International Journal of Rock Mechanics and Mining Sciences 100 (2017) 250–254

Contents lists available at ScienceDirect

International Journal of
Rock Mechanics and Mining Sciences
journal homepage: www.elsevier.com/locate/ijrmms

A numerical study of the impact of short delays on rock fragmentation MARK

a,⁎ b a a
Changping Yi , Jonny Sjöberg , Daniel Johansson , Nikolaos Petropoulos
a
Division of Mining and Geotechnical Engineering, Luleå University of Technology, Sweden
b
Itasca Consultants AB, Sweden

A R T I C L E I N F O

Keywords:
Numerical simulation
Blasting
Short delays
Fragmentation

1. Introduction
Blasting techniques are widely employed in mining in order to
fragment rock mass into smaller pieces to facilitate subsequent handling
(mucking, haulage, crushing, etc.). Short delay blasting techniques can
improve fragmentation compared to simultaneous initiation, but the
optimum delay time for blasting is still under discussion. The optimum
delay time to improve fragmentation has been studied by e.g. Tatsuya
et al.,1 Aldas et al.,2 Shi and Chen3 and Petropoulos et al.,4 but different
conclusions were obtained.
With the application of electronic detonators and with short delay
times, a hypothesis of achieving improved fragmentation through stress
wave interaction has been proposed by Rossmanith et al.5,6 In these
papers, a model was proposed to describe the stress wave interaction
between adjacent boreholes with Lagrange diagrams, which reveals
how a positive effect of the interaction of the stress waves could be
achieved with the assumption of an infinitely long charge length. With
the inspiration of Rossmanith and co-workers, Vanbrabant and Espi-
nosa7 stated that the delay times to match an overlap of the negative
tail of the particle velocity can improve fragmentation. They conducted
a series of field tests and claimed that the average fragmentation im-
proved by nearly 50%. Chiappetta8 also claimed that the very short
delays between holes, such as 2 ms, help to improve blast perfor-
mances. However, there are different opinions. Blair9 stated that the
delay time and initiation accuracy are not typical governing factors for
blast performances. Johansson and Ouchterlony10 investigated the in-
fluence of delay time on the fragmentation with a series of small-scale
tests. Their results showed no distinct differences in fragmentation
when there were shockwave interactions compared to no shockwave
interaction. The investigation of Katsabanis et al.11 indicated that se-
lecting a very short delay time for fragmentation optimization is
questionable. Schill and Sjöberg12 studied the influence of delay times
on the blasting effect in a two-hole model with the LS-DYNA13 code and
concluded that there was an effect of interacting stress waves. However,
this effect was local around the interaction plane, implying that very
short delay will not generate a dramatic increase in fragmentation
contrary to what was proposed by Rossmanith et al.5,6 The results of
Schill and Sjöberg12 also indicated that longer delay times (in which the
stress wave would have passed the neighboring boreholes) also resulted
in improved fragmentation. The quantitative results of stress wave in-
teraction between two adjacent boreholes were analytically and nu-
merically investigated by Yi et al.14 The analytical model based on the
assumptions used by Rossmanith et al.5,6 was compared with a nu-
merical model in LS-DYNA and the results were not consistent.
In the present paper, a four-hole model was built to study the pos-
sible effect of overlapping negative tails by using the 3D finite element
method. A method was presented to form fragments based on finite
element modeling results and damage concepts. An approach was
proposed to evaluate blast-induced fragmentation based on numerical
results.
2. Model and materials
A four-borehole model was constructed to model the field tests. The
model geometry and the sizes are shown in Fig. 1. The borehole diameter
is 310 mm. The depth of the borehole is 16 m, the length of subdrilling is
1 m and the length of stemming is 5 m. The green part in the model
geometry was selected to be evaluated after blasting. The model is dis-
cretized with hexahedral elements. The element size of the green part is 6
× 6 × 6 cm while the element size of the yellow part is 12 × 12 ×
12 cm. The green part and the yellow part are connected with transition
elements. The total number of elements is approximately 23 million.

⁎
Corresponding author.
E-mail address: changping.yi@ltu.se (C. Yi).

http://dx.doi.org/10.1016/j.ijrmms.2017.10.026
Received 2 May 2016; Received in revised form 14 March 2017; Accepted 23 October 2017
1365-1609/ © 2017 Elsevier Ltd. All rights reserved.
C. Yi et al. International Journal of Rock Mechanics and Mining Sciences 100 (2017) 250–254

Fig. 3. Stress limit surfaces of the RHT model, after Schill.16

Fig. 1. Geometry of the four-borehole model. (For interpretation of the references to color Table 1
in this figure, the reader is referred to the web version of this article). RHT model parameters for rock mass in Lagrangian elements.

