Not applicable.
The present invention relates generally to the evaluation of wellbore fluids, in particular to the determination of the composition of a multi-component fluid. More specifically, the present invention relates to the determination of the fluid composition of a multi-component fluid using measured physical properties of the fluid. This method is particularly useful in in-situ determination of the quality of wellbore fluid samples.
During drilling operations, a wellbore is typically filled with a drilling fluid (“mud”), which may be water-based or oil-based. The mud is used as a lubricant and aids in the removal of cuttings from the wellbore, but one of the most important functions of mud is well control. Hydrocarbons contained in subterranean formations are contained within these formations at very high pressures. Standard over-balanced drilling techniques require that the hydrostatic pressure in the wellbore exceed the formation pressure, thereby preventing formation fluids from flowing uncontrolled into the wellbore. The hydrostatic pressure at any point in the wellbore depends on the height and density of the fluid column of mud above that point. A certain hydrostatic pressure is desired in order to offset the formation pressure and prevent fluid flow into the well. Thus, it is well known in the art to control the mud density and it is often necessary to use high density “heavy” mud to achieve a desired hydrostatic pressure.
Whenever the hydrostatic pressure of the mud is greater than the pressure of the surrounding formation, drilling fluid filtrate will tend to penetrate the surrounding formation. Thus, the fluid in the formation close to the wellbore will be a mixture of drilling fluid filtrate and formation fluid. The presence of fluid filtrate in the formation can interfere with attempts to sample and analyze the formation fluid. As a fluid sample is drawn from the formation at the wall of the wellbore, the first fluid collected may comprise primarily drilling fluid filtrate, with the amount of filtrate in the mixture typically decreasing as collected volume increases. Because the fluid sought to be analyzed is the wellbore fluid, and not the drilling fluid filtrate, it is desirable to collect a sample containing as little drilling fluid filtrate as is possible.
Early formation testing tools were designed to draw in a fixed volume of fluid and transport that volume to the surface for analysis. It was soon realized that the fixed volume was not sufficient to collect a reasonable sample of formation fluid because the sample would be primarily drilling fluid filtrate. To solve this problem, formation testing tools were developed that were able to continuously pump fluid into the testing tool so that sample collection could be controlled by the operator. Using these types of tools, the operators attempt to avoid collecting filtrate in the fluid sample by pumping for a period of time before collecting the fluid sample. The amount of time used to obtain a filtrate-free sample is based on experience or intuition. The problem with this design is there is still no way to determine the quality of the collected sample without pulling the tool to the surface. Therefore, it is desirable to be able to determine the quality of the fluid sample in-situ, with the formation tester still in the well, in order to increase the efficiency and effectiveness of sample collection.
One method that has been used in an attempt to evaluate the quality of a fluid sample downhole is monitoring of a fluid property over time. One such fluid property is fluid density. There are tools available to measure fluid density downhole and plot the measured density as a function of time. As time increases, the measured fluid density in the sample volume changes until it levels out very close to the density of the formation fluid. This leveling out of the density is known as asymptotic convergence and the value of density at this point is the asymptotic value.
It is usually preferred to acquire a sample of the formation fluid when the measured properties of the sample fluid reach asymptotic levels, which indicates that the amount of filtration in the sample cannot be reduced further. The difficulty with this method is that, although an equilibrium between the amounts of formation fluid and drilling fluid filtrate entering the sample volume has been reached, the level of contamination of the fluid mixture may still not be known. For example, if drilling fluid filtrate is migrating into the formation faster than the sample is being drawn, the asymptotic level reached will still have a high percentage of drilling fluid filtrate. Therefore, there is still required a method for in-situ evaluation of a fluid sample that provides more than a qualitative, or low resolution (±10% or more) quantitative, measure of fluid composition.
One method of in-situ evaluation of fluid composition is described in WO 00/50876 and utilizes optical analysis of a fluid sample to evaluate fluid quality. The method utilizes optical density measurements to create data points to which are then used, by curve fitting techniques, to estimate an asymptotic value for the fluid density. This estimated asymptotic value is assumed to be the optical density of the wellbore fluid and is used to determine the amount of drilling fluid filtrate contamination in a sample. Optical density sensors work by transmitting a light of a specific wavelength through a fluid and measuring the optical absorption. The absorption of light as it travels through the fluid is greatly dependent on the type of hydrocarbons present in the fluid and on the wavelength of the light transmitted. Thus, the use of optical density sensors depends greatly on established knowledge of the type of hydrocarbons found in the formation fluid. This method also assumes an additional function for the change in optical density that is used to predict the asymptote and requires that the fluid be analyzed for a significant period of time before the end point (formation fluid property) can be established. Therefore, optical density evaluation does not provide real-time quantitative assessment of the fluid composition.
