PRESSURE ANALYSIS DETAILED FOR TWO-LAYER GAS POOL IN HORIZONTAL WELL
Gerd Meissner
Mobil Erdgas-Erdol GmbH
Celle, Germany
Kwaku O. Temeng
Mobil Exploration & Producing U.S. Inc.
Dallas
Pressure derivative plots for diagnosis and semi-log pressure plots for parameter estimation permitted evaluation of a two-layer gas reservoir in a horizontal well.
Although a precise type-curve match was not possible, the investigative approach used both qualitative and quantitative interpretation of the test data.
The subject well was drilled in a grainstone layer overlying a thicker, low-permeability dolomite. A prestimulation pressure buildup was conducted on the well for the purpose of characterizing both the well and the formation.
The analysis showed that the lower permeability layer in the reservoir was affected by the pressure transient, suggesting that drainage of the reserves in that layer is possible by the well.
The positive skin factor and the reduced effective well length suggest that the well might benefit from an acid job.
THE RESERVOIR
Mobil Erdgas-Erdol has been producing gas in North Germany since 1960 and is the second largest gas producer in Germany. Both sweet and sour gas are produced from relatively deep reservoirs between 2,500 and 5,000 m (8,200-16,400 ft).
The horizontal well Siedenburg Z-17, in the Siedenburg gas field, was drilled in 1990 about 50 miles northwest of the city of Hanover (Fig. 1).
Mobil, the operator, has 50% interest in the field with BEB Erdgas und Erdol GmbH owning the remaining 50% interest.
The Siedenburg sour gas reservoir, part of the Siedenburg/Staffhorst reservoir of the Permian age Zechstein main dolomite, went on production in 1966. Original gas in place was 40 billion cu m (1.5 tcf) at a reservoir pressure of 430 bar (6,235 psia) and temperature of 133 C. (271 F.). The reservoir depth is 3,400 m (11,155 ft) and the gas contains about 7% H-S and 9% CO2 After a production of 27.7 billion cu m (1.03 tcf), reservoir pressure in April 1991 was 125 bar (1,813 psia).
The Zechstein main dolomite (Fig. 2) in the horizontal well is comprised of two layers. The upper layer of oolitic grainstone, 33 m (108 ft) net thickness and 18% porosity, is underlain by a layer of dolomite that has 87 m (285 ft) net thickness but only 6% porosity.
The horizontal well was drilled in the top grainstone layer with a horizontal section of 413 m (1,362 ft). The well was completed with a slotted liner.
In May 1991, the production test included a two-rate, pressure-transient test whose objective was to determine horizontal and vertical permeabilities, damaged skin, the fraction of the well contributing to flow, and the degree of communication between the two layers.
TESTING THEORY
Several papers have described the transient pressure behavior of a horizontal well in a single-layer reservoir. The basic flow regimes for such a system have been identified as early radial, linear, and late pseudoradial.
References 7 and 8 describe the transient pressure behavior expected from a horizontal well in a layered system.
Siedenburg Z-17 produces primarily from the upper layer that receives pressure support from the adjacent, lower layer. The flow regimes resulting from a horizontal well in a two-layer system depend to a large extent on the relative magnitudes of the properties of the layers. At early time, following the well bore storage period, the pressure behavior is governed predominantly by the properties of the primary layer. At later times, the influence of the second layer increases.
The timing and the degree to which the second layer influences the pressure behavior are primarily a function of its storativity (octV), and vertical permeability, kv2 When the storativity and the vertical permeability values of the second layer are high, relative to the primary layer, the pressure behavior approximates a constant pressure system for a significant duration.
Eventually, when both vertical boundaries (top of upper and base of lower layer) are felt, the system may attain a late pseudoradial behavior corresponding to the kh of the total system, i.e., k1h1 + k2h2.
The parameters controlling pressure behavior of a horizontal well in a two-layer system include the effective well length, Lw, the vertical well location, Zw, the storage coefficient, C, and the individual layer values of horizontal permeability kh, vertical permeability, k,, thickness, h, and storativity, OCtV.
The large number of parameters involved in the two-layer system makes it difficult to obtain precise and unique interpretation of test data. In many cases all that may be accomplished is to place reasonable bounds on well and reservoir parameters.
PRODUCTION TESTING
The well was produced for 42 days at an average rate of In 16,000 cu m/hr (14.3 MMscfd) before being shut in for 3 days in preparation for the pressure transient test. In the vertical section of the well, a memory gauge/recorder was placed in a landing nipple at a measured depth of 3,479 m (11,414 ft).
During the two-rate test, the well flowed at 14,000 cu m/hr (12.5 MMscfd) for 3 hr, followed by a 9-hr flow at an average rate of 18,500 cu m/hr (16.5 MMscfd). The well was then shut in at the surface for a 3.5-day extended buildup.
Because of the high H2S content of the gas, the test was limited to 100 hr. A total of 30,000 pressure-time data pairs were recorded during the three periods (first drawdown, second drawdown, and buildup).
Fig. 3 shows the pressure history and the diagnostic log-log pressure and pressure derivative plots for the three test periods.
The pressure derivatives were computed with respect to the superposition time function, and thus account for rate changes prior to the test period of interest. These changes include the 42-day flow and 3-day shut-in periods prior to pressure measurements.
FIRST DRAWDOWN
Fig. 3b, the tog-log pressure and derivative plot of the pressure data from the first drawdown period, shows the erratic behavior of the pressure and derivative curves that reflect rate fluctuations caused by problems in stabilizing the first gas flow rate.
It was concluded that no meaningful information could be extracted from these data and no further analysis was performed.
