Pore-scale processes
A revolution in describing multiphase flow
Martin Blunt, Matthew Andrew, Branko Bijeljic, Sam Krevor,...
Ten-year, $70 million programme: 2008 – 2018.
To understand carbon dioxide storage in a Qatari context (carbonates). Major...
3
Geo Oct 2013
Status of Impact – Sea-level rise
4Abu Dhabi Environment Agency
2009
Abu Dhabi Environment Agency
2009
5
Motivation
Historically high oil prices, even at $40/barrel – peak oil per person in
1979 and current discoveries running ...
New tools
Multi-scale imaging – particularly ability to image the pore space of rock
and fluids at 10 nm to micron resolut...
What we can do
Original work on 3D X-ray microtomography by Flannery et al. (1987)
states in conclusion: “we believe that ...
Imperial College multi-scale imaging lab
Start with the fundamentals – understand processes experimentally at the
pore sca...
Micro-CT – Flow loop
10
Imaging and computing
Bench-top micro-CT scanners are
convenient, no time limitations and
modern systems have optics.
Sync...
Images and networks for carbonates
Estaillades Ketton Mount Gambier
Represent the pore space topologically and compute dis...
Transport – rocks and people
13
How to get to Imperial from
Heathrow airport?
Direct simulation: use a shallow
seismic ima...
Waterflooding and wettability
Complex displacement sequences, shown here for a single idealized
pore. What are the contact...
Water-wet two-phase predictions
Experimental data from Berea sandstone cores (Oak, 1990)
– No tuning of network (Øren and ...
The tyranny of scale
Typically have a million-fold variation in length scale, from 10 nm for
the smallest micro-pores to c...
Direct simulation and networks
• Cannot compute multiphase flow directly on images that can
resolve the smallest pores, an...
Back to the science - dispersion
Direct simulation on the pore-space images.
Stokes solver, streamline tracing, random mot...
Carbonate images and flow fields
5 mm
Ketton
Mt Gambier
EstailladesIndiana
ME1Guiting
Particle trajectories in the pore space
Combine analytical
streamline tracing with
a random hop to
represent diffusion.
So...
Concentration
profiles
Bentheimer
Sandstone
Bead pack Portland
Carbonate
Compare prediction of
concentration vs.
distance ...
Reaction with the solid: Dissolution regimes
22
Daccord et al.,
Chem. Eng. Sci, (1993)
Maheshwari et al.,
Chem. Eng. Sci, ...
Pore-scale dissolution experiments
Flow rate: 0.5 ml/min for 2.5 hrs [Pe ~103; Da ~10-4]
Brine composition: 1% KCl 5% NaCl...
SimulationExperimental
Model vs. experiment
Dissolution – parallel to flow direction
Ketton carbonate – chanelling Portland carbonate – compact dissolution
0.05 ml/min [Pe ~102; Da ~10-3]
Three-dimensional r...
1.3mm
0.67 mm
Small Pe regime only “face dissolution” - Whole grains are being dissolved
No significant impact in permeabi...
1.3mm
0.67 mm
Simulations: Estaillades Pe, Péclet = 50
Simulations: Estaillades Pe, Péclet = 280, fast flow
High Pe regime see more uniform dissolution, as the reactant can pene...
Trapped CO2 clusters – colour indicates size
Pentland et al., Geophysical Research Letters (2011)
How much is trapped and
...
Can study many systems – Bentheimer and Doddington
Can study many systems – Estaillades and Ketton
Can study many systems – Portland
Andrew et al.,
Geophysical Research
Letters (2011); IJGGC (2014)
Curvature, contact angle and validation
Can also use high-resolution images to
determine: curvature – capillary pressure,
...
Measurement of contact angle
Dynamic Tomography at Synchrotron Sources
35
Synchrotron Experimental
Team:
Matthew Andrew
Hannah Menke
Cat Reynolds
Kamal...
Connected pathway and ganglia flow
Scan time ≈ 20 s, Time step = 43 s,
10 PV
Interfacial curvature
Equilibrium capillary pressure change
38
Distal (non-local) snap-off
39
3D X-ray Micro-CT imaging of a rock sample
Does it matter?
40
Enhanced Oil Recovery
Carbon Storage
http://energy.gov/
Cont...
Conclusions
New tools – both experimentally and numerically allow us to
observe and model flow and transport in great deta...
Acknowledgements
Qatar Petroleum, Shell and the Qatar Science and Technology Park
under the Qatar Carbonates and Carbon St...
of 42

Pore scale dynamics and the interpretation of flow processes - Martin Blunt, Imperial College London, at UKCCSRC specialist meeting Flow and Transport for CO2 Storage, 29-30 October 2015

Pore scale dynamics and the interpretation of flow processes - Martin Blunt, Imperial College London, at UKCCSRC specialist meeting Flow and Transport for CO2 Storage, 29-30 October 2015
Published on: Mar 4, 2016
Published in: Engineering      
Source: www.slideshare.net