Density 2627 kg/m3 Ref. compressive stain rate 3.0e8

The explosives and the stemming material are modeled with Shear Modulus 18.6 GPa Ref. tensile strain rate 3.0e9
Eulerian elements. The rock surrounding the borehole is also modeled Pore crush B0 1.22 Failure tensile strain rate 3.0e22
Pore crush B1 1.22 Failure compressive strain rate 3.0e22
with Eulerian elements to accommodate for the large deformation in
Bulk Modulus T1 40 GPa Compressive strain rate 0.032
that region. The other rock parts are modeled with Lagrangian ele- dependence exponent
ments. The Eulerian elements are merged to the Lagrangian mesh. The Bulk Modulus T2 0 Tensile strain rate dependence 0.036
radius of the interface between the Eulerian elements and the exponent
Bulk Modulus A1 40 GPa Volumetric plastic strain 0.001
Lagrangian elements is 0.5 m. The initiation point is 1 m above the
fraction in tension
bottom of the blast holes. In order to model an infinite domain, non- Bulk Modulus A2 0 Compressive yield strength 200 MPa
reflecting boundaries are used on surfaces which are connected to Bulk Modulus A3 0 Tensile yield strength 7 MPa
continuing rock material. The top and front surfaces are modeled as free Failure surface A 2.618 Damage parameter D1 0.04
faces (Fig. 2). Failure surface N 0.7985 Damage parameter D2 1.0
Shear strength 36 MPa Minimum damaged residual 0.01
The Riedel-Hiermaier-Thoma (RHT) material model15 which is an
strain
advanced plasticity model for brittle materials such as concrete and Uniaxial tensile 10 MPa Residual surface parameter AF 0.873
rock was employed to describe the dynamic response of rock mass in strength
Lagrangian elements. This material model involves three limit surfaces Lode Angle Q0 0.567 Residual surface parameter AN 0.559
which describe the strength of the material, see Fig. 3. The first surface Lode Angle B 0.0105 Grunnisen Gamma 0
Compaction pressure 6 GPa Crush pressure 133 MPa
is the yield surface which is limited by a cap surface. Beyond this sur- Initial porosity 1.0 Porosity exponent 3
face the material starts to deform plastically with a linear hardening
description. When the material reaches the failure surface, the damage
of the material starts to evolve until the damage is equal to one. The were calibrated by Hansson based on the cylinder test.19
∆εp
damage level in this model is defined as D = ∑ f ,where ∆ε p is the The stemming material and the rock mass in Eulerian elements are
ε
accumulated plastic strain and ε f is the failure strain. The parameters of modeled by a soil material model (*MAT_SOIL_CONCRETE) since the
RHT material used in this paper are from Schill16 in which these RHT material model does not support the Eulerian solution technique.
parameters were calibrated based on the material tests presented by This model is a perfectly plastic, pressure dependent yield function.13 It
Haimson and Chang.17 The RHT material parameters are shown in also includes fracture and a residual strength surface where the material
Table 1. loses its ability to carry tension. The soil material parameters for rocks
The E682-b emulsion explosive was used and it was modeled with and stemming calibrated by Schill16 are given in Table 2.
an explosive material model in LS-DYNA and with the Jones-Wilkins-
Lee (JWL) equation of state.18
3. Fragmentation evaluation
w ⎞ −R1 V w ⎞ −R2 V wE
p = A ⎛1 −
⎜ e ⎟ + B ⎛1 − ⎜ e + ⎟ Fragmentation is one of the important indicators to evaluate blast
⎝ R1 V ⎠ ⎝ R2 V ⎠ V (1)
performance. It is complicated to directly evaluate fragmentation based
on finite element method. In this paper, after the calculation, the ele-
where p is the pressure, A, B, R1, R2 and w are constants; V and E are the
ments with damage level above 0.6 were blanked out to form cracks in
specific volume and the internal energy respectively. The density of
the rock mass. An instance for the case of 3 ms delay time is shown in
E682-b emulsion explosive is 1180 kg/m3. The velocity of detonation is
Fig. 4. Fig. 4(a) shows the damage distribution in the rock mass after
5866 m/s. The Chapman-Jouguet pressure (PCJ) of E682-b emulsion
blasting. Fig. 4(b) shows the overall crack pattern after the elements
explosive is 10.06 GPa. For E682-b emulsion explosive, A =
with damage level above 0.6 are blanked out. Hence, the rock mass is
285.73 GPa, B = 6.715 GPa, R1 = 4.933, R2 = 1.962, w = 0.52, The
separated into smaller fragments by these cracks.
detonation energy per unit volume E0 = 3.176 kJ/cc. These parameters
If the dimensions of these fragments can be determined, the frag-
ment size distribution can be evaluated. It is not an easy task in 3D, but
it is straightforward in 2D and a routine was implemented in LS-
PREPOST20 code which is an advanced interactive program for

Table 2
Parameters for rock mass and stemming in Eulerian elements.

