Geomechanics and Geoengineering

An International Journal

ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/tgeo20

Multiple regression model for prediction of rock

properties using acoustic frequency during core
drilling operations

Vijaya Kumar , Harsha Vardhan & Ch. S. N. Murthy

To cite this article: Vijaya Kumar , Harsha Vardhan & Ch. S. N. Murthy (2020) Multiple regression
model for prediction of rock properties using acoustic frequency during core drilling operations,
Geomechanics and Geoengineering, 15:4, 297-312, DOI: 10.1080/17486025.2019.1641631

To link to this article: https://doi.org/10.1080/17486025.2019.1641631

Published online: 22 Jul 2019.

Submit your article to this journal

Article views: 341

View related articles

View Crossmark data

Citing articles: 7 View citing articles

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=tgeo20
GEOMECHANICS AND GEOENGINEERING: AN INTERNATIONAL JOURNAL
2020, VOL. 15, NO. 4, 297–312
https://doi.org/10.1080/17486025.2019.1641631

Multiple regression model for prediction of rock properties using acoustic

frequency during core drilling operations
a
Ch. Vijaya Kumar , Harsha Vardhana,b and Ch. S. N. Murthya
a
Department of Mining Engineering, National Institute of Technology Karnataka, Surathkal, Mangalore, India; bMining, NITK, Surathkal,
Mangalore, India

ABSTRACT ARTICLE HISTORY

The primary purpose of this study is the quantification of rock properties uniaxial compressive strength Received 7 January 2019
(UCS), Brazilian tensile strength (BTS), density and abrasivity, using sound signal dominant frequencies Accepted 5 July 2019
produced during diamond core drilling operations. Rock drilling operations were performed on seven KEYWORDS
different types of rock samples, using a computer numerical control (CNC) drilling machine. Using the Rock drilling; regression
multiple regression analysis, satisfactory mathematical equations were developed for various physico- analysis; FFT; acoustic signal;
mechanical rock properties, as well as dominant frequencies of the sound level were generated during dominant frequencies
diamond core drilling operations. The developed prediction models demonstrated a good regression
coefficient between the rock properties and dominant frequencies i.e. the R2 values are 82.50%, 78.41%,
79.40%, and 93.24% for UCS, BTS, density and abrasivity, respectively. The performances indices are: (i)
root-mean-square error (RMSE) are 0.102754, 1.241652, 0.396727, and 0.697889 for UCS, BTS, density
and abrasivity, respectively; (ii) values account for (VAF) is 82.50008%, 78.41137%, 79.40137%, and
93.23596% for UCS, BTS, density and abrasivity, respectively. Presently, it is in the early stages of
development towards the prediction of rock properties using dominant frequencies with the help of
audio signal processing in the rock drilling operation. The developed prediction models can be utilised
at the early stages of mining and civil engineering projects, for the quantification of rock properties
using sound signal dominant frequencies.

1. Introduction Initially, the acoustic applications were implemented

by Obert (1941). Obert and Duvall (1942) investigated
Rock drilling is the process of extracting minerals, gas,
the occurrence of rock sub-audible noises in metal
and oil (in the mining industry) from the underground
mines. It was observed that, besides the rock noises,
space. Additionally, it is also used for tunnelling, excava-
other noises often present in the rock mass. The major-
tion, drainage, exploration, and other allied industrial
ity of rock noises usually travel less than 100 feet and
processes. Drilling is an integral step for mining/geotech-
these noises originate in areas under pressure. This was
nical projects. It is important for engineers/geologists to
demonstrated by probing an area suspected of being
be aware of the estimation of geo-mechanical rock prop-
under excessive pressure and correlating the number of
erties before beginning a project; without proper knowl-
rock noises with the occurrence of rock bursts. It was
edge of the geomechanical properties of rocks can lead to
concluded that in an area which subsequently bursts,
adverse effects in terms of cost and project duration.
the number of noises was relatively higher than in the
Generally, rock properties are not readily available at
surrounding rocks, and rock generates various noises
the sites and it is a time-consuming process. Various
under different stress states. The investigators utilised
techniques are available for the determination of geo-
sub-audible noises for quantifying the rock burst in
mechanical rock properties. For this, precious core rock
underground metal mines. Later, some studies reported
samples are required, and it is crucial to get these rock
detailed laboratory investigations on the characteristics
samples from raw rock mass/mining sites. In this study,
of carbonate rock and acoustic properties, with a vast
the reported quantification of rock properties is obtained
range of lithological and acoustic emission applications
using dominant frequencies and audio signal processing
in rock mechanics. Rafavich et al. (1984) investigated
in the rock drilling operation; this has not been reported
the relationship between the acoustic and rock proper-
anywhere by earlier investigators. However, acoustics
ties of brine saturated carbonate rocks. It was con-
application in the mining/geotechnical domain is limited.
cluded that the rock property results can be utilised

CONTACT Ch. Vijaya Kumar [email protected]

© 2019 Informa UK Limited, trading as Taylor & Francis Group
298 C. V. KUMAR ET AL.

to interpret porosity and lithology variations from lithological rock recognition, based on the vibro-acoustic
seismic sections. Hardy (1972) conducted an inclusive signal approach during rock drilling operations. The
review of acoustic emissions in the area of rock authors captured vibro-acoustic signals using vibration
mechanics. The authors conducted acoustic emission sensors and spectral wideband acoustic sensors. These
experiments related to mine design and the alleviation captured signals were analysed in terms of time domain
of rock burst in North America in the 1930s. It was and time-frequency domain for analysing the rock char-
concluded that the acoustic emission technique pro- acteristics. It was concluded that the spectral wideband
vides the behaviour of geologic rock materials’ defor- acoustic sensors were providing better signals to the noise
mation and failure. McNally (1990) studied the ratio than vibration sensors. Zhang et al. (2018a) intro-
quantification of coal strengths and elastic moduli in duced a new index for evaluating coal brittleness from
Queensland using sonic log techniques. The author fracture networks, using thee hydraulic fracturing process.
developed mathematical relationships between geo- The acoustical signals were captured during uniaxial
technical parameters and sonic log interval transit compression and triaxial compression tests; these cap-
times. Miklusova et al. (2006) and Krepelka and Futo tured signals were utilised and correlated with coal brit-
(2007) probed a rock disintegration process, using tleness. It was concluded that acoustical signals showed
acoustical signals during rotary drilling operations. It sudden changes when reaching the yield stress and peak
was concluded that the acoustical signal approach was strength, representing high brittleness. Zhang et al.
an effective tool for identifying the rock disintegration (2018b) reported that the acoustic emission technique
process. Gradl et al. (2012) attempted identifying the was employed to detect the initiation and evolution of
relationship between bit design and bit’s vibrational micro cracks of rocks during the laboratory investiga-
characteristics. Based on the vibro-acoustic frequency tions. Li et al. (2018), Jia et al. (2018), and Feng et al.
analysis, bit characteristics can be determined during (2019) conducted a laboratory investigation on hydraulic
drilling operations. In the following order, Zborovjan fracturing, using layered shale rock samples using the
(2001), Zborovjan (2002), Zborovjan et al. (2003), acoustic emission technique. The result concluded that
investigated the identification of rock types based on the characteristics of the injection pressure curve and
a hidden Markov model for rock drilling operations. acoustic emission response detected the hydraulic beha-
The Markov model recognised the particular acoustic viour growth in layered shale. Hu et al. (2019) conducted
signature of every rock type being drilled during the a laboratory experiment on the rock burst process of bore
operation. It is said that the maximum information was holes using the acoustic emission technique. It was con-
contained in the appropriate signal transferring and the cluded that sharp and high amplitude acoustical signals
rock drilling acoustic signature was found to be could be used for the quantification of the rock burst,
between 5000 and 8000 Hz. while micro cracks and splitting dominated the failure
Recently, a few studies investigated the prediction of process. He et al. (2019) investigated rock burst disasters
rock properties using the sound level produced during in thick coal seams and steeply inclined coal seams. The
the rock drilling/rock cutting operations. The authors authors used microseismic and acoustic emission techni-
derived the prediction from mathematical equations for ques as an early warning before the rock burst occurred
various geomechanical rock properties, with an accepta- during mining. Sheng et al. (2019) carried out a detailed
ble degree of accuracy (Vardhan and Murthy 2007, study about water jet rock drilling efficiency in relation to
Vardhan et al. 2009, Kumar et al. 2011a, 2011b, 2011c, acoustic emission. The authors correlated rock drilling
2013a, 2013b, Kahraman et al. 2013, Shreedharan et al. efficiency to acoustical signals during drilling operations.
2014, Kivade et al. 2015, Masood 2015, Delibalta et al. It was concluded that the high-frequency band reported
2015). Similarly, Karakus and Perez (2014) conducted a good correlation with the rate of penetration. Kumar
laboratory experiments using impregnated diamond et al. (2019a), (2019b) conducted laboratory experiments
core drilling operations. The authors developed linear for the prediction of rock properties, using sound signals
relationships between the acoustic emission signals and during core drilling operations. The authors developed
diamond drill bit wear. It was concluded that the devel- a simple linear regression model for the prediction of
oped linear relationships could predict the depth of the physico-mechanical rock properties with admissible accu-
cut, weight on the drill bit, wear of the drill bit, and racy. Apparently, it is evident that in rock mechanics very
torque on the drill bit, using the time domain of the few investigations have been conducted in this acoustic
acoustical signal. Rostami et al. (2015), Qin et al. (2018), area. Currently, it is the early stages of development,
Xiao et al. (2018), and Flegner et al. (2019) investigated working towards the prediction of rock properties using
 GEOMECHANICS AND GEOENGINEERING: AN INTERNATIONAL JOURNAL 299

