Importance of Non-Equilibrium Modelling for Compressors

2021 ◽  
Author(s):  
Robert Spencer ◽  
Pawel Przytarski ◽  
Paolo Adami ◽  
Patrick Grothe ◽  
Andrew Wheeler

Abstract This paper investigates the importance of non-equilibrium boundary layer modelling for three compressor blade geometries, using RANS and high fidelity simulations. We find that capturing non-equilibrium effects in RANS is crucial to capturing the correct boundary-layer loss. This is because the production of turbulence within the non-equilibrium region affects both the loss generation in the non-equilibrium region, but also the final equilibrium state. We show that capturing the correct non-equilibrium behaviour is possible by adapting industry standard models (in this case the k-omega SST model). We show that for the range of cases studied here, non-equilibrium effects can modify the trailing-edge momentum thickness by up to 40 percent, and can change the trailing-edge shape factor from 1.8 to 2.1.

2016 ◽  
Vol 139 (3) ◽  
Author(s):  
Chaitanya Talnikar ◽  
Qiqi Wang ◽  
Gregory M. Laskowski

High-fidelity simulations, e.g., large eddy simulation (LES), are often needed for accurately predicting pressure losses due to wake mixing and boundary layer development in turbomachinery applications. An unsteady adjoint of high-fidelity simulations is useful for design optimization in such aerodynamic applications. In this paper, we present unsteady adjoint solutions using a large eddy simulation model for an inlet guide vane from von Karman Institute (VKI) using aerothermal objectives. The unsteady adjoint method is effective in capturing the gradient for a short time interval aerothermal objective, whereas the method provides diverging gradients for long time-averaged thermal objectives. As the boundary layer on the suction side near the trailing edge of the vane is turbulent, it poses a challenge for the adjoint solver. The chaotic dynamics cause the adjoint solution to diverge exponentially from the trailing edge region when solved backward in time. This results in the corruption of the sensitivities obtained from the adjoint solutions. An energy analysis of the unsteady compressible Navier–Stokes adjoint equations indicates that adding artificial viscosity to the adjoint equations can dissipate the adjoint energy while potentially maintaining the accuracy of the adjoint sensitivities. Analyzing the growth term of the adjoint energy provides a metric for identifying the regions in the flow where the adjoint term is diverging. Results for the vane obtained from simulations performed on the Titan supercomputer are demonstrated.


Author(s):  
Y. Dong ◽  
N. A. Cumpsty

This paper follows directly from Part I** by the same authors and describes measurements of the boundary layer on a supercritical-type compressor blade with wakes from a simulated moving upstream blade row convected through the passage. (The blades and the test facilities togehter with the background are described in Part I.) The results obtained with the wakes are compared to those with none for both low and high levels of inlet turbulence. The transition process and boundary layer development is very different in each case though the overall momentum thickness at the trailing edge is fairly similar. None of the models for transition is satisfactory when this is initiated by moving wakes.


1990 ◽  
Vol 112 (2) ◽  
pp. 231-240 ◽  
Author(s):  
Y. Dong ◽  
N. A. Cumpsty

This paper follows directly from Part 1 by the same authors and describes measurements of the boundary layer on a supercritical-type compressor blade with wakes from a simulated moving upstream blade row convected through the passage. (The blades and the test facilities together with the background are described in Part 1). The results obtained with the wakes are compared to those with none for both low and high levels of inlet turbulence. The transition process and boundary layer development are very different in each case, though the overall momentum thickness at the trailing edge is fairly similar. None of the models for transition is satisfactory when this is initiated by moving wakes.


2021 ◽  
Author(s):  
Qiang Liu ◽  
Will Ager ◽  
Cesare Hall ◽  
Andrew P. S. Wheeler

