Modeling of the Snorre deep-divergence pilot NIOR-workshop – Stavanger - April 25th, 2016 Vegard R. Stenerud and Øivind Fevang

Outline

•  Snorre water diversion pilot

•  Mechanisms of Sodium Silicate system

•  Silicate modeling and History matching

•  Challenges and potential improvements

2

In-depth water diversion – Sodium silicate

Goal: Establish flow restrictions in flooded areas to improve lateral and vertical reservoir sweep

•  Silicate

•  Low temp – Flows like water

•  Reservoir temp - Gelling

•  Dynamic gelling:

•  2008-2009: Lab: > 25 bar

•  2011: Single well test: 80-120 bar

•  2013: Inter-well pilot

•  Intolerant to seawater (need preflush)

−  Accelerated gelling

−  Precipitation

Pilot area

Thief zone challenge

Initial injection water front speed: approx. 6 m/d (2002-2003)

Water tracer injection (2008) Tracer front speed: 9-11 m/d

4

2 km

Subsea water injector

Oil producer on Snorre A

Early and strong water break through

Modelling: Simulating and matching observed data

5

1E-3 0.01 0.1 1 10 100Time [hr]

1

10

Pre

ssur

e [b

ar]

E -­‐4H  paper  nov15.ks3  -­‐  Mar12  (ref)E -­‐4H  paper  nov15.ks3  -­‐  Mar14E -­‐4H  paper  nov15.ks3  -­‐  Vegard  Mar12E -­‐4H  paper  nov15.ks3  -­‐  Vegard  Mar14

Change in water-cut

Reservoir transmissibility : T=q/Δp

Black: Obs Green: Sim (silicate and HM) Red: Sim (no silicate)

Green: Observed Mar 2012 Orange: Observed Mar 2014 Purple: Simulated Mar 2012 Black: Simulated Mar 2014

Falloff-data

Tracer-data – Post-flush tracer

Silicate mechanisms Gelling:

Gelation time model from Stavland et al. (2011):

The pilot designed such that:

𝑡↓𝑔 =𝐴∙ 𝑒↑𝛼[𝑆𝑖] ∙ 𝑒↑𝛽[𝐻𝐶𝑙] ∙𝑒↑𝛾[ 𝐶𝑎↑2+ ] ∙ 𝑒↑𝐸𝑎⁄𝑅𝑇  

Dependent on: •  Silicate concentration •  HCl concentration (pH) •  Divalent ions (Ca2+, Mg2+) •  Temperature

Precipitation:

Precipitation of Mg(OH)2 from seawater in contact with silicate (high pH).

𝑡↓𝑔 ≈𝐵∙ 𝑒↑𝛼[𝑆𝑖] ∙ 𝑒↑𝐸𝑎⁄𝑅𝑇  

Pre and post flush: Desalinated water with KCl for ion exchange

During the operation: Desalinated water with HCl for pH stabilization

Temperature gradient

Simulation model

7

Simulated temperature

Simulated silicate slug

Simulated salt concentration

Permeability: thief zone scenario

Modeling of sodium silicate

•  History matching approach

−  For history matching

−  Restriction included by multiplier boxes in time

•  Silicate modeling approach – for prediction

−  Gelling: Temperature and silicate concentration dependent

•  Neglecting pH and salt concentration due to pilot design

−  Precipitation: salt and silicate concentration dependent

2015-12-03

8 Classification: Internal

Using BASRA for assisted history matching: Statoil internal software – SPE113390

Restrictions summary - Ensemble

10-0.3 16d

10-1.3

14d

10-1.3

39d

10-3.8

99d

10-2.3

172d

10-0.3

272d

10-1.9

228d

10-0.2

498d

On average introduced late, but it does not really matter when weak restrictions are introduced.