dominant frequencies with the help of audio signal pro- operation. The experimental room was covered comple-
cessing in the rock drilling operation. Thus, the present tely with glass and fibre panels with dimensions of
research work can be taken in this direction. 6 m length, 5 m width, and 9 m height. The CNC machine
had 450 mm x 900 mm-table size, a 6 bar-optimum air
pressure, and it was connected to 415 V 3-Phase 50
2. Rock samples used in the investigation Hz power supply.
The experiments were conducted using cubical rock
samples of 20 cm length x 20 cm width x 20 cm height 3.2 Data acquisition system
(Kumar et al. 2019b). A total of seven different rock
samples were used in these experiments, which were Data acquisition systems measure the pressure of
collected from different places of Andhra Pradesh and sound using computer interface. It consists of
Karnataka. Table 1 provides details about the location a computer, a microphone (model: GRAS type 40
and rock types collected. The collected rock samples PH), a data acquisition card hardware (model:
were inspected for any macroscopic (visible) defects, NIUSB-9234) along with a LabVIEW software. It is
fractures, and joints before conducting the experiment. a more effective measuring instrument, as compared
to traditional measurement systems (Forouharmajd
et al. 2015). The LabVIEW software is a flexible soft-
3. Equipment/instruments ware for the analysis of any measurement system.
3.1 CNC drilling machine The microphone and data acquisition cards (from
national instruments) were used for measuring the
All the rock drilling experiments were conducted using sound pressure level in the diamond core drilling
the BMV 45 T20 CNC (model: BMV 45 T20) machine, as operations. The microphone was highly sensitive for
shown in Figure 1, which involves a highly automated recording accurate sound pressure levels (audio sig-
nals). The microphone specifications are a follows:
Table 1. Different rock samples collected with location. Frequency range = 10 Hz to 20 KHz, dynamic range
Sl.no Location Rock type = 32 dB (A)–135 dB, sensitivity = 50 mv/pa. The
1 Veldurthy (Village)/Kurnool (District) Ochre specifications for the data acquisition card are as
Andhra Pradesh state, India follows: No. of channel = 4 analog input channels,
2 Khammam (Village)/Kothagudem (District) Bituminous coal
Telangana State, India ADC resolution = 24 bits, type of ADC = delta–
3 Padubidri (Village)/Udupi (District) Laterite sigma (with analog prefiltering).
Karnataka state, India
4 Bethamcherla (Village)/Kurnool (District) Pink limestone
Andhra Pradesh state, India
5 Bethamcherla (Village)/Kurnool (District) Black limestone 3.3 Uniaxial compressive strength
Andhra Pradesh state, India
6 Bethamcherla (Village)/Kurnool (District) Ironstone Uniaxial compressive strength (UCS) is the most
Andhra Pradesh state, India
7 Bethamcherla (Village)/Kurnool (District) Dolomite
essential physico-mechanical property of rocks, used
Andhra Pradesh state, India widely in the mining industry, petroleum industry,
tunnelling and other excavation projects. The com-
pression-testing machine (model: AIM-317E-Mu)
was used for the measurement of UCS in cylindrical
core rock samples, as per ISRM suggested methods
(Brown 1981). The maximum loading capacity of the
compression-testing machine was 2,000 kN.

3.4 Brazilian tensile strength

The Brazilian tensile strength (BTS) of the rock specimens
was determined indirectly by using a BTS testing machine, as
per the ISRM suggested methods by Brown (1981) & Mellor
and Hawkes (1971). Due to the unavailability of direct
determination of tensile strength in the Rock Mechanics
Figure 1. BMV 45 T20 CNC drilling machine. Laboratory of Department of Mining Engineering of NIT
300 C. V. KUMAR ET AL.

Karnataka, Surathkal, India, the BTS testing machine has Table 2. Physico-mechanical rock properties.
upper and lower jaws along with guide pin. A hydraulic jack Rock properties
(100 kN) is fixed in the middle, between the guide pin. This Name of rock UCS BTS Density (g/ Abrasivity
S. No sample (MPa) (MPa) cm3) (%)
hydraulic jack consists of an oil reservoir, an integral pump-
1 Ochre 14.77 1.02 2.38 70.84
ing unit, and an operating handle. On the top, the pressure 2 Bituminous coal 16.37 1.09 1.75 57.25
gauge (100 kN) was fixed to the jack for the display of load 3 Laterite 39.99 1.87 2.93 50.66
4 Pink limestone 51.49 2.32 2.49 23.52
on the specimen. 5 Black limestone 53.01 2.66 2.62 18.29
6 Iron stone 120.25 7.54 3.60 16.29
7 Dolomite 127.74 7.62 3.01 10.98

