10 results
Filters
Settings
Now showing 1 - 10 of 10
- Some of the metrics are blocked by yourconsent settings
Publication Access status: Open Access , Efficient high-order simulation of aeroacoustics from rigid body motion on massively parallel systems(2021)The numerical simulation of physical problems involves capturing various phenomena occurring simultaneously at different scales in a single simulation. For example, considering aeroacoustic problems, the noise generating flow and the propagation of the sound waves in the far-field need to be taken into account. The increasing computational capacities and the development of modern supercomputers allow for more detailed studies of complex multi-physics and multi-scale problems. In this work, the sound generation by moving obstacles and its propagation up to the far-field is of particular interest. For this purpose, a high-order Discontinuous Galerkin method is utilized to discretize the fluid dynamic equations. These high-order methods are exceptionally efficient as they only require a few degrees of freedom to represent smooth solutions. Therefore, they are often deployed for, e.g., the acoustics far-field, where a homogenous flow field can be found. From the computational perspective on modern high-performing architectures, few degrees of freedom are an exceptional advantage, with memory bandwidth being a bottleneck on modern systems. Additionally, the ratio between communication and computation is minimal due to the loose connection of computational elements at their respective interfaces, which is an additional advantage of these methods, when considering distributed and massively parallel computing systems. However, the high-order representation of complex geometries has been a critical limitation for their application in various fields. The representation of geometrical shapes has to be appropriate to preserve the quality of the numerical solution, which has been discretized with high-order. Incorporating meshing techniques such as body-fitted meshes might not be robust in the workflow for the simulation. They are required to withstand different scenarios that are common, such as scenarios with general complex geometries inside the simulation domain. They can become expensive in computation when involving simulations with multiple geometries and even more when geometries can move. In these cases, the embedded method, also known as immersed boundary method, provides a promising prospect. In this work, the Brinkman penalization technique is applied to model multiple complex and moving geometries. Moving rigid bodies are common in engineering applications. The sound emitted due to the motion and the flow disturbance by geometries is of particular interest, as awareness of environmental impact in society has grown in recent years. Therefore, predicting the produced noise is a common responsibility in different fields, such as the design of wind turbines. These simulations have a complex nature and require an efficient strategy to facilitate them feasibly. Therefore we deploy the partitioned coupling approach, where the complex and large simulation domain is decomposed into smaller subdomains. Each subdomain is configured such that the occurring phenomena can be precisely captured. It results in an efficient strategy allowing for the simulation of various scales and physics, such as the large-scale simulation in this work. The simulation includes the motion of an airfoil and the induced noise that spreads over a large domain.Source Type:Doctoral ThesisSeries:Simulation techniques in Siegen - STSVolume number:61 8 - Some of the metrics are blocked by yourconsent settings
Publication Access status: Open Access , Efficient coupling of fluid and acoustic interaction on massively parallel systems(2021)Multi-scale problems like the generation of sound in a flow field and its sound wave propagation in the far field have become increasingly important in the design phase of industrial devices: One example is noise reduction of aircrafts or wind turbines. Although the generation of sound as well as its propagation can both be described by the same governing equations, wave propagation is a linear phenomenon and its equations can be simplified, which results in less computational effort. Additionally, the generation of sound in a flow field occurs at small spatial scales, while its propagation in the far field has to be observed on a large spatial scale. These large differences in scales are particularly challenging for numerical simulations. Resolving the entire domain with the high resolution that is required for the small scales of the flow domain is impossible due to the vast computational demand. In this thesis, we propose a partitioned coupling approach for the efficient simulation of such problems on massively parallel supercomputers. In partitioned coupling the physical domain is split into smaller subdomains, each covering a different physical phenomenon. Their interaction is realized by exchanging data at the joint coupling interface. Subsequently, these subdomains can be solved with numerical methods, resolutions, and equations tailored to the local physical requirements. Even different solvers can be used. However, this approach also holds numerical challenges at the coupling interface: E.g. for a consistent data exchange in case of different spatial resolutions, direct data evaluation, or efficient interpolation methods are necessary. Within this work, two different approaches of partitioned coupling are implemented and compared: A blackbox and a white-box approach. The black-box approach is characterized by a flexible choice of numerical solvers which allows for a wide range of different applications. Its generality comes with limited access to information inside each solver and, therefore, with a potential loss of performance. However, a black-box approach only acts on point data at the coupling interface and therefore requires external interpolation methods for a consistent coupling in space which is expected to be less efficient than