Techniques for Assessment of Shale Gas and their Applicability to Plays from Early Exploration to Production Kurt Steffen, PhD
Outline Discuss techniques used by ExxonMobil to rapidly evaluate Shale Gas and Coal Bed Methane opportunities in the opportunity ID stage Uses Bayesian Belief Networks Integrated with Geographic Information Systems Discuss how these techniques can be integrated into a standardized volumetric assessment workflow Discuss techniques that use automated decline curve analysis to predict well EUR for unconventional wells 2
Key Observations of Shale Gas Reservoirs Many shale gas reservoirs are essentially gas-mature source rocks: OM Significant volumes of organic matter: 5 10 wt% significant volume of rock Parallel bedded connected network of OM OM = particles, ECP coatings of clays, DOM in φ waters High LOM: Expanded pores/microfractures around OM due to vol during oil generation Most clays transformed to illite, with an increase in µφ among clay minerals Oil-saturated pathways out of rock Retained oil in pores nearby OM cracks to gas Remnant OM = Petroleum Coke with significant porosity OM Biosilica OM Biosilica 3 1 mm
What Matters in Shale Gas?...The Summary Distinction between complicated and complex systems First Order: Thermal Maturity TOC res Reservoir Pressure Net Interval Thickness (and distribution) Areal Distribution degree of heterogeneity Fracability Modifiers: Fractures (and type) Adsorbed Gas vs Free Gas Types and amount of matrix porosity Depth OMT Lithology (silica vs carbonate vs clay) Non-hydrocarbon gas distribution 4
North American Shale Gas Plays Complex systems Unconventional Basins Shale Gas Plays are complex systems with interaction of multiple parameters Basin Areas Gas Play Areas Unconventional: Shale Gas Tight Gas, and Coal Bed Methane (CBM) E B H NA M Individual parameters vary widely Not all parameters need to be world class to have world class SG play Additional successful plays likely to expand range of parameters Parameter 1) Richness 2) Net thickness 3) Maturity 4) Depth Average Value in Core Area 5) Basin History 6) Pressure 7) Mineralogy 5
Why BBNs for Unconventional Gas Resource Evaluation? In Unconventional Resources (Shale Gas, Tight Gas, and Coal Bed Methane) the hydrocarbon system elements are largely controlled by the properties of a single lithology or closely spaced groups of lithologies. The ability to commercially extract natural gas from Unconventional Resources represents the primary risk in these resource types, and the commerciality of a particular play varies spatially within the region of hydrocarbon occurrence. It is critical to identify and evaluate the commerciality of Play Fairways (aka Sweet Spots) and differentiate these fairways from non-commercial areas. Although the properties of a single lithology can control Unconventional Resource commerciality, the properties of that single lithology represent the complex interaction of sedimentation and basin/tectonic evolution. Unconventional Resources require artificial stimulation (generally hydraulically induced fractures) in order to produce gas at commercial rates. Understanding/predicting commerciality in Unconventional Resources therefore requires understanding of a complex natural system and how that system will respond to engineering intervention. 6
7
8
9
10
11
12
13
14
15
Expert Knowledge (Explicit, Implicit, Tacit??) GIP / Mile 2 = SCF/Ton * Thickness* Density Empirical Relationships k = Theoretical Ae Ea/RT 16
Quantification of Expert Knowledge/ Bayesian Belief Networks (BBNs) Uses a Causal/Belief Net to describe the relationship between variables BBNs break understanding of the system into small, tractable piece BBNs are very flexible, we can use multiple sources to quantify the relationships as well as calibrate the results Acknowledges the multi-variable, complex nature of Unconventional Resources Because the process is automated, large numbers of polygons can be run thru the system, approximating the continuous nature of the continuous resource. Less sensitive to arbitrary cut-offs if well calibrated 17
Exploiting the BBN: Linking to Maps To exploit the power of BBNs as a predictive tool, numerous tools that link BBNs to Geographic Information System (GIS) data have been developed within ExxonMobil. These tools: Allow for output maps to be created based on BBN output and GIS based inputs Visualize probabilistic output. Aggregate probabilistic results to summarize resource estimates for areas of interest. A GIS enabled database of calibration data has been created that allows for the comparison of BBN predictions to actual results. These validation tools are used by the expert to improve predictive ability and gain insight into Unconventional Resources systems. 18
Example Oportunity ID Stage Few modern wells drilled into formation, minimal production information Five Input Maps Measured Depth (Feet) Net Isopach (Feet) Maturity (R0) Land Use (Rural vs. City) Well Cost (MUSD) Other Information (non-mapable, spotty) TOC (3 to 5 %) Operators getting Commercial Rates (IPs) Pressures Gradients Elevated (0.5-0.55 PSI/Foot) Normal Temperature Gradients (1.22 C/100 Feet) Siliceous Shale Type II Kerogen - Marine High Sg, Low CO2 19
Input Maps Depth Maturity Isopach Land Use Well Cost Other Information (non-mapable, spotty) TOC (3 to 5 %) Operators getting Commercial Rates (IPs) Pressures Gradients Elevated (0.5-0.55 PSI/Foot) Normal Temperature Gradients (1.22 C/100 Feet) Siliceous Shale Type II Kerogen - Marine High Sg, Low CO2 20
Output Maps Gas Filled Porosity SCF/TON GIP/Mile Well EUR EUR/Mile USD / KCF 21
Assessment Workflow GEOLOGIC ANALYSIS ArcGIS Play Assessment Tool Condensate Yield Recovery Factor Assessment Input Thickness Net/Gross Area Gas Content Coal Density Assessment Output Resource Volumes Seq. Strat. EOD Lithology Thickness Lateral Extent Gas Content TOC Maturity/Rank Structure Burial Depth Stress Field Hydrology Etc. Chance of Adequacy Reservoir presence Reservoir quality Source presence Source maturity Migration pathways Migration timing Trap seal Trap closure Hydrocarbon recovery + + Commodity 22 Volumetric Parameters 50 250 MOEB Probabilistic Distribution for Subplay 500 MOEB 3000 Probabilistic Distribution Entire Play
Techniques to Monitor Mature Plays Assessment techniques mature from BBN and volumetric to performance based as the play progresses Markov Chain Monte Carlo techniques are used to perform automated decline curve analysis to statistically predict well EUR Performs decline curve analysis on individual wells while simultaneously treating the wells as a population which improves EUR prediction for wells with short production periods Easily handles the many thousands of wells in unconventional plays Readily updated as new production data becomes availably (often monthly) allowing up to date situational awareness Intrinsically consistent 23
Well EUR Map West East Large Green Dots = High EUR Small Red Dots = Low EUR EUR estimated using automated, statistical technology created by ExxonMobil utilizing monthly production Monthly production from IHS 24
Variability within and Between Subplays Number of Wells Significant performance variation between and within Subplays EUR 25
EUR by Year Drilled Evaluation of Learning Curves -Locations -Operations 26
Prediction of Future Flow Streams Future production predicted if no more wells are drilled in play Population statistics can be used to predict flow streams with additional wells 27
EUR/Operator/Date Completed Performance trends are different for each operator 28
Outline Discuss techniques used by ExxonMobil to rapidly evaluate Shale Gas and Coal Bed Methane opportunities in the opportunity ID stage Uses Bayesian Belief Networks Integrated with Geographic Information Systems Discuss how these techniques can be integrated into a standardized volumetric assessment workflow Discuss techniques that use automated decline curve analysis to predict well EUR for unconventional wells 29