No abstract

Developments in numerical simulation of flows and high performance computing influence one another. More detailed simulation methods create a permanent need for more computational power, while new hardware developments often require changes to the software to exploit new hardware features.This dependency is very pronounced in the case of vector-units which are featured by all modern processors to increase their numerical throughput but require vectorization of the software to be used efficiently. We study the vectorization of a simulation method that exhibits an inherent level of vector-parallelism. This is of particular interest as SIMD operations will hopefully be available with std::simd in a future C++ standard.The simulation method considered here results in the simultaneous solution of multiple sparse linear systems of equations which only differ by their main diagonal and right hand sides. Such structure arises in the simulation of unsteady flow in turbomachinery by means of a frequency domain approach called harmonic balance.

Developments in numerical simulation of flows and high-performance computing influence one another. More detailed simulation methods create a permanent need for more computational power, while new hardware developments often require changes to the software to exploit new hardware features. This dependency is very pronounced in the case of vector-units which are featured by all modern processors to increase their numerical throughput but require vectorization of the software to be used efficiently. We study the vectorization of a simulation method that exhibits an inherent level of vector-parallelism. This is of particular interest as SIMD operations will hopefully be available with std::simd in a future C++ standard. The simulation method considered here results in the simultaneous solution of multiple sparse linear systems of equations which only differ by their main diagonal and right-hand sides. Such structure arises in the simulation of unsteady flow in turbomachinery by means of a frequency domain approach called harmonic balance.

The generation of data-driven turbulence models inherently requires the use of a sufficiently large database of high-fidelity reference data from DNS or LES. For technically relevant flows, such data is usually not readily available. However, in many cases there is a significant amount of experimental data available, though data points are mostly few and sparse. An approach which aims at deriving modelling errors by evaluating deviations from a given reference data set is the field inversion method proposed in [1]. Our aim is to verify this method as a tool which gives insight in the cause and nature of a given model’s inconsistencies. We therefore apply field inversion on the turbine cascade T106C for different reference data setups. We find, that for the investigated case, field inversion proved to qualitatively give the right hints towards the expected model correction, when only few data points were used as reference.

Being a comparably new technology, plasma actuation as a method to influence the flow in turbomachines has been investigated with a lot of effort in the past few years. Most of these investigations related to turbomachinery applications were carried out numerically using extensions to various RANS-codes and showed rather good improvements of the flow, particularly when using plasma actuators to suppress separations. Although most of these works gave a good impression of the capabilities of plasma actuators a comprehensive model of plasma actuation accounting for the impact of local static pressure, local temperature and the velocity of the flow passing the actuator was not presented yet. In this paper an extension to the proven and well validated RANS CFD code TRACE is presented which is able to model the influence of plasma actuation dependent on the local flow properties, i.e. dependent on the local static pressure, static temperature and Mach-number. This model extends the work of Lemire et al. [1] in a way that a comprehensive calibration based on open data is possible. In this paper, first the calibration is described followed by a validation in order to ensure that the model works correctly under different boundary conditions. The work is concluded by performing a numerical study applying a plasma actuator on different stator vanes along a generic high pressure compressor. Using this example the benefits of the extension to Lemire et al. [1], featuring the calibration for local static pressure, static temperature and Mach-number is explained and their impact on the potential of plasma actuation to suppress trailing edge separations is discussed.