Field challenges

The Casabe field, discovered by Shell in 1941, is located in the mid- Magdalena Valley basin. The structure is an asymmetrical anticline with moderate dip to the east, affected by transpresional faults. It includes 1,120 wells, which by December 2008 had accumulated 297 MMbbl of oil.

The productive formations include La Paz, Mugrosa, and Colorado, with depths ranging from 2,200 to 5,500 ft (670.5 to 1,676 m). Production peaked at 46,000 b/d in 1953 and achieved a primary recovery factor of 13% under natural mechanisms.

Production under primary mechanisms continued until 1985 in the northern area and 1989 in the southern, at which time waterflooding commenced. Waterflooding was conducted with regular five-spot patterns with up to four wells per injection location and up to two wells per production location (Figure 1).

Waterflooding raised recovery to 19.8% by 2007.Waterflooding was a challenge due to the heterogeneous nature of the reservoirs, sand continuity complexity, oil viscosity, sand production, and wells lost due to collapses.

Beginning in 2004, Ecopetrol, S.A. and Schlumberger combined efforts to revitalize this mature field to increase its value. The field redevelopment plan embraced 3-D seismic data acquisition, selective water injection, appraisal wells, and technology application, as well as facilities upgrades to handle incremental production and injection volumes.

History versus future

The Casabe field was initially under-saturated, and produced under primary mechanisms (mainly natural depletion) until the mid-1980s, when waterflooding was initiated (Figure 2). Most of the gas has already been produced, and oil rates were 10 b/d per well. The waterflooding tripled the oil rates per well, but remaining inefficiencies made it difficult to obtain better performance.

The field had been declining steadily at 7% per annum since its secondary production peak of 14,000 b/d of oil in the early 1990s. Figure 2 shows the Casabe production and injection behavior for the last four decades. Ecopetrol entered an alliance with Schlumberger in 2004, aimed at positively impacting field production, as the rate was declining at a sustained yearly rate.

Figure 2 also shows two periods of water injection rate peaks, in 1986 and 1991. These correspond to moments when waterflooding processes took place on the north and south areas, respectively. A common denominator is the sudden rate drop that took place two to three years after the peaks. This occurred at the end of fill-up once the free gas had been largely displaced from the injection region.

Injection loss also was a consequence of choking back injection due to water channeling, sand production, and lost wells due to casing collapse. Casabe is heterogeneous in nature, highly layered, and water goes preferentially through the most conductive sands. Water channeling is aggravated through the following mechanisms:

a. Shallower sands are usually unintentionally fractured during waterflooding. Other layers gradually plug due to water quality causing adverse injection profiles.

b. As water displaces the Casabe viscous oil, and once breakthrough takes place through a layer in the producer, injectivity— or more correctly injectivity-productivity index (IPI)—goes up in that layer. As a consequence the injection profile is controlled by that single layer, adversely affecting the vertical sweep efficiency. This is a distinctive feature during production operations.

Well production and injection histories were analyzed to establish the extent and causes of poor sweep. The principal findings were:

Poor vertical sweep due to:

• Large permeability variations;

• Single fracture in most injectors;

• Skin damage in injectors (unfractured zones);

• Poor mobility ratio;

• High IPI after breakthrough;

• Limited injection profile control;

• Sanding of injectors; and

• No production profile control.

Low injection rates and voidage replacement ratio due to:

• Damage in injectors;

• High-permeability layer shut-off;

• Choked-back injectors;

• Manifold pressure too low for deeper sands;

• Low IPI due to high viscosity; and

• Excessive injector shut-in during drilling. Poor areal sweep due to:

• Poor mobility ratio;

• Loss of wells causing pattern distortion;

• High, variable skin in producers;

• No pattern rate balancing;

• Highly variable voidage replacement; and

• Incomplete waterflood coverage due to lack of injection in some areas.

IPI

The IPI model was developed for Casabe to help explain injector behavior as well as water cut changes in producers. IPI combines the injectivity index of an injector with the productivity index of an associated producer. IPI is defined as the flow rate between an injector/producer pair divided by the difference in the bottomhole injection pressure and the bottomhole producing pressure.

The IPI model helps clarify some key features of waterflooding with a poor mobility ratio:

• Once water breaks through in one layer the IPI increases in that layer and dominates the flow, causing poor vertical sweep.

• The skin at the producers is multiplied by the high oil viscosity; this has a large effect on the IPI and therefore the injectivity.

• If skin damage in the four producers associated with an injector are unequal —e.g., due to gravel packs, fractures, etc.—then the IPIs from the injectors to the producers will be unequal, causing most injected water to flow towards the producers with the lowest skin, resulting in poor areal sweep.

• A Hall plot can be usefully modified to use the integral of bottomhole pressure difference (injector pressure minus producer pressure)—instead of the injection pressure—so that the slope gives the total IPI of the injector.

• A3-D IPI model can be used to determine the rates of the waterflood flow regulators required to optimize the injection profile for vertical sweep.

Key decisions made

Before creating a final plan, some strategic decisions were formulated. In regard to water management, these key decisions included the following:

• Improve vertical sweep efficiency, using selective completions;

• Improve areal sweep efficiency, by reestablishing patterns selectively; and

• Increment injection rates, to speed up the recovery.

The first two decisions would clearly impact and add reserves.

Improved use of technology and teamwork that took advantage of worldwide experiences became the basis for the field redevelopment.

Figure 2 shows the oil production and water injection rates after the alliance started in March 2004, with production continuously improving.

By December 2008, the production reached 12,200 b/d and by April 2009, production had already reached 13,000 b/d. The field had 5,200 b/d when the alliance started. Production would have been at 3,800 b/d had the field continued to decline at the same rate it was before the alliance. The improvement entailed water injection rate increased from 24,000 b/d of water to 60,000+ b/d of water.

In conclusion, a mature field was revitalized through an integrated approach that combines the best of two companies with different work concepts and business related philosophies. Improved water management was an important part of this effort.

Selective water injection has been key to increased recovery in the Casabe field. It also allowed reduction of investments by roughly 50% by drilling one injector instead of multiple injectors at a location.