3.5 Los angeles abrasion

The Los Angeles abrasion of the rock aggregates was 5. Experimental set-up
determined by the Los Angeles abrasion testing machine, The experimental set-up for carrying out drilling, using the
as per the ISRM suggested method, in the Rock Mechanics CNC machine on rock blocks and sound pressure level
Laboratory of Department of Mining Engineering of NIT measurement, is shown in Figure 2. The rock blocks were
Karnataka, Surathkal, India. It consisted of a hollow cylin- held at a particular position firmly during drilling, using
der with both ends closed. The inside length and diameter a vice and two long nuts and bolts. In this investigation,
of the cylinder are 508 ± 5 mm and 711 ± 5 mm. The cast drill bits of diameters 6, 10, 16, 18 and 20 mm (industrial
iron spheres were used for abrasion charge and weighed diamond core drill bit) at speeds of 150, 200, 250, 300 and
between 390 and 445 g. 350 rpm and penetration rates of 2, 3, 4, 5and 6 mm/min
were used for all the rock drilling operations. The CNC
4. Methodology machine was set to drill a 30 mm drill hole length using the
CNC programming, throughout the experiments. The
4.1 Determination of physico-mechanical microphone of the data acquisition system (DAQ-NIUSB
properties of rocks -9234) was held at a distance of 1.5 cm from the drill bit,
The UCS, BTS, and Los Angeles abrasion tests were deter- with the help of a magnetic holder. The DAQ microphone
mined, as per the ISRM suggested methods (Brown 1981, was connected to DAQ-NIUSB-9234 with a 1-channel
Mellor and Hawkes 1971). Whereas the dry density test cable, which was connected to the personal computer
was determined according to the ISRM (Ulusay and with windows operating system (64-bit) using the USB
Hudson 2007), with the equipment explained in section cable. The data recorded using the data acquisition system
3. The physico-mechanical properties of the rocks are could be visually seen on the data acquisition personal
provided in Table 2. computer (DAQ-PC). Figure 2 shows a dynamometer

Figure 2. Experimental set-up of (BMVK 45 T20) CNC drilling machine.

 GEOMECHANICS AND GEOENGINEERING: AN INTERNATIONAL JOURNAL 301

and dynamometer digital indicator, noise dosimeter with data were used for the frequency analysis. For the
a microphone, and a dosimeter PC, which were used to modelling of rock properties, four important para-
determine specific energy. However, the description about meters were considered for the analysis in the present
specific energy is beyond the scope of this study. investigation – (i) drill bit diameters in mm, (ii) drill
bit speed in RPM, (iii) penetration rate in mm/min,
and (iv) dominant frequencies (Hz).
5.1 Sound pressure level measurement
The sound pressure level data (dB) were obtained
The sound pressure level was measured for all bit-rock using the following equation:
combinations using the DAQ microphone by means of
The sound pressure level ðSPLÞ
the LabVIEW software. For all sound pressure level (1)
measurements, the microphone was set aside at ¼ 20 log10 ðPr:m:s =Pref Þ ðdBÞ
a distance of 1.5 cm from the outer edge of the drill where
bit diameter, as shown in Figure 2. The audio signals Pr.m.s = the sound pressure measured by the root mean
during drilling were recorded by the microphone to the square (r.m.s)
computer, using the DAQ NIUSB-9234 (data acquisi-
tion module from national instrumentation) 24-bit Pref ¼ reference sound pressure ð2  105 N m2 or 20 μ PaÞ
ADC (analog to digital converter), which captured
51,200 samples within one second. The resolution of
the response was maintained at 1 Hz by reading all the
5.2 Selection of dominant frequencies (DF)
samples, which were obtained in one second. This
module was connected to the system and the data The drilling audio signals were analysed using the FFTs,
were obtained using serial communication. Fast after capturing the audio signals from the rock drilling
Fourier transformation (FFT) was analysed to obtain operations. A total of 60 s (after penetration of drill bit
the resonant frequencies with their amplitude of sound into the rock samples) of time domain data was selected
pressure level. for the analysis of the corresponding combinations of
Acquired data from the microphone (Figure 3(a)) drill bit diameters: 6 mm, 10 mm, 16 mm, 18 mm, and
were put into a Fast Fourier analyser to convert the 20 mm at penetration rates of 2 mm/min, 3 mm/min,
time domain signals to frequency domain signals. 4 mm/min, 5 mm/min, and 6 mm/min, and speeds of
Append signals toolkit (Figure 3(b)) was also used to 150 rpm, 200 rpm, 250 rpm, 300 rpm, and 350 rpm.
understand the time domain response of whole signal, The peak amplitude of FFT was selected for the analysis
and the FFT was done based on peak amplitude of the 60-s time domain plots, i.e. where the highest ampli-
observed from the time domain plots. The FFTs were tude from the overall time domain plots with the highest
obtained after filtering the raw data obtained from the amplitude FFT were selected, corresponding to ochre,
experimental measurement. A Butterworth band pass bituminous coal, laterite, pink limestone, black limestone,
filter (Figure 3(b)) was used to ensure that the FFT was ironstone, and dolomite. As this peak amplitude carries the
free from noisy signals, with the help of the LabVIEW maximum energy in the noise spectrum, the drilling audio
software. Hence, the sound measurements are accurate recording sampling rate was 51.2 kHz per second. The test
and noisy signals have been avoided. The LabVIEW conditions (drill bit diameter of 18 mm, penetration rate of
block diagram code for sound signal measurement is 6 mm, and drill bit speed of 150 rpm) are shown in Figures
shown in Figure 3(a,b). 4 (a–g) and 5(a–g) as examples of spectrum signals in the
The audio signals produced during drilling were time and frequency domains. Figures 4(a–g) and 5(a–g)
measured for 60 s for various drill bit conditions, show the differentiation in-between time domain plots for
penetration rates, and spindle speeds (i.e. combinations various rock samples, and the selection of the dominant
of the drill bit diameters of 6 mm, 10 mm, 16 mm, frequency (Hz) between 5000 and 8000 Hz from the FFT
18 mm, and 20 mm at penetration rates of 2 mm/min, results for various rock samples. From the time domain
3 mm/min, 4 mm/min,5 mm/min, and 6 mm/min and plots, the peak amplitudes can be found at the 11th sec,
speeds of 150 rpm, 200 rpm, 250 rpm, 300 rpm, and 33rd sec, 24th sec, 4th sec, 33rd sec, 52nd, and 29th sec for
350 rpm) for various rock samples (Table 1). Hence, ochre, bituminous coal, laterite, pink limestone, black lime-
for each rock type, a total of 125 test conditions were stone, ironstone, and dolomite, respectively, as shown in
arrived at: For each rock = 125 test conditions, Total: 7 Figure 4(a–g), whereas, following the FFT results, the
rock types × 125 = 875. These 875 test conditions are dominant frequency (Hz) can be found between 5000
shown in Table 3. For each of these test conditions, and 8000 Hz at 5255 Hz, 5085 Hz, 5002 Hz, 5062 Hz,
audio signals were measured for 60 s, hence this 60-s 7931 Hz, 6778 Hz, and 6514 Hz for ochre, bituminous
302 C. V. KUMAR ET AL.

