Lithofacies Lithofacies and Associated Petrophysical Properties Classification Lithofacies Prediction: Tools, Methods, Results Rock properties data represent analyses from 33 wells (below) that have attempted to sample the complete range in porosity, permeability, geographic Lithofacies, Porosity, Permeability BA L- 8 Phyloid Algal Bafflestone Gamma Ray PE Effect distribution, and formational unit for each of the logk=Alogfi+B or k=10 fi: Keystone Wells PE vs. Delta Phi (N-D) by Lithofacies Fundamental to construction of the reservoir geomodel is the population Log Response and major lithofacies. Lithofacies were described for Steps to Predict 20 Permeability Permeability Permeability Standard Standard Lithofacies Digital Description: 3-8 / 7 L1 NM Silt&Sd L2 NM ShySilt core using a digital classification system to facilitate of cells with the basic lithofacies and their associated petrophysical 8 8 data management and because it offered the ability properties- porosity, permeability, and fluid saturation. Petrophysical Lithology Lithology Equation Equation Adjusted Error Error * Primary Depositional Environments: L- 8 Grainstones L3 Mar Sh&Silt L4 Mar Mdst Lithofacies Code A B R^2 (log units) (factor) L5 Mar Wkst L6 Mar Dol 7 7 L-7 Packstone and Packstone-Grainstone Lithofacies Shankle to use non-parametric categorical analysis. Digits L7 Mar Pkst L8 Mar Grnst&PA 1 NM Silt & Sand 7.861 -9.430 0.780 0.769 5.9 ) properties vary between the eight major lithofacies classified (see lower Lithofacies Digital Description: 3-8 / 6 n Thin Section Photomicrograph o 6 6 i Newby 2-28R t generally represent continuous variation of a s s 2 NM ShlySilt 5.963 -7.895 0.702 0.787 6.1 c 10 e Lithofacies Digital Description: 3-8 / 4-5 Primary Depositional Environments: e e r left). Mean and maximum porosities increase with increasing lithofacies Shrimplin i i Core Slab, 2992' Beaty r The eight lithofacies can be discriminated c c lithologic property that may be correlated with o 5 5 3 Mar Shale & Silt 8.718 -10.961 0.719 0.847 7.0 c Primary Depositional Environments: Shoals on mid to upper shelf in either regressive or transgressive phase. In Non-cored Wells a a f f e Cottonwood Limestone (B5) l a number for the limestones (mud- to grainstone; histograms below). In situ o o petrophysical properties. Final petrophysical trends 4 Mdst/Mdst-Wkst 7.977 -9.680 0.588 0.958 9.1 h effectively by four wireline log properties. Although 4 4 h h Shoals on mid to upper shelf in either regressive or transgressive s t t i i o n L L used the eight major lithofacies shown below 5 Wkst/Wkst-Pkst 6.260 -7.528 0.774 0.611 4.1 ( (stressed) porosities ( fi) were either measured or were calculated from phase, lagoons and tidal flats. Alexander D-2 M-CG Oncoid-Pellet Alexander the log distributions of the separate lithofacies 3 3 (selection process is discussed further on). 6 Sucrosic (Dol) 7.098 -8.706 0.643 0.673 4.7 Core Slab, 3024 Grainstone ) 0 routine helium porosity (f ) values using the developed correlation: f = Determine predictor variables, lithofacies % 2 2 routine i Beaty E-2 ( show overlaps in their ranges, their collective use 7 Pkst/Pkst-Grnst 6.172 -6.816 0.840 0.521 3.3 FG Pellet Cottonwood Limestone (B5) 8 Grnst/PA Baff 8.240 -8.440 0.684 0.600 4.0 Core Slab, 2782' categories, and optimal predictor tool in an Luke i within multivariate statistical analysis results in 1 1 Core Analysis Data 1.00froutine-0.68. Packstone-Grnst Stuart h Morrill Limestone (B4) P iterative, logic-based, trial and error process. good predictions of lithofacies identity as decisions All All Digital Description: 56-505-505-414-9434 a t At porosities below approximately 6% some facies exhibited higher Digital Description: 55-503-314-9424 Newby l -10 0.5 mm e Permeability is a function of several variables including primarily pore Routine Core Analysis: rooted in probability. 0 50 100 150 1 2 3 4 5 6 7 Routine Core Analysis: D permeabilities than predicted by the power-law function. For these Whole Core Porosity (%) 18.8 GR (API units) PE (barns) throat size, porosity, grain size and packing (which controls pore body size Whole CorePorosity (%) 11.8 Digital Description: 57-607-744-1534 Perm Max (md) 39.0 Optimize neural network parameters through facies the relationship between permeability and porosity was best Perm Max (md) 1.8 Routine Core Analysis: Plug Porosity (%) 21.8 Kimzey Shankle and distribution), and bedding architecture. Equations were developed to Gamma-ray log contrasts the generally low radioactivity within i3 Plug 1 Porosity (%) 20.6 Perm (md) 30.3 cross-validation and logic-based trial and error. predict permeability and water saturation using porosity as the represented by an equation of the form: logki = A logf + B. Plug Porosity (%) 13.7 -20 the carbonates to higher levels emitted from clay minerals and silt RT (apparent) Delta Phi (N-D) Perm(md) 3.3 Perm (md) 1141 (Keystone wells) fractions. Plug 2 Porosity (%) 15.4 2 3 4 5 6 7 Shrimplin independent variable because porosity data are the most economic and Beaty Perm (md) 73.1 one for Photo Electric Effect (barns) 8 abundant, and because porosity is well correlated with the other variables Standard error of prediction ranges from a factor of 3.3 to 9.1. At fi Select two neural network models, 8 Photoelectric effect is a direct function of the lithofacies for a given lithofacies. > 6% permeability in grainstone/ bafflestones can be 30X greater wells with PE curve and one for wells without PE curve. 