FLUSHED ZONE PROCESS HELPS CONTROL GAS MIGRATION IN PRIMARY CEMENTING

Robert R. Teichrob Husky Oil Operations Ltd. Calgary Gas migration during cementing operations can G be eliminated with optimal drilling practices and ultra-high filtrate loss drilling fluids or properly designed squeeze fluids. A method based on the flushed zone theory involves squeezing fluid into a potential gas-bearing formation to reduce the permeability to gas near the well bore. The process aims to suppress gas flow long enough for the cement to set.

Aug. 16, 1993

13 min read

Robert R. Teichrob
Husky Oil Operations Ltd.
Calgary

Gas migration during cementing operations can G be eliminated with optimal drilling practices and ultra-high filtrate loss drilling fluids or properly designed squeeze fluids.

A method based on the flushed zone theory involves squeezing fluid into a potential gas-bearing formation to reduce the permeability to gas near the well bore. The process aims to suppress gas flow long enough for the cement to set.

Paying close attention to how a well is drilled and prepared for cementing is at least as important to the success of the well (in terms of eliminating gas migration) as are exotic cement blends. The classical procedures to suppress gas influx use special cementing products and placement techniques, and to be successful these procedures often require the operator to have a prior knowledge of gas influx zones.

Statistics from Canada's Energy Resources Conservation Board (ERCB) show that about 4,500 wells in Alberta currently exhibit surface casing vent pressure. Soil leaks adjacent to wells with vent leaks are common, but these leaks are more difficult to quantify both in number and flow rate.

These leaks have led to concerns ranging from the ability to secure reclamation certificates (well abandonments and subsequent lease payment cessation) to unknown contributions to greenhouse effects. Thus, much effort has been directed toward solving the gas influx problem.

Various attempts to stop gas influx into freshly placed cement have been tried over the years; several theories on the interaction between liquid cement and gas are prevalent in the oil industry.

One theory is that the ability to impose hydrostatic pressure effectively on permeable and porous media containing gas is compromised as cement develops gel strength and shrinks because of hydration. Eventually, hydrostatic pressure degrades to a point where a pressure inversion occurs. Gas at the well bore face is no longer held in place by cement hydrostatic pressure and can now leak through any of several paths to the surface: between the casing and the cement, in the cement itself, between the cement and the hole wall, or through damaged rock adjacent to the well bore.

Much research has been geared to minimizing cement gel strength development time (transition time). Fundamentally, however, developing cement slurries with zero transition time has not been achieved. Furthermore, additives (such as surfactants) that coat the invading gas bubbles or that cause cement to generate gas internally for pressure equilibrium have had varying degrees of success.

FLUSHED ZONE THEORY

Most primary efforts to control gas influx into fresh cement have relied on cement composition, placement techniques, or a combination of the two. Furthermore, identifying distinct problem intervals that may introduce gas into a well bore can be virtually impossible. Thus, analyzing the entire hole interval before it is exposed to cement can help avoid certain cementing limitations.

The flushed zone process eliminates gas influx into new cement by forcing the gas radially away from the well bore, thereby creating a near-well-bore zone with reduced permeability to gas. The process is accomplished in the following way:

The well is drilled to a predetermined depth and all inhibited drilling fluids in the well are displaced by an appropriate squeeze fluid (Figs. 1-2).
The annulus is packed off and pressure is applied down the drill pipe (Fig. 3).
The volume of squeeze fluid pumped away is monitored, and the flush volume is estimated based on assumed permeability, porosity, and interval height. Once the squeeze fluid volume is pumped away, the remaining fluid is displaced back to surface with the original mud system (Fig. 4).
Drilling resumes, and the well is drilled to total depth, logged, and then abandoned or cased (Fig. 5).

As the hydrostatic pressure imposed by liquid cement decreases as a function of gel strength development, the altered permeability of adjacent gas-bearing intervals precludes gas movement into the setting cement.

The concept behind the flushed zone theory is basic: The gas is moved radially away from the well bore face, and because the formation's permeability to gas changes, the gas is kept away at least long enough to drill, log, and case the zone. Thus, cement can be placed without an initial concern for gas influx and subsequent cement contamination. Because the residual effects of reduced permeability to gas are unknown, cement permeability to gas in itself may play an important role in long-term solutions to controlling migrating gases through micro-annuli.

DEVELOPMENT

Field trials of the flushed zone theory were initially limited to two of four directional saltwater disposal wells drilled at the Husky Oil Operations Ltd. Bi-Provincial Upgrader in 1991. The initial work precluded mud displacement by a squeeze fluid and centered around proving the merit of using an inhibited drilling fluid as an economically viable alternative to a gel/chem system. An inhibited fluid was used because otherwise if water-sensitive shales adjacent to the well bore were damaged, the chances of securing a good hole with a good formation bond would be minimal. Thus, a potential conduit for gas migration to surface would be formed. The high spurt loss fluid (300 cc/3O min) would provide a pseudo buffer zone with some altered permeability to gas.

