基于非结构网格流场超大规模并行计算
|
摘要
大规模并行的计算流体力学已成为现代航空工业研发的核心手段之一。基于非结构混合网格和有限体积法,发展了适用于工业级复杂外形气动计算的并行流动数值模拟方法。文中首先介绍了紧致数值离散格式、基于Metis的分布式多核系统网格分区技术、并行边界虚拟单元技术和MPI并行实现等相关算法。采用网格量相对较小的旋成体构型绕流模型对比分析多核并行计算结果与单核计算结果以验证并行计算的正确性,比较了不同并行规模下并行效率和残差收敛情况。然后通过对上亿网格单元的运输机复杂构型绕流模拟,开展并行效率的测试,结果表明,本文方法并行加速性能高,直到多达18816核并行效率都保持在80%以上。
Massive parallel computational fluid dynamics is a vital tool for modern aviation industry.The parallel aerodynamics numerical simulation techniques based on unstructured mixed grid and cell-centered finite volume method(FVM)were presented in this paper.These techniques were developed for complicated configurations in industrial applications.The algorithms were discussed in detail for parallel implementation with message passing interface(MPI),including compact numerical discrete schemes,grid partition methods for multiprocessors based on Metis software package,ghost cells method for both physical and parallel boundary faces,and data transfer scheme.The parallel computing was verified by using a revolving model of relatively small amount of grid cells,and the convergence characteristics of aerodynamic force and flow variables residual were discussed.A large transport airplane configuration with about a hundred millions of grid elements was used for parallel test.An excellent performance of speed up was obtained,and the efficiency was above 80% with up to18816 processors.
Massive parallel computational fluid dynamics is a vital tool for modern aviation industry.The parallel aerodynamics numerical simulation techniques based on unstructured mixed grid and cell-centered finite volume method(FVM)were presented in this paper.These techniques were developed for complicated configurations in industrial applications.The algorithms were discussed in detail for parallel implementation with message passing interface(MPI),including compact numerical discrete schemes,grid partition methods for multiprocessors based on Metis software package,ghost cells method for both physical and parallel boundary faces,and data transfer scheme.The parallel computing was verified by using a revolving model of relatively small amount of grid cells,and the convergence characteristics of aerodynamic force and flow variables residual were discussed.A large transport airplane configuration with about a hundred millions of grid elements was used for parallel test.An excellent performance of speed up was obtained,and the efficiency was above 80% with up to18816 processors.
引文
[1]何开锋,钱炜祺,陈坚强,等.基于流体力学和电磁学方程数值求解的飞行器气动隐身一体化设计[J].空气动力学学报,2009,27(2):180-185.HE K F,QIAN W Q,CHEN J Q,et al.Integrated aircraft design of aerodynamic and stealthy performance with numerically solving fluid dynamics and electro-magnetics equations[J].Acta Aerodynamica Sinica,2009,27(2):180-185.(in Chinese)
[2]LI B,TANG J,ZHOU N C.Numerical simulation of missile launching with rudder deflection[J].Transactions of Nanjing University of Aeronautics&Astronautics,2013,30(2):107-115.
[3]Lyu Z,Martinsy J A.RANS-based aerodynamic shape optimization of a blended-wing-body aircraft[R].AIAA-2013-2586.
[4]唐静,邓有奇,马明生,等.飞翼气动优化中参数化和网格变形技术[J].航空学报,2015,36(5):1480-1490.TANG J,DENG Y Q,MA M S,et al.Parametrization and grid deformation techniques for flying-wing aerodynamic optimization[J].Acta Aeronautica et Astronautica Sinica,2015,36(5):1480-1490.(in Chinese)
[5]RONZHEIMERA,HEPPERLE M,BREZILLON J.Aerodynamic optimal engine integration for a business jet configuration[R].AIAA-2012-1841.
