GHAWAR FIELD GAS WELL COMPLETIONS-Conclusion
Saudi Arabian Oil Co. (Saudi Aramco) has completed the first phase of an accelerated project to develop Saudi Arabia's deep gas potential. The company has drilled and completed a large number of new gas producers, with many additional wells planned.
Acid fracturing and hydraulic propped fracturing stimulations have generated significant productivity increase from the Khuff carbonate and Jauf sandstone reservoirs. The company has finalized plans to stimulate most of the existing and future wells in the development.
This article concludes a three-part series on Ghawar field gas well completions. The first article, appearing in the Middle East Update special report, covered Saudi Aramco's efforts to optimize acid fracturing in the Khuff carbonate reservoirs (OGJ, Feb. 18, 2002, p. 69).
Published in OGJ the following week, the second article highlighted the use of diagnostics to evaluate the company's propped fracturing stimulation effectiveness in the Jauf, Prekhuff sandstone reservoir.
This third article features Saudi Aramco's work to equip gas wells with the tubing and wellbore completions to enable high-pressure hydraulic fracturing.
Several of the formations have downhole fracture initiation pressures that average 16,000 psi, with reservoir fluids containing hydrogen sulfide (H2S). It is critically important for wellbore completions to withstand these conditions.
The use of tubing stress analysis software supplied by Landmark Graphics Corp., Houston, known as WellCat, along with rigorous scrutiny of well completion components helped Saudi Aramco to identify a significant number of mechanical weak links that were previously overlooked by computer software developed in house.
The ability to perform multiple sensitivities and quickly evaluate "what if" scenarios was instrumental in providing solutions to complex problems.
Also, the company's ability to analyze well completions in intricate detail and model each fracture stimulation stage with realistic fluid rheology, allowed Saudi Aramco to identify weak links for removal and prevent potential problems.
Through close cooperation between several Saudi Aramco groups and Landmark Graphics, engineers implemented a systematic approach for evaluating existing and future well completions and issued specific recommendations to modify equipment.
Saudi Aramco put into motion an action plan for improving future completions and strengthening existing ones.
The company has performed fracture stimulation treatments in all 15 wells that were worked over to implement completion improvement recommendations; the wells remain trouble-free.
New wells, which were completed with recommended optimization changes, have experienced no completion integrity problems. Three of the six wells that were deemed borderline but not worked over, however, have since developed tubing-casing annular communication.
Several major initiatives are underway at Saudi Aramco further to optimize completions and operating practices in deep gas wells.
Wellbore conditions
The Khuff formation is a carbonate-evaporate sequence, with widely varying reservoir gas composition throughout the Ghawar field, with H2S content ranging 0-9 mole %, carbon dioxide ranging 0.5-4 mole %, and nitrogen ranging 7-14 mole %.
Average reservoir subsea datum depth is 11,000 ft, with average initial reservoir pressure of 7,500 psig and average temperature of 275° F.
The Jauf formation is a thick sandstone containing fibrous illite clays, which can contribute to formation damage and formation sand production.
The Jauf produces sweet gas with high condensate content. The average subsea reservoir datum depth is 13,000 ft, with average reservoir pressure of 8,500 psig and average temperature of 300° F.
The traditional stimulation methods of matrix acidizing and acid washing perforations to remove near-wellbore damage from the Khuff carbonate wells provided, until recently, the required gas rate deliverability.
The new areas of development, however, have shown reduced reservoir quality, which usually renders traditional stimulation methods ineffective. The Jauf reservoir has been found unconsolidated in many areas and fracturing for sand control is required.
Saudi Aramco successfully implemented a number of acid fracs in Khuff wells and hydraulic propped fracs in Jauf wells, proving the technique to be a viable productivity enhancement technique for the development.
This significant success early in the program supported the decision to move forward with a large-scale fracturing campaign covering most of the wells in the development.
Field experience and data, however, have shown that hydraulic fracturing is a challenging and complex task. Measured fracture initiation gradients in the Khuff can be quite high, ranging 1.0-1.4 psi/ft, due to an active tectonic stress environment.