Mixing laws are mathematical relationships that have been used to quantify the components of a mixture and describe a property of a mixture in terms of the properties of the constituents. Thus, a the application of a mixing law may allow for the property of the mixture to be predicted if the weighting functions for the constituents, and the properties of the constituents are known. For example, mixing laws have been applied to predicting a variety of properties, including the dielectric properties of mixtures, J. P. Calame, et al, First World Congress on Microwave Processing, Jan. 5–9, 1997, and density of a rock formation as described in http://appliedgeophysics.berkeley.edu:7057/gravity/grav23.pdf.
Consequently, there still exists a need for a method for real-time quantitative assessment of the composition of fluid in a wellbore where little or no information may be available about the formation fluid. The present invention is directed to improved methods for determination of fluid composition in a wellbore that seek to overcome these and other limitations of the prior art.
Accordingly, there is provided herein a method for determining the composition of a fluid by using measured properties of the fluid. One embodiment of a method generally comprises: selecting a fluid property that is related to the volume fraction of two or more fluids according to a mixing law; measuring the selected fluid property at a series of specific time intervals and plotting the measured property as a function of the selected property. The selected fluid property preferably has a linear, or approximately linear, relationship to fluid composition, but a property having a non-linear relationship to composition may also be selected if the relationship can be extracted from a single measurement (i.e. only one unknown). In effect, the present method entails plotting the measured property as a function of fluid composition in an arbitrary set of units. This allows for a in-situ qualitative evaluation of fluid composition by measuring a fluid property that has a known relationship to fluid composition.
Another embodiment of the present invention further comprises, establishing the endpoints of contamination and plotting the measured properties through these endpoints. Once the endpoints have been established, a quantitative evaluation of the fluid composition can be performed. The endpoints of contamination (100% drilling fluid filtrate and 100% wellbore fluid) can be determined using any known method.
The selected fluid property would preferably be measurably different for the two components of the fluid, but any fluid property that has a determinable response to levels of contamination may be used. Use of this method for evaluation of formation fluid enables operators to determine the quality and suitability of the sample being collected while the formation testing tool is in the well. Measurements of other fluid properties may also be taken and plotted as a function of the selected fluid property in order to determine the response of those properties to fluid contamination.
Thus the present invention comprises a combination of features and advantages that enable it to substantially advance the fluid evaluation art by providing methods for determining the composition of a fluid mixture using measured properties of the fluid. These and various other characteristics and advantages of the present invention will be readily apparent to those skilled in the art upon reading the following detailed description of the preferred embodiments of the invention and by referring to the accompanying drawings.
For a more detailed understanding of the preferred embodiments, reference is made to the accompanying figures, wherein:
As previously discussed, it is desirable to collect fluid samples after the composition of the fluid includes a sufficient amount of formation fluids. Accordingly, there is provided herein a method for determining the composition of a fluid using data collected by fluid measurement module 30 and analyzed by surface control system 36. The fluid measurement module 30 can be any device capable of measuring a property of the fluid mixture that has a measurable response to the change in fluid composition. Examples of such properties may include: density, resistivity, viscosity, chromatography, radioactivity, dielectric constant, optical density, and magnetic resonance, weight, acoustic impedance, and acoustic velocity. Further, the fluid measurement module 30 preferably has a resolution that allows a number of measurements to be taken between the end points of contamination, i.e. 100% drilling fluid filtrate and 100% formation fluid. These end point values can be assumed to be the maximum and minimum values of the selected fluid property for the given composition.
For fluid mixtures having two primary fluid components, many properties of the mixture are functionally related to the same properties of the component fluids in direct relation to the amount of each component fluid in the mixture. If these properties are known or measured for both the fluid mixture and the component fluids, the composition of the fluid mixture can be determined. A property of the mixture (Rmix) of a two component fluid mixture can be represented by a mixing law of the functional form;
Rmix=f(R1,R2,s1,s2) (1)
where Ri are the properties of the components and si are the relative fractions of each component.
