You are on page 1of 9

G R. Dysart, A.M.

Spencer, A L Anderson*
In December 1967, El Paso Natural Gas Con~pany,In

surfaces These explos~vesare very applicable t o exploslve fracturlng of 011 and gas wells
Renewed Interest 111 exploslve fracturlng has come
about because hydraul~cfracturlng IS not always successful In fields wlth unconnected permeablllty streaks
Explosive fracturlng can connect these streaks wlth flow
channels, and then, ~f deslred, these streaks may be
hydraulically fractured through these flow channels
Experience may show conventional high-explosive fracturlng t o be more convenient and less costly than
hydraulic fracturlng for certaln a p p l ~ c a t ~ o nand,
s In some
cases, the only way a reservoir can be stimulated

conjunction w t h the U S Government, detonated a ther-

monuclear devlce underground (Project Gasbuggy) t o

create extenslve fracturlng In the Plctured Chffs formatlon Tlus form of exploslve fracturlng ls a contlnuatlon
of one of the first well-st~mulatlon techn~ques-nltro
shootlng Nitro shootmg was a dangerous process In all
aspects-storage, handlmg, and appllcatlon
In recent years new high-energy exploslves have been
developed whlch are safe to handle, temperature~nsens~tlve,
and can be produced In varlous plastlc~tlesso
they can be made to conform to well-bore or fracture

On the mornlng of December 10, 1967, history was

made In the form of an underground nuclear exploslon
called project "Gasbuggy" Thls nuclear detonation,
whlch was the equivalent o f 26,000 tons of TNT,
occurred at 4,240 ft below ground level just 4 0 ft below
the base of the 300-ft thlck Plctured Cllffs gas sand In
RIO Arrlba County, New Mexlco T h ~ was
s the flrst j o ~ n t
government-industry-sponsored commerc~al apphcatlon
of a nuclear devlce to lncrease the productlv~ty of a
mlneral-producing format~on Even though only llnuted
data are presently avadable, gas pressure 1s u p slgnlficantly In the chimney and the formatlon was fractured
for hundreds of feet from the shot hole
Other types of exploslves have been In the news, too
Several patents have been Issued In recent years In the
explos~ve-fracturing area The U S Bureau of Mlnes has
done extenslve testlng In shallow rock formations New
formulations have been tested In 011 wells However, the
use of exploslves t o stmmulate petroleum-bearlng format ~ o n sn not a new art The first wells were shot wlth
black powder, and, m 1865, n~troglycerln was ~ n t r o duced wlth pronounced results In the boomlng Pennsylvanla od field
The advent of hydraul~c fracturlng III the 1950's
caused exploslve stlmulatlon of petroleum wells t o
decllne dramatically The many problems associated
w ~ t hnltro blasting, such as safety and the llm~tatlonsto
worlung In open holes, caused the decllne In the use of
nltro shootlng However, there a a requirement for thls
type of stlmulat~on, as evidenced by the extenslve
Interest expressed In "Gasbuggy" and other major
programs for certaln formations
*The Western Company, R~chardson,Texas
?Presented at the sprlng meetlng of the Mid-Cont~nent D~str~ct,
API D ~ v ~ s ~ofo Product~on,

Recent research work done In West Vlrglnla and West

Texas lrldlcates that hydraul~c fracturing, though an
excellent method of st~nlulatlngwells, IS llrnlted to
propagating a fracture only along the plane of weakness
In some types of rocks Thls leaves a large area whlch
drams e ~ t h e rlnefficlently or not at all because ~t 1s
untouched by stlmulatlon attempts Thls condltlon 1s
shown In F I ~1 In thls type of formatlon, Inltlatlng
unorlented fractures would be deslrable These fractures,
w h ~ can
h be created by explosions In rock, could then
be extended hydraulically t o further Increase the fracture dralnage area
The declslon t o use thls type of stlmulat~on1s based
on the need and abdlty t o cause a controlled exploslon
In a well It 1s the Intent of the authors t o ald the reader
In making declslons about uslng thls type of stlmulat~on
Thls paper presents background data In three areas
whlch affect these declslons, vlz
1 Types of explos~vesavadable
2 Fracture-radlus pred~ctlon
3 F ~ e l dappllcat~ons
Following black powder, three more powerful exploslves were used In well shootlng They were 1 , llquld
nltroglycer~n,2, 100-percent gelatln nltroglycerln exploslve, and 3 , varlous grades of d y n a m ~ t e Llquld
nltroglycerln was preferred and was used in a major~tyof
early 011-field shootlng jobs, although 100-percent gelatin exploslve also produced satisfactory jobs The varlous