Transcripts - Pore scale dynamics and the interpretation of flow processes - Martin Blunt, Imperial College London, at UKCCSRC specialist meeting Flow and Transport for CO2 Storage, 29-30 October 2015

  • 1. Pore-scale processes A revolution in describing multiphase flow Martin Blunt, Matthew Andrew, Branko Bijeljic, Sam Krevor, Catriona Reynolds, Ali Raeini, Hu Dong, João P. Nunes, Kamaljit Singh and Hannah Menke Department of Earth Science and Engineering Imperial College London and iRock Technologies, Beijing
  • 2. Ten-year, $70 million programme: 2008 – 2018. To understand carbon dioxide storage in a Qatari context (carbonates). Major experimental and modelling activity. Based at Imperial College. Work all published in the public domain. Multidisciplinary (Chem. Eng. / Earth Sci. & Eng.). Three major themes: rocks, fluids and rock-fluid interaction. Four dedicated lecturers, other faculty, post-docs and PhD students (some from Qatar): involves >70 researchers.
  • 3. 3 Geo Oct 2013 Status of Impact – Sea-level rise
  • 4. 4Abu Dhabi Environment Agency 2009
  • 5. Abu Dhabi Environment Agency 2009 5
  • 6. Motivation Historically high oil prices, even at $40/barrel – peak oil per person in 1979 and current discoveries running at half global production (30 billion stb/year). Need to produce more of the oil in existing fields. Exploitation of unconventional oil and gas. Wise use of groundwater. Global-scale CO2 storage. All involve understanding of flow of fluids in porous rocks.
  • 7. New tools Multi-scale imaging – particularly ability to image the pore space of rock and fluids at 10 nm to micron resolution. Public-domain availability of good-quality software for scientific computing – changes the way we develop computational models. What is digital rock analysis? A physically-based model for flow, based on pore-scale displacement. A nm – cm model (6 orders of magnitude in scale). A necessary complement and input to a field-scale geological/reservoir model (cm – km, or another 6 orders of magnitude).
  • 8. What we can do Original work on 3D X-ray microtomography by Flannery et al. (1987) states in conclusion: “we believe that it will be possible to study contained systems under conditions of temperature, pressure, and environment representative of process conditions.” Can now! Will discuss imaging and flow simulation: transport, reaction and multiphase flow. Flow Transport Reaction Structure
  • 9. Imperial College multi-scale imaging lab Start with the fundamentals – understand processes experimentally at the pore scale. Micron-to-metre imaging with in situ displacement at reservoir conditions.
  • 10. Micro-CT – Flow loop 10
  • 11. Imaging and computing Bench-top micro-CT scanners are convenient, no time limitations and modern systems have optics. Synchrotron sources. Bright, mono- chromatic and fast. Computationally, not interested in GPU, parallel, but better algorithms. Availability of excellent public- domain solvers: algebraic multigrid, OpenFoam Navier-Stokes solver. Fluid mechanics: unstructured adaptive grids. Blunt et al., Adv. Water Res. 2013
  • 12. Images and networks for carbonates Estaillades Ketton Mount Gambier Represent the pore space topologically and compute displacement semi- analytically through the network. Also accommodate micro-porosity.
  • 13. Transport – rocks and people 13 How to get to Imperial from Heathrow airport? Direct simulation: use a shallow seismic image of the subsurface of London?! London Underground map (the macro-pores) plus a local map (the micro-pores)
  • 14. Waterflooding and wettability Complex displacement sequences, shown here for a single idealized pore. What are the contact angles? Can now measure them in situ. Altered wettability surfaces after primary drainage: mixed-wettability. Relative permeability is governed by the interplay of displacement, structure and wettability, which can vary across the field
  • 15. Water-wet two-phase predictions Experimental data from Berea sandstone cores (Oak, 1990) – No tuning of network (Øren and Bakke, 2003) necessary – The fluids are water and oil – Water-wet data – predictions made with θa = [50°, 80°] 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 Water Saturation RelativePermeability Primary drainage 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 Water Saturation RelativePermeability Experimental Predicted p p rp p P Kk q   Secondary waterflooding Valvatne and Blunt, Water Resources Research (2004)
  • 16. The tyranny of scale Typically have a million-fold variation in length scale, from 10 nm for the smallest micro-pores to cms for whole cores. Need to upscale. No one method can capture complex displacement processes over this range of scales. Whole core – 1 cm Macro pore - 1 mm Micro pore - 10 m
