Professional Documents
Culture Documents
Cement Expansion Shrinkage
Cement Expansion Shrinkage
requested for testing were the standard 101 pcf bentonite Cement formulations used for long term testing are shown
extended cement and 118 pcf normal density cements that are below:
currently being used in our Khuff/pre-Khuff gas wells. Both • 71 PCF - HCS blend + 0.15% bwob (FLA)+ 0.06 gps
systems contained 35% (by weight of cement) silica flour. (dispersant)+ 0.025 gps (retarder) + 0.15% bwob
R&D scientists raised concern over HCS blends’ durability (anti-settling) + 2 pt/10 bbls (non-foamer) @ 8.80
due to the uncertainty of non-cement type materials in the gps total mixing fluid
blend and the possible lack of sufficient temperature stabilized • 101 PCF – Class G (HSR) + 1.9 % bwoc Bentonite
cement binding type material. Test conditions set by drilling (prehydrated) + 35% bwoc (silica flour) + 0.60%
and workover engineering were to test the cement for one year bwoc (FLA) + 0.05 gps (dispersant) + 0.025 gps
at 300 0F and 3,000 psi curing pressure. (retarder) + 2 pt/10 bbls (non-foamer) @ 11.47 gps
total mixing fluid
After discussion with Drilling and Workover Engineering • 118 PCF – Class G (HSR) + 35% bwoc (silica flour)
representatives the following tests were performed (table 1): + 0.60% bwoc (FLA) + 0.10 gps (dispersant) + 0.03
gps (retarder) + 1 pt/10 bbls (non-foamer) @ 6.26 gps
Table 1 total mixing fluid
Physical Properties Test Duration (test periods) Cementing Service Company provided a pre-blended sample
Compressive Strength API Crush 1, 7 days, 1,2,3,6,9 and and a special mixing and pressure conditioning procedure for
12 months HCS cement blend. The components of the sample were not
Permeability to gas 1, 7 days, 1,2,3,6,9 and revealed to R&D Center scientists. The mixing and pressure
12 months conditioning procedures were performed prior to performing
Shrinkage or Expansion Short term any test. Standard API6 testing and conditioning could not be
Young’s Modulus 1, 7 days, 1,2,3,6,9 and applied to test HCS blends for two reasons. First, some of the
12 months hollow spheres break in the blender at high RPMs. Second,
Poisson Ratio 1, 7 days, 1,2,3,6,9 and pressure applied when simulating down-hole hydrostatic and
12 months pump pressure breaks or cracks the hollow spheres. When
Rheology Short term these hollow spheres fail or crack the density increases. This
simultaneously produces an initial volume decrease. A typical
Settling Short term
71 pcf HCS blend should have an initial down-hole density of
75-81 pcf after the 10 minute conditioning time at a curing
Durable cement for long-term oil and gas well applications has
pressure of 4200 psi for a 9,000 feet casing job. The variance
not been fully defined. This is mainly due the complexity of
in slurry density is due to the inconsistency in crush strength
cement chemistry. Water to cement ratio, silica to cement ratio
from batch to batch of hollow ceramic spheres.
and exposure temperatures change the physical properties of
The uncertainty of the components in the HCS blend and the
set cement in various parts of the hole. Durable cement cured
resistance of the service company to provide the individual
at 300 0F and 3,000 psi for this application was set with the
components, led to our investigation of how the HCS blend
following criteria:
would react with respect to long-term high-temperature
• No or very little compressive strength durability.
retrogression after two months.
• No increase in permeability. Compressive Strength
• Minimal shrinkage (less than 2.0%) Samples of all three cements were prepared and API
• Constant or lowering Young’s Modulus with compressive strength cubes were made. The samples were
aging. prepared and placed in a HT/HP curing chamber at 217 0F and
• Increasing or constant Poisson ratio with 3000 psi for ten days. One day and one week samples were
extended curing. removed and the cubes were crushed according to API
specifications. After 10 days the temperature in the curing
Testing for this work required the use of three HP/HT curing chamber cells was raised to 300 0F, simulating the cement
chambers. The three curing chambers were run at 217 0F and heating due to the well flowing.