Fig 1 - Top Vlew of a Vertlcal Fracture

grades of dynamite were seldom used in well shooting,

pniiiardy because of handling problems
N~troglycerin [(C3~5)(N03)3] 1s a n unstable liquid
which decomposes t o gaseous products when subjected
t o a sharp shock Each pound of nitro yields 157 7 cu ft
of gases and generates a n adiabatic flame temperature of
6,280 F The detonation rate (the rate of propagation of
the reactlon) is 23,600 ft/sec Nitroglycerin has an upper
temperature limitation of approxlmately 140 F, where its
auto-decomposit~onrate becomes significant The usual
practice of loading nitroglycerin was t o put the material
into long cyllnderical contamers called torpedoes or shells
and lowering these t o the formation to be shot The
nitro was then detonated by means of a tlme bomb or
direct-firing detonator from the surface on a wireline
In recent years, new f a m d ~ e sof explosives have been
developed for the mlnlng industry These explosives
were designed specifically to cause rock fracturing and
rock breakage Their characteristics differ from those of
nltro in that theu detonation rates are considerably
slower and the energy release occurs over a greater time
interval This slowdown allows more efficient transfer of
the energy t o the formation rock These new famllies of
explosives started with the familiar fuel-oil-ammoniumnitrate explosives
The Texas City d~saster of 1947, caused by the
detonation of a sh~pload(3,000-4,000 tons) of ammonium nitrate (AN), brought t o the attention of the mining
~ndustry this low-cost, powerful explosive w ~ t hgreat
blasting potential that somehow had been overlooked by
the explosives industry. However, it was not until 1956
that the AN/FO (Ammonium Nitrate/Fuel011) explosive
system was characterized, standardized, and widely
utilized The 9-year delay was caused by a search for a
sensitizing agent for AN Sensitizing was finally accompllshed with fuel od, but a booster was stdl necessary t o
detonate the AN/FO system However, AN/FO explosives were not the answer t o all the explosive
problems Better mater~als w h ~ c h were safer, more
efficient, and able to operate under wider envuonmental
extremes were invented and patented These newer
materials inay be lumped into two broad classes slurry
blast~ngagents and slurry explosives
The term "blasting agent" is conventional for fuelsensitized ammonium nitrate, where the fuel alone a
non-explosive and the product requues powerful boosting to detonate it Slurry explosives are very sundar in
that they are "fuel"-sensitized AN with or without
oxidlzers (e g , s o d ~ u mnitrate, sodium perchlorate, or
ammonium perchlorate) in which the solid fuel and the
oxidlzers are dispersed m a continuous fluid m e d ~ u m
which 1s generally an aqueous solut~onwith a sohd-fuel
mutture containing h ~ g hexplosive (usually TNT) The
term "blasting agent", from the history and ~ n t e n tof
usage, ~rnpliesa much less hazardous material than those

