and A–B B–D, accuracy arithmetical check wasthese performed onfull theoflengths of theThe diagonals, A–C side was 345.0° from the North. The dimensions of the rectangular grid were measured through a 1.0 side was 345.0° from the North. The dimensions of the rectangular grid were measured through a beforecm the survey, to make sure that the rectangular was geometrically correct. We collected 14 1.0 profiles accuracy tape. An arithmetical check was performed on the lengths of the diagonals, A–C and B– cm accuracy tape. An2.0 arithmetical check wasthe performed on theand lengths of the diagonals, A–Ctoand B– (fromD,F01 to F14) with m spacing along X direction, 21 profiles (from F15 F34) before the survey, to make sure that the rectangular was geometrically correct. We collected 14with D, before the survey, to make sure that the rectangular was geometrically correct. We collected 14 3.0 mprofiles spacing along Y direction. length of each line X direction was equal to (from F01the to F14) with 2.0 mThe spacing along the X survey direction, andin21the profiles (from F15 to F34) profiles (from F01 to F14) with 2.0 m spacing along the X direction, and 21 profiles (from F15 to F34) 27.0 m; whereas, in thealong Y direction it was equal to 65.0 m. During theinsurvey, the locations of each with 3.0 m spacing the Y direction. The length of each survey line the X direction was equal with 3.0 m spacing along the Y direction. The length of each survey line in the X direction was equal to 27.0 m; whereas, in the Y direction it was equalintothe 65.0 m. During thethen survey, the locations of each antenna bumping were recorded in hand-writing field notes, to be later considered in data to 27.0 m; whereas, in the Y direction it was equal to 65.0 m. During the survey, the locations of each antenna bumping were recorded in hand-writing in the field notes, to then be later considered in data processing and interpretation. antenna bumping were recorded in hand-writing in the field notes, to then be later considered in data processing and interpretation. The benchand surface was surveyed through a GPR instrument of a four-wheel mounted dual processing interpretation. The bench surface was surveyed through a GPR instrument of a four-wheel mounted dual frequencyThe antenna (250 and MHz)through (Figurea 5). modelofofa the instrument, Opera bench surface was700 surveyed GPR The instrument four-wheel mounted dual Duo, frequency antenna (250 and 700 MHz) (Figure 5). The model of the instrument, Opera Duo, was frequency antenna 700 MHz) (Figure The The model of the Opera Duo, was was manufactured by (250 IDS and GeoRadar s.r.l., Pisa,5). Italy. grid wasinstrument, marked on the bench surface manufactured by IDS GeoRadar s.r.l., Pisa, Italy. The grid was marked on the bench surface every by IDS GeoRadar s.r.l., Pisa, Italy. Themovement grid was marked on the bench surface every everymanufactured meter, as shown in Figure 5, to ease the correct from one survey line to another. meter, as shown in Figure 5, to ease the correct movement from one survey line to another. meter, as shown in Figure 5, to ease the correct movement from one survey line to another.
Figure 4. Ground The Ground Penetrating Radar (GPR)survey survey design. The blue arrows refer to the direction Figure 4. The Penetrating Radar design.The The blue arrows refer todirection the direction Figure 4. The Ground Penetrating Radar(GPR) (GPR) survey design. blue arrows refer to the of the survey lines. of the survey lines. of the survey lines.
Figure 5. The GPR unit used, and the indications of points A and B on the bench surface.
Figure 5. The GPR unit used,and andthe theindications indications of B on thethe bench surface. Figure 5. The GPR unit used, ofpoints pointsAAand and B on bench surface.
