Load History and Buckling of the Production Casing in a Hpht Geothermal Well

8/10/2019 Load History and Buckling of the Production Casing in a Hpht Geothermal Well

1/9

PROCEEDINGS, Thirty-Sixth Workshop on Geothermal Reservoir EngineeringStanford University, Stanford, California, January 31 - February 2, 2011SGP-TR-191

LOAD HISTORY AND BUCKLING OF THE PRODUCTION CASING IN A

HIGH TEMPERATURE GEOTHERMAL WELLGunnar Sklason Kaldal1*, Magns . Jnsson1, Halldr Plsson1, Sigrn N. Karlsdttir2, Inglfur .

orbjrnsson2,3

1Faculty of Industrial Engineering, Mechanical Engineering and Computer Science, University of Iceland,Hjararhagi 2-6, Reykjavik, 107, Iceland

2 Innovation Center Iceland, Department of Materials, Biotechnology and Energy, Keldnaholt, Reykjavik, 112,Iceland

3Reykjavik University, Menntavegur 1, Reykjavik, 101, Iceland*e-mail: [email protected]

ABSTRACTThe production casing of a high temperaturegeothermal well is subjected to multiple thermo-mechanical loads in the period from installation toproduction. Temperature and pressure fluctuationsare large in high temperature geothermal wells, forexample during the first discharge the temperaturedifference from a non-flowing to a flowing well canbe on the range of hundreds of degrees centigrade.During installation, stimulation and production,problems can arise due to these loads and due to apossible corrosive geothermal environment. Plasticbuckling of the production casing is a problem thatcan occur. It results in a bulge in the wall of thecasing and is detrimental to the geothermal energyproduction and the lifetime of the well. The cost ofeach well is very high. Therefore, it is important toanalyze the structural environment of hightemperature geothermal wells in effort to avoidrepeated problems in the design and installmentphases of the casing.A finite-element model has been developed toevaluate the temperature distribution, deformationand stresses in a high temperature geothermal welland to evaluate the reasons for buckling in theproduction casing. The load history of the casing is

followed from the beginning of the installment phaseto the production phase.The results show that the load history and also thesequence of loading is important in order tounderstand the true structural behavior of wells.

INTRODUCTION

Geothermal wells consist of several concentric steelcasings and concrete sealant that is in contact withthe surrounding rock formation. Plastic buckling ofthe production casing is a problem that can occur.

The innermost casing, the production casing, buckles

and forms a bulge on the inside of the casing wall.This deformation of the casing can lead to reducedenergy output and in worst cases render the wellinoperative.A number of interesting cases of casing impairmentshave occurred in Iceland. There exist however somedifficulties in tracking the history of the wells. Forexample the load history of wells is not always fullyknown, as down hole P-T measurements are oftensparse and cannot be performed constantly. Icelandicwell drilling, operation and completion reports, fromthe National Energy Authority and IcelandGeosurvey, were used to gather information and dataon the load history and well completion processes.Casing failure as a result of trapped fluid in thecasing to casing annulus have been discussed as asuspected cause of casing collapse by for exampleBjrnsson (1978), Magneschi (1995) and Southon(2005). Southon lists casing failure modes ingeothermal wells and discusses the importance ofensuring that construction and design techniques aresound and carefully implemented. He also discussesthat pre-tension loads need to be determined to avoidcompression yielding when using buttress threadedcouplings. Euler buckling and helical buckling areaddressed by Leaver (1982) where analyses areperformed and equations developed for buckling ofan uncemented length of casing. Euler buckling isalso addressed by Rechard and Schuler (1983) wherebuckling models are produced.Chiotis and Vrellis (1995) list casing failuresobserved in Greek wells where wellhead movement,casing joint decoupling, buckling of a 9 5/8 in casingin 6 different places, tieback casing collapse andserious wellhead leakages associated with casingcorrosion are discussed. They conclude that the majorcasing failures observed are caused by thermal stress


2/9


3/9

faulty cementing job. Complete collapse of the casingcan occur if the pressure difference between the outerand inner wall exceeds the collapse resistance of thecasing, for example during cementing.High axial tension forces, for instance when negativetemperature change occurs, can lead to couplingfailures and in some instances casing body tear.When the casing depth is large and imperfections arepresent, the weight of an uncemented casing cancause a casing body tear or coupling failure in worstcase scenario. Axial compression, for example due toincrease in temperature, can cause a thread slip in thecoupling area.Corrosion can cause serious production casingfailures. It can be very different between geothermalregions and even different within a geothermalregion, for example from well to well or varying withdepth. For H2S rich environments sulfide stresscracking (SSC) and hydrogen embrittlement cancause problems depending on the material selectionfor the steel casing (Kane 1996). Other forms ofcorrosion, for example uniform corrosion, erosion,and cavitation can exist in geothermal wells. Nogeneral solution for corrosion in geothermal wellsexists and each case should be treated separately.

