Analysis of Critical Permeabilty, Capillary Pressure and Electrical Properties for Mesaverde Tight Gas Sandstones from Western U.S. Basins [report]

Alan Byrnes, Robert Cluff, John Webb, John Victorine, Ken Stalder, Daniel Osburn, Andrew Knoderer, Owen Metheny, Troy Hommertzheim, Joshua Byrnes, Daniel Krygowski, Stefani Whittaker
2008 unpublished
U.S. Energy Information Administration gas supply estimates predict that Unconventional gas sources will be the dominant source of U.S. natural gas supply for at least the next two decades. Mesaverde Group tight gas sandstones will play an important role. To understand the reservoir properties accurate tools for formation evaluation are needed. This project provides petrophysical formation evaluation tools. Tasks involved included a review of the research plan by DOE (Task 1); initial
more » ... ; initial technology assessment for DOE (Task 2); collection and consolidation of published advanced rock properties data into a publicly accessible digital database (Task 3.1); collection of 2216 (300 in original proposal) rock samples, with digital wireline logs, where available, from 44 wells in six basins (Washakie -11; Uinta -8; Piceance -8; Greater Green River -7; Wind River -4; Powder River -6; Sand Wash -2) including seven cores and wells contributed by six petroleum companies (3.2). Measurement of basic petrophysical properties (Task 4.1). Measurements on selected samples included: 1) drainage critical gas saturation (4.2); routine and in situ mercury intrusion capillary pressure analysis (4.3); porosity exponent and multi-salinity electrical conductivity measurements(4.4); geologic properties including core description, thin-section microscopy, including diagenetic and pointcount analysis (4.5); and standard wireline log analysis (4.6). The compiled published data and data measured in the study were input in a database (Task 5.1); and are provided online as a webbased database (5.2). Core and wireline log-calculated properties were compared and algorithms developed for improved calculation of reservoir properties from log response (Task 6). The scale dependence of critical gas saturation was evaluated (Task 7). An active web-based, publication, and short-course technology transfer program was conducted (Task 8) including presentation of all data on the project website (http://www.kgs.ku.edu/mesaverde). Advanced rock properties data were compiled from 88 published studies. A total of 2216 core plugs were obtained representing 1182 original plugs (A), 776 paired plugs (B), and 258 additional pair plugs (C). This sampling represents approximately four times more original plugs than the 300 core plugs proposed and six times as many paired plugs (proposed n=150). Core samples range in depth from 124-16,723 ft, reflecting the range in depth of the Mesaverde for the basins studied. The cores also represented the range of porosity, 0-25%, and in situ Klinkenberg permeability, 0.000001 mD-200 mD. Grain density distribution averages 2.653+0.04 g/cc exhibiting a slight difference in distribution among basins. In situ porosity was correlated with routine porosity and shown to follow a crack compressibility model (φ i /φ o =AlogPe+B) with compressibility increasing with decreasing porosity. The Klinkenberg constant, b, increases with decreasing permeability and in situ Klinkenberg permeability (mD) can be related to routine air permeability using either logk ik =1.34log k air -0.6 or logk ik =-0.0088(logk air ) 3 -0.0716(logk air ) 2 +1.366logk air -0.4574. Permeability can be predicted within an approximate standard error of +3.5X using: logk ik =C 1 φ i +C 2 RC2 log +C 3 where the coefficients are defined for three major lithofacies by basin. Critical gas saturation measurements, performed on 150 lithologically diverse samples of variable porosity and permeability, support the commonly applied assumption that S gc < 0.05 at core scale but is scale dependent. Heterolithic samples indicate the dependence of S gc on pore network architecture. Concepts from percolation theory and upscaling indicate that S gc varies among four pore network architecture models: 1) percolation (N p ); 2) parallel (N // ); 3) series (N ⊥ ); and 4) discontinuous series (N ⊥d ). Analysis suggests that S gc is scale-and beddingarchitecture dependent in cores and in the field. The models suggest that S gc is likely to be low in DE-FC26-05NT42660 Final Scientific/Technical Report xiv laminated and massive-bedded sandstones but in cross-bedded lithologies exhibiting series network properties, S gc can range widely but can reach high values (e.g., S gc < 0.6). Hysteresis drainage and imbibition capillary pressure measurements on 33 samples show that residual and initial non-wetting phase saturation can be related by a Land-type relation (1/S nwr *-1/S nwi * = C). Routine and in stu mercury injection capillary pressure analysis on 81 core pairs show that capillary pressure measurements on low-permeability sandstones are significantly influenced by confining stress, consistent with observed permeability changes. Threshold entry pressure increases with decreasing permeability. In situ and unconfined curves for high-permeability cores (k ik > 1 mD) are nearly identical. With decreasing permeability the difference between unconfined and in situ threshold entry pressure increases. For all pairs this difference is greatest at the threshold entry pressure and decreases with decreasing wetting-phase