Geosciences 2018, 8, 481 Geosciences 2018, 8, x FOR PEER REVIEW
5 of 15 5 of 15
4.4.Data DataProcessing Processingand andVelocity VelocityEstimation Estimation InInthis thiscase casestudy, study,only onlythe the700 700MHz MHzradargrams radargramswere wereconsidered consideredsince sincesuch suchfrequency frequencyprovided provided enough penetration depth with a higher resolution compared to the 250 MHz radargrams. Standard enough penetration depth with a higher resolution compared to the 250 MHz radargrams. Standard signal signalprocessing processingfunctions functionswere wereapplied appliedtotothe the700 700MHz MHzradargrams radargramsusing usingthe thesignal signalprocessing processing TM (manufactured software (manufacturedby byIDS IDSGeoRadar GeoRadars.r.l., s.r.l., Pisa, Italy). Firstly, the zero software package package GRED GREDTM Pisa, Italy). Firstly, the zero time time was corrected manually. The background noise was eliminated by applying a horizontal was corrected manually. The background noise was eliminated by applying a horizontal background background removal This command the Clear-Y filtering algorithmtotoremove remove continuous removal filter. Thisfilter. command appliesapplies the Clear-Y filtering algorithm continuous components componentsalong alongthe theYYaxis axis(horizontal (horizontaldirection) direction)using usingtwo twopreset presetparameters parameterstotoidentify identifythe thedepth depth ofofimplementation in the radargram. The minimum value of the preset parameters was equal to 0.0 implementation in the radargram. The minimum value of the preset parameters was equal to m 0.0 (depth), and the value value was equal 4.0 mto(depth). frequencies were filtered vertically m (depth), andmaximum the maximum was toequal 4.0 m The (depth). The frequencies were filtered through a vertical filters tool. Datatool. were bandpass filtered in the 200 to 1500 MHz vertically throughbandpass a vertical bandpass filters Data were bandpass filtered in the 200 to 1500band. MHz Aband. power of the amplitudes on the basis an basis automatically estimated linear attenuation Aequalization power equalization of the amplitudes onofthe of an automatically estimated linear was carried out gainthe function, followed by an automatic gain filter to attenuation wasthrough carried the outlinear through linear gain function, followed by smoothing an automatic smoothing equalize the power along a sweep for a moving window. Finally, by selecting a proper color scale and gain filter to equalize the power along a sweep for a moving window. Finally, by selecting a proper adjusting theand display gain,the it was possible lower amplitude regions. color scale adjusting display gain,toitspot was some possible to spot some lower amplitude regions. The Thebulk bulkpropagation propagationvelocity velocitywas wasestimated estimatedfrom fromthe thehyperbolic hyperbolicdiffractions diffractionsofofthe thesub-vertical sub-vertical fractures in different radargrams from the X and Y directions and within different subsurface fractures in different radargrams from the X and Y directions and within different subsurfacedepths. depths. ItItwas wasnot notpossible possibletotouse usethe the“depth “depthtotoaaknown knownreflector” reflector”calculation calculation[1,29] [1,29]totoestimate estimatethe theaverage average propagation thethe fitting hyperbola tool, 94 hyperbolic diffractions werewere fittedfitted usingusing the propagationvelocity. velocity.Using Using fitting hyperbola tool, 94 hyperbolic diffractions software package GRED™ (Figure (Figure 6). It was6).found thatfound the propagation velocity varied significantly the software package GRED™ It was that the propagation velocity varied insignificantly the surveyed mass body of the bench the finding in [30]. maximum in rock the surveyed rock mass bodyconfirming of the bench confirming the The finding in [30].and The minimum propagation velocities identified were equal to 136.0 Mm/s (Megameter/s) and 89.0 Mm/s, maximum and minimum propagation velocities identified were equal to 136.0 Mm/s (Megameter/s) respectively. Thisrespectively. study considered the average propagation velocity of 110.0 velocity Mm/s. Consequently, and 89.0 Mm/s, This study considered the average propagation of 110.0 Mm/s. the estimated bulk dielectric constant of theconstant limestone mass of the equal to 7.4. Consequently, the estimated bulk dielectric of rock the limestone rockbench masswas of the bench was This estimated dielectric constant valueconstant is consistent expectations, that, from bibliography equal to 7.4. This estimated dielectric valuewith is consistent with given expectations, given that, from data, the dielectric of limestone, in dry conditions,inshould range from 4.0 to 8.0, whilst wetto bibliography data,constant the dielectric constant of limestone, dry conditions, should range fromin4.0 conditions range fromit 6.0 to 15.0 [31]. from Using6.0 thetoestimated propagation 8.0, whilstitinshould wet conditions should range 15.0 [31].value Usingofthe estimated velocity, value of the two-way travel time converted true in theinto 34 radargrams giving maximum propagation velocity, thewas two-way travelinto time wasdepth converted true depth in the 34a radargrams penetration depth of penetration about 4.0 mdepth in the of 700 MHz4.0 radargrams. giving a maximum about m in the 700 MHz radargrams.
absolute frequency of the hyperbola fitted
22 20 18 16 14 12 10 8 6 4 2 0 89
93
97
101
105
109
113
117
121
125
129
133
137
propagation velocity (Mm/s) Figure 6. Histogram of the propagation velocities measured through the fitting hyperbola method. Figure 6. Histogram of the measured through fitting hyperbola method. The propagation velocities inpropagation the X axis of velocities the chart were plotted in a binthe of 4.0 Mm/s. The propagation velocities in the X axis of the chart were plotted in a bin of 4.0 Mm/s.
