Lithofacies Classification Based On Spectral Yields and Borehole Microresistivity Images

Authors: Kumar Anish; Kear, George R., Schlumberger Oilfield Services, 1515 Poydras Street, Suite 2700, New Orleans, LA 70112 USA. Abstract Lithofacies classification for borehole geology has, in general, been based on many different log responses and characteristics. Borehole image data are a proven, valuable resource for geological evaluation. Another useful device is the borehole spectroscopy tool, which generates the broad elemental composition of the rock and its quantitative geochemical lithology. The two outputs complement each other well for detailed geological interpretation. This paper introduces a novel lithofacies classification and processing scheme for use with borehole geological studies. The classification system uses the dry weight mineralogical output from wireline elemental capture spectroscopy data, and utilizes a ternary diagram approach with a corresponding set of rules to create dry-weight mineralogy-based lithofacies. This output is further refined to a more detailed level of formation representation using calibrated microresistivity image data and a second set of rules. The final results can then be presented as an output log as a detailed geological column displaying the calculated lithofacies. Numerical data from the final lithofacies results can then be used for additional computation as well as input for modeling. This new approach to lithofacies classification can be extremely valuable for stratigraphic interpretation such as depositional environment, detailed well-to-well geological correlation, and static reservoir modeling. The information may also be used in the planning, execution, and evaluation of borehole pressure and fluid sampling programs. Additional applications exist for wellbore stimulation procedures, completion and production strategies, and project economics. In addition to a summary of the tools and methods employed, selected case studies will be illustrated. Introduction and Background The challenge to geologists worldwide still exists to accurately describe subsurface geology, and is even greater in wells drilled with oil- and synthetics-based drilling fluids. Technological advances in the industry continually improve the interpretation of reservoirs and their lithologic characteristics. Among these advances are devices used to generate high-resolution images of the borehole (Cheung et al., 2001) and spectroscopy devices that provide the broad elemental composition of the rock and its quantitative geochemical lithology (Herron and Herron, 1997). The two sets of information complement each other well for any interpretation. Studies based solely on borehole images or only on borehole spectroscopy data are not able to provide comprehensive subsurface geological description and interpretation solutions. The paper summarizes a solution that addresses these limitations and adds value to the various exploration and production disciplines including geology, geophysics, petrophysics, reservoir engineering, and production engineering. Classification

Borehole micro-resistivity images are often used for geological interpretation. The high resolution of these images is critical for interpretation, however, an aspect lacking in these images is mineralogical information. A borehole spectroscopy tool generates a continuous output of mineralogical data. This tool measures the elemental yields of the principle elements in rocks (Grau et al., 1989). The measured values are used to compute dry weight percentages of total clay (WCLA), total quartz-feldspar-mica or QFM (WSAN), total carbonate (WCAR), total evaporates (WEVA), total pyrite, total coal, and total anhydrite. The biggest strength of the borehole spectroscopy data is the determination of total clay (WCLA), total QFM (WSAN), and total carbonate (WCAR). A major strength of the borehole microresistivity images is the high vertical resolution. The spectroscopy data, with a vertical resolution of 1.0-1.5 ft (30-46 cm), lacks this level of detail. The borehole images lack mineralogical information. Combining the strengths of these two types of data creates a very powerful interpretative solution that provides mineralogical data at a high resolution. The first step in the development of a system to integrate the two types of data and create a lithofacies column was to classify the mineralogical data from the spectroscopy device. Existing classification schemes were not based on mineralogical parameters and thus could not be used. Simple cut-offs to create some classes of rock types based on WCLA, WSAN, and WCAR did not account for all possible relative percentages of the three constituents. Classifications based on three components are most effectively done using a ternary diagram (see Hurlbut et al., 1977, for descriptions of ternary diagrams). Therefore, a ternary diagram was used to create a 100 % inclusive system of classification of rocks based on the selected spectroscopy outputs (see Fig. 1). Using the three end points of the ternary diagram, twelve lithofacies were defined (see Figs. 1 and 2) namely: Marl, Claystone, Shale, Sandy Shale, Shaly Sand, Sand, Clean Sand, Calcareous Sand, Calcareous Shale, Carbonate, Sandy Carbonate, and Shaly Carbonate. A salt lithofacies is computed using WEVA. As may be evident, this classification scheme is designed only for silici-clastic rocks. The mineralogical classification of the spectroscopy data, as mentioned above, lacked vertical resolution, hence a high vertical resolution input, inherent in the micro-resistivity images, was integrated with the outputs of the ternary diagram classification. Another rule base was created to accomplish the integration. Using a calibrated micro-resistivity value, two classes of rocks were assigned: shale and sand. Any bed too thin to be detected by the spectroscopy tool (a value of 12 inches (30.5 cm) was designated) would be described as shale or sand, based on selected cut-off values. Further, such beds would only be considered in zones not in that domain (as described by the above ternary diagram). iCore Lithofacies With the structure of the procedure in place, an algorithm was designed to apply the physics and mathematics of the system and perform the necessary processing and computations. The algorithm examines the data at each depth and assigns a specific lithofacies for that depth based on the input values. In this way, a lithofacies log is instantly generated for the entire interval where both sets of data are available. These lithofacies are called iCore* lithofacies. iCore lithofacies can be displayed as a graphical output with borehole images and other petrophysical data (see Fig. 3). The numerical output can subsequently be used as input into other interpretation and modeling software tools.

