Real-time fluid monitoring and classification using downhole spectrometer measurements

Abstract

The disclosed techniques relate to fluid monitoring and classification techniques that include a method that includes acquiring, via a spectral analysis module of a sampling system of a downhole well tool, spectral data for an unknown reservoir fluid received by the sampling system of the downhole well tool from a geological formation; determining, via the one or more processors, a fluid type of the unknown reservoir fluid in the sampling system of the downhole well tool based at least in part on projections of the spectral data for the unknown reservoir fluid onto at least the first two eigenvectors of spectral data of known fluids stored in a spectral database; and generating, via the one or more processors, a downhole well operation decision based on the determined fluid type of the unknown reservoir fluid in substantially real-time while the sampling system of the downhole well tool receives the unknown reservoir fluid.

Description

CROSS-REFERENCE TO RELATED APPLICATION

[0001] This application claims priority to and the benefit of U.S. Provisional Patent Application Ser. No. 63 / 583,970, entitled “REAL-TIME FLUID MONITORING AND CLASSIFICATION USING DOWNHOLE SPECTROMETER MEASUREMENTS,” filed Sep. 20, 2023, which is hereby incorporated by reference in its entirety for all purposes.BACKGROUND

[0002] The present disclosure relates generally to downhole tools. More specifically, the present disclosure relates to fluid monitoring and classification techniques to improve the accuracy of determining reservoir fluid types during downhole fluid sampling operations.

[0003] This section is intended to introduce the reader to various aspects of art that may be related to various aspects of the present disclosure, which are described and / or claimed below. This discussion is believed to be helpful in providing the reader with background information to facilitate a better understanding of the various aspects of the present disclosure. Accordingly, it should be understood that these statements are to be read in this light, and not as admissions of prior art.

[0004] The oil and gas industry includes a number of sub-industries, such as exploration, drilling, logging, extraction, transportation, refinement, retail, and so forth. During exploration and drilling, wellbores may be drilled into the ground for reasons that may include discovery, observation, and / or extraction of resources. These resources may include oil, gas, water, or any other combination of elements within the ground. Wellbores or boreholes may be drilled to, for example, locate and produce hydrocarbons. During a well development operation, it may be desirable to evaluate and / or measure properties of encountered formations, formation fluids and / or formation gasses.SUMMARY

[0005] A summary of certain embodiments disclosed herein is set forth below. It should be understood that these aspects are presented merely to provide the reader with a brief summary of these certain embodiments and that these aspects are not intended to limit the scope of this disclosure. Indeed, this disclosure may encompass a variety of aspects that may not be set forth below.

[0006] Certain embodiments of the present disclosure include a method that includes acquiring, via a spectral analysis module of a sampling system of a downhole well tool, spectral data for an unknown reservoir fluid received by the sampling system of the downhole well tool from a geological formation. The method also includes determining, via the one or more processors, a fluid type of the unknown reservoir fluid in the sampling system of the downhole well tool based at least in part on projections of the spectral data for the unknown reservoir fluid onto at least the first two eigenvectors of spectral data of known fluids stored in a spectral database. The method further includes generating, via the one or more processors, a downhole well operation decision based on the determined fluid type of the unknown reservoir fluid in substantially real-time while the sampling system of the downhole well tool receives the unknown reservoir fluid.

[0007] Other embodiments of the present disclosure also include a system having a spectral analysis module configured to acquire spectral data for an unknown reservoir fluid received by a sampling system of a downhole well tool from a geological formation. The system also includes a data processing system configured to determine a fluid type of the unknown reservoir fluid in the sampling system of the downhole well tool based at least in part on projections of the spectral data for the unknown reservoir fluid onto at least the first two eigenvectors of spectral data of known fluids stored in a spectral database; and to generate a downhole well operation decision based on the determined fluid type of the unknown reservoir fluid in substantially real-time while the sampling system of the downhole well tool receives the unknown reservoir fluid.

[0008] Other embodiments of the present disclosure also include a method that includes acquiring, via a spectral analysis module of a sampling system of a downhole well tool, spectral data for an unknown reservoir fluid received by the sampling system of the downhole well tool from a geological formation. The method also includes raining, via one or more processors, fluid monitoring and classification (FMC) algorithms based on reference fluid types of spectral measurements stored in the spectral database to determine eigenvectors of spectral data of known fluids stored in a spectral database. The method further includes determining, via the one or more processors, a fluid type of the unknown reservoir fluid in the sampling system of the downhole well tool based at least in part on projections of the spectral data for the unknown reservoir fluid onto at least the first two eigenvectors of the spectral data of the known fluids stored in the spectral database and based at least in part on the trained FMC algorithms. In addition, the method includes generating, via the one or more processors, a downhole well operation decision based on the determined fluid type of the unknown reservoir fluid in substantially real-time while the sampling system of the downhole well tool receives the unknown reservoir fluid.

[0009] Various refinements of the features noted above may exist in relation to various aspects of the present disclosure. Further features may also be incorporated in these various aspects as well. These refinements and additional features may exist individually or in any combination. For instance, various features discussed below in relation to one or more of the illustrated embodiments may be incorporated into any of the above-described aspects of the present disclosure alone or in any combination. The brief summary presented above is intended only to familiarize the reader with certain aspects and contexts of embodiments of the present disclosure without limitation to the claimed subject matter.BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Various aspects of this disclosure may be better understood upon reading the following detailed description and upon reference to the drawings in which:

[0011] FIG. 1A is schematic diagram of downhole drilling system include an optical spectrometer system, in accordance with aspects of the present disclosure;

[0012] FIG. 1B is a schematic diagram of downhole sampling equipment having various testing modules used to determine one or more characteristics of a subsurface formation fluid, in accordance with an embodiment of the present techniques;

[0013] FIG. 2 illustrates distribution of normalized eigenvalues of spectral data in a spectral database, in accordance with an embodiment of the present techniques;

[0014] FIG. 3 illustrates the first (top) and second (bottom) eigenvectors of the spectral data in the spectral database, in accordance with an embodiment of the present techniques;

[0015] FIG. 4 is a cross-plot of projections of spectral data in the spectral database on a first eigenvector versus projections of spectral data on a second eigenvector, in accordance with an embodiment of the present techniques;

[0016] FIG. 5 shows an example of field data where the break-through of formation fluid during a fluid sampling operation is clearly identified based on locations of projections with respect to the unit-circle, in accordance with an embodiment of the present techniques;

[0017] FIG. 6 is a plot of gas-oil ratio (GOR) of the fluids in the spectral database versus the angle of projections for the spectra in the spectral database, in accordance with an embodiment of the present techniques;

[0018] FIG. 7 is a cross-plot of projections of spectral data on a first eigenvector versus projections of spectral data on a second eigenvector for verification and validation (V&V) data sets, in accordance with an embodiment of the present techniques;

[0019] FIG. 8 illustrates a top plot of fluid monitoring and classification (FMC) processing results for sample line data of a first example well, and a bottom plot of FMC processing results for guard line data of the first example well, in accordance with an embodiment of the present techniques;

[0020] FIG. 9 illustrates a top plot of projections for sample line and guard line data with respect to the unit-circle for the first example well, and a bottom plot that illustrates the integration of sample line and guard line results for interpretation of the first example well, in accordance with an embodiment of the present techniques;

[0021] FIG. 10 illustrates a top plot of FMC processing results for sample line data of a second example well, and a bottom plot of FMC processing results for guard line data of the second example well, in accordance with an embodiment of the present techniques;

[0022] FIG. 11 illustrates a top plot of projections for sample line and guard line data with respect to the unit-circle for the second example well, and a bottom plot that illustrates the integration of sample line and guard line results for interpretation of the second example well, in accordance with an embodiment of the present techniques;

[0023] FIG. 12 illustrates recorded pressure and temperature in a sample line of the second example well, in accordance with an embodiment of the present techniques;

