Field examples of gas migration rates - Statistical Data Included

Recent research challenges traditional concepts of gas influx (kick) migration, as illustrated by five field case histories

In recent years, gas migration rates have become the center of considerable controversy. Historically, field personnel have used a "rule of thumb" that gas migrates at the rate of 1,000 ft/hr. Early basic research illustrated that the factors affecting the rate of bubble rise were quite complex. [1] For example, the rate of rise is affected by influx properties, mud properties, hole eccentricity and annulus size, to name a few.

The rate of rise is also affected by the manner in which the influx entered the wellbore. If the influx is dispersed in the mud as small bubbles, rise characteristics are different than if the influx enters the wellbore as a continuous bubble. A dispersed influx generally migrates much slower than a continuous bubble.

It was generally expected and observed that as an influx began to migrate toward the surface, the surface pressure would begin to increase. Those incremental changes in surface pressure were utilized to determine the rate of rise and predict travel of the influx. In addition, these techniques were used to model the well-control problem, expand the influx, protect the casing shoe, and otherwise analyze the problem. However, in many instances, the variables were too complex to permit accurate calculations and analyses in field operations--consequently, the old rules of thumb were utilized.

Recently, additional research has been presented that challenges the traditional concepts of influx migration. [2] It suggests that what had been done in the past was often in error, that the rate of rise was as high as 18,000 ft/hr, and pressure increases at the surface could not be relied on to properly analyze and predict influx migration and behavior.

To further illuminate this interesting and vital subject, this article chronicles several, well-documented instances of influx migration under a variety of conditions.

FIELD EXAMPLES

Described here are five field examples of wells incurring gas influx, including a directional well and a controlled blowout.

Example 1. One of the most interesting examples of influx migration--or the lack thereof--occurred at the E. N. Ross 1 near Jackson, Mississippi. [3] While on a trip at 19,419 ft, with I 7.4-ppg, oil-base mud, a 260-bbl, sour-gas influx was taken. The top of the influx can be calculated to be at 13,274 ft, with a shut-in surface pressure of 3,700 psi. The wellbore schematic is presented in Fig. 1A.

Since the influx entered the wellbore in a continuous bubble and was significantly less dense than the mud, it was expected that the migration would occur rapidly. However, that was not the case, as indicated by the fact that surface pressure remained essentially constant for the next 17 days while snubbing equipment was being rigged up. After 17 days, the surface pressure slowly began to decline to about 2,000 psi.

Migration is most often accompanied by an increase in surface pressure. However, in this instance, the influx was migrating from the 7-in, liner into the 9 5/8-in, casing. Therefore, the influx shortened and the surface pressure declined. As illustrated in Fig. lB. the top of the influx reached 10,200 ft. Six days later, during snubbing operations, analysis confirmed that the top of the influx was indeed at about 10,200 ft. In a total of 23 days, a significant influx had migrated a total of only 3,274 ft.

Example 2 (directional well). In this example, a directional well was being drilled with an 1l.2-ppg, gel-polymer mud system. Plastic viscosity was 29 and the yield point was 29 lb/100[ft.sup.2]. The well-bore schematic is shown in Fig. 2. As illustrated, 7-in. casing had been set at 8,563 ft and a 6-in, hole was being cored at 9,111 ft MD, 8,412 ft TVD. Wellbore angle was 38[degrees] at an azimuth of 91[degrees].

With the bit at 6,181 ft on a trip out of the hole, the well kicked and was shut in. The kick occurred at 0700 hr and a 12-bbl influx was recorded. Shut-in drill pipe pressure was 840 psi, and shut-in casing pressure was 1,040 psi. Analysis of the pressure data was conclusive. The top of the influx was at 4,685 ft. While operations were being conducted to strip back to bottom, the influx migrated, reaching the surface at 1,030 hr, for a migration rate of 1,339 ft/hr.

Example 3. At the Santa Fe Energy Bilbrey well, Fig. 3, a 200-bbl kick was taken while on a trip at 14,080 ft. [4] Density of the dispersed water-base mud was 11.7 ppg. As shown in Fig. 3, the top of the influx was at 8,422 ft. The kick occurred at 1600 hr. The influx reached surface at 1000 hr the following morning for an average migration rate of about 470 ft/hr.

Calculations based on the rate of pressure increase indicated the influx was initially moving at a rate of 290 ft/hr. Calculations based on incremental pressure increases near the end resulted in the estimation that the influx was moving at 400 ft/hr.

Example 4 (blowout). The ability to calculate influx behavior based on changes in surface pressure has been the subject of considerable debate. At a blowout near Abilene, Texas, the opportunity was presented to test these techniques, as well as to observe the migration rate of gas through water.

The wellbore schematic is presented in Fig. 4. The well was controlled by pumping mud down the drill pipe. However, the upper formations were supercharged during the blowout. During the kill operations, a hole developed in the drill pipe at 980 ft. After any pumping operation, gas would enter the drill pipe and migrate to surface.