[6]Roose D,Driessche R V.并行计算机和计算流体力学并行算法[J].力学进展,1998,28(1):111-125.ROOSE D,DRIESSCHE R V.Parallel computers and parallel algorithms for CFD[J].Advances in Mechanics,1998,28(1):111-125.(in Chinese)
[7]KNIGH D.Parallel computing in computational fluid dynamics[C]//AGARD Confernce Proceedings/Progress and Challenges in CFD Methods and Algorithms,1996.
[8]陆林生,董超群,王玲秋,等.并行程序概念设计方法研究[J].计算机学报,2003,26(9):1086-1093.LU L S,DONG C Q,WANG L Q,et al.Researh on parallel program conceptual designing method[J].Chinese Journal of Computers,2003,26(9):1086-1093.(in Chinese)
[9]陈刚,王磊,陆忠华,等.万核级并行飞机气动模拟软件CCFD研制[J].华中科技大学学报(自然科学版),2011,39(增刊1):99-101.CHEN G,WANG L,LU Z H,et al.Development of tenthousand-core parallel software CCFD for aircraft aerodynamics simulation[J].Journal of Huazhong University of Science and Technology(Natural Science Edition),2011,39(Sup.1):99-101.(in Chinese)
[10]WIART C,HILLEWAERK K.Development and validation of a massively parallel high-order solver for DNS and LES of industrial Flows[C]//IDIHOM:Industrialization of HighOrder Methods-A Top-Down Approach,2015,251-292.
[11]潘沙,李桦,夏智勋.高性能并行计算在航空航天CFD数值模拟中的应用[J].计算机工程与科学,2012,34(8):191-198.PAN S,LI H,XIA Z X.High-performance parallel computiation application for aerospace CFD numerical simulation[J].Computer Engineering&Science,2012,34(8):191-198.(in Chinese)
[12]HUANG X C,XIAO A.Iteration and parallel computation on computational fluid dynamics[C]//International Conference on Intelligent Computation Technology&Automation,2014:318-321.
[13]SHEKE S,KALYAN W.Parallel multigrid solver for NavierStokes equation using Open MPI[J].International Journal of Computer Science Trends and Technology,2015,3(5):131-134.
[14]BERGER M J,AFTOSMIS M J,MARSHALL D D.Performance of a new CFD flow solver using a hybrid programming paradigm[J].Journal of Parallel&Distributed Computing,2005,65(4):414-423.
[15]KARYPIS G,KUMAR V.A fast and high quality multilevel scheme for partitioning irregular graphs[J].SIAM Journal on Scientific Computing,1998,20(1):359-392.
[16]PULLIAM T H.Early development of implicit methods for computational fluid dynamics at NASA Ames[J].Computers&Fluids,2009,38(2):491-495.
[17]BLAZEK J.Computational fluid dynamics:principles and applications[M].Elsevier:Elsevier press,2015
[18]KIM J S,KWON O J.Improvement on block LU-SGS scheme for unstructured mesh navier-stokes computations[R].AIAA-2002-1061.
[19]SPALART S R.A one-equation turbulence model for aerodynamic flows[R].AIAA-92-0439,1992.
[20]ROEP L.Approximate Riemann solvers,parameter vectors,and difference schemes[J].Journal of Computational Physics,1981,43(2):357-372.
[21]HASELBACHER A,BLAZEK J.Accurate and efficient discretisation of the Navier-Stokes equations on mixed grids[J].AIAA Journal,2000,38(11):2094-2102.
[2]LI B,TANG J,ZHOU N C.Numerical simulation of missile launching with rudder deflection[J].Transactions of Nanjing University of Aeronautics&Astronautics,2013,30(2):107-115.
[3]Lyu Z,Martinsy J A.RANS-based aerodynamic shape optimization of a blended-wing-body aircraft[R].AIAA-2013-2586.