[0024] FIG. 13 illustrates a top plot of FMC processing results for sample line data of a third example well, and a bottom plot of FMC processing results for guard line data of the third example well, in accordance with an embodiment of the present techniques;

[0025] FIG. 14 illustrates a top plot of projections for sample line and guard line data with respect to the unit-circle for the third example well, and a bottom plot that illustrates the integration of sample line and guard line results for interpretation of the third example well, in accordance with an embodiment of the present techniques;

[0026] FIG. 15 illustrates a top plot of FMC processing results for sample line data of a fourth example well, and a bottom plot of FMC processing results for guard line data of the fourth example well, in accordance with an embodiment of the present techniques;

[0027] FIG. 16 illustrates a top plot of projections for sample line and guard line data with respect to the unit-circle for the fourth example well, and a bottom plot that illustrates the integration of sample line and guard line results for interpretation of the fourth example well, in accordance with an embodiment of the present techniques;

[0028] FIG. 17 illustrates a top plot of FMC processing results for sample line data of a fifth example well, and a bottom plot of FMC processing results for guard line data of the fifth example well, in accordance with an embodiment of the present techniques;

[0029] FIG. 18 illustrates a top plot of projections for sample line and guard line data with respect to the unit-circle for the fifth example well, and a bottom plot that illustrates the integration of sample line and guard line results for interpretation of the fifth example well, in accordance with an embodiment of the present techniques;

[0030] FIG. 19 illustrates a top plot of FMC processing results for sample line data of a sixth example well, and a bottom plot of FMC processing results for guard line data of the sixth example well, in accordance with an embodiment of the present techniques;

[0031] FIG. 20 illustrates a top plot of projections for sample line and guard line data with respect to the unit-circle for the sixth example well, and a bottom plot that illustrates the integration of sample line and guard line results for interpretation of the sixth example well, in accordance with an embodiment of the present techniques;

[0032] FIG. 21 is a plot of gas-oil ratio (GOR) of field fluid data versus angle estimates of corresponding spectral data determined using FMC techniques, in accordance with an embodiment of the present techniques; and

[0033] FIG. 22 is a flow diagram of a method of utilizing the FMC techniques described herein, in accordance with an embodiment of the present techniques.DETAILED DESCRIPTION

[0034] One or more specific embodiments will be described below. In an effort to provide a concise description of these embodiments, not all features of an actual implementation are described in the specification. It should be appreciated that in the development of any such actual implementation, as in any engineering or design project, numerous implementation-specific decisions must be made to achieve the developers' specific goals, such as compliance with system-related and enterprise-related constraints, which may vary from one implementation to another. Moreover, it should be appreciated that such a development effort might be complex and time consuming, but would nevertheless be a routine undertaking of design, fabrication, and manufacture for those of ordinary skill having the benefit of this disclosure.

[0035] When introducing elements of various embodiments of the present disclosure, the articles “a,”“an,” and “the” are intended to mean that there are one or more of the elements. The terms “comprising,”“including,” and “having” are intended to be inclusive and mean that there may be additional elements other than the listed elements. Additionally, it should be understood that references to “one embodiment” or “an embodiment” of the present disclosure are not intended to be interpreted as excluding the existence of additional embodiments that also incorporate the recited features.

[0036] As used herein, the terms “connect,”“connection,”“connected,”“in connection with,” and “connecting” are used to mean “in direct connection with” or “in connection with via one or more elements”; and the term “set” is used to mean “one element” or “more than one element.” Further, the terms “couple,”“coupling,”“coupled,”“coupled together,” and “coupled with” are used to mean “directly coupled together” or “coupled together via one or more elements.” As used herein, the terms “up” and “down,”“uphole” and “downhole”, “upper” and “lower,”“top” and “bottom,” and other like terms indicating relative positions to a given point or element are utilized to more clearly describe some elements. Commonly, these terms relate to a reference point as the surface from which drilling operations are initiated as being the top (e.g., uphole or upper) point and the total depth along the drilling axis being the lowest (e.g., downhole or lower) point, whether the well (e.g., wellbore, borehole) is vertical, horizontal or slanted relative to the surface.

[0037] As used herein, the term “computing system” refers to an electronic computing device such as, but not limited to, a single computer, virtual machine, virtual container, host, server, laptop, and / or mobile device, or to a plurality of electronic computing devices working together to perform the function described as being performed on or by the computing system. As used herein, the term “medium” refers to one or more non-transitory, computer-readable physical media that together store the contents described as being stored thereon. Embodiments may include non-volatile secondary storage, read-only memory (ROM), and / or random-access memory (RAM). As used herein, the term “application” refers to one or more computing modules, programs, processes, workloads, threads and / or a set of computing instructions executed by a computing system. Example embodiments of an application include software modules, software objects, software instances and / or other types of executable code.

[0038] In addition, as used herein, the terms “real time”, “real-time”, or “substantially real time” may be used interchangeably and are intended to describe operations (e.g., computing operations) that are performed without any human-perceivable interruption between operations. For example, as used herein, data relating to the systems described herein may be collected, transmitted, and / or used in control computations in “substantially real time” such that data readings, data transfers, and / or data processing steps occur once every second, once every 0.1 second, once every 0.01 second, or even more frequently, during operations of the systems (e.g., while the systems are operating). In addition, as used herein, the terms “continuous”, “continuously”, or “continually” are intended to describe operations that are performed without any significant interruption. For example, as used herein, control commands may be transmitted to certain equipment every five minutes, every minute, every 30 seconds, every 15 seconds, every 10 seconds, every 5 seconds, or even more often, such that operating parameters of the equipment may be adjusted without any significant interruption to the closed-loop control of the equipment.

[0039] In addition, as used herein, the terms “automatic”, “automated”, “autonomous”, and so forth, are intended to describe operations that are performed are caused to be performed, for example, by a computing system (i.e., solely by the computing system, without human intervention). Indeed, it will be appreciated that the data processing systems and control systems described herein may be configured to perform any and all of the data processing and control functions described herein automatically.

[0040] The embodiments described herein include robust and stable methods of processing spectral data acquired during fluid sampling jobs, which may be referred to as fluid monitoring and classification (FMC). Such methods perform data analysis in the eigenspace by projecting the spectral data onto the first and second eigenvectors, which are derived from a known spectral database containing various hydrocarbon fluids. The representation of spectral data in the eigenspace enables the visualization of important properties that lead to numerous applications such as:

[0041] 1. For hydrocarbon fluids, the representation of data projections would be on or close to the unit-circle, whereas for non-hydrocarbon fluids (e.g., water, mud and other contaminated fluid), their projections would be far away from the unit-circle. This property enables relative easy distinguishing between hydrocarbon fluids from non-hydrocarbon fluids. In the case of mixing water and hydrocarbon in the flow line, for example, this property enables the identification of a clean hydrocarbon spectra that can be used for estimating the water and oil fraction, and application of the composition ratio and gas-oil-ratio (GOR) calculations.

[0042] 2. For hydrocarbon fluids on the unit-circle, there is a natural separation of different types of hydrocarbon fluids, which leads to the fluid classification scheme described herein. This property also allows for identification of the oil-based mud (OBM) filtrate with specific projection angles.

[0043] 3. The evolution of data projections around the unit-circle shows a strong correlation with the GOR value of hydrocarbon fluids: the smaller the projection angles, the higher the GOR values. As a result, the evolution of data projections allows for continuous monitoring of the GOR trend and adjustment of the sampling operation accordingly.

[0044] 4. Integration of projection data from the sample line and the guard line enables determination of which flowline contains the cleaner fluid, and continuous monitoring of the efficiency of focused sampling.