Statistical distributions

Most representative (average) day and (average) strength based on ensemble

Temperature Security

Classification:

Restricted - Status: Draft

Sil. cons. 100d Sil. cons. 165d Sil. cons. 228d

Cross flow is one possibility

P2

P1

Salt (divalent cations) may back/cross flow into the thief zone during shut ins. That may give accelerated gelling for the silicate injected right after the shut-in. (Could be prevented/mitigated for future operations) Isolated good sands

P: Pressure ; P2 > P1 (Some gravity, friction,…)

Reservoir transmissibility

06/13 07/13 08/13 09/13 10/13 11/13 12/130

5

10

15

20

25

30

35

WTRANH WTRANS

WTRANH E-4HW SEP15_CTZ_FLT25 WTRANS E-4HW ROUND_12_11 WTRANS E-4HW SEP15_CTZ_FLT25

E-4HW shut in

Drop in transmissibility after shut in period

Pre-flush Siliscate injection Post-flush

E-4HW and P-15P shut in

Temperature triggered gelling (mainly)

06/13 07/13 08/13 09/13 10/13 11/13 12/13 01/14 02/140

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

Btc

nfin

2

BTCNFIN2 SEP15_MTZ_TR

BTCNFIN2 44 19 34 BTCNFIN2 44 20 34 BTCNFIN2 44 21 34BTCNFIN2 44 22 34

Simulated salt concentrations close to the injector

Concetration in wellblock

Concetration 1 block away from well

Concetration 2 blocks away from well

Concetration 3 blocks away from well

1

2

3

Thief zone layer

Shut in of the injector in August 2013

Predictive modeling

Gelling and precipitation parameters:

−  Silicate concentration triggering gelling

−  Temperature triggering gelling

−  Permeability reduction factors for gelling

−  Silicate concentration triggering precipitation

−  Salt concentration triggering precipitation

−  Permeability reduction factors for precipitation

Adsorption/loss parameter

•  Amount of adsorption of silicate (depends on silicate concentration)

13

Security Classification: Restricted - Status: Draft

7 generic silicate modeling parameters

Goal:

•  Calibrate by the pilot

•  Use for evaluating new field

candidates

Effect of cross-flow had to be modeled by multiplier boxes •  May be avoided for future operations

Calibrated silicate modeling parameters

14

Fits well with lab experiments - SPE169727 fig. 7

BHP - Producer

Ensemble of matching simulations (blue curves). The ensemble is also matching the other pilot responses.

Histogram for calibrated silicate modeling parameters – based on the matching ensemble. In this example - the logarithm of the permeability reduction factor for precipitation.

Blue: Matching simulations Black: Observed pressures Green: No silicate modeling

Predictive approach - Formed obstructions

15

Early cross-flow restrictions not in the figure.

Gelling and precipitation introduced by the Predictive approach by September 1st, 2015:

Red: Gelling Blue: Precipitation

Simulation: ROUND_16_21

Modelling: Estimated formed flow restrictions

16

Main flow restriction seems to have formed in this area between 100-230 days after the start of silicate injection.

Indications of early moderate flow restriction due to cross-flow of salts. Mitigating actions are possible for future operation.

Perm reduction factor ~0,1

Perm reduction factor

~0,0001

Consistency between the two modeling approaches

Summary – Pilot observations

•  Deep placement of flow restriction leading to flow diversion

−  Deep placement supported by: Transmissibility reduction, fall off-data, step-rate tests

−  Diversion supported by: Delay of water tracers, reduction in water cut

•  Reduced liquid throughput

−  Cross-flow of salts part of the reason

17

Summary – Silicate modeling •  History matching

−  Able to model location and strength of the introduced flow restrictions.

−  Able to history match pilot responses

−  Similar results using the two different approaches

•  Predictive approach

−  Calibrated silicate modelling parameters obtained

•  May be used in evaluation of new silicate-based water diversion candidates

•  Further calibration to lab experiment results and/or new field implementations

18

Challenges and improvements for the modeling

•  Modeling of the effect of salts on gelling (or preferably use a salt tolerant chemical)

−  Modeling the effect of potential cross-flow of salts

−  Determining size of pre flush

•  Better modeling of silicate consumption/loss

−  Integrated part of the silicate modeling

•  Still uncertainty in the silicate modeling parameters

•  Demonstrate good business cases for new field implementations

−  Use developed modeling techniques for finding candidates and maximize IOR volumes in an economical context

•  Low oil price and cost

19

Good ideas and contributions are welcome J

20

Thanks to: Statoil contributors: •  Kjersti Håland •  Kjetil Skrettingland •  Vilgeir Dalen •  Dag Chun Standnes •  Per Arne Slotte (now at NTNU)

The Snorre asset The Snorre partnership: •  Statoil, Petoro, ExxonMobile, Idemitsu, DEA,

Core Energy

Papers for the interwell pilot: •  Skrettingland et al. (2014) – SPE169727 •  Skrettingland et al. (2016) – SPE179602