PRODUCTION CASING LOAD HISTORY

Here the load history of the production casing istracked from the installment phase to the productionphase. The possible load cases considered occur atvarious phases, i.e. the (i) installation of theproduction casing, (ii) stimulation of the well, (iii)discharge of the well, and (iv) production.

Figure 1: Production casing loads.

(i) Installation of the production casingThe discussed load cases are summarized in Table 2.During the installation of the production casing,casing components are screwed together and lowered

down into the well one by one. The first load on thecasing, load case 1, is tensional force due to gravity,see diagram A in Figure 1. The tension increases withincreased depth, putting the highest strain on the lastinstalled casing component that supports the wholecasing before the concrete sets. While the casing isbeing installed, the well is kept full of cold water,which provides a buoyant force.

Table 2: Considered load cases. Loadcase

Description Load

(i) Installation of the production casing1 Casing hanging

from the top of thewell.

Gravity.

2 Cement slurry inplace.

Outer pressure fromcement slurry +pumping pressure.

3 Concrete setting. Temperature increasedue to heat ofhydration and wellsurroundings.

4 Production sectionof the well drilledwith cooling fluid.

Temperature decreasedue to cooling fluid.

(ii) Stimulation of the well5a-i Warm-up period. Temperature

increases.5a-ii Cooling, water is

pumped into thewell.

Temperaturedecreases due to coldwater.

5b Rock fractured withpressurized water.

Temperaturedecreases, pressureinside the well.

5c Fracture cleaningwith acid.

Can cause corrosion ifit comes in contactwith the casing.

5d Rock fracturedlocally by burningrocket-fuel.

No load subjected onproduction casing.

(iii) Discharge of the well6a Water column lifted

with air bubblesthrough drill-string.

Temperature increase,small pressure de-crease,

6b Water columnpushed down andreleased quickly

Rapiddepressurization andtemperature increase

(iv) Production7 Harmful flow

regimes.Local dynamicpressure change.

If centralizers are taken into account as a weightrelieving force, the relieving force has to be roughlyestimated. This is due to friction between thecentralizers and the outer steel casing wall and theformation below the outer casing shoe.


4/9

According to API SPEC 10D,Specification for Bow-Spring Casing Centralizers , the measured startingand running force of a previously run casing shouldbe less than the weight of 40 feet (12.2 m) of mediumweight casing (Mechanical Cementing Products2009). If one centralizer is placed on each threecasing components, then the maximum reduction ofthe load on the top of the casing should be less than1/3 of the casing weight. The pressure at the top ofthe casing cross section then becomes,

P = ( s C W ) g L pr

where s and w is the density of steel and water,g isgravity, L pr is the length of the casing andC is theweight reduction due to the friction between thecentralizers and the outer casing/formation. With thisapproachC has to be estimated but in all casesC should be larger than 2/3.The second load case occurs when the cement slurryis being pumped in place. The concrete is pumpedthrough the drill string, the casing collar and shoe,and up the annulus. The casing is full of water so thestatic pressure difference between the outer and innerwall of the casing is determined by the difference indensity between concrete and water, normally about1.6, see diagram B in Figure 1.When the slurry is pumped in place the outer pressureon the casing must not exceed the collapse resistanceof the casing. Pressure can build up for examplebecause of a blockage in the annulus which can leadto a casing collapse.The third load case deals with the reference

temperature conditions inside the well when theconcrete is setting. Heat of hydration, is releasedwhen cement comes in contact with water because ofthe exothermic chemical reaction in the cement(Portland Cement Association 1997). Temperatureincreases slightly as the concrete cures, a temperatureincrease of 12C of a 300 mm thick curing concretehave been recorded (Portland Cement Association1997). The annulus gap between casing andformation is much thinner so this temperature changecan be considered small compared to the temperatureconditions in a non-flowing geothermal well. Inaddition, when the cement has been placed and the

cooling of the well is stopped, the well heats upslowly due to the hot surroundings. When theconcrete bonds with the steel and solidifies thereference "zero" temperature of the casing-concrete isreached. After the bond between the casing andconcrete is made, the well could heat up slowly dueto the surroundings, but this depends on the rockformation, for example if there are hot fissurespresent.In the fourth load case the production section of thewell is drilled with cooling fluid or mud. This is the

first major cooling of the casing resulting in itscontraction. This leads to tensional forces in thecasing as the concrete reactional forces arecompressive, see diagram D in Figure 1.

