Ballistic Behavior of Boron Carbide Reinforced AA7075 Aluminium Alloy

Download as pdf or txt
Download as pdf or txt
You are on page 1of 7

+ MODEL

H O S T E D BY Available online at www.sciencedirect.com

ScienceDirect
Defence Technology xx (2015) 1e7
www.elsevier.com/locate/dt

Ballistic behavior of boron carbide reinforced AA7075 aluminium alloy


using friction stir processing e An experimental study and analytical
approach
I. SUDHAKAR a, G. MADHUSUDHAN REDDY b, K. SRINIVASA RAO c,*
a
Department of Mechanical Engineering, MVGRCE, Vizianagaram, India
b
Defence Metallurgical Research Laboratory, Hyderabad, India
c
Department of Metallurgical Engineering, Andhra University, Visakhapatnam, India
Received 18 March 2015; revised 20 April 2015; accepted 22 April 2015
Available online ▪ ▪ ▪

Abstract

High strength-to-weight ratio of non-ferrous alloys, such as aluminium, magnesium and titanium alloys, are considered to be possible
replacement of widely accepted steels in transportation and automobile sectors. Among these alloys, magnesium is self explosive and titanium is
costlier, and aluminium is most likely to replace steels. Application of aluminium or its alloys is also thought of as an appropriate replacement in
defence field, especially to enhance the easiness in mobility of combat vehicles while maintaining the same standard as that of conventional
armour grade steels. Hence most of the investigations have been confined to aluminium or its alloys as base material and open an era of
developing the newer composite materials to address the major limitation, i.e. tribological properties. The surface composites can be fabricated
by incorporating the ceramic carbides like silicon carbide, carbides of transition metals and oxides of aluminium using surface modification
techniques, such as high energy laser melt treatment, high energy electron beam irradiation and thermal spray process which are based on fusion
route. These techniques yield the fusion related problems, such as interfacial reaction, pin holes, shrinkage cavities or voids and other casting
related defects, and pave the way to need of an efficient technique which must be based on solid state. Recently developed friction stir processing
technique was used in the present investigation for surface modification of AA7075 aluminum alloy, which is an alternative to steels. In the
present investigation, 160 mm sized boron carbide powder was procured and was reduced to 60 mm and 30 mm using high energy ball mill.
Subsequently these powders were used to fabricate the surface composites using friction stir processing.
Ballistic performance testing as per the military standard (JIS.0108.01) was carried out. In the present work, an analytical method of pre-
dicting the ballistic behavior of surface composites was developed. This method was based on energy balance, i.e., the initial energy of impact is
same as that of energy absorbed by multi layers. An attempt also has been made to validate the analytical results with the experimental findings.
Variation between the analytical and experimental results may be accounted due to the assumptions considering such as isotropic behavior of
target and shearing area of contact as cylindrical instead of conical interface As the analytical model yields the ballistic performance in the closer
proximity of experimentally obtained, it can be considered to be an approximation to evaluate the ballistic performance of targets.
Copyright © 2015, China Ordnance Society. Production and hosting by Elsevier B.V. All rights reserved.

Keywords: AA7075 aluminium alloy; Friction stir processing (FSP); Surface metal matrix composite (SMMC); Boron carbide (B4C); Ballistic performance; Target

1. Introduction

* Corresponding author. A shift in paradigm has been witnessed from heavier ma-
E-mail addresses: [email protected] (I. SUDHAKAR), gmreddy_ terial to lighter one in the automobile sectors. It is always a
[email protected] (G. MADHUSUDHAN REDDY), [email protected] challenge imposed by corporate average fuel economy
(K. SRINIVASA RAO). (CAFE)  standards so that the major share of material used in
Peer review under responsibility of China Ordnance Society.

http://dx.doi.org/10.1016/j.dt.2015.04.005
2214-9147/Copyright © 2015, China Ordnance Society. Production and hosting by Elsevier B.V. All rights reserved.

