You must log in to edit PetroWiki. Help with editing

Content of PetroWiki is intended for personal use only and to supplement, not replace, engineering judgment. SPE disclaims any and all liability for your use of such content. More information


Reservoir pressure data interpretation

PetroWiki
Jump to navigation Jump to search

Once you have acquired bottomhole pressure data, you need to understand how to interpret the data received. Because reservoir pressure data has numerous applications, interpreting it incorrectly could skew results elsewhere. This article discusses normalization of data, static pressures, pressure-depth plots, and the effect of capillary pressure.

Applications of reservoir pressure data

Bottomhole pressure data are vital for understanding reservoir performance and predicting future behavior. Applications include:

Depth datum of pressure

Static pressures should be corrected to a fixed depth datum to eliminate the influence of the fluid pressure gradient for building isobaric maps, using bottomhole pressure to calculate inflow performance relationship (IPR) diagrams for multilayer pressure data sets, or interpreting vertical permeability barriers from a pressure differential between two reservoir layers.

Pressures are adjusted to a fixed datum by calculating the hydrostatic potential (also called the datum-corrected pressure) as follows:

RTENOTITLE....................(1)

The potentials (adjusted pressures) reflect the dynamics of fluid movement in the reservoir better than the raw pressure data can. Reservoir layers with different potentials flow into one another if put in communication (e.g., if they are completed in the same wellbore). Isobaric maps built on datum-depth-corrected pressures reveal flow within a specific reservoir layer if this layer shows different potentials in different regions of the reservoir. In addition, vertical permeability barriers are qualified in terms of potential differences between the two adjacent reservoir units separated by the barrier.

D0, the datum depth, can be arbitrary and has no influence on the interpretation of the hydrostatic potential, assuming a constant reservoir fluid density across the different wells or layers. Typically, a datum at the midpoint of the hydrocarbon column is selected to study pressure trends across the reservoir and phase behavior effects. A datum within a well may prove more useful for analyzing potential differences between multiple reservoir layers crossed by the well.

Static pressure

Static pressure measurements always result from some form of transient test, in which a large or small amount of fluid is withdrawn from the well before the pressures are allowed to stabilize. Static pressures are acquired during wireline testing at the rate of approximately one measurement every few minutes because only very small amounts of fluid samples are withdrawn. Conversely, static pressures take much longer to stabilize in conventional well testing because the much larger fluid samples withdrawn create much greater pressure disturbances.

Static pressure from buildup tests

The static pressure of a reservoir is one of the interpretation outputs of pressure transient tests. Many short-duration buildup tests (including wireline pressure "pretests") are designed solely for measuring the static reservoir pressure. The interpretation of buildup tests to determine the static reservoir pressure is discussed in fluid flow through permeable media.

Average reservoir pressure

The average reservoir pressure can be determined arithmetically by averaging the datum-corrected pressures of a given layer in all wells, with each pressure weighted by the net thickness of the reservoir at the well. A better average pressure is determined by recording the pressures, either actual or weighted, on a map of the area and drawing isobars from which the average pressure weighted for an area is determined by planimetery (or gridding) of the isobars.

Static pressure determined from the productivity index

The productivity index (PI) of a producing layer, J, is defined as the ratio of the downhole production rate of the layer to the pressure drawdown under which the layer produces:

RTENOTITLE....................(2)

On a plot of bottomhole flowing pressure vs. downhole flow rate, the PI is represented by the inverse of the slope of the IPR line describing the pressure-rate characteristics of the producing layer (Fig. 1).

Single producing layer

The static reservoir pressure and PI of a reservoir with a single producing layer can be determined with production logging measurements without the need to shut in the well. The well must be flowed at several different flow rates (typically three or four) and allowed to stabilize between successive rate changes. Bottomhole pressure and flow-rate measurements are performed for each value of the surface flow rate. The IPR is drawn through the data points on a pressure vs. rate plot, and extrapolation of the IPR line to a zero-flow condition gives the static pressure.