usually denoted as explosives For this reason, slurry

explosives also have been designated as blasting agents
For identification purposes then, slurries 'when sensltized with explosives (e g , trinitrotoluene-TNT), wlll be
classified as slurry explosives (SE) Those in which the
fuel is not a n explosive (e g , aluminum-Al, sulphur or
solid hydrocarbon) wdl be named slurry blast~ngagent
In 1956, an SBA which used A1 for sens~tizationwas
discovered, but the spontaneous alum~num-waterreaction and d~fficulties of controlling critical physical
factors were not solved until 1962 The slurry explosive
trin~trotoluene (SE-TNT) was discovered In 1957, and
gamed wide use by virtue of excellent reproducibdity of
explosive strength and ease of manufacture
In 1962, after the methods t o control the aluminumwater reactlon were developed, the SBA-A1 and the
SE-TNT were conlblned to produce the most powerful
commercial explos~venow avadable These SE-TNT/AI
explosives were designed specifically for rock fracturing
and rock breakage, and are most applicable t o hard rocks
which are often encountered 111 an od-bearing formation
Table 1 shows the compos~tionof t h ~ sSE-TNT/AI and
compares its performance t o a standard mditary explosme, Tritonal, SBA, and AN/FO It can be seen that the
slurry explosme, in volumes equal t o Tritonal, develops
2 2 percent greater cratering potential over a mditary
explos~vedesigned specifically for that purpose This can
be extended t o 7 3 percent by using sodium perchlorate
as the oxidlzer The relative seismlc strength measurements show SBA and SE t o be approx~matelyequivalent
t o Tritonal The addition of SP also increases selsrnlc
strength The S E contains 1 5 percent water, which is the
continuous phase of the explosive It is thickened and
often gelatinized by hydrophdic collo~dsand crosslinlung agents t o provide a w d e range of rheological
properties which can be adapted t o various envuonmental conditions The flow properties may be varied
from an almost sohdified consistency t o a material with
a viscosity of 160 c p which is easdy pumped The water
resistance of the slurry explos~vemay be closely regulated by the types of thickening agents used for the
aqueous phase of the slurries Guar gum, cross-l~nked
wlth borate ions, was the first successful hydrophdic
collo~dused for this purpose
The strength of a n explosive or its effectiveness may
be identified w ~ t hthe (thermodynam~c)maxlrnum avadable work (A) Generally A 1s approxunately the same as
the heat of explosion (Q) for most hgh-density explosives A plot of the Q value vs percent fuel for slurry
explos~ve TNT/AI, slurry explos~ve TNT, and slurry
blasting agents-AI, AN/FO, and nitroglycer~n-is shown
In Fig 2 For comparison, the Q of several other
non-aluminized explosives 1s also shown The Western

G R Dysart, A M Spencer, A L Anderson


Table 1
Cornpos~tlonand Properties of Explos~veMaterials


Fuel, percent
b Aluminum
c Fuel 011
Water, percent
Thickener, percent
Dens~ty,g / n ~ l
Deton~tlonPressure P2,kb
Relatlve Crater Volume
Relatlve Selsmc Strength
AN - Ammonium Nitrate
FO - Fuel Od
SBA - Slurry Blasting Agent




SE - Slurry Explosive
SN - Sod~urnNltrate
SP - Sodlum Perchlorate

Blasting Gelatin, PETN, NG

blastlng agent (SE-TNTIAl) 1s the most powerful commerc~alexploslve available and 1s represented by the top
llne In the figure
The SE-TNT/Al exploslves are temperature-stable to
180 F, and an SE-RDX/Al materlal can be produced
whlch will be stable to 330 F, thus ~nsurlngdetonation
control and preventing decomposltlon of exploslves In
much deeper wells than those shot previously It 1s
believed that even hotter wells (about 550 F) can be
successfully treated by sunple changes In treatlng technlques and nilnor adjustments In the chem~calformulat ion

Percent Fuel

Flg. 2

- Heat of Explosion vs. Percent Fuel

The slurry explos~veshave been successfully detonated under relatively hlgh pressures We have had
successful detonations In 5,000-ft wells Also, the
materlal has been detonated offshore at depths greater
than 2,000 ft
Sensitlv~ty is an unportant cons~deratlonfrom two
polnts of vlew safety In handling and controlled ~ n ~ t l a tlon The new slurry materials, as normally manufactured, w~llnot detonate in charge diameters of less
than 4 In, and requlre a booster for initlatlon Dynarn~te
can be detonated by a blastlng cap firlng through 118-111.
tluck plastic spacers The slurry exploslves may be
sensltlzed so that a blastlng cap unmersed m the slurry
w~lldetonate the nilxture Although the slurry explosive
reacts by surface burnlng of exploslve grains as convent~onal high explos~ves do, the slurry blasting agents
detonate by diffusion-controlled reactions enhanced by
free space and "hot spots" For thrs reason, aeration or
gasslng 1s requlred Some blastlng agents use sodium

