Parallel, Hierarchical Solution Algorithms for Diffusion Synthetic Acceleration of the Neutron Transport Equation PDF Download

Author: Benedict John O'Malley
Publisher:
ISBN:
Category :
Languages : en
Pages :

View

Book Description

Author: MUSA YAVUZ
Publisher:
ISBN:
Category :
Languages : en
Pages : 530

View

Book Description
performance of the new multigrid method degrades with an increasing number of processors, not because of the degradation in the convergence rate, but because of the domination of communication overhead and unemployment of some processors on the coarse grids.

Author:
Publisher:
ISBN:
Category :
Languages : en
Pages :

View

Book Description
The linear Boltzmann transport equation (BTE) is an integro-differential equation arising in deterministic models of neutral and charged particle transport. In slab (one-dimensional Cartesian) geometry and certain higher-dimensional cases, Diffusion Synthetic Acceleration (DSA) is known to be an effective algorithm for the iterative solution of the discretized BTE. Fourier and asymptotic analyses have been applied to various idealizations (e.g., problems on infinite domains with constant coefficients) to obtain sharp bounds on the convergence rate of DSA in such cases. While DSA has been shown to be a highly effective acceleration (or preconditioning) technique in one-dimensional problems, it has been observed to be less effective in higher dimensions. This is due in part to the expense of solving the related diffusion linear system. We investigate here the effectiveness of a parallel semicoarsening multigrid (SMG) solution approach to DSA preconditioning in several three dimensional problems. In particular, we consider the algorithmic and implementation scalability of a parallel SMG-DSA preconditioner on several types of test problems.

Author: Salli Moustafa
Publisher:
ISBN:
Category :
Languages : en
Pages : 0

View

Book Description
High-fidelity nuclear reactor core simulations require a precise knowledge of the neutron flux inside the reactor core. This flux is modeled by the linear Boltzmann equation also called neutron transport equation. In this thesis, we focus on solving this equation using the discrete ordinates method (SN) on Cartesian mesh. This method involves a source iteration scheme including a sweep over the spatial mesh and gathering the vast majority of computations in the SN method. Due to the large amount of computations performed in the resolution of the Boltzmann equation, numerous research works were focused on the optimization of the time to solution by developing parallel algorithms for solving the transport equation. However, these algorithms were designed by considering a super-computer as a collection of independent cores, and therefore do not explicitly take into account the memory hierarchy and multi-level parallelism available inside modern super-computers. Therefore, we first proposed a strategy for designing an efficient parallel implementation of the sweep operation on modern architectures by combining the use of the SIMD paradigm thanks to C++ generic programming techniques and an emerging task-based runtime system: PaRSEC. We demonstrated the need for such an approach using theoretical performance models predicting optimal partitionings. Then we studied the challenge of converging the source iterations scheme in highly diffusive media such as the PWR cores. We have implemented and studied the convergence of a new acceleration scheme (PDSA) that naturally suits our Hybrid parallel implementation. The combination of all these techniques have enabled us to develop a massively parallel version of the SN Domino solver. It is capable of tackling the challenges posed by the neutron transport simulations and compares favorably with state-of-the-art solvers such as Denovo. The performance of the PaRSEC implementation of the sweep operation reaches 6.1 Tflop/s on 768 cores corresponding to 33.9% of the theoretical peak performance of this set of computational resources. For a typical 26-group PWR calculations involving 1.02×1012 DoFs, the time to solution required by the Domino solver is 46 min using 1536 cores.

Author: Daniele Sciannandrone
Publisher:
ISBN:
Category :
Languages : en
Pages : 0

View

Book Description
The topic of our research is the application of the Method of Long Characteristics (MOC) to solve the Neutron Transport Equation in three-dimensional axial geometries. The strength of the MOC is in its precision and versatility. As a drawback, it requires a large amount of computational resources. This problem is even more severe in three-dimensional geometries, for which unknowns reach the order of tens of billions for assembly-level calculations.The first part of the research has dealt with the development of optimized tracking and reconstruction techniques which take advantage of the regularities of three-dimensional axial geometries. These methods have allowed a strong reduction of the memory requirements and a reduction of the execution time of the MOC calculation.The convergence of the iterative scheme has been accelerated with a lower-order transport operator (DPN) which is used for the initialization of the solution and for solving the synthetic problem during MOC iterations.The algorithms for the construction and solution of the MOC and DPN operators have been accelerated by using shared-memory parallel paradigms which are more suitable for standard desktop working stations. An important part of this research has been devoted to the implementation of scheduling techniques to improve the parallel efficiency.The convergence of the angular quadrature formula for three-dimensional cases is also studied. Some of these formulas take advantage of the reduced computational costs of the treatment of planar directions and the vertical direction to speed up the algorithm.The verification of the MOC solver has been done by comparing results with continuous-in-energy Monte Carlo calculations. For this purpose a coupling of the 3D MOC solver with the Subgroup method is proposed to take into account the effects of cross sections resonances. The full calculation of a FBR assembly requires about 2 hours of execution time with differences of few PCM with respect to the reference results.We also propose a higher order scheme of the MOC solver based on an axial polynomial expansion of the unknown within each mesh. This method allows the reduction of the meshes (and unknowns) by keeping the same precision.All the methods developed in this thesis have been implemented in the APOLLO3 version of the neutron transport solver TDT.