(ii) Stimulation of the well

If wells do not perform properly the relationshipbetween the well and the geothermal reservoir needsto be improved with stimulation methods.In load case 5, several stimulation methods arereviewed with regards to load on the productioncasing. Method 5a, where intermittent cold waterinjection is used with periods of thermal recovery, isone of the most common ones used for hightemperature wells in Iceland (Axelsson 2006). In thismethod cracking is caused in the rock with thermalshocking. Cyclic thermal loading and largetemperature changes can cause damage in theproduction casing and the surrounding concrete dueto thermal expansion/contraction of the steel, seediagram C and D in Figure 1.In method 5b pressurized water is used to clean outand fracture already present fissures. This cools downthe well causing contraction of the steel, see diagramD in Figure 1. This can be avoided by using inflatablepackers, where the stimulation can be focused onspecific intervals in the well rather than the wholeopen section (Axelsson 2006).Method 5c involves cleaning out fissures with acid.The acid must not come into direct contact with thesteel because of a possible corrosion risk.Method 5d was used recently in Iceland, whererocket fuel was burned at specified location a hightemperature geothermal well to create a shock wavewhich caused cracking in the rock(Sigursson 2010).This method separates the stimulation process fromthe well section above, minimizing the load on thecasing.

(iii) Discharge of the wellDischarge methods are used if the flow in the welldoes not start automatically when the well is opened.In load case 6, two discharge methods are described.In method 6a, flow is initiated with air that is pumpedthrough the drill-string creating air bubbles thatreduce the density of the water column above. In thismethod, increased temperature is the main load onthe casing as well as the pressure changes fromhydrostatic to flow conditions.In method 6b, air pressure at the wellhead is used topush the water column down into the reservoir. Thenafter some time the pressure is released and the welldischarges quickly. This causes a rapiddepressurization and temperature increase.


5/9

(iv) ProductionWhen the well is in production, harmful flow regimescould result in casing impairment. For example, plugflow can occur when the geothermal fluid boils,which could cause local dynamic pressure changesand cavitation.

FINITE-ELEMENT MODEL

The FE-model is a 3D thermal and structural model.The thermal model calculates the temperaturedistribution, or rather the temperature change fromthe reference cementing conditions. The referenceconditions (or zero condition) for the model is wherethe concrete sets and forms a connection to the steelcasings. The temperature distribution is firstcalculated through all casings, concrete and thesurrounding rock formation. The solution from thethermal model is then used as a load for the structuralmodel.

Figure 2: Element model geometry.

As mentioned before two couplings are included inthe production casing. For simplification, thecouplings are modeled as a solid body with nothreads included. The couplings are included to seethe steel-concrete interaction assuming no thread-slipin the coupling. For better efficiency half of the wellis modeled, which is possible because of symmetry.Three casings are included, a 13 3/8 in (outerdiameter) and 12.2 mm (0.48 in) thick productioncasing, 18 5/8 in and 13 mm (0.51 in) thick securitycasing and a 22 1/2 in also 13 mm (0.51 in) thicksurface casing.Figure 2 shows the model geometry and the includedcoupling. Imperfections are included in the concreteas a small variation in material properties. Thesevariations are shown as yellow elements.

Figure 3: The coupling of the production casing andthe surrounding concrete (transparent).

Material properties are defined separately for thesteel, the concrete and the ground, seeTable 3 . The

reference value for the compressive strength of theconcrete is 27.6 MPa. Stress-strain behavior of K55is used in the model for all three casings. In themodel four different stress-strain curves can be usedfor the steel. Defined steel grades for K55, X56, L80and T95 were obtained from tensile strength tests byKarlsdottir (Karlsdttir 2009).

Table 3: Reference values, material properties usedin the FE-analysis.

Steel Concrete RockYoung's modulus(E)

210 GPa 2,8 GPa 100 GPa

Poisson's ratio () 0.3 0.15 0.31Density () 7850kg/m3

1666kg/m3

2650kg/m3

Thermalconductivity (K)

46W/(mC)

0.81W/(mC)

2W/(mC)

Thermalexpansion ()

12e-6 /C 9e-6 /C 5.4e-6 /C

The bonding characteristics between steel andconcrete are one of the reasons for the numericalcomplexity of the model. During the solution process,the contact between the casing and the concrete isconstantly changing from bonded to sliding tosticking to debonding. This makes the problemextremely complicated and computation timebecomes considerably large.Maximum surface shear strength (Wallevik 2009)between the steel casing and the concrete, beforedebonding occurs, is used in the analysis. When thefriction stress reaches the maximum shear strength,the bond is broken and sliding begins.Furthermore, a maximum normal contact stress isused to control the debonding characteristics.


