Pressure test validity shown by comparing field tests and simulations

Formation-Pressure Test Tools - 2

Neal G. Skinner, Paul D. Ringgenberg
Halliburton Energy Services
Dallas
Mark A. Proett, Reddy Aadireddy, Wilson C. Chin
Halliburton Energy Services
Houston

Field tests and simulation comparisons have demonstrated the validity of pressure test analyses for a new early formation pressure system (EFPS), with results falling statistically within acceptable ranges.

A near-well bore, finite element, wire line formation tester simulator developed in the early 1990s was adapted to the EFPS system.¹

Because mudcake prevents excessive loss of the drilling fluid into the formation, as the mudcake forms and stabilizes to a nearly steady-state condition, a pressure gradient is established in the formation so that the pressure at the sand face is higher than formation pressure.²

While the simulator can be used to model supercharging, anisotropy, formation damage, invasion, and bed boundaries, it is necessary to assume a homogeneous formation. The supercharging effect can be corrected by similar techniques developed for wire line formation testers.³

Four simulations are presented in this article with the variables shown in Table 1 [49,645 bytes] held constant. The permeability drawdown volume and flow rates were changed in the four simulation runs listed in Table 2 [25,418 bytes].

The four simulations are compared to the exact solution statistically in Table 3 [20,583 bytes] and visually in Fig. 1 [30,766 bytes] and Fig. 2 [25,154 bytes]. The simulations shown in Table 3 and Figs. 1 and 2 range from 0.1 md to 10 md with two 1 md simulations run to compare a 1,000 cu cm drawdown volume (Table 2, Vdd) to a 146 cu cm drawdown volume.

This smaller volume was used in the field examples discussed in the next section. All drawdowns start at 120 sec, the estimated time to set the packers. The simulation can predict the pressure disturbance when the packers are inflated and set against the formation, not shown in these examples.3

As discussed in Part 1 of this two-part series, (OGJ, Jan. 5, 1998, pp. 48), a shape factor (l) is needed to correct the spherical source radius (rs) to the actual packer cylindrical source area. All of the results shown in Table 3 use the same geometric corrections determined from a regression analysis of the simulations.

Figs. 1 and 2 contains plots that compare the exact solution, shown as lines, against the simulations, shown as symbols, and confirm the statistical results shown in Table 2 as the experimental variance. The experimental variance is frequently referred to as the reduced chi-square test [12,645 bytes].

In Simulations 1-4, the formation permeability was varied by two orders of magnitude, and as the results in Table 3 show, the simulations had an experimental variance within 10-6 of the exact analytical model.

In all cases, the geometric shape factor was held constant. Additional simulations have shown that the shape factor depends primarily on the ratio of the packer spacing to the well bore diameter.

In order to fully characterize the shape factor, a number of examples should be run across a range of packer spacing in relation to well bore diameter ratios, and a lookup table needs to be used to determine the appropriate geometric factor.

Field test results

A field test was conducted with the early evaluation test tool system described in Part 1 of this series, Fig. 1-4. The test was conducted at the Rocky Mountain Oilfield Testing Center (Rmotc) near Casper, Wyo.

The well was drilled in an existing field but was not expected to become a producer. A total of eight tests were conducted on a single trip. Depths for the tests ranged from 4,184 to 4,317 ft.

Hole size was approximately 81/2 in. with a deviation of approximately 30°. Low formation permeability was expected, so the system pump was adjusted to provide a volume of 146 cu cm/stroke. The tool system was in the hole for about 20 hr.

Prior to the start of the testing program, the multiple circulating rates required to operate the tool must be calculated. The tool is triggered by modulating the circulating rate to produce a particular pattern of pressures in the ID of the tool.

Information required for this calculation includes mud weight, viscosity, number and sizes of nozzles in the bit, the OD, ID, lengths of all tools and pipe in the test string, as well as details on open hole size, and the casing program.

Fig. 3 [34,726 bytes] is a plot of formation and surface pressure during the first test. The depth of the zone of interest in this test was 4,317 ft. The tool was configured such that a low-rate followed by a high-rate circulation would signal the tool to perform a test.

Multiple high-rate circulations could be performed without triggering the tool. Note that a brief, high-rate circulation followed a brief, low-rate circulation. Next, the packers were inflated by applying pipe pressure. The increase in formation pressure was an indication that the packers were inflated.

A drawdown followed by a buildup was shown shortly after inflating the packers. The tool-system pump was adjusted to withdraw 146 ml from the formation with each stroke. After a buildup of approximately 6 min, the pipe pressure was released, the packers deflated, and the top tool reverted to its original state.