saturation. It can be interpreted that confining stress exerts principal influence on the largest pore throats and that pore throats accessed at non-wetting phase saturations below approximately 50% are not significantly affected by confining stress. A total of 907 resistivity measurements on 308 core samples were performed at various salinities. These data indicate that resistivity in these rocks is influenced by both conductive clays and pore architecture. Contrary to conventional models, the Archie porosity exponent decreases for all salinities with decreasing porosity below approximately φ i =6%. Over 550 core images were obtained and 150 thin sections micro-photographed and analyzed. Point-count data provided the basis for lithologic characterization and porosity typing. Rock lithologic properties were shown to correlate (and probably control/influence) petrophysical properties. Differences between properties of marine and fluvial rocks are evident. Both Standard and advanced log analysis was performed on all the primary wells. Wireline log-calculated properties (φ, S w , Lithology) were compared with core-derived properties A zoned grain density model based on geologic knowledge of the section, tied to core grain densities, offers the best approach for single-log porosity determination. Overall, the shale corrected density-neutron cross-plot porosity is the best predictor of in situ porosity. A series of models were analyzed that parametrically investigate the role of total bed thickness, thin high permeability bed permeability, and vertical permeability on cumulative gas recovery. The influence of a single 1-foot (0.3 m) thick higher permeability bed on cumulative gas production can be very significant. Over 9 gigabytes of data are available for download from the Project Website (http://www.kgs.ku.edu/mesaverde/) comprising 1) Excel workbooks containing tables of data from previous studies; 2) Excel workbooks containing data for all petrophysical measurements performed in this study including: 2,102 helium porosity, 2,075 routine air permeability, 2,062 in situ Klinkenberg permeability, 2,101 grain density measurements, 907 electrical resistivity measurements, 301 mercury intrusion capillary pressure analyses, 150 air-brine critical gas saturation measurements, 113 pore volume compressibility analyses, 310 air-brine in situ porosity measurements; 550 core slab images representing the range of lithofacies exhibited by the Mesaverde in the six basins studied; 750 thin-section photomicrographs from 41 wells; 6,447 feet (2,054 m) of digital core descriptions presented both in Excel workbook format and in graphical core descriptions for 42 wells from 6 basins; graphical core descriptions of core from 42 wells; 21 standard wireline log analyses; 21 advanced wireline log analyses; pdf files of all technical slide and poster presentations; pdf files of all technical quarterly reports. Two publications, seven technical presentations, a one-day workshop at the AAPG Annual Convention, and technical talks at several society lunches were presented. 5) Facies and Upscaling -Beyond investigating the above fundamental properties for representative lithofacies in the Mesaverde, it is necessary to know how critical gas saturation, capillary pressure, electrical properties, upscaling issues, and wireline log response and analysis change with more easily measured Mesaverde rock properties such as lithofacies, porosity, and permeability; and how flow properties, particularly critical gas saturation, upscale with lithofacies bedding architecture. In addition, accuracy and variance of petrophysical relationships are premised on sampling, the scale of sampling, measurement methodology, and the geostatistical or spatial distribution of the properties. Little published work is available that addresses how porosity or permeability change over short length scales (2.5-5 cm; 1-2 inches) 6) Wireline Log Interpretation -Petrophysical properties and relationships measured on core and at the core scale can provide critical reservoir characterization information, but core cannot reasonably, or economically, be obtained for most wells over entire intervals of interest. For this reason, core are used for calibration of wireline log response interpretation so that log algorithms can be used where core are unavailable. This requires that the wireline log response curves be correlated with core-measured petrophysical properties. These relationships can vary with such properties as rock lithology, petrophysical property, in situ conditions, log vendor, log vintage, log traces available in the logging suite, and the log algorithms developed and used. Algorithms can sometimes be developed that meet reasonable accuracy and precision standards but that require a suite of input logs that are unavailable for historical wells and/or are prohibitively expensive for new wells. Determining the number of unique lithofacies classes and the criteria for defining classes can involve four principal criteria: (1) maximum number of lithofacies recognizable using the available petrophysical wireline log curves and other variables; (2) minimum number of lithofacies needed to accurately represent lithologic and petrophysical heterogeneity; (3) maximum distinction of core petrophysical properties among classes; and 4) the relative contribution of a lithofacies class to storage and flow.
doi:10.2172/971248 fatcat:reedwuqhavasnadhbk6qm5pnji