[0045] With the foregoing in mind, FIG. 1A depicts an example of wellsite systems that may employ the techniques described herein. FIG. 1A depicts a rig 10 with a downhole tool 12 suspended therefrom and into a wellbore 14 of a reservoir via a drill string 16. The drill string 16 is rotated by a rotary table 24, energized by means not shown, which engages a kelly 26 at the upper end of the drill string 16. The drill string 16 is suspended from a hook 28, attached to a traveling block (also not shown), through the kelly 26 and a swivel 30 (e.g., rotary swivel) that permits rotation of the drill string 16 relative to the hook 28. The rig 10 is depicted as a land-based platform and derrick assembly used to form the wellbore 14 by rotary drilling.

[0046] While the depicted embodiment relates to a downhole tool 12 disposed in a wellbore 14, it should be understood that, at least in some instances, the disclosure techniques may be used in a logging-while drilling (LWD) tool. In such an embodiment, the formation fluid or drilling mud 32 (e.g., OBM or water-based mud (WBM)) may be stored in a pit 34 formed at the well site. A pump 36 delivers the reservoir fluid 52 to the interior of the drill string 16 via a port in the swivel 30, inducing the drilling mud 32 to flow downwardly through the drill string 16 as indicated by a directional arrow 38. The formation fluid exits the drill string 16 via ports of the downhole tool 12, and then circulates upwardly through the region between the outside of the drill string 16 and the wall of the wellbore 14, called the annulus, as indicated by directional arrows 40. The drilling mud 32 lubricates a drill bit and carries formation cuttings up to the surface as it is returned to the pit 34 for recirculation.

[0047] In certain embodiments, the downhole tool 12 includes a downhole analysis system. For example, the downhole tool 12 may include a sampling system 42 including a fluid communication module 46 and a sampling module 48. The modules may be housed in a drill collar for performing various formation evaluation functions, such as pressure testing and fluid sampling, among others. As shown in FIG. 1A, the fluid communication module 46 is positioned adjacent the sampling module 48; however, the position of the fluid communication module 46, as well as other modules, may vary in other embodiments. Additional devices, such as pumps, gauges, sensor, monitors or other devices usable in downhole sampling and / or testing also may be provided. The additional devices may be incorporated into the fluid communication module 46, the sample module 48, or disposed within separate modules included within the sampling system 42.

[0048] In some embodiments, the downhole tool 12 may be a formation testing downhole tool. For example, the downhole tool 12 may evaluate fluid properties of reservoir fluid 50, which is withdrawn into the downhole tool 12. Accordingly, the sampling system 42 may include sensors that may measure fluid properties such as GOR, mass density, optical density (OD), composition of C1, C2, C3, C4, C5, and C6+, formation volume factor, viscosity, resistivity, fluorescence, American Petroleum Institute (API) gravity, and combinations thereof of the reservoir fluid 50. In certain embodiments, the fluid communication module 46 includes a probe which may be positioned inside borehole. In addition, in certain embodiments, the probe includes one or more inlets for receiving the reservoir fluid 52 and one or more flowlines (not shown) extending into the downhole tool 12 for passing fluids (e.g., the reservoir fluid 50) through the tool. In certain embodiments, the probe may include a single inlet designed to direct the reservoir fluid 50 into a flowline within the downhole tool 12. Further, in other embodiments, the probe may include multiple inlets that may, for example, be used for focused sampling. In these embodiments, the probe may be connected to a sampling flowline, as well as to guard flowlines. In certain embodiments, the probe may be movable between extended and retracted positions for selectively engaging the wellbore wall 58 of the wellbore 14 and acquiring fluid samples from the geological formation 20. One or more setting accessories, standoffs, or rollers 64 may be provided to assist in positioning the fluid communication device against the wellbore wall 58.

[0049] In certain embodiments, the downhole tool 12 includes a spectral analysis module 68 configured to analyze optical spectral data to determine formation properties such as, for example, lithology, density, formation geometry, reservoir boundaries, among others. Sensors within the spectral analysis module 68 may detect optical spectral data and determine the geological characteristics of the geological formation 20 based at least in part on the optical spectral data, as described in greater detail herein.

[0050] FIG. 1B depicts an example of a wireline downhole tool 100 that may employ the systems and techniques described herein. The wireline downhole tool 100 is suspended in the wellbore 14 from the lower end of a multi-conductor cable 104 that is spooled on a winch at the surface 74. Similar to the downhole tool 12, the wireline downhole tool 100 may be conveyed on wired drill pipe, a combination of wired drill pipe and wireline, or other suitable types of conveyance. The cable 104 is communicatively coupled to the data processing system 76. The wireline downhole tool 100 includes an elongated body 108 that houses modules 110, 112, 114, 122, and 124 that provide various functionalities including imaging, fluid sampling, fluid testing, operational control, and communication, among others. For example, the modules 110, 112 may provide additional functionality such as fluid analysis, resistivity measurements, operational control, communications, coring, and / or imaging, among others. In particular, the modules 110, 112 may include a spectral analysis module similar to the spectral analysis module 68 of the downhole tool 12 of FIG. 1.

[0051] As shown in FIG. 1B, the module 114 is a fluid communication module 114 that has a selectively extendable probe 116 and backup pistons 118 that are arranged on opposite sides of the elongated body 108. The extendable probe 116 is configured to selectively seal off or isolate selected portions of the wellbore wall 58 of the wellbore 14 to fluidly couple to the adjacent geological formation 20 and / or to draw fluid samples from the geological formation 20. The probe 116 may include a single inlet or multiple inlets designed for guarded or focused sampling. In certain embodiments, the probe 116 may include a sample inlet fluidly coupled to a sample line and a guard inlet fluidly coupled to a guard line.

[0052] The reservoir fluid 50 may be expelled to the wellbore through a port in the body 108 or the reservoir fluid 50 may be sent to one or more modules 122, 124. The modules 122, 124 may include sample chambers that store the reservoir fluid 50. In the illustrated example, the data processing system 76 and / or a downhole control system are configured to control the extendable probe 116 and / or the drawing of a fluid sample from the geological formation 20 to enable analysis of the fluid properties of the reservoir fluid 50, as discussed above. In some embodiments, the wireline downhole tool 100 may include one or more light sources and / or light detectors disposed along a fluid conduit of the wireline downhole tool 100 to facilitate acquiring optical spectrometer data of the reservoir fluid 50.

[0053] In certain embodiments, the sensors within the downhole tool 12 may collect and transmit data 70 associated with the characteristics of the geological formation 20 and / or the fluid properties and the composition of the reservoir fluid 50 to a control and data acquisition system 72 at surface 74, where the data 70 may be stored and processed in a data processing system 76 of the control and data acquisition system 72.

[0054] The data processing system 76 may include a processor 78, memory 80, storage 82, and / or display 84. The memory 80 may include one or more tangible, non-transitory, machine readable media collectively storing one or more sets of instructions for operating the downhole tool 12, determining formation characteristics (e.g., geometry, connectivity, minimum horizontal stress, etc.) calculating and estimating fluid properties of the reservoir fluid 50, modeling the fluid behaviors using, for example, equation of state models (EOS). The memory 80 may store reservoir modeling systems (e.g., geological process models, petroleum systems models, reservoir dynamics models, etc.), mixing rules and models associated with compositional characteristics of the reservoir fluid 50, equation of state (EOS) models for equilibrium and dynamic fluid behaviors, and any other information that may be used to determine geological and fluid characteristics of the geological formation 20 and reservoir fluid 52, respectively. In certain embodiments, the data processing system 76 may apply filters to remove noise from the data 70.

