SPE

SPE7480
GEOTHERMAL
WELLS
WELLTESTNG INTWO-PHASE
byAlain C. (%ingarten, Member SpE-AIME, B.R.G.M.

~Copyright

1978, American

Institute

of Mining. Metallurgical.

and Petroleum Engineers. Inc

This paper was presented al the 53rd Annual Fall Technical Conference and Exhibition of the Society of Petroleum Engineers ot AIME. held in Houston. Texas. Oct. 1-3,1978 The malerial ISsubject to correction
by the author Permias!on 10 COPY la restricted to an abstract 01 not more than 390 worda. Write 6200 N. Central EXPY. Dallae. Texaa 75206

ABSTRACT
This paper deals with the possibility of using
usual oil well transient preaaure testing methods
in hot water wells when flashing occurs in the wellbore. Multiple rate analysis techniques have been
successfully applied to available data on a well drilled by B.R,G.M. in the former French Territory of
Afars and Issas. Practical recommendationsconcerning such teats are also included.
INTRC2UCTION
Well established pressure trans~ent analysis
techniques are routinely used in oil and gas wells
for the determination of reservoir parameterslz.
In recent years, improvements in these techniques
have allowed a better understandingof complex reservoir and well behaviore as in the presence of akine
wellbore storage3, or fracturesab.
Only recently have these methods been applied
to geothermal systems. For a number of years,
geothermal well testing was based on empirical methods5~ that, although useful for eatimsting geothermal field production pote~.tialegave limited knowledge on the reservoir itself.
Pressure transient techniques have been applied
successfully in vapor dominated geothermal reaervoira 6,7. It was found that gas well analysis methods apply to steam wells and that wellhead measurements were generally adequate for analysia, thus
eliminating the need for downhole instrumentation.
In water dominated systems, on the contrary,
transient well testing can be far more complex,
depending upon whether two-phase flow develops,
with flashing in the wellbore or in the formation.
Usual oil well methods have already been shown to
apply when no flashing develops8. On the other
hand, they apparently cannot be used when flashing
occurs in the formation. Very little information
was available until recently on the case of hot water wells in which flashing occurs at some depth in
the wellhcre. This is mainly due to the difficulty
References and illustrationsat~end of paper.

of bottom hole data gathering that often results

in mechanical damage of the measuring devices because of the high temperatures involvede and of the
high production ratea of boiling geothermal fluids.
A two rate flow teste conducted on one well
in the Cerro Prieto field was recently presented
by Riveratand RsmeylO. The data were satisfactorily interpretedb meana of models sug ested by
Russelll, Seliml1 , and Odeh and Jones!/
3e14.
To the authors knowledge, no other example of
transient test in a cwe-phasewell haa been published in the literature. Unpublished data, however, were available from a well drilled in 1975
by B.R.G.M. in the French Territory of Afars and
:ssas (now Republic of Djibouti).
Analysis of these data, which include pressure
build-up as well as two rate flow tests is presented hereafter.
ASAL WELL
TheAsal rifte located 80 km west of Djiboutieis
one of the active rifts-in rifttstructures of
the Afar depreasione a transition between the Gulf
of Aden and the Red Sea ridges.
Attention was drawn to this zone because of
the presence of a g:aben structure and of geochemical particularitiesof various hot springs.
Two wells were drilled on the S-W margin of the
rift, at locations chosen mainly from geological
considerations. The first hole (Asal 1) reached a
hot water geothermal reservoir, the second, one
kilometer away, was dry.
Both wells found, from top to bottom, a recent
basaltic seriese then a thick rhyolitic volcanic
aerieae and finally an old tectonice tiltede
basaltic series, where the reservoir is located,
The first well was drilled to a depth of 1130 me
where heavy mud losses occurede while the second
well reached 1550 m. An important normal fault
appears to separate the two wells.
A schematic of Asal I wellhead is presented
in figure 1. The well could be produced either
vertically through a 6 tube, or horizontally
.-