Many mixing laws are linear in terms of the relative fractions. In other cases transforms or approximations may be applied to obtain linearized responses. As examples, consider a linear mixing law given by
Rmix=sR1+(1−s)R2=s(R1−R2)+R2. (2)
This linear mixing law may be used to characterize properties such as density, as described below. In this case, if all three properties (Ri) are known or measured, the relative fraction (s) can be found using the following equation;
In cases, where one or more of the properties is not known and it becomes useful to present the property of the fluid mixture (Rmix) as a function of the relative fraction (f);
Rmix(s)=s(R1−R2)+R2. (4)
Taking the derivative of Equation 4 with respect to s gives;
which indicates that an incremental change in the fluid mixture property (Rmix) corresponds to an incremental change in relative fraction (s) by the difference between the two fluid component's properties. Although the exact value of either R1 or R2 may be unknown, their difference is constant, and therefore the measured data may be plotted on a scale of normalized contamination using the measured property.
Another example of a mixing law is
ln(Rmix)=sln(R1)+(1−s)ln(R2). (6)
Such a mixing law may be used to describe such properties as attenuation. As in the prior case, if all three properties (Ri) are known or measured, the relative fraction (s) can be found using the following equation;
In cases, where one or more of the properties is not known and it becomes useful to present the property of the fluid mixture (Rmix) as a function of the relative fraction (f);
ln(Rmix)=s(ln(R1)−ln(R2))+ln(R2). (8)
Taking the derivative of Equation 8 with respect to s gives;
which indicates that an incremental change in the fluid mixture property ln(Rmix) corresponds to an incremental change in relative fraction (s) by the difference between the logarithm of the two fluid component's properties. Although the exact value of either ln(R1) or ln(R2) may be unknown, their difference is constant, and therefore the measured data may be plotted on a scale of normalized contamination using the measured property.
Yet another example of a mixing law is
Rmix=(s(R1)n+(1−s)(R2)n)l/n. (10)
Such a mixing law may be used to describe such properties as viscosity. As in the prior case, if all three properties (Ri) are known or measured, the relative fraction (s) can be found using the following equation;
In cases, where one or more of the properties is not known and it becomes useful to present the property of the fluid mixture (Rmix) as a function of the relative fraction (f);
(Rmix)n=s((R1)n−(R2)n)+(R2)n. (12)
Taking the derivative of Equation 12 with respect to s gives;
which indicates that an incremental change in the fluid mixture property (Rmix)n corresponds to an incremental change in relative fraction (s) by the difference between the logarithm of the two fluid component's properties. Although the exact value of either (R1)″ or (R2)″ may be unknown, their difference is constant, and therefore the measured data may be plotted on a scale of normalized contamination using the measured property.
Other examples of properties which may be described by mixing laws include, but are not limited to, resistivity, viscosity, chromatography, radioactivity, dielectric constant, optical density, and magnetic resonance, weight, acoustic impedance, and acoustic velocity. When the relative fraction (s) can be determined from measurements or estimates of at least one property, it is possible to calculate other properties of one of the components using the relative fraction (s) determined previously, and measurements or estimates of the other property of the mixture and the second component. That is, if (s) can be determined from measurements of one property, some other property of the second component (R2) may be determined from this relative fraction (s) and measured of estimated values for (R1) and (Rmix) by application of the appropriate mixing laws.
One embodiment of the present invention uses the measured property of the sample fluid to provide a qualitative indication of the composition of the sample fluid. As shown in
A second embodiment of the present invention uses the measured property of the sample fluid in conjunction with known densities for one or both end point conditions to provide a quantitative indication of the composition of the sample fluid. This embodiment comprises the same steps as the first embodiment but includes using the known values of the end point fluid properties to provide a scale to the plot of measured fluid property versus normalized filtrate fraction. Therefore, a quantitative indication of the level of filtrate contamination in the fluid sample can be obtained.
One preferred embodiment relies on the selection of a fluid property that can be related to the volume fraction of two or more fluids according to a mixing law. This relationship may be linear but a non-linear relationship is also acceptable if the functional form is known. It is understood that any fluid property that has a relationship to composition that can be extracted from a single measurement (i.e. only one unknown) can be used. Additional unknown properties can also be analyzed provided that unique measurements allow for them to be determined. It is also preferred that the selected property have sufficiently different values for the drilling fluid filtrate and the formation fluid, so that the selected measurement module has sufficient resolution to obtain a reasonable amount of data points between the two endpoints. One preferred fluid property is density.
