Select topics and applications of probabilistic OCTG design - oil country tubular goods

Casing failure causes, focusing on operational problems and examples of successful applications, highlight need for different approach and potential benefits of non-conventional tubular design technology

Probabilistic oil country tubular goods (OCTG) design methods have been used successfully by ARCO on a "special project" basis for more than eight years. Specific instances of probabilistic design implementation are presented [TABULAR DATA FOR TABLE 1 OMITTED] here to illustrate the success of these approaches. The case histories summarize methods used to design both collapse- and burst-driven tubulars.

ARCO's approach to probabilistic design is primarily based on improved understanding and accurate quantification of material and dimensional tubular properties. The approach also includes increased scrutiny of assumed "maximum load" field conditions. Additional focus is placed on understanding the root causes behind tubular problems and failures and addressing them with appropriate remedial measures.

Through an integrated implementation approach, design savings achieved by more sophisticated characterization of pipe performance properties and field loads, and increased reliability by recognizing and dealing with the responsible failure causes, are simultaneously achievable. Efforts and specific progress by the manufacturing community relative to this program are overviewed. And finally, ongoing and future work to obtain supporting data for use in probabilistic design is described.

INTRODUCTION

ARCO presented probabilistic OCTG design concepts in a conceptual technical publication in 1989.[1] Although this publication is frequently referenced as one of the early papers in the field, background investigations have found fundamental papers on probabilistic tubular design issues published as early as 1945.[2] Clearly, the theory behind this design approach is therefore not only not new, but substantially mature.

Currently, focus needs to be placed on the key elements which will enable this technology to be broadly and efficiently used within the industry. Despite the long history of the subject matter in the literature, our industry is not [TABULAR DATA FOR TABLE 2 OMITTED] optimally poised to efficiently apply the technology. To focus on this issue, a number of implementations will be highlighted herein, based on applications within ARCO.

An important conclusion of the 1989 paper[1] was that conservatism is inherent to the design of tubulars since they are frequently designed for severe, and sometimes unlikely or impossible-to-occur loads. Designers thus use overly conservative resistance ratings. Initiatives now being pursued by the industry are not and should not be focused on removing that conservatism, but only on eliminating excessive conservatism so that pipe designs remain fit for purpose and are also cost effective.

TUBULAR FAILURE REVIEW

This section overviews specific tubular problems and failures incurred by ARCO in recent years. As shown in Table 1, review of the failures illustrates that the loading modes addressed by conventional design methods for burst, collapse, tension and yielding are not the predominant modes of failure. In fact, in all of ARCO's tubular failures over the recent years, only one failure occurred in a simple loading mode, and this failure (collapse) was the result of a field oversight to bleed off annular pressure.

Rather, failures have resulted from breakdowns in operating procedures and practices, or as a result of a combination of factors, each of which could have been individually controlled. These facts form the foundation of the premise that, through proper planning and execution, tubular designs can be refined and tubular-well reliabilities can be increased simultaneously.

Operation-related failures. As shown in Table 1, the following categories of problems have contributed to tubular problems and failures in recent years at ARCO:

* Casing wear due to tungsten carbide tool joint hardfacing in combination with rotating hours and significant dog legs

* Downhole connection back-out due to heavy drillstring-casing interaction

* Connection fatigue due to inadequate support and drillstring-casing interaction

* Connection leaks due to inadequate make-up

* Connection failures due to inadequate flush or high-clearance connections in the mid-1980s

* Connection jump-outs due to inadequate make-up or cross-threading

* Connection galling and/or drift problems due to excessive make-up.

Inadequate cementing. Recently, two ARCO wells in the Continental U.S. experienced 8 5/8-in. casing failures due to connection back-out. In both wells, the failures occurred due to back-out of the mill-end connection from the coupling.

In the first well, the 8 5/8-in. surface casing was cemented, the casing was tested, and the intermediate 7 7/8-in. hole was drilled to about 7,600 ft when problems were noted with the well fluids. Drilling was stopped, then resumed after a bridge plug was set. When the problems persisted, an impression block was run. It was noted that the casing had parted at about 4,825 ft in an uncemented interval. The impression block also indicated that the mill-end connection had backed off, and a pin connection was looking up.