Required treatment pressures for effective fracture growth average 16,000 psi, at reservoir depth, with treatment injection rates of up to 80 bbl/min. Wellhead tubing injection pressures often range 10,000-12,000 psi, which call for annular pressures of up to 7,000 psi for tubing support.
Well completions capable of withstanding such harsh conditions are an essential component of the development program.
Deep gas well completions in the Ghawar field to date are of two basic types: liner top polished-bore receptacles (PBRs) and packer top PBRs. Both completion types incorporate seal assemblies that allow upward movement of the tubing string.
The company has completed the majority of wells in the new development phase with either 41/2 in. or 51/2-in. tubing. Tapered 15.1 and 13.5-lb/ft, C-95 grade, 41/2-in. tubing with new VAM connectors is a common completion. The tubing strings typically incorporate several L-80 material components.
All wellbores include 7-in. or 7 x 41/2-in. cemented liners across the producing reservoirs. Fig. 1 shows typical wellbore completion cross-sections.
As new gas wells were drilled and in light of the present gas demand and development requirements, several factors had signaled to Saudi Aramco engineers that they needed to review past assumptions and completion practices.
These factors were that the company had adopted an hydraulic fracturing stimulation philosophy based on the positive results obtained so far, new wells were encountering higher-than-predicted minimum horizontal stresses in the formations resulting in higher-than-expected frac pressures, and detailed connector performance data had indicated there were problems with existing wells.
Engineers promptly completed reevaluations and issued detailed and specific recommendations for individual wells that had weak links. The company completed a workover campaign in a timely manner, achieving stringent completion-integrity upgrades.
Past completion practices
In the early field development, Ghawar oil producing wells did not require detailed tubing stress analysis. Most wells produced up the casing and had been completed with 27/8 or 23/8-in. kill strings.
The company developed the nonassociated deep gas fields after oil development in the area had been underway for a number of years.
Transferring to the new deep gas development program from the early oil development projects, most of the engineers brought the completion practices and expertise they had accumulated in the area with them.
Based on similar practices for gas wells in other fields around the world, Saudi Aramco engineers designed completions with floating seals inside of PBRs. Rigs had landed the tubing with enough pick up above the PBR so that the seals would float while keeping the locator sub above the PBR.
A short while later, gas well completions included packers with floating seals.
Engineers had estimated a total maximum upward movement of 12 ft, due to temperature change during predicted stimulations, production, and kill job loads.
Completion engineers generally disregarded tubing movement due to piston effects, incorrectly assuming, from lack of industry consensus at the time, that the piston effect forces would be negligible provided the seal and tubing diameters were similar.
Based on this maximum expected tubing movement, the company ordered 24-ft long PBRs for all of the new wells.
In addition, the company had purchased packers with only 10,000-psi rating, under the premise that higher rated equipment was unnecessary because no tubing-to-packer forces would be present with floating seals.
Furthermore, the general view at the time was that wells would only require matrix acid stimulation to achieve gas rate targets.
After Hammerlindl published a paper in 1977, which provided clarity to some of the issues the industry had struggled with, Saudi Aramco's engineers performed more detailed analysis and concluded that their original estimate of maximum tubing movement was optimistic.1
The company could not disregard movement due to piston effects.
Spacing out the tubing with the estimated required pick up would keep the locator sub from touching the PBR under tensile pressure loads but would make the floating seals pull out of a 24-ft long PBR under compressive pressure loads.
At this stage, it became necessary to rethink the original completion philosophy and use a new approach, but it had to be done with the added constraint that all the new downhole equipment had already been delivered and could not be exchanged.
The new approach was to calculate the maximum tubing-length change under compressive pressure loads and then space out the tubing in such a way that the floating seals would remain inside the PBR.
In-house software
The need for performing more detailed tubing stress analysis became evident. Discussions ensued over appropriate evaluation criteria and engineers reached consensus on two main issues.