To illustrate the above described method embodiments, reference will be made to acquiring formation fluid data from a wellbore filled with an oil-based mud. Fluid sampling may be performed with a conventional formation testing tool. One preferable formation testing tool is Reservoir Description Tool as manufactured by Halliburton. A preferred density sensor is the high temperature, high pressure density sensor, as described in U.S. patent application Ser. No. 09/482,793, entitled Downhole Densitometer, and incorporated herein by reference for all purposes.
For fluid mixtures having two primary fluid components, many properties of the mixture are functionally related to the same properties of the component fluids in direct relation to the amount of each component fluid in the mixture. If these properties are known or measured for both the fluid mixture and the component fluids, the composition of the fluid mixture can be determined. For example, the mixture density (ρmix) of a two component fluid mixture can be represented by;
where m, V, and ρ are the mass, volume, and density of the first and second components of the fluid. The above equation can be simplified using f as the volume fraction of one of the components, giving;
ρmix=fρ1+(1−f)ρ2−=f(ρ1−ρ2)+ρ2. (15)
Therefore, if all three densities are known the volume fraction (f) can be found using the following equation;
In many cases, one or more of the densities is not known and it becomes useful to present the density of the fluid mixture (ρmix) as a function of the volume fraction (f);
ρmix(f)=f(ρ1−ρ2)+ρ2. (17)
Taking the derivative of Equation 17 with respect to f gives;
which shows that an incremental change in the fluid mixture density (ρmix) corresponds to an incremental change in volume fraction (f) by the difference between the two fluid component's densities. Although the exact value of either ρ1 or ρ2 may be unknown, their difference is constant, and therefore the measured data may be plotted on a scale of normalized contamination using the measured density.
Consider, for example, the case of a wellbore fluid that is a mixture of a drilling fluid filtrate with a density of 1.0 g/cc and formation fluid with a density of 0.8 g/cc.
The end point values of density are the density of the drilling fluid filtrate and the density of the formation fluid. Determination of the density of the drilling fluid filtrate may be assigned based on surface measurements of the mud system. Alternatively, it is also possible to consider the initial fluid density measurement to be essentially the density of the drilling fluid filtrate because the composition of the initial sample will be almost all filtrate. The initial sample may contain a small fraction of formation fluid, so using the initial density measurement as an end point results in consistent overestimation of the level of contamination. Determination of the density of the formation fluid may be ascertained from pressure gradient data, or other methods known in the art.
For fluid mixtures having two primary fluid components, many properties of the mixture are functionally related to the same properties of the component fluids in direct relation to the amount of each component fluid in the mixture. If these properties are known or measured for both the fluid mixture and the component fluids, the composition of the fluid mixture can be determined. For example, the mixture viscosity (Rmix) of a two component fluid mixture may be represented by;
ln(Rmix)=s ln(R1)+(1−s)ln(R2) (19)
where R1 and R2 represent the viscosities of the first and second components of the fluid, respectively and Rmix represents the viscosity of the mixture and s represents the relative fraction of one of the components.
Therefore, if all three viscosities are known, the volume fraction (s) can be found using the following equation;
In many cases, one or more of the viscosities is not known. Taking the derivative of Equation 14 with respect to s gives;
which indicates that an incremental change in the fluid mixture property ln(Rmix) corresponds to an incremental change in relative fraction (s) by the difference between the logarithm of the two fluid component's viscosities. Although the exact value of either ln(R1) or ln(R2) may be unknown, their difference is constant, and therefore the measured data may be presented on a scale of normalized contamination using the measured viscosity.
To determine the relative amount of contamination of a fluid sample, the end point values of viscosity are the viscosity of the drilling fluid filtrate and the viscosity of the formation fluid. Determination of the viscosity of the drilling fluid filtrate may be assigned based on surface measurements of the mud system. Alternatively, it is also possible to consider the initial fluid viscosity measurement to be essentially the viscosity of the drilling fluid filtrate because the composition of the initial sample will be almost all filtrate. The initial sample may contain a small fraction of formation fluid, so using the initial viscosity measurement as an end point results in consistent overestimation of the level of contamination.
Consider, for example, the case of a wellbore fluid that is a mixture of a drilling fluid filtrate with a viscosity of 2 cp and formation fluid with a viscosity 10 cp. If a sample of a mixture of these components has a measured viscosity of 5 cp, the mixing law of Equation 20 may be applied to determine that the mixture consists of 43% filtrate. Similarly, if the mixture viscosity measures 2 cp, the composition is 100% filtrate.