bicarbonate and acetlc a c ~ das addltlves t o sens~tlzethe

mater~alsufficiently t o detonate Thls add~tlonalvanable
1s not requ~redIn slurry explos~ves
The energy transfer t o the format~on d u n g an
explos~vefracturlng job w ~ t hthe new exploslves 1s over a
longer t ~ m epenod, and the energy transfer effic~ency1s
consequently 111gher than that of formerly used explosrves For t h ~ sreason, less crushing of the rock around
the bore hole and more extenslve formation fracturlng 1s
t o be expected
Another explos~ve,a 11qu1d rocket-fuel product, has
been In the nat~onalnews because of ~ t reported
use as a
slniply placed anti-veh~cle mlne The mater~aln In 11qu1d
form and 1s poured onto the ground Even after settllng
Into the porous earth ~t can be detonated, whlch would
lndlcate ~ t posslble
usefulness In petroleuni reservolrs
Some attempts have been made t o use a product of t h ~ s
type In wells, but several drawbacks have prevented
more w~despreadtesting It requlres specla1 pumps whlch
must be remotely controlled for safety Tublng leaks can
be dangerous and tublng has t o be spec~allycleaned
when the mater~al1s pumped Because safety 1s of great
Interest In well-st~mulat~on
work, thls type of exploslve
would not quallfy In ~ t present

T o predlct the fracture radlus of an explos~on,~t 1s
first necessary t o cons~derthe mechanics of the problem
Probably the most popular concept of rock breakage by
exploslves has been that the gases generated by the
detonat~onof the exploslve push the rock away from the
bore hole, thus fracturlng ~t However, a study of the
literature deallng w ~ t htlie fracture of plastlcs and metal
plates under exploslve shock loadlng shows that the
reflection of stress waves from free surfaces accounts for
the fracture patterns observed m these materials
Fracturing of any materlal 1s prlmardy the result of
three stresses compresslve, shear, and tenslle Typ~cal
hard-rock strengths for these stresses vary In the ratlo of
100 10 1 When stress waves propagated by the exploslon exceed any of these levels, cracks occur The hlgh
gas pressure generated b y the exploslve detonat~on
produces these stress waves, but the expanding gases are
not dlrectly respons~blefor the fracturlng
Reflect~onTheory of Rock Breakage
Detonat~onof an exploslve charge In a bore hole In
rock creates a large quantlty of gas at high temperature
and pressure In a very short tlme Thls gas pressure,
actlng agalnst the rock and the column of stemmlng
(whlch may be alr or other materlal In the bore hole or
plpe), generates a compresslve stress and straln pulse
that travels radlally outward Into the surroundmg rock
wlth a speed equal t o or greater than that of sound m

the rock This compresslve pulse 1s characterized by a

slngle compresslon of short duratlon I~avlng a steep
front or rapld rlse tlme and a slower decay or fa11
time Assoc~ated w ~ t h the rad~al conipresslve straln
pulse 1s a tcnslle straln pulse
Near the explosion tlie alnplltude of the cornpresslve
stress pulse 1s greater than the strength of the rock In
compresslon, and the rock 1s crushed and fractured As
the pulse travels outward, ~ t amplitude
decreases rap~dly
because of the divergence of the wave and absorpt~onof
energy by the rock
'The gas pressure In the drdl hole also acts agalnst the
stenimlng materlal compressing ~t and generates a
compress~ve-stresspulse that travels u p the column of
stemrnlng Because of loss of heat and Increase of
volume (mainly resultlng from compaction of stemrnlng
niaterlal and crushlng of rock), the gas pressure In the
drdl hole decreases rap~dlyfrom ~ t peak
If no free boundar~eswere near the explos~onIn rock,
the only damage t o the rock would be the crushed zone
around the shot hole and some assoc~atedradlal crackIng However, the presence of a free boundary near the
shot polnt accounts for much addltlonal rock breakage
Most od and gas reservolrs wlll have many free surfaces
because of the heterogeneity of most reservoir rocks,
and because of natural fracture systems These changes
In character and stratigraphy of the rock will act as free
When the cornpress~ve-stresspulse m p a c t s on a free
surface, two reflected pulses are generated 1 , a tensllestress pulse, and 2, a shear-stress pulse The amount of
energy In each of these reflected stress pulses a a strong
functlon of the angle of lncldence of the ~mplnglngstress
pulse The effects of the reflected tensde-stress pulse are
felt first, as the propagation veloc~tyof these pulses 1s
always greater than that of shear pulses Also, as the
strength of rock In tens~on1s much less than the strength
of rock In compresslon (usually by a factor 5 0 or 100 or
more), t h ~ sreflected tenslle pulse 1s able to break the
rock In tens~onas ~t 1s reflected back Into the sohd rock
The process of tensde fracture by reflect~onof a
compresslve straln pulse n dlustrated In F I ~3 The
lncomlng cornpresslve straln pulse 1s shown just before ~t
lmplnges on a free surface The resultlng reflected tensde
pulse, after a gven mterval, 1s assumed t o equal the
tensde brealung straln of the rock, so that a crack
develops a t the polnt of m a x m u m tensde straln Thls
crack wdl act as a new free surface on which the
lmpactmg cornpresslve pulse wdl reflect Also, the slab
of rock produced wlll move forward because of the
energy trapped In the slab
After another t ~ m eInterval, the first slab has moved
forward, and the new free surface reflects the remalnlng
portlon of the cornpresslve straln pulse The resultlng
tensde straln 1s agaln enough t o crack the rock In


