egamls Inc. Principals: Eric Eslinger (Geologist/Petrophysicist) and Alan Curtis (Reservoir Engr/Geomodeller) What we do: Well and field studies using GAMLS software (plus GAMLS licensing) GAMLS focus: Integration of well log and core data for innovative reservoir characterisation GAMLS Modules: BASIC - Facies delineation and prediction using Multivariate Clustering Analysis MFBG - Mineralogy and chemistry-based volumetrics via Matrix-Fluids Balance in GAMLS PUD Probabilistic Upscaling and Downscaling (including grid-building) LVP Lithology-based Velocity and pore Pressure modelling SDVC - Seismic Depth and Velocity model Calibration CUSP - Classification and Upscaling of Saturation-dependent Properties CUSP 3D Plug-in to Petrel for modelling Saturation-dependent Properties SAA - Seismic Amplitudes and Attributes (recoding in progress) Contacts: Eslinger (USA): +1 518 852 4666 e.eslinger@egamls.com Curtis (Australia): +61 400 699 199 alan.a.curtis@egamls.com Website: egamls.com
BASIC - Facies delineation, correlation, and prediction via probabilistic Multivariate Clustering Analysis (MVCA) based modelling (BASIC contains a full suite of data-handling routines)
HCQL (HydroCarbon Quick Look) - a routine within BASIC (see also next view)
HCQL Output Summary (auto-output from previous view)
MFBG (Matrix and Fluid Balance modeling via GAMLS) - mineralogy-based forward modeling of log and core data to generate Por, Sw, Perm profiles (Calibrate using core data when present; build models using cored well; predict facies and RQ properties in non-cored wells) For a description and discussion of MFBG: Eslinger, E. and R. V. Everett, 2012, Petrophysics in Gas Shales, in J.A. Breyer, ed., Shale reservoirs - Giant resources for the 21 st century: AAPG Memoir 97, p. 419-451.
MFBG - Interaction Among Key Routines Cluster>>ModeAssign Modes (electrofacies) are assigned to rock types (ss, sltstn, sh, ls, ds ) Core>>RCA Core>>Min>>Stats Routine Core Analysis data is imported & stored Input Input cluster, well, logs, core data, cutoffs Mineral properties database Main>>Lith-Use Input of ECS log data & output of Grn Den, TPOR External Lith Genetic Algorithm solver to aid model balance Out Model output by sample depth AdsGas Profiles for adsorbed gas from adsorption data plus kerogen &/or TOC data Internal MFBG models (hard drive) Lith/Min files MFBG models GAMLS Projects GA Use mode min-suites from Lith or import from Min>>Stats Elm (simple) Core-to-Log correction Select min-suites for generic rock types (liths) Select mode-dependent constants (e.g., Rw ) Compare model output (por, perm, Sw, pay ) by mode with log & core data Depth>>CoretoLog Construct profiles for T, P, Sal, Rw, por, den Lith /Min files Min Main>>Balance Well>>Log Lab mineralogy (XRD, FTIR ) wt % by clustering mode Main>>Variables Rock Fluids For a description of how the genetic algorithm is used to obtain mineral and well log properties balance, see: Eslinger, E. and F. Boyle, 2011, Mineral and Well Log Balance for Computing Porosity in Argillaceous Rocks Application in Gas Shales, International Unconventional OilGas Conference, China University of Petroleum, Qingdao, China, July 4-5. Elm (robust) Input of modern elemental spectroscopy logs, & output of computed mineralogy into upstream MFBG for recomputation of TPOR MFBG components are inside the medium blue shaded area
Elm Robust (routine in MFBG) - Partitioning of spectroscopy log elements into minerals Tracks 1-3: red = core mineralogy; black = computed mineralogy from spectroscopy log; core mineralogy is used to calibrate the partitioning of elements into minerals The figure at right is from: Eslinger, E. and F. Boyle, 2013, Building a Multi-Well Model for Partitioning Spectroscopy Log Elements into Minerals Using Core Mineralogy for Calibration, SPWLA 54th Annual Logging Symposium, June 22-26.
Using GAMLS to integrate whole core CT data with standard well logs and whole core plug data These figures are from: Eslinger, E. et al., 2014, Probabilistic Facies Assignments in the La Luna Formation, Middle Magdalena Basin, Colombia, from Standard Well Logs Using Whole Core CT Scan Data as Initialization Input,, URTeC 1934648 (Unconventional Resources Technology Conference), Aug. 25-27, Denver, CO, 15 p.