3000 psi for ten days then heated to 300 0F and 3000 psi non-
stop for one year. Due to the size and number of samples only The early compressive strength of all three cement systems
one specimen was used for each test. All cements were mixed fell into expected values. The compressive strength data of
from the same batch/consignment number. HCS blend has shown that the submitted sample had
unacceptable compressive strength retrogression. The highest
recorded compressive strength was 2,858 psi at seven days.
The compressive strength after one year of curing has reduced
to 532 psi (fig. 1).
SPE 95816 3
1.2
6000
HCS 72 pcf
HCS 72 pcf 0.8
Bentonite 101pcf
Bentonite 101pcf
118 pcf Cement
Cement 118 pcf
3000 0.4
0
0 0 100 200 300 400
0 100 200 300 400 Time (days)
Time (days)
Fig. 3 – Long-term permeability of the samples tested show that
Fig. 1 – API crush strengths of all three systems are shown the 118 pcf cement is superior to either of the lightweight
above. The compressive strength of HCS blend has continuously cements. The long-term increase in the permeability of 101 pcf
declined from its peak strength of 2,858 psi in seven days to 532 bentonite cement and HCS blend was observed.
psi in one year.
Young’s Modulus
Compressive Strength Data Young’s Modulus for all three cements was declining the first
The following figures are data generated from the compressive six months. Young’s Modulus data for 72 pcf HCS blend
strength tester. The graph illustrates the programmed load and generally showed a decreasing trend. The Young’s Modulus
the maximum load versus time. Compressive strength (API for 101 pcf cement increased from 6 month to one year test
Crush) data for HCS blend for one year is shown below (fig. period. One year data for the 118 pcf cement shows a slight
2). decline in Young’s Modulus (fig. 4).
Compressive Strength API Crush of LiteCrete 72 Long Term Young's Modulus
pcf cured one year @ 300 0F / 3,000 psi @ 300 0F, 3,000 psi
1.0E+07
Young's Modulus (psi
3000
2500
Load (lbs)
2000 1.0E+06
1500 HCS 72 pcf
1000 Bentonite 101pcf
500 118 pcf Cement
1.0E+05
0
0:00:00 0:00:09 0:00:17 0:00:26 0:00:35 0:00:43
Time (sec) 1.0E+04
0 100 200 300 400
Load - Prog (Lbf) Load - Maximum (Lbf)
Time (days)
Fig. 2 - Compressive load to failure data for 72 pcf HCS blend Fig. 4 – Long-term Young’s Modulus data of all three systems are
cement for an API cement specimen that was cured at 300 deg F shown above.
and 3,000 psi for 1 year. Surface area for this sample was 3.902
sq. in.
Poisson Ratio
Permeability Poisson Ratio of HCS blend was initially considered good.
Samples were cured in the same curing chamber, at the same After the sample was cured for 9 months the Poisson Ratio
time, with the compressive strength samples. The samples decreased to an unacceptable limit having brittle
were given to the Petrophysics lab for gas permeability characteristics. The one year sample was similar. The 101 and
measurements. The initial permeability of HCS blend was 0.1 118 pcf cement also showed declining Poisson Ratio, the
mD. After the one-year curing period the permeability was values were considerably higher than the HCS blend (fig. 5).
0.77 mD. The initial permeability of the 101 pcf cement was Typical Poisson Ratio for concrete is 0.19.
0.60 mD and for one year the permeability was 1.07 mD. The
118 pcf cement had a initial permeability of 0.02 mD and a
one year permeability of 0.01 mD (fig. 3).
4 SPE 95816
0.38
Poisson Ratio
HCS 72 pcf
0.25
Bentonite 101pcf
118 pcf cement
0.13
0.00
0 100 200 300 400 Fig. 7 – The CEST cell has the same outside diameter as the
Time (days) standard UCA cells.