(a)

(b)
Figure 3. (a) LabVIEW block diagram code utilised for acquiring the sound signal data from microphone. (b) LabVIEW block diagram
code for sound signal analysis.

coal, laterite, pink limestone, black limestone, ironstone, were arrived at, as shown in Table 3) dominant fre-
and dolomite, respectively, as shown in Figure 5(a–g). quencies were extracted from the frequency domain of
The 125 (i.e. combinations of drill bit diameters of each selected rock, where the highest sound pressure
6 mm, 10 mm, 16 mm, 18 mm, and 20 mm at pene- level (dB) was determined. These frequencies were
tration rates of 2 mm/min, 3 mm/min, 4 mm/min, called dominant frequencies corresponding to every
5 mm/min, and 6 mm/min and speeds of 150 rpm, rock type. These frequencies were used for the multi-
200 rpm, 250 rpm, 300 rpm, and 350 rpm, total test ple regression analysis to predict the rock properties.
conditions: 125 × 7 rock types = 875 test conditions These dominant frequencies are in-line with
 GEOMECHANICS AND GEOENGINEERING: AN INTERNATIONAL JOURNAL 303

Table 3. Experimental results of 125 dominant frequencies (Hz) for each rock sample.
Test Drill bit diameter Penetration rate Drill bit Speed Bituminous Pink Black
conditions (mm) (mm/min) (rpm) Ochre Coal Laterite limestone limestone Ironstone Dolomite
1 6 2 150 5047 7360 6238 8000 5127 7313 7911
2 6 2 200 6024 8000 6000 7500 8000 7845 5401
3 6 2 250 5024 6520 6258 7412 6347 7680 5874
4 6 2 300 7014 7251 6741 8000 7541 7845 7004
5 6 2 350 5092 7529 6852 5214 7935 7999 6458
6 6 3 150 6000 5241 7000 6894 6012 7890 5252
7 6 3 200 5791 7899 5012 7553 5154 6201 7000
8 6 3 250 8000 6894 7285 5213 7899 7589 7676
9 6 3 300 7922 7836 7341 5000 6000 7845 8000
10 6 3 350 7500 7856 7417 6471 7985 7001 5250
11 6 4 150 8000 8000 7108 8000 5246 7685 7714
12 6 4 200 7771 5478 8000 5445 6999 7000 7685
13 6 4 250 5214 6998 7561 7456 5892 6993 7000
14 6 4 300 6999 6548 8001 7534 6007 8000 7891
15 6 4 350 8000 7993 7985 7641 7262 6457 7245
16 6 5 150 6247 7854 8000 7648 7856 7000 7658
17 6 5 200 5693 7841 7532 8000 7977 7498 7999
18 6 5 250 7958 5001 7214 7412 8000 7741 5879
19 6 5 300 8000 6641 8000 8000 5213 7769 6457
20 6 5 350 7142 8642 8001 5798 7845 8000 5281
21 6 6 150 5190 5731 5169 6000 5213 6589 6447
22 6 6 200 5999 6521 5214 7417 7008 7894 8001
23 6 6 250 5097 7854 5000 7888 7104 7733 7451
24 6 6 300 5896 5411 5891 8001 5998 8000 7481
25 6 6 350 7982 8000 5899 7885 5276 7450 8000
26 10 2 150 8000 8000 5201 7905 5782 6793 7704
27 10 2 200 7862 6987 8000 6451 5841 6000 4751
28 10 2 250 6807 8000 6208 6001 6089 6854 6652
29 10 2 300 7891 7541 6985 5214 6974 5689 8002
30 10 2 350 8000 8000 6748 5850 5787 5000 7474
31 10 3 150 7856 6471 8000 6008 7048 6214 8000
32 10 3 200 8000 5412 6259 6471 6998 5648 5481
33 10 3 250 7894 7814 6009 7401 6854 7474 5278
34 10 3 300 8000 7992 6785 5158 5047 5234 5894
35 10 3 350 6899 5689 6456 6320 7852 5000 8000
36 10 4 150 5446 5155 7031 7000 6689 7458 5417
37 10 4 200 8000 7451 7008 7412 8000 5234 6000
38 10 4 250 8000 6560 7496 7652 7548 7000 7412
39 10 4 300 7900 8000 8000 7810 8000 7897 7869
40 10 4 350 5408 6014 5689 7691 5214 5200 7568
41 10 5 150 6897 6475 6841 7891 7994 5940 7914
42 10 5 200 5286 5412 6001 7941 6568 6000 7365
43 10 5 250 8001 8974 8007 8001 6389 5412 7856
44 10 5 300 5031 7986 5687 6990 5133 6692 6874
45 10 5 350 6589 5641 6003 8000 8000 8000 8000
46 10 6 150 8000 6531 6999 7121 7854 5285 6987
47 10 6 200 7894 7894 7074 7235 5489 5867 7000
48 10 6 250 5442 7891 8000 7452 5989 7000 7104
49 10 6 300 7214 6457 6741 7861 6000 7984 7689
50 10 6 350 8000 6812 7187 6002 7977 7614 7467
51 16 2 150 6899 5481 5257 5474 5899 5006 7900
52 16 2 200 7086 8546 5332 6000 8000 7689 7842
53 16 2 250 7952 8000 5623 5789 7451 7771 7800
54 16 2 300 8000 7541 5404 6014 7823 8000 6987
55 16 2 350 8000 5017 5051 6741 6458 7568 6589
56 16 3 150 6948 6587 8000 6589 6004 7999 6325
57 16 3 200 7856 8000 5813 6984 8000 7025 6985
58 16 3 250 6852 6891 5974 7000 6241 7001 6999
59 16 3 300 7992 5142 5009 7456 5182 6541 7458
60 16 3 350 8005 8012 5278 8000 8000 6741 7986
61 16 4 150 7899 5004 5382 6201 6503 8000 7327
62 16 4 200 5269 7412 6007 5641 7856 7892 7958
63 16 4 250 6589 7059 6102 5894 6247 6845 7800
64 16 4 300 7009 7999 8000 5987 8000 8000 8000
65 16 4 350 8000 5033 5014 5999 8001 7965 6589
66 16 5 150 8001 6548 8000 6000 7923 7999 8001
67 16 5 200 7896 6888 5137 6741 6748 7356 6510
68 16 5 250 7021 7415 6897 6589 6999 7769 5001
69 16 5 300 5016 5430 5700 6475 7548 7665 6897
70 16 5 350 7693 6851 7845 6859 7654 7412 6000
71 16 6 150 7864 5147 6997 6999 7931 5983 5268
(Continued)
304 C. V. KUMAR ET AL.