WELL TESTING IN TWO-PHASE GEOTHERMAL WELLS

through a calibrated pipe. Four sizes were avail11and ln) which were used to
able (namely, 6, 4, 2
control the well flow rate. In?Idition, a 2
stainless steel tubing was lowez-d into the wellbore to a depth of 650 m, for wireline temperature
and pressure measurements with bourdon-tube type
gauges. As the well could not initially start by
itself, the same tubing was used for inducing it
:.ntoproduction by air lift. The initial water lev.1 in the well was at an approximate depth of 200 m
Jlow the well head.
The flow rate history of ASAL 1 well is presented on figure 2. After completion, the well
was produced through a 6 outlet for 6 days in preparation for a build-up test for obtaining reservoir parameters (flow period+1). Mechanical problems, however. were encountered with the stuffing
box on the lubricator attached to the 2 stainlese
steel tubing, and these required the closing of
the well for about 9 hours in order to retrieve
the measuring davices (flow period.#2). During
this unplanned pressure build-up, temperature and
pressure could only be measured at the wellhead
and are shown on figure 3,

Although no unique quantitative estimate of

the reservoir parameters can be deduced from figure 6, important quantitative information can be
obtained.
By comparing the log-log plot of figure 6
with the skin and storage type curves published in
ref. 3, it becomes apparent that none of the 6
flowing tests hatibeen run long enough for the
radial flow logarithmic approximation to apply.
On the other hand, that approximationdoes apply
to the 2 flowing test, which had a much longer
duration, and to the variom build-up tests.
The comparison also suggests that the early
time behavior of the well is of the changing wellbore storage type, going from liquid level controlled storage to compressibilitycontrolled storage when flashing or condensation occurs.
Figure 6 further indicates that the radial
flow logarithmic approximation applies sooner for
build-up than for drawdown tests. This is somewhat opposite to what is generally observed in
single phase wells, and may be due to the twophase nature of the flow.
Anothez comment regarding figure 6 is that,
although a dry well had been drilled one kilometer
away from ASAL 1, no impermeable boundary is apparent on the log-log plot.

It can be noticed on figure 3 that the wellhead temperature drops suddenly after 400 minutes.
This apparently corresponds to steam condensation
in the wellbore and defines the time limit after
which the well has to be induced into production.

Figure 6 was also used to estimate the mass

flow rates under resenroir conditions during the
various flow periods.

The well was again started by air-lift and

flown through the 6 outlet for approximately 4
days (flow period#3) before being shut-in for
20 hrs (flow period# 4). In order to obtain the
well characteristiccurve relating wellhead pressure and mass-flow rate, the well was again started by air-lift, flown for 43 hours with a 4
outlet (flow period~5), shut-in for 3 hours
(flow period#6), and open again with the 6
outlet. Because of the ehort 4 pressure buildup duration, the well was able to start by itself.
After about 3 hours of 6 production (flow period
# 7), the 6 outlet was changed into a 2 outlet;
the well was then produced for 3 days (flow period
# 8), shut-in for 7 hours (flow period#9), open
again with the 6 outlet (flow period#lO) which
was changed into a 1 outlet after 3 hours (flow
period#ll).
During this final teet, there was
evidence that the flow could not be sustained, and
no build-up was attempted.

As no separator was available, these were

first deduced from Jameslip pressure method15 :
James observed that, as a fairly large flow of
steam or steam-watermixture was expanded along a
pipe to the atmosphere, the pressure at the extreme end (lip) of the pipe was greater than the
atmospheric pressure, and proportional to the
mass flowing, and the enthalpy, as in the following formula :
~ E1*102
0.96

(1)

= 11 400.

L
where G is the mass velocity (in lb/sq ft.see),
E is the stagnation enthalpy (in Btu/lb, at reservoir conditions), and PL the critical lip pressure (in psi absolute).
Jamess formula was transformed for the present study into a more convenient system of units
to yield :

Although the primary objective of these last

tests was to obtain the well characteristic curve,
bottomhole measurements were taken throughout the
duration of most individual flow period, in order
to check the results of the eecond 6 build-up
test. Temperature and pressure measurements were
also performed at various depths in the wellbore,
and are shown on figures 4 and 5, respectively.
They indicate a reeervoir temperature of 253C,
and show the presence of a flashing front at
approximately870 m with the 6 outlet, 850 m
with the 4 outlet and 700 m with the 2 outlet.

w=692.33PL096

E-]02d2

(2)

where w is expressed in tons/hr*, PL in bars*, and

d in inches.
Data for mass flow rate calculations are summarized in Table 1. The enthalpy value used in
Eq. 1 was that for pure water, although the geothermal fluid salinity was of the order of 200,000
ppm. Enthalpy valuee for highly saline brines are
not readily available in the literature.