A similar set of events during drilling of the 7 7/8-in. hole led to discovery of a parted casing in the second well. As in the previous well, the parted casing was in an uncemented interval at similar depths. Both failures were a consequence of significant casing-drillstring interaction.

This mode of failure in two different wells in nearby fields, is caused by a section of poor cement in the surface hole, possibly causing casing buckling. Clearly, conventional design approaches do not address this type of failure.

Connection fatigue. Another failure mode from casing-drillstring interaction is the fatigue of connections. Payne, et. al. discuss four cases of connection fatigue and their causes.[4] Based on a review of earlier work[5,6] on the mechanics of load transfer in connections, the following conclusions were made:

* Fatigue failure of 8-round connections occurs when there is casing-drillstring interaction in uncemented sections of the well.

* Fatigue failures can occur in any grade of pipe, and in both STC and LTC versions of the API 8-Round threadform.

* Casing-drillstring interaction can be caused by drilling in buckled casing, drilling with air or gas without the beneficial vibration damping of a drilling mud, and misalignment of rotary equipment.

Buckling, wear. Without belaboring the details of all the failures listed in Table 1, it is sufficient to note that an overwhelming number of failures have occurred due to problems with connections or due to a breakdown in applying sound drilling practices. A significant contributing factor to several of the failure modes is drilling inside buckled casing. Buckling can always be eliminated or minimized through use of additional landing tension, higher cement tops, application of internal pressure while waiting on cement (WOC), or a combination of these measures. Ideally, landing procedures should be engineered for all drilling casings, for all wells in the industry. Unfortunately, this is not the case and failures result from time to time.

Wear failures, whereby the casing is simply worn away by the drillstring, continue, despite substantial evidence in recent years linking those failures to tungsten carbide hardfacing. Despite this, operators still tolerate contractors providing rigs with tungsten carbide hardfaced drillstrings. Several operators, including ARCO, have switched to raised chromium-based hardfacings, to which no known casing wear failures can be attributed. However, in a historical sense, use of chromium-based hardfacings is just beginning in the industry, and tungsten carbide hardfacing remains common.

Currently, casing wear, connection problems, and buckled [TABULAR DATA FOR TABLE 3 OMITTED] casing are, by far, the dominant drivers of OCTG failures. Hence, they are also the true drivers of tubular reliability in the industry.

It is critical to clearly understand the modes of tubular failure in the industry. By doing so, proper remedial measures can be pursued. Additionally, by realizing that the loading modes examined by classic OCTG design are not the critical performance modes, i.e., those leading to failures, the excessive conservatism carried along in conventional design practices can be more rationally eliminated through probabilistic approaches.

EXAMPLES OF SUCCESSFUL DESIGN APPLICATIONS

To illustrate benefits and methods of applying probabilistic design technologies, this section reviews several successful applications in a variety of operations.

Collapse-driven design of production liner. Described here are probabilistic methods used to design the 7-in. production liner and tubing for a major well construction project in the South China Sea.[7] The liner design required resisting a collapse load of 7,020 psi at 13,500 ft MD. The choice of 7-in, 29-ppf, L80-13Cr resulted in a "marginal" collapse design factor of 1.0 for this application using standard API performance properties. The next higher weight, 7-in., 32-ppf, L80-13Cr provided a higher design factor of 1.22, but adversely affected completion/production hydraulics.

Additionally, OCTG cost is a strong function of weight, and 13Cr is several times more expensive than alloy steels. In previous operations, such collapse problems have been solved using proprietary products and/or higher yield strengths. However, increasing yield strength range was not desirable here due to the possible presence of sour gas. Therefore, a higher collapse resistance had to be based on other methods.

This led to a focused examination of API collapse ratings accuracy and also the effect of improved dimensional/mechanical properties on performance properties. Past studies evaluated the conservatism of the API ratings and demonstrated the margin of safety between rating and actual collapse pressure of the tubular.[8] Major pipe mills have also been evaluated for their capabilities to produce higher quality pipe which justifies performance properties above API minimum levels.