Geosciences 2018, 8, x FOR PEER REVIEW Geosciences 2018, 8, 481
6 of 15 6 of 15
Since the GPR data were affected by a large amount of pseudo-vertical fractures, the reconstruction of the true shape and by geometry of the detected fractures was carriedthe out by applying Since the GPR data were affected a large amount of pseudo-vertical fractures, reconstruction data migration. Data migration requires an accurate estimate of the average propagation of of the true shape and geometry of the detected fractures was carried out by applying data velocity migration. the waves in therequires rock medium. The estimated average propagation velocityvelocity (110.0 Mm/s) used Data migration an accurate estimate of the average propagation of the was waves in for the data migration in order to collapse the diffractions and to re-position the sub-vertical and the rock medium. The estimated average propagation velocity (110.0 Mm/s) was used for the data dipping fractures. The propagation velocity used sufficientthe tosub-vertical collapse theand diffractions defining migration in order to collapse the diffractions and towas re-position dipping fractures. the sub-vertical implementation migration was significant in this study not only The propagationfractures. velocityThe used was sufficient of todata collapse the diffractions defining the sub-vertical due to collapsing diffractions, but also because many hyperbolic diffractions of fractures were fractures. The implementation of data migration was significant in this study not only due to collapsing intersected making it because difficult many to interpret the fractures in theofnon-migrated radargrams. diffractions, but also hyperbolic diffractions fractures were intersected making it difficult to interpret the fractures in the non-migrated radargrams. 5. Interpretations and Modeling 5. Interpretations andthe Modeling This study used interpretation and modeling method presented in [1]. The surveyed area of the bench was divided into nine volumes referring to the radargrams in the X and directions, as This study used the interpretation and modeling method presented in [1]. TheYsurveyed area shown in Table and Figure 7. The fractures in each volume were interpreted modeled of the bench was 1divided into nine volumes referring to the radargrams in the X andand Y directions, separately. fractures then traced and modeled the connections betweenand the volumes. as shown inThe Table 1 and were Figure 7. The fractures in eachinvolume were interpreted modeled Appendix AThe shows two volumes interpreted from the radargrams in the X between and Y directions. It is separately. fractures were then traced and modeled in the connections the volumes. worth mentioning in this case study where vertical or sub-vertical fractures intersectIt with the Appendix A shows that, two volumes interpreted from the radargrams in the X and Y directions. is worth bedding planes, interpretation of a vertical or sub-vertical fracture may not belong to one fracture but mentioning that, in this case study where vertical or sub-vertical fractures intersect with the bedding to several fractures (usually one or fracture per mechanical layer): One starts closebut to to theseveral point planes, interpretation of a vertical sub-vertical fracture may not belong to onevery fracture where the(usually other ends. fractures one fracture per mechanical layer): One starts very close to the point where the other ends. Table 1. The volumes used for the interpretation. Table 1. The volumes used for the interpretation.
