The Mensa field is a deepwater Gulf of Mexico (GOM) geopressured-gas field, producing from Upper Miocene sand with faulted four-way closure. During the early life, pressures fell sharply, with the field performing as a dominantly depletion-drive reservoir. The trend indicated isolated compartments. After a few years of production, pressures leveled off significantly. This midlife trend seemed to be in contact with originally isolated gas compartments and an aquifer. The late-life pressure trend indicates aquifer influx that effectively blocked the main producing wells from a significant gas area and pressure support.
Introduction
The Mensa field is a deepwater GOM subsea development in Mississippi Canyon (MC) Blocks 686, 687, 730, and 731 (Fig. 1). The field was discovered in 1987 with Well MC731-1, which penetrated 110 ft of net gas pay in the I sand. Appraisal Well MC730-1, drilled in 1988, penetrated 168 ft of net gas pay in the same sand.
On the basis of seismic interpretation and well information, the I sand was modeled as a thick high-net-/gross-pay-ratio (NTG) sand with 29 to 33% porosity, 250- to 1,000-md permeability, and a gas-in-place volume of up to 1.5 Tcf (Fig. 2). North/south-striking faults divide the main I sand into eastern and western fault blocks. The fault tips out to the south such that the fault blocks are in good pressure communication (confirmed later by depleted pressures recorded in Well A2). A basinwide aquifer was modeled in the south, with the original gas/water contact (GWC) set at 15,750 ft subsea. The I sand was assumed to produce by moderate-to-strong aquifer support.
The Mensa field faced many challenges.
- Deep water (approximately 5,300 ft) resulted in high subsea-well and system costs.
- Nearest facility was approximately 68 miles away (West Delta Block 143), involving high development cost to connect.
- Low ultimate recoveries because of the almost flat structure (approximately 5° dip), large aquifer, and high-permeability sand (wells water out immediately after water breakthrough).
- Poor economics because of the low condensate yield and low gas prices at that time (approximately USD 1.5/Mcf).
- High subsea-intervention cost and long time required (e.g., wireline job, acid treatment, or workover/sidetrack).
- High rock compressibility (30 to 60 microsip), which resulted in collapsed casing and sand-control issues.
A fit-for-purpose field-development plan, shown in Fig. 3, with three subsea producing wells, a subsea-manifold system, and 12-in.-diameter 63-mile-long flowline to platform WD143 was approved in 1995. Each well was connected to the subsea-manifold system through a 6-in.-diameter 5-mile-long flowline. Well A1 came on line in July 1997, producing the first gas from the Mensa field. Well A3 came on line in December 1997. Well A2 was drilled and completed in 1998, and production started in November 1998, as indicated in Fig. 4.
Well A2 penetrated thinner-than-expected I sand (per the seismic interpretation) and unexpectedly encountered approximately 24 ft of net gas pay in the H sand. Well A2 penetrated the depleted I sand, with pressure in alignment with pressures in Wells A1 and A3. The H‑sand pressure was recorded at 9,830 psi. Because the H sand was penetrated after production (and depletion) of the I sand, the lower pressure recorded in the H sand posed several scenarios. Geoscientists favored a scenario in which production and depletion from the I sand caused compaction, thus providing space for the overlying shale and H sand to expand. Data from Well A4 and 2D seismic recorded in mid-2005 supported this scenario. Well A2 also penetrated the K sand (below the I sand), but development was deferred pending additional information and evaluation.
Field History
Mensa gas has a high methane content (99 mol%) and very low condensate yield (1.6 bbl/MMscf). Pressure/volume/temperature analysis showed a dewpoint of approximately 5,300 psi at 176°F reservoir temperature. A glycol-dehydration system was installed on the WD143 platform to process Mensa gas. For the first 1½ years, the field produced intermittently because of topside issues (handling the produced condensate). Regular production from the field started in September 1998, and the peak rate of 300 MMscf/D was achieved in December 1998. Each of the three wells, when tested individually, produced more than 130 MMscf/D. However, because of the facility’s design capacity and to maintain well integrity, a conservative production philosophy was adopted to produce wells at rates less than their design rate of 100 MMscf/D.
A steep decline in reservoir pressure was observed from the start of production until early 2001 (i.e., early time of the Mensa field). The reservoir pressure depleted by approximately 1,500 psi after producing approximately 175 Bcf of gas, as shown in Fig. 5. This decline indicated that the producing wells were connected to a smaller reservoir volume than originally estimated (approximately 1.1 Tcf vs. initial estimates of approximately 1.5 Tcf) with insignificant or no aquifer influx. System energy came from fluid expansion and compaction.