A program to determine a proprietary collapse rating for the 7-in., 29-ppf, L80-13Cr product from a specific worldclass mill was therefore initiated. Twelve full-length joints were selected at random from three separate heats of initial manufacturing footage. Four collapse test specimens were cut from each of the 12 joints, generating 48 collapse test specimens.

Table 4. Conventional design factors for 5 1/2-in. production casing
to withstand frac job

Type csg., in., ppf    Burst    Tension    VME(*)

5 1/2, 17, N-80        1.55       3.64      1.68
5 1/2, 17, J-55        1.06       2.89      1.16
5 1/2, 15.5, J-55      0.96       2.65      1.05

* Von Mises Equivalent Stress

Table 2 and Fig. 1 show the results of collapse testing these specimens. The symbols T1, T2, B3 and B4 refer to top and bottom samples cut from the mother joint. The mean collapse failure for the test program was 9,114 psi. Collapse pressures ranged from 8,748 to 9,602 psi, yielding a variation of 854 psi. Based on overall ratings, average data set value ([Mu]) and standard data set deviation ([Sigma]) for collapse are 9,114 psi and 222 psi, respectively. Using [Mu]-2.574[Sigma] as the criterion for collapse (as per API 5C3) yields a rating of 8,542 psi, and a design factor of 1.22.

The 20+% margin between anticipated collapse load of 7,020 psi and the test rating of 8,542 psi implied that manufacturing processes and quality control procedures for this product resulted in a consistent and high quality product, for which higher performance properties were justified. The 7-in., 29-ppf, L-80-13Cr design for the production liners and tubing was finalized using this proprietary rating, and substantial cost savings were realized.

Burst-driven design of production tieback. This case illustrates a re-qualification of a production tieback in an HPHT exploration well in a remote international location. The tieback and other casing strings were designed based on conditions encountered in an earlier offset well. The 7-in., 38-ppf, T-95 tieback was designed to withstand a shut-in surface pressure of 10,800 psi. The API minimum internal yield pressure (MIYP) for this tubular is 12,280 psi, resulting in a conventional burst design factor of 1.14. When the well was drilled, it was found that the pressure at TD was higher than anticipated. Therefore, a re-qualification program for the tieback was initiated.

The 7-in. tubulars had been initially manufactured with a specified minimum wall thickness of 95% of the nominal wall, instead of the API standard 87.5% minimum wall. To facilitate re-qualification, the manufacturer reviewed documentation on the order and verified that virtually all joints actually met 100% of the specified nominal wall. The entire order was fully traceable with unique joint numbers on all mill certificates and ultrasonic test (UT) inspection charts. The small number of joints with minimum walls below 100% of the specified nominal were identified by serial number and culled, so that the entire final string provided 100% nominal wall.

Simultaneously, 5% of the joints were sampled to determine yield strength statistics. Test samples were cut from the pin ends of randomly selected joints. The pins on these joints were subsequently re-cut. The mechanical property tests were expedited and results were statistically analyzed. Distributions describing dimensional/mechanical properties were combined to statistically calculate Barlow Minimum Internal Yield Pressure, collapse resistance, and pipe body yield and tensile strengths. Improvements above the API values were then reviewed.

Results of the statistical analysis are shown in Table 3. These indicate a 36% margin between API MIYP and rupture pressure. Margins between yield loads and calculated limit loads are also substantial, i.e., 15.3% (1,067 to 1,230 kips) for axial tension, and 13.3% (16,280 to 18,930 psi) for internal pressure. With the re-qualification successfully executed, the exploration program proceeded without interruption with full assurance of fitness for purpose of the production tieback for the more severe conditions encountered by the well.

Burst-driven design of production casing. This section describes the evaluation of production casing in wells planned for a gas field in which the production zone is hydraulically fractured. The wells are shallow and completed without a packer. On reaching TD at about 4,000 ft, the production casing is run, cemented and perforated.

The well is fractured with an 8.3-ppg fluid and the production casing must withstand loads caused by the maximum applied surface pressure (MASP) of 5,000 psi. During the frac job, the casing is subjected to string weight, forces due to pressure, forces due to changes in temperature, and ballooning forces. Based on an analysis of these forces, 5.5-in., 17-ppf, N-80 grade casing was initially recommended for the production casing string.