G R Dysart, A M Spencer, A L Anderson

tenslon. This process continues untd the tensde stram

developed IS less than the tensde brealung strain of the
rock and no further crackmg IS produced. In t h
example, two slabs were produced. Rock fracturmg by
the foregoing process has been demonstrated expenmentally for plane waves reflected from the end of rock




Strain Pulse



-i l


Strain Pulse

Small t e n s i l e
Strain Pulse

o bv e s

Fig. 4

- Stram Record

N e w Free Surface


Fig. 3 - Tenslle Fracture by Reflect~onof

a Compress~ve-stra~n

Fracture Prediction
No sultable rnathematlcal model has been developed
which wll descnbe the stress waves m rock whlch are
responsible for rock fracturing. However, an equation
has been developed from experunental data whlch
enables one to predlct maxunum dlstance of fracture
propagat~on.The data were obtalned by shootlng explosive charges m various dlstances from the recorder. A
sample stramgage record ~sshown In Flg. 4.
The strain-gage record mdicates that nothmg occurs
after detonation untd the compressive wave, A, arrlves
and peaks at B w ~ t ha straln of C. The peak tensile stram
occurs at C. The fall stram Ef ~s measured from the
maxlmum compressive straln to the peak tensde stram
In a series of experunents, explosive charges were
detonated at various shallow depths m h e s t o n e and
sandstone rocks and the fa1 straln was measured at
various dlstances. These results are plotted m Fig. 5. The
ordlnate IS the actual dlstance divlded by the cube root
of explosive welght. Then by comparison of these
records wlth hgh-speed photographs of the propagatlon
of fractures to the surface, an estimated value of the
tensde straln at which the rock fractured was made

Scaled Distance


Fig. 5 - Fall Straln vs Scaled Distance

Scaled D l s t a n c e


which a w ~ t h i n 1 0 percent of L e w d experimentally

observed values
Table 2 presents the expected fracture distances for
varlous exploslve weights In both sandstone and lmlestone format~ons The fracture dtstance is belleved
rellable t o 20 percent, depend~ngupon the homogeneity
of the rock The practical lmltatlon on this process wlll
l m ~ t h s techn~quet o formatlons thlcker than 20 ft
For a formatlon less than 20 ft, ~t wlll be hard to
emplace enough exploslve t o be effectlve
The foregolng expermental der~vation 1s based on
data obtalned from surface testlng, and overburden
pressure 1s Ignored It 1s belleved that overburden
pressure may Indeed contribute t o fracture extension for
it wlll lnsure that formatlon rock is at m a x m u m denslty,
and stress wave propagation will not be greatly different
from the results of surface testing
Another polnt that was cons~dered is that the
foregolng data were generated with small cyllnderlcal
explosive charges, not long charges such as would be
encountered In a well bore T o determine ~f thls would
have a n effect, peak pressure data from spherical-shaped
exploslve charges were compared t o the peak pressure
off the slde of a long cyllndr~cal(Ime) exploslve charge
These data are plotted in F I ~7, and ~t is noted that the
11ne-charge peak pressure falls on the same curve as the
peak pressure for a spherical charge However, the peak
pressures from the ends of the charges are considerably
lower, thus offerlng some degree of protection t o
formatlons above and below the forlnat~onbelng fractured