solver-internal data mapping. In contrast, the white-box approach is fully integrated within one numerical framework. Accordingly, it can access solver-internal data mapping methods which promises better numerical results. This tight integration allows for the exploitation of knowledge about internal data structures and, therefore, yields performance benefits. On the other hand, it comes with less flexibility. Both strategies will be compared with respect to quality of data mapping at the coupling interface as well as performance on modern supercomputers. In order to achieve the best performance, the optimal load balancing strategy for a coupled setup is investigated. The benefits of the partitioned coupling approach are demonstrated on an industrial application of a 3D free-stream jet with a high Reynolds number showing that a multi-scale problem can be simulated using today’s compute resources.Source Type:Doctoral ThesisSeries:Simulation techniques in Siegen - STSVolume number:5363 134 - Some of the metrics are blocked by yourconsent settings
Publication Access status: Open Access , Framework for coupled simulation of electrodialysis processes(2021)Electrodialysis is an efficient process for seawater desalination that involves various interacting phenomena. In this process, ions are transported by flow, diffusion and an electric force and separated by selective membranes. For the optimization of this process, it is important to understand these interactions. This work presents rigorous mathematical models to describe the overall process and develops a numerical strategy for its simulation. With this approach it becomes possible to simulate the involved physical effects and their interactions in detail. To achieve this, the Maxwell-Stefan equations for mixtures are used. They take into account the electrical force and the multicomponent interactions with concentration dependent diffusivity coefficients and thermodynamic factors. Additionally, the usual assumption of local electroneutrality is not assumed to allow the nonideal effects in the electrical double layer near the membrane. For the numerical solution of these equations, the multicomponent lattice Boltzmann method (LBM) is developed and implemented in the solver Musubi. This model for the channel flow is coupled with an electric field and a model for the membranes. To obtain the electric field, the LBM that solves the Poisson’s equation is implemented in Musubi. The channels between the membranes are realized by spacers with complex geometry. A mesh generator (Seeder) on the basis of octrees is developed to ensure the appropriate discretization of the mesh for these channels. An essential part of this work is dedicated to the development of the parallel scaling coupling tool APESmate. APESmate is developed within the APES suite along with Seeder and Musubi on a central octree data structure that allows efficient handing of I/O on large scale distributed parallel computing systems. The developed software is used to compare the nonideal multicomponent model for various concentrations and surface potentials. The results show that nonideal effects increase with the concentration, especially in the electrical double layer. The spacers for various hydrodynamic angles and inflow velocities near and away from a sealed corner are investigated to find the design with reduced pressure drop and without low velocity zones. The highly resolved simulations show that the pressure drop increases with the hydrodynamic angle, while the extend of the low flow regions decreases.Source Type:BookSeries:Simulation techniques in Siegen - STSVolume number:43 3 - Some of the metrics are blocked by yourconsent settings
Publication Access status: Open Access , Parallel multi-scale simulations with octrees and coupled applications(2016)Physical simulations often require the consideration of many phenomena and scales. For example in aeroacoustic problems, both, the flow generating the noise and the sound wave propagation needs to be considered. This work investigates numerical approaches to such problems on large distributed and parallel computing systems. The coupling framework KOP is parallelized as far as possible and to overcome fundamental scalability limits a new framework APES is developed. Both implementations utilize high-order discretizations, as these allow for accurate simulations with less degrees of freedoms than lower order methods. This property of high-order methods is an important feature for modern supercomputing systems, as memory to represent degrees of freedom in a simulation is a scarce resource. The presented methods enable the transient simulation of multi-scale setups but detailed resolutions still require large amounts of computational resources. A focus is put on the efficient utilization of modern computing systems to address this need. Besides the scalability of the implementations, the importance of single core optimization and vectorization is illustrated. KOP uses discrete points to realize the coupling and allows for the interaction between domains with differing discretizations and solved equation systems. Arbitrary mesh configurations are supported and both, structured and unstructured mesh solvers are available in the framework. In both framworks explicit time integration methods are deployed to resolve the time dependent simulations. The coupling allows for a varying time step width over the participating domains by a sub-cycling method. Various conservation laws can be solved by the presented frameworks ranging from Maxwell’s equations and linearized Euler equations to full compressible Navier-Stokes equations. A fully distributed coupling approach is developed that allows for coupling of those in a large-scale simulation to solve, for example, aeroacoustic problems. APES enables high-order discretizations in the spectral regime. It involves a fully scalable toolchain for mesh-based simulations featuring a mesh generation and a post-processing tool to support the solvers. The common foundation of these tools is an Octree representation for the mesh, and this work specifically covers the generation of high-order geometry approximations in the developed mesh generator Seeder. This robust mechanism works for arbitrarily complex surfaces and offers a practical way to tackle engineering tasks with spectral element discretizations.Source Type:BookSeries:Simulation techniques in Siegen - STSVolume number:11 3 - Some of the metrics are blocked by yourconsent settings
