Application of the Connection Machine to flow and transport problems in three-dimensional heterogeneous aquifers

Abstract This paper explores the possibility of using the Connection Machine CM-2 for three-dimensional groundwater flow and transport problems in randomly heterogeneous aquifers. The steady-state flow problem is discretized using traditional finite differences and a conjugate gradient method is used to solve the resulting algebraic system. The transport problem is solved using the random walk particle tracking method (RWPTM); both reactive and non-reactive transport is examined. Although both methods appear to be well suited for the Connection Machine (CM), the flow problem, as has been demonstrated by other authors, is a clear winner on the CM, achieving speed-ups of close to 3 over the Cray Y-MP (single processor) for some problems. However, the RWPTM does not exhibit good performance because the heterogenity of the problem creates intense interprocessor communication when velocities are assigned to each particle. Because the particle distribution is non-uniform and the velocity field is large, this interprocessor communication is arbitrary. Simple vector valued subscripting was found to perform better than the several more complicated alternatives investigated for this communication step. Although the Cray Y-MP performance for the particle tracking problem was significantly better than the CM, it is nowhere near ideal because perfect vectorization is not achieved due to the large number of conditional statements (or IF statements) in the reactive part of the code.

[1]  David A. Yuen,et al.  Simulation of particle mixing by turbulent convective flows on the Connection Machine , 1992, Proceedings Supercomputing '92.

[2]  Nahil Sobh Preconditioned conjugate gradient and finite element methods for massively data-parallel architectures , 1991 .

[3]  G. Pinder,et al.  Computational Methods in Subsurface Flow , 1983 .

[4]  David E. Dougherty,et al.  Hydrologic Applications of the Connection Machine CM‐2 , 1991 .

[5]  Harvey Gould,et al.  Molecular dynamics simulation of liquids on the Connection Machine , 1991 .

[6]  Mary F. Wheeler,et al.  A modified method of characteristics technique and mixed finite elements method for simulation of groundwater solute transport , 1989 .

[7]  Ad Emmen A survey of vector and parallel processors for numerical applications , 1990 .

[8]  S. Ashby,et al.  A numerical investigation of the conjugate gradient method as applied to three‐dimensional groundwater flow problems in randomly heterogeneous porous media , 1989 .

[9]  D. E. Dougherty,et al.  Particle-grid methods for reacting flows in porous media with application to Fisher's equation , 1992 .

[10]  W. Daniel Hillis,et al.  The connection machine , 1985 .

[11]  Reuven Y. Rubinstein,et al.  Simulation and the Monte Carlo Method , 1981 .

[12]  D. E. Dougherty,et al.  Implementing the particle method on the iPSC/2 , 1990 .

[13]  R. Ababou,et al.  Implementation of the three‐dimensional turning bands random field generator , 1989 .

[14]  Gene H. Golub,et al.  Matrix computations , 1983 .

[15]  Andrew F. B. Tompson,et al.  Numerical simulation of solute transport in three-dimensional, randomly heterogeneous porous media , 1990 .

[16]  W. Kinzelbach,et al.  The Random Walk Method in Pollutant Transport Simulation , 1988 .

[17]  Albert J. Valocchi,et al.  Application of the random walk method to simulate the transport of kinetically adsorbing solutes , 1989 .

[18]  Bruce M. Boghosian,et al.  Computational physics on the Connection Machine , 1989 .

[19]  E. A. Sudicky,et al.  The Laplace Transform Galerkin Technique: A time‐continuous finite element theory and application to mass transport in groundwater , 1989 .

[20]  Henk A. van der Vorst,et al.  ICCG and related methods for 3D problems on vector computers , 1989 .

[21]  Henk A. van der Vorst,et al.  Iterative methods for the solution of large systems of equations on supercomputers , 1990 .

[22]  Giuseppe Gambolati,et al.  Is a simple diagonal scaling the best preconditioner for conjugate gradients on supercomputers , 1990 .