Fig. 6 - Fall Straln vs. Scaled D~stance

( Brealung Straln, DTBS) The DTBS

for sandstone is 560 p n /In and for lmestone 4 0 0
pin /In. If these are plotted on the fall straln curve (Fig
6), they correspond t o a scale distance of 5 for
sandstone and 7 for limestone
From these values a s m p l e empirical relatlonshlp
between fracture dlstance and exploslve welght has been
developed, vlz

D is the fracture dlstance, W is the we~ght of the

exploslve, and K 1s the expermentally derlved number
(5 for sandstone, 7 for lmestone) which is a function of
rock material, density, strength, sound velocity, and
crystal structure Addit~onalconfirnlat~onof this fracture distance predlctlon equation can be found In
Elemerzts of Mlnazg, by Lewis and Clark (see b~bllography), where it is polnted out that a 3-lb explosive
charge buried in llrnestone produces fractures as far as
10.8 ft from the exploslve. Using a k value of 7 in
Equation (I), a 10.1-ft fracture length 1s calculated,

How should the pr~nclplesof explosive f r a ~ t u r ~ nbeg
applled t o well stimulation? Tests made in quarries by
Eakln and MlUer,' uslng nltroglycerln (NC), proved that
under the test condltlons a detonat~oncould be inltlated
in slots or fractures as small as 1/32 In ~ t h e r s have
tested SBA in wells w ~ t hvaned results Ealun and others
Refe~encesare at the end of the paper

Table 2

Exploslve We~ghtand Fracture


Welght, Lb

Fracture Dlstance, F t




G R Dysart, A M Spencer, A L Anderson

Fig. 7

- Explosive Pressure Measurements

Peak pressure from spherrcal charge.

50-lb charge, 24 ft long, peak pressure off s~de
Pressure from end cyl~ndr~cal

r e ~ o m m e n d hydraulically fractur~ng tlie wells, then

d e t o n a t ~ n gan explosion t o lnLrease the recovery rate
B e ~ a u s e propagation of explosives In a fracture a
unreliable and hydraul~c fractures follow planes of
weakness, ~t would seem wlser to reverse the order The
smdU aniount of exploslve contalned In a t h ~ nfracture, ~f
detonated, w ~ l l not generate a s ~ g n l f ~ c a n trockdestru~tlve force under the confrn~ngpressure of the
overburden rock
Cons~derd vertlcal fmcture w ~ t h2 wings e x t e n d m
from tlie well-bore In the usually assuliled 180-deg
dlrectlolis Pdrt of the reservoir will be dralned through
these flow clidnnels, but a large volume wdl not be
connected to the well by flow channels (.Fig 1)
A d d ~ t ~ o n ahydraulic
fracturing wlll only extend the
exlstlnc., frdctures In the same two dlrect~ons
How can fractures be extended ~ n t othe untouched
portlon of the reservo~r?Explos~vesare the answer, for
the destruct~veforce does not follow natural p h n e s of
w e ~ k n e s sIn the rock The affected pattern 1s splierlcal
for J p o ~ n tclidrge Therefore, fractufes wlll extend ~ n t o
the Jreas whlcli would not be affected by hydraulic
frdcturlng The entlre system of fractures could then be
fractured by stages New temporary block111g agents can effectively seal the major fractures,
~ l l o w ~ ntreatment
of the others as shown In Flg 8 T h ~ s
a p p ~ o a c hwill be niost applicable where fractures are
l , In streaks, as In dense Innestone such
very d ~ r e c t ~ o n aor
JS tlie M~ss~sslpplan
In parts of Oklahoma
P r ~ o r to the performance of explos~ve fractur~ng
stllnulatron 111 3 well, the follow~ngcharacter~stlcs nus st
be cons~dered
I Cased or open-hole complet~on
?- Fluld level
3 Number of feet to be shot
4 C ~ s l n gor hole slze
5 Bottom-hole temperature
6 Bottom-hole pressure
These factors control the method of exploslve placement For example, the fall rate of a plastlc bag of slurry



