Professional Documents
Culture Documents
Spe 174080 Pa PDF
Spe 174080 Pa PDF
MD (GL)
–3900
Production Well path
18 5/8-in.
string Perforation/Open hole
41/2-in. 16-in. 741.2 m Stimulation treatment
ESP –4000
Monitoring
tubing 1200 m Upper Permian
(Hannover formation)
Depth (mTVD)
Upper Permian
13 3/8-in. (Dethlingen formation)
–4100
2381.5 m
KOP 2780 m
–4200 Upper Permian
1.5°/10 m (Havel formation)
Lower Permian
7-in. (Andesites)
3878 m –4300
5-in.
Total Depth 300 400 500 600 700
4400.4 m 4389 m Vertical Section (m)
Fig. 1—Simplified wellbore completion of the geothermal production well Gt GrSk 4/05 (A2).
assumed. From all particle shapes observed in the bailer samples, assumed to be lower because cold water is pumped into the annu-
a sphere was assumed to have the highest settling velocity. Fur- lus. The viscosity of water at lower temperatures is higher, resulting
thermore, the calculation was performed for pure water at 100 C. in a lower settling velocity. Exchanging pure water with formation
The average temperature along the CT during the cleanout is fluid would lead to a further viscosity increase and, therefore, a ve-
locity decrease. For the necessary fluid velocity during the clean-
1 out, we chose a safety factor of two to ensure the ascent of all
particles. For the reverse circulation, the pressure rating of the well-
1 head equipment and the inner diameter of the monitoring bypass
display the largest operational constraints; 70 bar for the inlet of
the power cable limited the possible flow velocity along the CT.
1 Despite the limitations mentioned in the preceding, the clean-
out operation was performed between 26 November 2012 and 6
1 December 2012. Although planned thoroughly, the CT operation
failed to remove all the solids from the well. This paper provides
Power cable a description of that operation. It also shows how a simple hydrau-
lic model can be used to identify the processes responsible for the
failure to clean out the wellbore properly.
4100
2. Gel/proppant fracture
3 3
2
3 3 3 4200 1. Gel/proppant fracture
MD (m)
4 4300
3 3
3 3
Water fracture
4400
5
0 4000 8000 12 000 16 000 20 000
Fluid Volume (m3)
Fig. 2—Wellhead of Well Gt GrSk 4/05 (A2). The numbers indi-
cate the size and pressure ratings of the flanges: 1 5 DN100, Fig. 3—Evolution of the wireline AD vs. the cumulative pro-
PN160; 2 5 183/4 in., 5,000 psi; 3 5 21/16 in., 3,000 psi; 4 5 203/4 duced-fluid volume. The depth interval of each stimulated sec-
in., 3,000 psi; and 5 5 203/4 in., 5,000 psi. tion is indicated.
3 F CT
8000 kg/m3
pCT
2X Settling Velocity
Required Flow Rate (L/min)
Settling Velocity (m/s)
2 3
6000 kg/m 300 pA BOP
CT
4000 kg/m3
Q CT
200
3000 kg/m3
1
2000 kg/m3 Pump
100
0 0
R
QCT
0.0001 0.001 0.01 0.1
Particle Diameter (m)
Wellbore
Coiled-Tubing (CT) Operation
Preliminary Considerations. A simple nozzle (bullnose) at-
tached to a 2-in. CT was chosen to perform the reverse cleanout. Fig. 5—Schematic of setup for the CT operation. Individual
Detailed information about the CT can be found in Table 1. The data, observed at different points within the system, are indi-
bullnose was equipped with one central 16-mm hole and four 8- cated. Data indicated in bold characters were available in real
mm holes at an angle of 45 to the axis of the CT. Because the time; all other data were available at time intervals longer than
coating fragments and some of the copper particles found in the 30 minutes. For details, see text. Figure modified from Michel
bailer sample were larger than the holes of the bullnose, a high- et al. (2004).
pressure jetting-tool operation was planned to crush and mobilize
the larger components during a first run and to clean the perfora-
tions of the production liner.