SECOND DRAWDOWN
As stated previously, in the log-log plot for the second drawdown period (Fig. 3c), the derivative computations account for the rate changes prior to that period.
The diagnostic plot for the second flow period has a more characteristic shape than the plot for the first drawdown period. The derivative shows a truncated unit slope followed by a transition period and then to an apparent stabilization that reflects radial flow in the reservoir (infinite-acting, semilog behavior).
The transition between full well bore storage and the infinite-acting period is determined by reservoir characteristics. Prior to stabilization, the derivative curve declines continuously over a period of about 3 hr. This decrease can result from an increase in either the system storativity, or the kh product, or both.
We attribute the behavior to the effects of cross flow from the tighter lower zone to the higher permeability upper layer. The stabilized value of the derivative corresponds to a semilog slope from which the total kh of the system can be obtained.
The slight upward kick in the derivative at about 0. 1 hr is believed to indicate the time at which the nearest boundary was encountered vertically. If the distance of the well from the nearest layer were known accurately, the time of first encounter could be used to estimate a value of vertical permeability for the primary layer.Buildup
The diagnostic log-log plot for the build-up period (Fig. 3d) is much smoother and possesses more distinctive features than the drawdown plots.
From Point 1 to Point 2 is the well bore storage dominated regime, followed by a smooth transition to Point 3.
The hump in the derivative between Points 1 and 3 suggests a slightly damaged well. At Point 3, the derivative shows a significant change. At this point, the upper boundary of the upper layer is first encountered. The timing of this occurrence shows that the horizontal well is not located in the center of the upper layer.
At Point 4 the boundary between the upper and the lower layers has been reached. the influence of cross flow from the lower layer becomes dominant and the derivative takes a downward turn.
Between Points 4 and 5 is a transitional pseudospherical flow as indicated by the negative half slope. We believe that pseudoradial flow for the total system occurs at Point 5 where the derivative appears to stabilize.
The reservoir and well parameters obtained from the build-up analysis are summarized as follows:
- Well bore coefficient C 0.29 st-tk bbl/psi
- Permeability of upper layer = 15 md
- Permeability of lower layer = 2.6 md
- Total system permeability = 6 md
- Skin factor = + 6
- Vertical distance of horizontal section to top of upper layer z, = 10 m ( 32.8 ft)
- Effective horizontal well length = 100 m (328 ft).
The well bore coefficient C was calculated from the first section of the log-log plot with a slope of 1. The semilog slope at Point 5 was used to compute the kh product for the total system and the skin factor.
The skin factor includes partial penetration, non-Darcy flow, and other effects. It was not possible to separate the different effects and therefore the computed skin factor should be seen as the upper limit of the mechanical well bore damage.
The well location, Z", and the upper layer vertical permeability, k, 1, were computed using Equations 1 and 2 for the times at Points 3 and 4 when the boundaries of the upper layer were encountered.
On the basis of core and log data and from an understanding of the depositional environment, it was assumed that the vertical permeability is equal to the horizontal permeability in both of the layers. Because of the estimated total kh, and upper layer vertical permeability, kv1, the precious assumption conveniently allows obtaining individual khs for the lower layer.
Finally using the time to late radial flow indicated by Point 5 on the derivative curve, the effective well length was estimated with Equation 3. The equation was derived for single-rate drawdown flow and, therefore, the calculation is merely a good approximation.
The calculated effective length of the well of 100 m (328 ft) is significantly less than the actual drilled length of 415 m (1,361 ft). This is probably caused by a combination of factors such as severe damage in sections of the well, frictional effects within the well, and changes in reservoir properties along the axis of the well.
The positive skin factor also suggests that even those sections of the well that are effective are also damaged.
ACKNOWLEDGMENT
The authors thank the management of Mobil Oil Corp. for the permission to publish this article.
REFERENCES
- Clonts, M.D., and Ramey, H.J. Jr., "Pressure Transient Analysis for Wells with Horizontal Drainholes," SPE Paper NO. 15116, 56th California Regional Meeting, Oakland, Calif., Apr. 2-4, 1986.
- Goode, P.A., and Thambynayagam, R.M., "Pressure Draw-down and Buildup Analysis for Horizontal in Anisotropic Media," Trans. AIME, Vol. 283, 1987, pp. 683-97.
- Daviau, F., Mouronval, G., and Curuchet, P., "Pressure Analysis for Horizontal Wells," SPEFE, December 1988, pp. 716-24.
- Ozkan, E., Raghavan, R., and Joshi, S.D., "Horizontal Well Pressure Analysis" Trans. AIME, Vol. 287, 1989, pp. 567-75.
- Odeh, A.S Ind Babu, D.K., "Transient Flow Behavior of Horizontal Wells, Pressure Drawdown and Buildup Analysis," SPEFE, March 1990, pp. 7-15.
- Kuchuk, F.J.. Goode, P.A., Wilkinson, D.J., and Thambynayagam, R.K.M. "Pressure Transient Behavior of Horizontal Wells With and Without Gas Cap or Aquifer," SPEFE, March 1991, pp. 68-94.
- Kuchuk, F.J., "Pressure Behavior of Horizontal Wells in Multilayer Reservoir, with Crossflow," PE Paper No. -731, 66th Annual Technical Conference and Exhibition, Dallas, Oct. 6-9, 1991.
- Suzuki, K. and Nanba, T., "Horizontal Well Pressure Transient Behavior in Stratified Reservoirs," SPE Paper No. 22732, 66th Annual Technical Conference and Exhibition, Dallas, Oct. 6-9, 1991.
Copyright 1993 Oil & Gas Journal. All Rights Reserved.