Tuesday, May 15, 2018

Renew an Old Field with a Horizontal Well

Maraven, S.A., one Venezuela's three national oil companies, is going for the hard-to-get roe in Block 1 of lake Maraicabo, Venezuela. Forty years of production there has left isolated pockets of hydrocarbon, known as attic oil, in the tops of structural and stratigraphic traps. Recovering this attic oil with vertical wells is not usually cost-effective because the thin layer of oil in place increases the likelihood of water coning. 

 Taking a new approach, an integrated team of geoscientists from Maraven and Schlumberger planned, drilled and completed VLA-1035 - Lake Maracaibo's first successful horizontal well- gaining an eight-fold increase in oil production over vertical wells in the same reservoir.

 The motivation for VLA-1035 was provided when Maraven's parent company Petroleos de Venezuela, S.A. (PDVSA) launched a development program for Lake Maracaibo. The plan called for generating 11 billion barrels of additional oil reserves through new wells, horizontal development and reworking of older wells. Although horizontal drilling had been considered in Lake Maracaibo since 1986, attempts by other companies to drill horizontal wells were unsuccessful because of the complex geology or completion problems.

Yet, horizontal drilling seemed the only way to produce from Block 1. A vertical well typically produced 150 barrels of oil per day (BOPD). Most older wells had been shut in as uneconomic, and the wells that were on line typically produced no more than 150 barrels of oil. Some recent wells began producing water immeadiately, others made water within two months. Early breakthrough of water was inevitable because of the reduced vertical heigh of pay, reduced reservoir pressure and increased relative permeability of water to oil.

The Planning Stage

In early 1992, Maraven began assessing the economic and tehnical feasibility of drilling a horizontal drainhole to recover remaining reserves.  Reservoir engineers evaluated production histories to identify regions with recoverable oil and later modeled drainhole performance. Geophysicist used three-dimensional (3D) seismic data, having vertical resolution of tens to hundreds of feet, to obtain a big picture of the reservoir and identify prospective sands. Geologist and sedimentologist examined cores and logs, with vertical resolutions on the order of inches to one foot, to identify sands and model their orientation, continuity and distribution. Petrophysicst working with sedimentologist integrated log and core data with drilling records, including bit and mud data, for 33 wells in the area. This provided an understanding of the mechanical stability of the formation, fluid distribution, oil-water contact location, and flagged possible drilling difficulties.

They targeted reservoir VLA-8 in Block 1, bound on the west by the Icotea fault. It contains a region of low dips (2 degree to 10 degree) called El Pilar and a region of high dips (30 degree to 45 degree) called the Attic. Since 1954, VLA-8 has produced 42 million barrels of the estimated 118 million barrels of oil in place. This production reduced reservoir pressure from 3200 psi to 1800 psi at 6700 ft [2040m] in some areas and raised the oil-water contact. 

Water coning has been a problem from the beginning, with the average water cut in the field increasing from 20% in 1960 to 85% by 1991. The influx of water moves hydrocarbons toward the top of traps, creating isolated pockets of oil. Because of the extensive production in the field, normally desirable high permeabiltiy zones had water, whereas low-permeability zones contained oil.

The attic is considered the last opportunity  for development in Block 1. Three-dimensional seismic data, shot in 1990 and covering 235 square km, revealed the structural complexity of the fold and fault systems that bound the reservoir, and also stratigraphic features within the pay sands. The steeply dipping flanks are difficult to image seismically because a mud layer at the bottom of Lake Maracaibo absorbs high-frequency seismic energies.

 Well-tie sections, time slices and 3D cube displays from Schlumberger's Charisma workstation contributed to understanding the structure. Productive sands in the attic are in the C-6 and C-7 horizons, which have each been divided into three intervals-upper, middle and lower. In addition, seismic attribute sections were generated on the workstation and interpreted. Seismic attributes, such as signal phase and polarity, can reveal subtle characteristics of a seismic trace. In this case, instantaneous phase sections were particularly helpful in confirming the continuity of the C-7 structure. But the steep dip of the beds prevented determining an exact location of the C-7 reservoir.

Maraven was especially interested in the massive C-7 sands, 60 to 200 ft thick, products of deltaic and fluvial depositional environments. Target sands appeared to be the C-7 upper and lower interevals, with the initial preference by Maraven for the lower one.

Their next step was putting the seismic, log and core data into a reservoir model that would help identify the drainhole's position in the sand for maximum production. Modeling performace of the proposed horizontal drainhole in the C-7 sands was accomplished with a black-oil reservoir simulator.

 Based on log and core analysis, the model comprised a partially anisotropic reservoir with a horizontal permeability 250 md and a vertical to horizontal permeability ratio, Kv/Kh, of 0.5. In the model , the reservoir was bounded on one side by an aquifer and on the other by the Icotea fault. Assuming that no more vertical wells would be shut in and that water cut would stabilize, Maraven calculated that existing conventional wells would recover only 18% of the remaining reserves.

