ANCIENT LOG BASICS
The term "ancient logs" is usually applied to the electrical survey (ES), microlog (MLC), and gamma ray neutron (GRN) that became available in the early 1930's and were in common use well into the late 1950's. At this time, induction (IES) and sonic (SL) logs gradually supplanted the ES and GRN. Early forms of the laterolog (LL7 and LL3) and microlaterologs (MLLC) were also developed to replace the ES and MLC in salt mud environments. In the early 1960's, induction, sonic, and density log presentations were fairly primitive by today's standards, and some people consider them to be ancient logs also. A brief description of all logging tools, including ancient logs, can be found in the Tool Theory section of this Handbook.
By the mid 1960's, compensated neutron and density logs (scaled in porosity units), as well as borehole compensated sonic logs, were beginning to augment porosity determination. By the mid 1970's, "ancient" logging tools disappeared from most parts of the World, but ancient tools are still run in China, the former Soviet Union republics, and other areas that could not afford newer technology.
To put this in perspective, here is a brief history timeline.
Many newer tools have evolved from the older ones since 1985. Various authors have specified alternate dates for these events - I have usually chosen the earliest.
Since the well files of the world are full of these ancient logs, we must learn to glean what we can from them. This article is a self contained coverage of how to analyze ancient logs to obtain shale volume, porosity, water saturation, permeability, and average reservoir properties. Methods that were once common in pre-computer days have been excluded because we now have better ways to do the same job.
Logarithmic scaler for reading porosity from an un-scaled neutron log. Draw vertical line on log at low porosity point (say porosity = 0.05) and another line at high porosity point (usually a shale - say porosity is 0.30). Align scaler between the two lines, setting 0.05 on scaler at low porosity line, and 0.30 on scaler on high porosity line. Skew scaler to obtain good fit. Mark other porosity points on log. Enlarge or reduce scaler in copier to fit smaller or larger logs. High and low porosity points are a matter of good judgment tempered by core or log analysis results from modern wells.
Typical GRN log with gamma ray (GR) and un-scaled neutron log (NEUT). Use the scaler to draw a porosity scale on this log.
There are special rules for picking the formation resistivity from the long normal (RESD in this book, Rt in most of the literature). These are empirical rules that work reasonably well and circumvent the need for bed thickness and borehole corrections. The rules are shown in the top half of of the illustration below..
The lateral curve on an ES log is not symmetrical and leaves a blind spot at the top or bottom of a resistive bed (pay zones). The normal practice in most parts of the world is to run the electrode arrangement, called the inverse lateral, that puts the blind spot on the bottom of the reservoir. The blind spot is the thickness of the tool spacing, 18'8" for typical tools. The blind spot shows low resistivity when it ought to show high resistivity, so the bottom 19 feet of the pay zone looks wet. If the other electrode arrangement is used, the top 19 feet look wet, as illustrated at right.
(dotted curve) and original lateral (solid curve)
The lateral curve can be used for handpicked data by following the special rules shown above. Because the lateral curve has a blind spot over the top or bottom of every resistive zone, it cannot be used in computer aided log analysis unless it is pre-processed with resistivity inversion software. However, even this sophisticated software is a poor solution, as it tends to draw a straight line through the blind zone, which you or I could do faster and cheaper with a red pencil.
In ancient wells, the logs available for shale calculation are more limited than in modern wells. The usual curves are gamma ray, spontaneous potential, and shallow resistivity. Many ancient wells have been re-logged through casing with gamma ray, neutron, and thermal neutron decay (TDT) logs. There may even be modern logs such as spectral gamma ray, sonic (compressional and shear), compensated neutron, even resistivity. There may be a large number of suitable curves to choose from. Density logging through casing is exceedingly rare, so the density neutron crossplot method for shale volume will be unavailable.
Shale volume estimation is the first calculation step in a log analysis. All other calculations depend on the shale volume being known from this step.
1: Calculate shale volume from all available methods:
NOTE: Trim values between 0.0 and 1.0. If too many values fall outside this range, check the clean and shale parameters. Do not calculate methods which fail to pass all usage rules listed below.
2: Adjust gamma ray method for young rocks, if needed:
3: Take minimum of available methods:
If SP is missing, flat, or noisy, we can calculate a
replacement SP. In hydrocarbon bearing sand shale sequences ONLY,:
must be from a 64 inch normal or a laterolog (or an old induction
log) which are symmetrical curves, and NOT from a lateral curve,
which has a blind spot on top or bottom of pay zones, depending on
electrode arrangement and spacing.
If log analysis porosity is too low, calculated shale volume may be too high (or vice versa).
The shale in the zone may not have the same properties as nearby shales seen on the log. Therefore, some adjustments to shale properties might be necessary.
Shale can be structural, dispersed, or laminated. Shale volume calculations give averages over several feet. Different distributions will affect resistivity, porosity, and permeability differently, so these calculations will be affected by assumptions about distribution. Special rules for laminated shaly sands are required and are covered elsewhere.
equation is used to find total porosity. For a dual receiver tool:
CASE 2: Correct
each layer for gas effect, ONLY IF PHIsc > PHItrue and gas is known
NOTE: It may be necessary to combine
Cases 1 and 2 to obtain a single correction factor.
Comparison of single receiver sonic (Y-axis) with 2-receiver sonic log, showing higher DTC of single receiver version
Many ancient sonic logs give unrealistic porosity values, others have chronic cycle skips, mostly to high values.
Changes in borehole size also cause spikes on the log, to both the high and low travel time directions. These are not cycle skips, but are due to the unequal travel time to each detector through the mud in the brehole.
Example of single transmitter sonic log with spikes (crosshatched areas) caused by variations in hole size. These should be edited or trimmed off before using the log data for porosity calculations.
usual density to porosity transform is used:
An alternate to the semi-log High-Low method was used with early neutron and density logs. This involves "relative count rate excursions" and service company charts of these values versus the desired rock property (porosity, density). The problem, of course, is that you need the chart unique to each tool and sufficient patience to do the constructions. Below is the instruction set developed by Lane Wells from their 1964 Technical Bulletin on their Densilog Tool.
POROSITY FROM ANCIENT NEUTRON LOGS
A large number of charts for specific tools, spacings, borehole conditions and rock types were available from service companies, such as the one shown below. These may no longer be easily found today, and the semi-logarithmic approach described below works well except in very low porosity .
There were three source types used (RaBe, PuBe, and AmBe) and several source - detector spacings (15.5 and 18.5 inches were common), combined with hole size, mud weight, and casing variations, leading to a plethora of transforms. Some service companies didn't have a lot of faith in their charts - one used the term "Strata Index" instead of "Porosity" on the Y-axis.
If no appropriate chart exists, or if you don't believe in them, it is expedient to use the "High porosity- Low porosity" method.
1. Select a high porosity point on the log, usually a shale, and assign it a porosity based on offset wells with scaled logs or a local compaction curve. This is PHIHI.
2. Pick the count rate on the neutron log at this point - this is CPSHI, even though it is a low numerical value.
3. Choose a low porosity point on the log. Assign this a porosity value, again based on offset scaled porosity logs or core porosity. This is PHILO. Tight lime stringers or anhydrite are best but you need some imagination if there are no truly low porosity streaks.
4. Pick the corresponding count rate on the log. This CPSLO, even though it is a larger number than CPSHI.
5. Plot these points on semi-log graph paper as shown below. Read porosity for any other count rate from the graph.
To use this plot in a calculator or computer
instead of on a graph:
Correct scaled neutron porosity for shale effect
Semi-log crossplot of count rate versus porosity for a group of Russian log data
To calibrate to core porosity, adjust PHIHI, PHILO, PHINSH or Vsh to obtain a better match by trial and error. Appropriate crossplots may assist.
Scaled neutron logs are also common in ancient wells, having been run through casing sometime after the original logs were run. They will have a GR curve and a neutron porosity curve (PHIN in this Handbook), the latter may have lithology, borehole, or casing corrections already applied. If it does not have these corrections, service company charts are used to apply the corrections. Read the log heading carefully to determine what has already been done.
CAUTION: In dolomite zones, many so-called compensated cased hole neutron logs did not present a rational value for porosity. This appears to have been fixed in recent years. Always compare results in carbonates with offset open hole logs or core data.
There are hundreds of charts used to perform borehole and bed thickness corrections to these curves. For typical fresh mud in an 8 inch (200 mm) borehole in a bed thicker than 8 feet, these corrections are small enough to be ignored. Charts for these corrections can be obtained on request from service companies.
simplest porosity from resistivity method is to use the shallow
resistivity and assume that the flushed zone water saturation
is near 1.0.
Choosing PHIMAX from a plot of Vsh vs PHIe from offset well - high porosity at high Vsh on this plot is from bad density data in rough borehole .
hole, bad cement, high shale volume, and statistical variations
can cause erratic results when a scaled or un-scaled neutron log
is used, Values of porosity from any method should be trimmed
by the following:
This material balance prevents the sum of shale volume, porosity, and rock matrix from exceeding 100%, and prevents porosity in the sand fraction of a shaly sand from reaching ridiculous values. It is useful for estimating porosity in shaly sands where only an SP or gamma ray log is available.
CAUTION: Bear in mind that this approach provides a porosity value based only upon the shale content and the analyst's assumed maximum possible porosity. With offset well data for control this is not a bad approach for wells with a very limited log suite. It is often used in computer analysis of ancient logs. Because of its gross assumptions, a warning note should be annotated on the results, if the method is used in this manner.
- CONVENTIONAL METHODS
Prior to about 1950, there were no laterologs, so you are stuck with ES curves. In large boreholes or salty mud, environmental corrections could be large and necessary. You will need appropriate service company chartbooks or resistivity inversion software. I recommend the latter, using SP and 16" Normal and the tool dimensions, to correct the 64" Normal.
If porosity is still doubtful, try the model in the next Section.
Calculate water saturation
SWrt becomes SWe if there is no other method available.
Calibrating Ancient to
Modern Logs (Shaly Sand)
Although the induction resistivity is focused better than a 64 inch normal, this example shows that the PHIMAX method is quite suitable in a shaly sand sequence.
Lake Maracaibo (Shaly Sand)
During a project to analyze the log and core data on 150 wells in the Western Flank Reservoirs offshore in Lake Maracaibo, we developed a technique to determine accurate values of porosity, water saturation, and permeability from old ES logs. The depositional environment is a complicated sequence of superimposed fluvial channels, resulting in many isolated channels that were not fully drained by nearby wells. It was therefore necessary to obtain a quantitative reservoir description for all wells in the project area, even if the log suite did not lend itself to direct calculation with traditional log analysis methods.
These highly detailed reservoir properties from log analysis were augmented by similarly detailed seismic and stratigraphic correlations, and integrated together in a reservoir simulator to provide an accurate historical and predictive model for production optimization. We would not have been able to do this to a useable level if only the wells with full porosity log suites were used.
The method used requires calibration to conventional and special core data and/or modern porosity log suites. Conventional core analysis data, electrical properties, and capillary pressure data was provided in paper form. This data was entered into a spreadsheet database for processing and was placed in each well file for depth plotting with the log data. Core data was depth shifted to match well log depths.
Our objective was to define a method that would utilize all available log and core data while providing the most consistent results between old and new well log suites. A detailed foot-by-foot analysis was required to allow summations of reservoir properties over each of many stratigraphic horizons.
Shale volume (Vsh) was calculated from the gamma ray (GR), spontaneous potential (SP), and deep resistivity (RESD) responses. The minimum of these three values at each level was selected as the final value for shale volume. A unique clean sand and pure shale value for GR, SP, and RESD were chosen for each zone in each well. A linear relationship was applied to the Vsh from GR. The resistivity equation for Vsh is similar to the GR equation, but uses the logarithm of resistivity in each variable.
Where a full suite of porosity logs was available, effective porosity (PHIe) was based on a shale corrected complex lithology model using PEF, density, and neutron data. The method is quite reliable in a wide variety of rock types. No matrix parameters are needed by this model unless light hydrocarbons are present. Shale corrected density and neutron data are used as input to the model. Results depend on shale volume and the density and neutron shale properties selected for the calculation. Therefore, the porosity from this stage is compared to core porosity where possible, and parameters are revised until a satisfactory match is obtained.
In wells with an incomplete suite of porosity logs, we used a model based on the shale corrected density log, shale corrected neutron log, or the shale corrected sonic log. Again, a comparison with core or nearby offset wells with a full log suite is necessary to confirm shale and matrix parameters.
In wells without any porosity logs, porosity was based on the shale corrected total porosity model, where total porosity (PHIMAX) was derived from offset wells with porosity logs or from nearby core analysis. The equation used was PHIe = PHIMAX * (1 - Vsh). This step was the most important contribution to the project as it integrates all available data in all wells in a consistent manner.
The value for PHIMAX was derived from a map of the average of the total porosity of very clean sands in modern or cored wells. The map was inspected and a transform created which varied the PHIMAX value from south to north through the project area. The effectiveness of this method is demonstrated by the close match between core and log analysis porosity in well LMA 11, shown in Figure 1. Another way to see this relationship is in a crossplot of log derived shale volume versus core porosity as in Figure 2.
In modern wells, PHIMAX is also used to limit the porosity results. This limit is needed because rough hole conditions or sonic cycle skips can cause erroneous porosity values to be computed. PHIMAX is computed as above, but modified by adding 0.03 to the result. This higher value for PHIMAX prevents the reduction of those few legitimate porosity results which are slightly higher than usual on the logs.
From this stage onward, both old and new wells were treated identically, with water saturation, permeability, and mappable reservoir properties being derived in a uniform and consistent manner.
Water resistivity (RW) was varied with depth to account for the temperature gradient over the computed interval. These values were confirmed by the obvious water zones in the lower sands in a number of wells. Care must be taken to segregate swept zones from original water zones when checking the RW value. Swept zones show residual oil on log analysis of between 20 and 60 percent. Back calculation of RW in a swept zone will lead too high a value for RW.
Water saturation (Sw) was computed with a shale correction using the Simandoux equation and with the Waxman-Smits equation. Both equations reduce to the Archie equation when shale volume is zero. Simandoux and Waxman-Smits methods gave very similar results in this project area. The resistivity curves used were the long normal from ES logs, the deep induction, or the deep laterolog.
The shale resistivity (RSH) needed for these equations was chosen by observation of the logs and crossplots. RSH was varied from well to well to account for differences in response between electrical logs, induction logs, and laterologs in shale. Resistivity anisotropy and hole size or mud resistivity effects cause these differences. The range of values used is small, between 4.0 and 5.0 ohm-m.
Values of A, M, and N of 1.00, 1.80, and 2.00 were input, based on special core analysis crossplots. The effect of overburden pressure on M and N was compared to non-overburden data on the plots where such data was available. The regression lines for M were pinned at A = 1.0 because the free regression lines vary too much, due to the small range in porosity of the core plugs.
Saturation results were confirmed by comparison to porosity vs capillary pressure water saturation crossplots derived from the special core data (Figure 3). When this data is missing in a project area, it is very difficult to refine the saturation calculation. If a mismatch does occur, the electrical properties and/or RW and temperature data must be reviewed and modified if possible, to obtain a better match to capillary pressure data.
Zones swept by production from older offset wells are evident on all newer wells in this project. These zones should not be confused with the original water zones. Swept zones will produce water if perforated, but contain 20 to 60 percent residual oil. On raw logs, the difference in resistivity between a swept zone and an original water zone may be very small (eg 0.4 vs 0.2 ohm-m in an extreme case).
An irreducible water saturation (SWir) was calculated based on a curve fit to the capillary pressure data, using the following: IF PHIe > 0.10 THEN SWir = 0.20 / (PHIe - 0.10) ELSE SWir = 1.00. This equation represents a skewed hyperbola through the porosity vs saturation data.
SWir was also limited by the Simandoux water saturation such that SWir could not exceed the Simandoux result. This means that SWir is the lower of the actual log derived water saturation and the SWir calculated above. The swept zones are most easily seen on depth plots by comparing SWir to the Simandoux or Waxman-Smits water saturation. Where large differences occur, the zone is likely swept.
Crossplots of core porosity vs core permeability (Figure 4) gave: Perm = 10 ^ (23.0 * PHIe - 3.00). Detailed crossplots of each zone in each well, composite plots of each zone for all wells, and a composite plot of all zones in all wells were made. Differences between zones and between wells were negligible. Regression analysis to predict permeability from porosity produces a good average permeability within a zone. It may not always honour every peak and valley seen on real cores.
Crossplots of permeability vs capillary pressure water saturation were also made. These show a semi-logarithmic straight line relationship. The plots show that water saturation and permeability are closely related. High water saturations indicate fine grained, more poorly sorted, lower permeability, and often shalier zones.
Crossplots of permeability vs residual oil saturation also show a semi-logarithmic straight line relationship with higher permeability having lower residual oil saturations. This is a normal occurrence, and allows a check of the residual oil saturation seen in swept zones by log analysis.
older wells, previous work used a two step correlation of oil
saturation (So) times porosity (PHI) to the short normal resistivity
(SN) and mud resistivity (RM), of the form:
This method was developed by Dr Ovidio Suarez and is documented in internal reports provided by the client. The parameters A through D were derived from correlations with hydrocarbon pore volume (HPV) estimated from core analysis. The method does not account for borehole effects, invasion, or variations in grain size, sorting, or shaliness, all of which influence HPV from this type of correlation. It also does not generate a porosity value, so results cannot be compared easily to core data and cannot be used to calculate permeability. Large differences in results between adjacent wells were noted, leading to the conclusion that these inconsistencies should be addressed in our new work.
In the porosity track of Figure 37.17 (above), the green line is porosity from SOPHI based on the SWe derived in our study: PHIrt = SOPHI / (1- SWe). This well shows a good agreement between the two methods but others do not, because the short normal is not always a good indicator for RT.
It should be noted, however, that at the time the method was invented, it was the best approach available for un-cored intervals, since modern porosity indicating logs had not yet appeared on the scene.
results of this study will lead to a significant change in original
oil-in-place compared to the value determined from a strict use
of the prior petrophysical analysis. In addition, all by-passed
pay zones are identified and can become targets for specific in-fill
wells. The reservoir simulation based on this new reservoir description
will have greater predictive power and will be easier to history
match because both reservoir volume and flow capacity are better
Page Views ---- Since 01 Jan 2015
Copyright 2022 E. R. Crain, P.Eng. All Rights Reserved