Christine A. Ehlig-Economides, Peter Hegeman
Schlumberger Oilfield Services
Houston
Gavin Clark
Schlumberger Oilfield Services
Aberdeen
Testing sequences in two very different wells illustrate the wide range of objectives that are met with modern testing procedures.
The first example is a drill stem test in an exploration well. 1 The second test is in an established producing well. 2
The exploration well test incorporated tubing-conveyed perforating, fluid sampling, production logging, and matrix stimulation to evaluate and properly treat near-well bore damage, as well as to investigate reservoir volume and characterize boundaries.
The test on the established producer evaluated whether a workover could remedy lower than expected productivity. Production logging was combined with stationary transient measurements.
This is the concluding article of a three-part series that began last month (OGJ July 18, p. 33).
EXPLORATION WELL
An Elf Congo exploration well drilled into a reservoir characterized as limestone overlying sandstone at about 7,200 ft. In the area, both the limestone and sandstone typically have 20-22% porosities and about 100 md permeabilities.
The modem well test sequence addressed two objectives. First, the completion was optimized and a strategy for subsequent field development was established by:
- Perforating the well underbalanced with tubing-conveyed guns
- Determining needed stimulation
- Evaluating post-completion well flow efficiency for both the limestone and sandstone layers.
Second, an extended test evaluated reservoir limits to indicate how many additional wells should be planned to adequately drain the observed reservoir volume.
These operations used fullbore testing equipment, run in at the start of the job and maintained in position until completion of the entire 130-hr job sequence.
TEST DESCRIPTION
Fig. 1 shows the surface flow rate and the downhole pressure recorded during the job. Four pressure gauges were used during the test, including a strain gauge, two quartz gauges, and one sapphire gauge.
The dashed portion of the line in Fig. 1 represents the downhole pressures obtained through surface readout. The solid line shows pressures stored in downhole memory and read after retrieving the drill stem test tool.
Underbalanced perforating, Point 1, was designed to minimize well bore damage as the immediate post-shot pressure impulse (Point 2) initiated perforation clean up. The recorded pressures reveal the initial underbalance and the moment of perforation.
The well was then flowed to surface for near-well bore formation clean up. This was followed by shutting in the well to run a surface pressure readout device, Point 3. The downhole pressure record signals this event.
Following a 6-hr flow (Point 4), the first test was a buildup with surface pressure readout. At Point 5, the radial flow portion of the transient data was observed as shown in Fig. 2. Calculations determined a skin of 1 and a formation permeability of 85 md.
The estimated flow efficiency indicated that matrix acidizing could significantly improve well productivity. Production resumed after pulling the surface readout device.
After a short flow period, a production logging string was lowered through the fullbore drill stem tool and run alongside the perforations (Point 6). The flow profile indicated that unlike nearby wells, the limestone had considerably less productivity than the sandstone. the sandstone productivity was as expected.
This confirmed the need to acidize.
The well was shut-in (Point 7), acidized, and then flowed back to surface for clean up (Point 8). Then for a reservoir limits test, the well was shut in again (Point 9) to reattach the surface readout device to the drill stem test tool.
To investigate reservoir volumes, boundaries, and acid-job effectiveness, a 24-hr flow period began at Point 10. The post-acid build-up data from the surface readout (Point 11), when compared to the preacid build-up data (Fig. 3) showed the effectiveness of the acid job. While the preacid skin value was 1, the post-acid data indicated a substantially reduced skin of 2.
In Fig. 3, the post-acid period has the long, 60-hr build-up with the late-time data reflecting reservoir characteristics distant from the well bore. However, once the post-acid radial flow hit a plateau and its derivative moved up and then down into a "noisy" point spread, the test was stopped.
This noise generally indicates the influence of gauge resolution limits, and further useful data cannot be obtained. Point 12 marks the removal of the surface readout device from the well bore.
After completing all pressure testing procedures, a formation fluid sampler was run (Point 13) on wire line through the fullbore drill stem tool and positioned adjacent to the perforations. The well was allowed to flow to surface at a downhole pressure just above the bubblepoint.
Next, the sampler was retried and a second production logging profile taken while the well flowed at a higher rate (Point 14). A comparison of the first and second flow profiles indicates that the acid stimulation of the limestone significantly increased the flow rate (Fig. 4).