To find the most productive drainhole location, Maraven modeled performance for four horizontal drainholes, with lengths of 584 ft , 884 ft, 1200 ft and 1600 ft, in the upper, middle and lower sands. The 1200-ft drain-hole in the C-7 lower sands performed best. Overall, reservoir modeling showed that a horizontal well would recover 40% to 160% more oil than a vertical well. 

After analyzing the seismic interpretation and the reservoir simulation, Maraven geoscientist concluded that a horizontal drainhole could not be drilled without additional information from a pilot well. First, they needed to pinpoint the top and thickness of C-7 with respect to the Icotea fault. Second, they needed to better define the oil-water contact. At this point, they negotiated with Schlumberger to manage drilling the pilot well and the subsequent drainhole under Maraven supervision. As Maraven and Schlumberger geoscientist worked together on the project, specialist from both companies refined the initial geologic and reservoir engineering studies.

Coordinating the project was Anadrill's Bill Lesso, who had worked with Schlumberger's Horizontal Integration Team (HIT), which pioneered an integrated-services approach to drilling horizontal wells. The HIT group found that a coordinator was essential to facilitate communication betwen disciplines and act as a catalyst for decision making. 

The program for drilling and completing the horizontal well took about 10 weeks. Each task in the program was listed chronologically with its projected duration and status. This helped identify both progress and problem areas. Once drilling began, these weekly meetings gave way to daily sessions at a "mission control" center in Lagunillas, 20 miles from the offshore rig but linked to it by phone, fax and data transmission lines. Every morning at 9:00, the team met to discuss drilling or completion operations- whatever was planned for the next 48 hours. The team needed to achieve a consensus on drilling decisions and be on call round the clock during critical operations. To keep team members and interested parties informed, the project coordinator prepared and distributed weekly updates - one page summaries that highlighted progress and issues to be resolved.

Prompt and frequent communication was critical for weaving togehter the expertise of Maraven and Schlumberger specialist. 

Drilling the Pilot Hole

The plan called for an 8 1/2- inch pilot hole deviated 55 degree , with three possible drainhole trajectories to follow.  Log data from the pilot well would be used to pick the best drainhole location. In addition to determining the drainhole trajectory, drilling the pilot hole gave the team an opportunity to learn how directional drilling equipment behaved in the VLA-8 formation.

The well was drilled vertically to a kickoff at 5350 ft [1630 m] , then with a build of 6 degree/100 ft to 50 degree deviation using a steerable bottomhole assembly (BHA). CDR compensated Dual Resistivity and CDN compensated Density Neutron measurements were added to correlate in real time with log data from nearby wells. LWD logs were later complemented by a suite of wireline measurements comprising a resistivity log, two porosity logs, a gamma ray log and the DSI Dipole Shear Sonic Imager log. Tool sticking in the build section of the pilot well, attributed to overbalance caused by low reservoir pressure, precluded logging with a dipmeter tool. The lack of dip information near the well created a formidable challenge when it came time to drill the horizontal drainhole.

Log data from the pilot well were fed into the ELAN Elemental Log Analysis progam, which fits openhole log measurements to a formation model comprised of mineral and pore fluid combinations. The ELAN results showed that the C-7 upper sand,  with higher clay content than the other sands, had lowest effective porosity, but the highest hydrocarbon saturations. Logs of the shaly C-7 upper sand indicated oil in the top 40 ft, with a water leg in the clead sand section below. Consequently, the team directed their attention to the C-7 upper rather than lower sand. 

Next, petrophysicst used the impact Integrated Mechanical Properties Analysis Computation Technique program to evaluate whether the C-7 upper sand could support a horizontal drainhole. The Impact program processes a variety of data - including bulk volume analysis from the ELAN output, vertical and horizontal stresses derived from logs and core measurements, and density logs - to calculate the stress field at the borehole wall for a given well inclination and direction. 

More importantly, it establishes safe mud weights along the trajectory in the borehole. The mud-weight range indicates the degree of difficulty and expense associated with drilling a horizontal well.

Vertical stress was derived from log measurements of the cumulative density of overlying sediments. Horizontal stresses were obtained using diferential strain-curve analysis. In this technique, strain gauges are attached to a core sample, which is then encased in a silicone plug and compressed hydrostatically. Hydrostatic compression closes microcracks that developed when the core was removed. Measuring strain while these cracks close gives the ratio of the horizontal stresses.

Analysis of DSI data gave the compressional and shear velocities needed, along with the bulk density, to compute the dynamic elastic moduli. These computations matched the elastic moduli measured on cores prior to strain curve analysis. The Impact analysis showed the zone to be competent and drillable at high angles. 

