See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/259976017

Aerothermodynamics of tight rotor tip

clearance flows in high-speed unshrouded
turbines
Article in Applied Thermal Engineering April 2014
DOI: 10.1016/j.applthermaleng.2014.01.015

CITATIONS

READS

134

4 authors, including:
Cis De Maesschalck

Sergio Lavagnoli

Purdue University - von Karman Institute fo

von Karman Institute for Fluid Dynamics

16 PUBLICATIONS 33 CITATIONS

35 PUBLICATIONS 72 CITATIONS

SEE PROFILE

SEE PROFILE

Guillermo Paniagua
Purdue University
189 PUBLICATIONS 573 CITATIONS
SEE PROFILE

All in-text references underlined in blue are linked to publications on ResearchGate,

letting you access and read them immediately.

Available from: Guillermo Paniagua

Retrieved on: 25 September 2016

Applied Thermal Engineering 65 (2014) 343e351

Contents lists available at ScienceDirect

Applied Thermal Engineering

journal homepage: www.elsevier.com/locate/apthermeng

Aerothermodynamics of tight rotor tip clearance ows in high-speed

unshrouded turbinesq
C. De Maesschalck a, b, *, S. Lavagnoli a, G. Paniagua a, N. Vinha a
a
b

Turbomachinery and Propulsion Department, von Karman Institute, Chausse de Waterloo 72, 1640 Rhode-Saint-Gense, Belgium
Faculty of Applied Sciences, Department of Mechanical Engineering, Vrije Universiteit Brussel, Triomaan 43, B-1050 Brussels, Belgium

h i g h l i g h t s

Publication on the thermal effects of tight gaps in rotating uid machinery.

Tight clearances (0.1% of the airfoil height) revealed novel reversed ow topology.

Heavily altered thermal eld due to increased importance of the viscous effects.

Identied optimal clearance height for suction side heat load.

Accurate prediction of the adiabatic efciency through isothermal simulations.

a r t i c l e i n f o

a b s t r a c t

Article history:
Received 26 August 2013
Accepted 11 January 2014
Available online 23 January 2014

The inevitable clearance between stationary and rotating parts in any uid machinery gives rise to
leakage ows, which strongly affect the overall performance of the machine. In modern gas turbine
engines, the existing gap between the rotor airfoil tip and the shroud is responsible for about one third of
the total aerodynamic losses. Additionally, this leakage ow induces erce unsteady heat loads onto the
rotor casing and provokes signicant thermal stresses at the airfoil tip. One can attempt to curtail these
detrimental effects by running tight clearances; however, the meager number of publications on this
topic presents an obstacle to exploiting the design opportunities.
This paper presents the outcome of an extensive numerical investigation of a high pressure turbine
stage operating at engine-representative non-dimensional parameters (Reynolds and Mach number,
temperature ratios). RANS calculations were performed using the Numeca FINE/Turbo suite, adopting the
keu SST turbulence model to investigate the aerodynamic and heat transfer characteristics in the tip
region. Five clearances, ranging from 0.1% to 1.9% of the rotor channel height, were simulated at adiabatic
and isothermal (Ttotal,in/Tw 1.57) conditions. The detailed ow analysis revealed an unexpected aerodynamic ow topology at tight clearances (h/H < 0.5%), characterized by a reverse ow over a signicant
part of the tip gap region. The heat transfer on the airfoil tip, shroud and near-tip regions was examined
in detail, with emphasis on the different driving phenomena. This elaborate numerical study provides a
deeper insight into the complex aerothermal physics of leakage ows occurring for tight clearances in a
high-speed environment relevant to any uid machinery design and analysis.
2014 The Authors. Published by Elsevier Ltd. All rights reserved.

Keywords:
Turbine aerothermodynamics
Tight clearances
Heat transfer
Tip leakage ows
Rotor gap
Transonic

1. Introduction

q This is an open-access article distributed under the terms of the Creative

Commons Attribution-NonCommercial-ShareAlike License, which permits noncommercial use, distribution, and reproduction in any medium, provided the
original author and source are credited.
* Corresponding author. Turbomachinery and Propulsion Department, von Karman Institute, Chausse de Waterloo 72, 1640 Rhode-Saint-Gense, Belgium.
Tel.: 32 23599768; fax: 32 23599600.
E-mail address: cis.demaesschalck@vki.ac.be (C. De Maesschalck).