7 7 aggregate atomic number and can be related to mineral content. 6 6 Alexander than mudstones and >100X greater than marine siltstones of similar s PE vs. Core Grain Density by Lithofacies s e i e i Kipling c 5 porosity. Differences in permeabilities between nonmarine silt/ c 5 In situ Klinkenberg (high-pressure gas or liquid-equivalent) gas a f Develop code to automate process of 2.9 a is sensitive to both pore volume and pore fluid f L1 NM Silt&Sd L2 NM ShySilt Resistivity log o CMAC o 4 sandstones and shaly siltstones range from 3.3X at 12% porosity to 4 h t permeability (k) exhibits a log-log correlation, or power-law, relationship h i generating the Marine-Non Marine curves and neural L3 Mar Sh&Silt L4 Mar Mdst t content in the carbonates as well as conductivity caused by clay i Luke L L5 Mar Wkst L6 Mar Dol L 3 3 Stuart with porosity though the relationship changes in some facies at porosities 7X at 18%. Regression analysis required careful data filtering such mineral cation-exchange properties in the shaly lithofacies. network predictions. L7 Mar Pkst L8 Mar Grnst&PA 2 2 below ~6%. Each lithofacies exhibits a relatively unique k-f correlation as to removed data from fractured samples. Full-diameter cores 0.5 mm 0.5 mm ) c 2.8 c 1 1 frequently exhibit permeabilities as great as 50X plug permeabilities / Newby that can be represented using equations of the form: Delta Phi transform (neutron minus density porosity) is Thin Section Photomicrograph Close-up Core Slab Close-up Core Slab Thin Section Photomicrograph Close-up Core Slab due to stress relief fracturing. Generate predicted lithofacies and probability m highly correlated with the grain density of the lithofacies and is All All g curves and output LAS format files through batch ( y 0.1 1 10 100 -20 -10 0 10 20 30 t Kimzey i 2.7 Rta (ohms) Phi Units (%) processing of input LAS files. s n PE Rock L-5 Wackestones and Wkst-Packstones e Mudstones-Bafflestones 5.1 Limestone Porosity Histogram Porosity Histogram D 5.1 Anhydrite L1 NM Silt & Sd L5 Mar Wkst (Key wells are named) n 100 Lithofacies Digital Description: 3-8 / 2-3 i NM Clastics Limestones 1.8-6.3 Shale (~3.4 is common) 30% a 25% r 2.6 ) Primary Depositional Environments: Mdst/Mdst-Wkst L4 3.1 Dolomite NM Silt & Sand L1 G d Automation Mostly in carbonate dominated mid shelf, some lagoon and tidal flat 1.8 - 2.7 Sandstone (Qtz = 1.81) L2 NM Shaly Silt L6 Mar Dol s Wkst/Wkst-Pkst L5 NM Shaly Silt L2 s L-4 Mudstones and Mdst- Wackestones Kipling/CMAC tesselation of n 25% m n 10 Significant custom programming by Bohling has made Density Rock ( o 20% Pkst/Pkst-Grnst L7 o i by: Alan Byrnes, Martin Dubois i Newby 2-28R predictor variable (well log) space Digital Rock Classification System t 2.88 Dolomite (2.83 on logs) t Lithofacies Digital Description: 3-8 / 0-1 y Algal-Mixed Skeletal possible the generation of the Marine-Nonmarine predictor a Grst/Grst-Baff L8 a l t l 20% Core Slab, 2986' in two and three dimensions. This 2.5 2.71 Limestone L3 Mar Silt & Sh L7 Mar Pkst Primary Depositional Environments: i DIGIT # 1 2 3 4 5 6 u u l Wackestone variable LAS curve through a Visual Basic routine in MS p 15% i Cottonwood Limestone (B5) 2.65 Sandstone Rock Dunham/Folk Consolidation/Fracturing Argillaceous Grain Principal p CODE Siliciclastic or carbonate dominated mid shelf. non-parametric classification 2 3 4 5 6 7 o Type Classification Content Size Pore Type o 1 b Access that allowed the mining of the KGS Oracle data 2.98 Anhydrtite P 9 Evaporite cobble conglomerate unconsolidated Frac-fill 10-50% vcrs rudite/cobble congl (>64mm) cavern vmf (>64mm) P 15% technique was considered but not a Photo Electric Effect (barns) L4 Mar Mdst L8 Mar Grnst & PA f 8 Dolomite sucrosic/pebble conglomerate poorly cemented, high porosity Frac-fill 5-10% med-crs rudite/pebble congl (4-64mm) med-lrg vmf (4-64mm) f Digital Description: 52-081-014-1413 Variable Shale e base, automated cross-validation exercises via R-language o o Thin Section Photomicrograph 7 Dolomite-Limestone baffle-boundstone/vcrs sandstone cemented, >10% porosity, highly fractured Shale >90% fn rudite/vcrs sand (1-4mm) sm vmf (1-4mm) 10% Silty Wackestone used in favor of the neural net.