Below bubblepoint pressure

In most gas wells and in oil wells drawn below the bubblepoint pressure, the IPR may not be linear. Although the same procedure can be used, the IPR shape should fit the curved nature of the data (e.g., a quadratic fit if turbulence is the cause of the nonlinearity of the IPR). Once properly fitted, the y-axis intercept (zero flow) of the modeled IPR gives the static pressure, and the x-axis intercept (atmospheric pressure) gives the absolute open flow (AOF) potential. The AOF potential of a gas well is generally a better indicator of its performance than its PI because PI is not constant and the IPR is represented by a curved line. The AOF of a gas well is determined by plotting the gas potential, m(p), as a function of the flow rate of each flow period of an isochronal test. In some cases of low reservoir pressure, Δp2 can be used instead of m(p).

Multilayer completions

The same procedure applies to multilayer completions—plotting bottomhole pressure vs. rate for each layer of the system. To interpret the pressure data of combrble layers, however, the pressures must be corrected to a common arbitrary datum depth to readily differentiate whether the layers belong to the same hydraulic system. The results of the procedure, called selective inflow performance (SIP), include the static pressure and the PI per layer. The SIP procedure has become very popular for commingled producing systems, especially in gas wells because of the shorter stabilization times involved. SIP overcomes a fundamental limitation of commingled producing systems where the layer static pressures are not available by direct measurement, not even by shutting in the well, unless all the reservoir layers are in a strict hydraulic equilibrium.

Fig. 2 shows this technique applied to a multilayer reservoir comprising four layers: A, B, C, and D. The "Total" curve represents the global performance of the whole system, intersecting the pressure axis at a value that represents the wellbore pressure when shutting in the well. Obviously, this shut-in pressure differs from the pressure of each of the individual wells because the whole system is not at hydraulic equilibrium. Crossflows develop when shutting in this well, and high-pressure Layers A and B flow into depleted Layers C and D.

Pressure-depth plots

Vertically distributed wellbore and formation pressures, such as those measured by a wireline pressure tester, can be used to build mud and reservoir pressure profiles. If the measured interval is sufficiently thick, accurate pressure gradients may be established. As already mentioned, the gradients can in turn be used to spot permeability barriers and reservoir fluid contacts and to determine the reservoir fluid density.

Thick beds have a greater pressure change from top to bottom than thin beds. Therefore, the resolution of the pressure gauge becomes increasingly important the thinner the beds are. Another important factor is the number of pressure measurements taken within the bed of interest. Fig. 3 shows that increasing the number of pressure points greatly reduces the statistical error in determining the true gradient.

In some plots, the recorded pressures may not fall on a linear gradient. One example of this condition is when pressure points are not taken in a uniform depth-increasing or depth-decreasing sequence. This situation favors dispersion of the pressure measurements because of gauge hysteresis and lack of temperature stabilization. A procedure to help determine the reservoir fluid density consists of comparing the fluid density with the mud density over a set of tests taken with a wireline tester. As shown in Fig. 4, if the fluid pressures vary by Δpfl and the mud pressures vary by Δpm over the depth interval ΔD and a vertical well is assumed, then the following can be written:

RTENOTITLE....................(3)

RTENOTITLE....................(4)

then by elimination:

RTENOTITLE....................(5)

Because mud pressures are consistent over greater depth intervals, ρm is usually known. Eq. 5 then can be used to improve the reservoir fluid density determination.

Virgin reservoirs

In virgin reservoirs, the static reservoir pressures are unaffected by fluid withdrawal and the observed gradients therefore reflect the density of the original fluids. The "breaks," where the slope changes in the gradient, reflect the original fluid contacts as illustrated in Fig. 5.

Permeability barriers can also be identified as illustrated in Fig. 6. The barrier is indicated in Fig. 6a by the hydrostatic potential difference between the layers above and below the detected permeability barrier of approximately 20 psi. The line with a gradient of 0.497 psi/ft represents the mud pressure, which was measured in the same trip in the well while acquiring the formation pressure. In Fig. 6b, the reservoir fluid gradients differ across the permeability barrier. Nevertheless, a potential difference of approximately 140 psi across the barrier is interpreted as indicating a no-flow barrier. Zero permeability is implied. Otherwise, the pressure would have equilibrated on both sides of the barrier over geologic time.