The tight clearance in between two mechanical components

offers various aerodynamic advantages to regulate mass ows
(such as control valves) or to enhance the mixing effects (e.g. atomizers). In turn, heat transfer aspects within such narrow passages are vital to develop novel miniaturized heat sinks or to assess
the thermal phenomena inside slotted coaxial rotating shafts [1,2].
Leakage ows, encountered in many mechanical machines, are the
prime source of aerodynamic losses with a consequent performance drop (e.g. reciprocating engines [3]) and may provoke severe

1359-4311/$ e see front matter 2014 The Authors. Published by Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.applthermaleng.2014.01.015

344

C. De Maesschalck et al. / Applied Thermal Engineering 65 (2014) 343e351

thermal loads onto the side surfaces [4]. The abatement of the
detrimental effects demands for enhanced sealing methods and
compels to run at tighter gap sizes without deteriorating the mechanical integrity and reliability of the machine.
In gas turbines, the pressure difference between both sides of
the rotor blade forces a portion of the ow to travel over the tip
through the clearance gap, giving rise to a leakage jet and a leakage
vortex. The cross ow inside the tip gap generates erce unsteady
heat loads onto the rotor casing and severe thermal stresses in the
blade tip; consequently, the aerothermal issues have been the
subject of many noteworthy experimental and numerical investigations in the past [5]. Several tip design approaches have
been proposed [6] to mitigate tip leakage effects at the behest of the
turbine durability [7] and aerothermal performance improvement.
Squealer tips adopting rims are proven to enhance the aerodynamic
performance [8] and reduce the average heat loads [9] compared to
a plain tip. To further increase the efciency, a winglet shaped tip
[10] can be considered as a suitable alternative. Due to the mechanical simplicity, robustness, ease of cooling and machinability,
tight running at tips are desirable solutions to turbine manufacturers. Bunker [11] presented an extensive review of the research,
mainly performed in low-speed environments, on the aerothermal
performance of the plain turbine tips in a high pressure environment. Subsonic tip gap ows are characterized by the formation of a
recirculation bubble followed by a ow reattachment onto the
blade tip. Creating thereafter a mixing zone up to the suction side of
the blade, this leakage ow is mainly dominated by the pressure
loading across the tip and the maximum channel ow contraction
over the bubble.
Current turbine designs tend to adopt higher stage loadings and
rotational speeds which result in a supersonic overtip leakage ow.
Oblique shock waves originate at the reattachment point and are
reected between the rotor tip and casing [12], setting an effective
limiter on the leakage mass ow [13]. Enhanced heat transfer levels
are observed at the reattachment line on the tip pressure side [14]
and on the rotor shroud, just above the pressure side edge of the
airfoil [15]. Zhang et al. [16] observed parallel lines of high heat
transfer due to the shock impingement. Additionally, an increase in
clearance height was found to give rise to increased leakage ow

velocities in the overtip region, resulting in an enhanced gap mass

ow and a leakage vortex core further detached from the airfoil
suction side.
To the authors knowledge, there are no publications addressing
the physics of overtip gap ows for clearances below 0.5% of the
turbine passage height. The current paper presents the results of an
elaborate numerical study of a high-speed turbine covering tight
tip clearances as low as 0.1% of the channel height. The aerothermal
eld investigation revealed remarkable changes between tight and
large running clearances. This research thus yields insight into the
complex ow phenomena occurring in the overtip region. The
description of the physics, needed for the future design of tight
running uid machinery, will contribute to reducing the machine
performance loss and therefore the engines impact on the environment will be minimized.
2. Numerical methodology
2.1. Computational domain and mesh
Fig. 1(a) presents the cross section of the investigated turbine
stage, containing 42 stator vanes and 64 rotor blades. Such geometry was selected as to be representative of modern gas turbine
technology for aero-engines and power generation. In order to
allow a sufcient development of the ow, the outlet of the
computational domain was placed 2 chord lengths downstream of
the trailing edge of the rotor blade. The stationary stator passage
and rotating rotor domain are connected through a mixing plane.
Only one periodic section of the turbine stage is modeled (i.e. 1/42
of the stator annulus and 1/64 of the rotor annulus). A at rotor tip
geometry was considered along with ve different clearances, h,
between the rotor tip and the stationary shroud, ranging from 0.1%
to 1.9% of the rotor channel height, H. The relevant geometrical
parameters are summarized in Table 1.
Fig. 1 additionally displays the structured grids created with
Numeca AutoGrid5, dividing radially the stator vane and rotor
blade computational domain into 61 and 80 ow paths respectively. In order to be able to resolve the detailed gap ow characteristics, the clearance height was further divided radially into 33

Fig. 1. Turbine stage geometry and computational mesh: (a) meridional view and mesh at the mid-height meridional surface; (b) rotor shroud mesh; (c) rotor blade mesh, including
detail of the tip gap mesh.