Please cite this article in press as: SUDHAKAR I, et al., Ballistic behavior of boron carbide reinforced AA7075 aluminium alloy using friction stir processing e
An experimental study and analytical approach, Defence Technology (2015), http://dx.doi.org/10.1016/j.dt.2015.04.005
+ MODEL

2 I. SUDHAKAR et al. / Defence Technology xx (2015) 1e7

vehicle outfit must be replaced by light metals, such as


aluminium, magnesium and titanium [1,2]. Being magnesium
self explosive and titanium costlier, aluminium and its alloys
paves a way to encroach several domains where ferrous alloys
have dominated [3e8]. The mechanical properties of steel,
such as high energy absorbing property, high strength, greater
notch toughness and high hardness, make as it a globally
accepted primary material in defense [9e14]. Even though
aluminium and its alloys possess all the requisite mechanical
properties to become a potential armour material, its lower
melting point, less strain rate sensitivity and poor tribological
property forbids it's usage [15,16]. Surface layering of Fig. 1. Schematic diagram of ballistic test [28].
monolithic materials with harder material and subsequent
existence of tougher material to dissipate the kinetic energy of tested with 7.62 mm-diameter lead projectiles located at 10 m
projectile have been found to be the effective measures to- from the projectile exit region. The striking velocity of the
wards improving the penetration resistance of targets in de- projectile was measured using infrared light emitting diode
fense [17e20]. Several methods, such as plasma spray, photo-voltaic cells by measuring the time interval between the
chemical vapour deposition (CVD), physical vapor deposition interceptions caused by the projectile running across two
(PVD), electron beam welding(EBM), plasma transferred arc transverse beams at fixed distance a part. The probes were
welding (PTAW), laser surfacing, are generally used to pro- placed at 6 m and 8 m from the nozzle of the gun barrel. The
duce surface composite [21e26] by changing the surface first probe activates the timer and the second probe de-
morphology of substrate. These surface modification tech- activates it. Few numbers of preliminary experiments were
niques are based on fusion technique and not free from some performed, and the adjustments were made to obtain the
limitations like poor interfacial bond integrity, casting defects required impact velocity of the projectile onto the target. The
such as pores, and shrinkages voids. Hence a newly developed velocity of projectile was measured to be 830 ± 10 m/s.
solid state process, i.e. friction stir process [27], has been After ballistic test, the targets were subjected to macro-
chosen for fabricating the targets. Keeping the above facts in scopic examination to evaluate the ballistic performance by
view, high strength AA 7075 aluminium alloy has been chosen considering the perforation depth of projectile. In the present
as substrate for fabricating the surface composites or targets by work, the depth of penetration was calculated analytically by
incorporating boron carbide using friction surfacing. the conservation of energy balance, i.e., kinetic energy
possessed by projectile during impact is equal to energy spent
2. Material and methods in multi-layers. Shearing area, i.e., outer surface area of pro-
jectile, was taken into consideration for the analytical findings.
Base metal or AA7075-T6 aluminium alloy, i.e., substrate Shearing strength was taken as half of the ultimate tensile
having chemical composition (% of weight), consists of 5.6% strength. Ultimate tensile strength was estimated from the
zinc, 2.5%magnesium, 1.6%copper, 0.23% chromium, 0.3% equivalent conversion of Vicker hardness values obtained for
manganese, 0.2%titanium, 0.4%silicon and 0.5% iron, and the the targets. Schematic diagram is represented to find the depth
remaining is aluminium. The initial thickness of base metal of penetration into the target using an analytical approach, as
was 40 mm. 160 mm-sized boron carbide powder of 99.92% shown in Fig. 2. For better illustration, a model calculation is
purity was received and reduced to 60 mm and 30 mm using given below.
high energy ball mill, and subsequently these powders were
ð½Þ mass of projectile  velocity2
used to fabricate the surface composites or targets with 40 mm
in thickness using friction stir processing. ¼ Work done against first layer
Trial runs were made by varying the rotational speeds and þ work done against supporting layer:
feed rate to obtain the sound surface composites on substrate
using friction stir processing. Surface composites, i.e., targets,
were fabricated by reinforcing 160 mm, 60 mm and 30 mm-
sized boron carbide powders in two stages using a flat tool and
tool with pin. The metallographic samples were prepared from
both base and targets using Kellar reagent (95% H2O, 2.5%
HNO3, 1.5% HCl, and 1% HF) as etchant to understand the
particle distribution of boron carbide (B4C). Hardness of tar-
gets was measured using Vicker hardness tester under a load
of 0.3 kgf. Ballistic testing was carried out on targets in a
ballistic testing tunnel as per the military standard
(JIS.0108.01) and the standardized testing conditions. An
experimental setup is shown in Fig. 1. The target plates were Fig. 2. Model for analytical assessment of ballistic performance.