  • 17. Direct simulation and networks • Cannot compute multiphase flow directly on images that can resolve the smallest pores, and processes within them. • Direct simulation would require of order 1021 grid blocks. No, not even the fastest in-the-future computer will ever be able to do this. • Need to combine methods: direct simulation for pore-scale events, “simple” images; network modelling to upscale behaviour and capture the correct displacement sequence.
  • 18. Back to the science - dispersion Direct simulation on the pore-space images. Stokes solver, streamline tracing, random motion for diffusion. Sandpack Sandstone (Bentheimer) Carbonate (Portland)
  • 19. Carbonate images and flow fields 5 mm Ketton Mt Gambier EstailladesIndiana ME1Guiting
  • 20. Particle trajectories in the pore space Combine analytical streamline tracing with a random hop to represent diffusion. Solute particles travel longer distances for larger Pe number. 𝑃𝑒 = 𝑣𝐿 𝐷 𝑚 = advection diffusion v = velocity; L = grain/pore size; Dm = molecular diffusion coefficient. Include reaction by allowing particles within a diffusion distance to react, including solid. Probability of reaction relates to reaction rate.
  • 21. Concentration profiles Bentheimer Sandstone Bead pack Portland Carbonate Compare prediction of concentration vs. distance for different times and rock types against NMR experiments. Can make first principles predictions once the pore geometry is imaged. Bijeljic et al. PRL (2011); PRE (2013); WRR (2013). Time
  • 22. Reaction with the solid: Dissolution regimes 22 Daccord et al., Chem. Eng. Sci, (1993) Maheshwari et al., Chem. Eng. Sci, (2013) compact uniform wormhole 𝑃𝑒 = 𝑣𝐿 𝐷 𝑚 = advection diffusion Da = reaction advection Compare pore-scale experiments and models. In the models if a particle hits solid in the diffusive step, dissolve solid after a given number of hits: determines reaction rate.
  • 23. Pore-scale dissolution experiments Flow rate: 0.5 ml/min for 2.5 hrs [Pe ~103; Da ~10-4] Brine composition: 1% KCl 5% NaCl brine saturated with CO2 at 10 MPa and 50oC [pH=3.1] Ketton carbonate - homogeneous Portland carbonate - heterogeneous Menke et al., EST (2015)
  • 24. SimulationExperimental Model vs. experiment Dissolution – parallel to flow direction
  • 25. Ketton carbonate – chanelling Portland carbonate – compact dissolution 0.05 ml/min [Pe ~102; Da ~10-3] Three-dimensional results (low flow rate)
  • 26. 1.3mm 0.67 mm Small Pe regime only “face dissolution” - Whole grains are being dissolved No significant impact in permeability. Simulations: Estaillades Pe, Péclet = 1, slow flow
  • 27. 1.3mm 0.67 mm Simulations: Estaillades Pe, Péclet = 50
  • 28. Simulations: Estaillades Pe, Péclet = 280, fast flow High Pe regime see more uniform dissolution, as the reactant can penetrate the rock before reacting. As seen experimentally.
  • 29. Trapped CO2 clusters – colour indicates size Pentland et al., Geophysical Research Letters (2011) How much is trapped and how much can be stored? Results in sandstones (Doddington, Bentheimer and Berea). After drainage After waterflooding 20 mm 0.0 0.2 0.4 0.6 0.0 0.5 1.0 Snwr Snwi C. Pentland (2011) @ 70 C Rehab results @ 70C
  • 30. Can study many systems – Bentheimer and Doddington
  • 31. Can study many systems – Estaillades and Ketton
  • 32. Can study many systems – Portland Andrew et al., Geophysical Research Letters (2011); IJGGC (2014)
  • 33. Curvature, contact angle and validation Can also use high-resolution images to determine: curvature – capillary pressure, and local pressure for each ganglion; and surface contacts to determine contact angles. Andrew et al., AWR (2014) Residual oil in a mixed-wet system Direct simulation (volume of fluid) of trapping
  • 34. Measurement of contact angle
  • 35. Dynamic Tomography at Synchrotron Sources 35 Synchrotron Experimental Team: Matthew Andrew Hannah Menke Cat Reynolds Kamal Singh Branko Bijeljic Martin Blunt
  • 36. Connected pathway and ganglia flow Scan time ≈ 20 s, Time step = 43 s, 10 PV
  • 37. Interfacial curvature
  • 38. Equilibrium capillary pressure change 38
  • 39. Distal (non-local) snap-off 39
  • 40. 3D X-ray Micro-CT imaging of a rock sample Does it matter? 40 Enhanced Oil Recovery Carbon Storage http://energy.gov/ Contaminant Transport http://www.euwfd.com/html/groundwater.html Shale oil and gas
  • 41. Conclusions New tools – both experimentally and numerically allow us to observe and model flow and transport in great detail from the pore scale upwards. Huge practical challenges also drive the science. We are on the cusp of a revolution.
  • 42. Acknowledgements Qatar Petroleum, Shell and the Qatar Science and Technology Park under the Qatar Carbonates and Carbon Storage Research Centre

Related Documents