Five Things I Wish a Geologist Had Taught Me Confessions of a Frac Engineer Mike Vincent mike@fracwell.com Fracwell LLC Microseismic image: SPE 119636 Proposal G&G folks often have tremendous advantages over PEs Familiar with how rocks break Understand reservoir laminations and compartments Ability to visualize proportions Less contaminated by simplified models and established rules of thumb Many others I ve screwed up a bunch of fracs in the past. I wish you had helped me. There are more than 5 things I failed to understand, but that is all I have time to review today! 1
Do we model fracs correctly? We picture fracs as perfect vertical planes without restriction to hydrocarbon flow We have created hydraulic fracs 22 ft half-length but less than.1 inches wide Fracs are very narrow ribbons, massively long! Frac length frequently thousands of times greater than the wellbore diameter 3 SPE 128612 Observations of Fracture Complexity Physical evidence of fractures nearly always complex NEVADA TEST SITE - HYDRAULIC FRACTURE MINEBACK 2
Multiple Fractures Initiation At Perforations Multiple Perforations Provide Multiple Entry Points For Fracture Initiation Five Separate Fractures Are Visible In These Fractures Initiated From Horizontal Wellbore 12 Perforations Total 6 Top & Bottom I would have modeled/predicted a single frac with much higher conductivity than 5 narrow fracs added together [This actually is a bad outcome!] Multiple Strands in a These fractures are narrow, you are looking Propped Fracture at an angle to the exposed frac face (Vertical Well) NEVADA TEST SITE HYDRAULIC FRACTURE MINEBACK 3
Multiple Strands in a Propped Fracture (Vertical Well) Mesaverde 7 MWX test, SPE 22876 71 ft TVD [216m] 32 Fracture Strands Over 4 Ft Interval HPG gel and fine (pulverized) sand residue glued some core together (6-7 elapsed years) Gel residue Physical coated evidence every of surface fractures nearly always A second fractured complex zone with 8 vertical fractures in 3 ft interval observed 6 feet away (horizontally) Fracture Complexity Due To Joints Physical evidence of fractures nearly always complex NEVADA TEST SITE HYDRAULIC FRACTURE MINEBACK 4
Laminated on every scale? Figure 2 On every scale, formations may have laminations that hinder vertical permeability and fracture penetration. Shown are thin laminations in the Middle Bakken [LeFever 25], layering in the Woodford [outcrop photo courtesy of Halliburton], and large scale laminations in the Niobrara [outcrop and seismic images courtesy of Noble] 9 SPE 146376 (pending publication) Rational Expectations? Some reservoirs pose challenges to effectively breach and prop through all laminations Failure to breach all lamina? Will I lose this connection due to crushing of proppant in horizontal step? Narrower aperture plus significantly higher stress in horizontal steps? Woodford Shale Outcrop Our understanding of frac barriers and k v should influence everything from lateral depth to frac fluid type, to implementation 5
Fractures Intersecting Stacked Laterals Bakken Three Forks Inability to create an effective, durable fracture 3 feet tall?! Drill redundant well in each interval since frac has inadequate vertical penetration/conductivity?! Lateral separation 25 feet at toe/heel, crossing in middle 11 23 ft thick Lower Bakken Shale Frac ed Three Forks well ~1MM lb proppant in 1 stages 1 yr later drilled overlying well in Middle Bakken; K v <.,,1D (<.1 µd) k v /k h ~.25 even after fracing! Modified from Archie Taylor SPE ATW Aug 4 21 Uniform Packing Arrangement? Pinch out, proppant pillars, irregular distribution? 12 Is this ribbon laterally extensive and continuous for hundreds of meters as we model? 6