C. De Maesschalck et al. / Applied Thermal Engineering 65 (2014) 343e351

Table 1
Summary of the turbine geometry and mesh parameters.
Geometrical parameters
Stator

Mesh parameters
0.1, 0.2, 0.4, 1.9%a
0.9%a

Rotor

No. of blades
42
64
Stator divisionsb
Inlet channel
50.7
51.97 Rotor divisionsb
height (H) [mm]
Aspect ratio (H/C)
0.704 1.053 No. of cells
a
b

61
80
33 (gap)
2.68  106

61
80
53 (gap)
3.1  106

% h/H.
In the radial direction.

sections (53 for the largest clearance). In the surface at a constant

radius, an O4H multiblock topology was used, consisting of ve
blocks: an O-mesh block surrounding the blade and captured inside
four H-mesh blocks. The mesh inside the tip gap is generated using
a buttery topology (an H-block surrounded by an O-block), as
visualized in Fig. 1(c). The nal mesh contains 16 blocks and guarantees y values between 0.05 and 0.3 in the tip region.
A grid sensitivity analysis was performed by doubling the
amount of radial divisions inside the clearance and halving the rst
cell thickness inside the boundary layer. Further renement of the
mesh did not alter the aerothermal results in the overtip region,
and hence, a nal grid of approximately 2.7e3.1 million cells was
selected with the parameters listed in Table 1-right.
2.2. Solver parameters and evaluated cases
Steady RANS calculations were performed using the Numeca
FINE/Turbo suite 8.9.2, adopting the keu SST turbulence model. The
computations allow simulation of the actual turbine conditions in
terms of Reynolds and Mach numbers [19]. An axial airow was
assumed at the inlet of the turbine, imposing the total pressure,
Ptotal, and temperature, Ttotal, together with a turbulence intensity
of 5% [20]. At the outlet of the computational domain, the static
pressure, Ps, was set in the mid-height of the channel (imposing
radial equilibrium) to give a total-to-static pressure ratio of 2.93.
The turbine rotor blade and its hub section are rotating at 6790
RPM. On the sidewalls, periodic boundary conditions ensure that
the simulation of one sector is representative of the whole annulus.
To mimic the heat transfer characteristics in a true turbine environment [21,22], the gas-to-wall temperature ratio Ttotal,in/Tw was
set to 1.57. For each evaluated rotor clearance, an adiabatic and an
isothermal computation were performed. Table 2 summarizes the
turbine operating conditions.
3. Identication of a novel overtip ow topology
3.1. Aerodynamics of a reversed tip ow
Fig. 2 shows the streamlines at the mid-gap together with the
tip surface pressure contours for the different clearances,
Table 2
Turbine operating conditions.
Turbine conditions
Ptotal,in/Ps,exit
Ttotal,in/Tw

ain

Inlet turbulence intensity

Rotational speed
Mtip peripheral Macha
Mid-gap Reb
a
b

Based on aexit.
Isothermal, based on W and h.

2.93
1.57
0
5
6790
0.78
350e14,000

e
e
deg.
%
RPM
e
e

345

decreasing from left (h/H 1.9%) to right (h/H 0.1%). For the two
largest clearances (i.e. 0.9% and 1.9% of the blade span), the tip
leakage ow shows a conventional ow pattern with streamlines
crossing the blade tip region from pressure to suction side, driven
by the differential blade loading across the tip gap. For such
clearances, the ow angle across the gap can be predicted using the
correlation developed by Heyes and Hodson [23]. However, at
clearances lower than 0.9% a striking feature is observed: a part of
the ow enters the gap from the suction side and bends back to
leave the gap again at the suction side. For the two smallest
clearances, the ow direction over the front part of the blade prole
is fully reversed, even though the driving pressure gradient is still
directed in the opposite direction, from the pressure to the suction
side.
This novel ow behavior is induced by the effect of the casing
proximity which becomes signicant only at very tight clearances.
The uid near the casing has a quasi-zero velocity in the absolute
frame of reference due to the presence of the inlet velocity
boundary layer. In the rotating frame this corresponds to a relative
velocity nearly equal in magnitude to the rotor tips absolute peripheral velocity, but opposite in direction, i.e. from the blade
suction side to the pressure side. Fig. 3(a) shows the tangentially
(mass ow) averaged aerodynamic and thermal boundary layer
proles close to the casing endwall, 10% Cax,r upstream the rotor LE.
These proles were hardly affected by the gap size; the maximum
variation in temperature and velocity between the lowest and
highest clearance remains below 1% at the considered radial positions. Therefore, only the radial proles for the largest gap are
displayed in the gure. The ow exits the stator vane with a total
absolute velocity V to which corresponds a tangential component
Vt. As the ow approaches the rotating blade, the uid heading
towards the tip gap is subjected to a momentum component
generated by the relative movement of the casing endwall. The
casing applies a force on a uid element in the tip channel along the
peripheral direction and so proportional to the rotor speed. The
relative movement of the casing wall therefore acts in opposition to
the airfoil rotation which causes the uid to cross the overtip gap
from the pressure towards the suction side. The casing effect on the
clearance ow becomes more signicant as the tip gap gets tighter.
Additionally, Fig. 3(a) shows that at a clearance of about
h/H 0.55%, the absolute tangential velocity approaching the rotor
leading edge is balancing the peripheral velocity u. As the tip
clearance grows beyond that, the tangential velocity of the uid
entering the rotor gap becomes higher than the rotor peripheral
speed. Hence, the relative movement of the rotor casing is not
sufcient to counteract the effect of the blade loading which pushes
uid over the tip from the pressure side to the suction side. This
case describes the physics of tip leakage ows at relatively large
tip gaps, as commonly found in gas turbine engines and reported on
in scientic publications.
However, when the tangential ow momentum at the rotor tip
gap inlet and the blade tip peripheral speed are of comparable
magnitude (the ratio Vt/u is close to unity), parts of the overtip gap
ow cannot travel across the entire airfoil thickness. Hence, the
uid enters the gap suction side, moves along the channel and exits
again at the airfoil suction side further downstream. This scenario
has been observed for the investigated tip clearances smaller than
0.5% of the channel height, as shown in Fig. 2. This ow pattern is
particularly evident near the leading edge of the rotor blade, where
the pressure differential between the pressure and the suction side
is weak.
As the clearance is reduced further below 0.2%, the uid above
almost the whole blade tip surface undergoes fully reversed ow
(Vt/u < 1). An explanatory sketch representing the three distinct
overtip ow topologies is provided by Fig. 3(b). The momentum