In finalizing the horizontal trajectory, the team correlated pilot log data with offset data from two nearby wells, which showed that the C-7 upper dipped up about 5 degree from the pilot well, then flattened out and eventually started dipping down. A 6 degree/ 100 feet build to 95 degree was planned to intersect the target sand at 6380 ft true vertical depth (TVD). Markers that could be identified with the LWD gamma ray or resistivity sensors were chosen to verfiy the approach to horizontal.

Drilling the Horizontal Drainhole

The drainhole was geosteered with an LWD system, providing real-time gamma ray and resistivity logs. Density and neutron porosity data were recorded in downhole memory and used to locate gas, which, if detected, would affect completion strategies. A sedimentologist at the rig analyzed drill cuttings to monitor the location of the drainhole in the target.

The pilot hole was plugged and opened up to allow setting 9 5/8 in. casing at 6062 ft. When drilling began, the close interaction between Maraven and Schlumberger geoscientists was important in allowing the team to respond swiftly to new developments. 

The pilot-hole logs were used to construct a dipping, "layer-cake" resistivity model that could simulate the LWD resistivity response in a drainhole for any depth and deviation. These simulated tool responses would guide the LWD interpreter in advising the driller once real-time LWD logs were available. When simulated and measured resistivities differ, the model is modified by adjusting the dip of the bed with respect to the borehole angle or the depth of the structure. This process is repeated until the simulated and measured resistivity measurements match, indicating the correct model for the depth and dip of the structure.

As the horizontal section began, early correlations between the simulated and real-time LWD measurements indicated a steeper dip than expected. To compensate, the team increased build angle from 6 degree/ 100 ft up to 16 degree / 100 ft. Even so, the drainhole exited the oil section at 6748 ft, striking water. 

  The driller jacked up the inclination to 100 degree to steer the wellbore up toward the oil bottom. Once it was found, the hole was plugged back. LWD logs from this drainhole provided the team with dip information crucial for revising the formation model. The team accounted for the effects of azimuthal changes and high transverse dips relative to the well path, caused by the beds dipping up about 35 degree toward the fault. Any change in azimuth to the left would cause the drainhole to lose elevation in the oil section. A turn to the right would cause a gain in elevation.

 The revised strategy was to land the drainhole in the upper section of the target. Once there, a 95 degree inclination would follow the dip until the CDR curve indicated the well path exiting into overlying shale. The azimuth would be closely controlled.

The new drainhole entered the top of the sand at 6750 ft [2057 m] with an inclination of 87.6 degree. Logs across the target sand from the pilot were then used to navigate the drainhole, with gamma ray and resisitivity measurements from the LWD tool as indicators. The top of the C-7 upper interval contains a series of thin sands and shales each with an identifiable gamma ray signature  . Alternating changes in sand and shaliness as found from the pilot were assigned letters from "a" to "k" -"a,c,e, " indicated sands , "b,d,f..." indicated shales. As the wellbore progressed, correlations indicated that the drainhole had penetrated sandy "a" through shaly "f". Then the team decided to steer up to avoid hitting water. This correlation of sedimentological facies between pilot and drainhole proved to be a powerful geosteering technique. The drainhole reached its planned displacement with 1112 ft [339 m] of net pay sand.

Completion and Production

Several factors influenced the completion of VLA-1035. The overall strategy was to produce through slotted liner, but this hinged on the ability to slide the slotted section to the total depth (TD). An openhole gas section below the 9 5/8 in. shoe found with the CDN log needed to be hydraulically sealed from oil production. In addition, low reservoir pressure of 1200 psi required artificial lift for production.

Consequently, 7-in. casing with the lower section slotted for production, rather than liner, was set to surface. This would provide the pushing power to reach total depth and guarantee gas isolation. CemCADE cementing design and evaluation software analysis was used to determine pump rates, fluid volumes, surface pressures and centralizer calculations for cementing. The 7-in. casing was run to the bottom of the hole, and an inflatable external casing packer was placed above the slotted section to isolate the gas. A port collar placed at the top of horizontal slotted section directed the cement first into the packer and then up the annulus between the 7-in. casing and the openhole and 9 5/8 in. casing. After the packer was inflated, cement was pumped 1500 ft above the 9 5/8 in. shoe to provide the hydraulic seal. Finally, 3 1/2 in. tubing equipped with two gas lift mandrels was run.

During the first two weeks of production, chokes ranging from 3/8 in. to 1 in. were tested, with the largest diameter yielding 2456 BOPD. With a 5/8 in. choke, the well averaged 1400 BOPD and a 4% water cut in the first five months and continues to produce 1000 BOPD with a 12% water cut today.

Maraven has since drilled two additional horizontal wells with Anadrill in Lake Maracaibo, including a reentry well, and is studying the optimal length of a horizontal drainhole. In 1994, Maraven plans to drill 11 horizontal wells, building to 36 in 1999. By the year 2000, horizontal wells are predicted to account for 20% of oil production in Venezuela. 



No comments:

Post a Comment