Opinions differed about the potential for the well to develop a leak. However the experience with operational problems (such as mud costs and cleanup) associated with inhibited drilling fluids were used to optimize other drilling fluid programs.

Of more interest regarding intensive field trial of this process were four wells recently drilled in the Vermilion/Wildmere, Alberta, area. The wells were drilled in the following order: Husky Wildmere (16B-31-48-6 W4M), Husky Vermilion (71-50-5 W4M), Husky Windy (6-27-49-4 W4M), and Husky Islay (11-3-50-4 W4M).

These wells are in an area with a high incidence of gas migration (70% of wells in the area have either vent or soil leaks), providing an excellent opportunity to test the flushed zone theory further. Several drilling/geological parameters were closely watched to yield as much useful information as possible, and particular attention was paid to the following:

Upper hole shale gas was monitored with sensitive gas detection equipment to pick a representative unproductive shale interval for coring and subsequent reservoir flow characteristic tests.
Upper hole shale samples were accurately caught and geologically interpreted on site.
Proper casing running and cementing procedures were closely followed.
The drilling fluid inhibitive properties were monitored.

The first three wells (16B-31, 7-1, and 6-27) were drilled as programmed with an aluminum sulfate (Al2(SO4)3) system. A lack of efficient solids control quickly became an operational problem requiring immediate attention. By the second well, solids control problems were alleviated with the use of higher volumes of pretreated mud as opposed to the less economically viable solution of using a centrifuge. The fourth well (11-3) was drilled with a conventional gel/chem mud system. Because of the use on a water-based gel mud system on the fourth well, a conclusive observation based on gas detection levels could be made on the effectiveness of the filtrate sweep and the correlation of gas-source upper hole shales.

Another reason for using gas detection equipment was to try to correlate similar spiked intervals in upper shales to pinpoint potential sources. Gas detection readings during drilling were in the order of 20 units, with an obvious deflection point to 200 units by 290 m measured depth. All wells drilled with Al2(SO4)3 showed definite suppressed gas levels compared to the well drilled with the conventional gel/chem mud system, verifying the positive effects of the filtrate sweep.

Although identifying a gas source may be of academic interest, the solution to gas migration problems in the area must deal with gas that one way or another has charged upper intervals. Shutting off a gas source to upper zones may make perfect sense in an area with limited development, but this approach does not solve the problems of dealing with 30 years of uncontrolled interzonal communication.

FIELD EXPERIENCES

Sandwich squeeze (a Dowell Schlumberger process) surface cement jobs were performed on all four wells. In these wells, there were considerable intervals of gravel and unconsolidated till through the upper portions of the surface hole. Despite the presence of these zones that were operationally difficult to cement Husky Oil was still able to squeeze volumes of 0.4-1.2 cu m and hold pressures of 4,000-4,500 kPa, thus exceeding industry-accepted fracture gradients by almost 100% (18 kPa/m industry-accepted fracture gradient compared to 32 kPa/m actual fracture gradient). The extra integrity at the shoe and the cooperation of Alberta Environment resulted in a reduced surface casing depth requirement.

Well 16B-31 was drilled with the inhibitive Al2(SO4)3 mud system and subsequently cased and cemented. Drilling fluid water loss was decreased from 350 cc/30 min to 20-30 cc/30 min prior to entering the productive sands,

The prognosis called for 15 m of hole to be drilled into the Paleozoic formation. At 13.9 m into the Paleozoic, total lost circulation occurred. The Al2(SO4)3 system was then replaced with a freshwater gel system to finish logging and to run production casing.

Well 7-1 was also drilled with an Al2(SO4)3 mud system and subsequently cased and cemented. The drilling fluid water loss was decreased from 350 cc/30 min to 20-30 cc/30 min prior to entering the productive sands. Because of the ability of the flushed filtrate to mask logging tool response, a drill stem test (DST) was necessary to evaluate the potential of the well as a gas producer; routine DSTs for potential gas zones are not an economic option for these Lloydminster-type wells. Therefore, it was imperative that water loss be decreased to 10-15 cc prior to drilling into any potentially productive sands.

A dynamic fluid caliper was run before cementing production casing in both 16B-31 and 7-1 wells. The dynamic fluid caliper is based on actual displacement efficiency. The conventional approach is to use integrated hole volume logs and increase volume by 20% (per ERCB G-9). By using the dynamic fluid caliper, Husky Oil saved an additional 10 tons of cement that would have otherwise ended up in the sump. Also, competent cement returns to surface were ensured, thus eliminating the possibility of remedial cementing through macaroni string from surface or running a costly log to locate the top of cement.