346

C. De Maesschalck et al. / Applied Thermal Engineering 65 (2014) 343e351

Fig. 2. Representation of the streamlines entering the tip region in the middle of the gap and static pressure distribution on the tip surface.

thickness q of the airfoil inlet boundary layer (0.55% of the blade

span) can also be used to estimate the clearance size for which
reverse ow starts to appear on the tip surface. For large clearances
(h/q > 1), the uid ingested by the rotating gap has enough
momentum to cross the tip channel. When the gap gets smaller
(h/q < 1), reverse ow over the rotating tip surface is observed.
In dry air (Pr z 0.7) the thermal boundary layer has a comparable extent as the aerodynamic boundary layer and hence, in the
case of tight clearances, both the velocity and the thermal eld are
heavily inuenced by the characteristics of the incoming boundary
layer outside the gap. Fig. 4 displays the distribution along the
radial direction of the ow velocity in the rotor frame of reference
within the gap, extracted at a location K, taken at 40% of the
distance from the suction to pressure side. The velocity in the rotor
frame of reference was decoupled into a pressure driven component along the machine axis (Waxial, Fig. 4(a)), and a component in
the tangential direction (Wt, Fig. 4(b)) which includes the effect of
the casing relative movement (u 280 m/s). The smaller the
clearance is, the smaller the (signed) tangential component of the
velocity vector becomes, and thus the inuence of the casing increases. Nevertheless, a signicant portion of the ow presents a
negative tangential component along the radial direction, which
moves from the suction towards the pressure side of the blade
(reverse ow).
3.2. Thermal eld modication for tight clearances
The aerodynamic ow pattern of tight clearances has a signicant effect on the thermal eld. The coupling between the aerodynamic and the thermal eld was analyzed using the energy
equation that expresses the stagnation enthalpy of a uid particle

per unit time and mass along a streamline. In Equation (1), let us
consider as reference the stationary frame. The rst term in the
right-hand side, vQi/vxi, expresses the inuence of the heat transfer.
The second one, v(sijvi)/vxj, concerns the viscous effects, and the
third part, vPs/vt, represents the impact of the static pressure.
Indeed, the stagnation enthalpy of an isentropic uid is modulated
by the temporal variation of the static pressure. In this case, a xed
point in the stationary frame, e.g. on the casing, will rst sense the
pressure eld from the rotor tip SS and then from the PS. One can
get an estimation of the contribution of the pressure term to the
stagnation enthalpy by monitoring the temporal pressure variation
(in the xed frame of reference) at the different locations along a
streamline, or similarly, on the absolute total temperature (generally negative in the blade passage indicating the turbine work
extraction, but positive in the tip region). It is exactly this inherent
unsteadiness of the pressure eld which is at the base of the work
extraction in rotating uid machinery [17].


Dhtotal
1 vQi 1 v sij vi
1 vPs

r vxi r vxj
r vt
Dt

(1)

The second term in the right-hand-side of Equation (1) represents the viscous work done on or by the uid. In the casing frame
of reference, the shroud cannot perform any work. By contrast, near
the blade, the tip movement contributes to the work process. Previous investigations did not elaborate on this term, arguing that
since the tip moves with a velocity less than the uid, the inuence
on the total temperature would be negative. However, in modern
turbines operating at high peripheral speeds and tight clearances,
the effect of the moving wall may represent a signicant contribution to the temperature eld. Since a considerable portion of the
ow is traveling in the rotor frame of reference from the suction to

Fig. 3. (a) Aerodynamic and thermal boundary layer 10% Cax,r upstream of the rotor LE; (b) effect of clearance height on overtip ow topology.