Table 4 shows the burst, tension and Von Mises Ellipse (VME) design factors for that tubular and two other choices, 5.5-in., 17-ppf (lower grade) and 5.5-in., 15.5-ppf, J-55 (lower weight and lower grade). The design factors shown are based on resistances calculated with nominal values of geometrical and material properties. The burst design factor is the ratio of the API MIYP and the MASP (in this case 5,000 psi). The tension design factor is the ratio of the maximum tensile load (which accounts for forces due to temperature changes, pressures and ballooning) at the top of the production string and nominal yield load. The VME design factor is based on comparing the maximum VME stress on the inner pipe surface with nominal yield.

Table 5 shows mean and standard deviation of wall thickness, yield strength and tensile strength for 5 1/2-in, 17-ppf and 15.5-ppf, J-55 casing. Mean and standard deviations were calculated directly from mill data. Based on tubular property statistics, distributions for the burst ratings according to Barlow, Lame and Plastic Failure Pressure (PFP)[9] criteria were calculated as presented in Table 5. The table also calculates the [Mu]-2.574[Sigma] value for each of these burst ratings. The resulting design factors indicate the adequacy of the 17-ppf tubular based on this "uniaxial" analysis of the burst load.

Since the frac job creates the worst load on the casing at the surface, effect of tension on burst rating should also be considered. Fig. 2 illustrates distribution of burst ratings [TABULAR DATA FOR TABLE 5 OMITTED] when the tubular is subjected to tension of 50% nominal yield load. The burst pressure is based on the Von Mises stress calculation on the tubular's inner surface in the presence of tension. Finally, Table 5 summarizes the distribution in the presence of tension, along with the internal pressure, for 15.5-ppf and 17-ppf tubulars. Fig. 2 and Table 5 indicate that, based on the actual mill properties and statistical analyses, the J-55 grade casing can withstand the maximum surface pressure of 5,000 psi.

Casing corrosion or wear. This example considers the effect of casing corrosion on a tubular's burst resistance. A well had to be re-entered through the production casing and re-drilled to a new sidetrack target. The re-entry procedure required the production casing to withstand a MASP of 3,500 psi. The MASP was based on gravel packing requirements and was therefore adequately controlled.

Since the well had been abandoned for several years, a caliper survey was performed to evaluate the 7-in., 29-ppf, N-80 production casing string; a total 226 joints were surveyed over a 9,650-ft MD. The survey indicated that portions of the casing had experienced as much as 50% wall reduction. Fig. 3 shows that burst ratings according to different criteria predict approximately similar values for thin walls. In this case, the figures indicate that 4,500 psi is the approximate internal pressure capacity of a casing joint that has experienced 50% reduction in its wall thickness.

Table 6 shows mean and standard burst rating deviation by assuming that thickness of the production casing is equal to the minimum thickness measured in the caliper survey and its standard deviation is zero. The yield strength statistics are based on mill data. The figures indicate that the probability of the burst pressure of a single joint of the casing being less than 3,500 psi is very small.

However, the production casing string comprises 226 joints, several of which have experienced significant wall loss. Clinedinst[2] and Bradley[10] have shown that the probability of failure of a string increases with the number of joints. By assuming that failure of a given number of joints in a string can be modeled as a binomial distribution, Bradley shows that the probability of string failure is given 1 - [(1 -[Rho]).sup.n], where ([Rho]) is the probability of failure of a single joint of casing at the pressure in question and (n) denotes the number of joints in the string.

Fig. 4 shows probability of failure of the production string in this well for different pressures, assuming all joints experienced the same wall loss as the worst joint. The figure illustrates that the pipe is almost certain to "fail," i.e., exceed the rating basis, when pressure is greater than 5,000 psi, and that there is very little chance of failure for pressures below 4,000 psi.