Please cite this article in press as: SUDHAKAR I, et al., Ballistic behavior of boron carbide reinforced AA7075 aluminium alloy using friction stir processing e
An experimental study and analytical approach, Defence Technology (2015), http://dx.doi.org/10.1016/j.dt.2015.04.005
+ MODEL

I. SUDHAKAR et al. / Defence Technology xx (2015) 1e7 3

Fig. 3. Friction stir processed surface with the cracks along the tool travel and the voids along the transverse section.

Fig. 4. Sound surface composite fabricated using FSP.

ð1=2ÞmV 2 ¼ pðDÞðlÞðshear stress of first layerÞðtÞ x ¼ 0:03914 m; or x ¼ 39:14 mm


þ pðDÞðlÞðshear stress of supporting layerÞðxÞ Depth of penetration ð39:14 þ 3Þ ¼ 42:14 mm
where m ¼ 0.0506 kg, V ¼ 830 m/s, D (diameter of A similar approach was adopted for evaluating the pene-
projectile) ¼ 7.62 mm, length of projectile (l ) ¼ 98 mm, tration depth in SMMC B4C-60 mm and SMMC-30mm, and the
thickness of first layer (t) ¼ 3 mm, i.e., fabricated surface penetration depth of 40.5mmand 38.17 mm were obtained,
composite layer thickness which is maintained constant for all respectively. An attempt was made to introduce an empirical
targets, and x ¼ traveling distance of projectile in the sup- constant such that analytical obtained depth of penetration
porting layer. approach experimental one.
Empirical constant was calculated by taking the average of
Shear stress of first layer ¼ ultimate tensile stress of layer 2 =
ratio between experimental depth to analytically obtained
values and found to be 0.707.

Shear stress of supporting layer ¼ ultimate tensile stress of supporting layer 2 =

3. Results and discussion


Ultimate tensile stress for target is calculated empirically
from hardness conversions tables. Hence, monolithic AA 7075 3.1. Fabrication of surface composite using friction stir
aluminium alloy having ultimate tensile strength of 339 MPa processing
acts as a supporting layer for other fabricated targets.
Ultimate tensile strength of SMMC B4C-160 mm is Higher rotational speed (greater than 1200 rpm) of rotating
527 Mpa. tool resulted in intense plastic deformation which led to the
The above expressions are used to calculate the penetration surface cracks along the tool traveling direction and the voids
depth of projectile intoTarget-1: in the transverse section of processed surface composite, as
shown in Fig. 3. Similarly, the rotational speed less than
 
527 750 rpm during friction stir processing caused an improper
ð½Þ0:0506  830 ¼ pð7:62Þð98Þ
2
ð0:003Þ mixing of B4C particles and yielded the earlier results.
2
  Higher transitional feed rate as wells as larger plunging
339
þ pð7:62Þð98Þ ðxÞ: speed during fabrication of surface composite using friction
2 stir processing resulted in tool breakage. It was observed that
the threaded portion of tool used for mixing the hard boron
17429:17 ¼ 1855:33 þ 397809:72x carbide particles during the second stage of friction stir

Please cite this article in press as: SUDHAKAR I, et al., Ballistic behavior of boron carbide reinforced AA7075 aluminium alloy using friction stir processing e
An experimental study and analytical approach, Defence Technology (2015), http://dx.doi.org/10.1016/j.dt.2015.04.005
+ MODEL