On one occasion, after pumping fresh water down the chill pipe, the gas migrated to surface in one hour and 15 mm., for a migration rate of 784 ft/hr. On another occasion, as the gas migrated, water was pumped in four, 2-bbl increments and the resulting change in surface pressure was noted. The volume of water was carefully measured in the suction tank of a service company cement pump truck.

In the 4 1/2 in. drill pipe, two barrels of water represents 142 ft of hydrostatic head, or 122 psi. On each occasion, the surface pressure declined by 120 psi when two barrels were pumped. It should be noted that although the water being pumped was incompressible, the other fluids in the wellbore were extremely compressible. The annulus contained some combination of oil, gas, mud, water and cement. The drill pipe below 980 ft contained water, cement, mud and gas.

Example 5. The wellbore schematic for this example is presented in Fig. 5. As illustrated, 7-in. casing was set at 7,434 ft. After coring to a total depth of 7,667 ft in 6-in. hole, a trip was commenced to retrieve the core. The gel polymer mud density was 10.1 ppg, plastic viscosity 14 cp, yield point 16 lb/100 [ft.sup.2], and funnel viscosity 40 sec/qt. The drill-string was pulled to 2,484 ft, where a gain of 3 bbl was observed. The well was shut in with a total 6 bbl gained. The shut-in drill pipe and shut-in casing pressure were equal at 350 psi. The kelly was picked up, the choke opened and the well circulated. When the well was shut in the second time, the total gain was 115 bbl and shut-in drill pipe and casing pressure were equal at 1,320 psi.

It seems ironic that the influx did not migrate when the top of the influx was only 4,380 ft from surface in 10.1-ppg mud. For three days, attempts were made to lubricate mud into the hole. However, shut-in surface pressures remained stable at 1,320 psi, indicating the influx failed to migrate.

During the next 24 hr, the drillstring was stripped back to bottom. A total 7 bbl gas was bled from the annulus during the stripping process. The remainder of the gas was at total depth and had to be circulated to the surface.

SUMMARY AND CONCLUSIONS

Influx migration is a complicated phenomenon. As illustrated in these examples, the influx may or may not migrate. In most instances, it can be anticipated that the influx will migrate and, if it does, its rate can and usually will vary throughout the process. Normally, the migration rate will increase as the influx approaches the surface.

It can generally be anticipated that surface pressure will increase as the influx migrates toward the surface, and the increase in surface pressure can be used to analyze wellbore conditions. However, as illustrated in Example 1, it is possible that-depending on wellbore geometry and physical properties of the influx--surface pressure can actually decrease as the influx migrates upward.

In the examples presented, the highest rate of migration was 1,339 ft/hr, which was observed in the wellbore, which was inclined to 38[degree]. In the vertical wells, the highest migration rate observed was 784 ft/hr in fresh water. On two occasions, one in 17.4-ppg, oil-base mud and one in 10.7-ppg, water-base mud, the influx did not migrate.

In the examples studied and presented, surface pressures could be relied on to predict influx behavior and migration rate. In all cases, well-control personnel must rely on the conditions at the well to make every effort to analyze and model the well's condition. In the experience of these writers, these techniques, analyses and models have proven reliable for designing well-control procedures. Failure to understand these situations can cause a serious well-control problem to further deteriorate into a major disaster.

THE AUTHORS

Robert D. (Bob) Grace, president of GSM Enterprises, Inc., holds BS and MS degrees in petroleum engineering from the University of Oklahoma. He served as head of the petroleum engineering department at Montana College of Mineral Science and Technology. He has more than 30 years of on-site experience as a consultant in blowouts, fires, well control and deep drilling/completing operations. He has authored numerous papers on blowouts, well control and drilling practices, and has conducted seminars worldwide in well control and drilling practices since 1968.

Jerry Shursen, previously a partner in GSM and now an independent consultant for a major oil company working offshore in the U.S. Gulf of Mexico, holds a BS in chemical engineering from Oklahoma University. He has extensive foreign operations experience and has conducted more than 150 seminars on drilling practices, pressure control and well completion operations in domestic and foreign locations.

LITERATURE CITED

(1.) Rader, D. W, A. T. Bourgoyne and R. H. Ward, "Factors affecting bubble rise velocity of gas kicks," Journal of Petroleum Technology, May 1975, pp. 571-585.

(2.) Tarvin, J. A., et al., "Gas rises rapidly through drilling mud," IADC/SPE Drilling Conference, February 1994.

(3.) Grace, R. D., Advanced blowout and well control, Gulf Publishing Co., 1994.

(4.) Grace, R. D., M. Barton and B. Cudd, "Mud lubrication-a viable alternative its in well control," SPE/IADC Well Control Conference, 1995.

COPYRIGHT 2001 Gulf Publishing Co.
COPYRIGHT 2001 Gale Group