C. De Maesschalck et al. / Applied Thermal Engineering 65 (2014) 343e351

347

Fig. 4. Radial distribution, within the tip gap, of the axial and tangential components of the ow velocity in the rotor frame of reference for the investigated clearance heights.

the pressure side, in the opposite direction to the rotational speed,

the tip exerts work on the uid, contributing to an increase of the
total ow temperature. Fig. 5 presents the contours of static pressure and stagnation temperatures in the middle of the gap for three
different clearances (h/H 1.9%, 0.2%, and 0.1%), considering two
different solid boundary conditions in the tip region. It should be
noted that the pressure eld is barely affected by a change in the
thermal solid boundary condition. The rst row shows the mid-gap
static pressure distribution together with the streamlines. The
second row shows the absolute total temperature considering
adiabatic walls for the thermal boundary condition. In this case the
total temperature can only be altered through viscous work or by
means of a temporal pressure variation (Eq. (1)). For a fully reversed
ow, streamline a for h/H 0.2%, both the temporal pressure
variation and the viscous work contribute to a rise in the absolute
total temperature. The streamline labeled b enters the gap region
through the suction side and exits further downstream near location c, resulting in a large rise in total temperature due to this
combined effect. While this rise is absent for clearances greater

than 0.5% of the blade span, tighter clearances, on the contrary,

show an important effect on the total temperature. For the case
where h/H 0.1%, even total temperatures above 1000 K can be
observed.
The third row in Fig. 5 shows the absolute total temperature in
the case of isothermal walls at 300 K. In this case, all three components on the right-hand side of Equation (1) are non-zero. One
can conclude that the thermal balance at the airfoil tip region is
largely dominated by heat transfer at the uid-casing interface in
the most conned gaps (i.e. h/H 0.2% and 0.1%), rather than by the
effect of the viscous and pressure terms. This results in a total
temperature eld dominated by the imposed wall temperature
distribution.
The fourth row shows, for the adiabatic case, the total temperature variations in the rotor frame of reference. Since the static
pressure can be seen as steady in the relative frame of reference, the
variation in temperature can only be attributed to the viscous effects in the relative frame of reference. For the highest clearance
this total temperature remains fairly constant, indicating almost no

Fig. 5. Ps and Ttotal eld in the mid-gap region for different tip clearances and boundary conditions.

348

C. De Maesschalck et al. / Applied Thermal Engineering 65 (2014) 343e351

viscous contribution. However, for smaller clearances (i.e. <0.5%)

there is a considerable rise in total temperature which can be solely
attributed to the viscous effects in the rotating reference frame.
Comparing the last row with the second one shows a remarkable
similarity of the temperature eld, both in terms of magnitude as
well as in the topology. This justies that, even though different
velocities (i.e. the absolute and relative) should be used to evaluate
this viscous term in both of the reference frames, the importance of
the viscous contribution surpasses the effect of the pressure
gradient for tight clearances.
4. Inuence of clearance size on the near-tip heat transfer
Fig. 6(a) presents the heat transfer over the rotor domains most
relevant regions, namely the tip, shroud and suction side as
sketched in Fig. 6(b), for all the investigated clearances. The rst
and second column present respectively the heat transfer to the tip
and shroud for the isothermal cases. From a comparison of the
aerodynamic ow topology (Fig. 2) against the heat load contours,
it can be concluded that enhanced heat transfer levels are found in
the regions where the uid enters the tip gap. For the largest
clearances this happens in the proximity of the airfoil leading edge
and pressure side, while for tighter clearances zones of high heat
transfer appear on the near-tip suction side surface. Fig. 6(c) shows
the heat transfer onto the blade tip and shroud together with the
Mach number isolines and density gradient contours in the tip
channel for the 1.9% case, at the cut section A of Fig. 6(b). Through

the turning of the ow around the entrance corner of the at tip, a

separation bubble is formed, which effectively creates a throat. At
the reattachment point of the ow on the tip surface a region of
high heat transfer is found. In the region close to the throat
(characterized by M z 1), the heat transfer on the shroud is
signicantly altered due to the sudden decrease in boundary layer
thickness combined with the acceleration of the ow. In the aft part
of the blade for the two largest gap cases, a sonic throat is created at
the maximum bubble height and the ow is further accelerated up
to supersonic velocities with Mach numbers above 1.4. Consequently, at the reattachment point, a shock originates and propagates back and forth between the tip and casing (Fig. 6(c)). When
the shock impinges on the shroud, the boundary layer separates,
creating a zone of lower heat transfer downstream.
The third column of Fig. 6(a) shows the heat load on the upper
part of the blade suction side (from 75% to 100% of the blade
height). For the two largest clearances (h/H > 0.5%), the suction side
leakage vortex, indicated with an arrow, enhances the heat transfer
onto the blade surface. For the highest clearances, due to the faster
gap exit ow, the leakage vortex size is larger, with its core located
further away from the blade surface. As a consequence, the heat
transfer levels for a clearance of h/H 1.9% are slightly lower (2%)
and the region of enhanced heat transfer extends towards a lower
radius compared to the case where h/H 0.9%. For lower clearances
(h/H < 0.4%) the tip leakage vortex has almost fully vanished and no
large heat transfer levels are observed in the aft part. On the contrary, a zone of high heat transfer emerges near the tip surface

Fig. 6. Heat transfer for the 5 different clearances: (a) heat ux to the tip, shroud, and upper part of the blade suction side (75%e100% of blade height); (b) the extraction zones; (c)
Mach contours and density gradient at cut section A together with tip and shroud heat ux.

