Because depletion drive and partial water drive gas-condensate reservoirs diminish in pressure with production, flow conditions are continuously changing. Choosing the optimum tubing size for this type of wells can be made easier by use of a nodal analysis program to aid in identifying production velocities. More specifically, the seemingly intuitive conclusion that smaller tubulars will always be required during the latter stages of production in order to unload liquids may not be correct. In particular, when flowing bottom hole pressures on high productivity wells reach the low values of 500 to 1,000 psi, fluid loading considerations can become insignificant when compared to frictional losses caused by the unusually high velocities found at these low pressures. Nodal analysis is indispensable for identifying these situations. By performing a few basic sensitivity runs on bottom hole pressure and liquid yields, a confident recommendation can generally be made on whether one tubing size will serve the life of the well or if a size change is advised later in the well life.

Presented is a case history of the Goodrich #1 well located in East Bayou Postillion Field, Iberia Parish, Louisiana. The well produced from a partial water drive gas-condensate reservoir known as the Planulina P4 Sand which occurs at a depth of 14,300 feet in the subject well. The production interval is approximately 200 ft. thick and the well initially produced at rates in excess of 25 MMCFPD through 3½" tubing. After approximately 15 years of service, the tubing failed in 1990 and the well was worked over to replace the 3½" tubing with 2⅞" tubing. The smaller tubing was believed to be capable of depleting the reservoir from the now 900 psig bottom hole pressure down to roughly 500 psig abandonment pressure. However, after only two years of service, the 2⅞" tubing developed a leak and the well had to be worked over again to replace the tubing. Prior to the second workover, the bottom hole pressure was measured at 680 psig and a diagnostic caliper survey was run. The survey revealed that the leak was not a fluke and that significant wall loss existed over the entire tubing length. Nodal analysis was run prior to the second workover which revealed that, since the previous workover, the well had been producing from the 2⅞" tubing at rates far in excess of erosional velocity. Further, the nodal results indicated that 3½" tubing was still the optimum size to deplete the reservoir in spite of the greatly reduced reservoir pressure. The 2⅞" tubing was then pulled and replaced with 3½" tubing and the well was returned to production at rates 40% higher than before. From that point, another 1,800 MMCF was produced before final abandonment in late 1995 at 480 psig bottom hole pressure.

Fig. 1 is a graph of the production history for the Goodrich #1 which is one of fourteen wells that have produced a combined cumulative of over 267 BCF from the Planulina P-4 Sand Unit. The reservoir had been on production for nearly six years when the Goodrich #1 was drilled and completed in 1975. Therefore, it is fairly safe to state that the reservoir was known to be a depletion drive (later it would be determined that the expandable water layer below was ineffective in sustaining reservoir pressure, but nonetheless eventually moved into each producer causing them to water out). By the time the subject well was completed, several of the offset wells had been potentialed at rates indicative of a high productivity reservoir. Therefore the water and condensate yields could be estimated with a fair degree of accuracy. Given this data then, it might have been obvious that a well capable of production in excess of 25 MMCFPD would require 3½" tubing. However, when the well went off production in mid-1990 with a bottom hole pressure of 900 psig, the decision on what size tubing to replace the 3½" tubing with was probably not so obvious. Before going off line in 1990, the well was producing 2.7 MMCFPD plus 12 BF/MMCF and had produced a cumulative of 47,000 MMCF, 568 MBC and 100 MBW. Although finding a 14,300' deep sand producing with only 900 psig shut in bottom hole pressure may be a rarity, the techniques used to choose the replacement tubing size are the same (aside from mechanical safety margins) regardless of depth or pressure. It is not known whether the prior operator examined the problem using a nodal program, but in any case the smaller 2⅞ tubing was chosen and the well was returned to production at a greatly diminished rate of 1.8 MMCFPD plus associated fluid. Because the smaller tubing was used, another problem was accentuated to problematic proportions: the produced gas contained approximately 1.9% CO2 At first inspection, this may not appear to be a major concern since the first tubing string (N-80 grade) had lasted 15 years. One may expect similar service from the 2⅞", N-80 replacement string. But, when the well was returned to production at such high velocities, it is believed that whatever inhibitive effects an oxide coating may have provided the tubing were being constantly stripped away. This effect combined with whatever actual erosion was taking place is believed to be the cause of the premature tubing failure after 2 years. The 2⅞" tubing had produced a cumulative of 1,120 MMCF, 8 MBC, and 7 MBW. This failure may have been avoided by using larger tubing.