Applications and Case Study iCore lithofacies are a very useful resource in various phases of the planning and development of reservoirs, as well as procedures performed at the well site. Applications: Depositional environment interpretation iCore visual output highlights facies and sequences within a well so that an accurate interpretation is possible. Geological modeling iCore generated outputs include numerical values based on mathematical computations (as discussed above). These can be used as input into geological modeling software. Reservoir modeling iCore lithofacies computed in multiple wells provide clear lithological information, and the numerical outputs from iCore provide a very useful input for more detailed static reservoir models. Input for mechanical properties and earth models The mineralogical basis for the classification allows for a better understanding of the mechanical properties of formation, and the mathematical outputs provide a useful earth model input. Subsurface correlation of wells iCore run in multiple wells provides a very useful means of subsurface correlation of wells. The detailed resolution and lithofacies information can be used to delineate hydrocarbon zones as well as potential reservoir seals, and field markers, and appropriate stimulation and production procedures can be applied across the field (also see below). Three-dimensional (3D) reservoir mapping Taking iCore lithofacies to the next dimension, an iCore based 3D view of hydrocarbon reservoirs in a field allows improved exploration and production operations. Reservoir characterization iCore provides enhanced understanding for different geological and engineering aspects and greatly improves reservoir characterization. Seismic attribute analysis iCore output provides an added refinement to regional and field-wide classification for the reservoir to recognize and record seismic attributes. Wellbore stimulation The Total Clay input as well as iCore lithofacies on the final presentation provides mineralogical information that is critical to using correct reservoir stimulation and sand control. This may save from losing a well completely if incorrect procedures are used. Perforating and Completion design

Knowing the applicable rock types and their vertical extents enables improved and optimized perforating and completion design. Further, because of better decisions, there is reduced remedial and completion expenses. Field planning and development All the above are critical components of field development planning and hydrocarbon recovery. The increased understanding of the nature and characteristics of the reservoir allow optimization of the number of wells, their placement and trajectory. This also impacts facility sizing and placement for drilling and for production in the field. Overall project economics Every data stream that helps design and implement more efficient and economic procedures adds to the economic bottom line of the project. iCore results aid in and enhance so many facets of hydrocarbon exploration and production that the overall project economics are very greatly improved. Case Study: The well used in this case study is a deepwater Gulf of Mexico (GOM) well. Figure 4 shows a structure and seismic amplitude map of the case study area. iCore lithofacies have been used for some of the geo-scientific analyses conducted thus far. In this well, Sand A was cored. Wireline logs acquired included logs from a microimager for oil-based muds and a spectroscopy sonde. Sand Interval with Core Figure 5 shows an interval in the well that was cored. The iCore lithofacies indicate that this interval consists of thin interlaminated sand-shale sequences with different degrees of shaliness, as represented by the brown shades in the lithofacies track. The core photograph track also indicates that there are no thick sand or shale beds. Rather, the interval is a very finely interlaminated sand-shale sequence, and the shaliness at any depth is determined by the density of shale laminations per unit lithologic thickness. These correlations point to the ability of iCore lithofacies to identify lithology as might be observed in a core. Further, these verifications demonstrate that iCore lithofacies can be used with certainty in non-cored intervals. Pressure Barrier and Depositional Environment Interpretation in Sand Intervals Without Core Figures 6 and 7 are sand units that were not cored. In Fig. 6, iCore lithofacies indicate that this is also an interlaminated sand-shale sequence with an overall fining upwards, i.e., waning energy, trend. There are three less-laminated sand beds in the lower part. In this section, a pressure difference of 100 psi was noted from pressure measurements taken between 17,650 ft and 17,700 ft. None of the standard logs could clearly identify the pressure barrier until the iCore lithofacies were generated. From the iCore track in Fig. 6, it is very clearly evident that a shale lithofacies (brown colors) exists at 17,661 ft. Thus with the help of iCore lithofacies, the pressure barrier was identified, accurately and with a high degree of confidence. The information obtained from iCore can be used to aid in design of formation pressure and fluid sampling programs in offset wells to determine the lateral extent of these subtle, but considerable, pressure barriers. Ultimately, the information can be incorporated