Flg. 8 - Hydraulic Fracturlng after

Explosive Fracturlng

1s much greater In a dry well than In one full of fluld

The bag dragglng agalnst dry plpe a t a h ~ g hveloc~tyWIU
rupture easlly
There are several ways to place the slurry In a well It
can be made w ~ t l ia low v ~ s c o s ~ tso
y that ~t can be
pumped down the tublng, but the t u b ~ n gmust be
removed afterwards There 1s also the disadvantage of
unbalancing the hydrostatic head c o n d ~ t ~ o nwhlch
cause loss of the SE-TNT/AI t o fractures, also causlng
contam~natlon and dllutlon The SE-TNT/AI can be
pumped down the caslng, w ~ t hcement plugs below and
above ~t Tlie top plug can then be fished out by
wirel~neThe t o p of the exploslve WIU b e at the top open
perfordtlon In t h ~ case
The plug can be left In the hole,
~f a tlme bomb 1s placed wlthln the explos~ve.but ~f the
top plug does not go t o the planned depth, the blast
could take place at the wrong ~ n t e ~Ina lthe plpe
Tlie niater~alhas also been dropped Into the wells I I ~
plastic bags, as mentioned, especially slniple In shallow
wells Placement wlth a dump bader 1s also posslble
Another method offenng good control 1s the one stlll
used by the nltro shooters A thln-walled metal tube, or
e shell IS made wlth a
shell, 1s lowered o n a w ~ r e l ~ nThe
closed bottom and a ball handle on the t o p whlch WIU
fall t o one s ~ d ewhen released The depth of placement,
when uslng shells, IS known from wlrel~nemeasurements
Each of these placement methods has been used, but
each well must be planned rndrv~dually A slngle routlne
cannot f ~ every
W ~ t hthe SE-TNT/Al exploslve In place, a procedure
t o detonate the charge must be a p p l ~ e d Prunacord,
blastlng caps, RDX, and comblnatlons of these have
been lowered and set off by wlrellne However, the
presence of w ~ r e l ~ nIne the well prevents placement of a
cement tamp, although ~t works well with a sand tanip
If no tanip 1s used, as In a dry well, the wlrel~rlewill be
twlsted and blown u p the hole, whlch usually necessltates a f ~ s h ~ job
A t ~ m ebomb 1s preferred, and well shooters have
been uslng such a bomb for years wlth excellent results
The bomb conslsts of two separdte clocks w ~ r e dto a
blastlng cap, and boosted by gelled nltroglycer~n
Tests were conducted uslng varlous stenimlng
niaterlals Sand 1s eas~ly placed and used by n ~ t r o
shooters, but ~t can brldge easlly In the plpe ~t can be
blown up the hole, and ~t can leave open places In the
crlt~calzone above the exploslve Such an open place or
v o ~ d wlll lead to darnaged plpe Sand can often be
more costly
reversed out, however, e l l m ~ n a t ~ n the
dr~lllngprocess A q u c k - s e t t ~ n gcement 1s better because
~t bonds to tlie plpe wall lnipart~nggreater res~stanceto
the explos~on Drllllng out the plug and cleanlng u p the
d e b r ~ scan be costly, but ~t IS far better than damaged
and collapsed plpe A competent cement job IS compulsory above the zone to be shot when shootlng In
casing The caslng 1s dlsrntegrated In the area of the


G R Dysart, A. M Spencer, A L Anderson

explosion by the extrenlely hlgh energy release, and the

plpe above t h n p o ~ n t1s subjected to a tremendous
pressure whlch only good external cement can protect
After clean~ngout the well, a fracturlng job should be
performed In the nlanner outl~nedThe cost of thls type
bf st~rnulatlon work 1s ~ncreased over hydraul~c fractunng, not because of the expense of the explos~ve
treatment, but due pr~nclpallyto dr~lllngout the plug
and clean~ngout the well T h ~ excludes
potentrally poor
producers because of the long payout, but u n l ~ k enltro
shoot~ng,the blast-fracturing treatment can be used In
deep wells where production potential a often much
h~gherthan In shallow wells