They agreed that forces acting on tapered strings and tubing to locator sub crossovers could be ignored because it was generally assumed that bending stress would be small due to their short length.
They also agreed to assume that the seals would eventually stick in the sealbore when the locator sub bottomed out on top of the PBR or packer.
This was obviously an oversimplification because the seals could in fact stick in any position.
Engineers also ignored forces acting on the completion section below the locator sub, because compressive loads are insignificant at flowing conditions. They overlooked this significance under stimulation conditions, however.
With no production data available before completing the first wells, it was necessary to make assumptions about temperatures and pressures at predicted stimulation, production, and shut-in conditions.
A senior Saudi Aramco completions engineer wrote a mainframe computer program in the early 1980s to fulfill the need for faster and more detailed tubing stress analysis. The company had used the program until mid-1999.
Based on the assumptions and criteria discussed, all wells met completion-integrity requirements for predicted load cases.
Engineers identified significant program limitations, however, once the company had realized that fracturing would be necessary to achieve targeted gas rates, rendering the computer program obsolete.
The program could not effectively estimate tubing bending stresses nor assess connector strength under compressive loads.
The latter became a significant issue for deep gas completions when, in recent times, tubing-casing annulus communication problems developed with regular frequency in old and new wells alike.
Engineers identified loss of a gas-tight seal in the new VAM connectors, used in most of the wells, as the likely cause.
Extensive consultations with the manufacturer confirmed that the connections were not tested in high compressive-pressure load conditions and, therefore, were only rated to 40% of the pipe's axial strength in compression.
Initial gas well completions did not take this connector compressive-load limitation into account.
The combination of these issues motivated Saudi Aramco to search for alternative computer programs with the ability to perform rigorous tubing stress analysis for a wide variety of load conditions.
Following systematic evaluation of different software packages, the company decided to license the Landmark software.
Simultaneously, Saudi Aramco completed a new testing protocol for specifically qualifying field-proven connection designs to higher compression ratings under combined loading situations.
New approach
With newly acquired ability to perform detailed and thorough tubing stress analysis, company management and technical staff recognized the need to reanalyze existing wells that had previously been found to meet completion integrity requirements.
Saudi Aramco formed a small team consisting of drilling and production engineers and a technical consultant from Landmark. The team was given the task of identifying weak links in existing and future wells and issuing specific recommendations to modify the completions.
The team had to complete its work under a tight deadline because full mobilization for a fracture stimulation campaign had already started and new wells were being drilled and completed at a rapid pace.
The team devised a systematic approach and methodology to ensure rigorous analysis and implemented the following actions:
1. The team reviewed all drilling and completion records meticulously to verify actual completion components and wellbore conditions during tubing space out operations.
Field personnel, on location during completion operations, provided their account of events when reporting discrepancies emerged so that actual vs. reported events could be reconciled.
2. The team contacted vendors, service companies, and contractors and asked them to provide detailed specifications for completion components they had supplied.
The team required component type, grade, weight, length, IDs and ODs, pipe and connector burst, collapse, and axial load ratings, temperature rating, tubing-to-packer load ratings, thread type and rating, and other vendor-specific pertinent information.
3. The team met with the new VAM connector manufacturers to obtain connector performance details for different load scenarios and conditions.
Engineers established the sequence of load conditions, which would cause a connector to lose tight-gas seal, and generated Von-Mises (VME) envelopes to identify key operating constraints.
4. The team met with reservoir engineers and technical staff involved with hydraulic fracturing stimulation design to discuss implementation procedures, treatment objectives, expected maximum injection pressures, treatment volumes, number of stages, flowback procedures, etc.
Engineers evaluated designs and results of all previous stimulation treatments pumped in existing deep gas wells to identify common trends and ranges. They obtained stimulation designs for wells already scheduled for treatment and used historical treatment data to develop load cases for every frac stage.
5. The team analyzed the production history and fluid PVT properties from existing deep gas wells and built a hydrocarbon composition database including fluids present in the Ghawar field.
6. Service companies provided detailed rheology data for the fluids in each treatment stage, at expected temperature and injection-rate conditions during treatments.