Sometimes the gradients must be extrapolated to confirm fluid contacts. The gas/water contacts in Fig. 7 cannot be identified by the pressure profile of Well 1 or Well 2. By extrapolating the water gradient of Well 1 and the gas gradients of Well 2, however, it is possible to determine the position of the gas/water contacts in three zones. This extrapolation shows that pressure readings taken near the wellbore in this case reflect pressures that exist deep within the formation. From the gradient interpretation, the fluid in the upper formation is water, and there are two gas/water contacts in the lower formation.

It is important to note when extrapolating gradients from reservoir pressures in low-permeability reservoirs that the pressures may be affected by supercharging. Supercharging is caused by the nonzero, small value of the mudcake permeability. This permeability allows a finite continuous flow of filtrate across the mudcake. In a low-permeability formation, the resistance to fluid flow created by the mudcake can be on the same order of magnitude as the resistance of the formation to accepting fluid. A standard wireline pressure measurement is therefore insufficient to measure the pressure of the virgin formation because a residual finite pressure difference remains between the formation at the mudcake interface and the virgin formation some distance away. Supercharged points plot to the right of a normal fluid gradient line.

Developed reservoirs

Differential depletion is most likely to occur in developed reservoirs, destroying the original gradients. In addition, differential depletion generates vertical flow in the reservoir. Vertical flow may also result from partial completion effects that superimpose the corresponding pressure gradients on the fluid density gradients. A typical pressure-depth profile in a well drilled in a field under production is in Fig. 8. The well was completed in an interval in Zone 1. The pressure profile taken some time after initial completion clearly shows that the pressure in Zone 1 has been drawn down by fluid withdrawal. The pressure in Zones 2, 3, 4, and 5, which were not perforated, has also been affected by vertical flow through the reservoir. The measured gradients reflect the pressure drop created by vertical flow. The sharp pressure drop across Zone 2 reflects the very low permeability of this zone.

In spite of the blurring of fluid gradients in developed reservoirs, vertically distributed reservoir pressures are still useful for correlating formations hydraulically from well to well. The initial correlation made on the basis of openhole logs (left side) in Fig. 9 had to be modified because of the reservoir pressure data. Although time equivalent and present in both Wells 1 and 2, Zones A and B show different pressure regimes in the two wells and are not in hydraulic communication (right side).

Pressure probes in duplex or triplex

Taking pressure points with a multiple-probe wireline tester eliminates the uncertainty of the depth measurement for the set of points taken at a tool station. Modern wireline testers include a multiple-probe system that can measure pressure at a sink, or flowing, probe, at the same depth at a "horizontal probe" opposite the sink probe, and at a "vertical probe" at some vertical distance on a generatrix (i.e., parallel to the tool axis) with the sink probe. Both data density and data consistency increase greatly when this probe arrangement is used.

Effect of capillary pressure

Several studies have shown that a wireline formation tester actually measures the pressure of the continuous phase in the invaded region around a wellbore; typically this is the drilling fluid filtrate. The measured tester pressure is thus different from the reservoir pressure by the amount of capillary pressure. The capillary pressure affects the saturation of the wetting phase in the reservoir. The combined effects of rock wettability and capillary pressure can be reflected as changes in the pressure gradient, fluid contact level, or both on pressure-depth profiles, especially those recorded with oil-base mud in the borehole.[1][2]

The first and most conspicuous effect of capillary pressure on wireline tester pressure gradient profiles is the creation of a break in the gradient at the FWL that may not coincide with the OWC interpreted from other measurements such as resistivity logs. The depth difference increases as the displacement pressure (a function of pore-throat diameter) increases. Fine-grained reservoirs with small pore throats are most likely to exhibit potentially large depth differences between the OWC and FWL.