Volume
Volume1 1 2 3 4 5 6 7 8 9
2 3 4 5 6 7 8 9
Radargrams Radargrams F01-F02-F03-F04 F05-F06-F07-F08 F01-F02-F03-F04 F05-F06-F07-F08 F09-F10-F11-F12 F09-F10-F11-F12 F13-F14 F13-F14 F15-F16-F17-F18 F15-F16-F17-F18 F19-F20-F21-F22 F19-F20-F21-F22 F23-F24-F25-F26 F23-F24-F25-F26 F27-F28-F29-F30 F27-F28-F29-F30 F31-F32-F33-F34 F31-F32-F33-F34
Direction Direction Y Y Y Y Y Y Y Y X X X X X X X X X X
interpretation (top (top view). view). Figure 7. Graphical illustration of the volumes used for the interpretation
Geosciences 2018, 8, 481
Geosciences 2018, 8, x FOR PEER REVIEW
7 of 15
7 of 15
During During the the survey, survey, an an operator operator marked marked on on aa printed printed paper paper grid grid where where the the antenna antenna mechanically mechanically vibrated (antenna bumping) due to the stone debris or the uneven ground surface in aa few few locations. locations. vibrated (antenna bumping) due to the stone debris or the uneven ground surface in This This was was necessary necessary to to avoid avoid incorrect incorrect interpretation interpretation of of the the fractures fractures in in the the radargrams, radargrams, particularly particularly after data migration (Figure 8), since the mechanical vibration of the antenna causes a ringing noise in after data migration (Figure 8), since the mechanical vibration of the antenna causes a ringing noise the radargrams. in the radargrams.
Figure 8. 8. The The antenna antenna bumping bumpingeffect effectshown shownthrough throughananexample exampleextracted extracted from first meters Figure from thethe first sixsix meters of of radargram F13 (original length equal to 65 m). (a) Radargram F13 processed without data radargram F13 (original length equal to 65 m). (a) Radargram F13 processed without data migration, migration, thediffractions hyperbolicofdiffractions of fracture a sub-vertical surface highlighted in the showing theshowing hyperbolic a sub-vertical surfacefracture highlighted in the yellow rectangle. yellow rectangle. The blue rectangles refer to the zone of the antenna bumping. (b) The radargram The blue rectangles refer to the zone of the antenna bumping. (b) The radargram after data migration. after data migration. (c)fracture Interpretation thereflection fracture surface. The reflection the bumping zone is (c) Interpretation of the surface.ofThe in the bumping zone isin not considered. not considered.
In all the radargrams, a layering of the surveyed rock mass body was detected. The radargrams In all radargrams, a layering of the surveyed rock mass body was Thewere radargrams allowed usthe to observe that the upper and lower surfaces of the detected fourdetected. rock layers not flat. allowed us to observe that the upper and lower surfaces of the detected four rock layers were not flat. Furthermore, the method allowed us to detect the wavy feature of the layers’ surfaces (see an example Furthermore, the method to detect the feature of considered the layers’ surfaces in Figure 9). However, theallowed averageus thicknesses of wavy the layers were (Table 2).(see an example in Figure 9). However, the average thicknesses of the layers were considered (Table 2).
Geosciences 2018, 8, 481
8 of 15
Geosciences 2018, 8, x FOR PEER REVIEW
8 of 15
Figure Figure 9. 9. (a) (a) Radargram Radargram F03 F03 processed processed without without any any interpretation; interpretation; (b) (b) Radargram Radargram F03 F03 processed processed with with the interpretation of the detected layering (black dashed lines) and the detected sub-vertical the interpretation of the detected layering (black dashed lines) and the detected sub-vertical fractures fractures (different colors). (different colors). Table 2. The thickness thickness of Table 2. The of the the detected detected rock rock layers. layers. Rock RockLayer Layer
layera a layer layer layerb b layer layerc c layer d
layer d
from Z(m)totoZ(m) Z(m) Thickness Thickness (m) from Z(m) (m) 0.00.0 m–0.8 0.8 0.8 mm m–0.8mm 0.8 m–2.0 m 1.2 0.8 m–2.0 m 1.2 m m 2.0 m–2.8 m 2.0 m–2.8 m 0.8 0.8 mm 2.8 m–3.6 m 0.8 m 2.8 m–3.6 m 0.8 m
The fractures fractureswere weremodeled modeledinina arock rock mass body measuring 27.0 m65.0 X, 65.0 Y, and Z. mass body measuring 27.0 m X, m Y,mand 4.0 m4.0 Z.m The The polygon file format (PLY) fracture dataand andthe thesurveyed surveyedvolume volumewere were imported imported in polygon file format (PLY) [32][32] of of thethe fracture data ParaView [33] [33] for foraa3D 3Dvisualization visualization(Figure (Figure 10). As shown in Figure 10, the two sets of fractures 10). As shown in Figure 10, the two sets of fractures were were somehow orthogonally intersected. The fractures in the radargrams of the X and Y somehow orthogonally intersected. The fractures detecteddetected in the radargrams of the X and Y directions directions are represented in the model(Set in yellow 2) and (Set 1), respectively. Asthe shown in the are represented in the model in yellow 2) and (Set red (Set 1), red respectively. As shown in visualized visualized model, fractures belonging to a fracture set had noticeable variations in space. However, model, fractures belonging to a fracture set had noticeable variations in space. However, the general the general the dip direction, angle, and spacing still clearly thefracture two fracture sets. trend of thetrend dip of direction, dip angle,dip and spacing still clearly defineddefined the two sets. The The visualized model provided some indications that Set 2 was the dominant one. visualized model provided some indications that Set 2 was the dominant one.
