Edmund Kügeler

GASPI – A Partitioned Global Address Space Programming Interface

At the threshold to exascale computing, limitations of the MPI programming model become more and more pronounced. HPC programmers have to design codes that can run and scale on systems with hundreds of thousands of cores. Setting up accordingly many communication buffers, point-to-point communication links, and using bulk-synchronous communication phases is contradicting scalability in these dimensions. Moreover, the reliability of upcoming systems will worsen.

Appropriate Turbulence Modelling for Turbomachinery Flows using a Two-Equation Turbulence Model

The simulation quality of numerical flow simulations depends on the choice of physical modelling as well as an appropriate numerical treatment. In this study, a standard two-equation turbulence model has been extended for compressible, rotational flow as it occurs in turbomachinery and subsequently applied to different turbomachinery relevant flows of varying complexity. A number of different numerical schemes has been employed to evaluate their impact on the solution.

HICFD: Highly Efficient Implementation of CFD Codes for HPC Many-Core Architectures

The objective of the German BMBF research project Highly Efficient Implementation of CFD Codes for HPC Many-Core Architectures (HICFD) is to develop new methods and tools for the analysis and optimization of the performance of parallel computational fluid dynamics (CFD) codes on high performance computer systems with many-core processors. In the work packages of the project it is investigated how the performance of parallel CFD codes written in C can be increased by the optimal use of all parallelism levels. On the highest level Message Passing Interface (MPI) is utilized. Furthermore, on the level of the many-core architecture, highly scaling, hybrid OpenMP/MPI methods are implemented. On the level of the processor cores the parallel Single Instruction Multiple Data (SIMD) units provided by modern CPUs are exploited.

Deterministic Stress Modeling for Multistage Compressor Flowfields

Unsteadiness is one of the main characteristics in turbomachinery flows. Local unsteady changes in static pressure must exist within a turbo-machine in order for that machine to exchange energy with the fluid. The primary reason for unsteady effects lies in the interaction between moving and stationary blade rows. The industrial design process of aero-engines and gas turbines is still based on Reynolds-averaged Navier-Stokes (RANS) techniques where the coupling of blade rows is carried out by mixing-planes. However, this methodology does not cover deterministic unsteadiness in an adequate way. For standard aero-optimization, detailed unsteadiness is not essential to the designer of turbomachines but rather its effect on the time averaged solution. The time averaged deterministic unsteadiness can be expressed in terms of deterministic stresses. The present paper presents two different modeling strategies for deterministic stresses that constitute an improvement of the conventional steady mixing-plane approach. Whilst one of the presented models operates with deterministic flux terms based on preliminary unsteady simulations, the other one, a novel transport model for deterministic stress, is a stand-alone approach based on empirical correlations and a wide range of numerical experiments. A 4.5 stage transonic compressor is analyzed regarding blade row interaction effects and their impact on the time averaged solution. The two models are applied to the compressor and their solutions are compared to conventional mixing-plane, time accurate and experimental data. The results for the speedline, the wake shapes, the radial distributions and the rotor blade loadings show that the deterministic stress models strongly improve the RANS solution towards the time accurate and the experimental methods.

AMANDA - A Distributed System for Aircraft Design

In the AMANDA project a component-based framework for the integration of coupled technical applications, running distributed in a network, is developed. It is designed to deal with parallel and sequential programs and massive data exchange between the integrated programs. Two pilot applications will be implemented to show the feasibility of the chosen approach: a trimmed, freely flying, elastic airplane and an aircooled turbine. Beside using the integration systems, the MpCCI library is used for the coupling of the codes, each simulating single physical processes.

Adaption of Giles Non-Local Non-Reflecting Boundary Conditions for a Cell-Centered Solver for Turbomachinery Applications

In contrast to external flow aerodynamics, where one-dimensional Riemann boundary conditions can be applied far up- and downstream, the handling of non-reflecting boundary conditions for turbomachinery applications poses a greater challenge due to small axial gaps normally encountered. For boundaries exposed to non-uniform flow in the vicinity of blade rows, the quality of the simulation is greatly influenced by the underlying non-reflecting boundary condition and its implementation. This paper deals with the adaptation of Giles’ well-known exact non-local boundary conditions for two-dimensional steady flows to a cell-centered solver specifically developed for turbomachinery applications. It is shown that directly applying the theory originally formulated for a cell-vertex scheme to a cell-centered solver may yield an ill-posed problem due to the necessity of having to reconstruct boundary face values before actually applying the exact non-reflecting theory.In order to ensure well-posedness, Giles’ original approach is adapted for cell-centered schemes with a physically motivated reconstruction of the boundary face values, while still maintaining the non-reflecting boundary conditions. The extension is formulated within the original framework of determining the circumferential distribution of one-dimensional characteristics on the boundary. It is shown that, due to approximations in the one-dimensional characteristic reconstruction of boundary face values, the new approach can only be exact in the limiting case of cells with a vanishing width in the direction normal to the boundary if a one-dimensional characteristic reconstruction of boundary face values is used. To overcome the dependency on the width of the last cell, the new boundary condition is expressed explicitly in terms of a two-dimensional modal decomposition of the flow field. In this formulation, vanishing modal amplitudes for all incoming two-dimensional modes can easily be accomplished for a converged solution. Hence we are able to ensure perfectly non-reflecting boundary conditions under the same conditions as the original approach. The improvements of the new method are demonstrated for both a subsonic turbine and a transonic compressor test case.Copyright

Assessment of Transition Modeling for the Design of Controlled Diffusion Airfoil Compressor Cascades