Groundwater from a near-surface well was used as circulation The disposed fluid was a mixture of circulation fluid and for-
fluid for the cleanout. This bicarbonate-rich water was condi- mation fluid (brine). For the system circulation fluid/brine, a sim-
tioned to a pH of 4 by use of 15% hydrochloric acid (HCl) to pre- ple empirical correlation was established in the laboratory to
vent iron carbonate precipitation downhole. Furthermore, the fluid correlate fluid density and electrical conductivity for different volu-
was loaded with potassium chloride (KCl) to obtain a 4% solution metric ratios of fluid and brine (Fig. 6). The data were measured at
to prevent the formation clay minerals from swelling. ambient conditions and fitted to a third-order polynomial function:
Different data were acquired during the CT operation. A sim- q ¼ 0:000011813 r3 0:0021306
plified job setup is displayed in Fig. 5. At the reel, the CT pressure . . . . . . . . . . . . . ð1Þ
r2 þ 0:65769 r þ 996:01:
(pCT), flow rate (QCT), and depth (zBHA) of the bottomhole assem-
bly (BHA) were measured. The CT weight (FCT) was measured at Because electrical conductivity was easier to measure during the
the injector head and the wellhead-pressure ( pW) data were operation than density, the fluid density was calculated according
acquired at the blowout preventer (BOP). Additionally, the pres- to the relationship in Eq. 1.
sure in the annulus (pA) between the production string and the
casing was measured. To understand the hydraulic situation
within the CT during the reverse cleanout, the returned-fluid vol- Field Operation. The different stages of the intervention can be
ume was measured at the fluid disposal. Measurements of electri- summarized as follows:
cal conductivity (r), density (q), solid content (w), and the flow 1. Run in hole (RIH) with rotating high-pressure jetting nozzle.
rate of the disposed fluid ðQRCT Þ were performed. While data from 2. CT weight loss because of contact with fill at 4111 m.
the CT were available in real time throughout the intervention, 3. Reciprocate three times between that depth and 4170 m.
data from the fluid disposal were available only at time intervals 4. Pull out of hole (POOH) and redress to reverse-circulation
exceeding 30 minutes. nozzle.
5. RIH with reverse-circulation nozzle and implement for-
ward circulation until 4121 m.
6. Reverse circulation down to 4149 m, CT blocked, block-
Outer Nominal Minimum age removed by forward circulation.
Diameter Thickness Thickness Segment 7. Reverse circulation down to 4157.5 m.
(in.) (in.) (in.) Length (m)
8. CT blocked, blockage not removable by forward
2.000 0.134 0.128 903.86 circulation.
2.000 0.145 0.138 3828.29 9. POOH, blockage at the nozzle removed at surface.
2.000 0.156 0.148 1920.24 10. RIH and apply forward circulation down to 4165 m.
11. Unstable return-flow rate during reverse circulation at
Table 1—CT specification; thick-walled interval: uphole end; thin- 4100, 4050, 3900, and 3700 m.
walled interval: downhole end. 12. POOH and redress to rotating high-pressure jetting nozzle.
1200 6
5
Density (kg/m³)
4
1120
3
1080 Density of returned fluid
2
Solid-content of returned fluid
1040 0.4
0.2
400 0
Flow rate 4400
300 Depth
4200
200 4000
3800
100
3600
0
Pumped volume
200 Returned volume
Volume (m³)
100
20 CT pressure
Wellhead Pressure
Pressure RR
Pressure (MPa)
15
10
0
0
00
00
00
00
00
00
00
00
:0
0:
8:
5:
3:
:
9:
4:
14
19
13
2
2
2
2
2
12
12
12
/1
/1
/1
/1
12
/1
0/
/2
/1
/1
/2
30
0/
0/
2/
/3
12
12
12
12
/3
/3
/
/
12
11
11
11
11
Time
Fig. 7—Hydraulic situation during CT operation. Top: Density of the returned fluid together with the solid content of the returned
fluid. The marks indicate the different measurements. Upper Middle: Flow rate together with the measured depth of the CT. Lower
Middle: Total pumped- and returned-fluid volume. Bottom: Annular, CT, and wellhead pressures. CT blockages are indicated with
arrows.
was not significantly affected by a change of the straight length. be made for the density of the fluid column within the annulus
The correlation displayed in Fig. 9 shows a good fit of the entire outside the CT. Here, a simple linear mixing law has been
data set. assumed for pumping circulation fluid with an average calculated
density of 1025 kg/m3 and brine with a density of 1180 kg/m3
(Francke et al. 2013).