Parameters Rock mass Stemming material

Density (kg/m3) 2770 1650

Shear Modulus (GPa) 26.1 10
Bulk Modulus (GPa) 37.6 3.6
Pressure Cutoff (MPa) 2.67 0.01
Fig. 2. The boundary conditions.

251
C. Yi et al.
Fig. 4. Overall damage distribution and crack pattern for the 3 ms delay time case.
preparing input data for LS-DYNA and processing the results from LS-
DYNA analyses and in which fragments delineated by cracks were
identified, and the area of each such fragment was determined. A
number of 2D cuts were selected to be evaluated and the area of each
fragment at the cut calculated with the routine. After the fragment area
was calculated, some area sizes resembling the sieve mesh sizes were
defined to obtain intervals for different fragment areas. Then the ex-
tended Swebrec function21 was employed to fit the fragment area dis-
tribution. The function is
P (x ) = 1/{1 + a [ ln(x max / x )/ ln(x max / x50)] b+ (1 − a)[(x max / x − 1)/(x max / x50 − 1)]c }
(2)
where P(x) is a cumulative distribution function and may take any
value in the range 0–1, if x is less than xmax, where xmax is the maximum
fragment size and x 50 is the median fragment size, and a, b and c are
constants. For instance, the fragmentation of one cross-section of the
model in Fig. 4 is shown in Fig. 5(a). The fragment area distribution and
the fitting curve are shown in Fig. 5(b).
Using this method it is possible to study the accumulated area for
different fragment areas and compare the fragmentation between dif-
ferent cross-sections and simulations. The accumulated area plot is
constructed to resemble a mass passing (or “sieve curves”) which are
commonly used in the blasting community to evaluate fragmentation.
Fig. 5. (a) Fragmentation at the chosen cross-section; (b) Fragment area distribution and
the fitting curve with Swebrec function.
252
International Journal of Rock Mechanics and Mining Sciences 100 (2017) 250–254
Fig. 6. Vertical cuts used in the results presentation.
Fig. 7. Overall crack pattern for the case of 0 ms delay.
4. Model results and discussion
Four inter-hole delay cases were modeled in the paper. The inter-
hole delay times were 0 ms, 1 ms, 3 ms and 6 ms respectively. When the
delay time is 6 ms, there is no interaction of stress waves from adjacent
boreholes according to the stress wave speed in the rock mass. The
initiation sequence is from right to left. The simulations of the 0 ms
delay case and the 1 ms delay case were run to 15 ms. The simulation of
the 3 ms delay case was run to 18 ms and the simulation of the 6 ms
delay case was run to 25 ms. Six cuts (cross-sections) were selected to
be evaluated (Fig. 6). Fig. 7 presents the overall crack pattern of the
case of 0 ms delay time. Because the delay time is 0 ms, it appears that
the results are symmetric. Fig. 8 presents a horizontal cut just above the
explosives. The results show that the fragmentation is fine around the
boreholes and near the free face. This is due to the high pressure and
quick unloading process of the blast load around the boreholes and the
effect of the reflected tensile stress waves from the free face. The radial
cracks and the hoop cracks are also obvious, which is consistent with
the experimental results.22 The fragment area distribution curves for
the 0 ms delay case show that the fragmentation at the cuts near the
borehole is finer than the cuts at the middle of two adjacent boreholes
(Fig. 9).
According to the concept of Vanbrabant and Espinosa,8 cuts X22
and X32 are the most interesting cuts because they are certain that the
stress waves meet and interact in the 0 ms delay case. If the negative
tails of the stress waves overlap at the area between the second bore-
hole and the fourth borehole with very short delay time such as 1 ms
and 3 ms, the fragmentation at Cut X22 and Cut X32 should be finer
than the case that there is no stress wave interaction because of the long
delay time such as 6 ms. For cut X22, the finest fragmentation is found
in the simulation with 6 ms delay (Fig. 10). There is no interaction of
stress waves when the delay time is 6 ms, which indicates that there are
other factors rather than stress wave interaction that affect the frag-
mentation the most.
Fig. 8. Horizontal cut of fragmented region.
C. Yi et al.
Fig. 9. Accumulated area plot for the case of 0 ms delay time.