4 I. SUDHAKAR et al. / Defence Technology xx (2015) 1e7

Fig. 6. SEM micrograph. (a) SEM image of stir zone; (b) transverse section of
SMMC.
Fig. 5. Microstructure of the elongated grains. (a) Microstructure of base; (b)
surface composite.
boron carbide particles in the stir zone is significantly smaller
than received powder, which can be witnessed from Fig. 5(b).
processing was worn. Tool wear may be attributed to the Friction stir processing involves stirring action of pin with
presence of harder boron carbide particles which are respon- high rotational speed which generates frictional heat between
sible for wearing out the threaded portion of pin. After a tool and workpiece. This heat is responsible for softening the
number of trials for fabrication of surface composite, the material around the tool and helps in the distribution of rein-
friction stir processing parameters, such as tool rotational forcement particles to form a surface metal matrix composite.
speed of 925 rpme1000 rpm, a longitudinal feed rate of The effect of stirring on the fabrication of surface composite
50 mm/min and plunging speed of 30 mm/min, resulted in the and the distribution of boron carbide particles can be easily
formation of surface metal matrix composites without voids witnessed from SEM micrograph, as shown in Fig. 6(a). The
along the transverse section and surface cracks, as shown in degree of boron carbide particle distribution is dependent on
Fig. 4. It was also found that the tracks of rotational tool have pin height of processing tool. A pin with 3 mm in height was
specific pattern in the form of onion rings, as shown in Fig. 4.
Tool movement from right to left about horizontal axis was
maintained to obtain a defect-free surface composite during
surfacing.

3.2. Microstructure

Base metal consists of the elongated grains along the roll-


ing direction as evident from the optical micrograph shown in
Fig. 5(a).
During fabrication of surface metal matrix composite
(SMMC), the stirring action of tool results in changing the
dimensions of boron carbide particles either by rounding off
the sharp corners or breaking off the sharp edges and leads to
finer particle size. Hence, the estimated average size of the Fig. 7. Hardness of targets.

Please cite this article in press as: SUDHAKAR I, et al., Ballistic behavior of boron carbide reinforced AA7075 aluminium alloy using friction stir processing e
An experimental study and analytical approach, Defence Technology (2015), http://dx.doi.org/10.1016/j.dt.2015.04.005
+ MODEL

I. SUDHAKAR et al. / Defence Technology xx (2015) 1e7 5

Fig. 8. Front view of targets after ballistic test. (a) Base; (b) B4C-160 mm; (c) B4C-60 mm(d); B4C-30 mm.

used during the fabrication of surface composite, which is outcomes obtained during ballistic test, the depth of penetra-
identical with the magnitude of interface, as shown in tion (DOP) of targets was taken to evaluate the ballistic per-
Fig. 6(b). formance. Fig. 8 shows the macro photographs of targets after
From Fig. 6(a) and (b), it can be interpreted that FSP causes ballistic test under the different conditions of targets.
the homogeneous dispersion of boron carbide particles in the From Fig. 8, it can be interpreted that the base, i.e.,
stir zone and leads to the formation of defect-free interface and AA7075 aluminium alloy, exhibited the fragments, the pro-
adherent B4C particles in substrate. jectile was almost penetrated, and other targets exhibited
petaling failure. It can also be noticed that the craters are in the
3.3. Hardness testing base. From the front ends of targets in other cases, the pres-
ence of projectile inside the targets can be witnessed.
The hardness data of base metal and friction stir processed From Fig. 9, it can be inferred that the degree of penetration
alloy under different conditions is presented in Fig. 7. There is is more significant in the base metal compared to other targets
significant improvement in hardness by the addition of boron in all the trials for evaluating ballistic performance of base
carbide particles into substrate during friction stir processing. metal, It was found that the base metal was perforated in all
The stirring action of rotating tool results in the mechanical the trials for evaluating ballistic performance of base metal,
rupture of boron carbide, and a large number of high angle while in very few trials (about 5%) the projectile caused
boundaries are produced. These high angle boundaries impede damage to a depth of 80% of base metal, as shown in Fig. 9(a).
the free movement of dislocations and enhance the strength As the armour materials are fabricated for high safety,
and hardness. Fine dispersion of carbide particles in the AA7075 aluminium alloy cannot be suitable for fabrication of
AA7075 aluminium alloy matrix hinders the free movement of combat vehicles. Depth of penetration of projectile was
dislocations and enhances the hardness of surface composites, measured from the cross section of each target and presented
which is well validated with Orwan mechanism [29]. in Table 1. Similarly, the analytic results were also presented
in Table 1. Fig. 10 shows the graphical representation for
3.4. Ballistic behavior better illustration.
From Table 1 and Fig. 10, it can be witnessed that the depth
Lead projectile was used to evaluate the ballistic perfor- of penetration is found to be very close to each other and are
mance of fabricated target under different conditions. Damage, well validated. However, the variation in the results may be
in the form of a perforation hole with measurable depth of attributed to the assumption that the targets are considered to
penetration, width and cracks around the hole, was noticed in be isotropic and the perfect shearing, i.e. between projectile
the targets after ballistic test. Among these noticeable surface and inner portion of target is taken into account during

Please cite this article in press as: SUDHAKAR I, et al., Ballistic behavior of boron carbide reinforced AA7075 aluminium alloy using friction stir processing e
An experimental study and analytical approach, Defence Technology (2015), http://dx.doi.org/10.1016/j.dt.2015.04.005
+ MODEL

6 I. SUDHAKAR et al. / Defence Technology xx (2015) 1e7

Fig. 9. Cross section of targets after ballistic test. (a) Base (b) B4C-160 mm; (c) B4C-60 mm; (d) B4C-30 mm.