TYPE - CURVE ANALYSIS

Aaa 1 bottomhole pressure data that weru available for flow periods 4, 6, 7, 8, 9, IO and 11
are ehown on the log-log plot of figure 6. For
each individual flow period, the difference
between the bottomhole pressure during the test
and that at the beginning of the test is plotted
versus the time since the beginning of that test.

SPE 7480

As indicated in Table 1, the lip pressure was

uniquely defined only with the 6 outlet, corresponding to a rate of approximately 83 tlhr.
With the 4 and 2 outlets, on the contrary, the

Metric units are used throughout this paper :

1 bar = 105 Pa
1 ton = 103 kg;

.-*
xn

.&.(.
. . -.

TI.on

I*OU

lip pressure was oscillating between two extreme

values, because of the slug nature of the flow, and
only a range of rate values could be obtained by
applying James formula (71 -83 tlhr and 40-54
tlhr, respectively). According to James 16, such
pulsations could have been eliminated by using a
glycerine-dampedpressure gauge with a needle valve.
By throttling the valve, the lip pressure stabilizes
at the correct value, between the two extreme ones.
However, by comparing the relative position of
the different curves on figure 6, it appears that
the r~tes with the 4 and 2 outlets were more likely 71 ad 23 tlhr, respectively. In the same way,
the rate ~ith the 1 outlet seems to be equal to
11 t/hr.
MULTIPLE RATE ANALYSIS
Quantitative analysis of Asal 1 bot~~-zihole
pressure data was performed with a variable drawdown model, similar to that proposed lJyOdeh and
Jones 13. The well was col.sideredas ..laving
been
produced at different flow rates, q), qz,... q .
Aseuming the radial flow Iogarithm;,capproxima/ion
valid for each one of the flow pe:iods,the bottomhole well.pressure during the nt~ test is equal to :

n q.-q.
~~

qnu
P
Wf

Pi-

411kh 1
j=ln

+ ln~=+

0.80907 + 2S

(3)
\

With geothermal metric units, Eq. (3) becomes :

P=
Wf
n
n
j~l

(Wj-wj-l)

wnv!J
+ Pi - 0.228 F(ln*

in

60

(tn-tj-l)

- 8.434 + 2S)(4)

In Eq.(4), k is expressed in darcy, t in minutes, and rw in meters.

Therefore, by graphing ~wf versus E (wj~j-l)
In 60 (t - t. ,) in carteszan coordinates
one shoul~ obt&
a straight line, the slope of
which is equal to :
m=

0228 E

A straight line of slope in= 3.6 10-3 is clearly evidenton figure 7. Substituting=
3.6 10-3
into Eq. (4) and taking v= 1.25 10-3 m./kg and
u = 0,2 cp yields :
kh = 15,9 Dm.
(P = 0,2 cpwas obtained by extrapolatingpublished
viscosity curves1 to a temperature of 253C and a
salinity of 200,000 ppm).
The calculated initial pressure is 76,6 bars,
which compares reasonably well with the value of
77,4 bars measured before the test.
It can be observed in figure 7 that data points
for build-up times greater than 6 hours deviate
from the straight line. This is a consequence of
the condensation %ffect already noticed on figure 3
during the first 6build-up (flow period#-2).
The early time deviation (for build-up times
less than 20 minute?) is caused by wellbore storage
effects.
In the same way, analysis of the 4 build-up
test (flow period#6) yields a kh value of 15,0 Dm
and a pi value of 76,4 bars (figure 8).
kh and p. from the 2 build-up (flow period #9)
are equal to 16,8 Dm and 77,2 bars, respectively
(figure 9).
Flowing
------

in Darcy units.

-0.228 ~kh

.J

In (tn-t. )
J-l

12RThllURT!?hl
--------------

(5)

This slope is independent of the flow rate.