Fig. 10. Accumulated area plot for Cut X22.
Fig. 11. Accumulated area plot for cut X32.
The accumulated area plot for Cut X32 is shown in Fig. 11. The
finest fragmentation is found for the simulation with the 3 ms delay
time case, and the 6 ms delay time case yields the second finest frag-
mentation.
Figs. 10 and 11 show that 3 ms and 6 ms delay time can improve
fragmentation compared to simultaneous initiation, but very short
delay time such as 3 ms and 1 ms cannot ensure finer fragmentation
than the long delay time such as 6 ms. The fragments at Cut X32 with
different delay times are shown in Fig. 12.
5. Field test results
To test the hypothesis that very short delay time could improve
fragmentation, a series of field tests have been performed at the Aitik
open pit mine by Petropoulos et al.4 A total number of six benches were
assigned for trials with different inter-hole delay times. Two of the
benches were blasted with pyrotechnic caps with 42 ms delay time and
were used as references for further comparisons. Two benches were
blasted with 1 ms of inter-hole delay time by use of electronic deto-
nators. Two other benches were also blasted by electronic detonators,
but with 3 ms and 6 ms of inter-hole delay time respectively. One of the
test benches is shown in Fig. 13. Image analysis was used to analyze the
fragmentation of the rock after blasts. A series of images were shot from
253
International Journal of Rock Mechanics and Mining Sciences 100 (2017) 250–254
0ms delay 1ms delay 3ms delay 6ms delay
Fig. 12. Fragments at X32 cut of different cases.
Fig. 13. Field trial bench (Taken by Petropoulos).
Fig. 14. Median and oversize vs. Delay time (after Petropoulos et al.5).
trucks carrying the ore and were later analyzed to obtain the frag-
mentation for each bench and the results are shown in Fig. 14.
The tests showed that the inter-hole delay time of 3 ms resulted in
the finest fragmentation among all benches; all examined values i.e.
x50, x80 and xmax showed improvements upon other benches. Compared
to reference trials, the 3 ms delay yielded 10% finer fragmentation. The
improvement was not as significant as stated by Rossmanith5 and
Vanbrabant and Espinosa.7 However, the crushing energy of the ore
from this trial was the highest among all.
Two trials with 1 ms inter-hole delay time did not result in any
significant variation compared to reference benches. The difference in
x50 values were ignorable, the same is true for crushing energy of the
mentioned trials.
The bench with inter-hole delay time of 6 ms resulted in the lowest
crushing energy among other trials. However, the bench gave more
boulders and coarser fragmentation compared to reference benches.
It should be noted that different specific charges were used in these
field trials. For instance, the specific charge is 1.07 kg/m3 for the 3 ms
delay case while it is 1.0 kg/m3 for the 6 ms delay case. Authors4
concluded that specific charge was found dominant over the delay time
regarding fragmentation.
C. Yi et al.
6. Conclusion
A method to form cracks in rock mass due to blasting based on
damage concept and an approach to evaluate the fragment size dis-
tribution based on finite element method were presented in this paper.
These methods were used to investigate the impact of short delays on
rock fragmentation. Judging from the numerical results, it may be
stated that the problem is more complex than described by Rossmanith
et al.5,6 The fact that fragmentation increased for longer delay times
contradicts the previous theory, which suggests that factors other than
stress wave interactions govern the fragmentation. The cause of this is
not yet known, and need further study. The results of the field trials
cannot definitely support the hypothesis proposed by Rossmanith
et al.5,6 either. The following conclusions may be drawn.
The approach for evaluation of fragmentation provides a tool for
comparing modeling results with respect to expected fragmentation.
The numerical results cannot support the hypothesis that very short
delay time could improve fragmentation.
The field trials lead to different results from numerical simulation.
The results of field trials and numerical simulations show the need for
further investigation of the short delay times.
Acknowledgements
The authors wish to thank Vinnova (the Swedish Governmental