Table 6. Pipe properties and burst ratings for corroded 7-in., 29
ppf, N-80 production casing

                  Nom.      Mean       SD

OD, in.           7.000     7.000    0.000
WT, in.           0.408     0.208    0.000
Yield, psi       80,000    87,200    2,320
Barlow, psi       9,326     5,182     138
Lame, psi         9,865     5,336     142
PFP, psi         10,205     5,422     144
API MIYP, psi       -       8,160      -

Note that the probabilities of string failure (or absence of it) is driven by the tails of the distributions. Therefore, the probabilities of failure or success must still be treated fairly qualitatively due to limited accuracy of the tail characterization; and probabilities should be interpreted to mean either very low or high failure probabilities. The exact values of the failure probabilities are not necessarily important here, since the aim of the analysis is to determine the ability of the production casing to withstand the MASP of 3,500 psi.

The above sections highlight successful applications of probabilistic design techniques for both burst- and collapse-driven designs on new wells, as well as re-certification or assessment efforts based on situations arising from new information concerning operational or tubular conditions.

MANUFACTURING TECHNOLOGY IMPROVEMENTS

Many advances have been made by the OCTG manufacturing community with regard to consistency/quality of both mechanical and dimensional tubular product properties. In many cases, these advances justify use of elevated performance properties relative to standard API performance properties. As an example, Fig. 5 shows a schematic of the principal manufacturing stages in seamless mandrel and plug mills. Process control systems are an integral part of current manufacturing plants at these key stages.

Process control advances. Since dimensional/ mechanical tubular properties are strong functions of the manufacturing processes, it is critical to understand the process control mechanisms which contribute to manufacturing reliability. Significant improvements in the quality of tubulars have resulted from advances in steel manufacturing; these include:

* Improved chemistry control during the casting stages, e.g., use of argon steering and calcium injection to enhance steel cleanliness. Here "clean" steel refers to superior grain structure and absence of defects, voids and impurities.

* Advances in heat treatment: Improved furnaces provide manufacturers with the ability to obtain and track temperatures more accurately, and thus yield better material properties. Improved furnace designs use "walking beam" or "conveyor-driven" systems to ensure uniform heating of the tubes as they move through the furnace. New furnace types include computer-controlled baffling systems to evenly distribute heat throughout the oven. These computerized enhancements ensure adequate soak times for austenitizing/tempering. This is a significant improvement over batchtype furnaces, which can suffer from poor temperature control and non-uniform heat distribution.

* Improvements in post heat treatment processes such as "hot rotary straightening" have contributed to reduced residual stresses in tubulars. This is an improvement from the common practice of straightening, which can induce residual stresses and subsequently result in reduced mechanical properties.

Improved final product inspection. After final heat treating, manufacturers perform mechanical testing to determine yield and tensile strengths. Many manufacturers now have in-house collapse testing capabilities and offer proprietary "high-collapse" products. Changes in API requirements now specify that manufacturers measure coupling toughness for almost all grades, using the Charpy impact test. Generally, properties in excess of minimum API requirements are recorded in all of these tests.

Table 7. Pipe body inspection methods

Grade                 Visual    EMI    UT         MPI
                                             (Circ. field)

H-40, J-55, K-55,        R       N     N           N
N-80 (N,N & T)

N-80 (Q & T), L-80,      R       A     A           A
C-95

P-110                    R       A     A           -

C-90, T-95, Q-125        R       B     C           B

R = Required as specified in API 5CT.

N = Not required.

A = One method or any combination of methods shall be used.

- = Not applicable.

B = At least one method (excluding visual method) shall be used in
addition to UT to inspect outside surface.

C = UT shall be used to inspect outside surface.

Other methods used to verify mechanical properties include hydrostatic testing. Depending on the grade of the material, API requires the product to be tested to a maximum 10,000 psi, or 80% of the minimum yield strength, whichever is less. In some manufacturing facilities, agreements have been established to perform hydrostatic testing above API-recommended test pressure.

Manufacturers who produce API tubulars are required to perform several types of inspections to ensure that the tube body is free of defects; the required inspections are described in Table 7. Several manufacturers have instituted additional quality programs to locate and eliminate inherent defects in tubulars. For example:

* Full-length eddy current testing is performed in conjunction with visual inspection to search for visually undetectable defects on the OD.

* Gamma Ray wall thickness inspection is normally combined with automated electromagnetic inspection (EMI) systems. This provides a full length measurement of wall thickness, and locates thinning and eccentricity.