The second, potentially more deleterious effect of capillary pressure on wireline tester pressure gradient profiles is that the measured pressure may differ from the true formation pressure. The difference results in an unrecognized shift of the gradient to the right or left of the true reservoir fluid gradient. The effect of the gradient shift, which is equal to the amount of capillary pressure, is to displace the observed break in the gradient. Interpreting this break as the FWL yields an erroneous depth, located either above or below the true FWL, depending on the conditions described next.

Fig. 10 presents pressure-depth plots and capillary pressure profiles in a water-wet reservoir drilled with water-base mud and oil-base mud. Fig. 10c shows the capillary pressure profile expected in a water-wet oil-bearing section of the reservoir. The capillary pressure is greater in the virgin zone because oil is the nonwetting fluid. The wireline tester measures a filtrate pressure in the invaded zone that is lower because of the absence of capillary pressure. The result is a gradient shift to lower pressure values and the FWL is interpreted above its true location. There is no shift in the water-bearing section of the reservoir because the capillary pressure is zero in both the virgin zone and invaded zone. See Capillary pressure models.

When oil-base mud is used in a water-wet reservoir, the effect of the capillary pressure causes the measured pressures to differ from the true formation pressures only in the water-bearing section of the reservoir (Fig. 10b). Fig. 10d shows the capillary profile in the water-bearing section of the reservoir for this case.

Similar data and effects for water- and oil-base muds used in oil-wet reservoirs are shown in Fig. 11.

One possible method to correct for wettability and capillary pressure effects on wireline formation tester pressures is the Leverett J-function[3]:

RTENOTITLE....................(6)

Laboratory measurements of pc, k, and ϕ are used to develop a relationship for a reservoir. The amount of capillary pressure determined by the J-function is added to the measured pressure:

RTENOTITLE....................(7)

where pc(Sxo) is the capillary pressure in the filtrate-invaded zone, for which the water saturation is traditionally called Sxo.

Alternatively, if a nuclear magnetic resonance log (NMR) is available, the in-situ capillary pressure correction can be performed directly. NMR logs have the capability to model the pore-size distribution. The method also makes use of laboratory experiments on cores to calibrate the correction.[4][5]

References

  1. Elshahawi, H., Fathy, K., and Hiekal, S. 1999. Capillary Pressure and Rock Wettability Effects on Wireline Formation Tester Measurements. Presented at the SPE Annual Technical Conference and Exhibition, Houston, Texas, 3-6 October 1999. SPE-56712-MS. http://dx.doi.org/10.2118/56712-MS.
  2. Elshahawi, H., Samir, M., and Fathy, K. 2000. Correcting for Wettability and Capillary Pressure Effects on Formation Tester Measurements. Presented at the SPE Annual Technical Conference and Exhibition, Dallas, Texas, 1-4 October 2000. SPE-63075-MS. http://dx.doi.org/10.2118/63075-MS.
  3. Leverett, M.C. 1941. Capillary Behavior in Porous Solids. Trans. of AIME 142 (1): 152-169. http://dx.doi.org/10.2118/941152-G.
  4. Lowden, B. 2000. Some Simple Methods for Refining Permeability Estimates From NMR and Generating Capillary Pressure Curves. DiaLog, The On-Line Newsletter of the London Petrophysical Society 8 (1). http://www.lps.org.uk/site_pages/dialog.htm#top.
  5. Marschall, D. et al. 1995. Method for Correlating NMR Relaxometry and Mercury Injection Data. Trans., Intl. Symposium of the SCA, San Francisco, 12–15 September.

Noteworthy papers in OnePetro

Use this section to list papers in OnePetro that a reader who wants to learn more should definitely read

External links

Use this section to provide links to relevant material on websites other than PetroWiki and OnePetro

See also

Reservoir pressure and temperature

Acquiring bottomhole pressure and temperature data

Bottomhole pressure and temperature gauges

PEH:Reservoir_Pressure_and_Temperature