Table 1 analytical calculation. Shear strength is considered to be half


Analytical and experimental results of ballistic test. that of ultimate tensile stress, and the ultimate tensile strength
Target DOP DOP DOP by is obtained from conversion of Vickers's hardness data.
experimental/mm analytical/mm incorporating It can be observed that SMMCs/targets have exhibited
constant/mm
better ballistic performances compared to the base metal. It
Base 34 43.8 30.8 may be attributed to the frictional characteristic of SMMC and
B4C-160/mm 30 42.15 29.67
B4C-60/mm 29 40.5 28.51
projectile. During travel of projectile inside the armor, a strong
B4C-30/mm 24 38.17 26.87 friction between the contact pair causes damage to the surface
of the projectile tip due to abrasive action of harder (B4C)
particles on SMMC and leads to the macro deformation and
breaking of the projectile tip. Hence, the depth of penetration
of projectile into the armour material/targets (SMMC) is
reduced. The fine grade target, i.e., SMMC B4C-30 mm, con-
sists of more number of boron carbide particles compared to
the coarse grade target. The intactness of boron carbide par-
ticles, which is the measure of surface hardness, of the targets
favors abrasive action damages the projectile surface and re-
duces the travel of projectile inside the armour material.
From the above, it can be established that the friction stir
processed AA7075 alloy plates/targets successfully stopped
the projectile whereas the uncoated base metal was perforated
completely. The targets with finer grade boron carbide parti-
cles have better ballistic performance compared to the base
Fig. 10. Analytic and experimental values of ballistic test. metal. The observed result is in agreement with hardness data.

Please cite this article in press as: SUDHAKAR I, et al., Ballistic behavior of boron carbide reinforced AA7075 aluminium alloy using friction stir processing e
An experimental study and analytical approach, Defence Technology (2015), http://dx.doi.org/10.1016/j.dt.2015.04.005
+ MODEL