Table 3. (Continued).
Test Drill bit diameter Penetration rate Drill bit Speed Bituminous Pink Black
conditions (mm) (mm/min) (rpm) Ochre Coal Laterite limestone limestone Ironstone Dolomite
72 16 6 200 7894 8000 7986 7478 6992 5007 5000
73 16 6 250 8000 8214 5056 8000 6097 6928 6974
74 16 6 300 7999 8000 7771 5471 6000 6348 5000
75 16 6 350 5009 5144 5023 5497 5496 7009 5864
76 18 2 150 8000 5142 5186 6485 5007 5128 5870
77 18 2 200 5698 5647 7496 5684 8000 5012 5010
78 18 2 250 6891 6584 5065 7814 7897 6598 5478
79 18 2 300 5645 7814 8000 8201 5689 6002 8000
80 18 2 350 5000 5632 7921 6580 7855 6354 6570
81 18 3 150 7894 6582 7354 8009 6999 5031 6915
82 18 3 200 8000 5892 5053 6254 5623 6573 6748
83 18 3 250 6258 6354 7962 8001 6349 5209 6352
84 18 3 300 5024 5253 6090 7881 6798 5312 5684
85 18 3 350 7999 7414 7847 5641 5289 5796 5989
86 18 4 150 7589 5096 5641 7458 6788 5647 6000
87 18 4 200 5296 7896 5647 8560 7096 5858 6589
88 18 4 250 5649 5869 5131 8000 7900 5974 5601
89 18 4 300 7859 8000 5789 7451 8000 5674 5748
90 18 4 350 8000 5265 7898 7548 6201 6231 8000
91 18 5 150 6895 6987 8000 7992 6784 8000 5874
92 18 5 200 8000 8000 5163 8000 5556 7149 8000
93 18 5 250 7771 5899 5647 6547 8000 8000 5847
94 18 5 300 5202 5058 5060 7012 6899 5947 6947
95 18 5 350 6989 8000 5200 7999 7585 8000 5472
96 18 6 150 5255 5085 5002 5069 7931 6778 6514
97 18 6 200 7895 7899 5874 6301 6428 7008 6871
98 18 6 250 7458 8000 5050 6000 5789 7587 6412
99 18 6 300 8000 6008 5389 6197 5647 7566 8000
100 18 6 350 6997 5139 5104 6783 5142 7441 6893
101 20 2 150 8000 5341 7106 6314 5082 5040 6898
102 20 2 200 5269 6958 7000 7412 7412 5104 8000
103 20 2 250 5826 5994 6999 7631 6932 6289 7845
104 20 2 300 6845 8524 8000 7014 7124 5794 8000
105 20 2 350 7175 7200 6826 5418 5124 5014 5647
106 20 3 150 7000 5264 8000 8001 6000 6060 8000
107 20 3 200 6999 6891 6911 6521 5834 5121 6588
108 20 3 250 8001 5233 7000 6978 5667 6000 7450
109 20 3 300 7122 5833 6777 5555 7946 6589 6695
110 20 3 350 7849 7000 6948 6147 8000 6001 6541
111 20 4 150 8000 5516 6944 8000 5696 7000 8000
112 20 4 200 6524 7589 7000 4321 6890 8520 7450
113 20 4 250 5248 6999 6604 6589 6589 7442 6989
114 20 4 300 5896 7414 8001 5820 7899 7410 5898
115 20 4 350 7170 7796 6623 7410 8000 8000 7004
116 20 5 150 6000 5246 5117 6004 6977 7532 7878
117 20 5 200 6478 6845 6675 8001 5989 7854 7965
118 20 5 250 8000 6852 7321 5628 5748 7698 5681
119 20 5 300 7415 5781 6581 8521 6589 5705 5555
120 20 5 350 6475 6999 7648 8001 8000 8000 6235
121 20 6 150 7047 5361 5174 5646 7845 6740 7415
122 20 6 200 8000 7895 6765 8741 5801 5246 6000
123 20 6 250 5781 8000 5702 8000 7452 8000 5014
124 20 6 300 8000 7584 7999 6547 8000 7869 6021
125 20 6 350 7274 5109 6760 5049 5106 7546 6416

(Zborovjan 2001, 2002, Zborovjan et al. 2003), 6. Results and analysis

wherein they state that the relevant signals can be
6.1 Multiple regression analysis
found between 5000 and 8000 Hz in the rock-
porting acoustic signature during rock drilling opera- Multiple regression analysis is one of the most power-
tion. The basic statistical evaluations of dominant ful tools for estimating the relationship between the
frequencies and experimental rock properties’ tests several dependent and independent variables, using
results are shown in Table 4. mathematical relations. It provides information about
 GEOMECHANICS AND GEOENGINEERING: AN INTERNATIONAL JOURNAL 305

(a)

(b)

(c)

(d)

Figure 4. (a)–(g) Differentiation between time domain plots for various rock samples, i.e. (a) ochre, (b) bituminous coal, (c) laterite,
(d) pink limestone, (e) black limestone, (f) ironstone, and (g) dolomite at drill bit diameter of 18 mm, penetration rate of 6 mm/min,
and drill bit speed of 150 rpm.

the independent variables and their uses to make much operational parameters represent a quadratic model.
more powerful and accurate predictions of the depen- The quadratic model involves ‘n’ independent vari-
dent variables. The mathematical modelling for domi- ables, namely Xi, Xj . . . Xn, and the multiple regression
nant frequencies, produced during rock drilling, were equation is generally expressed as:
carried out using Minitab17. Dominant frequencies
produced during drilling depend on a number of para- X
n X
n X
n
Y ¼ a0 þ a i Xi þ aij Xi 2 þ aij Xi XJ þ error
meters, such as drill bit diameter, spindle speed, and i¼0 i¼1 i<j
penetration rate. For the development of the prediction
(2)
models, four important operational parameters were
used as the dependent variables (input parameters) – where Y denotes dependent variables, Xi, Xj, . . . Xn indicate
drill bit diameter (mm), spindle speed (RPM), penetra- the independent variables, ai is the linear parametric effect
tion rate (mm/min), and dominant frequencies (Hz). of xi, and aij represents the quadratic effect, and the third
The responses (output) were UCS, BTS, density, Los and fourth terms represent the combination of both. The
Angeles abrasion, and abrasivity. Hence, the detailed regression model includes linear, squared, and cross
306 C. V. KUMAR ET AL.

(e)

(f)

(g)

Figure 4. Continued.

(e)

(f)

(g)

Figure 5. Continued.
 GEOMECHANICS AND GEOENGINEERING: AN INTERNATIONAL JOURNAL 307

(a)

(b)

(c)

(d)

Figure 5. (a)–(g) Selected dominant frequency (Hz) between 5000 and 8000 Hz from FFT results at various rock samples, i.e. (a)
ochre, (b) bituminous coal, (c) laterite, (d) pink limestone, (e) black limestone, (f) ironstone, and (g) dolomite at drill bit diameter of
18 mm, penetration rate of 6 mm/min, and drill bit speed of 150 rpm.

product terms. A multiple regression model was developed statistically significant, and the corresponding data were
for UCS, BTS, density, and abrasivity considering 60 s of noted down, as shown in the regression Table 5.
drilling, after the penetration of the drill bit into the rock Development of regression models:
mass, as shown in Equations (3) and (6). To evaluate the Regression equation for UCS:
model, a backward elimination method was used as the test
procedure. Analysis of variance was performed to observe UCS ¼ 10:5  13:035 DD  0:0377 RPM
the essential parameters carried out in the statistical model þ 0:02758 FR þ 0:7698 DD  DD
for UCS, BTS, density and abrasivity, with a significance of
 0:000002 FR  FR þ 0:000006 RPM  FR
95% confidence interval. Influence of the parametric level
of the UCS, BTS, density and abrasivity were compared (3)
using ANOVA with the Minitab17. Where the p-values
equal or smaller than 0.005 were considered to be Regression equation for BTS:
308 C. V. KUMAR ET AL.

Table 4. Statistical values of physico-mechanical rock properties and range of dominant frequencies.
Rock Properties Ochre Bituminous coal Laterite Pink limestone Black limestone Iron stone Dolomite
UCS (MPa) Min 13.220 15.370 37.990 50.150 53.010 120.25 126.14
Max 14.770 16.370 39.990 51.490 54.220 122.25 128.74
SD 0.501 0.408 0.691 0.503 0.520 0.667 0.692
Mean 14.269 16.135 39.586 51.254 53.331 120.58 127.58
BTS (MPa) Min 1.0200 1.0900 1.0700 2.320 2.0600 6.540 7.2200
Max 1.1200 1.2900 1.8700 3.320 2.6600 7.540 7.6200
SD 0.0350 0.0632 0.2989 0.316 0.2461 0.327 0.1476
Mean 1.0360 1.1100 1.7300 2.420 2.5100 7.540 7.6200
Density (g/cm3) Min 2.1800 1.4500 2.9300 2.1900 2.1200 2.600 3.0100
Max 2.3800 1.7500 2.9300 2.5900 2.8200 3.610 3.2100
SD 0.0675 0.1160 0.000000 0.1059 0.1776 0.317 0.0675
Mean 2.3500 1.6800 2.9300 2.4600 2.6200 3.601 3.0100
Abrasivity (%) Min 70.640 57.250 50.610 21.520 18.280 16.290 10.480
Max 71.840 58.150 51.660 23.740 19.290 16.850 10.980
SD 0.333 0.282 0.319 0.642 0.316 0.176 0.162
Mean 70.840 57.259 50.654 23.525 18.292 16.292 10.903
Dominant Frequencies (Hz) Min 5000 5001 5000 4321 5007 5000 4751
Max 8005 8974 8007 8741 8001 8520 8002
SD 1066 1114.7 1049.5 987.1 1014.7 1015.4 961.4
Mean 6978.6 6777.3 6566.7 6886.7 6752.2 6819.8 6827.2

Table 5. Statistical analysis of significant regression models for various rocks.