RF (Rock-Fluids) - Routine within BASIC ; generates profiles for rock and fluid properties above and below well-logged intervals (Output used in MFBG, LVP, CUSP, )
LVP - Lithology-based Velocity and pore Pressure modeling) What It Does: Uses seismically-determined or modeled velocities Evaluates pore pressure using Bower s equation Permits comparison of seismic-based and log-based pressures Lithology-based: uses clustering-derived rock types during modelling Permits incorporation of drilling and engineering events GAMLS Interface - Index View
LVP (Lithology-based Velocity and pore Pressure modeling) Schematic Note: P is used for fluid pore pressures; S is used for directed stresses (in solids)
LVP - Work Flow
LVP - Establishing general velocity profile (independent from seismic) _ Sobn r y l_ar Ph_RF IntVe RF Athy porosity RaymerWyllie travel time h At o yp it s o a critical factor for modeling adjust porosity & travel time inputs so that modeled DT merges with well log DT (while also generating a reasonable porosity profile) GR R B O H
LVP - Selected output; note lithologic control from MVCA of wireline well logs adding Ssh_eff & Psh Psh = Sobn Ssh_eff Legend includes rock type assignments of MVCA results (from Cluster>>Mode Assign)
LVP - Selected output Depth vs psi (left) (right) Depth vs ppg
CUSP (Classification and Upscaling of Saturation-dependent Properties) Classification at all Scales to Ensure Properties Representativeness and to Permit Upscaling Representative Core Plug Cored Wireline Log Section Classification of Core Plugs into Petrotypes Representative High Res Sub-sample Representative Whole Core Classification of Plug Scoping Scans into Millitypes Construction of Pore Network Model Petrotype Log Classification of Classification of Wireline Logs into Whole Core into Electroclasses Electrotypes Classification of Whole Core into Petrotypes Derivation of Petrophysical Properties
CUSP and PUD (Probabilistic Upscaling and Downscaling) Properties Distribution from Representative Cells at the Fine Scale and Upscaling to Coarser Scale Representative Core Plugs Model using Cored Intervals Population of Plug Properties into Whole Core, by Petrotype Upscale from Millitypes to Petrotype Representative High Res Sub-sample Population of Core Plug Scoping Scans with Properties (m) Upscale to m Representative Whole Core Wireline Log Data Points Propagation of Properties along rest of Logs, by Electroclasses Population of Properties into Cored Intervals, by Electrotypes Upscale Petrotype to Electrotype Population of HR Sub-samples with Properties Sub-sample Petrophysical Properties from PNM
CUSP - Upscaling Pc(Sw) curves from Petrotype (plug scale) to Electrotype (well log scale) and development of a Rep (Representative) Curve for one of the Electrotypes Pc(Sw) curves (all depths) after upscaling and MVCA classification into six Electrotypes Rep Curve (bold dashed line) for orange Electrotype
CUSP - Depth profiles (wrapped) showing MVCA results for two Electrotype models (Tracks 1 and 2), and profiles for Sw and Swir (blue and green, Track 3) assuming a FWL at 3550 m. The saturation profiles were developed within CUSP, completely independent from well log resistivity measurements.
Calibration of Depth Imaged Seismic Data using an Interval Velocity Scaling Method SDVC (Seismic Depth and Velocity Calibration) recalibration of a previously imposed seismic velocity model to force seismic markers to tie with well log tops; adjusted velocities are distributed within 3D volume via scaling factors Seafloor Salt Depth Upper Formation Lower Formation 64th EAGE Conference, Florence Images from: Curtis, A. A., et al., 2002, Calibration of depth imaged seismic data using an interval velocity scaling method, 64th EAGE Conference, Florence, Italy.
SDVC Calibration of Depth Imaged Seismic Data Well-based Scaling Procedures - Scaling Factor Derivation interval velocity (ft/s) and scaling factor (x 10000) Well Velocities and Scaling Factor Derivation 4000 4000 6000 6000 8000 Well-based interval velocity 10000 12000 Interval velocity scaling factor 8000 Seismic Seismic imaging imaging interval interval velocity velocity (grid lines at 2000 ft increments) depth (TVDss) 10000 12000 Calibrated interval velocity 14000 Geological markers 16000 18000 20000 64th EAGE Conference, Florence 1 2 3 4 5 22000 Using IVS Using well-based adjustment velocity only Tied Adjusted Original imaging imaging imaging markers markers markers
SDVC Calibration of Depth Imaged Seismic Data Well-based Scaling Procedures - Calibration Line Disposition Seafloor Salt Depth Upper Formation Lower Formation 64th EAGE Conference, Florence Vertical Calibration Lines
SAA - Seismic Amplitudes and Attributes (recoding in progress) (MVCA of seismic amplitude and attributes and integration with core and log data) Examples of MVCA results. Top (left and right): using only SAs as clustering variables. Bottom (left and right) using a combination of well log properties (density, neutron, gamma ray...) and core plug properties (porosity and permeability) as clustering variables [Gonzalez, R.J., S.R. Reeves, E. Eslinger, and G. Garcia, 2007, Development and application of an integrated clustering/geostatistics approach for 3D reservoir characterization, SACROC Unit, Permian Basin, SPE 111453-PP, 41