Comparison runs of predicted vs. actual pressure data collected during previous treatments and made for calibration purposes showed very close pressure and temperature matches.
7. For modeling purposes, team members split each completion into as many components as necessary, so that axial loads and stress effects could be observed clearly and weak links identified.
They split the tubing into several sections according to weight, material, and connector type, identifying individual components such as, nipples, flow couplings, crossovers, locator subs, spacer bars, seal units, pup joints, etc.
Table 1 shows how a typical completion was defined and modeled.
8. The team reached consensus on the new software modeling criteria and load cases to evaluate. They agreed on:
- Required safety factors.
- Ratings for new VAM and other connector types.
- Buckling constraints.
- Axial load transmittal in different completion components.
- Forecasted gas production.
- Condensate and water rate ranges.
- Maximum operating pressure drawdown.
- Maximum treatment injection and annular pressures.
- Flowing and shut-in wellhead pressures.
- Temperature ranges.
- Annular pressure at flowing and shut-in conditions.
9. The team agreed on packer-depth criteria for different completion types and load cases (Fig. 1).
For packerless or liner top completions, engineers assumed packer depth was at the bottom of the seal assembly for both stimulation and production load cases, which was based on how tensional and compressive forces are transmitted throughout the completion.
For packer top PBR completions, engineers assumed the packer depth was at the top of the packer for stimulation load cases and at the bottom of the locator sub for production load cases.
10. Engineers included buckling constraints into the modeling, in an attempt to emulate actual tubing movement conditions, particularly under compressive pressure loads.
When exposed to compressive pressure loads, a section of the sealbore assembly pulls out of the PBR but another section remains inside. Under this scenario, the section outside the PBR experiences significant bending stress, whereas the section inside does not since there is only a small cross-sectional area change.
The length of the exposed section varies as a function of axial load, so that engineers assumed the worst-case scenario conditions for modeling purposes, when maximum movement occurs.
The software subtracts the maximum movement length from the sealbore assembly length and applies a buckling constraint to the length differential. Equation 1 calculates the buckling restriction ID, which the computer software requires as an entry into the program (see accompanying box).
Engineers also apply a buckling constraint to the locator sub because they do not expect bending stress to affect it, being a sturdy, short component, and having large OD.
11. The team selected design factors (Table 2) for various aspects of the completion modeling. The software accounted for the connector compression deratings as maximum axial load values in the proprietary connections spreadsheet.
12. The team completed Saudi Aramco's testing protocol to qualify several connection designs for higher compression ratings.
Analysis, results
The team's actions established the ground rules for rigorous tubing-stress analysis, reducing the potential for obtaining different interpretations of the new computer software modeling results.
The ground rules simplified the process of evaluating the model's results and issuing recommendations for individual wells.
The company deemed that wells deviating from agreed modeling criteria lacked the required wellbore completion integrity to withstand expected loading conditions.
The team modeled loading cases, which fell into the following categories:
- Hydraulic acid fracturing loads.
- Hydraulic proppant fracturing loads.
- Matrix acid treatment loads.
- Short-term production loads.
- Long-term production loads.
- Cold, shut-in loads.
- Hot, shut-in loads.
- Pressure testing loads.
Engineers modeled each stage for typical stimulation treatments as a separate case with fluid volumes, fluid rheology, proppant concentration, and injection rates equivalent to those they expected to be pumped in each of the wells evaluated.
Engineers evaluated matrix acid loads that were similar to treatment designs, pumped previously in wells with similar characteristics to those being evaluated.
The team also modeled screen-out load cases for propped hydraulic fracture treatments and based production, shut-in, and pressure testing loads on Saudi Aramco's operating philosophy, requirements, and constraints.
The team scrutinized and modeled 21 existing and completed wells in detail, consistently identifying a number of weak links in most of the wells, which rendered their completions incapable of meeting the required completion integrity.
As weak links, they identified new VAM connectors, locator subs, spacer bars, crossovers, and seal units, most of which were L-80 material components.