Geosciences 2018, 8, 481 Geosciences 2018, 8, x FOR PEER REVIEW
9 of 15 9 of 15
Figure 10.10. The 3D3D fracture orientationsinside insidethe thesurveyed surveyed bench Figure The fracturemodel modelvisualized visualizedin intwo two different different orientations bench body. Vertex A of the survey grid is indicated in the model. The bedding planes were filtered for body. Vertex A of the survey grid is indicated in the model. The bedding planes were filteredbetter for visualization of the fracture model. model. better visualization of the fracture
6. Discussion 6. Discussion The tests analysisperformed performed demonstrated that use in ofquarries GPR incanquarries can be The testsand and analysis demonstrated that the usethe of GPR be technically technically feasible even in and stratified and orthogonally fractured Thefrequency use of high frequency feasible even in stratified orthogonally fractured benches. Thebenches. use of high antennas in antennas in compacted dry materials can lead to a high penetration depth. The 3D variation of compacted dry materials can lead to a high penetration depth. The 3D variation of the dielectricthe
Geosciences 2018, 8, 481
10 of 15
dielectric constant vertically and horizontally in rock masses is a critical issue in GPR measurements. Minimization of the resulting error is required, when possible, by the depth to a known reflector calculation or averaging many hyperbolic diffractions at different locations. Division of the whole area surveyed into volumes is necessary for large area interpretation and modeling so as to minimize human error. The Paraview program is still successful in visualizing modeled fracture data in a large area. The implemented fracture modeling strategy in a large area of bench is promising, considering the geologic conditions of the bench studied. However, the limitation of missing small fractures in the non-surveyed cells of the grid exists. When possible, intensive survey grid is much more favorable. The most common method of fracture detection in quarrying applications is the manual method that assumes all fracture planes are persistent, giving advantage to relatively less time consumption in the modeling phase. The modeling strategy used in this paper aimed at modeling 3D surfaces of fractures, rather than planes, as is usually the case. The assumption that fractures are persistent can be feasible in other applications. In the application for quarrying production optimization, the irregularity of curved fracture surfaces, being persistent or not based on the detected reflections, can play a significant role in quantification of non-fractured blocks that can vary from one layer to another when the cutting direction changes, as our current research indicates. The issue of time consumption in the implementation of the fracture model on a large area of a bench is a critical aspect, particularly with regard to the manual picking of the coordinates of the detected fractures from the radargrams. The aim is to integrate all of the components of the method (data processing, fracture interpretation, and model formulation) into a single software package, as this would reduce time consumption. From the resulting 3D fracture model, it is possible to roughly estimate the dimensions of the exploitable non-fractured blocks from the studied bench. The average spacing of the fractures in Set 1 and Set 2 was estimated to be about 3.0 m and 2.0 m, respectively. Here, there is no relation between the spacing between survey lines selected for the survey grid and the estimated spacing of fracture sets. The spacings of survey lines were 3.0 m and 2.0 m for radargrams orthogonal to Set 2 and Set 1, respectively. The spacing of survey lines mainly controls the accuracy (more geometric details) of modeling 3D curved surfaces of fractures. Since the technical manager of this quarry reported that the target is to cut blocks with the largest surface area possible (X–Y plane), and due to the variable thickness of the layers, any reasonable thickness (Z) of the extracted blocks can be acceptable. Since the radargrams allow for the estimation of the thickness of the layers, the expected average size of the exploitable block for each layer is equal to about 3.0 X (m), 2.0 Y (m), Z (m) of the layer. This is quite a rough estimation when considered in regards to quarrying. Therefore, the authors are developing a quarrying optimization algorithm based on the mentioned GPR fracture modeling strategy to find the optimum size of a non-fractured block as a function of the optimal cutting pattern in order to maximize production and minimize waste. 