The intercept of the straight line can be used to
calculate the skin, if pi and @ are known. In
the case of a build-up test (wn = O), the intercept
is equal to the initial pressure pi.

test
analgsis------------

A plot of p f versus the multiple rate function

for the two 6 f~owing tests (flow periods&7 and
10) is shown on figure 10. The slope for the first
one (+#7) is equal to 7,2 10a , which yields
kh = 8 ~, while that for the second one (#IO) ia
equal to 10,2 10-3 (kh = 5,6 Din). Aa expected from
figure 6, these kh values are much lower than those
obtained before, and are not correct. On the other
hand, the analyais shown on figure 11 of the 2
flowing test (flow period#O) yields a slope equal
to 3,5 10-3 , corresponding to kh = 16,3 ~, which
ia consistent with the results of the ?mild-up test
analyses.
The sudden increase in bottomhole pressure
indicated on figure 11 is likely to be caused by a
decrease in the flow rate. The same phenomena
appears on figure 12, in the analysis cf the 1
flowing test (flow period~ll), during which the
well was not able to sustain the flow : the last
recorded pressure is roughly equal to the initial
pressure. All the other pressure points fall on a
straight line, whose slope is equal to 3,2 10-3.
The correspondingkh is equal to 17,8 Dm, which
agrees well with the build-up testa results.
SKIN EVALUATION
It is not possible to obtain the sk~ factor
directly from the multiple rate analysis plot,
because of insufficient reservoir information. The
skin factor, however, can be evaluated from a type
curve match, with a dimensionlesspressure given by ~

D =

AP
kh
0.228 Wn::VIJ=~

(6)

for a build-up test, or

Build-u~
-------

test
analysis----------.-

A plot of pwf during the second 6 build-up

test (flow period~4), versus the multiple rate
function, (expressedin tonafhr), is shown on
figure 7.

D =

AP
kb
0.228 (wn - Wn-l)vp=

for a flowing test.

AP

2(wn Wn-l)m

(7:

WRT,T.

TFXTTNG TN

TW(I-PMASR

The type curve match is shown on figure 13, and

indicates a skin approximately equal to 25 . Thie
high skin value is probably due to rocks plugging the
bottom of the hole : although the well was drilled to
1130 m, the pressure gauge could only be lowered to a
depth of approximately 1050 m. This would also explain why, although the formation is known to be
fractured, no fracture controlled flow period is apparent on the log-log plot of figure 6.
CONCLUSIONS
The following conclusions can be reached from
the present study :
1) Two phase geothermal well testing can be
carried out with standard bourdon-tube
type instruments for extended periods of
time at temperatures ae high as 253*C and
salt content of the order of 200,000 ppm.
2) Usual multiple rate techniques can be used
in two phase geothermal wells for obtaining reservoir characteristicsfrom bottom
hole transient pressure data. Type curve
matching also provides useful qualitative
and quantitative resemoir information.
3) Early time bottom hole pressure appear to
be wellbore storage controlled. There
are indications that wellbcre storage
changes from liquid level type to compressibility type, as a consequence of water
flashing or steam condensing in the wellbore. These wellbore storage effects
appear to last longer in drawdown and
double-rate tests than in build-up tests
(200mn compared to 20mn for the well
Aeal I studied in this paper). This has
to be taken into account when planning
such tests.
4) When the well is not artesian, there is
no practical avantage of running long
build-up tests, because of the effects of
steam condensation in the wellbore (6
hours was the practical limit in the case
of well Asal 1 ).

---- f3?fWFTTi!RMAT.
----------

C2=

=
=

total fluid compressibility,bar-1

enthalpy, kcallkg
formation net thickneaa,m
permeability , darcy
slope of straight line
constant rate intervals
pressure, bar
volumetric flow rate, m3/hr
wellbore radius, m
skin factor
producing time, min
mass flow rate, tlhr
viscosity, cp
epecific volume, mslkg
poroeity, fraction

initial condition
constant rate interval
wellbore or water
flowing conditions at well bottom

*I..

Vt,Qn
,-?

The work reported in this study was supported

in part by the Commission of the European Communities (Contract 081-76 EGF, Geothermal Energy Program). The author,is grateful to the Conunissionof
the European Communities and to the management of
the Bureau de Recherches G6010giques et MiniSres for
permission to publish this paper.
REFERENCES
1.

Matthews, C.S., and Russell, D.G. :

!Ipressurebuild-up and flow tests in welle
Monograph series, Society of Petroleum Engineers of AIME, Dallas,(1967),Volume 1.

