1,720,973 research outputs found
Flow and heat transfer mechanism for gas turbine internal cooling: DNS & LES study
No full textGas turbine are used for aircraft propulsion and in land-based power generation or industrial application. One of the key to increase turbine efficiency (or to reduce fuel consumption) is to increase turbine inlet temperature (TIT). It is crucial to design an efficient cooling system for turbine blade, to avoid penalties in the overall turbine efficiency. There is still an unsatisfactory comprehension of the mechanisms of turbulence and heat transfer in those configurations, considering also that, most of the times, they are designed using simple empirical correlations. The objective of this thesis is to analyse flow configurations representing the most complex parts of a turbine cooling system, using high-fidelity approaches (i.e. DNS and LES). These approaches are unique tools for a clear understanding of the physics behind the cooling process, but are unlikely to be appealing in the following years for industries, since they require large computational power and a lot of time to be performed (usually months). Increasing computing power does not necessarily make DNS or LES attractive, since some other issues may arise (e.g. the high storage required for the simulated data). Starting from the above considerations, URANS models, tailored to account for rotation and non linear turbulent flow features, were developed and validated in this research, to overcome all the issues related to DNS and LES
Effect of wall curvature of an asymmetric subsonic impinging jet: DNS study
No full textWe report on Direct Numerical Simulation (DNS) of a subsonic confined jet impinging on a curved surface. The configuration resembles a turbine leading edge cooling system. The bulk Reynolds number (based on bulk velocity Ub and jet diameter D) is 3300. The impinging wall was kept at a constant temperature higher than the jet bulk temperature. Comparisons with the jet impinging on a flat plate are carried out.
Our major findings are that the turbulent flow field is affected by the shape of the impinging wall, leading to an asymmetric development of turbulent structures. The frequencies which dominates the excess heat transfer mechanism change with respect to the reference case, changing the average heat transfer into the hot wall
Effects of rotation on flow in an asymmetric rib-roughened duct: LES study
No full textWe report on large-eddy simulations (LES) of fully-developed flow in a duct of a rectangular
cross-section in which square-sectioned, equally-spaced ribs oriented perpendicular to the flow
direction, were mounted on one of the walls. The Reynolds number, based on hydraulic di-ameter Dh,
bulk velocity U0 and air properties, is 15,000. The duct was subjected to clock-wise (stabilizing) and
anti-clock-wise (destabilising) rotation, at a moderate Rotational number, Ro=0.3. The LES results
showed good agreement with the experimental results of Coletti et al. (2011) available in the
mid-span plane. We analyzed the effects of stabilizing and destabilizing rotation on the flow, its vortical
structures and turbulence statistics by comparing results of the rotating configurations with the nonrotating
case with a focus on the im-portance of the secondary motions and the scales of the turbulent
structures. To determine the existence of very long structures, we considered two different domains: a
single rib spacing, confirmed in the experiment to be fully periodic after the 6th rib and a double
domain extend-ing over two successive ribs. Comparisons of the two cases in terms of turbulence
statistics and vorticity field confirmed no significant differences. The modifications of the turbulence
intensity were interpreted on the basis of the rotation-induced production terms in the Reyn-olds
stress equation
LES analysis of flow and heat transfer in a rib-roughened duct in clockwise and anti-clockwise rotation regimes
No full textWe report on an LES study of effects of stabilising/destabilising rotation on heat transfer over a ribbed surface in a rectangular duct at Re = 15000. The duct bottom wall, ribbed by flow-normal, equally-spaced square-sectioned ribs, was uniformly heated (except for the ribs) by a constant heat flux. The duct was rotated with angular velocity corresponding to the rotation number of 0.3, around an axis parallel to the ribs in counterclockwise (clockwise) direction, thus destabilising (stabilising) the ribbed-wall adjacent flow. These well-resolved LES gave some new insight into the rotation effects on flow and heat transfer providing information that are not easily accessible by experiments. An attempt was made to identify the heat transfer effects due to the rotation-induced modifications of the secondary motion
URANS study of flow and heat transfer in a rotating rib-roughened internal cooling channel
No full textThe flow field and heat transfer in a rotating ribbed duct is investigated by means of U-RANS. The duct has an almost
square cross-section. Square-sectioned, equally spaced ribs, oriented perpendicular to the flow direction, are mounted
on one of the walls. The bulk Reynolds number is 15,000 and the Prandtl number 0.7. This configuration is analysed
in a rotating configuration aiming increasing turbulence on the ribbed surface (i.e. destabilizing configuration).