Another embodiment provides an improvement over the conventional reporting of data when one or both component properties are unknown. In current practice, the measured mixed sample properties would be reported relative to the time at which they were acquired or relative to the cumulative volume of sample which has been acquired. In this embodiment of the method, successive measurements of the mixed sample property may be presented in terms of normalized contamination. As an example, consider the case above in which fluid filtrate with a viscosity of 2 cp and formation fluid with a viscosity 10 cp are mixed, but neither of these values is known at the time of the measurement. The difference between the logarithms of two measures of the mixture viscosity is related to a fixed step change in the amount of filtrate in the sample. Further, any other pair of differences in logarithms which is equivalent to the first difference represents the same change in filtrate content as the first pair of measurements.
For example, if two measurements of the viscosity of the mixture yielded 2 cp and 3 cp. Then, applying the mixing law of Equation 21, the same reduction in filtrate contribution as provided an increase from 2 cp to 3 cp, would yield a mixture viscosity of 4.5 cp, and two times this reduction would yield a mixture of 6.75 cp.
As an aid to understanding the concept presented above the following table can be readily constructed by application of the mixing law of Equation 20 with the component viscosities stated above (fractional differences are the result of rounding errors).
Two or more mixing laws may also be used in combination to calculate formation fluid properties. For example, in a reservoir sampling application, the density of the formation fluid may be determined (for example by pressure gradient analysis) and the filtrate density may be determined from surface or subsurface measurements of the drilling fluid filtrate. These properties may be used, ether with an appropriate mixing law (e.g. Equation 14) to determine the amount of filtrate(s) present in the mixed sample (see, for example, Equation 16).
By taking measurements of another property for which a mixing law is known, the value of this property may be calculated for the formation fluid. For example, if the viscosity of the mixed sample is measured at the same time as the density, the filtrate viscosity (determined by surface or subsurface measurement of the filtrate) and the filtrate level (s) determined from an independent measurement (as the density measurement presented above) may be used in an appropriate mixing law (for example Equation 19) to calculate the viscosity of the formation fluid (R2).
Thus, by determining the relationship of a property of a mixture to the mixture's components, a real-time, quantitative, indication of the composition of the mixture can be determined
The embodiments set forth herein are merely illustrative and do not limit the scope of the invention or the details therein. For example, while one preferred measured property is density, it is understood that other fluid properties that have a measurable response to fluid contamination can be used. For example, other fluid properties that may find use in the present invention include resistivity, viscosity, chromatography, radioactivity, dielectric constant, optical density, and magnetic resonance, weight, acoustic impedance, and acoustic velocity. It will be appreciated that many other modifications and improvements to the disclosure herein may be made without departing from the scope of the invention or the inventive concepts herein disclosed. Because many varying and different embodiments may be made within the scope of the inventive concept herein taught, including equivalent structures or materials hereafter thought of, and because many modifications may be made in the embodiments herein detailed in accordance with the descriptive requirements of the law, it is to be understood that the details herein are to be interpreted as illustrative and not in a limiting sense.
The present application is a Continuation-in-Part application Ser. No. 10/166,195 filed Jun. 10, 2002 filed under 35 U.S.C. § 120, claiming priority to U.S. Pat. No. 6,748,328, issuing Jun. 8, 2004, and titled “Determining Fluid Composition from Fluid Properties,” and hereby incorporated herein by reference for all purposes.
Number | Name | Date | Kind |
---|---|---|---|
4994671 | Safinya et al. | Feb 1991 | A |
5266800 | Mullins | Nov 1993 | A |
5331156 | Hines et al. | Jul 1994 | A |
5518996 | Maroy et al. | May 1996 | A |
5519214 | Houwen et al. | May 1996 | A |
5644076 | Proett et al. | Jul 1997 | A |
5741962 | Birchak et al. | Apr 1998 | A |
6350986 | Mullins et al. | Feb 2002 | B1 |
6378364 | Pelletier et al. | Apr 2002 | B1 |
6539795 | Scherpenisse et al. | Apr 2003 | B1 |
6748328 | Storm, Jr. et al. | Jun 2004 | B2 |
Number | Date | Country |
---|---|---|
WO 0050876 | Aug 2000 | WO |
Number | Date | Country | |
---|---|---|---|
20040254732 A1 | Dec 2004 | US |
Number | Date | Country | |
---|---|---|---|
Parent | 10166195 | Jun 2002 | US |
Child | 10862594 | US |