I. SUDHAKAR et al. / Defence Technology xx (2015) 1e7 7

Hence it can be concluded that the friction surfacing can be [5] Sako R. Nihon Parkerizing technical report, vol. 9; 1996. p. 28e37.
successfully used to deposit the hard boron carbide powders [6] Mizuno H, Uchiyama K, Sato S, et al. Nihon Parkerizing technical
report, vol. 12; 2000. p. 53e9.
on the surface of armour grade AA7075 aluminium alloy and [7] Ishii H, Furuyama O, Tanaka S. Metal Finish 1993;91(4):7e13.
enhance the ballistic resistance of selected AA7075 [8] Kr Aune T, Westengen H. Autom Eng 1995;103(8):87e93.
aluminium alloy. [9] Ade F. Ballistic qualification of armour steel weldments. Weld J
1991;70:53e4.
4. Conclusions [10] Madhusudhan Reddy G, Mohandas T. Ballistic performance of high
strength low-alloy steel weldments. J Material Process Technol
1996;57:23e30.
1) Friction stir processing has uniformly distributed the boron [11] Madhusudhan Reddy G, Mohandas T, Tagore GRN. Weldability studies
carbide particles in AA7075 aluminum alloy matrix and of high strength low- alloy steel using austenitic fillers. J Mater Process
enhances its surface hardness compared to that of base. Technol 1995;49:213e28.
2) Increase in ballistic efficiency may be attributed to the [12] Tore Borvika, Clausena Arild H, Odd Sture Hopperstada,
Magnus Langsetha. Perforation of AA5083-H116 aluminium plates with
frictional characteristics of armour surface and projectile, conical-nose steel projectiles d experimental study. Int J Impact Eng
which favors the damage to the surface of the projectile tip 2004;30:367e84.
by abrasive action of harder (B4C) particles on the target [13] Backman EM, Goldsmith W. The mechanics of penetration of projectiles
and leads to the macro-deformation and breaking of the into targets. Int J Eng Sci 1978;161:99e103.
projectile tip. [14] Gooch WA, Burkins MS, Squillacioti RJ. Ballistic testing of commercial
aluminum alloys and alternate processing techniques to increase the
3) Analytic approach to estimate the penetration depth has availability of aluminum armor. In: 23rd International symposium on
been done for the first time and the results obtained are ballistics. Tarragona, Spain 16e20 April; 2007.
closer to the experimental finding. Variation in the findings [15] Backman EM, Goldsmith W. The mechanics of penetration of projectiles
may be accounted to the assumptions of considering the into targets. Int J Eng Sci 1978;vol. 16:1e9.
isotropic nature of target, and the shear strength is ob- [16] Corbett GG, Reid SR, Johnson W. Impact loading of plates and shells by
free-flying projectiles: a review. Int J Impact Eng 1996;vol. 18:141e9.
tained on hardness data. [17] Johnson W. Impact strength of materials. London: Edward Arnold; 1972.
4) For the first time, the present work demonstrated suc- [18] Ozsahin Evren, Tolun Suleyman. On the comparison of the ballistic
cessfully that the friction stir processing route is an response of coated aluminum plates. Mater Des 2010;vol. 31:3188e93.
effective strategy for enhancement of ballistic perfor- [19] Radin J, Goldsmith W. Normal projectile penetration and perforation of
mance of AA7075 aluminium alloy which finds wider layered targets. J Impact Eng 1988;vol. 7:229e38.
[20] Ozsahin E, Tolun S. Influence of layer sequencing on ballistic resistance
range of defence applications. of polyethylene supported AA 7075 T651 plates. J Istanbul Technol Univ
2009;8(2):72e80.
[21] Pobolo IL, Shipko A. A first and second generation electron beam sur-
Acknowledgement face engineering. Mater Manuf Process 1999;14:321e30.
[22] Yamana H, Shimizu S. Alloying of metals and carbides on an aluminium
surface by electron beam welding. Weld Int 1998;12:160e4.
The authors would like to thank Director, Defence Metal- [23] Vreeling JA, Ocelik J, Pei YT, Van Agterveld DT, Hosson TM. Laser
lurgical Research Laboratory and Hyderabad, India for his melt injection of ceramic particles in metals. Acta Mater 2000;48:4225.
continued encouragement and permission to publish this work. [24] Abboud JH, West DRF. Microstructure of titanium injected with SiC
Financial assistance from Armament research board, New particles by laser processing. J Mater Sci Lett 1991;10:1149.
Delhi, India is gratefully acknowledged. [25] Pei YT, Ocelik V, Hosson JT. Futuristic materials. In: Jha BB, Galgai RK,
Misra Vibhuti N, editors. Bhubaneswar: Allied Publishers, Indian Insti-
tute of Minerals Engineers; 2004 [chapter].
References [26] Thomas WM, Nicholas ED, Needham JC, Murch MG, Templesmith P,
Dawes CJ. Patent application No.9125978.8, Dec. 1991.
[1] Statistical Data of Automotive Materials Committee, Japanese Society of [27] Mishra RS, Ma ZY. Mater Sci Eng R 2005;R50:1e78.
Automobile Industry, May 1997. [28] Sudhakar I, Madhu V, Madhusudhan Reddy G, Srinivasa Rao K. Defence
[2] North American Passenger Car and Light Truck. 1999. Technol 2014:1e8.
[3] Aluminum Content Report. Ducker Research Co. TM30. Advanced [29] Madhusudhan Reddy G, Sambasiva Rao A, Srinivasa Rao K. Friction stir
materials and processing. Southfield Metals 1999(3):10e7. processing for enhancement of wear resistance of ZM21 magnesium
[4] Takeuchi K, Isobe Y, Toyama T. J Japan Inst Light Metals alloy. Trans Indian Inst Metal 2012:28e35.
1998;48(9):446e58.

Please cite this article in press as: SUDHAKAR I, et al., Ballistic behavior of boron carbide reinforced AA7075 aluminium alloy using friction stir processing e
An experimental study and analytical approach, Defence Technology (2015), http://dx.doi.org/10.1016/j.dt.2015.04.005

You might also like