14-17 May 2018
New Orleans
US/Central timezone

A parallel boundary element method for subsurface flow problems in three-dimensional fracture networks

17 May 2018, 11:41
15m
New Orleans

New Orleans

Oral 20 Minutes MS 2.03: Challenges in flow and transport simulations in poro-fractured media: numerical methods and modeling Parallel 10-C

Speaker

Dr Yin Feng (University of Louisiana at Lafayette)

Description

Modeling subsurface flow in three-dimensional (3D) discrete fracture networks (DFN) is of interest of many engineering problems, such as CO2 sequestration, natural gas production and geothermal energy extractions. However, the recent grid-based models describing flow behaviors in 3D DFN are still suffering from the complex gridding issue and high computational burden. In this work, a meshless approach is presented for the steady-state flow in DFNs with arbitrary geometries and penetrated wellbores.

A DFN consists of planar polygonal fractures with random orientation, size and flow transmissivity. Based on a parallel BEM approach, the flow solution of a large DFN can be decomposed into a series of sub-domains by the domain decomposition method (DDM). The problem on each sub-domain can be solved independently from each other, and iterations are performed until the continuities of pressure and flux balance are satisfied at fracture intersections. Additionally, BEM solution of each planar fracture only requires the discretization along fracture edges, and penetrated wellbores are simulated by only adding point source nodes instead of the local grid refinement as in conventional grid-based methods. Thus the presented parallel BEM approach significantly reduces the overall storage and computational burden.

The presented method is verified against a commercial finite element method (FEM) simulator on several synthetic examples from simple to complex fracture network geometries. The resultant pressure field shows a good agreement with the fine grid FEM models. The method shows nearly a linear CPU-time scaling dependency to the number of fractures. The performance of the presented method is also investigated in terms of convergence properties and accuracy with respect to several fracture parameters, such as the number of fractures, fracture flow transmissivity, penetrated wellbores and mesh density.

The method has potentials to provide an efficient, accurate and parallel framework to account for large and complex DFNs with numerous penetrated wellbores.

Source code of PyDFN3D will be provided .

References

[1] Kamiya, N., Iwase, H. and Kita, E., 1996. Parallel implementation of boundary element method with domain decomposition. Engineering Analysis with Boundary Elements, 18(3), 209-216. doi: 10.1016/S0955-7997(96)00050-

[2] Berrone, S., Pieraccini, S., Scialo, S. and Vicini, F., 2015. A parallel solver for large scale DFN flow simulations. SIAM Journal on Scientific Computing, 37(3), pp.C285-C306. doi: https://doi.org/10.1137/140984014

[3] Tao, S., Cheng, J. and Mosallaei, H., 2016. An integral equation based domain decomposition method for solving large-size substrate-supported aperiodic plasmonic array platforms. MRS Communications, 6(2), pp.105-115. doi: https://doi.org/10.1557/mrc.2016.11

[4] Lenti, V. and Fidelibus, C., 2003. A BEM solution of steady-state flow problems in discrete fracture networks with minimization of core storage. Computers & geosciences, 29(9), pp.1183-1190. doi: https://doi.org/10.1016/S0098-3004(03) 00140-7.

[5] Berrone S, Fidelibus C, Pieraccini S, Scialo S., 2014. Simulation of the steady-state flow in discrete fracture networks with non-conforming meshes and extended finite elements. Rock mechanics and rock engineering. 47(6):2171-82. doi: 10.1007/s00603-013-0513-5

[6] Hyman, J.D., Karra, S., Makedonska, N., Gable, C.W., Painter, S.L. and Viswanathan, H.S., 2015. dfnWorks: A discrete fracture network framework for modeling subsurface flow and transport. Computers \&Geosciences, 84, 10-19. doi: 10.1016/j.cageo.2015.08.001

Acceptance of Terms and Conditions Click here to agree

Primary authors

Mr Bin Wang (Louisiana State University, University of Louisiana at Lafayette) Dr Yin Feng (University of Louisiana at Lafayette) Stefano Berrone (Politecnico di Torino) Dr Stefano Scialo (Politecnico di Torino) Dr Sandra Pieraccini (Politecnico di Torino) Dr Corrado Fidelibus (Politecnico di Bari)

Presentation Materials

There are no materials yet.