With what certainty can we explain this production? 2 18 Actual Production Data 2 18 Stage Production (mcfd) 16 14 12 1 8 6 4 16 14 12 1 8 6 4 Cumulative Production (MMscf) 2 2 1 2 3 4 5 6 Production Days 13 SPE 16151 Fig 13 Production can be matched with a variety of fracture and reservoir parameters Nice match to measured microseismic, eh? 2 Actual production data 2 Stage Production (mcfd) 18 16 14 12 1 8 6 4 Long Frac, Low Conductivity 5' Xf, 2 md-ft,.5 ud perm, 23 Acres 4:1 aspect ratio 18 16 14 12 1 8 6 4 Cumulative Production (MMscf) 2 2 1 2 3 4 5 6 Production Days 14 SPE 16151 Fig 13 Production can be matched with a variety of fracture and reservoir parameters 7
Is this more accurate? Tied to core perm Stage Production (mcfd) 2 18 16 14 12 1 8 6 4 2 Actual production data Long Frac, Low Conductivity 5' Xf, 2 md-ft,.5 ud perm, 23 Acres 4:1 aspect ratio Medium Frac, Low Conductivity 1' Xf, 2 md-ft, 5 ud perm, 11 Acres 4:1 aspect ratio 2 18 16 14 12 1 8 6 4 2 Cumulative Production (MMscf) 1 2 3 4 5 6 Production Days 15 SPE 16151 Fig 13 Production can be matched with a variety of fracture and reservoir parameters Can I reinforce my misconceptions? 2 Actual production data 2 Stage Production (mcfd) 18 16 14 12 1 8 6 4 2 Long Frac, Low Conductivity 5' Xf, 2 md-ft,.5 ud perm, 23 Acres 4:1 aspect ratio Medium Frac, Low Conductivity 1' Xf, 2 md-ft, 5 ud perm, 11 Acres 4:1 aspect ratio Short Frac, High Conductivity, Reservoir Boundaries 5' Xf, 6 md-ft, 1 ud perm, 7 Acres 4:1 aspect ratio Even if I know it is a simple planar frac, I cannot prove whether it was inadequate reservoir quality, or inadequate completion with a single well History matching of production is surprisingly non-unique. Too many knobs available to tweak We can always blame it on the geology 18 16 14 12 1 8 6 4 2 Cumulative Production (MMscf) 1 2 3 4 5 6 Production Days 16 SPE 16151 Fig 13 Production can be matched with a variety of fracture and reservoir parameters 8
Removing the Uncertainty If we require a production match of two different frac designs, we remove many degrees of freedom lock in all the reservoir knobs! The difference in production must be explained with the difference in the FRAC descriptions, not the reservoir description, right? 17 We are 99.9% certain the Pinedale Anticline was constrained by proppant quality Production Rate 1 days post-frac (mcfd) 9 8 7 6 5 4 3 2 1 Effect of Proppant Selection upon Production Averages based on 95 stages ISP- BS and 54 stages ISP 2/4 Versaprop ISP-BS CarboProp ISP 2/4 LL3 LL2 LL1 MV5 MV4 MV3 MV2 MV1 SPE 16151 and 18991 Reservoir Sub-Interval (Lower Lance and Mesa Verde) MV Average 9
Evidence to convince your engineer SPE 119143 2 fields in which alternative frac designs were compared SPE 13433 143 fields in which refrac results were published Compelling evidence that formations are often more permeable than we think and fracs are not optimized 19 Can we learn from refracs? Gas Condensate wells in DJ Basin up to 5 restimulations Rangely oilfield 17 refracs 1947-1989. Most wells have received 3-4 refracs yet remain viable restimulation candidates. Pembina oilfield Conductivity was understood to degrade over time, with production falling to unstimulated rates in 6-7 years. Pagano, 26 1
Does Conductivity Degrade? McDaniel, SPE 1567 All published lab data show proppants continue to crush, compact, rearrange over time and lose conductivity. SPE 12616, 14133, 1567, 11451,128612, 13433, 136757, Hahn, Drilling Vol 47, No 6, April 1986 Some proppants are more durable than others. But none are constant Why don t engineers recognize this? Production from Fracture (bfpd) 4 35 3 25 2 15 1 5 Increase Conductivity in Refracs? Dozens of examples in literature First Refrac Incremental Oil exceeds 65, barrels Incremental Oil Exceeds 1,, barrels Second Refrac May-84 May-86 May-88 May-9 May-92 May-94 May-96 May-98 May- Date Original Fracture (2/4 Sand) Phase I refrac (2/4 Sand) Phase III refrac (16/2 LWC) Pospisil, 1992 6 years later, 2 md oil. 12 ) y a 1 /d s e n 8 ( to te 6 a R n 4 tio c u d 2 r o P Initial Frac Refrac Well A Well B Well C Well D Well E Dedurin, 28, Volga-Urals oil Gas Rate, MCFD 35 3 25 2 15 1 Initial Frac in 1989: 48, lb 4/7 sand + 466, lb 12/2 sand Gas Water Stabilized Rate (MSCFD) 25 2 15 1 5 May 1995 Frac: 5, lb 1 mesh + 24, lb 2/4 Sand Pre Frac 1, gal 3% acid + 1, lb glass beads May 1999 Frac: 3, lb 2/4 LWC 8, gal + 1, lb 2/4 sand Ennis, 1989 sequential refracs, tight gas 5 45 4 35 3 25 2 15 75, gal + 12, lb 2/4 ISP Water Rate, BWPD 5 Jan-9 Jan-91 Jan-92 Jan-93 Jan-94 Jan-95 Jan-96 Jan-97 Jan-98 Jan-99 Jan- Jan-1 1 5 Shaefer, 26 17 years later, 22 tight gas Vincent, 22 9 years later, CBM 11
1) Many rocks are laminated Please Teach Us Terrible vertical perm and resistance to frac penetration In tight reservoirs, you ve got to frac it if you want to drain it Conventionally implemented fracs are KNOWN to fail to drain the entire productive section Ramifications: Much better height containment than anticipated We aren t even draining the entire hydrocarbon-bearing interval In horizontal wells, landing depth matters! Many refrac opportunities to target bypassed pay [SPE 13433, 136757] Please Teach Us 2) Fracs can provide tremendous reservoir contact, but have a tenuous connection with the wellbore Help us visualize a frac 2 feet long,.1 inch wide, 5 feet high Perhaps 1 million to 1 million ft 2 of reservoir contact achieved with multiple transverse fracs [upcoming SPE 146376 and SPE DA series to discuss] Transverse fracs provide only a tiny intersection with the wellbore Ramifications: Hydrocarbons move at least a million times faster in a propped frac than in the reservoir rock [SPE 11821, 128612] You should evaluate wider fracs with better proppant near-wellbore Be concerned about overflushing gelled stages! 12
3) Fracs are not simple, vertical planes within homogenous reservoirs. Production models typically mislead us A homogenous reservoir model incorrectly predicts all mobile hydrocarbons will find a perforation (regardless of frac) We touch more rock than expected, but are challenged to place a frac with adequate conductivity and continuity There is more stress applied to proppant in horizontal steps than in vertical sections of the fracture Ramifications: Please Teach Us Frac designs are not optimized [SPE 119143] We should not anticipate hydraulic continuity after pumping low proppant concentrations in viscous fluids When fracs succeed in placing a durable conduit into previously undrained lamina, fantastic increases in production are possible Please Teach Us 4) Fracs are not as durable as previously thought In most reservoirs, unpropped fracs heal [SPE 115766] Even in reservoirs in which unpropped fracs work, propped fracs often provide superior production [SPE 13433] All the lab data indicate that proppants continue to crush, compact, rearrange over time [SPE 136757] Ramifications: We often mistakenly interpret frac degradation as poor reservoir quality, or very short frac lengths Might reconsider/avoid overflushing proppant in some reservoirs Might evaluate more durable proppants Many refrac opportunities 13
Please Teach Us 5) Engineers do not know what we think we know Interpretations are non-unique [SPE 16151] Disappointing production has frequently been blamed on poor rock quality, when the actual cause is later proven to be inadequate frac performance Carefully designed field trials can eliminate uniqueness problem and distinguish between reservoir and fracture performance [SPE 18991, 119143] Ramifications: Don t walk away from a prospect if the failure was in frac design or implementation There are tremendous opportunities to improve production from most reservoirs Five Things I Wish a Geologist Had Taught Me Confessions of a Frac Engineer Mike Vincent mike@fracwell.com Fracwell LLC Microseismic image: SPE 119636 14