C. De Maesschalck et al. / Applied Thermal Engineering 65 (2014) 343e351

between 25% and 40% of the Cax,r due to the scraping of the
boundary layer and uid ingestion into the suction side marked as
zone C in the gure. The zones labeled D exhibit low levels of heat
ux and are located at the inection point of the streamlines that
enter from the suction side and also exit at the suction side.
5. Aerothermal performance budget
The clearance between the rotating blade and the stationary
casing induces a leakage ow traveling over the tip, resulting in a
loss of the potential work that could be extracted by the rotor and
hence, a decrease in efciency. Consequently, keeping the rotor
clearance as tight as possible is of utmost importance to limit the
leakage mass ow and the induced aerodynamic losses. The aerodynamic losses are shown in Fig. 7(a), where the relative total
pressure contours are shown in a cutting plane located at 95% of the
rotors axial chord (Cax,r). The clearance ow creates a leakage
vortex near the suction side which causes a considerable entropy
rise. There is a signicant reduction of the leakage vortex size with
decreasing clearance height. This effect is also visible in the circumferentially averaged relative total pressure values downstream
of the rotor (Fig. 7(b), left). The tip leakage ow governs the pressure levels even down to about 60% of the blade height. While the
vortex for the largest clearance penetrates the passage up to 25% of
the blade height and gives rise to a spanwise-averaged pressure
decit of over 15%, the vortex in the h/H 0.4% case has almost
vanished and its core barely affects the overall total pressure losses.
Moreover, the leakage vortex alters signicantly the exit ow swirl
(Fig. 7(b), right) which will be felt by the following turbine stage.
Lowering the clearance from 1.9% to 0.4% of the blade span reduces
the absolute outlet ow angle variations in the upper half of the
passage by more than a factor of two.

349

Fig. 7(c) displays the (total-to-total) efciency of the turbine

stage in function of the clearance. The efciency has been evaluated
using different approaches. Conventionally, in the absence of heat
transfer, the aerodynamic efciency is assessed through the ratio of
the extracted power (i.e. total temperature drop) to what would be
isentropically possible across the same total pressure ratio. This
performance is further on called the adiabatic efciency hAD. The
adiabatic efciency for the design point is hAD 91.1%, while the
experimental adiabatic efciency was 91.8%, with an estimated bias
error of 1.1% [24]. However, for investigations performed at
isothermal conditions (like in transient rigs or design and optimization studies) the heat dissipated to the endwalls must be
considered, e.g. by subtracting the energy lost through heat transfer
from the actual drop in total enthalpy across the stage (hISO). Atkins
and Ainsworth [18] recently proposed a comprehensive estimation
(hISO,corr.) of the adiabatic efciency through a non-adiabatic
simulation or experiment by taking into account the temperature
at which the heat transfer takes place for the entropy creation:

hISO;corr:

"
_
_
Ttotal;3 Q_ stator
W
shaft Q stage

Dhtotal;is
Dhtotal;is Ttotal;1
#
Q_ rotor

1

2 Ttotal;2 Ttotal;3

(2)

In this case, the modied approach of Atkins and Ainsworth

allows one to predict the adiabatic efciency hAD from isothermal
evaluations with a difference below 0.17%, while the conventional
approach gives rise to an error of about 1.3%. Examining the efciency for the adiabatic case (hAD), an efciency gain of 2% can be
noticed for every 1% decrease in clearance [25].

Fig. 7. Aerodynamic performance for different clearances: (a) contours of Ptotal,relative at 95% Cax,r; (b) pitchwise-averaged Ptotal,relative and absolute ow angle downstream of the
rotor; (c) losses, efciency and pressure-side gap mass ow.

350

C. De Maesschalck et al. / Applied Thermal Engineering 65 (2014) 343e351

Fig. 8. Turbine heat transfer budget: (a) breakdown for the entire stage; (b) heat transfer levels on the rotor surfaces as a function of the tip clearance, normalized by the heat
transfer for the h 1.9% H case.