7. Conclusions In this study, a dual-frequency GPR system (250 MHz and 700 MHz) was used to survey a large-scale area of a limestone bench. The 700 MHz antenna proved to be effective in penetrating rock material studied at a large depth (4.0 m), providing a high resolution. This allowed us to image the thickness of the different rock layers in the studied bench and to detect the position, geometry, and dip of the fracture traces in the radargrams. The bulk dielectric constant of the rock mass was calculated by fitting several hyperbolic diffractions of the fractures detected at different depths as an alternative solution, since the “depth to know reflector” method could not be applied. The modeling approach implemented in this paper allowed us to deterministically model and visualize fractures on a large area by dividing volumes. The implementation of the modeling strategy in this paper can highlight further implementations, either in small or large areas, in different GPR applications, not only in quarrying, where subsurface planner anomalies exist. The resulting 3D model in the case study allowed us to
Geosciences 2018, 8, 481
11 of 15
roughly estimate the average exploitable size of non-fractured blocks within the rock layers of the bench: 3.0 X (m), 2.0 Y (m), Z (m) of layer. The development of a large area 3D deterministic model of fractures before cutting benches into blocks can be strategic, as it can help us to maximize quarrying production whilst minimizing waste. This work proves that a 3D fracture model can be obtained through a GPR survey. Moreover, the authors are currently conducting further research with the aim to make a step forward: The realization of a production optimization algorithm to find the optimum cutting pattern capable of maximizing the number of non-fractured blocks for each layer. Author Contributions: M.E. is the main contributor of the paper, he contributed in conceptualization, investigation, GPR survey, data processing, modeling, software design, 3D visualizations, analysis, and writing the first draft. F.T. contributed in GPR survey from both administrative and technical sides, project administration and instrumentation. S.K. contributed in GPR survey from both administrative and technical sides. F.G. contributed in investigation, project administration and instrumentation. S.B. contributed in conceptualization, investigation, modeling, software design and implementation, 3D visualizations, project administration and instrumentation. R.B. supervised the work. All authors contributed in reviewing and editing the paper. Funding: This research received no external funding. Acknowledgments: The authors are grateful to Michele Augelli, “Augelli Marmi” quarrying company, for allowing the GPR test in their quarry area and to Carlo Cormio (SERENGEO Srl, a spin-off company of the University of Bologna) for the technical suggestions provided. A special word of thanks goes to the EU-METALIC II—ERASMUS MUNDUS program that funded the lead author’s PhD scholarship. Conflicts of Interest: The authors declare no conflicts of interest.
Appendix A The interpretations of the fractures in the radargrams of the 700 MHz antenna in two volumes, in the X and Y directions, are provided below.
Geosciences 2018, 8, 481 Geosciences 2018, 8, x FOR PEER REVIEW Geosciences 2018, 8, x FOR PEER REVIEW
Figure The F01, F02, F02, F03, and and F04. Figure A1. A1. Volume Volume 1: 1: The interpreted interpreted radargrams radargrams Figure A1. Volume 1: The interpreted radargrams F01, F01, F02, F03, F03, and F04. F04.
12 of 15 12 of 15 12 of 15
Geosciences 2018, 8, 481 Geosciences2018, 2018,8,8,xxFOR FORPEER PEERREVIEW REVIEW Geosciences
13 of 15 13 of 15 15 13 of
FigureA2. A2.Volume Volume7:7:The Theinterpreted interpretedradargrams radargramsF23, F23,F24, F24,F25, F25,and andF26. F26. Figure F26.
References References 1. 1.
Elkarmoty, M.; M.; Colla, Colla, C.; C.; Gabrielli, Gabrielli, E.; E.; Bonduà, Bonduà, S.; S.; Bruno, Bruno, R. R. Deterministic Deterministic three-dimensional three-dimensional rock rock mass mass Elkarmoty, fracturemodeling modelingfrom fromgeo-radar geo-radarsurvey: survey:A Acase casestudy studyin inaasandstone sandstonequarry quarryin inItaly. Italy.Environ. Environ.Eng. Eng.Geosci. Geosci. fracture 2017, 23, 314–331, doi:10.2113/gseegeosci.23.4.314. 2017, 23, 314–331, doi:10.2113/gseegeosci.23.4.314.