Fig. 2 is the composite of the caliper survey run in 1993 just before the second workover in which the 2⅞" tubing was replaced. This figure exhibits that the tubing wall loss was not limited to any one section of the tubing. Localized CO2 corrosion usually occurs in the upper portions of the tubing string where condensed water forms carbonic acid which then refluxes in the gas stream. The extreme velocities of the low pressure gas appear to have greatly accelerated metal loss throughout the tubing string. At the time of this second workover, the bottom hole pressure was measured at 620 psig.

Fig. 3 is a graph of the inflow and tubing performance curves generated by the nodal program being used in this case1. Examined in this presentation are the generic tubing sizes: 2⅜", 2⅞", 3½", and 4" which are plotted against an array of reservoir pressures ranging from 620 psig down to 300 psig. From this presentation, the water and condensate unloading rates can be seen to occur at surprisingly low flow rates ranging from 150 MCFPD for 2⅜" tubing to 350 MCFPD for 4" tubing. These points are based on the program default values of 50 ft/sec for erosional velocity, 7 ft/sec to unload water, and 4 ft/sec to unload condensate1. As previously stated, the 2⅞" tubing, if allowed to produce at full capacity of 2,100 MCFPD exceeded the erosional limitations flag denoted by the circle at 1,600 MCFPD. Also according to the plot, the 2⅞" tubing would continue to produce above erosional velocity until the reservoir pressure drops below 500 psig. From the P/Z curve, not shown, the well could produce another 1.5 BCF before reaching 500 psig. Based on the previous tubing performance coupled with this new evidence, the 2⅞" tubing is ruled out as the replacement size. The 3½" tubing also exceeds erosional velocity at current estimated capacity, but only for the next 30 psig of reservoir pressure. The 4" is an obvious choice, but once the reservoir pressure drops below roughly 590 psig, the 3½" tubing would again be the more logical choice based on using velocity as the primary selection criteria. The risk of a similar tubing failure in the 3½" tubing is therefore deemed minimal since only an estimated 350 MMCF will be produced at velocities slightly above acceptable limits. Another option would be to simply choke the well back slightly during the production of the next 350 MMCF. In either case, the 3½" tubing was chosen with one other consideration: L-80 grade was run to aid in minimizing corrosion. Of course this entire analysis assumes that the 19 BBL/MMCF fluid yield will remain constant which appears reasonable judging from the production plot (Fig. 1).

To summarize, once loading considerations were shown to be negligible, the problem was reduced to minimizing frictional losses in the most cost effective manner. Approaching the problem of tubular selection in this manner should result in increased flow rates and lower abandonment pressures.

Ultimately the production was restored by installing the 3½", L-80 tubing and an additional 1.9 BCF plus 16 MBC was produced before the economic limit was reached at 480 psig. As a side note, there was a landing nipple located below the Model D production packer located just above the perforations in which a plug was set while the tubing was round tripped during both workovers. Had it not been for this completion design, workover fluid and a bridging agent may have been required to contain the well during the workover. At this depth and low pressure, the likelihood of restoring production following such measures would have been greatly reduced.

High velocity in gas wells is not always a bad condition. It can be beneficial to production when harnessed. Presented in Table 1 is the necessary data to outline a basic problem often encountered in gas wells: surface hydrates (or freezing).