[4]唐静,邓有奇,马明生,等.飞翼气动优化中参数化和网格变形技术[J].航空学报,2015,36(5):1480-1490.TANG J,DENG Y Q,MA M S,et al.Parametrization and grid deformation techniques for flying-wing aerodynamic optimization[J].Acta Aeronautica et Astronautica Sinica,2015,36(5):1480-1490.(in Chinese)
[5]RONZHEIMERA,HEPPERLE M,BREZILLON J.Aerodynamic optimal engine integration for a business jet configuration[R].AIAA-2012-1841.
[6]Roose D,Driessche R V.并行计算机和计算流体力学并行算法[J].力学进展,1998,28(1):111-125.ROOSE D,DRIESSCHE R V.Parallel computers and parallel algorithms for CFD[J].Advances in Mechanics,1998,28(1):111-125.(in Chinese)
[7]KNIGH D.Parallel computing in computational fluid dynamics[C]//AGARD Confernce Proceedings/Progress and Challenges in CFD Methods and Algorithms,1996.
[8]陆林生,董超群,王玲秋,等.并行程序概念设计方法研究[J].计算机学报,2003,26(9):1086-1093.LU L S,DONG C Q,WANG L Q,et al.Researh on parallel program conceptual designing method[J].Chinese Journal of Computers,2003,26(9):1086-1093.(in Chinese)
[9]陈刚,王磊,陆忠华,等.万核级并行飞机气动模拟软件CCFD研制[J].华中科技大学学报(自然科学版),2011,39(增刊1):99-101.CHEN G,WANG L,LU Z H,et al.Development of tenthousand-core parallel software CCFD for aircraft aerodynamics simulation[J].Journal of Huazhong University of Science and Technology(Natural Science Edition),2011,39(Sup.1):99-101.(in Chinese)
[10]WIART C,HILLEWAERK K.Development and validation of a massively parallel high-order solver for DNS and LES of industrial Flows[C]//IDIHOM:Industrialization of HighOrder Methods-A Top-Down Approach,2015,251-292.
[11]潘沙,李桦,夏智勋.高性能并行计算在航空航天CFD数值模拟中的应用[J].计算机工程与科学,2012,34(8):191-198.PAN S,LI H,XIA Z X.High-performance parallel computiation application for aerospace CFD numerical simulation[J].Computer Engineering&Science,2012,34(8):191-198.(in Chinese)
[12]HUANG X C,XIAO A.Iteration and parallel computation on computational fluid dynamics[C]//International Conference on Intelligent Computation Technology&Automation,2014:318-321.
[13]SHEKE S,KALYAN W.Parallel multigrid solver for NavierStokes equation using Open MPI[J].International Journal of Computer Science Trends and Technology,2015,3(5):131-134.
[14]BERGER M J,AFTOSMIS M J,MARSHALL D D.Performance of a new CFD flow solver using a hybrid programming paradigm[J].Journal of Parallel&Distributed Computing,2005,65(4):414-423.
[15]KARYPIS G,KUMAR V.A fast and high quality multilevel scheme for partitioning irregular graphs[J].SIAM Journal on Scientific Computing,1998,20(1):359-392.
[16]PULLIAM T H.Early development of implicit methods for computational fluid dynamics at NASA Ames[J].Computers&Fluids,2009,38(2):491-495.
[17]BLAZEK J.Computational fluid dynamics:principles and applications[M].Elsevier:Elsevier press,2015
[18]KIM J S,KWON O J.Improvement on block LU-SGS scheme for unstructured mesh navier-stokes computations[R].AIAA-2002-1061.
[19]SPALART S R.A one-equation turbulence model for aerodynamic flows[R].AIAA-92-0439,1992.
[20]ROEP L.Approximate Riemann solvers,parameter vectors,and difference schemes[J].Journal of Computational Physics,1981,43(2):357-372.
[21]HASELBACHER A,BLAZEK J.Accurate and efficient discretisation of the Navier-Stokes equations on mixed grids[J].AIAA Journal,2000,38(11):2094-2102.