2.

Earlougher, R.C. : Advances in well test analysis Monograph series, Society of Petroleum
Engineers of AIME, Dallas, (1977), Volume 5.

3.

Agarwal,R.G.,Al-Hussainy,R.ad Ramey, H.J., Jr:

t~~ investigation of wellbor~ storage and skin
in transient liquid flow - 1 analytic treatment, Sot. of Pet. Eng. J, Sept. 1970, 279.

4.

Gringarten, A.C., Ramey; H.J., Jr and Rag avan:

Applied pressure analysis for fractured wells
Pet. Tech., July 1975, 887.

5.

James, R. : Factors controlling borehole performanceU.N.Symposium cn the Development and

Utilization of Geothermal Resdurces, Piss
Proceeding (Geothermics,Spec.Iss.2) V.2 pt.2
p 1502.

6.

Ramey, H.J., Jr and Gringarten A.C. : Effect

of high volume vertical fractures on geothermal
steam well behavior, Proceedings, 2nd U.N.
Symposium on the Development and use of geothermal resources, San Francieco , California
(USA), MsY 20-29, 1975, V.3, p. 1759.

7.

Barelli, A., Msnetti, G., Celati, R. and

and
baCk-pX.e66ure
ests
NERI, G. : llBuild-up
on italian geothermal wells, Proceedings,
2nd U.N. Symposium on the Development and use
of geothermal resources, San Francisco,
t%lifornia (USA), May 20-29, 1977, V.3, p.1537,

8.

Witherspoon, P.A., Narasimhan, T.N. and Mc

Edwards, D.G. : Results of interference tests
from two geothermal reservoirs, SPE paper 6052
51st Annual Fall Meeting, New Orleans, Ott.3-6,
1976.

9.

Gulati, M.S. : Pressure and temperature buildup in geothermal wells Proceedings, Stanford
geothermal workshop, Stanford University,
Dec. 15-17, 1975.

10.

Rivers, J.R. and Ramey, H.J., Jr : Application

of two-rate flow tests to the determinationof
geothermal reservoir parameters, SPE paper
6887, 52nd Annual Fall Meeting of SPE, Denver,
=9-12,
1977.

11.

Russell, D.G. : )JDete~inationOf fo-tion

characteristicsby two-rate flow tests,
J. Pet. Tech., Dec. 1962, 1349.

Subscri@s
--------

i=
j=
w=
Wf

cm7

.ACKJ70WLEDGEMENT

NOMENCLATURE
E=
h=
k=
m=
n=
P=
q
r=
Sw
t=
w=
u=
v=
Q=

------

WRT.T.R

12.

A.C. GRINGARTEN

SPE 7fb80

Odeh, A.S. and Jones, L.G. : Two-rate flow

J. Pet. Tech,, Janv.
test, variable rate case,
1974, 93.

14. Odeh,A.S. and Jones, L.G. : Pressure drawdown

analysis, variable rate casa, J. pet. Tech.,
Aug. 1965, 960.

13. Selim, M,A, : Modification of the two-rate

flow nethod for determination of reservoir
parameters, J. Institute of Petroletaq,
V. 53, N527, NOV, 1967, 343.

15. James, R. : Alternativemethods of determining

enthalpy and mass flow Proceedings, U.N.
Conference on new sources of energy, RO=,1961.

[[

16. J-es, R. : Sprivateco~unication or16an8

(France), Nov. 1977.

TABLE

MASS FLOWRATE MEASUREMENTS

---------------. -----

OUTLET
DIAMETER
(inch)
[pgy[~sj(fr=~l,l]~?-.----..-- .--------- ------....-.--6
4
2
,11

0.5
2-2.5
6-8.5
not
available

1.5
3-3.5
7.9.5

83
71 - 83
40- 54

II

83
71
23
11

5,

=4-!=

4111

!:1
7

5/8

13 3/8

Fig. 1 - Schematic of ASAL 1 well.

100

TIME

,OAYS

Fig. 2 - Mass flow rate changes during ASAL 1 tests.

\
w H TEMPERATuRE

wELL

IWAO PRESSURE

,
WILO -UP

..