A characteristic hook in the pressure history at the end of the buildup indicates the tool returned to its original position. Note that the entire test lasted only 19 min from the start of the first low-rate circulation that initiated the test sequence, to the unseating of the packers at the end of the test.

The drawdown/buildup analysis plot for the first test is shown in Fig. 4 [28,941 bytes] where the actual data are shown with a dashed line and the exact model curve fit is shown as a solid line.

Multiple regression was used to determine a shut-in pressure of 1,792.6 psi (Pf), permeability of 0.158 md (Kf), and a borehole fluid compressibility of 3.8x10-6 1/psi (Table 3, cf).

The reduced chi-square variance (x) shown in Table 3 is based on the buildup. The drawdown is short in this test example, and there is a poor correlation near the beginning of the buildup.

This is probably because of the dynamic effects the packer elements have on the effective compressibility during the sharp drawdown. During the buildup, the pressure transients are more stable, and an excellent correlation is shown in the enlarged buildup plot.

A chi-square variance of less than 10-7 was determined, making this test example within the same accuracy as the numerical simulations.

Seventh field test

For validation purposes, an additional field test is described below. Fig. 5 [31,537 bytes] is a plot of formation and surface pressure on the seventh test performed at Rmotc. This test was conducted at a depth of 4,210 ft and is similar to earlier tests except two drawdown-buildups were performed.

Note the unusual looking hump in the formation pressure data between the end of the first buildup and the start of the second drawdown. To reset the pump to perform the second drawdown, weight is applied to the pipe string.

Because the inflatable packers isolated the zone of interest from the well hydrostatic, when this weight was applied, pressure inside the inflatable packers increased in response to the applied load.

Some of this pressure increase was transmitted through the packers to the formation. This formation pressure increase is shown in Fig. 5 as the hump. For the second drawdown/buildup in this test sequence, the analysis plot is shown in Fig. 6 [29,818 bytes] for the second drawdown-buildup in this test sequence.

Multiple regression on this test was used to determine a shut-in pressure of 1,614.5 psi, permeability of 0.770 md, and a fluid compressibility of 2.77x10-6 1/psi (Table 3).

As with the first example, the buildup chi square variance was less than 10-7, showing the same accuracy as the first.

Both field test examples and the numerical simulations suggest that a larger pretest could have been used. The smaller pretest volume was used because the storage volume was originally underestimated. The field test indicated that the volume of fluid used to fill the packers influences the storage volume of the tests.

This resulted in smaller drawdowns than expected. Longer and deeper drawdowns also improve the accuracy of the test when there is a significant difference between the well bore compressibility (cs) and formation fluid compressibility (cf ).

A simulated test for a 1 md formation with a 147 cu cm test volume is shown in Fig. 1. This can be compared to the 1 md, 1,000 cu cm test volume also shown in Fig. 1, which has a much longer and deeper drawdown.

The 1,000 cu cm simulation continues to draw down where the 147 cu cm stops after a few seconds, much like the second field test example. If the field test example were conducted at 1,000 cu cm, the test would have taken approximately 1 min longer, but the quality of the test would have been improved.

Compared to a WFT, the EFPS system should yield faster buildups with improved data quality, particularly for zones less than 1 md. Wire line testers use 5-10 cu cm pretest chambers, resulting in a small radius of investigation, while the EFPS can easily use 1,000 cu cm test volume and still produce a fast buildup with a large radius of investigation.

This result was evident from the first set of EFPS system field tests. While high permeability zones (10-1,000 md) were not presented in this article, simulations have shown that the EFPS 1,000 cu cm test chamber should also improve the accuracy of pressure data and permeability estimates over a WFT.

References

Waid, M.C., Proett, M.A., Vasquez, R., Chen, C.C., and Ford, W. T., "Interpretation of Wire line Formation Test er Pressure Response, Mud Super charging and Mud Invasion Effects," paper SPWLA-92-HH paper, presented at the 33rd Annual Symposium, Oklahoma City, June 1992.
Chin., W.C., Formation Invasion, With Applications To Measurement While Drilling, Time Lapse Analysis And Formation Damage, Gulf Publishing Co., Houston, pp. 93-98, 1995.
Proett, M.A., and Chin, W.C., "Supercharge Pressure Compensation Using a New Wire line Testing Method and Newly Developed Early Time Spherical Flow Model," SPE paper 36524, presented at the 71st SPE Annual Technical Conference and Exhibition, Denver, Oct. 6-7, 1996.