Agency for Innovation Systems) (2008-00862) for funding support.
LKAB, Boliden Mineral AB and Luleå University of Technology are ac-
knowledged for supplementary funding. The authors would also like to
thank Mikael Schill at DynaMore Nordic AB who also was involved in
this project. Professor Emeritus Finn Ouchterlony, formerly at Luleå
University of Technology, is acknowledeged for initiating this project.
References
1. Tatsuya H, Gento M, Kou SQ. Optimum delay interval design in delay blasting. Int J
Blast Frag. 2000;4:139–148.
2. Aldas GGU, Bilgin HA, Esen S. Timing simulation for the selection of optimum delay
time. In: Proceedings of the 17th International Mining Congress And Exhibition of
Turkey, Ankara; 19–22 June 2001:317-322.
3. Shi X, Chen S. Delay time optimization in blasting operations for mitigating the vi-
bration-effects on final pit walls' stability. Soil Dyn Earthq Eng. 2011;31:1154–1158.
254
International Journal of Rock Mechanics and Mining Sciences 100 (2017) 250–254
4. Petropoulos N, Beyglou A, Johansson D, Nyberg U, Novikov E. Fragmentation by
blasting through precise initiation: full scale trials at the Aitik copper mine. Blasting
Fragm. 2014;8:87–100.
5. Rossmanith HP. The use of Lagrange diagrams in precise initiation blasting. Part I:
two interacting blastholes. Int J Blast Frag. 2002;6:104–136.
6. Rossmanith HP, Kouzniak N. Supersonic detonation in rock mass: particle displace-
ments and velocity fields for single and multiple non-delayed and delayed detonating
blastholes. Int J Blast Frag. 2004;8:95–117.
7. Vanbrabant F, Espinosa A. Impact of short delays sequence on fragmentation by
means of electronic detonators: theoretical concepts and field validation. In:
Proceedings of the 8th International Symposium on Rock Fragmentation by Blasting.
Santiago; 7–11 May 2006:326–331.
8. Chiappetta F. Combining Electronic Detonators with Stem Charges and Air Decks;
2010. Available at: 〈http://www.iqpc.com/redForms.aspx?id=414254&sform_id=
473344〉. Accessed 10 August 2012.
9. Blair DP. Limitations of electronic delays for the control of blast vibration and
fragmentation. In: Proceedings of the 9th International Symposium on Rock
Fragmentation by Blasting. Granada; 13–17 August 2009:171–184.
10. Johansson D, Ouchterlony F. Shock wave interactions in rock blasting-the use of
short delays to improve fragmentation in model-scale. Rock Mech Rock Eng.
2013;46:1–18.
11. Katsabanis PD, Tawadrous A, Braun C, Kennedy C. Timing effects on the fragmen-
tation of small scale blocks of granodiorite. Int J Blast Frag. 2006;10:83–93.
12. Schill M, Sjöberg J. Finite element simulations of blasting and fragmentation with
precise initiation. In: Proceedings of the 12th International LS-DYNA User Conference.
Dearborn; 3–5 June 2012:1–10.
13. Hallquist J. LS-DYNA Keyword User’s Manual. Livermore Software Technology
Corporation, LSTC. Vol. 1 & 2; 2012.
14. Yi CP, Johansson D, Nyberg U, Beyglou A. Stress wave interaction between two
adjacent blast holes. Rock Mech Rock Eng. 2016;49:1803–1812.
15. Riedel W, Thoma K, Hiermaier S, Schmolinske E. Penetration of reinforced concrete
by BETA-B-500, numerical analysis using a new macroscopic concrete model for
hydrocodes. In: Proceedings of the 9th International Symposium on Interaction of the
Effects of Munitions with Structures. Berlin; 3–7 May 1999:315–322.
16. Schill M Finite Element Simulations of Blasting and the Effects of Precise Initiation on
Fragmentation. Swebrec Report, No. 2012:2; 2012. ISSN 1653-5006.
17. Haimson B, Chang C. A new true triaxial cell for testing mechanical properties of
rock, and its use to determine rock strength and deformability of Westerly Granite.
Int J Rock Mech Min Sci. 2000;37:285–296.
18. Lee EL, Horning HC, Kury JW. A Diabatic Expansion of High Explosives Detonation
Products. Report. Lawrence Livermore National Laboratory, University of California;
1968.
19. Hansson H. Determination of Properties for Emulsion Explosives Using Cylinder
Expansion Tests and Numerical Simulation. Swebrec Report, No: 2009:1. ISSN 1653-
5006.
20. Available from: URL: 〈http://www.lstc.com/lspp/content/overview.shtml〉.
21. Ouchterlony F. The Swrbrec function: linking fragmentation by blasting and
crushing. Min Tech. 2005;114(1):29–44.
22. Banadaki M. Stress-Wave Induced Fracture in Rock Due to Explosive Action. [Ph.D.
Thesis]. Toronto: University of Toronto; 2010.