LOGS DETERMINE WATER FLOW BEHIND PIPE IN ALASKA

DeWayne R. Schnorr
Schlumberger Wireline & Testing
Anchorage

Several case studies from waterflood wells in Alaska illustrate the effectiveness of impulse oxygen activation logging to determine water flow in both injection and producing wells.

The log provides a qualitative measurement of water injection and/or production rates by zone, in cased wells completed with either single or dual tubing strings.

The technique has been used in Alaskan waterfloods since 1990.

Prior to this log, obtaining adequate water data from injection wells was difficult, if not impossible. The technique also is useful in producing wells.

Schlumberger's water flow log (WFL) can measure water velocity directly behind pipe and thereby eliminate the need for radioactive tracers. The new data provide insight into ways to modify injection profiles so that oil recovery can be improved across a field.

METHOD

The WFL tool employs a pulsed neutron generator in noncontinuous mode and gamma ray detectors in various configurations (Fig. 1). The WFL measurement uses an impulse oxygen activation technique. Logging consists of a series of station measurements performed at various depths depending upon the logging objective.

When the generator is turned on, oxygen atoms in and surrounding the well bore are activated. When turned off, the activated oxygen nuclei, which now is a radioactive isotope of nitrogen, decays back into an oxygen isotope with a half-life rate of 7.13 sec.

If no water is flowing past the neutron generator toward the gamma ray detectors, the oxygen in the water around the detectors is stationary and the decay, as measured by the tool's 1-ft and 2-ft detectors, is exponential and predictable.

The tool's 6-ft and 20-ft detectors are far enough from the generator so that the background is constant and predictable. However, if the activated oxygen in the water is flowing toward the detectors, its movement will be tracked as it passes by. Knowing the generator/detector spacing and the time required for the activated water to reach the detectors, allows the water velocity to be calculated.

CAPABILITIES

A WFL log can define the water injection profile in the following cases:

• Behind tubing in wells having a selective single completion.

• In wells having two strings of tubing, allowing simultaneous injection, at different rates and pressures, into two different intervals.

• In wells having scale or particles that plug spinners.

• For each zone flowing water in a producing well.

By combining oxygen activation measurements with thermal-decay time log measurements, the WFL log provides knowledge of water saturation changes, water/oil and gas/oil contacts behind casing, and where and how much water is flowing in and around the casing.

On a single trip into a well, water can be detected flowing up the well (upflow mode) or down the well (downflow mode) with a standard tool configuration (Fig. 1), or in both directions at the same time with a modified tool. Wells flowing in two directions also can be measured with a standard tool by making two trips. Measurements can be made in flow velocities ranging from less than 2 fpm to more than 400 fpm.

If water is simultaneously flowing in the same direction both inside and outside of tubing or casing, and the velocities of these flows differ by at least a factor of three, each flow can be measured during the same trip into a well. This is possible by using three differently spaced detectors (Fig. 1), 1, 2, and 15 ft from the neutron generator, in a standard tool, and 2, 6, and 20 ft in a modified tool. The 1-ft detector is not used in the modified tool.

All detectors are activated simultaneously from the time that the neutron generator is turned off plus half of the neutron burst time. As the different slugs of activated water, inside and outside of the tubing or casing, flow at their own velocities toward the various detectors, these anomalies are measured and then their velocities can be determined.

Importantly, this technique measures and calculates a static zero flow for each detector, so that the well does not have to be shut in for a static zero-flow calibration. Often, wells that are shut-in at the surface are still cross flowing from one zone to another at the bottom of the well. This calculation is made at the same time that each water flow measurement is taken, because the zero-flow signal can change when:

• The pipe size and thickness changes from one zone of interest to another.

• The well fluids change in a well bore from the time of the no flow signal to the time of the flowing signal.

• The well bore size varies around the casing OD.

• The fluids/materials in the annulus change from gas to fluid to solid, with well depth change.