In summary ~tcan be stated that

I The varlous fam~llesof explos~vesusable In 011and gas-well s t ~ m u l a t ~ owork
have been evaluated and
the most appl~cableexploslve 1s the slurry exploslve wlth
TNT and alumlnum
2 A study of the rock mechan~csand forces ~nvolved
In exploslve fracturlng has been performed, leadlng t o
an equatlon for pred~ctlrlgthe approxlrnate r a d ~ u sof a
fracture caused by a down-hole explosion-D = K W ' I 3
3 Many methods of f~eld enlplacenlent e x ~ s t ,
depend~ngupon well cond~tlons
4 Sternrnlng 1s reqi~lredto m1nmi17eplpe damdge
5 Explos~ve~ n ~ t l a t ~1sobest
n accon~pl~shed
by use of
a tlme bomb
6 A process ~nvolvlng hydraulic fracturlng after
exploslve fractur~ng(blast-fracturing) 1s recommended
for niaxlmum effect~vestlrnulatlon of reservoirs
The authors wish t o express thelr appreclatlon to the
o nThe Western Conlpany for perResearch D ~ v ~ s ~of
m~sslon to prepare and present t h ~ spaper Thanks are
also extended to shooters Boyce Wllllams of Bartlesv~lle,
Okla and D L Darden of Borger, Texas, who alded In
some of the field work

' Eakln, J L and Mlller, J.S Explos~vesResearch to

Improve Flow Through Low Permeabll~tyRock, J Petr
Tech, 1431, Nov (1967)
'Hurst, R E Verbal presentat~onsto Amardlo sect ~ o nof SPE of AIME, Oct 38, 1966, and Oklahoma Clty
sectlon, July 20, 1967

3Levey, Davld Exploszve Stzmulatzon Report, The

Western Company publlcat~on,Aug 30, 1967
Anderson, A L Blast-Frac Report and Recommendatzons, The Western Company publlcat~on, May 3,
Brandon, C W Method of Explos~velyFractur~nga Product~ve011and Gas Forniatlon, 0. S Patent 3,066733, Dec 4 , 1963
Anon Space-Age Explos~veMay Revlve Well Shoot~ng,
O d G a s J , Sept 19 (1966)
Brewer, B Stlrnulat~onof 011Product~onby the Use of
Explos~ves After H ydraul~c Fracturing, PLoducers
Monthly, Feb (1 957)
Cook, M A m e Sclence o f Hzgh Exploszves, Re~nhold
P u b l ~ s h ~ nCorp
New York, N Y , 1958
Explosrons, Dover PubllCole. Robert H
catrons, Inc , New York, N Y , 1948
Coffer, H F, Bray, B G , Knutson, C F , and Rawson,
D E Effects of Nuclear Explosions on 011Reservolr St~mulatlon,J Petr Tech , May (1964)
Atch~son, Thomas C, Duval, Wllbur 1, and Pugl~ese,
Josepll M Effect of Decoupllng on Explos~onGcnerated Stral~lPulses In Rock, USBM 6333(1964)
Atch~son,Thomas C and Pugllese, Joseph M Comparat ~ v eS t u d ~ e sof Explos~vesIn Limestone, USBM 6395
(1 964)
Cook, M A, Cook, V 0, Clay, R B, Keyes, R T, and
Udy, L L Behavlor of Rock Durlng Blast~ng,Trans.
Soc Mznzng Engrs , Dec (1 966)
Duvall, Wllbur I and A t c l ~ ~ s o Thomas
C Rock Breakage by Explos~ves,USBM 5356 ( 1957)
Jacobs, Slgmund J The Energy of Detonat~on,(I. S
Naval Ordnance Laboratory Pub. 4366, Sept 1 7
(1 956)
Thompson, A A A Cornpar~sonof the D y n a m ~ cand
the S t a t ~ cStress-Stram Curve In Sand under Conf ~ n e d and Unconfined C o n d ~ t ~ o n sBallistic
Laboratones Report 1262, Aprll ( 1 960)
Cook, M A Explos~ves--A Survey of Techn~cal Advances, Ind Eng Chem , July (1 968)
Baum, F A Physzcs o f an Exploszon, Flxmatglz,
Moscow, 1959
Lewls, R G and Clark, G B ElementsofMznzng, John
Wlley & Sons, Inc , New York, N Y , 3 r d Ed , 1964