No abstract

The regeneration cycle of streaks and streamwise vortices plays a central role in the sustainment of near-wall turbulence. In particular, the streak breakdown phase in the regeneration cycle is the core process in the formation of the streamwise vortices, but its current understanding is limited particularly in a real turbulent environment. This study is aimed at gaining fundamental insight into the underlying physical mechanism of the streak breakdown in the presence of background turbulent fluctuation. We perform a numerical experiment based on direct numerical simulation, in which streaks are artificially generated by a body forcing computed from previous linear theory. Upon increasing the forcing amplitude, the artificially driven streaks are found to generate an intense fluctuation of the wall-normal and spanwise velocities in a fairly large range of amplitudes. This cross-streamwise velocity fluctuation shows its maximum at λ + x ≈ 200 − 300 (λ + x is the inner-scaled streamwise wavelength), but it only appears for λ + x 3000 − 4000. Further examination with dynamic mode decomposition reveals that the related flow field is composed of sinuous meandering motion of the driven streaks and alternating cross-streamwise velocity structures, clearly reminiscent of sinuousmode streak instability found in previous studies. Finally, it is shown that these structures are reasonably well aligned along the critical layer of the secondary instability, indicating that the surrounding turbulence does not significantly modify the inviscid inflectional mechanism of the streak breakdown via streak instability and/or streak transient growth.

The high order spectral/hp element methods implemented in the software framework Nektar++ are investigated for scale-resolving simulations of LPT profiles. There is a growing demand for high fidelity methods for turbomachinery to move towards numerical “experiments”. The study contributes at building best practices for the use of emerging high fidelity spectral element methods in turbomachinery predictions, with focus on the numerical details that are specific of these classes of methods. For this reason, the T106A cascade is used as a base reference application because of availability of data from previous investigations. The effects of polynomial order (p-refinement), spanwise domain extent and spanwise Fourier planes are considered, looking at flow statistics, convergence and sensitivity of the results. The performance of the high order spectral/hp element method is also assessed through validation against experimental data at moderately high Reynolds number. Thanks to the reduced computational cost, the proposed methods will have a strong impact in turbomachinery, paving the way to its use for design purposes and also allowing for a deeper understanding of the flow physics.

The recent development and increasing integration of high performance computing, scale resolving CFD and high order unstructured methods offers a potential opportunity to deliver a simulation-based capability (i.e. virtual) for aerodynamic research, analysis and design of industrial relevant problems in the near future. In particular, the tendency towards high order spectral/hp element methods is motivated by their desirable dispersion-diffusion properties, that are combined to accuracy and flexibility for complex geometries. Previous work from the Authors focused on developing guidelines for the use of these methods as a virtual cascade for turbomachinery applications. Building on such experiments, the present contribution analyzes the performance of a representative industrial cascade at moderate Reynolds number with various levels and types of inflow disturbances, adopting the incompressible Navier-Stokes solver implemented in the Nektar++ software framework. The introduction of a steady/unsteady spanwise-nonuniform momentum forcing in the leading edge region was tested, to break the flow symmetry upstream of the blade and investigate the change in transition mechanism in the aft portion of the suction surface. To provide a systematic synthetic turbulence generation tool, a parallelised version of Davidson’s method is incorporated and applied for the first time in the software framework to a low pressure turbine vane. The clean results of the cascade are compared to various levels of momentum forcing and inflow turbulence, looking at blade wall distributions, wake profiles and boundary layer parameters. Low levels of background disturbances are found to improve the agreement with experimental data. The results support the confidence for using high order spectral methods as a standalone performance analysis tool but, at the same time, underline the sensitivity at these flow regimes to disturbances or instabilities in the real environment when comparing to rig data.

In this paper, we investigate the feasibility of using DNS data and machine learning algorithms to assist RANS turbulence model development. High-fidelity DNS data are generated with the incompressible Navier–Stokes solver implemented in the spectral/hp element software framework Nektar++. Two test cases are considered: a turbulent channel flow and a stationary serpentine passage, representative of internal turbo-machinery cooling flow. The Python framework TensorFlow is chosen to train neural networks in order to address the known limitations of the Boussinesq approximation and a clustering based on flow features is run upfront to enable training on selected areas. The resulting models are implemented in the Rolls-Royce solver HYDRA and a posteriori predictions of velocity field and wall shear stress are compared to baseline RANS. The paper presents the fundamental elements of procedure applied, including a brief description of the tools and methods and improvements achieved.