The reservoir behavior observed during early time prompted remapping of the I sand with new high-frequency seismic data. The interpretation showed the I sand as a two-lobe sand, Ia and Ib, isolated by a thin shale break. The shale was visible in well logs, but in the original model the shale was assumed insignificant. The updated maps also suggested that the aquifer in the south is attached to the Ib lobe. The Ib lobe was present and partially perforated in Well A3, but it was inferred that the Ib lobe, which has approximately 200 Bcf of gas in place and an attached aquifer, was isolated and not contributing to production. The north block, where the I sand is significantly thinner (confirmed later with Well A6), has complex faulting. The area has an estimated 175 Bcf of gas in place (Ia sand), and its communication with the main I sand was uncertain. Production and pressure history observed during the early time posed a serious question about ultimate recoveries from the field and caused a downward reserves revision.
In mid-2000, the reservoir pressure began deviating from the early-time trend, indicating an additional energy source, possibly from a combination of sources (gas lobes/compartments, an aquifer, or rock compaction). During this midtime period, which extended to early 2007, the field produced approximately 500 Bcf of gas and experienced approximately 1,500-psi pressure depletion (recall that the field produced only 175 Bcf for similar pressure drop in the early time). Important events of the Mensa field during the midtime period were
- Well A3 indicated slightly higher reservoir pressure during early 2001 compared with the updip Well A1.
- Well A4 came on online in early 2003, increasing field production to approximately 300 MMscf/D.
- Well A4 was completed with downhole gauges, providing direct reservoir monitoring and management.
- In early 2003, Wells A2 and A3 started producing sand and were choked back to produce at sand-free levels (approximately 30 MMscf/D per well).
- In February 2005, Well A3 went off line because of sand failure.
- Early in 2005, Well A1 started producing sand and was choked back to produce at a sand-free level.
- Two time-lapse 2D-seismic lines were shot over the Mensa field in mid-2005.
- In August 2005, the Mensa field host facility (WD143) shut down for 3 months after Hurricane Katrina, resulting in the Mensa field being shut in for 93 days.
- In July 2006, Well A1 went off line because of sand failure.
- In December 2006, Well A5 was drilled to a position low on structure and slightly updip of appraisal Well 731-1, which originally found pay from top to base, and found that this reservoir area was swept fully by the aquifer.
During the midtime period, the reservoir-pressure-depletion rate slowed down, indicating that an external source of energy was in contact with the reservoir, primarily from an aquifer or another gas compartment. However, how much energy was coming from each source and which reservoir areas were initially baffled and later came into communication after pressure depletion (approximately 1,500 psi, as indicated by history match) has never been quantified. Well A3, which was perforated over the entire Ia layer and the top of the Ib reservoir layer, showed higher pressure at the beginning of the midtime period, indicating that some source of energy was becoming available, possibly from the lower Ib layer.
Interpretation of time-lapse 2D-seismic data indicated visible compaction effects in the I-sand gas body, but compaction effects in the aquifer were not apparent. The change in seismic amplitude and thickness corresponded qualitatively to the decrease in porosity and the aquifer influx as a result of 8 years of production (approximately 600 Bcf). There also was considerable change predicted by seismic to the overlying strata related to complementary shale swelling, which was confirmed later with drilling. The change in seismic amplitude was predicted by modeling and was the result of compaction and aquifer influx. The change in the 2D-seismic surveys showed no distinct water front. Two interpretations of the 2D seismic emerged.
- Stratigraphically, the Mensa field is a collection of amalgamated channels stacked in high-NTG sands. Some of these sands are more connected to the wellbore than others, consistent with the Katrina pressure-buildup event and a second energy source. In the first interpretation, it was argued that the total volume of the reservoir was underestimated, and that there was little or no aquifer influx. Therefore, material and potentially stranded hydrocarbons existed downdip. The connected aquifer in this scenario was very small. A 35% chance was assigned to this possibility.
- Shortcomings in the time-lapse 2D-seismic-data acquisition were recognized, and the 2D acquisition was meant only to be a proof of concept to propose a later 3D shoot. The two 2D lines that were acquired did not span the entire reservoir. Also, it was identified that if the completion failure at Well A3 was related to a changing stress regime when water hit, the aquifer could have invaded the entire reservoir area imaged on the 2D line. No separate water front, independent of compaction, should have been recognized in the data. The Mensa seismic bright spot associated with gas had a much higher signal/noise ratio than the seismic response in the aquifer, and it was recognized that compaction changes in the aquifer would be difficult to discriminate because of noise. In this second scenario, the reservoir volume predicted from well control and seismic inversion was taken to be correct, and the second source of energy was the aquifer. A 65% chance was assigned to this possibility.