The impulse oxygen activation technique also measures background each time a water flow measurement is taken. The difference between the flowing and non-flowing oxygen activation conditions is then computed (Fig. 2). Specific information about this technique can be found in SPE Paper No. 20586, by McKeon, et al. (SPE Formation Evaluation, September 1991).

As stated, water flow velocity is determined by the time it takes activated water to flow from the neutron generator to the various detectors (Fig. 1). Impulse oxygen activation duration times of 2 or 10 sec are chosen such that velocity sensitivity and signal strength are maximized, given local needs and conditions.

Strength of the detected signal varies as a function of activation time, detector spacing, water velocity, distance from the tool to the water that is flowing inside or outside the tubing/casing, and well geometry (pipe thickness, liner laps, etc.).

Multiple cycles of data are recorded and averaged before processing because a weighted-average velocity has been determined to yield the most precise results with the greatest dynamic range. Water velocity profiles often are not constant over the generator-detector interval because of the spacing between them. Further, interpretation software automatically corrects the measured velocities to account for signal decay and loss of some signal at high velocities.

When using the tool, the maximum/minimum range of velocities that each detector can adequately measure must be taken into account (Table 1). And if one wants to generate a water profile with a vertical resolution of less than generator-detector spacing of 2, 6, or 20 ft, overlapping stations are taken at intervals that are less than the spacing.

The measured velocities are used in a constrained, weighted least-squares regression analysis to derive the velocity profile with a vertical resolution corresponding to the station intervals.

The regression analysis takes into account the source-detector spacing and changes in the flow area between the neutron generator and detector. The reconstructed flow rate may be constrained to be either non-increasing injection wells or non-decreasing producing wells. It also can be calculated without constraint.

However, if the solution is not constrained, small errors in the measured velocities can result in unrealistic flow rate changes when station measurements have been taken close together.

See the accompanying box for flow determination calculations.

KUPARUK INJECTOR

A Kuparuk injection well illustrates water flowing in the same direction at two different velocities. The oxygen activation technique and WFL tool profiled both of these flow paths, inside and outside the tubing, a task that previously was impossible.

The logs in Fig. 3 are from a Kuparuk River field, Alaska North Slope water injection well that has a selective single completion. Water is injected into two perforation intervals.

One of the intervals is below the tubing tail pipe and the other is behind tubing and isolated by a packer below and above the interval.

Water is injected into the perforated interval behind tubing through three gas injection mandrels located above and below the interval. At the time of logging with the WFL tool, all mandrels were closed and water was being injected through a hole that had eroded through the blast rings at a depth of 6,845 ft.

The logging survey was to determine the water injection profile before a polymer was injected into the interval behind tubing. The polymer would modify the injection profile by plugging the high-permeability intervals, thereby forcing the water into other intervals. A post-injection log was planned to determine the modified profile.

The initial survey immediately pointed out that the mandrels above the hole in the tubing were leaking. The 2-ft detector measured a water rate of 9 fpm in the annulus, while the water velocity inside the tubing, as measured by the 15-ft detector, was 233.5 fpm. Thus, the upper mandrels were leaking.

The hole in the tubing at 6,845 ft was obvious on the log because once the 15-ft detector moved below the hole, the water velocity slowed down dramatically. Clearly, most of the water was going through the tubing hole and up into the perforations from 6,841 to 6,847 ft into a very high-permeability zone.

The remainder of the injected water in the annulus was flowing down the annulus and into three different zones within the perforations behind the casing. Also, about 25 bw/d were being injected down into the tubing and into the interval below the tubing tail pipe, as detected by the 1-ft detector.

ENDICOTT PRODUCER

The well described by Fig. 4, from Endicott field, Alaska, produces water with oil and free gas. The logging suite included a WFL log and a monitor, Dual-Burst thermal decay time (TDT) log. Both logs were obtained during a single trip into the well because the WFL tool and thermal-decay time tool are combinable.