[0055] To process the data 70, the processor 78 may execute instructions stored in the memory 80 and / or storage 82. For example, the instructions may cause the processor to compare the data 70 (e.g., from the logging while drilling and / or downhole analysis) with known reservoir properties estimated using the reservoir modeling systems, use the data 70 as inputs for the reservoir modeling systems, and identify geological and reservoir fluid parameters that may be used for exploration and production of the reservoir. As such, the memory 80 and / or storage 82 of the data processing system 76 may be any suitable article of manufacture that can store the instructions. By way of example, the memory 80 and / or the storage 82 may be ROM memory, random-access memory (RAM), flash memory, an optical storage medium, or a hard disk drive. The display 84 may be any suitable electronic display that can display information (e.g., logs, tables, cross-plots, reservoir maps, etc.) relating to properties of the well / reservoir as measured by the downhole tool 12. It should be appreciated that, although the data processing system 76 is shown by way of example as being located at the surface 74, the data processing system 76 may be located in the downhole tool 12. In such embodiments, some of the data 70 may be processed and stored downhole (e.g., within the wellbore 14), while some of the data 70 may be sent to the surface 74 (e.g., in real time). In certain embodiments, the data processing system 76 may use information obtained from petroleum system modeling operations, ad hoc assertions from the operator, empirical historical data (e.g., case study reservoir data) in combination with or lieu of the data 70 to determine certain parameters of the reservoir 15.Singular-Value Decomposition (SVD) of an Optical Spectral Database

[0056] During recent years, an optical spectral database has been acquired that consists of approximately 500 optical spectra of a wide variety of petroleum fluid samples measured at various temperatures (up to 347° F.), and pressures (up to 25,000 psi) with their corresponding compositions (e.g., weight fractions derived from gas chromatography analysis) and pressure-volume-temperature (PVT) properties. The fluid samples are roughly classified into oil, gas condensate, and gas. This database has been used to develop the mapping of matrices of a new fluid composition algorithm. As described herein, the database may be used to extract characteristic features of optical spectra for the classification of fluid type and various applications that involve monitoring fluid properties and contamination estimation during sampling operations.

[0057] First, the embodiments described herein perform singular-value decomposition (SVD) of the entire spectral database. The SVD of a matrix is a mathematical procedure that factorizes the matrix into the product of three matrices. Here, the matrix, denoted as A, is an N-by-M matrix, where N is the number of spectra in the database and M is the number of optical channels in each spectrum. In other words, the matrix A essentially contains the entire optical spectral database. Mathematically, the SVD of A may be written as:A=UDVT(1)where the columns of U and V are orthonormal eigenvectors and the matrix D is diagonal with positive real entries known as eigenvalues. The factorization in (1) is typically arranged such that the eigenvalues (i.e., di) in D are ordered from the largest to the smallest.

[0059] For this application, the part of the spectra in the hydrocarbon absorption region is of primary interest and, therefore, the wavelength channels from 1500 nm to 1800 nm are selected for the SVD spectral decomposition. In certain embodiments, this corresponds to channel 8, 9, 10, 11, 12, 13, 14, and 15 (i.e., 1500 nm, 1600 nm, 1650 nm, 1671 nm, 1690 nm, 1725 nm, 1760 nm, and 1800 nm). The OD spectra are first decolored and normalized by subtracting the OD values at 1600 nm in these channels.

[0060] SVD may be used to extract the dominant features hidden in the data. It is often possible to deduce a few significant factors to prescribe the underlined characteristics of data under consideration. For example, the eigenvector associated with the largest eigenvalue is the most dominant factor, while the eigenvector associated with the second largest eigenvalue, which is orthogonal to the first eigenvector, is the next dominant factor, and so on. In addition, the eigenvalues are used to quantify the relative significance of each eigenvector (or each factor) for characterizing the database. The normalized eigenvalue is defined as:di=λi∑ j=18⁢λj,i=1,2,... ,8(2)

[0061] The normalized eigenvalue distribution from the largest to the smallest is illustrated in FIG. 2. In particular, FIG. 2 shows that the first eigenvalue 126 contributes approximately 76% of the total, the second eigenvalue 128 contributes approximately 20% of the total, whereas the remaining eigenvalues contribute only a few percent of the total. In other words, using the first and second eigenvectors that correspond to the first and second eigenvalues 126, 128 enables the capture of more than 96% of content in the spectral database. In the next section, the techniques for extracting the spectral features using the first and second eigenvectors for classifying the fluid type and monitoring the fluid during sampling jobs are presented.

[0062] In addition, FIG. 3 shows the correspondent first (top) and second (bottom) eigenvectors 130, 132 of the spectral data in the spectral database that correspond to the first and second eigenvalues 126, 128. Note that these two eigenvectors 130, 132 are orthogonal (e.g., the inner product of two vectors is equal to zero) to each other and have the unit length (e.g., the L2-norm of each vector is equal to 1). The first eigenvector 130 shows the shape with the peak at 1725 nm, which appears to capture the primary response of hydrocarbon fluids. In comparison, the second eigenvector 132 shows a reverse polarity of gaseous channels (e.g., 1650 nm, 1671 nm, and 1690 nm) and the dead-oil channels (e.g., 1725 nm, 1760 nm, and 1800 nm). As a final note, the values of the first and second eigenvectors 130, 132 at 1500 nm are relatively small and yet important because they add some discriminating power against water (e.g., the OD value of water at 1500 nm is relatively large while that of hydrocarbon fluids (after de-coloring) at 1500 nm is relatively small or negligible).Projections of Spectra Data on the First and Second Eigenvectors

[0063] The first and second eigenvectors 130, 132 are referred to the 1st and 2nd column of V in Equation (1). They may also be referred to as the principal components or loading vectors of the data space. Prior to the projection of spectral data of unknown fluids, the following preprocessing may applied to the spectral data (xi, i=1, . . . . M):si=xi-xk∑ i=1M⁢(xi-xk)2(3)

[0064] where xi is the spectral data at the ith channel and xx is the spectral data at 1600 nm. In the present case, M=8, and i=1, . . . . M, corresponds the wavelength channels of 1500 nm, 1600 nm, 1650 nm, 1671 nm, 1690 nm, 1725 nm, 1760 nm, and 1800 nm. Denote s=[s1 s2 . . . sM] and then the projections (i.e., p1 and p2) of spectral data on the first and second eigenvectors 130, 132 are:p1=sv1T(4)p2=sv2T(5)where v1 and v2 are the 1st and 2nd eigenvectors, respectively. The computations in (4)-(5) are essentially the inner product of two vectors s and v1 (or v2).

[0066] FIG. 4 shows a cross-plot of projections of spectral data in the spectral database on the first eigenvector 130 versus projections of spectral data in the database on the second eigenvector 132. Each data point in this eigenspace corresponds to one spectrum in the database and is labeled with the fluid type that was pre-classified in the database. A natural separation of oil, gas condensate, and gas may be seen in this eigenspace. One point on the plot 134 was pre-classified as an oil in the database but it actually has a GOR of 3,457 scf / bbl. Therefore, it should be classified as gas condensate based on established fluid classification techniques. Within the gas / gas condensate group, there are some overlapping between the two. The overlapping gas samples may come from the category of wet gas, which has similar GOR values as lean gas condensate. However, most gas sample data are concentrated on the lower left-hand corner of FIG. 4—indicative of separation from other gas condensate samples.

[0067] The dashed line 136 illustrated in FIG. 4 is the unit-circle with origin at (0,0) and radius equal to 1. Since hydrocarbon fluids can very well be characterized by the first two eigenvectors 130, 132, their representation on this space should fall on or close to the unit-circle, as seen in FIG. 4. On the other hand, the representations of other non-hydrocarbon fluids like water or mud will not do so. FIG. 4, for example, shows the projections of water spectra (e.g., on the left side of FIG. 4), which are far away from the unit-circle.

[0068] FIG. 4 also shows the projections of J26 and synthetic oil-based mud (SOBM) of various temperatures and pressures. Their representations also fall on the unit-circle but at different locations. Note that each point on the circle can be uniquely represented by the polar coordinate angle θ, which can be utilized to identify fluids. For example, θ of SOBM is about 19°.θ=tan-1⁢p2p1

[0069] As a final note, the clockwise arrow in FIG. 4 also indicates the GOR trend, which may be used for the fluid classification as discussed in greater detail in the next section.