Author: Laurent Graziano
Publisher:
ISBN:
Category :
Languages : en
Pages : 0

View

Book Description
The purpose of this PhD is the implementation of an axial polynomial approximation in a three-dimensional Method Of Characteristics (MOC) based solver. The context of the work is the solution of the steady state Neutron Transport Equation for critical systems, and the practical implementation has been realized in the Two/three Dimensional Transport (TDT) solver, as a part of the APOLLO3® project. A three-dimensional MOC solver for 3D extruded geometries has been implemented in this code during a previous PhD project, relying on a piecewise constant approximation for the neutrons fluxes and sources. The developments presented in the following represent the natural continuation of this work. Three-dimensional neutron transport MOC solvers are able to produce accurate results for complex geometries. However accurate, the computational cost associated to this kind of solvers is very important. An axial polynomial representation of the neutron angular fluxes has been used to lighten this computational burden.The work realized during this PhD can be considered divided in three major parts: transport, acceleration and others. The first part is constituted by the implementation of the chosen polynomial approximation in the transmission and balance equations typical of the Method Of Characteristics. This part was also characterized by the computation of a set of numerical coefficients which revealed to be necessary in order to obtain a stable algorithm. During the second part, we modified and implemented the solution of the equations of the DPN synthetic acceleration. This method was already used for the acceleration of both inners and outers iteration in TDT for the two and three dimensional solvers at the beginning of this work. The introduction of a polynomial approximation required several equations manipulations and associated numerical developments. In the last part of this work we have looked for the solutions of a mixture of different issues associated to the first two parts. Firstly, we had to deal with some numerical instabilities associated to a poor numerical spatial or angular discretization, both for the transport and for the acceleration methods. Secondly, we tried different methods to reduce the memory footprint of the acceleration coefficients. The approach that we have eventually chosen relies on a non-linear least square fitting of the cross sections dependence of such coefficients. The default approach consists in storing one set of coefficients per each energy group. The fit method allows replacing this information with a set of coefficients computed during the regression procedure that are used to re-construct the acceleration matrices on-the-fly. This procedure of course adds some computational cost to the method, but we believe that the reduction in terms of memory makes it worth it.In conclusion, the work realized has focused on applying a simple polynomial approximation in order to reduce the computational cost and memory footprint associated to a Method Of Characteristics solver used to compute the neutron fluxes in three dimensional extruded geometries. Even if this does not a constitute a radical improvement, the high order approximation that we have introduced allows a reduction in terms of memory and computational times of a factor between 2 and 5, depending on the case. We think that these results will constitute a fertile base for further improvements.

Author: Dylan Scott Hoagland
Publisher:
ISBN:
Category :
Languages : en
Pages : 354

View

Book Description

Author:
Publisher:
ISBN:
Category :
Languages : en
Pages :

View

Book Description
A class of acceleration schemes is investigated which resembles the conventional synthetic method in that they utilize the diffusion operator in the transport iteration schemes. The accelerated iteration involves alternate diffusion and transport solutions where coupling between the equations is achieved by using a correction term applied to either the diffusion coefficient, the removal cross section, or the source of the diffusion equation. The methods involving the modification of the diffusion coefficient and of the removal term yield nonlinear acceleration schemes and are used in k/sub eff/ calculations, while the source term modification approach is linear at least before discretization, and is used for inhomogeneous source problems. A careful analysis shows that there is a preferred differencing method which eliminates the previously observed instability of the conventional synthetic method. Use of this preferred difference scheme results in an acceleration method which is at the same time stable and efficient. This preferred difference approach renders the source correction scheme, which is linear in its continuous form, nonlinear in its differenced form. An additional feature of these approaches is that they may be used as schemes for obtaining improved diffusion solutions for approximately twice the cost of a diffusion calculation. Numerical experimentation on a wide range of problems in one and two dimensions indicates that improvement from a factor of two to ten over rebalance or Chebyshev acceleration is obtained. The improvement is most pronounced in problems with large regions of scattering material where the unaccelerated transport solutions converge very slowly.

Author: B. D. Ganapol
Publisher:
ISBN: 9789264990562
Category : Criticality (Nuclear engineering)
Languages : en
Pages : 269

View

Book Description

Author: Volodymyr Kindratenko
Publisher: Springer
ISBN: 3319065483
Category : Computers
Languages : en
Pages : 404

View

Book Description
This book brings together research on numerical methods adapted for Graphics Processing Units (GPUs). It explains recent efforts to adapt classic numerical methods, including solution of linear equations and FFT, for massively parallel GPU architectures. This volume consolidates recent research and adaptations, covering widely used methods that are at the core of many scientific and engineering computations. Each chapter is written by authors working on a specific group of methods; these leading experts provide mathematical background, parallel algorithms and implementation details leading to reusable, adaptable and scalable code fragments. This book also serves as a GPU implementation manual for many numerical algorithms, sharing tips on GPUs that can increase application efficiency. The valuable insights into parallelization strategies for GPUs are supplemented by ready-to-use code fragments. Numerical Computations with GPUs targets professionals and researchers working in high performance computing and GPU programming. Advanced-level students focused on computer science and mathematics will also find this book useful as secondary text book or reference.