Validation. To validate the hydraulic model, the density of the
returned fluid was calculated on the basis of a 30-minute average
of the return-flow rate, taking into account the measured solid Results. Fig. 10 shows the fluid density calculated from Eq. 1
content (Eqs. 2 and 4). The density difference between the inside with that evaluated from the simulation described in the preced-
and the outside of the CT, together with wellhead and CT pres- ing. According to the simulation, the correlation between calcu-
sure, determines the pressure difference at zBHA available to gen- lated density and measured density is reasonable for the available
erate a return-flow rate. For the calculation, an assumption has to data. Not only is the absolute value similar, but the trend of lower
(a) pCT
pW
(b) Density (kg/m3) Temperature (°C)
900 1000 1100 1200 40 80 120 160
Wellbore
0
CT
pA pCT
1000
3000
4000
Mud
Δp
5000 Brine
pBH
Fig. 8—(a) Hydraulic model for the simulation of the CT operation. Input pressure data and the density of the fluid column are indi-
cated. (b) Comparison of fluid temperatures and densities at different depths for circulation and formation fluid (after Francke
et al. 2013).
density values on 30 November at 1300 hours and 1 December at higher than expected for brine, although the measured solid con-
1400 hours could also be reproduced. Right after times of forward tent is low, and the CT should be filled with circulation fluid right
circulation, only circulation fluid was filling the CT. Therefore, after periods of forward circulation.
the calculated density is lower until the CT filled with brine dur-
ing reverse circulation. Even the large solid load on 2 December
at approximately 0300 hours can be reproduced by the calcula- Discussion
tion. For later times, however, the calculated density is much Coiled-Tubing (CT) Operation. Michel et al. (2004) provide a
checklist of actions to be taken to perform a reverse-cleanout
operation safely. Although it attempted to meet the required pre-
requisites, the CT operation failed to remove all solids from the
100 wellbore. The operation at the Grob Schönebeck site was opti-
mized for a high return-flow rate within the limits authorized by
the pressure rating of the wellhead. The fluid flow from the CT to
the fluid disposal was not choked, resulting in a pressure reading
Pressure Increase at CT (MPa)
/1 00
00
00
0
:0
:0
13
7:
/1 1:
7:
6:
19
12
is underestimated (see Fig. 7). At certain times, the fluid-disposal
2
12 /12
2
2
2
/1
/1
2
/1
/1
/1
0
/1
/2
0
/2
/3
12
12
/3
11
12
11
11
in physics and geology from University of Cologne and a PhD PhD degree in geophysics from the Technische Universität
degree in petroleum engineering from the Clausthal University Berlin.
of Technology.
Jan Henninges is a research scientist at GFZ German
Simona Regenspurg is a research scientist at GFZ German Research Centre for Geosciences. He holds a diploma
Research Centre for Geosciences. She holds a diploma degree in geology from the University of Tübingen and a PhD
degree in hydrogeology from Ludwig Maximilian University, degree in applied geophysics from the Technische Universität
Munich, and a PhD degree in hydrology from the University of Berlin.
Bayreuth.
Günter Zimmermann is a research scientist at GFZ German
Elvira Feldbusch is a PhD degree student at GFZ German Research Centre for Geosciences and a professor at the Tech-
Research Centre for Geosciences. She holds a diploma nische Universität Berlin. He holds a diploma degree in physics
degree in chemistry from the University of Potsdam. from the University of Bonn and a PhD degree in geophysics
Ali Saadat is research scientist at GFZ German Research from the Technische Universität Berlin.
Centre for Geosciences. He holds a diploma degree and a
Ernst Huenges is head of the International Geothermal
PhD degree in mechanical engineering from the Technische
Research Centre at GFZ German Research Centre for Geo-
Universität Berlin.
sciences and professor at the Technische Universität Berlin. He
Kemal Erbas is a research scientist at GFZ German Research holds a diploma degree in physics and geology and a PhD
Centre for Geosciences. He holds a diploma degree and a degree in rock physics from the University of Bonn.