* Manufacturers of ERW tubulars ultrasonically inspect the seam in search of defects. In most cases, this type of inspection is performed immediately after welding and again after final heat treatment and hydrostatic testing.

* Many manufacturers have capability in their quality system to perform a full-length ultrasonic inspection instead of the electromagnetic inspection. This is normally provided when the tubes have a body wall thickness greater than a specific thickness ([greater than]0.450 in.).

Current manufacturing/inspection technologies must be more carefully accounted for and taken advantage of in design and use of OCTG. As illustrated in the above case studies, substantial margins for design optimization can be pursued when actual pipe properties are defined, as opposed to designing with assumed "minimum" properties. A number of significant advances in steel/pipe manufacturing and inspection justify use of elevated performance properties.

CONCLUDING STATEMENTS, FUTURE WORK

The availability of supporting data on geometrical/mechanical properties of the tubulars is fundamental to reliability-based design. One barrier to efficient implementation of probabilistic design techniques is the ability to obtain that information from manufacturers in a cost-effective manner. A recent joint-industry survey of data acquisition capabilities of worldwide casing manufacturers revealed that a majority are not yet capable of furnishing statistical data on the dimensional/mechanical properties of their products.[8] API standards set minimum requirements and some testing frequencies, but testing frequency remains under study, and API minimum requirements are inadequate to support these new design technologies.

Work, or proposals for work, are now active within both API and ISO (International Standards Organization) to begin addressing probabilistic design approaches for tubular goods. Within API, an active agenda item is studying the use of products specification levels (PSL) to enable users to invoke enhanced specification levels on selected tubular products. Within ISO, a proposal is active, and will be decisioned in the near future, to activate work on the modernization of its performance property standard.

Clearly, work remains for both users and manufacturers to realize the full potential of these more advanced design techniques. From the user's perspective, work remains around definitive methods to generate performance properties based on statistical pipe information, closer scrutiny of field load descriptions, and increased reliabilities for field procedures based on quality assurance systems and processes for the entire tubular life cycle and particularly, for rig-site procedures.

From the manufacturer's perspective, focus is needed on better use of available inspection and computing technologies to measure, record and report salient pipe properties in a statistically significant and cost-effective manner. Future pipe deliveries should be accompanied by computer disks with comprehensive mechanical/dimensional information. Better still, the information will be electronically transmitted to the involved user(s) or distributor(s). Such information will characterize quality/capabilities of different manufacturers. This will allow more informed comparisons between manufacturers, enable informed buying decisions, and promote higher quality for the whole industry. These statistical product characterizations will, along with price, delivery and service, become the basis of future OCTG market competition.

Applications of probabilistic design techniques have been successful in reducing tubular and well costs while maintaining or increasing tubular reliability and operational integrity. This should be expected since the techniques are based on generating the optimal design with enhanced information characterizing the product. These design techniques are not and should not be based on introducing substantial risks into the OCTG or well design process.

An overview of ARCO's tubular failure history, and others to come, shows that tubular reliability is in fact driven by a variety of operational and procedural issues, but rarely design driven issues. With the exception of a properly engineered landing procedure, virtually none of these failure causes are addressed by classic OCTG design methodology. The industry must and will move forward on these issues. ARCO looks forward to actively participating and leading these changes. Based on its experiences, a dual purpose approach is recommended whereby tubular designs are optimized, and reliabilities are simultaneously increased.

ACKNOWLEDGMENTS

The authors thank the management of ARCO Exploration and Production Technology for their support of the preparation and presentation of this paper. They also thank all personnel in ARCO's operating companies, subsidiaries and joint-ventures who have contributed to the success of the probabilistic design applications described herein. This article was prepared from paper SPE 48324, of the same title, by the same authors, presented at the SPE Applied Technology Workshop on Risk Based Design of Well Casing and Tubing, The Woodlands, Texas, May 7-8, 1998.

LITERATURE CITED

1 Payne, M. L. and J. D. Swanson, "Application of probabilistic reliability methods to tubular design," paper SPE 19556, presented at the 64th Annual SPE Technical Conference and Exhibition, San Antonio, Texas, Oct. 8-11, 1989, and the 13th SPE Production Technology Symposium, Lubbock, Texas, Nov. 13-14, 1989. Published in SPE Drilling Engineering, December 1990.