6/9


7/9

The temperature distribution solution from thethermal model for load case 6a can be seen in Figure5. The temperature distribution is then used as atemperature difference load on the structural model.In Table 5 the maximum von Mises stress is listed forthe production casing and surrounding concrete forall load cases. The maximum radial displacement ofthe casing is displayed in Table 6 and the maximumaxial displacement of the casing is displayed in Table7.

Table 5: Maximum von Mises stress (MPa).Steel Casing Concrete Load

Case Value Location Value Location1 31.2 Coupling

border- -

2 97.0 Nearcoupling

- -

3 160 Couplingborder

2.87 Couplingborder

4 284 Couplingborder

4.96 Couplingborder

5a-i 329 Couplingborder

30.2 Couplingborder

5a-ii 433 Betweencouplings

15.7 Couplingborder

6a 374 Nearcoupling

34.5 Couplingborder

Table 6: Maximum radial displacement of the casing(mm).

Steel Casing LoadCase Value Location1 -0.00757 Near coupling2 -0.0638 Near coupling3 0.136 Near coupling4 -0.567 Casing body5a-i 0.726 Near coupling5a-ii -1.740 Between couplings6a 0.897 Outer radius

Table 7: Maximum axial displacement of the casing(mm).

Steel Casing LoadCase Value Location

1 2.834 At lower end2 2.798 At lower end3 -0.0938 Coupling border (inner)4 0.0583 Coupling border (outer)5a-i -0.689 Coupling border (inner)5a-ii -7.77 Between couplings6a -7.814 Between couplings

From these results it can be seen that the casingsuffers the highest strain when it is cooled downduring the supposed stimulation process in load case

5a-ii. The highest stress in the concrete occurs inwarm-up periods at the coupling borders.It is interesting to see that the highest inward radialdisplacement of the casing occurs during this coolingperiod.

Figure 6: Stress reduction/increase in couplings inload case 6a.

Near the couplings, the stress increases in theconcrete and reduces in the steel couplings, seeFigure 6. Since there are no threads included in thecouplings in the analysis the coupling failures can notbe predicted precisely with this model, but this givesan indication of how the steel and concrete react.Figure 7 shows that debonding of the productioncasing and concrete is progressing and a small gap isbeginning to form, increasing the risk of bucklingnext time the well is heated up.

Figure 7: Contact gap between casing and concretein load case 5a-ii.


8/9

The stimulation method where the rock is fracturedwith cyclic thermal shocking can cause damage to thecasing if the difference in temperature is high and ifthis is done repeatedly.

Buckling

Since buckling did not occur in the load historyanalysis above, a load of a high temperature issubjected on the casing to see at what temperaturebuckling occurs.

Table 8: Maximum stress and displacements - casewithout water pocket.

Steel casing Maximum Value LocationVon Mises stress [MPa] 358 Casing bodyRadial displacement [mm] 2.20 Near couplingAxial displacement [mm] 1.09 Near coupling

Table 9: Maximum stress and displacements at thebuckling point - case with water pocket.Steel casing

Maximum Value LocationVon Mises stress [MPa] 440 At water pocketRadial displacement [mm] -106 At water pocketAxial displacement [mm] 63.7 At water pocket

In the case without the water pocket buckling doesnot occur despite the high temperature change. Theresults show that the casing expands radially pushingup against the concrete and causing no debondingfrom the concrete. The maximum von Mises stress in

the concrete is 35.7 MPa at the coupling boundary.

Figure 8: Buckling, radial displacement (meters).

In the case including the water pocket, bucklingoccurs at 40% of the load, i.e. at about 300C and -2

MPa inside pressure. The maximum stress of thecasing reaches the yield strength of the steel at thebuckling point and the maximum von Mises stress inthe concrete reaches 44 MPa at the boundary of thewater pocket. The radial displacement and thebuckling shape can be seen in Figure 8.