3000
Corr. Displ. (%)
-1
2000
Fig. 6 – The CEST shown above in horizontal orientation, test the
real time shrinkage or expansion of cement cured at temperatures -2
0 1000
up to 400 F and pressures up to 10,000 psi.
0 -3
The cell is the same diameter as the standard size Ultrasonic
0 3 6 9 12 15 18
Cement Analyzer (UCA). (fig. 7) Time (hrs)
Temperature (°F) Pressure (psi) Displacement
Shrinkage of 118 PCF Cement @ 190 deg F Shrinkage of 71 PCF HGS 5000, 190 0F, 3,700 psi
4000 1 4000 10
3000
Pressure (psi)/
Temp (deg F)
0
Pressure (psi)/
Temp (deg F)
2000 -1 2000
-10
1000 -2 1000
0 -3 0 -20
0.0 3.0 6.0 9.0 12. 15.
0 3 6 9 12 15 18
Time (hrs) Time (hrs)
Temperature (°F) Pressure (psi) Displacement (%) Temperature (°F) Pressure (psi) Displacement (%)
Fig. 9 - Shrinkage of 118 pcf cement was similar to published data Fig. 12 - A request was made from drilling to also test the HGS.
for these types of cements. Similar to HCS, high shrinkage was observed. The majority of
shrinkage occurred after the cement was set.
Shrinkage of 72 pcf HCS - 100 psi / 74 deg F
120 1
The high shrinkage of the cements with both hollow
ceramic and glass spheres is due to the additional breaking of
0
spheres after heat and pressure are applied. Shrinkage was also
Temp (deg F)
80
Press. (psi)/
-1
100 psi with the temperature constant at 210 0F. It is also
postulated that many hollow spheres in the blend have been
40
-2
cracked during the initial mixing and pressurization. When
cement is placed in the well, continued hydrostatic pressure or
0 -3
formation pore pressure along with temperature cause more
0
spheres to fail. This leads to additional void spaces in the
Time - total curing time 572 hours cement matrix. These void spaces in the cement matrix
Temperature (°F) Pressure (psi) Displacement (%) produce a reduction in cement pore pressure, which is ideal for
Fig. 10 HCS cement blend was tested for a total of 572 hours at bulk shrinkage.
room temperature and 100 psi test pressure in the horizontal During the slurry preparation, special care was taken to
position. Continuous shrinkage was noted throughout the test remove the released air from the broken spheres that occurred
period.
during the initial pressurization. This was accomplished by
placing a vacuum on the sample after the initial pressurization.
Shrinkage of 72 pcf HCS, Well A-2, 190 0F The samples were then placed in the CEST with pressurization
4000 10 and increasing temperature curing to simulate well conditions.
The graphs above show that the additional shrinkage occurred
3000 after the cement had reached the final curing temperature. The
Temp (deg F)
0
apparatus used to test the shrinkage does not allow additional
2000
water to come in contact with the cement. Normal lab
-10 procedures allow for additional water to enter the cement
1000
while being cured in a hydraulically pressurized curing
chamber.
0 -20
0 11 22 33 44
Corrected displacement, for the purposes of this work, is
Time (hrs) determined at the time when the cement has been heated to
Temperature (°F) Pressure (psi) Displacement (%)
final curing temperature. At this point in the curing time, the
displacement is corrected to zero displacement. This was
Fig. 11 – Temperature and time were the main parameters that chosen because the expansion of water has reached a
contributed to the shrinkage of this 72 pcf HCS cement blend.
maximum value and the cement has not yet set. Initial
expansion is subtracted out from the total displacement of the
This data led to the investigation of Hollow Glass Spheres
piston from the start of the test till the final temperature time is
(HGS) that were rated for 5,000 psi for use in in lightweight
reached, which is approximately four hours. The movement of
cements. The service company’s HCS blend testing
the floating piston is responsible for determining the shrinkage
procedures were used for testing the HGS. Similar shrinkage
or expansion of cement.
occurred as compared to the HCS blend (fig. 12).