Fig. 7(c) also shows the rotor (mass ow averaged) aerodynamic losses u (dened as the drop in relative total pressure
across the rotor row normalized by the difference between the
total relative and static pressure downstream the rotor) and the
amount of leakage mass ow passing through the pressure side
section. Both trends are similar and illustrate the essentially mass
ow-driven loss mechanism. Moreover, for very tight clearances
(e.g. 0.1% and 0.2% of the blade span), the mass ow going through
the gap becomes considerably small and even negative. This can
be explained through the reversed ow (cfr. Fig. 2), which (over)
compensates for the ow entering through the pressure side
section of the gap.
Fig. 8(a) presents the distribution of the heat transfer onto the
different turbine parts for a tip clearance of h/H 1.9%. The rst bar
considers the whole stage, the second bar only the rotor domain,
and the third bar shows how the heat load is distributed over the
rotor blade. Even though the wetted surface of the stator is similar
to that of the rotor, the stator extracts 60% more heat than the rotor
section due to the higher driving temperature difference. Within
the rotor section, the airfoil contributes about 66% of the total heat
transfer. Of this, 6% goes to the tip surface while the rest is almost
equally distributed amongst the different pressure and suction side
surfaces.
Fig. 8(b) illustrates the effect of the clearance on the rotor heat
ux across the surfaces of interest, namely the upper part of the
rotor blade, the tip surface, and the shroud. The heat load onto
the other surfaces (PS and SS) showed almost no variation (<3%)
with clearance. One can observe a signicant decrease in blade tip
and shroud heat transfer in the gap region (the same zones as
presented in Fig. 6(a)), due to the combined effect of the ingested
low energy uid and the low aspect ratio of the gap channel.
Hence, for the largest clearance, the tip and shroud heat transfer
is more than twice the value for the lowest gap size. However, the
overtip shroud component shows a peak value at h/H 0.9%
while the tip heat load shows a monotonic increase with the gap
height. This is due to the reduced heat transfer on the shroud in
the case of the highest clearance, where a signicant ow
detachment takes place (cfr. Fig. 6), resulting in lower heat
transfer levels in the aft part of the blade. The heat transfer to the
pressure side remains practically unchanged (variations <2%),
while the suction side reveals a minimum (7%) for h/H 0.4%.
This is the clearance size at which the leakage vortex is diminished signicantly (which is the main reason for the heat transfer
at higher clearances) while the high heat transfer levels due to
the scraping of the boundary layer (important for very low
clearances) are not yet signicant.

6. Conclusion
This paper presents an elaborate numerical investigation of a
high-speed gas turbine stage, operating in the as-yet uninvestigated tight running regime (tip clearances less than 0.5% of the
blade span). In such a regime, the streamlines in the tip gap reveal a
reversed ow (i.e. from the suction side to the pressure side) due to
the low-momentum and -energy uid from the upstream boundary layer ingested through the gap suction side. Moreover, this
aerodynamic ow topology alters signicantly the thermal eld
conventionally expected in the tip region of a rotor blade, through
an important increase of the viscous effects.
Heat transfer distributions show enhanced levels on the shroud
and tip where the uid enters the gap. In the suction side near-tip
region, the inuence of the leakage vortex is predominant for
clearances above 0.5%. However, for lower gap sizes, an important
zone of high heat transfer emerges at 25% Cax,r due to the scraping
of the boundary layer. The dependency of the heat load on clearance size show different trends for the suction side and tip surface,
with a minimum at h/H 0.4% in the former and a maximum at h/
H 0.9% in the latter. The aerodynamic (adiabatic) efciency
proved to be mainly driven by the leakage mass ow and was
predicted accurately from isothermal calculations with an error
lower than 0.17%.
The results showed a novel change in the aerothermal characteristics in the overtip region for tight clearances and provide new
physical insights, indispensable to design efcient rotating machines adopting small leakage paths.
Acknowledgements
The authors gratefully acknowledge the nancial support of the
Agency for Innovation by Science and Technology in Flanders
(IWT), the support of Johan Prinsier for the numerical simulations
and John Mcclean for the enhancements in the nal manuscript.
Nomenclature
a
speed of sound
C
blade chord
h
rotor tip clearance height
Dh
enthalpy difference
H
rotor channel height
LE
leading edge
M
Mach number
P
pressure
Pr
Prandtl number

C. De Maesschalck et al. / Applied Thermal Engineering 65 (2014) 343e351

PS
Q
RANS
Re
RPM
SS
t
T
TE
u
v
V
W
y

pressure side
heat transfer
Reynolds-averaged NaviereStokes
Reynolds number
revolutions per minute
suction side
time
temperature
trailing edge
rotor peripheral speed
velocity vector
absolute velocity
relative velocity
dimensionless wall-normal height of rst cell at wall

Greek symbols
absolute ow angle
efciency
boundary layer momentum thickness
density
shear stress
(Ptotal,relative,2  Ptotal,relative,rotor exit)/(Ptotal,relative,rotor
exit  Protor exit)

a
h
q
r
s
u

Subscripts
AD
adiabatic
ax
axial component
corr.
corrected to account for the temperature at which the
entropy is generated
exit
outlet of a turbine stage
in
inlet of a turbine stage
ISO
isothermal
PS
pressure side
r
rotor
s
static quantity
sh
turbine casing
SS
suction side
t
tangential component
w
at the wall
1,2,3
inlet, statorerotor interface and outlet of the turbine
stage respectively
References
[1] M. Fnot, Y. Bertin, E. Dorignac, G. Lalizel, A review of heat transfer between
concentric rotating cylinders with or without axial ow, Int. J. Therm. Sci. 50
(7) (July 2011) 1138e1155.