CONCLUSION

A finite-element model was developed to calculatethe stresses in a casing that is subjected to thermo-mechanical loads.The results show that the production casingexperiences a peak in stress when the casing is cooledduring a supposed stimulation process, whereas theconcrete suffers the highest stress during heatingperiods. The stress in the concrete is increased nearthe couplings, whereas the inverse occurs in the steelcouplings.During cooling periods the casing contracts axially

and it is interesting to see how much it contractsradially resulting in debonding between the steel andconcrete. This leads to higher risk of buckling whenthe well heats up again because of reduced supportfrom the concrete, which shows that the load historyand the sequence of load cases is important. Inaddition the load history is important because ofcumulative stresses and plastic strains that occur inthe casing.The results show buckling when a water pocket isincluded in the concrete surrounding the productioncasing, whereas a case without the water pocketshows radial expansion of the casing and nobuckling. This shows that a water pocket that isenclosed in the casing-to-casing annulus clearly hasan effect on the buckling phenomenon.It is clear that further work needs to be done to gain abetter knowledge of how the production casingbehaves as a whole in a high temperature geothermalwell and to gain a better insight into the failuremodes that cause problems. In addition, values for acomplete load history of a real failure case would bepreferred to use as an input in the model. There aremany uncertainties regarding what leads to casingimpairment. It is apparent that a combination offactors are causing casing failures, which couldinclude; imperfections and production flaws incasings, casing thickness deviation, ovality of thecasing, casing centralization, concrete mix properties,quality of the cementing job, and various loadingscenarios. In further work it would also be interestingto compare different casing sizes, the effect ofconcrete gap or water pocket size on the types ofbuckling, as well as different stimulation anddischarge procedures and methods.


9/9

ACKNOWLEDGMENT

This work was supported by the TechnologyDevelopment Fund at RANNIS - The IcelandicCentre for Research, GEORG - Geothermal ResearchGroup and the Innovation Center Iceland. Theirsupport is much appreciated.

REFERENCES

Axelsson, G., Thrhallsson, S., Bjrnsson, G.Stimulation of geothermal wells in basaltic rock inIceland. Kartause Ittingen, Zurich: ENGINE ENhanced Geothermal Innovative Network forEurope, Workshop 3, Switzerland, 2006.Bjrnsson, G., Ragnars, K., Sigfsson, S., Karlsson,. Styrkleiki furrra hhitaborholum (OS JHD7805) (in Icelandic). Reykjavk: Orkustofnun, 1978.Kane, R. D. Evaluation of Geothermal Productionfor Sulfide Stress Cracking and Stress CorrosionCracking. Denver, Colorado: InterCorr/96, 1996.Karlsdttir, S. N., Thorbjornsson, I. O. HighTemperature Geothermal Wells Center ofExcellence in Iceland - Phase I: Corrosion testing ofsteel in high temperature geothermal wells inIceland. Technical Report for RANNIS (TheIcelandic Centre for Research), Reykjavik, Oct.,2009.Leaver, J. D. Failure Mode Analysis for Casing andLiners in Geothermal Production Wells. 1982.Magneschi, P., Bagnoli, C. , Lazzarotto, A. ,

Ricciardulli, R. Structural Models for the Analysisof Stresses in the Casings of Geothermal Wells.World Geothermal Congress, 1995.

Mechanical Cementing Products. Products manual,Houston, Texas: Weatherford International Ltd.,2009.Plmason, G. Jarhitabk (in Icelandic). Reykjavk:slenskar orkurannsknir og Orkustofnun, 2005.Peng, S., Fu, J., Zhang, J. Borehole casing failureanalysis in unconsolidated formations: A case study.

Journal of Petroleum Science and Engineering 59 ,2007: 226-238.

Philippacopoulos, A. J., Berndt, M. L.Characterization and Modeling of Cements forGeothermal Well Casing Remediation. SanFrancisco, California: Geothermal ResourcesCouncil, 2000.Philippacopoulos, A. J., Berndt, M. L. Structualanalysis of geothermal well cements. ElsevierScience Ltd. , 2002.

Portland Cement, Concrete, and Heat of Hydration.Concrete Technology Today (Portland Cement

Association) , nr. Volume 18/Number 2 (July 1997).Rechard, R. P., Schuler, K. W. Euler Buckling ofGeothermal Well Casing. Albuquerque, NewMexico: Sandia National Laboratories, 1983.

Release 11.0 documentation for ANSYS. SAS IP, Inc.,2007.Sigursson, . Njung vi rvun borholna slandi. Frttaveitan, HS-orka newsletter (in

Icelandic) , 2010: 10-11.Southon, J. N. A. Geothermal Well Design,Construction and Failures. Auckland, New Zealand:Proceedings World Geothermal Congress 2005,2005.Vrellis, G., Chiotis, E. Analysis of Casing Failuresof Deep Geothermal Wells in Greece.Geothermics,

Elsevier Science Ltd. , 1995: 695-707.Wallevik, S. O. High Temperature GeothermalWells Center of Excellence in Iceland - Phase III:Rheological behavior and mechanical properties ofcement paste for high temperature geothermal wells.Technical Report for RANNIS (The Icelandic Centrefor Research), Reykjavik, Oct., 2009.

Documents

Load History and Buckling of the Production Casing in a Hpht Geothermal Well