The modeling work predicted that the failure mode in the overwhelming majority of cases, was loss of the tubing connector's gas tight seal in loads exceeding the connector's compression load rating (Fig. 2).
This condition is likely to cause tubing-casing annular communication problems once the connections are exposed to a sequence of tensile and compressive pressure loads.
Pumping stimulation treatments at the high injection pressures and rates required to initiate fractures, would have exposed weak links that were identified in several of the scrutinized wells to axial loads in excess of the triaxial stress limit (Fig. 3).
High axial loads generated by bending forces acting on short lengths were common in most of the completions evaluated.
This became a relevant factor because some of the historical completion design decisions made were based on the assumption that bending stresses were low in short completion components.
Fig. 4 shows a good example of instantaneous doglegs caused by bending, in a completion that was evaluated. The dramatic shifts in axial load occur as the model shifts from one component in the completion to the next and as the seals exit the seal housing.
Actions required
The team issued recommendations to work over 15 of the 21 wells they evaluated.
Engineers classified the remaining wells as borderline, because the completions either failed to meet the established integrity criteria for screen-out loads during propped fracture stimulations or exceeded the connector axial-load rating only slightly.
Team members specified required changes for each well in detail and provided several options where possible.
Most of the changes recommended specified that existing connectors be replaced with connectors of higher compressive load rating and new completion components with higher axial compressive load rating material be installed.
Saudi Aramco promptly initiated a workover campaign because a quick turnaround was required to meet a tight schedule for delivery of gas rate targets.
Team recommendations called for replacing all L-80 material components with C-95 and higher, but it soon became apparent that manufacturing time would take several months, and other recommended changes would not be possible.
Team members performed additional and extensive computer modeling in search of alternative options that could be implemented quickly while ensuring that completion integrity requirements were met.
The results indicated that a few simple completion modifications would go a long way in meeting the requirements, which were quickly implemented.
Some of these modifications include:
- The company installed new, locally manufactured locator subs, with P-110 material and connectors rated at axial loads of 100% of the pipe body yield strength, in the majority of wells that were worked over.
- Saudi Aramco installed new, locally manufactured spacer bars with C-95 material and 100% load rated connectors, in most of the wells worked over.
- The company added a short section of P-110 pipe below the spacer bar in completions where the rig could not replace the seal assembly nor rethread the connectors.
The objective was to ensure the spacer bar had higher rated connectors at both ends, thus turning the short section of pipe below it into the weak link.
Tubing movement calculations showed, however, that this short pipe section should remain inside the PBR when under compressive pressure loads, and the combination of being exposed to low bending forces and having a high load rating ensured that all completion integrity requirements were met.
Also, thanks to the changes described it no longer became necessary to modify the seal units, which was another component identified as a weak link in most wells.
Fig. 5 and 6 show the results of computer model runs highlighting the effect of loads in one of the weak completion components before and after changes were implemented.
- Saudi Aramco's machine shop rethreaded thousands of feet of pipe with 100% load rated connectors, which was run and installed in wells during workovers.
Acknowledgments
he author thanks Allen Blanke for providing historical perspective and Steve Smith for information on the testing protocol for qualifying connectors to higher ratings.
Reference
- Hammerlind, D.J., "Movement, Forces, and Stresses in Packers," Journal of Petroleum Technology, February 1977.
The author
J. Ricardo Solares is a petroleum-engineering specialist with Saudi Aramco in Udhailiyah, Saudi Arabia. He is currently involved with the massive propped hydraulic and acid fracturing programs for the Ghawar field gas development projects. He has 18 years of diversified oil industry experience, with reservoir and production engineering assignments in Middle East, Gulf of Mexico, Alaska, and South America. He joined ARCO Oil & Gas Co. in 1984 and moved to BP PLC in 1989, where he worked until joining Saudi Aramco in 1999. He holds a BS in geological engineering and MS in petroleum engineering from the University of Texas in Austin and an MBA in finance from Alaska Pacific University, Anchorage.