Regression Adjusted
Standard Tabulated Tabulated coefficient coefficient
Model Variable Coefficient error t-value t-value F-value F- ratio (%) (%)
Equation Constant −10.5 33.5 −0.31 ±1.962 682.00 2.11 82.50 82.38
(1) Drill bit diameter(DD) −13.035 0.881 −14.79
Spindle speed (RPM) −0.0377 0.0558 −1.97
Frequency (FR) 0.02758 0.00945 2.92
(Drill bit diameter(DD))2 0.7698 0.0337 22.82
(Spindle speed (RPM))2 −0.000002 0.000001 −3.23
Spindle speed (RPM) X Frequency (FR) 0.000006 0.000008 1.97
Equation Constant 0.58 2.33 0.25 ±1.962 525.44 2.11 78.41 78.26
(2) Drill bit diameter(DD) −1.1561 0.0614 −18.82
Spindle speed (RPM) −0.00278 0.00389 −2.72
Frequency (FR) 0.001899 0.000659 2.88
(Drill bit diameter(DD))2 0.06017 0.00235 25.58
Frequency (FR)X Frequency (FR) 0.000001 0.00001 −3.20
Spindle speed (RPM) X Frequency (FR) 0.000003 0.000001 2.75
Equation Constant 2.915 0.792 3.68 ±1.962 207.41 2.02 79.60 79.40
(3) Drill bit diameter(DD) −0.2413 0.0268 −9.00
Spindle speed (RPM) −0.00127 0.00214 −2.02
Frequency (FR) 0.000219 0.000214 2.02
(Drill bit diameter(DD))2 0.011513 0.000753 15.30
Frequency (FR)X Frequency (FR) −0.000001 0.000001 −2.74
Drill bit diameter(DD)X Frequency (FR) 0.000005 0.000003 2.87
Spindle speed (RPM) X Frequency (FR) 0.000001 0.000001 2.06
Equation Constant 112.6 10.3 10.89 ±1.962 2998.03 2.38 93.24 93.20
(4) Drill bit diameter(DD) −4.651 0.252 −18.44
Frequency (FR) −0.00527 0.00301 −2.75
Frequency (FR)X Frequency (FR) 0.000001 0.000001 2.47
Drill bit diameter(DD)X Frequency (FR) 0.000088 0.000036 2.42

BTS ¼ 0:58  1:1561 DD  0:00278 RPM Density ¼ 2:915 0:2413 DD 0:00127 RPM
þ 0:001899 FR þ 0:06017 DD  DD þ 0:000219 FR þ 0:011513 DD  DD
 0:000001 FR  FR þ 0:000003 RPM  FR  0:000001 FR  FR þ 0:000005 DD  FR
(4) þ 0:000001 RPM  FR
(5)
Regression equation for density:
Regression equation for abrasivity:
 GEOMECHANICS AND GEOENGINEERING: AN INTERNATIONAL JOURNAL 309

Table 6. Results of ANOVA for various rocks.

Source DF Seq.SS Adj SS Adj MS F-Value P-Value
UCS
Equation (1) Regression 6 1,308,442 1,308,442 218,074 682.00 0.000
DD 1 1,136,383 69,965 69,965 2188.81 0.000
RPM 1 0 146 146 0.46 0.050
FR 1 359 2724 2724 8.52 0.004
DD*DD 1 168,212 166,496 166,496 520.70 0.000
FR*FR 1 3327 3345 3345 10.46 0.001
RPM*FR 1 160 160 160 0.50 0.005
Error 868 277,547 277,547 320 – –
Total 874 1,585,989 – – – –
BTS
Equation (2)
Regression 6 4899.61 4899.6 816.60 525.44 0.000
DD 1 3854.94 550.35 550.35 354.12 0.000
RPM 1 0.00 0.80 0.80 0.51 0.054
FR 1 1.11 12.92 12.92 8.31 0.004
DD*DD 1 1026.90 1017.06 1017.06 654.43 0.000
FR*FR 1 15.78 15.88 15.88 10.22 0.001
RPM*FR 1 0.87 0.87 0.87 0.56 0.055
Error 868 1348.99 1348.99 1.55 – –
Total 874 6248.60 – – – –
Density
Equation (3)
Regression 7 230.625 230.625 32.9464 207.41 0.000
DD 1 191.702 12.852 12.8521 80.91 0.000
RPM 1 0.000 0.165 0.1655 1.04 0.008
FR 1 0.355 0.166 0.1664 1.05 0.006
DD*DD 1 37.353 37.167 37.1675 233.99 0.000
FR*FR 1 0.488 0.479 0.4791 3.02 0.053
DD*FR 1 0.549 0.556 0.5560 3.50 0.002
RPM*FR 1 0.178 0.178 0.1784 1.12 0.007
Error 867 137.718 137.718 0.1588 - -
Total 874 368.343 - - - -
Abrasivity
Equation (4)
Regression 4 391,573 391,573 97,893.3 2998.03 0.000
DD 1 391,228 11,100 11,099.8 339.94 0.000
FR 1 88 100 100.3 3.07 0.080
FR*FR 1 67 71 70.7 2.17 0.001
DD*FR 1 191 191 190.8 5.84 0.016
Error 870 28,408 28,408 32.7 - -
Total 874 419,981 - - - -

Abrasivity ¼ 112:6 4:651 DD 0:00527 FR characteristics and identify the percentage contribution
þ 0:000001 FR  FR þ 0:000088 DD  FR ratio of each process factor on the output response.
Initially, ANOVA was performed between the input
(6)
parameters i.e. drill bit diameter (mm), penetration rate
where (mm/min), spindle speed (RPM), dominant frequencies
DD = diameter of the drill bit (mm) (DF) and rock properties (i.e. UCS, BTS, density, abra-
SS = spindle speed (rpm) sivity). All the insignificant terms (p > 0.05) from the
PR = penetration rate (mm/min) obtained model were removed, and the regression analy-
FR = dominant frequency (Hz) sis was carried out again, this time with significant terms
in the model. Table 6 shows the data obtained through
experimental analysis. Whereas, the P-value was found to
6.2 Analysis of variance (ANOVA)
be significantly good, at less than 0.005.
Regression analysis was carried out to obtain second-
order models for UCS, BTS, density, and abrasivity
during diamond core drilling operations. This test 6.3 Validation of the derived model
helps find the input parameters that considerably affect
the desired output response in the model. ANOVA is The statistical results of the UCS, BTS, density and
most the popular analysis tool for studying the signifi- abrasivity models for various rock types are demon-
cant parameters that influence the quality stated in Figure 4. The correlation coefficients of
310 C. V. KUMAR ET AL.

