Field Examples of Activating Shale as Well Barrier by Exposing it to a Rapid Pressure Drop SPE Stavanger: Efficient P&A with Formation as Barrier Tron G. Kristiansen Operations Geology & Rock Mechanics Manager 20 March 2017

Outline Shale creep Other ductile failure modes Methods to activate shale barriers Simulation and laboratory observations Field cases of activating and verifying shale barriers in a new well Summary

Creeping Shales Vs. Activating the Shale ? Primary and secondary creep is quite normally observed Depending on the creep rate the shale may or may not form a barrier around the casing Tertiary creep is observed in rocks, but then typically when the rock load is close to the non-elastic limit If we can help the rock to achieve tertiary creep, we would have a good barrier in all cases By activating the shale barrier in an engineered operation we think we can make a more consistent barrier, a more engineered barrier

Rock Yield Surface in 3D Brittle shear failure Ductile pore collapse failure mode (wire mesh) Brittle tensile failure If one tests the same rock with all possible combinations of stresses, the elastic area of the rock behavior will be enclosed in a 3D surface named the yield surface of the rock. Yield surface is in the principal stress space. It is often represented by a simpler 2D representation. Yield surface can be thought of as the PVT envelope of the rock (an analogue). Remember: whether the rock fails in a brittle mode (tesnile failure or shear failure) or a ductile mode (pore collapse) is determined by the stress path the rock specimen is following and where it hits the yield surface. The specific surface model in the illustration was co-developed by Arlo Fossum during his time at Sandia. If the same rock is tested with all possible combinations of stresses, the elastic area of the rock behavior will be enclosed in a 3D surface named the yield surface of the rock. Yield surface: Is in the principal stress space. Is often represented by a simpler 2D representation (examples are Mohr Coloumb diagram, p’-q diagram) Can be thought of as the “phase behavior” of the rock (an analogue) For weak shales we have done experiments indicating that fractures from brittle shear failure will close under the right conditions 4 4

Methods to Activate Barriers Pressure induced Temperature induced Combined mechanisms Focus in this talk We have the best experience so far with using a rapid pressure drop in the annulus Our experience with chemical methods may take some more time than we typically need with the pressure drop in annulus method We think temperature may be a good method in many cases, we are working on this in AkerBP (eith Sintef and Interwell)

Activating Shale Barriers with Rapid Pressure Drop Hole deformation Rapid pressure drop After the casing was removed from the hole Shale deformation response, shale is closing annulus around casing* *Experiment performed by SINTEF in Shalle Barrier JIP sponsored by Aker BP, BP, CoP, Shell, Statoil and Total

Numerical Simulation of Pressure Drop in Annulus Shear Stress Mean Stress Contact Force on Casing

Stress Path & Favourable Rock Strength Properties This plot is showing the stresspath of an element at the fringe of the no shear stress zone The point is at the boundary of no shear stress Points outside this region follows a distinctly different deformation mode to those within the no shear stress zone Material within this no shear stress zone could be regarded as liquefied as the material temporarily can not carry any shear stress Materials with low strength and low internal friction angle are the most prone to show this behavior

G-9 Well: Activation and Verification of Shale Barrier During New Well Construction Well IS currently Being drilled!! Drilling sequence with drilling liner (minor changes since 1993) Equipment and sequence for barrier verification PS OHAP PC Openhole annulus packer (OHAP) Wireless pressure sensors (PS) Port collars (PC) Drilling and geo-stopping 5-8 m TVD over depleted reservoir Pick up drilling liner assembly and drill into depleted reservoir Goes on losses and the pressure drop rapidly in the annulus Stop losses on inside with a ball/plug, losses in annulus most often cure themselves Set liner top packer

G-9 Results Operationallly Went on increasing losses as we pentrated reservoir as predicted Pressure drop not as high as planned for (2000 psi +), but believed to be sufficient (around 1500 psi) The wireless gauge system did not work Lost the ability to use pressure points along the wellbore and their pressure history with time to better understand the process A large volume (20 barrels +) XLOT (more than 10 times the volume we typically use) was pumped in 2nd cycle after a breakdown test Two linear pumpcycles, cycle two slightly stiffer Cycle one stopped prior to breakdown and fracture proppagation Cycle two breakdown ~15.8, fracture proppagation 15.76 - 15.9 ppg Closure pressure ~ 15.72 ppg

Very good barriers identified on bond logs at depth of deepest port collar where the XLOT was performed

G-20 Operationally Well predicted to have higher reservoir pressure than G-9, so risk of shale not being activated before logging No significant loss rate increase observed when drilling into the reservoir It turned out that the pressure in the reservoir was much higher than predicted (compartment in reservoir) Fracture and closure pressure in reservoir measured to be high enough to not result in losses as we drillled into reservoir, i.e. ~0 psi pressure drop Run a bond log and found very limited shale bond, nothing circumferential that could act as a well barrier Opened deepest port collar and attempted a two cycle XLOT, but large differences in the two cycles as well as low leak off in second cycle much lower than first cycle and not indicating bond Good argument for finalizing the construction of the drawdown tool we have started on ......... Squeezed a cement job through the port collar as planned for contingency, good results based on bond log afterwards We also re-loged this section ~3 weeks later before running reservoir liner to check effect of creep and the port collar cement job

Results from G-20 Cycle 1 peak pressure: 535 psi No barrier based on XLOT No barrier based on bond log

Re-logged 19 days later Changes in bond log indicated over the 19 day period Most significantly the grey rectangle which is from the cement job through the port collar This was a surprise to many Also more bonding in Lista is developing Som increased bonding in Horda 5 and 6 as well No verified barrier from log 19 days later

Summary Understanding of the activation of shale barriers using pressure drop in annulus is maturing Field trials indicate that we can predict how barriers can be activated in the field based on laboratory experiments and numerical simulation One well with sufficient pressure drop formed a barrier and was verified with XLOT and bond logs One well with no or very limited pressure drop did not form a barrier and failed XLOT and bond log verification A pressure activation tool need to be finalized for cases with no natural pressure drop Shale cuttings from weak shale sections seems to be a good barrier material under confinement and is easily outperforming permeability of standard cements based on laboratory experiments A combination of scaled experiments, numerical modeling and instrumented field trials seems to be an effective way of qualifying new barrier materials for P&A P&A technology do also have applications in new well design as illustrated in these cases Temperature activation needs more work to understand limitations (and to potentially design a temperature activation tool)