TABLE 1

EXAMPLE VERTICAL WELL

TYPE WELL GAS 
PERFORATION DEPTH 6,000' 
EXPECTED PRODUCTION RATE 2,000-3,000 MCFPD 
FLUID YIELD 15 BBL/MMCF 
TUBING SIZE 2⅜" 
SHUT IN TUBING PRESSURE 2,300 PSIG 
FLOWING TUBING PRESSURE 1,200 PSIG 
TYPE WELL GAS 
PERFORATION DEPTH 6,000' 
EXPECTED PRODUCTION RATE 2,000-3,000 MCFPD 
FLUID YIELD 15 BBL/MMCF 
TUBING SIZE 2⅜" 
SHUT IN TUBING PRESSURE 2,300 PSIG 
FLOWING TUBING PRESSURE 1,200 PSIG 

The relatively low fluid yield coupled with the expected flowing tubing pressure found in this problem alert us that this well is likely to have a surface hydrate problem and will require a line heater. Depending on pressure decline and possible water production, the line heater may not be needed for very long. Therefore one would rather avoid the expenditure of setting and later having to remove this temporary piece of surface equipment. An alternative would be to complete the well with a production profile (nipple) just below the packer in which a down hole choke can be later installed using slick line. Once the well has been perforated and tested using a portable line heater, the proper down hole choke size can be calculated. Proper sizing is important to reduce the flowing tubing pressure enough to eliminate freezing problems at the surface while not unnecessarily restricting production. As an added precaution, flow couplings are generally run on either side of any nipple that may be used as a choke position to aid in eliminating possible erosion. A down hole choke in this well will also result in higher tubing velocities which in turn will aid in moving the liquids. This not only keeps the well unloaded but is beneficial in reducing hydrate formation at the surface by eliminating slugging. The steady flow of liquids at the surface will yield a constant transfer of heat carried from down hole. Later in the well life, the bottom hole pressure may decline or the well may begin producing more fluid. In either event, the choke can then be removed with slick line and the full capacity of the tubing realized.

One of the points demonstrated by this example is that the tubing size was intentionally chosen to be too large for current service. But by employing a down hole choke to increase velocity in the early stages of well life, the full capacity of the tubing could be utilized later without a workover. Another point to make here is that no loss in production volume was sacrificed in this example since a surface choke would have been necessary had the down hole choke not been used. Instead, a necessary pressure drop was repositioned downhole where the cooling effects could be offset by the virtually infinite heat source available at that point.

Out-dated notions about loading tendencies can often result in tubular selections that are too small. These choices can not only hamper current cash flow, but can also result in lost reserves on the tail end of production life; both of which are attributable to excess frictional losses. Some of the previous thinking on tubular sizing has been that," It may be better to choose tubing one size too small than one size to large. That way the well will at least flow smoothly without slugging." Although this thinking may not be altogether wrong, the thought process should incorporate nodal analysis to address the interaction of fluid loading and frictional losses and how they are related to production velocities. Even the most experienced individuals can be surprised occasionally.

Obviously, not all gas wells that are economical fall under the classification of high productivity. In low volume gas wells, deep or shallow, the mass flow rate is, by definition, low. Here it is equally important to understand what production velocities will be created by the tubulars of choice. A low productivity well may perform best with smaller tubing due to the smaller cross sectional area of the tubing which creates velocities sufficient to stay unloaded However, we may find that these low volume recompletions exist as a through-tubing plug back to much higher productivity zone that is now depleted. One way to achieve the desired effect of smaller tubing without a rig workover is to run a siphon string inside the existing tubing. Flow can come up the inside string or up the annulus created by the two strings depending on the sizes involved. The inside string generally ranges from 1" coiled tubing to 2-1/16", again depending on the well. Loading tendencies in annular flow can be difficult to predict. A nodal program is essential for analyzing such projects and many can have applications that specifically address coiled tubing designs.

