Parallelization strategies for an implicit Newton-based reactive flow solver. (English) Zbl 1253.80022
Summary: The solution of reactive flows using fully implicit methods on distributed memory machines is investigated in detail. Three different parallel implementations of Newton’s method are described and tested on the solution of two-dimensional laminar axisymmetric coflow diffusion flames. Each implementation has different computational requirements, both in the amount of communication among the processes and in the computational overhead due to the calculation of physical quantities at the interfaces between subdomains. An effective trade-off is established between communications and calculations so that the most communication-intensive implementation results in computational speed-up only if the network is sufficiently fast.
Benchmark results are presented for a variety of chemical mechanisms, grid decomposition techniques, and hardware. Parallelization efficiencies of about 80% and speed-ups of 20–100 are reported for most test cases. The method developed here is well-suited for complex chemistry problems with very large mechanisms; in particular, the numerical solution of a laminar axisymmetric JP-8/air coflow diffusion flame with a 222-species mechanism is made possible using this approach.
Benchmark results are presented for a variety of chemical mechanisms, grid decomposition techniques, and hardware. Parallelization efficiencies of about 80% and speed-ups of 20–100 are reported for most test cases. The method developed here is well-suited for complex chemistry problems with very large mechanisms; in particular, the numerical solution of a laminar axisymmetric JP-8/air coflow diffusion flame with a 222-species mechanism is made possible using this approach.
MSC:
| 80A25 | Combustion |
| 80A32 | Chemically reacting flows |
| 76D05 | Navier-Stokes equations for incompressible viscous fluids |
| 80A20 | Heat and mass transfer, heat flow (MSC2010) |
| 80M25 | Other numerical methods (thermodynamics) (MSC2010) |
| 60H10 | Stochastic ordinary differential equations (aspects of stochastic analysis) |
| 65Y05 | Parallel numerical computation |
Keywords:
Newton’s method; reactive flows; distributed memory; parallel algorithm; computational fluid dynamicsReferences:
| [1] | DOI: 10.1016/j.proci.2008.06.218 · doi:10.1016/j.proci.2008.06.218 |
| [2] | DOI: 10.1080/00102200215080 · doi:10.1080/00102200215080 |
| [3] | DOI: 10.1021/ie049318g · doi:10.1021/ie049318g |
| [4] | DOI: 10.1016/j.proci.2004.08.046 · doi:10.1016/j.proci.2004.08.046 |
| [5] | DOI: 10.1016/j.combustflame.2008.05.020 · doi:10.1016/j.combustflame.2008.05.020 |
| [6] | DOI: 10.1115/1.1377871 · doi:10.1115/1.1377871 |
| [7] | DOI: 10.1080/00102208708947079 · doi:10.1080/00102208708947079 |
| [8] | DOI: 10.1177/109434209100500404 · doi:10.1177/109434209100500404 |
| [9] | DOI: 10.1016/j.proci.2008.06.182 · doi:10.1016/j.proci.2008.06.182 |
| [10] | DOI: 10.1006/jcph.1997.5856 · Zbl 0936.76064 · doi:10.1006/jcph.1997.5856 |
| [11] | DOI: 10.1177/109434209500900301 · doi:10.1177/109434209500900301 |
| [12] | DOI: 10.1016/0045-7930(96)00009-6 · Zbl 0900.76349 · doi:10.1016/0045-7930(96)00009-6 |
| [13] | DOI: 10.1016/j.jcp.2009.04.037 · Zbl 1172.80303 · doi:10.1016/j.jcp.2009.04.037 |
| [14] | DOI: 10.1016/j.combustflame.2006.06.007 · doi:10.1016/j.combustflame.2006.06.007 |
| [15] | DOI: 10.1080/00102208908924063 · doi:10.1080/00102208908924063 |
| [16] | DOI: 10.1016/j.combustflame.2009.01.020 · doi:10.1016/j.combustflame.2009.01.020 |
| [17] | DOI: 10.1137/0915044 · Zbl 0807.65027 · doi:10.1137/0915044 |
| [18] | DOI: 10.1006/jcph.1996.0117 · Zbl 0853.76091 · doi:10.1006/jcph.1996.0117 |
| [19] | DOI: 10.1016/S0010-2180(01)00306-6 · doi:10.1016/S0010-2180(01)00306-6 |
| [20] | DOI: 10.1016/0021-9991(82)90032-8 · Zbl 0502.76040 · doi:10.1016/0021-9991(82)90032-8 |
| [21] | Guj G., J. Comput. Phys. 106 pp 286– (1993) · Zbl 0770.76045 · doi:10.1016/S0021-9991(83)71108-3 |
| [22] | DOI: 10.1002/fld.204 · Zbl 1009.76059 · doi:10.1002/fld.204 |
| [23] | DOI: 10.1007/BF00933755 · Zbl 0487.65045 · doi:10.1007/BF00933755 |
| [24] | Ern A., Vorticity-velocity modeling of chemically reacting flows (1994) |
| [25] | DOI: 10.1016/S0010-2180(01)00352-2 · doi:10.1016/S0010-2180(01)00352-2 |
| [26] | DOI: 10.1016/j.pecs.2008.10.002 · doi:10.1016/j.pecs.2008.10.002 |
| [27] | DOI: 10.1007/BF01406969 · Zbl 0313.65070 · doi:10.1007/BF01406969 |
| [28] | Curtis A., J. Inst. Math. Appl. 13 pp 117– (1974) · doi:10.1093/imamat/13.1.117 |
| [29] | DOI: 10.1080/00102209108951734 · doi:10.1080/00102209108951734 |
| [30] | DOI: 10.1108/02644400810881392 · Zbl 1257.74157 · doi:10.1108/02644400810881392 |
| [31] | DOI: 10.1137/0611030 · Zbl 0711.65034 · doi:10.1137/0611030 |
| [32] | Knuth D. E., The Art of Computer Programming 1 (1968) · Zbl 0191.17903 |
| [33] | Grama A., Introduction to Parallel Computing (2003) |
| [34] | Saad Y., SIAM J. Sci. Statist. Comput. pp 856– (1986) · Zbl 0599.65018 · doi:10.1137/0907058 |
| [35] | DOI: 10.1017/CBO9780511615115 · doi:10.1017/CBO9780511615115 |
| [36] | Cooke J. A., One- and two-dimensional simulations of diffusion flames using complex fuels and distributed-memory computing clusters (2006) |
| [37] | DOI: 10.1080/01495730208941444 · doi:10.1080/01495730208941444 |
| [38] | DOI: 10.1016/S0010-2180(01)00254-1 · doi:10.1016/S0010-2180(01)00254-1 |
| [39] | DOI: 10.1016/S0010-2180(97)00282-4 · doi:10.1016/S0010-2180(97)00282-4 |
| [40] | DOI: 10.1016/S0360-1285(00)00013-7 · doi:10.1016/S0360-1285(00)00013-7 |
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.