Geosciences 2018, 8, 481
14 of 15
References 1.
2. 3. 4. 5. 6. 7.
8.
9. 10.
11. 12. 13.
14.
15.
16.
17.
18.
19.
20.
Elkarmoty, M.; Colla, C.; Gabrielli, E.; Bonduà, S.; Bruno, R. Deterministic three-dimensional rock mass fracture modeling from geo-radar survey: A case study in a sandstone quarry in Italy. Environ. Eng. Geosci. 2017, 23, 314–331. [CrossRef] Gillespie, M.R.; Barnes, R.P.; Milodowski, A.E. British Geological Survey Scheme for Classifying Discontinuities and Filings; Keyworth: Nottingham, UK, 2011. Montani, C. Stone 2002—World Marketing Handbook; Gruppo Editoriale Faenza Editice: Faenza, Italy, 2003. Montani, C. Stone 2008—World Marketing Handbook; Gruppo Editoriale Faenza Editice: Faenza, Italy, 2008. Vidi´c, D.; Gali´c, I.; Farkaš, B. The profitability of dimension stone deposit exploitation in relation to the coefficient of utilization. Min.-Geol.-Pet. Eng. Bull. 2012, 25, 123–130. Gali´c, I.; Vidi´c, D.; Jembrich, Ž. The influence of the coefficient of utilization of reservoir feasibility of dimension stone production and improvement opportunities. Min. Geol. Bull. 2011, 15, 117–130. Zanzi, L.; Hojat, A.; Ranjbar, H.; Karimi-Nasab, S.; Azadi, A.; Arosio, D. GPR measurements to detect major discontinuities at Cheshmeh-Shirdoosh limestone quarry, Iran. Bull. Eng. Geol. Environ. 2017, 1–10. [CrossRef] Lualdi, M.; Zanzi, L. 2D and 3D experiments to explore the potential benefit of GPR investigations in planning the mining activity of a limestone quarry. In Proceedings of the Tenth International Conference on Ground Penetrating Radar (GPR2004), Delft, The Netherlands, 2004; pp. 613–616. Rey, J.; Martínez, J.; Vera, P.; Ruiz, N.; Cañadas, F.; Montiel, V. Ground-penetrating radar method used for the characterisation of ornamental stone quarries. Constr. Build. Mater. 2015, 77, 439–447. [CrossRef] Martínez, J.; Montiel, V.; Rey, J.; Cañadas, F.; Vera, P. Utilization of Integrated Geophysical Techniques to Delineate the Extraction of Mining Bench of Ornamental Rocks (Marble). Remote Sens. 2017, 9, 1322. [CrossRef] Seol, S.J.; Kim, J.-H.; Song, Y.; Chung, S.-H. Finding the strike direction of fractures using GPR. Geophys. Prospect. 2001, 49, 300–308. [CrossRef] Luodes, H.; Selonen, O.; Pääkkönen, K. Evaluation of dimension stone in gneissic rocks—A case history from southern Finland. Eng. Geol. 2000, 52, 209–223. [CrossRef] Elkarmoty, M.; Colla, C.; Gabrielli, E.; Bonduà, S.; Bruno, R. A combination of GPR survey and laboratory rock tests for evaluating an ornamental stone deposit in a quarry bench. Procedia Eng. 2017, 191, 999–1007. [CrossRef] Elkarmoty, M.; Colla, C.; Gabrielli, E.; Kasmaeeyazdi, S.; Tinti, F.; Bonduà, S.; Bruno, R. Mapping and modelling fractures using ground penetrating radar for ornamental stone assessment and recovery optimization: Two case studies. Min.-Geol.-Pet. Eng. Bull. 2017, 32, 63–76. [CrossRef] Elkarmoty, M.; Colla, C.; Gabrielli, E.; Bonduà, S.; Bruno, R. Application of low frequency GPR antenna to fractures detection and 3D visualization in a new quarry bench (E-poster). In Proceedings of the International Conference on Geophysics, Orlando, CO, USA, 6–7 October 2016. And Published Abstract in J. Geol. Geophys. 