Rheology and Settling
Rheology values for 101 and 118 pcf cements tested were in
the expected ranges. Rheologies of pressurized HCS blend for
well A-1 showed some higher than planned values. The
surface density was planned to be 79 pcf with the planned
6 SPE 95816
down hole density to be 82 pcf. Note the inconsistency of Hussain Al-Shammary who performed many experiments in
density and rheology between the two field samples shown the laboratory in the support of this work.
below (table 2).
Table 2 REFRENCES
Well- Slurry 600 300 200 100 6 3 1. Mukhalalaty, T., Al Suwaidi, A., Shaeen, M., “Increasing
A-1 Density RPM RPM RPM RPM RPM RPM Well Life Cycle By Eliminating The Multistage Cementer
4200 PCF And Utilizing A Light Weight High Performance Slurry,”
psi SPE 53283, Middle East Oil Show held in Manama,
No. 1 82.0 210 127 92 56 17 14 Bahrain 20-23 February, 1999
No. 2 90.5 300+ 238 195 139 37 29
2. Rucker, R., Sauvabgeau, K. and Randolph, T. “A New
Approach to an Old Problem – Lightweight Liquid
Settling tests of HCS blend exhibited some density Concentrate,” SPE 67193, SPE Production and Operations
segregation (four pcf from top to bottom). Settling tests Symposium held in Oklahoma City, Oklahoma, 2001
specimens left in the curing chamber showed significant
density increases. Additional settling tests performed for well- 3. Baret, J. F., Garnier, A., Rashad, B., “High Performance
A-2 had densities as high as 94.3 pcf for a 71 pcf cement Water Reduced Cement Slurries Prepared With Low Cost
mixed at surface. Another HCS blend for well 10 had a high Optimized Blend,” IADC/SPE 35088, 1996
density of 93.2 pcf for 71 pcf cement mixed at surface. This
4. Sorgard, E., Villar, J. P., “Reducing the Environmental
increase in density clearly shows that the hollow ceramic Impact by Replacing Chemistry with Physics,” SPE 66551,
spheres in the HCS blend are failing or being consumed by the 2001
cement reaction with the silica sphere shells. This allows
water and cement crystals to fill the pore spaces that were 5. Revil, P., Jain, B., “A New Approach to Designing High-
once occupied by air. In field situations where no extra water performance Lightweight Cement Slurries for Improved
is allowed to enter the cement matrix, the void spaces from the Zonal Isolation in Challenging Situations,” IADC/SPE
broken spheres allow cement crystalline growth and 47830, 1998
shrinkage.
6. American Petroleum Institute: “API Recommended
Practice 10B,” 22nd Edition, December 1997
Conclusions
1. Compressive strength of HCS blend declined 81% in 7. Chandler Engineering: “Cement Expansion/Shrinkage
11.75 months when cured at 300 0F and 3,000 psi. Cell,” Instruction Manual, Rev. 1, June, 2003
2. Permeability of HCS blend has increased 7.7 fold in one
year. 8. American Petroleum Institute: “Shrinkage and Expansion
3. Young’s modulus of HCS blend has decreased 81% over in Oilwell Cements,” API Technical Report 10TR2, 1st
1 year. Edition, July 1997. PP 17.
4. One year testing of Poisson ratio of HCS blend reveals the
9. Backe, K. R., Lile, O.B., Lyomov, S. K., Elvebakk, H. and
cement has deteriorated to brittle failure.
Skalle, P., “Characterizing Curing-Cement Slurries by
5. Shrinkage of HCS blend after initial setting is Permeability, Tensile Strength and Shrinkage,” SPE
unacceptably high. Drilling & Completion 14 (3), September 1999
6. Normal density (118 pcf) cement was the most durable of
the three cements tested. Nomenclature: