It is the function of Well site Geologists to integrate processed data produced prior to and dur ing the drilling operation With their own geological observations. For this reason, it is necessary that geologists appreciate some of the. Well Logging and Formation Evaluation. Well Logging and Formation Evaluation by M. Universitetet i Bergen. Well Logging Handbook by Oberto Serra.
Geophysical Well Logging by Jay Tittman. Practical Formation Evaluation by Robert C. Formation Evaluation by Edward Joseph Lynch. The Log Analyst by Anonim. At the top are the types of reactions and processes plotted adapted after a igure from Hearst and Nelson, In general, elastic scattering is dominated by pore luids.
Neutrons can be captured when they are thermalized. With this process a compound nucleus in an excited state is formed. When it returns to the ground state, it and emits gamma radiation with an exactly deined energy see Table Diferent mineral composition and diferent pore luids result in a deviation from the porosity deined as the ratio of pore volume divided by sample volume.
Figure shows a simpliied picture of a porous rock with the various solid and luid components. Figure Neutron response to diferent minerals Baker Atlas, Chart book X-axis shows the measured porosity using a limestone-calibrated tool.
If the logged formation is a limestone, then this is identical true rock porosity y-axis. If the logged formation is a dolomite, then measured limestone-referenced porosity for example 20 p. If the logged formation is a sandstone, then measured limestone-referenced porosity for example 20 p.
Note that such charts are referenced to a speciic tool series. For a more detailed and reliable characterization of the mineralogical composition, elemental analyses based on spectral nuclear measurement have been developed. Neutron gamma spectrometric methods using pulsed neutron generators can deliver information about the concentration of various elements from gamma rays produced either by inelastic scattering or by in neutron capture events.
Elements occurring in diferent rock components are listed in Table Fundamental papers were written, for example, by Hertzog et al. For carbonates the measurement of magnesium and sulphur can be used for discrimination of calcite and anhydrite and for anhydrite volume estimate underestimating anhydrite content results in underestimation of grain density and underestimation of total porosity.
Results give information about porosity, lithology, and supports seismic exploration. Velocities vp and vs decrease with increasing porosity porosity efect; see Figure , 2. Diferent lithology sandstone, limestone, dolomite show diferent magnitudes of velocity for same porosity value matrix efect , 3. A change of pore luid from gas to water results in a strong increase of vp but a small decrease of vs pore luid efect , 4.
Velocities increase nonlinearly with increasing efective pressure pressure efect. From the strong correlation between compressional wave slowness and porosity Figure , Wyllie et al. Table gives some mean values for the components. Presence of gas can give erroneous results Asquith and Krygowski, Poor consolidation or low efective stress results in high slowness values and therefore an overestimate of porosity. For carbonates with secondary porosity, vugs, fractures with much larger dimensions than primary pores the wave propagation is still controlled mainly by primary porosity.
In deep and medium invaded reservoir zones and in water zones compressional wave propagation is controlled by mud iltrate as pore luid and the magnitude of is approximated by water slowness. From this boundary incident wave propagates as relected and refracted waves. At the so called critical angle a refracted wave propagates with the velocity of the formation along the borehole.
Diference of the arrival time divided by receiver distance gives the slowness of the formation if the tool is centered 5. A third arrival not marked is the Stoneley wave. Data from all receivers are processed for a precise and robust determination of wave arrivals compressional, shear, Stoneley ; see Figure A powerful part of these sophisticated systems is the application of correlation techniques Semblance technique. Figure Modern digital sonic tools left ofer the possibilities of a signal correlation right An important part of the tool is the transmitter.
In such cases a special transmitter dipole-shear creates a shear wave in the formation. In order to control if the cement is adhering solidly to the outside of the casing the log uses the variations in the amplitude of an acoustic signal traveling down the casing wall between a transmitter and receiver to determine the quality of cement bond on the exterior casing wall.
Hydrogen nuclei protons have a magnetic and an angular momentum. Hydrogen nuclei are a component of luid molecules of water and hydrocarbons in the pore space. Irreversible interactions result in a decrease of the echo amplitude and describe the signal decay.
Two pieces of information are extracted from the spin echo sequence: 1 Initial signal amplitude: he initial signal amplitude is controlled by the number of hydrogen nuclei associated with the pore luids in the measurement volume — thus, it gives the porosity. The envelope of the spin echo maxima decays exponentially with the time constant T2. The extrapolation of the spin echo envelope to time zero yields the net magnetization. After calibration, the net magnetization is a direct measure of formation porosity Appel Figure shows the principle of an NMR borehole tool.
For a given magnitude of the ield, the Larmor frequency is deined. Realizing the measurements within this frequency band relates the measurements to an exactly deined radial distance response space, sensitive volume.
Measured data are in the time domain. The inversion process results in a partitioning of the individual relaxation contributions bulk volumes of the three regions clay bound water CBW , capillary bound water BVI , and free movable water BVW. Regions are separated by cutofs. The discrete points on the igure are the core data. The porosity and BVI data are displayed on the far right track, The Swirr data irreducible water saturation are presented on the 2nd track from the right, and the permeability data are displayed on the 3rd track from the right.
NMR luid typing techniques are based on diferences of T1 and T2 relaxation time, difusion coeicient D , and derived properties for diferent luids Table ; properties — particularly for gas — are dependent on temperature and pressure.
As a result, data acquisition delivers a view into luid properties using not only one but also two or three parameter-domain dimensions 2D and 3D NMR ; Figure shows an example. Images have an azimuthal orientation and are characterized by extremely high resolution.
Pictures carry valuable information about bedding dip, faults, vugs and pore types, fractures etc. A gyro-based system additionally gives the azimuth and produces an oriented scan of the borehole wall. With the simultaneously measured mud velocity, a detailed oriented caliper can be derived from the two-way traveltime. Electrical Resistivity Imaging is based on the registration of the electrical resistance at the vicinity of a number of electrode buttons ixed on a pad scratching along the borehole wall conductive mud required.
Each individual button measures the contact resistance. Depending on the electrode array and borehole diameter a system of four or more electrode pads scans in several sectors. Modern tools overlap these sectors and deliver the full circle. Resistance is plotted along with orientation mostly colour-coded which again looks like a picture.
Frequently both systems are combined as one tool Figure Two connected practises are discussed: 1. Quick look interpretation for a irst scanning of logs and information about zones of interest. Algorithms and inputs are the same for any computer supported interpretation.
Data processing realizes important steps like corrections and extraction of true parameters from the measured data set for example and. Sophisticated interpretation concepts result in a combined methodology with corresponding computer programs in order to inally ind a consistent rock model.
Detailed quantitative analysis in terms of porosity, saturation, shale content, mineral composition, and permeability. Fundamentals and tools for log interpretation are: a Knowledge of characteristic log responses, b mpirically derived equations for quantitative interpretation, c Model-derived equations for quantitative interpretation, d Implementation of all information geology, cuttings, cores, … …and the experience of the interpreter.
Logs have a particular sensitivity with respect to reservoir properties, based on their physical principle Figure Figure Logging measurements used to determine reservoir properties. Blue — direct measurement of property; green — provide partial information that is combined with other measurements; brown — sensitive to property, but do not provide to property determination Andersen, Figure Permeable bed identiication — Example Baker Atlas, Porosity calculation requires the matrix properties and the luid properties as expressed in the equations let side with Greek symbols, right side with Latin alphabet as additional input: are the measured logged data.
Porosity calculation from Neutron and Gamma-Gamma-Density measurement needs luid properties. If the invasion luid front gets deep into the reservoir, the neutron and density porosity measurements approach the true porosity for the assumption of mud iltrate density and neutron response as luid properties. Under these conditions, the density porosity estimate is the true value, while the neutron porosity estimate is still low.
A root mean square RMS equation is recommended for gas reservoirs: 5. Figure Multiple porosity determination — Cases and principle. If the calculated porosities are diferent in case 2, the conclusion is that the used interpretation model matrix mineralogy, luid is not correct. For combined interpretation the input must be modiied or approximated.
If, for example, Neutron porosity limestone-calibrated tool and Density-derived porosity calculated also under assumption of calcite as matrix with density 2. Important for overlay analysis is to understand the scales for both methods. In this case there is only one scale or two identical scales. If there is dolomite or anhydrite in the formation, then density is higher than 2.
Calculation of limestone- referenced porosity therefore results in an apparently negative porosity value on the right-hand end of the scale. If the density of the formation is higher than 2. Type 2 is illustrated in Figure The logs allow a irst interpretration using following criteria: Shale: High GR, crossover of Density red curve and very high apparent Neutronporosity, Gas-bearing zone: Low GR, crossover of Density and very low Neutronporosity; high resistivity, Oil-bearing zone: Low GR, low crossover of Density and Neutronporosity; high resistivity, Water-bearing zone: Low GR, no crossover of Density and Neutronporosity; low resistivity.
If it is known that the formation is composed of diferent lithologies for example a sandstone-shale formation and this overlay technique is applied, the limestone-referenced scale delivers a deviation of the two curves. For the combination of Neutron and Densitylog, the Neutron readings limestone- calibrated porosity from the LAS-ile are plotted on the x-axis and the measured Gamma-Gamma- Density in reverse direction on the y-axis for each depth step Figure Measured data are plotted into the designed curve set.
For shaly sand a speciic cross plot can be designed with a sand point, a wet-shale point, and a water point. If colour indicates low shale content low Gamma , then the point represents a clean rock in the sandstone, limestone or dolomite region. If color indicates high shale content and the point plots, for example, in the porous dolomite region then this is not really a highly porous dolomite but could be a shaly sandstone.
Crossplot — Shaly sand A speciic application is the interpretation of shaly sands. In a Neutron-Density crossplot data points of clean sandstone are arranged along the sandstone line.
With increasing shale content the data points are shited to higher Neutron porosity clay-bound water efect. If data points are color-coded with Gamma intensity GR this efect is visible as demonstrated in Figure It characterizes the shale including the clay bound water. Figure Shaly sand model. Vi is the volume fraction of component i, i his is a system of independent linear equations; it can be solved for components deterministic solution.
Valuable input for this step comes from geological knowledge of the investigated formation, core data with respect to minerals, and cuttings data with respect to minerals. Particularly in carbonates with mixture mineralogy some minerals with very strong efect on the measured properties for example high density of anhydrite can be present.
Calcualtion of water saturation applies electrical methods, because formation water is an electrolytic conductor and hydrocarbons are insulators. Electrical methods cannot distinguish between gas and oil but the Neutronlog separates oil and gas. Tools with diferent radial depth of investigation deep-reading logs and micrologs evaluate water saturation in the non-invaded virgin zone Sw and saturation with mud iltrate Sxo in the invaded zone.
For a correct Sw—calculation from resistivity logs this second conductivity efect therefore must be eliminated. A systematic overview and discussion is published by Worthington Resistivity tools cannot resolve individual layers but measure the efect of a parallel conductor assuming current lines are parallel to the layering Figure Figure shows the result of a forward calculation of formation resistivity as a function of water saturation for diferent laminar shale content.
It demonstrates that the efect of shale is very strong at low water saturation and increases with increasing water resistivity. Figure Formation resistivity Rt as function of water saturation Sw for four diferent laminar shale contents Vsh. Neglecting the shale results in a non-realistically high water saturation or low hydrocarbon saturation. Figure demonstrates the efect of the shale upon the result of a Sw-calculation for three shaly sand equations. Shale content Vsh , 2.
Shale and distribution type Vsh-lam, Vsh-disp , 3. Shale properties R-shale,. In general, borehole measurements for water exploration apply a small number of logs and tools and depth of the wells is smaller compared with hydrocarbon exploration. Two examples may illustrate typical applications.
Sand aquifer layers are marked and the completion design screened sections is derived. Ater completion a second series of logs was measured right part of Figure : Temperature log and diferent modes of lowmeter. Flowmeters measure the vertical low speed of the water in the well. Frequently an impeller lowmeter is used in water exploration.
Below aquifer A the measured signal results only from aquifer B and from the speed of the tool. And below aquifer B there is no inlow — the measured signal represents the tool speed efect. If aquifers are not artesian a pump above the uppermost aquifer can activate the inlow. If the tool passes a permeable screened section, a part of the displaced water lows behind the screen crossing the permeable gravel pack. If the gravel pack is not permeable then the step is less steep and indicates the well completion is of low hydraulic quality.
All screens are at the correct depth position and have a full hydraulic function indicated by high permeability. A Flowmeter measurement pump was located at a depth of m conirms the inlow in this section; spikes on the log result from caliper irregularities in the open well.
In chalk reservoirs, compaction may well be an issue during field life. While this has the advantage of providing an additional pressure support mechanism, extensive studies will be required during the design phase of any installations, particularly offshore.
I have worked in some fields having varying amounts of limestone, marl, anhydrite, dolomite, siderite, pyrite, quartz, and clays where a conventional approach using deterministic equa- tions is not reliable. The basic way programs using this tech- nique operate is as follows: 1. The various minerals and fluids to be included in the model are determined.
Overall, this sounds like a very rigorous approach to a conventional sort of deterministic evaluation, which would be preferred in all cases. Reasons why it is not are as follows: 1. The fact that the program is capable of calculating back the correct log responses does not mean that the results are necessarily correct. The program does not have any depth-dependent reasoning capability.
Many of the tool responses for minerals in the formation are not known accurately and recourse is made to standard tables of typical minerals. Porosities derived by the program will typically come about from a combination of the density, neutron, and sonic responses, plus what- ever assumption is made about the relative saturations in the invaded zone as opposed to the virgin zone.
I am dubious about how correct such a porosity really is. The programs are very sensitive to log quality and noise on the log traces. Where even one log is reading wrong, the volume fractions will be affected and the output may be completely unreliable.
Overall, my impression is that there are some cases in which statistical models offer real advantages over conventional interpretations. A good example is in a sandstone reservoir having variable amounts of siderite or pyrite. The program is also useful in a normal sandstone reservoir where there are limestone stringers intermittently present. In situations where the mineralogy is not well known or the logs are of poor quality, I am extremely dubious about the quantitative correctness of the output.
Even when neither of these situations arises, it is still my experience that it is necessary to make dozens of runs, investigating the effects of minor changes to the input parameters, before a solution can be produced in which one can have any confidence. All the main logging contractors are able to offer software for statistical analyses, which can be run either by the contractor or in-house by the oil company.
The values for typical parameters for various minerals are usually built into the software as default values, so will not be repeated here. In this respect it offered to solve one of the problems occasionally confronting petrophysicists, namely, low-resistivity pay. Advanced Log Interpretation Techniques 77 The basic principle by which the tool operates is as follows. The tool is assumed to respond only to hydrogen nuclei in water, oil, and gas in the porespace.
The hydrogen nuclei which are just protons in the pore- fluids have a spin and magnetic moment that may be affected by an exter- nal magnetic field.
In the absence of an atomic field, these moments are aligned randomly. When an external field B0 is applied, a process occurs whereby the orientation of the nuclei changes so that a proportion of them align in the direction of the applied field H. The reason they do not all immediately align in the direction of the field is that two adjacent nuclei are in a lower energy state when they are aligned in opposite directions.
In the absence of any further fields being applied to the horizontal magnetic components, the individ- ual nuclei will be randomly distributed and sum to zero. Now consider what happens if, after a period denoted by Tw wait time , a horizontal magnetic field is applied at a frequency equal to the Larmor frequency.
After a time given by t the echo time , a pulse is given at degrees to the direction of B0 called a p pulse. The nuclei start to align themselves in the opposite direction. However, because of differences in their horizontal relaxation times, the magnetic moment building up in the opposite direction will be less than during the first pulse.
A third pulse is then applied in the original direction of B0 with correspondingly even less buildup of moment.
In practice, not all the fluid in the pores will relax according to the same T2. Those lying close to the pore wall will relax more quickly than those in the center of the pore. This means that there is a series of contributions to Mh, each decaying at a different rate. In addi- tion, because of diffusion of the nuclei within the pores, nuclei that may not initially be close to the wall may move toward the wall during the measurement and relax more quickly.
Because different fluids oil, gas, water have different values of D, if measurements are done at different values of t, there is the possibility of differentiating fluid type. This is the basis for what is called time domain analysis TDA. This is the basis for the T2 cutoff commonly used: 33 ms for water-wet sandstones and ms for carbonates.
In fact the relaxivity may vary consider- ably among different types of rock. Values as low as Times as long as ms have also been seen in sandstones drilled with OBM. If the appropriate T2 cutoff is not a constant but is facies dependent, significant problems are caused in determining permeability accurately.
In fact, a conventional poroperm approach normally works quite well if a different relationship is used according to facies type. A lot of the poten- tial benefits from NMR are removed if one requires both core T2 mea- surements for all facies types and a means for determining which facies is being logged.
Applying the gas correction may affect the permeabili- ties by a factor of up to The total porosity of the sample is related to the strength of the initial signal occurring from the tool following the first transverse pulse during a T2 acquisition.
This is referred to as incomplete polarization, to which polar- ization correction may be applied. The HI is influenced by temperature, pres- sure, and salinity, as well as the fluid type water, oil, or gas. Because clay-bound water relaxes very fast T2 of a few ms , a special mode of acquisition is required to measure total porosity.
In a normal acqui- sition mode, the tool will respond to only capillary-bound and free fluids. It should be noted that it is normally assumed that the rock is water wet. This means that any short T2 arrivals are the result of the wetting phase relaxing close to the pore wall. Other fluids, such as gas and water, are assumed to be far from the pore wall, so that one sees only their bulk fluid relaxation times. Any kind of TDA, which exploits the differences in D, exploits this fact.
Even where WBM water-based mud has been used, any mixed wettability in the formation will tend to result in anomalous results. The tool will measure a decaying magnetic amplitude vs. Because the fixed magnet is located in the borehole, with the magnetic field decreasing with distance from the borehole, this will define the zone of investigation of the tool. Having measured the transverse signal as a function of time, the next step is to invert these data into the corresponding distribution of T2 values that make up the signal.
This would be a straightforward mathematical operation were it not for the presence of noise in the signal. In fact, without some additional form of constraint, at the noise levels typically encoun- tered in the tool it is possible to produce wildly different T2 distributions that can all honor the original decay curve.
One constraint that is com- monly applied, called regularization, is that the T2 distribution must be smooth. This results in a more stable solution, although there is no par- ticular reason why the T2 spectrum should indeed be smooth.
Needless to say, unless the inversion is correct, the results of the tool will be com- pletely useless. This is worth bearing in mind in situations where the tool gives results that cannot be explained in terms of known properties of the lithology based on core data. The maximum value of T2 that can be measured is determined by the time allowed for the signal to be mea- sured.
This in turn is related to the logging speed. The proportion of non-clay-bound or non-capillary- bound fluid to the total volume. Above the transition zone, i. Below this Pc value, FFI will comprise both the free water and the free hydrocarbon. Fluid differentiation with the tool may be performed by either the dif- ferential spectrum method DSM or the shifted spectrum method SSM.
DSM works by varying the wait time Tw, thereby exploiting the different T1 times for different fluid types. SSM works by varying the echo time Te, thereby exploiting differences in the diffusivity D between different fluids.
Often a few stationary measurements, with very long wait times and numbers of echoes, are acquired as a check that the logging speed being used is not too fast. NMR properties of reservoir fluids vary with pressure, tempera- ture, salinity, and viscosity.
Table 5. Through further TDAs, the software can produce the oil and gas satu- ration. If resistivity data are input into the software, a secondary mea- surement of hydrocarbon saturation is made. Limitations in the physics of early-generation tool were: 1. The difficulty of generating a uniform magnetic field over the parts of the formation from which the measurements were being made the T1 and T2 times are dependent on the strength of the static field 2.
Problems with the static magnetic field magnets at downhole temperatures 3. Problems with logging speed when the relaxation times could be as long as many seconds Table 5.
The difficulty of obtaining sufficient depth of investigation for the measurements to be useful for saturation determination. Most of the early drawbacks with the tool have been overcome.
These will have a great effect on the relaxation times of the hydrogen nuclei. Hence, there is additional rig time, adding to the total cost. An LWD logging while drilling version of the tool is still in the test phase.
Others point to the fact that NMR logging has been around for 15 years and has offered few real advantages in most fields. I have also seen permeabilities differ by a factor of 10 or more when com- pared with core-calibrated values derived from poroperm relationships. However, I have also seen the tool explain why some zones, with high total water saturation, are capable of producing dry oil.
In this chapter, the basic technique will be explained, together with a worked example to illustrate the principle. Consider a sit- uation in which one is using a GR gamma ray log to discriminate sand and shale. With the conventional approach, one would determine a cutoff value below which the lithology should be set to sand and above which it should be set to shale.
To use fuzzy logic, one would do the following: 1. Over the interval defined by the learning set, one would separate all the bits of GR log corresponding to sand and shale, respectively. For the sand facies, a histogram would be made of all the individual GR readings. The same would be done for all the shale values, generating a new membership function with its own mean and spread. Both membership functions would be normalized so that the area underneath them is unity.
The resulting distributions would look like Figure 5. The interval would be assigned to the one having the greatest probability. Moreover, one can assign a confidence level based on the relative probabilities. Having understood the principle of fuzzy logic with one variable, it is easy to see how it might be extended to more than two classes e.
Since it is not practical to plot more than two variables on a graph, the actual allocation is performed in a computer program in the N dimensional space corresponding to the N variables. Obviously for the method to work well, it is necessary that the membership function not 0. Also, the method does not work well with parameters that vary gradually with depth. The advantage of the method over other approaches such as neural net- works is that one is able to see, through plotting the membership func- tions with respect to a certain variable, whether or not it is applicable to include a certain variable or not.
If there is, the membership functions may be used as input to a seismic cube for allocating facies types to parts of the seismic volume, thereby showing up potential hydrocarbon zones. Fuzzy logic may also be useful to allocate certain facies types to the logs, as for instance a basis of applying a different poroperm model.
In my experience with using fuzzy logic, I have often found that one starts out with too many facies, which then are found to overlap each other.
Also, the effect of adding more log types as variables, which may be only loosely related to the properties one is interested in, is generally detri- mental.
In many respects, fuzzy logic is similar to the statistical analyses packages described earlier. In common with these, it has the advantage over deterministic techniques in that it can handle a lot of variables impar- tially and simultaneously. However, also in common with those packages, it can easily generate rubbish unless great care is taken with the input.
Fuzzy Logic 1. Set up a fuzzy logic model to distinguish between net and non-net on the basis of GR using the data from the core as a learning set. Apply the model to the lower half of the entire logged interval. Problems arise where the variations initiate a nonlinear response in the tool used to evaluate them. The most common example of this is the effect of thin shale beds on the resistivity log. When saturations are calculated, they may make the zone appear predominantly water bearing and not perforated, when in fact the water saturations in the individual sand beds are very low and the zone is capable of producing dry oil.
If the laminae are on the millimeter scale, then they will not be resolved by the density log and other approaches should be adopted. First of all it is necessary to identify the laminated zones and to deter- mine the proportion of sand to shale. The most reliable way to identify laminated sand is through direct inspection of the core. Measurements should be made to determine the relative thicknesses of the sand and shale layers as a function of depth.
Where the borehole is inclined to the bedding, the resistivity is affected less, and there are published equations for determining the resistivity measured as a function of the orientation angle and Rshale, Rsand. Another technique that has been applied is the Thomas-Stieber plot. This will now be explained. Start by considering a clean sand with poros- ity fi and containing water with a hydrogen index of HIw. There are three ways in which shale can be introduced Figure 5.
Dispersed shale may fill the existing porespace until the point at which the pores are completely filled with shale. The total volume percent- age of shale is given by Vsh. Structural shale may replace the quartz grains while the primary porosity remains the same. The total volume percentage of shale is given by Vsh. The way these processes will affect fd and fn may be predicted as follows.
Let the shale porosity be denoted by fsh and the clean sand poros- ity be denoted by fcsa. The HI of the shale is denoted by HIsh. Depending on the nature of the shale, the behavior can be seen to follow different trends.
Note that since oil and water have a similar HI, a similar behavior would be observed in an oil reservoir. A greater deviation would be observed in a gas reservoir since the HI for gas is much less than that for water. For instance, a similar behavior would be expected if the total porosity PHIT from the density fd were plotted against compressional velocity Vp or GR.
This should lead to a plot as shown in Figure 5. Above we have two equations in two unknowns Vlam, Vdis , which can be solved provided that GRsa, GRsh, fcsa, and fsh are all known or can be picked from the crossplot. In some areas it is also possible to impose an additional empir- ical constraint relating GRsa to GRsh. Having determined fsa, a conventional Archie, Waxman-Smits, or capillary curve approach may be used to determine water saturation, Sw.
Clearly it is preferred to run the tool in a packer-type mode when testing such zones. The only way to be completely sure whether a zone might be producible is through production testing. In this event I would recommend perforating the longest zone possible to give the best possible chance of encountering producible zones. In one field I have worked in, the oil contained in missed laminated sequences was such that some blocks in the field had a larger cumulative production than the calculated STOIIP.
However, when the field was reevaluated, it was found that using conventional petrophysics but remov- ing the cutoffs that had previously been applied had the effect of more than doubling the STOIIP.
In many cases it may be true that the effect of including the nonreservoir shale laminae as net sand roughly compensates for the oil volume lost from overestimating Sw in the sands caused by the effect on Rt of the shale laminae.
Advanced Log Interpretation Techniques 93 Exercise 5. Thin Beds 1. Make a Thomas-Stieber plot using fd and fn. Identify the clean sand and shale points and establish the types of clay that are present struc- tural, dispersed, laminated. Do you consider that your evaluation in this well is affected by thin beds? Also make a plot using fd and Vp. Do you learn anything additional from this plot? These neutrons get captured by atoms in the for- mation, most principally chlorine, which then yield gamma ray pulses that may be detected in the tool.
Through the use of multiple detectors, the tool is able to differentiate between the signal arising from the borehole and that of the formation. The components of the formation may be distinguished on the basis of their neutron capture cross-sections, measured in capture units c.
The contractors provide charts to predict the values of S for different rock and fluid types. Clearly the accuracy of the tool in differentiating oil and water is dependent mainly on the contrast in S between the oil and water. Hence the tool works well in saline environments and poorly in fresh environments. Even in a saline environment it might be found that small changes in the input parameters result in a large change in Sw. Hence the tool can give very unreliable results unless some of the water saturations are already well known in the formation.
The tool also has a limited depth of investigation, sufficient to pene- trate one string of casing but not always two. It is essential to have a completion diagram of the well available when interpreting the tool, so that the relevant positions of tubing tail, casing shoes, and tops of liners are known. Where the tool is used in time-lapse mode in a two-fluid system, clearly the variables relating to the nonmovable fluids drop out and changes in Sw can be calculated on the basis of only Sw - Shydrocarbon.
An example of an interpreted TDT is shown in Figure 5. What is Sw? What is the uncertainty in Sw resulting from this? In practice, one is trying to deter- mine the properties based on measurements performed in a number of wells in the field, each subject to measurement error.
Hence it is impor- tant to realize that there are two completely different and independent sources of error in petrophysical properties across a field. Firstly, there are errors arising from tool accuracy, sampling, and the petrophysical model, which will affect zonal averages as measured in individual wells. We will first deal with errors in the zonal average properties as measured in a particular well.
I believe the most rigorous way of dealing with measurement error is through the use of Monte Carlo analysis. This method has the advantage of not requiring any difficult mathematics and is easily implemented in a spreadsheet. In this example, we will attempt to estimate the error in the average properties for a simple sand that is assumed to follow an Archie model.