TEST RESULTS
After completing all testing procedures, data stored in downhole memory were retrieved and analyzed. A comparison of the variances in the four pressure gauges showed that after normalization for different gauge depths, the readings were within 12 psi (Fig. 5).
The late-time behavior of the data in the second buildup period revealed that two different models compared favorably with the derivative results.
One model included a single sealing boundary (possibly a fault) and a more distant constant pressure boundary (possibly an aquifer). The other possible match used a composite radial model. A final interpretation was reconciled by integrating the well test results with geoscience data.
Thus, in 5 days, with only one trip into the hole, this exploration well was successfully tested and the information obtained led to:
PRODUCING WELL
An established production well in the Middle East experienced lower productivity than surrounding wells. A test, therefore, was designed to evaluate the well's lower zone. This zone in other wells was the main producing layer in the commingled, two-layer formation.
In particular, the objective was to determine if the inferior well performance was caused by damage to the lower zone or by reservoir boundaries.
The strategy called for measuring both transient flow rate and pressure while flowing the well.
A production logging tool was positioned just above the perforations for the lower zone (Fig. 6). From estimated formation properties of nearby wells, design calculations indicated that 4 hr of transient data would be sufficient to obtain the objectives. Fig. 7 shows the test sequence.
Following an extended shut-in, a series of three flow rate changes preceded the lowering of the production logging tool (Point 1). The production log flow profile (Point 2) indicated the percentage of flow from the lowest zone at stabilized flow conditions. For preceding flows, this same percentage was assumed. These production rates provided the recent flow-rate history for quantitative analysis of the subsequently transient data.
After the well was shut-in, Point 3, a production log survey, determined if cross flow existed in the well bore. Then to obtain transient-flow data, the production logging tool was positioned, as in Fig. 6.
Fig. 8 shows the measured transient-flow rate and pressure. On this plot, data quality is indicated by the orderly way the pressure and flow rate follow very similar trends in a mirror image. In Fig. 9a, however, the log-log diagnostic plot of pressure change and derivative exhibits considerable noise in the derivative response, obscuring the flow-regime identification.
Small variations in surface flow experienced by established wells are usually due to wellhead back pressure changes caused by communication with nearby wellheads, manifolded through a common gathering system. The remedy is to replace pressure change and its derivative data with rate-normalized pressure and derivative data 3 derived from both flow rate and pressure measurements (Fig. 9b).
Following a brief response dominated by well bore storage, the flow regime identification (FRID) tool (described in the first part of this series of articles) indicates radial and linear flow in the late-time transient period. From the radial-flow regime, calculated skin is 1.47 and permeability is 105 md.
The linear-flow trend indicated that distances from the well to parallel-flow barriers (probably sealing faults) needed more detailed interpretation techniques. After identifying that the transient response model should include well bore storage and an elongated reservoir shape, the analysis found that nonlinear parameter estimation provided the detailed pressure match as shown in Fig. 10. This match suggests that the well is nearby and centered between two parallel boundaries about 190 ft on either side.
This match with the pressure data is called a convolution-type curve.' It employs a convolution computation that accounts for measured flow rate variations and simulates the pressure behavior during the flowing transient. The boundaries detected by this test implied an elongated drainage area (parallel sealing faults confirmed by the seismic interpretation) for this well. This explained the lower-than-expected well productivity.
The drainage area geometry also indicated that reducing the skin of 1.5 with a stimulation treatment would be short lived, and therefore not cost effective. Thus, this relatively short test prevented doing a stimulation treatment on the well.
ACKNOWLEDGMENTS
The authors thank Elf Aquitaine for the exploration well test data, and Schlumberger management for supporting the production of this three-part series.
REFERENCES
- Deruyck, B., Ehlig-Economides, C.A., and Joseph, J., "Testing design and analysis,' Oilfield Review, April 1992, pp. 28-45.
- Ayestaran, L., "A two-pronged approach to testing," MEA, Well Evaluation Review, No. 6, 1989, pp. 34-45.
- Kuchuk, F.J., "Applications of convolution and deconvolution to transient well tests, SPE FE, December 1990.
- Ayestaran, L., Minhas, H.N., and Kuchuk, F.J., "The use of convolution type curves for the anal),sis of drawdown and buildup tests," SPE Paper No. 18535.
Copyright 1994 Oil & Gas Journal. All Rights Reserved.