[0070] FIG. 5 shows an example of field data where the break-through of formation fluid during a fluid sampling operation is clearly identified based on the locations of projections with respect to the unit-circle. The subplot 138 on the top left-hand corner is a variable density log (VDL) of field spectral data whereas the subplot of radius of projections is shown on the lower left-hand corner. Relative locations of projections are shown on the subplot 140 on the right-hand side. At the beginning and the end of the fluid sampling, station fluid in the flowline is contaminated fluid or mud indicated by the radius of projections much less than one, shown as the cloud of blue dots 142 inside the unit-circle 144. Once the formation fluid has broken through at about 20 minutes, the data projections begin falling on the unit-circle 144 and with continued pumping, the data projections move downward, indicating fluid cleaning with increasing GOR, but remain on the unit-circle 144.Classification of Hydrocarbon Fluids

[0071] The database also contains the GOR of fluids measured in a PVT laboratory. However, the GOR measurements may only be available for fluids classified as oil and gas condensate. For those classified as gas, there may not be any GOR values reported, even though some of them may be wet gas with measurable GOR. FIG. 6 shows a plot of GOR of the fluids in the spectral database versus the corresponding angle of projections for the spectra in the spectral database. A strong correlation between the GOR and angle of the projections may be noted: for example, the larger the projection angles, the smaller the GOR values. Also shown in FIG. 6 are the different GOR zones 125 based on reservoir fluid classification in terms of GOR given in Table 1.TABLE 1Fluid classification based on GOR.GOR (in scf / bbl)Black Oil (125A)GOR < 1500 Volatile Oil (125B)1900 < GOR < 3200Gas Condensate (125C)3200 < GOR < 1500Wet Gas (125D)15000 < GOR < 100000Dry Gas (125E)GOR > 100000

[0072] Based on FIG. 6, the current fluid classification is accomplished by using the angle of projections as follows:TABLE 2Fluid classification based on angle of projections.Angle of projectionsBlack Oil (125A)3.9°< 0    Volatile Oil (125B)−1.5°< 0 < 3.9°Gas Condensate (125C)−20.3°< 0 <−1.5°Wet Gas (125D)−36.8°< 0 <−20.3°Dry Gas (125E)0 <−36.8°

[0073] It should be noted that refinement of angle range for classification may be continually improved. For example, further sub-division of the black oil into heavy and medium oil may be possible by including the coloration information of the spectrum.Validation Using FISO V&V Field Data

[0074] FISO verification and validation (V&V) data sets, which were recorded in 2015, are presented herein for validation of the embodiments described herein. It should be noted that the embodiments described herein may be applied to any spectrometer. In particular, the embodiments described herein may be applied to any optical transmission type of spectrometers. FIG. 7 shows the projections on the first and second eigenvectors 130, 132 for eight different hydrocarbon fluids: Oil21, Oil22, Oil23, Oil24, Cond6, Cond9, Wetgas, and Drygas. For each case, there are multiple date sets recorded for various temperatures and pressures. For example, the pressure ranges from 5 k to 20 k psi whereas the temperature ranges from 75 to 200° C. The dashed line 136 in FIG. 7 is the unit-circle and the band 146 is +2% around the unit-circle 136. The different color zones 125 indicate the different fluid classification regions based on Table 2. As expected, the projections of all fluids nearly fall on the unit-circle 136 but at different locations following the GOR trend. It should be noted that the PVT GOR of each fluid except Wetgas and Drygas is indicated with the fluid on the figure. The classification scheme also identities each fluid correctly (i.e., Oil22, Oil21, and Oil24 as black oil; Oil 23 as gas condensate; Cond6, Cond9, and Wetgas as wet gas; and Drygas as dry gas). The FISO field data is used in the below examples to demonstrate the FMC techniques described herein.

[0075] It will be appreciated that the FMC techniques described herein include machine learning (ML) techniques, artificial intelligence (AI) techniques, or a combination thereof, to increase the accuracy of the determined fluid types, as described in greater detail herein. Further, such FMC algorithms may be trained using reference spectral data, such as the field data described in greater detail below. In particular, the embodiments described herein train with a known database and apply the features / patterns (e.g., eigenvectors) to unknown fluid spectra for classification.Field Example 1: OBM, Black Oil

[0076] FIGS. 8 and 9 show the FMC processing results in a well drilled with OBM. The formation fluid is expected to be black oil. The top plot 148 of FIG. 8 shows the projections 150, 152 of data in the sample line (L1). Also shown on the subplot 154 above the projections is the VDL of sample line spectral data. At the beginning, the estimated radii 156 of the projections on the first and second eigenvectors 130, 132 are way below 1, which indicates the presence of non-hydrocarbon fluids (e.g., water) or contaminated fluids (e.g., mud) in the flowline. At about 75 minutes, the formation hydrocarbon fluid broke through and started filling the flowline, which is indicated by the radii 156 quickly raising to 1 (i.e., on the unit-circle). From 75 minute to the end, the estimated radii 156 of projections remain close to 1. Specifically, once the formation fluid broke through, the projection 150 on the first eigenvector 130 quickly rises from below 0.9 to about 0.98 and then stabilizes thereafter, while the projection 152 on the second eigenvector 132 shows a reverse trend (i.e., dropping from above 0.3 to about 0.16). Since the estimated radii 156 are close or equal to 1, the spectral data in this interval (75 min to the end) are on the unit-circle and move downward clockwise, meaning the GOR of hydrocarbon fluid in the flowline is getting higher and higher (i.e., the sampling fluid in the flowline is getting cleaner and cleaner with pumping.) The bottom plot 158 of FIG. 8 shows the same presentation of applying the FMC processing to the guard line (L2) data. Similar trends as the results of the sample line data are seen.

[0077] One unique feature is to integrate and combine the sample line and guard line results for interpretation. FIG. 9 shows such an integration. The top plot 160 of FIG. 9 shows the evolution of projections 162, 164 of the sample line and guard line data, respectively, on the unit-circle 136. Initially, both sample line and the guard line data are slightly off the unit-circle 136 and with more pumping, both projections 162, 164 move downward clockwise and merge with the unit-circle 136. Note that the color zones 125 of the band 146 on the plot 160 represent different regions for fluid classification (based on Table 2), which indicate both flowlines are filled with black oil.

[0078] A different way to present the integration is shown in the bottom plot 166 of FIG. 9. Initially, as illustrated in subplot 168, during the commingle flow (about 70-95 minutes), the estimated angles 170, 172 of the sample line and guard line, respectively, are similar. As soon as split flow (at about 95 minutes) is established, the estimated angles 170, 172 for the two flowlines diverge and show a relatively large difference. These differences subsequently diminish with continued pumping with split flow and focused flow. This pattern is indicative of an efficient focused-sampling operation. At the end of the sampling operation, at about 300 minutes, the estimated angles 170 of the sample line L1 remain smaller than the estimated angles 172 of the guard line L2. This means that the fluid in the sample line L1 has a higher GOR value than the fluid in the guard line and, therefore, likely contains cleaner fluid at low contamination. After about 310 minutes, the commingle valve was opened and the fluid in the sample and guard line are mixed, which is clearly shown by the matching estimated angles 170, 172 from both flowlines. The bottom two subplots 174, 176 show the fluid classification in both flowlines and throughout the entire sampling operation, both flowlines contain black oil. Label scheme: BO (=1) stands for black oil, VO (=2) stands for volatile oil, GC (=3) stands for gas condensate, WG (=4) stands for wet gas, and DG (=5) stands for dry gas.Example 2: (OBM, Volatile Oil)

[0079] This example is from a well drilled with OBM and the formation fluid is expected to be volatile oil. FIGS. 10 and 11 show the results of FMC processing for the sample line (L1) and the guard line (L2), which are displayed in the same manner as shown in Example 1 (e.g., FIGS. 8 and 9).