The WFL log determined the amount of water produced from the different perforations. Comparing the TDT log run with a previously run TDT log helped identify reservoir gas/oil contacts and changing water saturations across the zones being waterflooded. Thus the two TDT logs yield a better understanding of how the varied zones are depleted and the sweep efficiency of the waterflood.

The WFL log indicates that half of the total produced water is from the bottom perforations. This was not a surprise. However, it was surprising that the remaining water was from the two upper zones. The combination of the two logs led to this discovery.

Also surprising, the second set of perforations above the well bottom were not producing water at all, and the comparison of the TDT logs indicated no water saturation change. This zone probably does not produce anything and is not being supported by the flood program. It may be limited in size or isolated by faults.

INJECTION WELL

A water injection well (Fig. 5) had been surveyed in the past with flowmeters to define the water injection profile. However, after injected water quality deteriorated (produced water contains suspended particles) scale began to build up and bacteria became a problem, causing subsequent flowmeter (spinner) surveys to fail due to plugged tools.

Impulsed oxygen activation logging conducted with the WFL tool can obtain a water profile in this case because the tool has no moving parts that can be plugged. The tool also covers a wide range of injection rates, from 100 to 9,600 bw/d inside 7-in. casing. There is no radioactive material to handle with this procedure and an unlimited number of velocity shots can be taken.

The WFL survey found the squeeze perforations in this well to be leaking (Fig. 5). The interval from 11,418 to 11,444 ft was taking 3,065 bw/d, while the zone from 11,545 to 11,557 ft was taking 1,265 bw/d, or 61.3%, and 25.3%, respectively, of the total 3,000 bw/d being injected.

The remaining zones were taking only small amounts of water. Therefore, the injection profile needed to be modified to balance the water sweep efficiency and pressure maintenance of the different zones.

KUPARUK PRODUCER

In the Kuparuk River field a selective single completion produces water with oil and free gas from intervals below the tubing tail pipe and behind the tubing (Fig. 6). The producing intervals behind the tubing flow up the annulus, between the tubing and casing, and into the tubing through a production mandrel.

A WFL survey was conducted to determine what intervals produce water. The log indicated that the intervals below the tail pipe produce 90 bw/d, while the interval behind tubing (8,025-8,040 ft,) produces 2,050 bw/d. In a part of a larger zone (8,010-8,060 ft), a 15-ft interval is highly permeable and obviously has flood water breakthrough.

The stick plot of the measured data shows water flowing from below the tubing tail pipe, as well as behind tubing. Note that the zone behind tubing shows a steady velocity increase across the water entry, and then a decease as the water flows up the annulus. This is due to a change in annulus area.

Production enters about where the 4.05-in. OD blast rings are located on the tubing exterior. It then moves up into a wider annulus area alongside the 3.3-in. tubing itself.

COOK INLET INJECTOR

A well located on Grayling platform in the Cook Inlet has two tubing strings (Fig. 7). The two strings allow for simultaneous water injection into two different intervals at different rates and pressures.

In the past, only the zones below the longest tubing string would have been profiled to locate zones taking water. Zones above the tail pipe of the longest tubing were not profiled.

Now with the oxygen activation technique, long and short-string water injection profiles can be evaluated on one trip into the well. First, the long-string water injection profile is measured by running the WFL tool through the long tubing string and below the tubing tail pipe. This allows multi-station measurements to be taken above and within the perforated zones.

From these data, the long-string water injection profile is determined. Long-string injection is then shut in and water flow in that tubing string is allowed to become static.

Next, the short-string water injection profile is acquired with the WFL tool still in the long string. This is possible because the WFL can activate oxygen in the water that is flowing on the outside of the tubing, in the annulus.

Survey results showed that water injected below the long string is going into a zone behind a 5-in. isolation casing patch. There is a hole in this patch.

The perforated zones that are below this patch are taking almost none of the water. It was concluded that it would be better to shut-in the long tubing string rather than continue to inject water into the wrong zone. The short-string injection profile showed even distribution in all of the perforations.

Copyright 1993 Oil & Gas Journal. All Rights Reserved.