Horizontal Shale Well Refracs: The Next Evolutionary Step © 2010 Baker Hughes Incorporated. All Rights Reserved. SPE Ft Worth Section Monthly Meeting March 26, 2015

Enhanced Crude Oil Recovery Potential – An Estimate T. Doscher; F. Wise - Shell Oil Company; SPE #5800; 1976 © 2010 Baker Hughes Incorporated. All Rights Reserved. 2 “The foremost conclusion of this report, in common with many others, is that large, new domestic sources of crude oil need to be discovered and developed soon.” “The API data leave little doubt that there is and will be a very large amount of oil discovered that will not be produced with the technology that we have developed to date.” “In addition to the need for large, new domestic sources of petroleum to be discovered, there is a potential for a sizeable addition of domestic reserves if recovery efficiency can be increased.”

Historical Crude Oil Prices – 2014 Dollars From MacroTrends.net © 2010 Baker Hughes Incorporated. All Rights Reserved. 3

Historical US Crude Oil Production From EIA © 2010 Baker Hughes Incorporated. All Rights Reserved. 4

Horizontal Well Inventories Major US Shale Formations © 2010 Baker Hughes Incorporated. All Rights Reserved. 5

Refrac Theory © 2010 Baker Hughes Incorporated. All Rights Reserved. 6

Why Refrac? Reconnect to existing fractures that have closed or been damaged – Unpropped frac area – Proppant Crushing/Embedment – Shale creep (Partial Monolayer? For how long?) Closing of natural fractures opened by original treatment or unpropped area Unpropped near wellbore restrictions – Spalling and fines plugging Bypassed Pay – Unproductive Perforations – Some % of perfs do not contribute or are poor performers? Create more surface area – Target perforated intervals previously under-stimulated – Contact new reservoir rock if there is stress reorientation © 2010 Baker Hughes Incorporated. All Rights Reserved. 7

Stress Reorientation Mechanical Stress Changes – The result of physically separating the rock and permanently displacing it Poroelastic Stress Changes – The result of changing the pore pressure in the rock matrix itself – Non-uniform depletion Helping Hands – Relatively low stress anisotropy – High pressure drawdowns © 2010 Baker Hughes Incorporated. All Rights Reserved. 8

Stress Reorientation Vertical Wells © 2010 Baker Hughes Incorporated. All Rights Reserved. 9 Area of stress reversal Original Frac Refrac

Stress Reorientation Horizontal Wells © 2010 Baker Hughes Incorporated. All Rights Reserved. 10

Stress Reorientation Horizontal Wells © 2010 Baker Hughes Incorporated. All Rights Reserved. 11

Stress Reorientation Horizontal Wells © 2010 Baker Hughes Incorporated. All Rights Reserved. 12

Stress Reorientation and Natural Fractures Horizontal Zipper Fracs © 2010 Baker Hughes Incorporated. All Rights Reserved. 13

Far Field Stress Reorientation SPE # ; ExxonMobile Upstream Research and XTO © 2010 Baker Hughes Incorporated. All Rights Reserved. 14 Well #1 Far Field Offset w/ SR

Candidate Selection (Empirical/Statistical) Early completions where lessons were still being learned Prematurely terminated frac stages Production Signatures – Proxy for reservoir quality IP and decline rates Relative to completion variables such as stage count, # proppant, gal fluid, etc. Cross-Plots of completion efficiency, production response, and reservoir quality Time since original completion Mechanical integrity © 2010 Baker Hughes Incorporated. All Rights Reserved. 15

16 © 2014 Baker Hughes Incorporated. All Rights Reserved. Refracture Viability Poorer Candidate Metric Better candidate Measure High Depletion LowCum prod Low Reservoir quality HighHigh IP, phi*h Low Distance from nearest neighbor HighDistance High Anomalous production relative to neighbors LowCum prod Low Effective reservoir contact High% well in zone High Past fracture intensity Low Lb/ft, ft/stage, clusters/stage Good Past fracture job execution Poor No. of screenouts Good Past fracture design Inadequate Comparison to current practices Poor Annular isolation Good Completion type Smaller Available ID LargerID Poor Wellbore condition Good Well age, corrosion issues Reservoir Past Fracture Wellbore

17 © 2014 Baker Hughes Incorporated. All Rights Reserved. Consolidating the data together

Candidate Selection (Geomechanical Stress Modeling) Integrated reservoir modeling Geomechanical Stress Simulation – Modeling vector stress changes with production and existing fracture network Confirm with microseismic © 2010 Baker Hughes Incorporated. All Rights Reserved. 18 SvSv S hmin S Hmax From Saudi Aramco UCS Pp

Lithofacies Seismic 2D,3D Well info Tops Casing Point Seismic 2D,3D Well info Tops Casing Point Static Reservoir Model Geomechanical Model Geomechanical Model Dynamic Reservoir Model Frac Model Shale Asset Evaluation Work Flow Geology Structure, Zonation Geology Structure, Zonation Petrophysics TOC, Ro, Porosity, Hydrocarbon Saturation, GIP Rock mechanics Density, Acoustic, Image, cores Mineralogy Petrophysics TOC, Ro, Porosity, Hydrocarbon Saturation, GIP Rock mechanics Density, Acoustic, Image, cores Mineralogy Microseismic Geomechanical effects Predict Performance Geomechanical effects Predict Performance Predict Well Performance Predict Well Performance Proppant Placement

Methodologies Cement squeeze the lateral and reperf Slim-hole liners with cement or viscous gel OR expandable inner liners Stick pipe or coiled tubing with straddle packers Diverters – Near wellbore – Far field – Critical logistics © 2010 Baker Hughes Incorporated. All Rights Reserved. 20 At least you know where in the wellbore its going You have an idea of where its going; It sure would be nice to have microseismic or at least tracer logs

Great Expectations Public data source results show 50-75% EUR increases per well are reasonable expectations – 100% EUR increases are possible in some instances; but… – Reports are success biased – Little information on failures – Theft production from offsets? – SPE # Tell-tale signs that new reservoir rock is being exposed – ISIP comparable to original completions – Unusually high friction pressures (tortuosity from fracture reorientation) – Lower GOR, approaching original production values © 2010 Baker Hughes Incorporated. All Rights Reserved. 21

Ongoing Challenges Predicting outcomes Placement techniques – Fluid selection – Pump schedule – Pressure management – Leak-off control – Logistics How to pump 3-5 million lbs of proppant in 18 hrs? – Long continuous pump times for diverter refracs © 2010 Baker Hughes Incorporated. All Rights Reserved. 22 Lower maximum STP? Effect on rate? Effect on fluid choice? Proppant choice? Effect of all on frac development?

Final Thoughts Ultra low recovery rates from primary completions leave substantial reserves unrecovered Need to renew existing fracture paths and create new ones to improve recovery Empirical and Statistical Review of well population to identify candidates Geomechanical stress modeling, confirmed or modified with microseismic and history matching to further improve our understanding of reservoir dynamics Stage frac designs are unlikely to look like the primary completions due to changed well conditions and dynamics Incremental EUR of 50% - 75% are reasonable expectations and more is possible – Further understanding of the negative consequences are still needed © 2010 Baker Hughes Incorporated. All Rights Reserved. 23