In August 2005, Hurricane Katrina caused damage to several offshore facilities, including WD143, resulting in the Mensa field being shut in for 93 days. Most of pressure data were lost because of the power outage. Limited data that were recorded during the shutdown indicated that the reservoir pressure was increasing by approximately 7 psi/D during the first few days of shut-in and by approximately 2 psi/D after 93 days. From observations, it was inferred that the energy source was finite.
Well A5, drilled in December 2006, was drilled near appraisal Well 731-1, downdip of Well A3. It penetrated wet Ib sand, confirming aquifer influx. Pressures could not be recorded in the well because of borehole problems. Therefore, no information about the GWC could be inferred at that time. The well was sidetracked (as planned) updip of Well A4 as a producer. Well A5 was drilled with several objectives.
- Optimize field production.
- Drain volumes updip of Well A4 that would be trapped after Well A4 waters out.
- Produce in conjunction with Well A2, which could not produce by itself because of a subsea-system flow-assurance limit.
Despite comparatively lower incremental volumes (Wells A5 and A2 combined), Well A5 was justifiable economically with the higher gas price and associated condensate production at that time and the benefit of delaying field and subsea-system abandonment for several years. The well was kept away from the north fault block because of uncertainty of its communication with the main Ia sand.
In early 2007, the field started showing steep reservoir-pressure decline, which extended to early 2009, referred to as the late-time period (Well A5 came on line in July 2007). Reservoir simulation supports the pressure decline in that the two I-sand fault blocks (Well A5 in the west fault block and Well A2 in the east fault block) may have been isolated by the encroaching aquifer, making the reservoir a two-tank system with pressure communication through the aquifer. Higher withdrawal rates from the west fault block (Wells A4 and A5 produced an average of 185 MMscf/D) resulted in a rapid reservoir-pressure decline. The phenomenon also was confirmed with reservoir simulation. From pressure and production behavior observed during the late-time period, it was inferred that the north fault block is in communication with the main I sand. The late-time period ended in March 2009 when water breakthrough occurred in Well A4 and the field lost 110 MMscf/D of production. The Ia-sand pressure started increasing after Well A4 went off line because of lower withdrawal rates from the reservoir.
At the end of the late-time period, Well A2 was still producing at 25 MMscf/D. The pressure and production history collected during the late-time period and the consistent performance of Well A2 provided the basis for another development well, A6, the last well in the field. The north fault block was targeted with Well A6, which was justified with value drivers similar to those of Well A5. The well was economically attractive when combined with continued production from Well A2 and the delayed field-abandonment costs. Well A6 came on line in December 2010, and Well A5 went off production in January 2011.
Besides its own production, Well A5 provided flow-assurance support for Well A2, which produced an additional 17 Bcf between March 2009 and January 2011 (after Well A4 went off line because of water breakthrough) and delayed field abandonment by approximately 2 years. Had Well A5 not been drilled, the field was a candidate for abandonment after March 2009. The Mensa field was still producing at the time this paper was written because Well A6 provides support to Well A2 in maintaining the flow-assurance limit.
Well A6 penetrated the depleted I sand (on pressure trend with downdip Well A5) and the partially depleted H sand (approximately 9,100 psi). Well A6 experienced steep reservoir-pressure decline. Pressure-transient analysis indicated that the well penetrated a small compartment, which is in baffled communication with the main Ia sand. Well A6 produces approximately 10 MMscf/D, much lower than expected. The compartment pressure stabilized at approximately 3,000 psi, while the reservoir pressure recorded at Well A5 indicates approximately 0.43-psi/D continued recharging.
The Mensa field has produced approximately 16 Bcf (Wells A6 and A2 combined) since Well A5 went off line. The field started producing water, although the gas-flow rate has not changed. Because of system and reservoir-performance limitations, it is impossible to determine which well is producing the water. The subsea system has 12,000 bbl of liquid in equilibrium state. Months are required to collect a representative sample. The wells cannot be tested individually. Water analysis indicates approximately 10,000 ppm chlorides, which is more likely condensed water (seawater and formation water have chlorides content of approximately 16,000 ppm and 120,000 ppm, respectively). As yet, water-production rates have not increased.
The field is in its final stage. There are some shallow opportunities that are targeted for uphole recompletion or short sidetrack completions.
This article, written by Senior Technology Editor Dennis Denney, contains highlights of paper SPE 159741, “Mensa Field, Deepwater Gulf of Mexico—Case Study,” by Muhammad Razi and Peter Bilinski, Shell Exploration & Production, prepared for the 2012 SPE Annual Technical Conference and Exhibition, San Antonio, Texas, 8–10 October. The paper has not been peer reviewed.