By analyzing the result- ing distributions, we can estimate an uncertainty range for each. From inspection of the GR we may conclude that in fact the GR cutoff could have been chosen as lying anywhere between 40 and Next we import the GR for the interval into a spreadsheet down one column.
Over suc- cessive columns we will set the cell to 1 or 0 depending on whether the log is determined to be sand or shale at the depth increment. Above the second column, the value of the cutoff chosen randomly for that run will be calculated. Hence in the cell will be calculated a cutoff value lying randomly between 40 and Let this result be denoted as GRco.
In the cells below we will determine whether or not the depth incre- ment should be designated as sand or shale. This formula is then copied down all the cells. The mean is likely to be close to that deter- mined using a nonrandom cutoff.
The next step is to calculate the uncertainty in porosity. At this stage it is advisable to remove all the intervals that are designated as non- reservoir. Determine allowable ranges for rm and rf and at the top of each column determine the values to be used in each run. If the allowable range for rm is, say, 2. Copy column B 50 times across the spreadsheet. The uncertainty in the average porosity may be taken as two standard deviations.
Since the porosity equation is linear, the mean porosity should be the same as that calculated through fixed fluid and matrix densities. Finally we will deal with saturation in an identical manner, although there are a few complexities. Advanced Log Interpretation Techniques 99 Above each column, it is necessary to derive randomly generated values for rm, rf, m, n, and Rw between allowable ranges. At this point it should be noted that applying a range to both m and Rw is not really fair if they have been determined through a Pickett plot, since any error in one will probably be corrected by adjusting the other so that the points still go through the waterline.
Hence I would apply a range to one of the two parameters only if no water sand had been available for calibration and Rw has been chosen purely from produced water samples or regional correlation. At the bottom of each column, average the Swpor and then take the mean of all the runs and the standard deviation as before. I do not believe it is necessary to take into account error in the poroperm rela- tionships, Swirr, or fluid densities, since these would be compensated for when making the log J -log Swr plot.
The second stage of the process involves looking at the uncertainties in the mean values of these parameters for individual reservoir units over the entire field. At this stage it is probably useful to digress a bit and cover some elements of basic sampling theory. The best way to estimate M is to take the mean of the IQs measured on the sample of n people, denoted by Mn. Take the mean and standard deviation of the various average porosities as measured in all the wells.
From the Monte Carlo analyses, one has denoted an uncertainty in the individual zonal average porosities of d. Most programs either require only a minimum and a maximum value for the parameters or require a mean, standard deviation, min, and max. All values within this range are treated as being equally likely.
A few concluding remarks about error analysis. With regard to the Monte Carlo analyses, it is always necessary to use good judgment when considering whether the uncertainties resulting are to be considered rea- sonable. An experienced petrophysicist should already have a good feel for the uncertainties in the average zonal parameters he is presenting, which should roughly agree with those derived from the spreadsheets.
The sampling theory presented above assumed that n, the number of samples, is large. If this is not the case e. Care should also be taken in the event that all the wells are crowded in one part of the field and there are large areas unpenetrated.
This effectively means that the sampling is not random. Advanced Log Interpretation Techniques In many field static models currently being developed, all the net sand, porosity, and permeability are input from each well in the form of logs and geostatistics applied to determine the fieldwide averages, sometimes also using stochastic approaches e.
In this case the sampling part of the above becomes redundant. Such models will typically be upscaled for reservoir simulation. Sometimes it is the case that the whole model is completely wrong; for instance, based on one sample, the reservoir is assumed to be oil filled when in fact there is a gas column occupying most of the reservoir.
This hand guide in the Gulf Drilling Guides series offers practical techniques that are valuable to petrophysicists and engineers in their day-to-day jobs. The primary functions of the drilling or petroleum engineer are to ensure that the right operational decisions are made during the course of drilling and testing a well, from data gathering, completion and testing, and thereafter to provide the necessary parameters to enable an accurate static and dynamic model of the reservoir to be constructed.
This guide supplies these, and many other, answers to their everyday problems.
0コメント