Well 6-27 was drilled with the inhibitive Al2(SO4)3 SYStem and subsequently abandoned. The main hole was drilled to 500 m measured depth, and then a resin squeeze was performed. The Al2(SO4)--4 system was displaced from the well with a Permabloc fluid (a type of resin).

Approximately 70 1. were squeezed away at 1,600 kPa surface pressure. This volume equates to 1.08 linear meters of porous silt/shale with a porosity of 10% being flushed to a radial distance of 75 mm. Once the squeeze was performed, the well was displaced with the original drilling fluid, and drilling operations resumed. Subsequent log analysis showed the well to be uneconomic. From an operational perspective, displacing the entire well from one fluid to another and then back to the first went smoothly.

Unfortunately, however, the abandoned well eventually leaked.

Well 11-3 was drilled with a freshwater gel system and subsequently abandoned. Based on an analysis of gas detector data, a core point from an uphole gas source was picked at 253 m measured depth.

The presence of gas in this interval was a mild surprise to the geologists. This interval helped substantiate the problem of uphole gas without an identifiable source. Two cores were cut from 253 m to 264 m, with 9.5 m total recovery. Drilling resumed and the well reached total depth on program. The well was dry and subsequently abandoned. Again, the abandoned well leaked.

RESULTS

Of the four wells drilled, only the first two (16B-31 and 7-1) were cased. Some of the reports as to whether 16B-31 is leaking have conflicted. A recent Husky Oil study stated "approximately a bubble every 20-30 sec" as a positive show of vent gas. More recent reports show the well has stopped leaking altogether.

The following is a possible explanation for the leak, assuming the well was leaking at a minimal rate: Complete lost circulation and subsequent regaining of circulation with a freshwater gel system at the end of operations probably displaced the inhibitive fluids from the near-well-bore shales and ultimately caused the water-sensitive shales to hydrate.

A successful casing cement job needs competent shales to achieve a bond. The zone of hydrated shale adjacent to the well bore could provide a conduit for gas to migrate to surface. Well 7-1 was drilled similarly and with no lost circulation problems; this well shows no signs of gas leaking to surface to date, thus helping to verify this supposition.

Wells 6-27 and 11-3 were drilled and subsequently abandoned. Although both wells are leaking gas to surface, the leaks are significantly different from each other according to surface observations. Well 11-3 was drilled with a freshwater system for datum information. This well developed a "vigorous gas bubble every 5 sec" immediately after abandonment plugs were run. Well 6-27 did not start leaking until roughly 8 days after abandonment plugs were run.

The original design of the flushed zone theory was to avoid contamination of the cement slurry by supplying an additional 16-48 hr of lag time before the gas migrated back to the well bore face. Based on the results, it is accurate to say that the initial objectives were achieved and that optimized squeeze fluid based on injectivity tests on cores may have stopped leaks from 6-27 altogether. Also, plug design in the absence of optimized squeeze fluids must cover the precharged shales of the Lea Park formation.

The flushed zone theory was not 100% successful in the Vermilion area wells; however, the following conclusions were drawn:

There A,as a remarkable difference in the way gas migrated to surface in Well 6-27 and Well 11-3. Because the abandonment plugs and intervals were virtually identical, the way the wells were drilled played a major role in the severity of the gas leaks.
Wells 16B-31 and 7-1 were drilled identically with the exception of 16B-31 losing circulation at total depth. Well 16B-31 developed a leak almost too small to measure; thus, the inhibited fluid displaced with freshwater in 16B-31 must have contributed to the eventual leak.

RECOMMENDATIONS

Injectivity tests with different fluids should be performed on samples taken from the core. Optimal drilling and squeeze fluids should then be designed based on these tests.

The squeeze process should be tested on more wells. Once the migration mechanism has been pinpointed, the process should be adapted to the ongoing abandonment of existing wells.

At first glance, it appears as though the success of the flushed zone theory was marginal at best. A detailed analysis, however, quickly shows that definite advantages were realized by using the inhibited, high water loss mud system.

The next step is the development of optimum drilling and squeeze fluids based on physical evidence rather than theory. The core recovered will be used for extensive fluid optimization. Further field testing is necessary for the process to be applied ultimately to rectifying the significant remedial problems in the area.

ACKNOWLEDGMENT

The author thanks Husky Oil Operations Ltd. for permission to publish this article. The author would also like to thank Barry Wagner, Stu Butler, Ryan Cox, Bob Baird, and John Ulrich for their support throughout the project.