[0080] From the radius-plot of the projections 150, 152 in FIG. 10, the formation fluid broke through at about 120 minutes. With the estimated radii 156 quickly rising to 1 and remaining at approximately 1 (i.e., on the unit-circle), hydrocarbon fluid is filled in both flowlines throughout the entire job (up to about 810 minutes).

[0081] The evolution of data projections (e.g., the top plot 160 of FIG. 11) show that the data are initially farther away from the unit-circle 136 and gradually move clockwise close to the unit-circle 136. Eventually, they merge with the unit-circle 136 and continue with downward movement (i.e., increasing GOR trend). At the end of the station, the data projections fall into the volatile oil zone 125B for both flowlines L1, L2. The estimated angles 170, 172 versus elapsed time plot (bottom of FIG. 11) also indicate volatile oil in the sample and guard line L1, L2. Furthermore, the estimated angles 170 of projections in the sample line L1 are slightly lower than the estimated angles 172 of projections in the guard line L2—an indication of a cleaner volatile oil (i.e., higher GOR value) in the sample line L1.

[0082] It should be noted that the estimated angles 170, 172 of both flowlines L1, L2 show a small drift upward with increasing time, which is mostly caused by increasing temperature with pumping and station time. FIG. 12 shows the recorded pressure and temperature in the sample line (L1). The temperature increases like a ramp from 300 to 800 minutes, which coincides with the drift upward in the estimated angles 170, 172 illustrated in FIG. 11.Example 3: (OBM, Gas Condensate)

[0083] This example is from the same well as Example 2 but at a different station and depth. The formation fluid is expected to be gas condensate. FIGS. 13 and 14 show the results of FMC processing for the sample line (L1) and the guard line (L2), which are displayed in the same manner as the previous examples.

[0084] Except for the three intervals that show the flowlines L1, L2 filled with contaminated fluid like mud, the spectral data in the sample line L1 and the guard line L2 show relatively clean data (e.g., the top plot 148 of FIG. 13). The radii 156 of the projections for both flowlines L1, L2 are relatively close to 1 in the clean-data zones, indicating that the fluid in both flowlines L1, L2 is hydrocarbon.

[0085] The evolution of data projections (e.g., the top plot 160 of FIG. 14) with respect to the unit-circle 136 shows a similar pattern as Example 2-data initially are farther away from the unit-circle 136 and eventually merge into the unit-circle 136. While continuously moving downward clockwise, the data projections fall into the gas condensate zone 125C at the end. The bottom plot 166 of FIG. 14 shows the estimated angles and detected fluid types as a function of elapsed time. Over the entire sampling interval, the detected fluid type is gas condensate. A similar drift upward trend in estimated angles 170, 172 for both flowlines L1, L2 as Example 2 is observed. Once again, this is caused by the ramping up of fluid temperature and, in this case, the ramp-up is relatively large: from about 90° C. to about 120° C. according to the recorded temperature in the tool.Example 4: (OBM, Gas)

[0086] This example is from a gas sampling job with OBM as drilling fluid. FIGS. 15 and 16 show the results of FMC processing for the sample line (L1) and the guard line (L2), which are displayed in the same manner as the previous examples.

[0087] The top plot 148 of FIG. 15 shows the results of projections 150, 152 of sample line data. It should be noted that, over the entire operation interval, the data projections 150, 152 on the first and second eigenvectors 130, 132 display relatively large variations. The projections 152 on the second eigenvector 132, for example, drop from positive to negative values and continue a downward trend until the end, corresponding to the movement of data along the unit-circle 136 in the lower right-hand quadrant. Even though the projections 150, 152 show relatively large variations, the estimated radii 156 are close and nearly 1 over most of the interval. This indicates that the fluid in L1 is hydrocarbon but exhibits the characteristics of a different fluid type during pumping and cleanup. The results of projections of guard line data (e.g., the bottom plot 158 of FIG. 15) show the similar trend as the results of sample line data.

[0088] The evolution of data projections (e.g., the top plot 160 of FIG. 16) with respect to the unit-circle 136 shows that the data in both flowlines L1, L2 start in the black oil zone 125A, then cross the volatile oil and gas condensate zones 125C, and eventually end up in the wet gas zone 125D. This is more elucidated by examining the integration of sample line and guard line results plot 166 of FIG. 16. Based on the estimated angles 170, 172, the identified black oil interval (up to 60 minute) is the OBM filtrate with the estimated angles 170, 172 at approximately 19°. With more gas coming into the flowlines L1, L2 and mixed with OBM filtrate, the fluid type transitioned to volatile oil 125B, gas condensate 125C, and finally wet gas 125D. At the end, both flowlines L1, L2 were filled with nearly identical wet gas indicated by the same angles of projections.Example 5: (WBM, Black Oil)

[0089] This example is from a well drilled with WBM and the formation fluid is black oil. FIGS. 17 and 18 show the results of FMC processing for the sample line (L1) and the guard line (L2), which are displayed in the same manner as the previous examples.

[0090] Since the well was drilled with WBM, the presence of water in the flowlines L1, L2 is inevitable and this produces a strong interference to the hydrocarbon absorption port of spectra. This is clear from the VDL of sample line (L1) and guard line (L2) spectral data illustrated in FIG. 17. Therefore, the projections 150, 152 of data on the first and second eigenvectors 130, 132 are relatively noisy and not useable until approximately 350 minutes, where the estimated radii 156 of both flowlines L1, L2 rise to above 0.95.

[0091] The evolution of data projections (e.g., the top plot 160 of FIG. 18) with respect to the unit-circle 136 shows that the data in both flowlines L1, L2 are still far away from the unit-circle 136, even though they exceed the threshold of 0.95. However, for both flowlines L1, L2, the data projections show a tendency to converge to the unit-circle 136. With more pumping and diminishing water in the flowlines L1, L2, the data projections merge with the unit-circle 136 like other cases. The integration of sample line and guard line results plot 166 of FIG. 18 shows rapidly decreasing angles of projections corresponding to increasing GOR in both flow lines L1, L2. Based on the data in this section, the fluid in both flowlines L1, L2 is identified as black oil 125A.Example 6: (WBM, Gas Condensate)

[0092] The last example is from a well drilled with WBM and the reservoir fluid is gas condensate. Again, FIGS. 19 and 20 show the results of FMC processing for the sample line (L1) and the guard line (L2), which are displayed in the same manner as the previous examples.

[0093] The presence of water is ubiquitous from the VDL of sample line and guard line spectral data in FIG. 19. For the first 120 minutes, the estimated radii 156 for the sample line L1 and the guard line L2 are way below 1-indicative of no hydrocarbon fluid in the flowlines L1, L2. Soon after, the data show a segregated or slug flow pattern with alternating water and hydrocarbon fluid passing the sensor windows. This flow response is clearly indicated by the data projections 150, 152 and estimated radii 156. Note that whenever hydrocarbon fluid flows passing by the FISO window, the estimated radii 156 jump to the value close to 1 and subsequently drop way below 1 when water pass by. For the rest of the interval, the estimated radii 156, whose values are close to 1, identify different zones with hydrocarbon fluid.

[0094] The evolution of data projections (e.g., the top plot 160 of FIG. 20) with respect to the unit-circle 136 shows that the data in both flowlines L1, L2 start in the black oil zone 125A, then cross the volatile oil zone 125B, and eventually end up in the gas condensate zone 125C. The evolution of fluid can be more elucidated by examining the integration of sample line and guard line results plot 166 of FIG. 20. In the slug flow interval (e.g., 120-190 minute), the fluid type is alternatively switch between volatile oil and gas condensate. At the end, both flowlines L1, L2 are filled with gas condensate 125C.