The aim of this paper is to investigate whether correlation-based transition models can be used for the design of CDA profiles. To this end, a CDA compressor cascade has been widely experimentally investigated at DLR Cologne. Off-design measurements have been carried out and the influence of the variation of four flow parameters has been investigated: The inlet Mach number, the incidence, the chord-based Reynolds number and the free-stream turbulence intensity. The inlet Mach number has been varied from 0.5 up to 0.8. The incidence was varied over the whole working range and beyond. Realistic values of the Reynolds number and of the free-stream turbulence intensity have been attained. Hence, the test case apt to assess the capacity of the DLR’s in-house turbomachinery specific CFD code TRACE to design modern compressor blades. In this paper, computations simulating the influence of those four parameters on the performance of the CDA profile are presented and compared to the measurements. Two transition models are used for this study: an in-house model denoted MultiMode model and the γ-ReΘ model. In addition, two turbulence models (Wilcox k-ω and Menter k-ω SST) and their turbomachinery extensions have also been used for this study. The results between the different numerical simulations and the measurements are discussed in term of loss coefficients and Mach number distributions. The computed losses are close to the experimental values and the physics of the flow is also well reproduced. Bypass transition as well as laminar separation bubbles have been simulated in accordance with the experimental observations. Hence, the TRACE code is able to predict the onset of transition over a wide range of flow conditions.© 2013 ASME

On the Simulation of Unsteady Turbulence and Transition Effects in a Multistage Low Pressure Turbine: Part III — Comparison of Harmonic Balance and Full Wheel Simulation

This is the third part of a series of three papers on the simulation of turbulence and transition effects in a multistage low pressure turbine. The third part of the series deals with the detailed comparison of the harmonic balance calculations with the full wheel simulations and measurements for the two-stage low-pressure turbine. The harmonic balance calculations were taken into account turbulence and transition either only once in the 0th harmonic and the other times in all harmonics. The same Wilcox’s two-equation k − omega turbulence model along with Menter and Langtry’s two-equation  Gamma− Re_Theta transition model is used in the Harmonic Balance simulation as in the full wheel simulations. The measurements on the second stator of the low-pressure turbine have been carried out separately for downstream and upstream influences. Thus, a dedicated comparison of the downstream and upstream influences of the flow to the second stator is possible. In the Harmonic Balance calculations, the influences of the not directly adjacent blade, i.e. the first stator, were also included in the second stator. In the first analysis, however, it was shown that the consistency with the full wheel configuration and the measurement in this case was not as good as expected. From the analysis of the full wheel simualtion, we found that there is a considerable variation in the magnitude of the unsteady values in the second stator. In a further deeper consideration of the configuration, it is found that modes are reflected in downstream rows and influences flow in the second stator. After the integration of these modes into the harmonic balance calculations, a much better agreement was reached with results the full wheel simulation and the measurements. The second stator has a laminar region on the suction side starting at the leading edge and then transition takes place via a separation or in bypass mode, depending on the particular blade viewed in the circumferential direction. In the area of transition, the clear difference between the calculations without and with consideration of the higher harmonics in the turbulence and transition models can be clearly seen. The consideration of the higher harmonics in the turbulence and transition models brings about an improvement in the consistency.

On the Simulation of Unsteady Turbulence and Transition Effects in a Multistage Low Pressure Turbine: Part II — Full-Wheel Simulation

This is the second part of a series of three papers on the simulation of turbulence and transition effects in a multistage low pressure turbine. In this second part, the investigated two-stage low pressure turbine is described and results of a nonlinear full-wheel time-domain simulation are presented, analyzed and compared with the available experimental data. Furthermore, recent improvements to the CFD solver TRACE are described in brief that lead to significantly reduced wall-clock times for such large scale simulations. The utilized models match those used in the Harmonic Balance (HB) based simulations that are presented in the third paper of this series, such that the full-wheel result can be utilized to validate the HB result. Transition, flow separation and wall pressure fluctuations on the stator blades of the second stage are analyzed in detail. A strong azimuthal Pi-periodicity is observed, manifesting in a significantly varying stability of the midspan trailing edge flow with a quasi-steady closed separation bubble on certain blades and highly dynamic partially open separation bubbles with recurring transition and turbulent reattachment on other blades. The energy spectrum of fluctuating wall quantities in that regime shows a high bandwidth and considerable disharmonic content, which is challenging for the HB based simulations.

On the Simulation of Unsteady Turbulence and Transition Effects in a Multistage Low Pressure Turbine: Part I — Verification and Validation

This is the first part of a series of three papers on the simulation of turbulence and transition effects in a multistage low pressure turbine. In this first part, the extension, verification and validation of a Harmonic Balance (HB) method recently proposed by the authors to fully include established turbulence and transition models in the method is presented. As an alternating frequency/time-domain type method the implemented HB solver has the advantage of being able to utilize models (e.g. boundary conditions or residual functions) formulated in either the frequency or time domain. On the one hand this allows highly accurate nonreflecting boundary conditions formulated in the frequency domain to be used along entry, exit or Interface boundaries, and on the other hand complex nonlinear terms formulated in the time domain to be used to describe nonlinear effects. Nevertheless, the wish to minimize the number of harmonics used to describe a given time periodic unsteady flow, coupled with the often highly nonlinear nature of turbulence and transition models makes the full inclusion of such models in the HB method challenging. In this work the integration of Menter’s SST two-equation k − w turbulence model along with Menter and Langtry’s two-equation gamma − ReTheta transition model in the context of a general framework for transport equations in the CFD solver TRACE is described in detail. Following the basic verification of the underlying transport equation framework, the implemented models are used to compute the well known high lift, low pressure turbine airfoil T106C and results are compared with the available experimental data as well as results from more conventional time-domain simulations. Alongside the basic validation of the models this testcase is furthermore used to investigate the importance of including higher harmonics, as opposed to only the zeroth harmonic, of the turbulence and transition models for the accurate prediction of the time-mean flow.