Publication Efficient high-order simulation of aeroacoustics from rigid body motion on massively parallel systems(2021-12-29)This work presents an efficient strategy to facilitate large-scale simulations of aeroacoustics induced by rigid body motion. A Discontinuous Galerkin method, providing low dispersion and dissipation errors, is used to propagate the flowinduced acoustics. An embedded boundary method is deployed to model complex moving geometries. Both, flow field and geometry modeling can use the same discretization method. The combination with a partitioned coupling approach allows for efficient massively parallel simulations of multiple scales. The modeling method and the partitioned coupling are thoroughly investigated and used in a detailed simulation of a real-world three-dimensional engineering problem of a moving airfoil that disturbs the fluid flow, resulting in acoustic waves.Source Type:DissertationSeries:Simulation Techniques in Siegen - STS2 2 - Some of the metrics are blocked by yourconsent settings
Publication Transition to Turbulence in Physiological Flows: Direct Numerical Simulation of Hemodynamics in Intracranial Aneurysms and Cerebrospinal Fluid Hydrodynamics in the Spinal Canal(2016-10-05)Numerical simulations using the Lattice Boltzmann Method are conducted on massively parallel supercomputers to explore the detailed dynamics of transitional flows in patient specific anatomical geometries. The methodology is validated against the literature by simulations of stenotic flows and conditions for transition in a purely oscillatory flow are explored. The characteristics of transitional flow in physiological cases are quantified in detail and the implications of such a flow regime on the pathology and pathophysiology of intracranial aneurysms and Chiari I malformation are discussed.Source Type:DissertationSeries:Simulation Techniques in Siegen - STS2 2 - Some of the metrics are blocked by yourconsent settings
Publication Framework for Coupled Simulation of Electrodialysis Processes(2021-05-10)This thesis presents the development of a scalable coupled simulation framework to simulate the interactions of different physical phenomena. It employs rigorous models and develops a numerical strategy for coupled simulations implemented in the APESmate tool. This coupling strategy maintains good scalability on large, distributed computing systems and ensures accurate interpolation between meshes. APESmate enables multiphysics simulations with any physics supported by solvers in the APES framework. One specific multiphysics application this thesis discusses is the electrodialysis process. Here, ions in a multicomponent flow are driven apart by an electric field. To this end, the presented work implements a Maxwell-Stefan model in the Lattice-Boltzmann solver Musubi and investigates different geometrical designs for electrodialysis.Source Type:DissertationSeries:Simulation Techniques in Siegen - STS4 1 - Some of the metrics are blocked by yourconsent settings
Publication Parallel Multi-Scale-Simulations with Octrees and Coupled Applications(2016-10-03)Aero-acoustic phenomena pose a multi-scale problem. To tackle this class of problems, a general coupling tool for structured and unstructured meshes is first parallelized to scale on thousands of processes. However, as there remains a principle bottleneck in the treatment of meshes with this approach, an entirely new framework on the basis of octrees is then developed. This restriction in the mesh representation allows for fully parallel simulations of arbitrary large setups.Source Type:DissertationSeries:Simulation Techniques in Siegen - STS1 3 - Some of the metrics are blocked by yourconsent settings
Publication Efficient coupling of fluid and acoustic interaction on massively parallel systems(2021-12-15)This work presents an efficient coupling strategy to tackle the multi-scale problem posed by technical devices that radiate sound. Two different approaches, a flexible multi-solver approach (with the coupling library preCICE) and an optimized integrated approach (called APESmate), are established and compared in terms of quality, load balancing, and performance. Load balancing is a crucial aspect regarding simulations in the field of high performance computing where many processes are used in parallel and, in particular, when considering coupled simulation where different partitions are solving different physical phenomena and numerical discretization. The benefits of the partitioned coupling approach and good load balancing are demonstrated on an industrial application of a 3D free-stream jet with a high Reynolds number showing that a multi-scale problem can be efficiently simulated using today‘s computing resources.Source Type:DissertationSeries:Simulation Techniques in Siegen - STS3 3 - Some of the metrics are blocked by yourconsent settings
Publication Efficient Lattice Boltzmann Simulations on Large Scale High Performance Computing Systems(2017-06-19)This thesis presents efforts to attain efficient Lattice Boltzmann simulations on large scale parallel High Performance Computing (HPC) systems. To enable the fluid simulation on non-uniform meshes, a three-dimensional compact interpolation scheme is presented, which greatly reduces the computational effort. For an efficient utilization of the parallel and distributed computing resources, the key issues in software implementation are discussed. The presented methods are implemented in the Lattice Boltzmann solver – Musubi, and available under a permissive open source license.Source Type:DissertationSeries:Simulation Techniques in Siegen - STS2 2