Horizontal gas wells are probably more dependent on knowledgeable tubular selection and velocity profiles than their vertical counterparts. When the horizontal well is successfully completed with minimal skin damage, the frictional losses to move gas from the perforations to the well head (up the tubing only) are most often several times greater than the pressure loss across the completion (from the formation to the inside of the gravel pack). This is generally true with any high capacity completion, but more so here where the combination of high permeability and lateral reach in the pay section result in capacities that are many fold that of a vertical completion. With less pressure expended across the completion, there is more pressure available to effect draw down, provided it is not wasted on overcoming frictional losses in the tubing. Basically this problem can be visualized as one of energy allocation. But in high productivity wells, the flow rate per psi drawdown is significantly higher, so one must reach a decision on the optimum trade off between frictional loss and production rate. Keying in on velocity through nodal analysis can be very lucrative in horizontal wells.

It is not unreasonable to expect 4½" tubing to be the optimum size for completing in a 2,000' horizontal gas sand provided that surface facilities have been designed to handle 10-15 MMCFPD per well. Fig. 4 is an inflow performance curve1 for a 2,000' TVD dry gas well with 400' or lateral section. The curve is similar to Fig. 3 in that various tubing sizes are plotted against selected reservoir pressures. By selecting the initial flowing tubing pressure as 600 psi to accommodate single stage compression, the expected production can be seen to range from 5,500 MCFPD for 2⅜" tubing to 30,000 MCFPD for 4½" tubing. If the fluid yields are less than 1 or 2 BBL/MMCF, previously published literature suggests that the effects of erosion on the tubulars can be neglected. But even if erosional limitations are considered, the expected range of production is 4,500 MCFPD to 18,000 MCFPD for the same tubing sizes. Similar to the Goodrich case presented earlier, the 4½" tubing can also be shown to be the size choice should this well be depletion drive. If on the other hand, large amounts of fluid production emerge later in the well life, a 2-1/16" string can be inserted and hung off in the 4½ tubing.

  1. Some of our fundamental "rules of thumb" concerning gas well depletion may be lacking, i.e. that smaller tubulars will always allow lower field abandonment pressures. Velocity is the basis for loading and frictional loss calculations. When minimal unloading velocities are not calculated or at least estimated, current industry thinking seems to gravitate to tubulars that are too small. This practice can result in needlessly excessive tubular friction losses and restrictive flow rates as well as possible tubing failure due to production rates in excess of erosional velocity.

  2. Nodal analysis programs are an excellent source for gaining insight into velocity and frictional loss calculations. Unfortunately, it is surprising to find through general industry feedback that this valuable resource is still greatly under utilized.

  3. When the tubing is found to be too large, down hole chokes should be considered as an alternative to running smaller tubing. Most nodal packages can model down hole chokes. But in the absence of a program, some equipment specialty companies will perform the necessary calculations at no charge to ensure proper choke sizing Multiple down hole chokes may even be necessary to avoid erosional velocities.

  4. Sizing tubulars only for current loading considerations without some serious regard to future well conditions is almost a guarantee of future regrets. Some forethought should be given to how a reservoir is expected to deplete. Sensitivity runs on pressure and fluid yields are a must in well planning. This procedure is even more critical in horizontal wells where most of the total pressure drop can be frictional losses in the tubing.

  5. Every production system is limited either by the formation, the completion, the tubulars or the surface facilities. A logical approach to tubular selection process will help assure that efforts expended on a successful completion will not suffer from a lack of flow capacity in the tubing.

This paper was selected for presentation by an SPE Program Committee following review of Information contained In an abstract submitted by the author(s). Contents of the paper, as presented, have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material, as presented, does not necessarily reflect any position of the Society of Petroleum Engineers, Its officers, or members. Papers presented at SPE meetings are subject to publication review by Editorial Committees of the Society of Petroleum Engineers. Permission to copy Is restricted to an abstract of not more than 300 words. Ilustrations may not be copied. The abstract should contain conspicuous acknowledgment of where and by whom the paper was presented. Write Librarian, SPE, P. O. Box 833836, Richardson, TX 75083-3836, U.S.A., fax 01-952-9435.

Thanks is extended to Callon Petroleum Company for fostering the open-minded management practices that keep new technology at our finger tips. Thanks also to Dwights Energydata for supporting this article.

"
SAM & GLD Program Reference Manual
",
Dwight's Energydata, Inc.
(
1985-93
)