2 Clinedinst, W.O., "Collapse safety factors for tapered casing strings," Materials, 1945, pp. 181-184.

3 Payne, M. L., W. T. Asbill, H. L. Davis and P. D. Pattillo, "Joint-industry qualification test program for high-clearance casing connections, paper SPE/IADC 21908, presented at the 1991 SPE/IADC Drilling Conference, Amsterdam, March 11-14, 1991, published in SPE Drilling Engineering, December 1992.

4 Payne, M. L. R. E. Leturno and C. A. Harder, "Fatigue failure of API 8-Round casings in drilling service," paper SPE 26321, presented at the 1993 SPE Fall Conference, Houston, Oct. 3-6, 1993. Published in Oil and Gas Journal, Nov. 28, 1994.

5 Asbill, W. T., P. D. Pattillo and W. M. Rogers, "Investigation of API 8-Round casing connection performance, Part I: Introduction and method of analysis," Journal of Energy Resources Technology, Vol. 106, 1984, pp. 130-136.

6 Asbill, W. T., R D. Pattillo and W. M. Rogers, "Investigation of API 8-Round casing connection performance, Part II: Stresses and criteria," Journal of Energy Resources Technology, Vol. 106, 1984, pp, 137-143.

7 Payne. M. L., J. L. Zerbi and D.C. Sims, "Tubular design optimization for ARCO China's Yacheng 13-1 development," paper SPE 29926, presented at the International Meeting on Petroleum Engineering, Beijing, PR China, Nov. 14-17, 1995.

8 Anon., Drilling Engineering Association, Project 111, "Assessment of the data acquisition capabilities of major world tubular mills and processors," coordinated by Oil Technology Services, Houston, November 1996.

9 Payne, M. L. and D M. Hurst, "Heavy wall production tubing design for special alloy steels," paper SPE 12622, presented at the Deep Drilling and Production Symposium, Amarillo, Texas, April 3, 1984. Published in SPE Production Engineering, July 1986.

10 Bradley, W. B., "The effect of casing wear on the burst strength of casing, Part 2: Statistical burst strength of worn casing strings," Transactions of the ASME, Journal of Engineering for the Industry, Vol. 98, 1976, pp. 686-691.

The authors

Dr. M. L. (Mike) Payne is an advisor with ARCO Exploration and Production Technology in Plano, Texas. He holds a BSME degree from Rice University (1982), an MS degree in petroleum engineering from the University of Houston (1984) and a PhD in mechanical engineering from Rice University (1992). Previously a consultant, he has 17 year's industry experience and has been with ARCO for 14 years in positions of increasing responsibility in drilling operations/research. He was seconded to BP to work on the Wytch Farm ERD project, and he is continuing work on ERD, HPHT drilling and high-performance well construction. Dr. Payne has published numerous technical papers for SPE, IADC, ASME and several trade journals. He is a registered professional engineer, a member of the SPE Editorial Review Committee, Vice-Chairman of API Subcommittee 5 and Covenor of ISO/TC67/SC5/WG2. He served as an SPE Distinguished Lecturer during the 1995-'96 season on the subject of ERD.

Udaya B. Sathuvalli, research engineer with the drilling and completion group in ARCO Exploration and Production Technology, has a BS degree in electrical power engineering from Mangalore University, an MS degree in electrical engineering from the Indian Institute of Science and a PhD in mechanical engineering from Rice University. His areas of specialization include solid and fluid mechanics, heat transfer and electromagnetic field theory. Current interests include mechanics of tubulars, application of probabilistic methods to OCTG design, well-bore heat transfer and coiled tubing.

Stephen Crabtree is the founder of Technical & Quality Solutions, Inc., a consulting firm based in Houston. He has 15 year's oilfield experience, previously positioned as Sr. VP at T. H. Hill Associates, Inc., in charge daily operations of the firm. He has consulted with companies around the world on quality and mechanical requirements for equipment in all areas of drilling and completion.

COPYRIGHT 1998 Gulf Publishing Co.
COPYRIGHT 2000 Gale Group