Petrophysics
PetrophysicsIntroduction
All conventional well logs record some average of the petrophysical properties of the rocks in the logged interval. Logs have a huge range of volumetric resolution. For example, the MSFL log is a non-statistical tool that records the electrical properties in a few cubic inches of formation. At the opposite extreme, the SP and deep-reading resistivity tools record the properties of the rocks in hundreds of cubic feet.
Most logging companies also routinely apply smoothing filters to "noisy" logs, such as Gamma Ray, neutron and density logs to make these curves visually attractive.
One consequence of the smoothing effect of the logs is to make thin, clean sands appear to be shaly. If the sands contain hydrocarbons, then water saturation in these thins sands appears to be high. This problem increases as the sands become thinner.
Core and high resolution logs usually show that sand-shale boundaries are much sharper that they appear from conventional logs. The smoothing effect of the logs makes the reservoir boundaries apppear to be shaly with high water saturation.
Long established petrophysical analysis methods and reservoir summary procedures often result in pessimistic reserves estimates. For example, since wireline logs cannot completely resolve petrophysical properties near bed boundaries, the use of porosity, shale fraction and water saturation cutoff criteria to determine "net" reservoir properties will always result in pessimistic reserves estimates. This happens because values close to the cutoff value are included in the reservoir average. The problems associated with zone averaging increase as the reservoir beds become thinner.
The use of "effective porosity" estimates is very common in log analysis but is fraught with danger. The common assumptions that the shales have no porosity and that the clay/shale content occurs as pore-occluding disseminations are the most pessimistic assumption for estimates of "effective porosity". If a significant portion of the shale occurs as framework grains and/or laminations then reservoir porosity will be only slightly affected by the shale content.
Examples of Using "Misleading" Cutoff Criteria on Reserves Calculations:
Note: In the following discussion "Reserves" is in-place oil volume (millions of reservoir barrels) per square kilometer. One cubic kilometer of rock equals 6.29 million barrels. In the following three examples we illustrate the reserves impact of wireline log smoothing of reservoir properties and various reservoir summation methods.
The following simulated reservoirs are modelled as being homogenous. The apparent porosity and water saturation values are modelled as being affected by the adjacent shale beds.
Example 1: "Thick", homogenous reservoir bed
Bed thickness = 1 metre,
Water saturation (Sw) = 10 percent (Hydrocarbon saturation = 90 percent),
Porosity = 30 percent.
The "reservoir" has 10 log "samples" (simulating log analysis results calculated at 10 centimetre spacing),
Reserves: = 1.698 million barrels (Porosity * hydrocarbon saturation * 6,290,000)The following tables show the apparent "sample" values resulting from wireline log averaging at the reservoir boundary.
Apparent porosity: .10 .30 .30 .30 .30 .30 .30 .30 .30 .10 Average apparent porosity: 26.0 percent Apparent water saturation: .60 .10 .10 .10 .10 .10 .10 .10 .10 .60 Apparent average Sw: 20.0 percent The following table shows the apparent in-place oil reserves calculated using various methods.
Method Apparent reserves Underestimation Notes Exclude "non-net" samples 1.359 20 percent 8 "net" samples Zone average porosity and Sw 1.308 23 percent Depth-integrate porosity and Sh 1.409 17 percent
Example 2: "Thin", homogenous reservoir bed;
Bed thickness = 0.5 metres,
Sw = 10 percent,
Porosity = 30 percent.
The reservoir has 5 "samples" (simulating log analysis results calculated at 10 centimetre spacing),
Reserves: 0.849 million barrels (Porosity * Hydrocarbon saturation * Thickness * 6,290,000)The following table shows the apparent "sample" values resulting from wireline log wireline log averaging at the reservoir boundary
Apparent porosity: .10 .30 .30 .30 .10 Average apparent porosity: 22 percent Apparent water saturation: .60 .10 .10 .10 .60 Apparent average Sw: 30 percent The following table shows the apparent in-place oil reserves calculated using various methods
Method Apparent reserves Underestimation Notes Exclude "non-net" samples 0.509 40 percent 3 "net" samples Zone average porosity and Sw 0.484 43 percent Depth-integrate porosity and Sh 0.560 34 percent
Example 3: "Very Thin", homogenous reservoir bed;
Bed thickness = 0.3 metres,
Sw = 10 percent,
Porosity = 30 percent.
The reservoir has 3 "samples" (simulating log analysis results calculated at 10 centimetre spacing),
Reserves: 0.509 million barrels (Porosity * Hydrocarbon saturation * Thickness * 6,290,000)The following table shows the apparent "sample" values resulting from wireline log wireline log averaging at the reservoir boundary
Apparent porosity: .10 .30 .10 Average apparent porosity: 16.7 percent Apparent water saturation: .60 .10 .60 Apparent average Sw: 43.3 percent The following table shows the apparent in-place oil reserves calculated using various methods.
Method Apparent reserves Underestimation Notes Exclude "non-net" samples .170 67 percent 1 "net" sample Zone average porosity and Sw 0.178 65 percent Depth-integrate porosity and Sh 0.220 57 percent
Conclusions:The application of zone average properties and the use of cutoff criteria are both commonly used in reservoir summary calculations. However, both of these result in very significant underestimation of hydrocarbon reserves.
Depth-integration of hydrocarbon-pore-volume is the method least likely to result in pessimistic reserves estimates.
Problems associated with log averaging and inappropriate petrophysics methods increase as bed thickness decreases.
Reserves estimates can be improved by including the apparent EHT in the shale beds adjacent to the reservoirs.