D*O.00

40000

3WO0

20090

IWRO

TIME , MINUTES

Fig. 3- ASAL 1 pressure buiId-up test (6 outlet, flow period #2).

oor
~
Zwlw

1
++

00.00

6X1 70

(tOUTLET
)

2X1 n

(t OuTbEl )

31.x
70($

28 X.75($ WTIST)

21.X.?6 (d OUTLET}

0UTL27)

1?. x.76[s7ATIO)

AO.IX.70R7A71C)

WRWA G31LLIW0

WEALLOWwSLL9 1972

aw.w

Fig. 4-

4oom

W090
0EP7H, METERS

8CW0

Iwono

Temperature profiles in well ASAL 1.

1800.00

800.00

2.W.7S (IWTL2T)

O 3.X1.75 (*OUTLET)
7om

51.X.7614OU1LE1)

o W.x .75(0 OLWMT )

80.00

0 t4.x.70

(qullm

D ID.X.TS (mATIC )
00.00

a 20. IX.70 (mm

II

mm

2W.00

40000
mmn

IOw.w

8W.W

Uoo.w

, M2TCX2

Pressure profiles in well ASAL 1.

Fig. 5-

VALIDITY

OF SEMI-WQ

METHOD
-t-1

BUILD-UP

oRAW;OWN
~

II

2UL0 .UP (n04)

OROO

+.+

..+

4W1L0
-...+.--.

-+-+-H

,,+,-.
*4
\*

, ~:

.-

l OWAWWWN
(0111
-UP [II*6)
+...++
+.+.+++
-+-++
k ORMIOOWN(,*8)

@ +
;

2 IuILOAW

(*91

+
;

.
/
..i
/+
..
I
I

,.2

10
&t,

Fig. 6-

MlhUTES

ASAL 1 pressure tests.

Id

o
0

0
0

0
0
0

0
0

W.w
26000

aoo.oo

moo

40000

2W.00

Soo.oo

42000

ew.w

wow

7W.00

7&Xlo

MUL71PL2 RATE FUNCTION

Fig. 7-

ASAL 1 pressure build-up test (6 outlet, flow period #4).

Ww

PIS76.6

m. 3,0 to -2

73W

\.

a
a 4

%6
6

row

n
aA
n

:
..

A
n

A
00

24W
I

%.

b
3

o
0

2om

A
o
n

Woo

2920W0

22000

5WO0

40000

4WQ0

Sww

Swm

2WW

Wo.oo

70000

WJL?WLC RATE FUNCTION

Fig.

8-

ASAL 1 pressure build-up test (4 outlet, flow period #6).

7wno

sem

70.00
0

moo
!.
II W.w
!4
I 00.CO

sow

m.m
0

,
Oaoo

,
10090

160.W

200.00
MuLnPLE

Fig. 9-

800Q0

mm
mm

mom

400.CQ

480.00

Sw.oo

FUHCTION

ASAL 1 pressure build-up test (2 outlet, flow period #9).

ao.oo

70.00

mao
i

6090
1
3

~ 0.00

0600

Oosm
Wooo

4mw

Wo.m

700Q0

7m.oo
WLTIPU

Fig. 10 - ASAL 1

wow
RATE

mono

Wow

Wom

1000.00

FUNCTION

(6 outlet, flow perids #7

pressure
flowing
tests

and 10).

IMOBO

$.

o
00

,
m

4mm

4mm

9V2)J0
MLwl?lx

Omao
mom
KATEFLwc71an

7m.eo

Wow

lmm

Ftg. 11 - ASAL 1 pressure flowing test (2 outlet, flow period #8).

9090

e
7D.W

v
VVG

mm

Zwm

am>

mwv

Me.m
wLTIRC

4aam

4wile

mono

worn

mom

RAIE fW310N
\

Fig. 12- ASAL 1 pressure flowing test (1 outlet, flow period #11).

Uvne

C=s

[ ----------

---------

*U

-----

.....

... ... ... ..

.....

-----

...........

-----

-----

.....

... ........

.....

.....

...........

-----

..... a

.... .......

-----

----.21

---------

D .. . ..

10

--------- ----

----1s

...........

s
----=

-----

--p~isjj~
--;;::..::
.-. .--.

--.-----

-0-.

1.---

10

105

I
106

109

10

Fig. 13 -Type curve matching with skin and storage type curves (ref.3).

-.