(a) (b)

(c) (d)

Figure 6. Measured UCS, BTS, density and abrasivity vs. predicted UCS. BTS, density and abrasivity.

these models i.e. the R2 values are 82.50%, 78.41%, (Alvarez and Babuska 1999, Finol et al. 2001,
70.40% and 93.24% for UCS, BTS, density and abra- Gokceoglu 2002, Yilmaz and Yuksek 2008, 2009,
sivity, respectively. Figure 6 shows the measured Yilmaz and Kaynar. 2011, Kumar et al. 2011b). The
value versus predicted value, corresponding to the performance indices of the developed regression model
UCS, BTS, density, and abrasivity. The scatter plots are shown in Table 7.
show that the values obtained from the multiple  
Varðy  y1Þ
regression models and the value measured from VAF ¼ 1   100 (7)
the experimental work, both are fairly close with Varð yÞ
less than 10% error. It can be said that the devel- sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
oped models are reasonably good. 1X n
RMSE ¼ ðy  y0 Þ2 (8)
N i¼0

6.4 Performance prediction of derived models

N  
(The prediction of performance is a better indicator for 1X  Ai  Pi
MAPE ¼  100 (9)
the developed prediction models, to assess the coeffi- N i¼1  Ai 
cient of correlations between the predicted and mea-
sured results. The value adjustment factor (VAF) and
the root-mean-square error (RMSE) were calculated for Table 7. Performance Indices of the developed regression
the predictive ability of the models, whereas, y and y1 model.
and Ai and Pi are the measured and predicted values, as Performance indices
shown in Equations (7) to (9). If VAF is equal to 100 Variable RMSE VAF (%) MAPE
UCS 0.102754 82.50008 0.027281
and RMSE is equal to zero then the model will be good. BTS 1.241652 78.41137 0.028388
Similarly, the mean absolute percentage error (MAPE) Density 0.396727 79.40137 0.007817
demonstrates the accuracy fit value of the statistics Abrasivity 0.697889 93.23596 0.00521
 GEOMECHANICS AND GEOENGINEERING: AN INTERNATIONAL JOURNAL 311

7. Conclusions physico-mechanical properties of rocks. Fluctuation and

Noise Letters, 14 (1), 1550006-1-11. doi:10.1142/
An attempt was made in this investigation to quantify rock S0219477515500066
properties using sound level at dominant frequencies, dur- Feng, X.T., et al., 2019. ISRM Suggested method for in situ
ing diamond core drilling operations. A satisfactory math- acoustic emission monitoring of the fracturing process in
ematical relationship between the physico-mechanical rock rock masses. Rock Mechanics and Rock Engineering, 52 (5),
1395–1414.
properties – UCS, BTS, density abrasivity, and the sound
Finol, J., Guo, Y.K., and Jing, X.D., 2001. A rule based fuzzy
level at dominant frequencies from rock drilling operations model for the prediction of petro physical rock parameters.
was determined. The developed equations are extremely Journal of Petroleum Science and Engineering, 29, 97–113.
essential for the estimation of rock properties in the early doi:10.1016/S0920-4105(00)00096-6
stages of mining and civil engineering projects. The devel- Flegner, P., et al., 2019. Evaluating noise sources in a working
oped equations for the quantification of rock properties can environment when disintegrating rocks by rotary drilling.
Polish Journal of Environmental Studies, 28 (5), 1–10.
be used in the field, with an acceptable degree of accuracy. doi:10.15244/pjoes/94848
In this study, the quantification of physico-mechanical Forouharmajd, F., et al., 2015. Sound pressure level tools design
rock properties was carried out using the peak amplitude used in occupational health by means of Labview software.
of the sound signal technique, and it was also used for the International Journal of Environmental Health Engineering, 4
quantification of rock properties during rock core drilling (1), 1–6.
Gokceoglu, C., 2002. A fuzzy triangular chart to predict the
operations. It needs to be emphasised that, in this research
uniaxial compressive strength of the Ankara agglomerates
work, specific energy (required time to remove a unit from their petrographic composition. Engineering Geology,
volume of rock mass) was not reported using the sound 66, 39–51. doi:10.1016/S0013-7952(02)00023-6
level produced during the diamond core drilling operation. Gradl, C., Eusteslll, A.W., and Thonhauser,G., G., 2012. An
The specific energy is an essential parameter to estimate analysis of noise characteristics of drill bits. Journal of
the required time to remove a unit volume of the rock Energy Resources Technology, ASME, 134 (013103), 1–6.
doi:10.1115/1.4005324
mass during excavation, tunnelling, quarrying, and other
Hardy, H.R., 1972. Application of acoustic emission techni-
allied industrial processes. It is recommended that further ques to rock mechanics research acoustic emission, ASTM
investigation be conducted in this direction, that is the STP 505. In: R. Liptai, D. Harris, and C. Tatro, eds.
estimation of specific energy using sound level produced American Society for Testing and Materials. Technical
during rock core drilling operations. Report, 41–83. doi:10.1520/STP35381S
He, S., et al., 2019. Precursor of spatio-temporal evolution
law of MS and AE activities for rock burst warning in
Acknowledgments steeply inclined and extremely thick coal seams under
caving mining conditions. Rock Mechanics and Rock
The authors would like to express their gratitude to the reviewers Engineering, 52 (7), 2415–2435.
for their comments and suggestions, which helped us improve Hu, X., et al., 2019. Experiment on rock burst process of borehole
the manuscript. Furthermore, the authors duly acknowledge the and its acoustic emission characteristics. Rock Mechanics and Rock
generous support provided by the Centre for System Design Engineering, 52 (3), 783–802. doi:10.1007/s00603-018-1613-z
(CSD): A centre of excellence at NITK, Surathkal. Jia, S.Q., et al., 2018. Investigating fracture growth and source
mechanisms in shale using acoustic emission technique. In:
52nd US Rock Mechanics/Geomechanics Symposium, Settle,
Disclosure statement WA, 17–20 June. American Rock Mechanics Association.
No potential conflict of interest was reported by the authors. Kahraman, S., Delibalta, M.S., and Comakli, R., 2013. Noise
level measurement test to predict the abrasion resistance
of rock aggregates. Fluctuation and Noise Letters, 12 (4),
ORCID 1350021-1-12. doi:10.1142/S0219477513500211
Karakus, M. and Perez, S., 2014. Acoustic emission analysis for
Ch. Vijaya Kumar http://orcid.org/0000-0001-6926-7852
rock–bit interactions in impregnated diamond core drilling.
International Journal of Rock Mechanics and Mining Sciences,
References 68, 36–43. doi:10.1016/j.ijrmms.2014.02.009
Kivade, S.B., Murthy, C.S.N., and Vardhan, H., 2015. ANN
Alvarez, G.M. and Babuska, R., 1999. Fuzzy model for the models for prediction of sound and penetration rate in per-
prediction of unconfined compressive strength of the rock cussive drilling. Journals of Institution of Engineers, India,
samples. The Journal of Rock Mechanics Mining Sciences, Series D, 96 (2), 93–103. doi:10.1007/s40033-015-0067-7
36, 339–349. doi:10.1016/S0148-9062(99)00007-8 Krepelka, F. and Futo, J., 2007. Acoustics aspects of techno-
Brown, E.T., 1981. Characterization testing and monitoring. logical process in the rock identification. Acta
Pergamon, Oxford: International society for rock Montanistica Slovaca, 12 (1), 25–28.
mechanics (ISRM) Suggested methods. Kumar, B.R., Vardhan, H., and Govindaraj, M., 2011a. A new
Delibalta, M.S., Kahraman, S., and Comakli, R., 2015. The usabil- approach for estimating of rock properties of meta-
ity of noise level from rock cutting for the prediction of morphic rocks. The International Journal of Mining and
312 C. V. KUMAR ET AL.