into static and dynamic reservoir models.In Fig. 7, a mere glance at the iCore lithofacies track is sufficient to make the interpretation of the sand body. Unlike the sands discussed above, this unit is not interlaminated. The interval shown is a stacked sequence of turbidite sheet sands or lobes (yellow colors) separated by distinct shale interbeds (brown colors). Marker Beds The ability of iCore to resolve and assign lithofacies to marker zones is very useful for detailed well-to-well correlation. Figure 8 shows an interval that contains some thin carbonate and marl streaks (seen in the blue colors) in the iCore lithofacies track at 18,286 ft, 18,296 ft, 18,332 ft and 18,338 ft. These marker beds represent stratigraphic sequence boundaries or maximum flooding events and are ideal for well-to-well correlation across the field, as well as seismic stratigraphic interpretation. Modeling As also mentioned in previous sections, iCore lithofacies are available in a numerical format, allowing for input into reservoir modeling. Numerical outputs from iCore lithofacies in the sand unit observed in Fig. 6 will be used for reservoir modeling. Summary and Conclusions Lithofacies classification for borehole geology has, typically, been based on many different log responses and characteristics. Studies based solely on borehole images or borehole spectroscopy data are not able to provide comprehensive subsurface geological description and interpretation solutions, however, the two types of data strongly complement each other for detailed geological interpretation. Dry-weight mineralogical data from a wireline elemental capture spectroscopy device is classified utilizing a ternary diagram with a corresponding set of rules and definitions to create dry-weight mineralogy-based lithofacies. This output is further refined to a higher resolution using calibrated micro-resistivity image data and a second set of rules and definitions. The final results can then be presented in an output log as a detailed geological column displaying the calculated lithofacies. Numerical data from the final lithofacies results can also be used for additional computations as well as modeling. This new approach to lithofacies classification is extremely valuable for stratigraphic interpretation such as depositional environment, detailed well-to-well geological correlation, and static reservoir modeling. The information may also be used in the planning, execution, and evaluation of borehole pressure and fluid sampling programs. Additional applications exist for wellbore stimulation procedures, completion and production strategies, and project economics. References Cheung, P., Pittman, D., Hayman, A., Laronga, R., Vessereau, A., Ounadjela, A., Desport, O., Hansen, S., Kear, R., Borbas, T., Borbas, T., and Wendt, B., 2001, Field Test Results of a new oil-base mud formation imager tool, Transactions of SPWLA Fortysecond Annual Symposium, Paper XX. Grau, J. A., Schweitzer, J. S., Ellis, D. V., and, Hertzog, R. C., 1989, A Geological Model for Gamma-ray Spectroscopy Logging Measurements, Nuclear Geophysics, Vol. 3, No. 4, pp., 351-359. Herron, Michael M., and Herron, Susan L., 1997, Log Interpretation Parameters

Determined from Chemistry, Mineralogy and Nuclear Forward Modeling, Proceedings 1997 International Symposium of the Society of Core Analysts, September 8 10, Calgary Paper SCA 9727. Hurlbut, Cornelius S., Jr., and Klein, Cornelis, 1977, Manual of Mineralogy (after James th D. Dana), 19 Edition, John Wiley & Sons, New York, Chichester, Brisbane, Toronto, Singapore, 533 pp. Acknowledgements The authors would like to thank Shell Exploration and Production Company, with special appreciation to Michael F. Medeiros and Richard J. Stocker whose tremendous input and effort made this project viable, and also Schlumberger Technology Corporation for their support of this work and paper. The authors would also like to express their gratitude to Oswaldo Viloria and David Williamson without whose immense contributions iCore would be non-existent. The authors acknowledge the support, collaboration and encouragement from Schlumbergers North Gulf Coast Interpretation Development Team.

* Mark of Schlumberger

WSAN
0.9 0.8 0.7 0.6
Calcareous Sand Shaly Sand Clean Sand

0.1 0.2 0.3 0.4

Sand

LA WC

WS AN

0.5

0.5

0.4 0.3
Sandy Carbonate Shaly Carbonate Sandy Shale

0.6 0.7
Shale

0.2 0.1

0.8
Carbonate Marl Calcareous Shale Claystone

0.9

WCAR
0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1

WCLA

WCAR

Figure 1. Ternary diagram with dry weight mineralogical inputs and lithofacies classification. The three end points are WCLA, WSAN, and WCAR. Shale-sand ranges are from the dark brown to the yellow colors.