Comparisons with non-rotating configuration are carried out to investigate influence of Coriolis Force. Such
configuration mimics the flow in internal cooling channels in the high pressure (HP) rotor and vane blade. The bottom
walls are uniformly heated, except the rib walls. The turbulent flow field is modelled by the elliptic-relaxation-type k-
ε-ζ-f model. This model was modified by the authors, to take in account the turbulence anisotropy derived by the
Coriolis force through the sensitizing of turbulent viscosity to angular velocity vector.
Results demonstrate heat transfer enhancement due to the rib induced secondary flows. The ribs cause an increase of
turbulence, thus increasing heat transfer. In addition to it, rotation promotes turbulence close to the heated wall.
Consequently, the heat transfer is found to be higher in the rotating case than in a non-rotating case.
Secondary motions play an important role in removing hot fluid from the ribbed surface, enhancing mixing of hot and
cold fluid. The Taylor-Görtler vortices produced by the Coriolis force bring hot gas far from the heated wall to the
centre of the channel, increasing Nusselt number
DNS study of fusion reactor dust particle mobilization induced by a transonic jet incoming in a vacuum container
No full textThe flow induced motion of wall deposited particles
is highly linked with the instantaneous fluid structures.
Here we perform a two-phase flow DNS to analyze
the resuspension of solid particles from a surface
hit by a transonic jet into a low pressure container
in conditions similar to those which occur in a fusion
reactor vacuum vessel during a Loss of Vacuum Accident
(LOVA). The initial condition of pressure and
temperature were set to 49.5 mbar and 373 K, with
a Reynolds number of 3300 on a 512 512 512
grid properly refined in regions where high gradients
are present. The Thornton and Ning impact/adhesion
model is adopted, whereas an advanced resuspension
model, which also takes into account the dynamics
(rolling and sliding) of particles at the wall, is here
implemented. The initial deposited particles follow a
log-normal distribution with a count median diameter
of 2.21 μm, geometric standard deviation of 2.93 and
constant density of 8527 kg/m3. It has been found that
the resuspension phenomenon mostly affect particles
of the biggest diameters. Moreover, the jet-deposit interaction
is for the most part confined within a circumference
around the jet of radius equal to the jet diamete
LES of heat transfer in an asymmetric rib-roughened duct: influence of rotation
No full textWe report on an LES study of effects of destabilising
rotation on heat transfer over a ribbed surface in
a rectangular duct at Re = 15000. The duct bottom
wall, ribbed by flow-normal, equally-spaced squaresectioned
ribs, was uniformly heated (except for the
ribs) by a constant heat flux. The duct was rotated with
angular velocity corresponding to the rotation number
of 0.3, around an axis parallel to the ribs in counterclockwise
direction destabilising the ribbed-wall adjacent
flow.
These well-resolved LES gave some new insight
into the rotation effects on flow and heat transfer providing
information that are not easily accessible to experiments.
An attempt was made to identify the heat
transfer effects due to the rotation-induced modifications
of the secondary motion, and the direct effects on
the turbulence statistics, especially the budgets of the
temperature variance and turbulent heat flux. It turned
out that the former is predominant in the recirculation
zone, whereas the latter prevails just after it
LES scrutiny of non-linear k-eps-zeta-f model sensitized to rotation
No full textWe scrutinize an updated version of the non-linear (quadratic) k-ε-ζ-f aiming at sensitizing the model to the effect of rotation. This objective was obtained by imposing that C coefficient depends on the strain and vorticity tensors, the latter explicitly including solid body rotation. The model was tested on plane channel and square-sectioned duct flows. Results are assessed against DNS literature data and properly developed LES computations. We demonstrate that, when considering the channel flows, the developed formulation is able to accurately reproduce flow and turbulent variables at various angular velocity regimes. Good predictions are also obtained for the duct flow, where the flow is subjected to the mutual influence of rotation and near-wall turbulence anisotropy. In particular, the non-linear rotation-sensitized model is able to reproduce the near-wall turbulent kinetic energy distribution close to the suction side, returning a zero value in the mid-span and a small peak close to the vertex on the suction side. Budgets analysis of turbulent kinetic energy demonstrates that the proposed model is able to properly reproduce any of the terms in the k-equatio
- …