2016, 5, 47. [CrossRef] Luodes, H.; Sutinen, H. Evaluation and modelling of natural stone rock quality using Ground Penetrating Radar (GPR). In Geoscience for Society: 125th Anniversary Volume; Nenonen, K., Nurmi, P.A., Eds.; Special Paper 49; Geological Survey of Finland: Espoo, Finland, 2011; pp. 83–90. Porsani, J.L.; Sauck, W.A.; Júnior, A.O.S. GPR for mapping fractures and as a guide for the extraction of ornamental granite from a quarry: A case study from southern Brazil. J. Appl. Geophys. 2006, 58, 177–187. [CrossRef] Arosio, D.; Deparis, J.; Zanzi, L.; Garambois, S. Fracture characterization with GPR: A comparative study. In Proceedings of the 16th International Conference Ground Penetrating Radar, Hong Kong, China, 13–16 June 2016. [CrossRef] Arosio, D.; Munda, S.; Zanzi, L. Quality control of stone blocks during quarrying activities. In Proceedings of the 14th International Conference Ground Penetrating Radar, Shanghai, China, 4–8 June 2012; pp. 822–826. [CrossRef] Arosio, D. Rock fracture characterization with GPR by means of deterministic deconvolution. J. Appl. Geophys. 2016, 126, 27–34. [CrossRef]
Geosciences 2018, 8, 481
21. 22. 23. 24. 25. 26. 27. 28.
29.
30. 31. 32. 33.
15 of 15
Elkarmoty, M.; Tinti, F.; Kasmaeeyazdi, S.; Bonduà, S.; Bruno, R. 3D modeling of discontinuities using GPR in a commercial size ornamental limestone block. Constr. Build. Mater. 2018, 166, 81–86. [CrossRef] Daniels, D.J. Ground Penetrating Radar; The Institute of Electrical Engineers: London, UK, 2004. Reynolds, J.M. An Introduction to Applied and Environmental Geophysics; John Wiley & Sons, Ltd.: West Sussex, UK, 2011. Yelfm, R.J. Application of ground penetrating radar to civil and geotechnical engineering. Electromagn. Phenom. 2007, 7, 103–117. Grandjean, G.; Gourry, J.C. GPR data processing for 3D fracture mapping in a marble quarry (Thassos, Greece). J. Appl. Geophys. 1996, 36, 19–30. [CrossRef] Grasmueck, M. 3-D ground-penetrating radar applied to fracture imaging in gneiss. Geophysics 1996, 61, 1050. [CrossRef] Grasmueck, M.; Quintà, M.C.; Pomar, K.; Eberli, G.P. Diffraction imaging of sub-vertical fractures and karst with full-resolution 3D Ground-Penetrating Radar. Geophys. Prospect. 2013, 61, 907–918. [CrossRef] Reina, A.; Loi, M. Environmental background in Apricena quarries (Apulia, Southern Italy). In Engineering Geology for Society and Territory—Volume 5. Urban Geology, Sustainable Planning and Landscape Exploitation; Lollino, G., Manconi, A., Guzzetti, F., Culshaw, M., Bobrowsky, P., Luino, F., Eds.; Springer International Publishing: Cham, Switzerland, 2015; pp. 1–1400. [CrossRef] Maerz, N.H.; Aqeel, A.M.; Anderson, N. Measuring orientations of individual concealed sub-vertical discontinuities in sandstone rock cuts integrating ground penetrating radar and terrestrial LIDAR. Environ. Eng. Geosci. 2015, XXI, 293–309. [CrossRef] Elkarmoty, M.; Colla, C.; Gabrielli, E.; Papeschi, P.; Bonduà, S.; Bruno, R. In-situ GPR test for three-dimensional mapping of the dielectric constant in a rock mass. J. Appl. Geophys. 2017, 146, 1–15. [CrossRef] Cassidy, N.J. Electrical and magnetic properties of rocks, soils and fluids. In Ground Penetrating Radar Theory and Applications; Elsevier: Amsterdam, The Netherlands, 2009; pp. 41–72. [CrossRef] Bourke, P. PLY-Polygon File Format. 2011. Available online: http://paulbourke.net/dataformats/ply/ (accessed on 19 March 2018). ParaView Website. n.d. Available online: http://www.paraview.org/ (accessed on 1 May 2018). © 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