Claystone

Figure 2. Based on the ternary diagram in Figure 1, the above lithofacies are computed. Salt input is taken directly from the spectroscopy processed results.

20

25

Figure 3. Snapshot of a portion of the iCore lithofacies track (within the black boundary). From left, the tracks are spectroscope composite track, microimage log quality control track, iCore lithofacies track, static normalized microimage track, bed boundary dip tadpoles track. The spectroscope composite track seen here consists of dry-weight percentages from 0% on the left to 100% on the right. The red shading is total clay. The yellow shading is total QFM and is measured from the edge of the red shading and on towards the right, and the light blue shading is total carbonate and is measured from the edge of the yellow shading and on towards the right. The microimage log quality control track consists of four bars representing the four tool pads, and each bar consists of five sub-bars each representing the five button pairs on each pad. The LQC track displays one of four colors for each button-pair: green, yellow, blue or red. A green color implies good quality data, a yellow color implies marginal quality data, blue implies poor data, and red means pad lift-off from the borehole wall. The iCore lithofacies track consists of a continuous display of colors and symbols of the lithology as computed by the iCore algorithm. Yellows are sands and browns are shales. This lithology representation is bound on the left by the total clay curve. The total clay dry-weight percentages range from 0% on the left (no clays) to 100% on the right (all clay). This gives the lithofacies column a geological outcrop appearance, where the shales are weathered and the sands stand out. The static normalized image track displays an image that has been normalized using a static normalization as applied across the entire logged interval. The bed boundary dip tadpoles track displays tadpoles showing true dip determined by hand-picking bed boundaries from the image. The track represents true dip magnitudes of 0 on the left to 90 on the right on a tangential scale. The tail of the

tadpole points in the direction of true dip azimuth and is on a 0 to 360 orientation, with 0 or north being the top of the page.

10

Figure 4. Seismic Amplitude and Structural Map of the upper surface of the sand unit observed in Figure 5. The figure illustrates the complex nature of the reservoirs in this area, and the need for any and all means of correlation of wells as well as depositional environment interpretation.

X XX
Interpreted channel depositional axes Cored Well

XX

SALT

X
GOM DEEPWATER EXAMPLE

N
SEISMIC AMPLITUDE & STRUCTURE

XXX
11

2000'

Figure 5. Snapshot of a portion of the entire width of a complete geological interpretation presentation including the iCore lithofacies track. From left, the tracks are borehole attributes track, depth track, resistivity curves track, neutron-density track, spectroscopy composite track, microimage log quality control track, core photo track, iCore lithofacies track, static normalized microimage track, bed boundary dip tadpoles track, dynamic normalized microimage track, and structure-deleted bed boundary dip tadpoles track. The blue dashed lines have been drawn to emphasize correlation between the core and iCore lithofacies. This figure shows an interval in the well that was cored. The iCore lithofacies indicate that this interval consists of thin interlaminated sand-shale sequences with different degrees of shaliness as represented by the brown shades in the iCore lithofacies track. The core photograph track also indicates that there are no thick sand or shale beds. Rather, the interval is a very finely interlaminated sand-shale sequence, and the shaliness of a section is determined by the density of shale laminations per unit lithologic thickness.

12

Pressure Barrier

Figure 6. This is an interval that was not cored. iCore lithofacies indicate that this is also an interlaminated sand-shale sequence with an overall fining upwards, i.e. waning energy, trend. There are three less laminated sand beds in the lower part. A pressure difference of 100 psi was noted from pressure measurements taken between 17,650 ft and 17,700 ft. None of the standard logs could clearly identify the pressure barrier until the iCore lithofacies were generated. From the iCore track above it is very clearly evident that a shale lithofacies exists at 17,661 ft. Thus, with the help of iCore the pressure barrier was identified.

13

Figure 7. This is also a sand interval that was not cored. A glance at the iCore lithofacies track is sufficient to make the interpretation of this sand body. Unlike the sands discussed above, this unit is not interlaminated. This is a stacked sequence of turbidite sheet sands or lobes (yellow colors) separated by distinct shale interbeds (brown colors).

14

Figure 8. Here carbonate and marl streaks (seen in the blue colors) are noticeable in the iCore Lithofacies track at 18,286 ft, 18,296 ft, 18,332 ft and 18,338 ft. These marker beds represent stratigraphic sequence boundaries or maximum flooding events and are ideal for well-to-well correlation across the field, as well as seismic stratigraphic interpretation.

15