Mineral Engineering, 3 (2), 109–123. doi:10.1504/ Rostami, J., et al., 2015. Rock characterization while drilling and
IJMME.2011.042426 application of roof bolter drilling data for evaluation of ground
Kumar, B.R., Vardhan, H., and Govindaraj, M., 2011b. Prediction conditions. Journal of Rock Mechanics and Geotechnical
of uniaxial compressive strength, tensile strength and porosity Engineering, 7, (3), 273–281. doi:10.1016/j.jrmge.2015.01.006
of sedimentary rocks using sound level produced during rotary Sheng, M., et al., 2019. Frequency analysis of multi-sources
drilling. Rock Mechanics and Rock Engineering, 44, 613–620. acoustic emission from high-velocity waterjet rock drilling
doi:10.1007/s00603-011-0160-7 and its indicator to drilling efficiency. International
Kumar, B.R., Vardhan, H., and Govindaraj, M., 2011c. Sound level Journal of Rock Mechanics and Mining Sciences, 115,
produced during rock drilling Vis-à-vis rock properties. 137–144. doi:10.1016/j.ijrmms.2019.01.005
Engineering Geology, 123 (4), 333–337. doi:10.1016/j. Shreedharan, S., et al., 2014. Acoustic fingerprinting for rock
enggeo.2011.09.009 identification during drilling. International Journal of
Kumar, B.R., et al., 2013b. Artificial neural network model for Mining and Mineral Engineering, 5 (2), 89–105.
Prediction of rock properties from sound levels produced during doi:10.1504/IJMME.2014.060193
drilling. An International Journal of Geomechanics and Ulusay, R. and Hudson, J.A., 2007. The complete ISRM suggested
Geoengineering, 8 (1), 53–61. doi:10.1080/17486025.2012.661469 methods for rock characterization testing and monitoring
Kumar, B.R., et al., 2013a. Regression analysis and ANN 1974–2006. Ankara, Turkey: Compilation arranged by the
models to predict rock properties from sound levels pro- ISRM Turkish national group. ISBN: 978-975-93675-4-4. 628.
duced during drilling. International Journal of Rock Vardhan, H., Adhikari, C.R., and Govindaraj, M., 2009.
Mechanics and Mining Sciences, 58, 61–72. doi:10.1016/j. Estimating rock properties using sound levels produced during
ijrmms.2012.10.002 drilling. International Journal of Rock Mechanics & Mining
Kumar, C.V., Vardhan, H., and Murthy, C.S., 2019a. Sciences, 46 (2009), 604–612. doi:10.1016/j.ijrmms.2008.07.011
Quantification of rock properties using frequency analysis Vardhan, H. and Murthy, C.S.N., 2007. An experimental
during diamond core drilling operations. Journal of the investigation of jack hammer drill noise with special
Institution of Engineers (india): Series D, 100 (1), 1–15. emphasis on drilling in rocks of different compressive
Kumar, C.V., et al., 2019b. Estimating rock properties using strengths. Noise Control Engineering Journal, 55 (3),
sound signal dominant frequencies during diamond core 282–293. doi:10.3397/1.2737667
drilling operations. Journal of Rock Mechanics and Xiao, Y., et al., 2018. Investigation of active vibration drilling
Geotechnical Engineering. doi:10.1016/j.jrmge.2019.01.001 using acoustic emission and cutting size analysis. Journal
Li, N., et al., 2018. Acoustic emission response of laboratory of Rock Mechanics and Geotechnical Engineering, 10 (2),
hydraulic fracturing in layered shale. Rock Mechanics and 390–401. doi:10.1016/j.jrmge.2017.10.002
Rock Engineering, 51 (11), 3395–3406. doi:10.1007/s00603- Yilmaz, I. and Kaynar., O., 2011. Multiple regression, ANN
018-1547-5 (RBF, MLP), and ANFIS models for prediction of swell
Masood, 2015. Estimation of sound level produced during potential of clayey soils. Expert Systems with Application,
drilling of igneous rock samples using a portable drill set- 38, 5958–5966. doi:10.1016/j.eswa.2010.11.027
up, global challenges policy frame work & sustainable Yilmaz, I. and Yuksek, A.G., 2008. Technical note an exam-
development for mining of mineral and fossil energy ple of artificial neural network (ANN) application for
resources (GCPF). Procedia Earth and Planetary Science, indirect estimation of rock parameters. Rock Mechanics
11, 469–482. doi:10.1016/j.proeps.2015.06.047 and Rock Engineering, 41 (5), 781–795. doi:10.1007/
McNally, G.H., 1990. The prediction of geotechnical rock s00603-007-0138-7
properties from sonic and neutron logs. Australian Yilmaz, I. and Yuksek, A.G., 2009. Prediction of the strength
Society of Exploration Geophysics (ASEG), 21 (1/2), and elasticity modulus of gypsum using multiple regres-
65–71. doi:10.1071/EG990065 sion, ANN,ANFIS models and their comparison.
Mellor, M. and Hawkes, I., 1971. Measurement of tensile strength International Journal of Rock Mechanics and Mining
by diametral compression of discs and annuli. Engineering Sciences, 46 (4), 803–810. doi:10.1016/j.ijrmms.2008.09.002
Geology, 5, 173–225. doi:10.1016/0013-7952(71)90001-9 Zborovjan, M., 2001. Identification of minerals during dril-
Miklusova, V., et al., 2006. Acoustic signal – new feature in ling process via acoustic signal. Metallurgy and Foundry,
monitoring of rock disintegration process. Contributions Krakow, Poland, 26 (4), 367–374.
to Geophysics Geodesy, SAS36, SI6SGK, 36, 125–133. Zborovjan, M., 2002. Identification of minerals from sound
Obert, L., 1941. Use of sub audible noises for prediction of during drilling. Semestral project TU-Kosice. Poland: Acta
rock bursts – part I. Washington, DC: U.S. Dept. of the Montanistica Slovaca.
Interior, Bureau of Mines, (1941) R.I.3555. Zborovjan, M., Lesso, I., and Dorcak, L., 2003. Acoustic
Obert, L. and Duvall., W., 1942. Use of sub audible noises for identification of rocks during drilling process. Journal of
prediction of rock bursts – part II.Washington, DC: U.S. Dept. Acta Montanistica Slovaca, 8 (4), 191–193.
of the Interior, Bureau of Mines, R.I.3654. Zhang, J., et al., 2018a. Energy-based brittleness index and acoustic
Qin, M., et al., 2018. Analysis of signal characteristics from emission characteristics of anisotropic coal under triaxial stress
rock drilling based on vibration and acoustic sensor condition. Rock Mechanics and Rock Engineering, 51 (11),
approaches. Applied Acoustics, 140, 275–282. doi:10.1016/ 3343–3360. doi:10.1007/s00603-018-1535-9
j.apacoust.2018.06.003 Zhang, R., et al., 2018b. Failure characterization of three
Rafavich, F., Kendall, C.H., and Todd, T.P., 1984. The rela- typical coal-bearing formation rocks using acoustic emis-
tionship between acoustic properties and the petrographic sion monitoring and X-ray computed tomography
character of carbonate rocks. Geophysics, 49 (10), techniques. Rock Mechanics and Rock Engineering, 52 (6),
1622–1636. doi:10.1190/1.1441570 1945–1958.