Abstract This paper investigates the surface boundary layer and wake development of a compressor blade at a range of low Reynolds number from 45000 to 120000. Experiments in a miniature linear compressor cascade facility have been performed with detailed surface pressure measurements and flow visualization to track variations in the separation bubble size. These have been combined with high resolution pneumatic pressure and hot wire probe traverses in the downstream wake. High fidelity DNS simulations have been completed on the same compressor blade section across the same range of operating conditions. The results show that large laminar separation bubbles exist on both blade surfaces. As Reynolds number increases, these separation bubbles shorten in length and reduce in thickness. Correspondingly, the downstream wake narrows, although the peak wake loss coefficient remains approximately constant. As the Reynolds number is increased from 45000 to 120000 the bubble length on the suction side reduced from 48% to 28% chord and on the pressure side reduced from 35% to 20% chord, while the loss coefficient reduced from 9% to 5%. The flow features are examined further within the high-fidelity computations, which reveal the dependence of the wake turbulence on the laminar separation bubbles. The separation bubbles are found to generate turbulent kinetic energy, which convects downstream to form the outer part of wake. As Re increases, a shorter bubble produces less turbulence in the outer part of the boundary layer leading to a narrower wake. However, the trailing edge separation is largely independent of Reynolds number, leading to the constant peak loss coefficient observed. The overall loss is shown to vary linearly with the total turbulence production, and this depends on the size of the separation bubbles. Overall, this research provides new insight into the connection between the blade surface flow field and the wake characteristics at low Reynolds number. The findings suggest that changes that minimize the extent of the blade separation bubbles could provide significant improvements to both the steady and unsteady properties of the wake.


Author(s):  
Chaitanya Talnikar ◽  
Qiqi Wang ◽  
Gregory M. Laskowski

High fidelity simulations, e.g., large eddy simulation are often needed for accurately predicting pressure losses due to wake mixing and boundary layer development in turbomachinery applications. An unsteady adjoint of high fidelity simulations is useful for design optimization in such aerodynamic applications. In this paper we present unsteady adjoint solutions using a large eddy simulation model for a vane from VKI using aerothermal objectives. The unsteady adjoint method is effective in capturing the gradient for a short time interval aerothermal objective, whereas the method provides diverging gradients for long time-averaged thermal objectives. As the boundary layer on the suction side near the trailing edge of the vane is turbulent, it poses a challenge for the adjoint solver. The chaotic dynamics cause the adjoint solution to diverge exponentially from the trailing edge region when solved backwards in time. This results in the corruption of the sensitivities obtained from the adjoint solutions. An energy analysis of the unsteady compressible Navier-Stokes adjoint equations indicates that adding artificial viscosity to the adjoint equations can dissipate the adjoint energy while potentially maintain the accuracy of the adjoint sensitivities. Analyzing the growth term of the adjoint energy provides a metric for identifying the regions in the flow where the adjoint term is diverging. Results for the vane from simulations performed on the Titan supercomputer are demonstrated.


Author(s):  
Marion Mack ◽  
Roland Brachmanski ◽  
Reinhard Niehuis

The performance of the low pressure turbine (LPT) can vary appreciably, because this component operates under a wide range of Reynolds numbers. At higher Reynolds numbers, mid and aft loaded profiles have the advantage that transition of suction side boundary layer happens further downstream than at front loaded profiles, resulting in lower profile loss. At lower Reynolds numbers, aft loading of the blade can mean that if a suction side separation exists, it may remain open up to the trailing edge. This is especially the case when blade lift is increased via increased pitch to chord ratio. There is a trend in research towards exploring the effect of coupling boundary layer control with highly loaded turbine blades, in order to maximize performance over the full relevant Reynolds number range. In an earlier work, pulsed blowing with fluidic oscillators was shown to be effective in reducing the extent of the separated flow region and to significantly decrease the profile losses caused by separation over a wide range of Reynolds numbers. These experiments were carried out in the High-Speed Cascade Wind Tunnel of the German Federal Armed Forces University Munich, Germany, which allows to capture the effects of pulsed blowing at engine relevant conditions. The assumed control mechanism was the triggering of boundary layer transition by excitation of the Tollmien-Schlichting waves. The current work aims to gain further insight into the effects of pulsed blowing. It investigates the effect of a highly efficient configuration of pulsed blowing at a frequency of 9.5 kHz on the boundary layer at a Reynolds number of 70000 and exit Mach number of 0.6. The boundary layer profiles were measured at five positions between peak Mach number and the trailing edge with hot wire anemometry and pneumatic probes. Experiments were conducted with and without actuation under steady as well as periodically unsteady inflow conditions. The results show the development of the boundary layer and its interaction with incoming wakes. It is shown that pulsed blowing accelerates transition over the separation bubble and drastically reduces the boundary layer thickness.


Sign in / Sign up

Export Citation Format

Share Document