[0095] As a final note, only the results with the estimated radii 156 exceeding the threshold of 0.99 are plotted in the integration plot. Setting the threshold on the radii 156 acts like a filtering step which eliminates noisy and non-hydrocarbon data in the presentation.

[0096] Table 3 shows a summary of FMC processing results from FISO field testing for several example wells. The second column is the angle estimates determined by FMC at the end of each job, whereas the third column is the fluid classifications determined by FMC. The last column is the PVT GOR. The PVT reports for the jobs shaded in red in Table 3 are not available. Table 3 is sorted by the angle estimates (i.e., the second column) in descending order. By comparing the second column to the last column, a correlation trend between the angle estimates and GOR values can clearly be seen: In particular, the larger the angle estimates, the smaller the GOR values. FIG. 21 shows a plot of the GOR of field fluid data versus the angle estimates of corresponding spectral data from Table 3, which clearly elucidates the strong correlation between the two.TABLE 3Summary of FMC on FISO field dataWellAngle (End)FMC ClassificationPVT GOR (scf / bbl)113.4Black oilNA212.3Black oilNA311.7Black oilNA49.5Black oil20359.1Black oil31269.0Black oil32178.9Black oil46285.4Black oil109694.9Black oil1202104.8Black oil1220112.7Volatile oil210512−6.3Gas Condensate398613−25.3Wet gas1875814−26.6Wet gasNA15−30.1Wet gasNA16−30.5Wet gasNA17−33.3Wet gasNA

[0097] Contamination—From the examples described above, it is frequently observed that the initial data projections 150, 152 are far away from the unit-circle 136 and move close to the unit-circle 136 with more pumping. Eventually, the data projections 150, 152 merge onto the unit-circle 136. This pattern of data evolution is related to sample fluid clean-up and, therefore, is worth exploring for the contamination estimation problem.

[0098] GOR Models—As described above, there is a strong correlation between PVT GOR and the projections 150, 152, as shown in FIGS. 6 and 21. Specifically, to the first order, the logarithm of GOR values can be related to the projections 150, 152 on the first and second eigenvectors 130, 132 and, therefore, a GOR model approximating the following relationship may be used:log⁡(GOR)∼f⁡(p1,p2)=a1+a2⁢p1+a3⁢p2+a4⁢p1⁢p2(6)where p1 and p2 are the projections 150, 152 on the first and second eigenvectors 130, 132, and a1, a2, a3 and a4 are unknown model constants. Note that Equation 6 also includes a cross-term p1p2 in the model. Alternatively, the following model using the projection ratio and the quadratic term may be used:log⁡(GOR)∼g⁡(p1,p2)=a1+a2(p2 / p1)+a3(p2 / p1)2(7)In certain embodiments, other approaches relating the logarithm of GOR to the angles of projection and including higher order terms may yield a model with better performance.

[0101] FIG. 22 is a flow diagram of a method 178 of utilizing the FMC techniques described herein. In certain embodiments, the method 178 includes acquiring, via a spectral analysis module 68 of a sampling system 42 of a downhole well tool 12, spectral data for an unknown reservoir fluid 52 received by the sampling system 42 of the downhole well tool 12 from a geological formation 20 (step 180). In addition, in certain embodiments, the method 178 includes determining, via the one or more processors 78, a fluid type 125 of the unknown reservoir fluid 52 in the sampling system 42 of the downhole well tool 12 based at least in part on projections of the spectral data for the unknown reservoir fluid 52 onto at least the first two eigenvectors 130, 132 of the spectral data of the known fluids stored in the spectral database (step 182). One relatively important embodiment described in greater detail herein is to take the acquired unknown spectral data in step 180 and projecting acquired unknown spectral data onto at least (if not only, in certain embodiments) the first and second eigenvectors 130, 132 in order to determine the fluid type.

[0102] As described in greater detail herein, in certain embodiments, the method 178 may also include training, via the one or more processors 78, fluid monitoring and classification (FMC) algorithms (e.g., training ML, AI, or both) based at least in part on reference spectral measurements stored in the spectral database (step 184); and determining, via the one or more processors 78, the fluid type of the unknown reservoir fluid 52 based at least in part on the trained FMC algorithms. In certain embodiments, the training of the FMC algorithms may include determining the eigenvectors 130, 132 of the spectral data of the known fluids stored in the spectral database. In addition, in certain embodiments, determining the fluid type 125 of the unknown reservoir fluid 52 includes categorizing the unknown reservoir fluid 52 as one of: black oil 125A, volatile oil 125B, gas condensate 125C, wet gas 125D, and dry gas 125E.

[0103] In addition, in certain embodiments, the method 178 includes generating, via the one or more processors 78, a downhole well operation decision based on the determined fluid type 125 in substantially real-time while the sampling system 42 of the downhole well tool 12 receives the unknown reservoir fluid 52 (step 186). For example, a primary objective of sampling in a job is to promptly identify the clean and representative reservoir fluid 52 present in the flowline. To achieve this objective, the real-time FMC algorithms described herein play a key role by empowering operators to, for example, manipulate valves of sampling system 42 and collect the reservoir fluid 52.

[0104] In addition, in certain embodiments, the method 178 includes continuously monitoring, via the spectral analysis module 68 of the downhole well tool 12, changes in the spectral data for a subsequent reservoir fluid 52 during a downhole well operation; determining, via the one or more processors 78, a fluid type 125 of the subsequent reservoir fluid 52 based at least in part on the subsequent projections of the spectral data; and adjusting, via the one or more processors 78, the downhole well operation decision based on the determined fluid type 125 of the subsequent reservoir fluid 52. In addition, in certain embodiments, the method 178 includes automatically adjusting, via the one or more processors 78, a downhole well operation in accordance with the downhole well operation decision.

[0105] In certain embodiments, the method 178 may include determining, via the one or more processors 78, the fluid type 125 of the unknown reservoir fluid 52 in the sampling system 42 of the downhole well tool 12 based at least in part on projections of the spectral data for the unknown reservoir fluid 52 onto only the first two eigenvectors 130, 132 of the spectral data of the known fluids stored in the spectral database. In other words, in certain embodiments, the third and higher order eigenvectors 130, 132 of the spectral data may be ignored entirely, thereby simplifying and accelerating the FMC processing results.

[0106] In addition, in certain embodiments, determining the fluid type of the unknown reservoir fluid 52 may include utilizing cross-plots of projections of the first and second eigenvectors 130, 132 of the spectral data relative to their position on a unit-circle 136 of 1. In addition, in certain embodiments, the method 178 may include determining, via the one or more processors 78, a gas-oil-ratio (GOR) of the unknown reservoir fluid 52 based on angles formed by the projections of the first and second eigenvectors 130, 132 of the spectral data; and determining, via the one or more processors 78, the fluid type of the unknown reservoir fluid 52 based at least in part on the determined GOR.

[0107] As such, the embodiments described herein utilize data relating to a wide array of hydrocarbon fluids from various locations worldwide, totaling several hundred samples. These fluids were carefully analyzed at a laboratory to assess their properties, such as fluid type, GOR, density, viscosity, and other relevant characteristics. Simultaneously, spectrometers of downhole well tools 12 were utilized to measure the spectral data for each fluid.

[0108] This collection of fluid samples, along with their associated spectral data and properties, formed the known spectral database described herein. The next step involved determining the eigenvectors from the spectral data stored in the spectral database. This process included utilizing a machine-learning procedure aimed at establishing a model represented by the eigenvectors. Similar to other AI / ML methodologies, once the eigenvectors were derived, they remain fixed, and may be stored in memory of downhole well tools 12 and used for the analysis of spectral data of unknown fluids acquired downhole by spectrometers of the downhole well tools 12. In general, these eigenvectors are not subject to any changes or adjustments during downhole operations.

[0109] The embodiments described herein include various key features including, but not limited to:

[0110] 1. Distinguishing between Hydrocarbon and Nonhydrocarbon Fluids: This involves the identification of nonhydrocarbon fluids such as water, mud, and other contaminants from the hydrocarbon fluid.

[0111] 2. Detecting OBM Filtrate and Formation Fluid Breakthrough: During the sampling process, it is important to identify the presence of OBM filtrate and the breakthrough of formation fluids.

[0112] 3. Monitoring a Fluid Cleanup Process based on GOR Trends: Throughout the sampling procedure, the fluid cleanup process is monitored by observing the GOR trend, which helps in determining the appropriate timing to take a clean fluid sample.

[0113] 4. Evaluating Dual Flow-line Efficiency for Focused Sampling: The efficiency of the dual flow-line system, which includes the sample line and guard line, is continuously monitored and adjusted to ensure effective fluid sampling.

[0114] These four aspects form the foundation of a successful fluid sampling job, and can be efficiently executed in real-time using the FMC algorithms described in greater detail herein.

[0115] The techniques presented and claimed herein are referenced and applied to material objects and concrete examples of a practical nature that demonstrably improve the present technical field and, as such, are not abstract, intangible or purely theoretical. Further, if any claims appended to the end of this specification contain one or more elements designated as “means for [perform]ing [a function] . . . ” or “step for [perform]ing [a function] . . . ”, it is intended that such elements are to be interpreted under 35 U.S.C. 112(f). However, for any claims containing elements designated in any other manner, it is intended that such elements are not to be interpreted under 35 U.S.C. 112(f).

Claims

1. A method, comprising:acquiring, via a spectral analysis module of a sampling system of a downhole well tool, spectral data for an unknown reservoir fluid received by the sampling system of the downhole well tool from a geological formation;determining, via the one or more processors, a fluid type of the unknown reservoir fluid in the sampling system of the downhole well tool based at least in part on projections of the spectral data for the unknown reservoir fluid onto at least the first two eigenvectors of spectral data of known fluids stored in a spectral database; andgenerating, via the one or more processors, a downhole well operation decision based on the determined fluid type of the unknown reservoir fluid in substantially real-time while the sampling system of the downhole well tool receives the unknown reservoir fluid.

2. The method of claim 1, comprising:training, via the one or more processors, fluid monitoring and classification (FMC) algorithms based on reference fluid types of spectral measurements stored in the spectral database; anddetermining, via the one or more processors, the fluid type of the unknown reservoir fluid based at least in part on the trained FMC algorithms.

3. The method of claim 2, wherein training, via the one or more processors, the FMC algorithms comprises determining the eigenvectors of the spectral data of the known fluids stored in the spectral database.

4. The method of claim 1, comprising:continuously monitoring, via the spectral analysis module of the downhole well tool, changes in the spectral data for a subsequent reservoir fluid during a downhole well operation;determining, via the one or more processors, a fluid type of the subsequent reservoir fluid based at least in part on the subsequent projections of the spectral data; andadjusting, via the one or more processors, the downhole well operation decision based on the determined fluid type of the subsequent reservoir fluid.

5. The method of claim 1, comprising automatically adjusting, via the one or more processors, a downhole well operation in accordance with the downhole well operation decision.

6. The method of claim 1, comprising determining, via the one or more processors, the fluid type of the unknown reservoir fluid in the sampling system of the downhole well tool based at least in part on projections of the spectral data for the unknown reservoir fluid onto only the first two eigenvectors of the spectral data of the known fluids stored in the spectral database.

7. The method of claim 6, wherein determining the fluid type of the unknown reservoir fluid comprises utilizing cross-plots of the projections of the first and second eigenvectors of the spectral data relative to their positions on a unit-circle of 1.

8. The method of claim 7, comprising:determining, via the one or more processors, a gas-oil-ratio (GOR) of the unknown reservoir fluid based on angles formed by the projections of the first and second eigenvectors of the spectral data; anddetermining, via the one or more processors, the fluid type of the unknown reservoir fluid based at least in part on the determined GOR.

9. The method of claim 1, wherein determining the fluid type of the unknown reservoir fluid comprises categorizing the unknown reservoir fluid as one of: black oil, volatile oil, gas condensate, wet gas, and dry gas.

10. A system, comprising:a spectral analysis module configured to acquire spectral data for an unknown reservoir fluid received by a sampling system of a downhole well tool from a geological formation; anda data processing system configured to:determine a fluid type of the unknown reservoir fluid in the sampling system of the downhole well tool based at least in part on projections of the spectral data for the unknown reservoir fluid onto at least the first two eigenvectors of spectral data of known fluids stored in a spectral database; andgenerate a downhole well operation decision based on the determined fluid type of the unknown reservoir fluid in substantially real-time while the sampling system of the downhole well tool receives the unknown reservoir fluid.

11. The system of claim 10, wherein the data processing system is configured to:train fluid monitoring and classification (FMC) algorithms based on reference fluid types of spectral measurements stored in the spectral database; anddetermine the fluid type of the unknown reservoir fluid based at least in part on the trained FMC algorithms.

12. The system of claim 11, wherein training the FMC algorithms comprises determining the eigenvectors of the spectral data of the known fluids stored in the spectral database.

13. The system of claim 10, wherein the spectral analysis module is configured to continuously monitor changes in the spectral data for a subsequent reservoir fluid during a downhole well operation, and wherein the data processing system is configured to:determine a fluid type of the subsequent reservoir fluid based at least in part on the subsequent projections of the spectral data; andadjust the downhole well operation decision based on the determined fluid type of the subsequent reservoir fluid.

14. The system of claim 10, wherein the data processing system is configured to automatically adjust a downhole well operation in accordance with the downhole well operation decision.

15. The system of claim 10, wherein the data processing system is configured to determine the fluid type of the unknown reservoir fluid in the sampling system of the downhole well tool based at least in part on projections of the spectral data for the unknown reservoir fluid onto only the first two eigenvectors of the spectral data of the known fluids stored in the spectral database.

16. The system of claim 15, wherein determining the fluid type of the unknown reservoir fluid comprises utilizing cross-plots of the projections of the first and second eigenvectors of the spectral data relative to their positions on a unit-circle of 1.

17. The system of claim 16, wherein the data processing system is configured to:determine a gas-oil-ratio (GOR) of the unknown reservoir fluid based on angles formed by the projections of the first and second eigenvectors of the spectral data; anddetermine the fluid type of the unknown reservoir fluid based at least in part on the determined GOR.

18. The system of claim 10, wherein determining the fluid type of the unknown reservoir fluid comprises categorizing the unknown reservoir fluid as one of: black oil, volatile oil, gas condensate, wet gas, and dry gas.

19. A method, comprising:acquiring, via a spectral analysis module of a sampling system of a downhole well tool, spectral data for an unknown reservoir fluid received by the sampling system of the downhole well tool from a geological formation;training, via one or more processors, fluid monitoring and classification (FMC) algorithms based on reference fluid types of spectral measurements stored in the spectral database to determine eigenvectors of spectral data of known fluids stored in a spectral database;determining, via the one or more processors, a fluid type of the unknown reservoir fluid in the sampling system of the downhole well tool based at least in part on projections of the spectral data for the unknown reservoir fluid onto at least the first two eigenvectors of the spectral data of the known fluids stored in the spectral database and based at least in part on the trained FMC algorithms; andgenerating, via the one or more processors, a downhole well operation decision based on the determined fluid type of the unknown reservoir fluid in substantially real-time while the sampling system of the downhole well tool receives the unknown reservoir fluid.

20. The method of claim 19, comprising automatically adjusting, via the one or more processors, a downhole well operation in accordance with the downhole well operation decision.

Images