351

[2] M. Fnot, E. Dorignac, A. Giret, G. Lalizel, Convective heat transfer in the entry
region of an annular channel with slotted rotating inner cylinder, Appl. Therm.
Eng. 54 (1) (May 14 2013) 345e358.
[3] J.M. Bergada, S. Kumar, D.Ll. Davies, J. Watton, A complete analysis of axial
piston pump leakage and output ow ripples, Appl. Math. Modell. 36 (4) (April
2012) 1731e1751.
[4] M. Pelosi, M. Ivantysynova, Heat transfer and thermal elastic deformation
analysis on the piston/cylinder interface of axial piston machines, J. Tribol. 134
(4) (2012) 041101.
[5] M. Papa, R.J. Goldstein, F. Gori, Numerical heat transfer predictions and mass/
heat transfer measurements in a linear turbine cascade, Appl. Therm. Eng. 27
(4) (2007) 771e778.
[6] N.W. Harvey, Aerothermal implications of shroudless and shrouded blades, in:
T. Arts (Ed.), Turbine Blade Tip Design and Tip Clearance Treatment, Von
Karman Institute for Fluid Dynamics Lecture Series, von Karman Institute for
Fluid Dynamics, Brussels, 2004.
[7] M. Malak, J. Liu, E. Zurmehly, Turbine Shroud Durability Analysis Using Time
Unsteady CFD and SieC Testing, 2011. ISABE paper No. 2011e1706.
[8] F.J.G. Heyes, H.P. Hodson, G.M. Dailey, The effect of blade tip geometry on the
tip leakage ow in axial turbine cascades, J. Turbomach. 114 (1992) 643e651.
[9] J.S. Kwak, J. Ahn, J. Han, C. Pang Lee, R.S. Bunker, R. Boyle, R. Gaugler, Heat
transfer coefcients on the squealer tip and near-tip regions of a gas turbine
blade with single or double squealer, J. Turbomach. 125 (2003) 778e787.
[10] N.W. Harvey, K. Ramsden, A computational study of a novel turbine rotor
partial shroud, J. Turbomach. 123 (2001) 534e543.
[11] R.S. Bunker, A review of turbine blade tip heat transfer in gas turbine systems,
Ann. N. Y. Acad. Sci. 934 (2001) 64e79.
[12] A.P.S. Wheeler, N.R. Atkins, L. He, Turbine blade tip heat transfer in low speed
and high speed ows, J. Turbomach. 133 (2011) 041025.
[13] Q. Zhang, L. He, Overtip choking and its implications on turbine blade-tip
aerodynamics performance, J. Propuls. Power 27 (5) (2011) 1008e1014.
[14] P.J. Newton, G.D. Lock, S.K. Krishnababu, H.P. Hodson, W.N. Dawes, J. Hannis,
C. Whitney, Heat transfer and aerodynamics of turbine blade tips in a linear
cascade, J. Turbomach. 128 (2006) 300e309.
[15] A.L. El-Gabry, A.A. Ameri, Comparison of steady and unsteady RANS heat
transfer simulations of hub and endwall of a turbine blade passage,
J. Turbomach. 133 (2011) 031010.
[16] Q. Zhang, D.O. ODowd, L. He, M.L.G. Oldeld, P.M. Ligrani, Transonic turbine
blade tip aerothermal performance with different tip gapsepart I: tip heat
transfer, J. Turbomach. 133 (2011) 041027.
[17] R.C. Dean, On the necessity of unsteady ow in uid machines, ASME J. Basic
Eng. 81 (1959) 24.
[18] N.R. Atkins, R.W. Ainsworth, Turbine aerodynamic performance measurement
under nonadiabatic conditions, J. Turbomach. 134 (6) (2012) 061001.
[19] S. Lavagnoli, G. Paniagua, C. De Maesschalck, T. Yasa, Analysis of the unsteady
overtip casing heat transfer in a high speed turbine, J. Turbomach. 135 (3)
(2013) 031027.
[20] T. Yasa, G. Paniagua, R. Denos, Application of hot-wire anemometry in a blowdown turbine facility, J. Eng. Gas Turbines Power 129 (2007) 420e427.
[21] The Jet Engine, Rolls-Royce plc, 2005, ISBN 978-0-902121-23-2.
[22] R.S. Bunker, Gas Turbine Cooling: Moving from Macro to Micro Cooling, 2013.
ASME paper no. GT2013e94277.
[23] F.J.G. Heyes, H.P. Hodson, Measurement and prediction of tip clearance ow in
linear turbine cascades, J. Turbomach. 115 (1993) 376e382.
[24] T. Yasa, Efciency of a HP Turbine Tested in a Compression Tube Facility, PhD
thesis, von Karman Institute for Fluid Dynamics, 2008. ISBN: 978-0-90212123-2.
[25] S. Yoon, E. Curtis, J. Denton, J. Longley, The Effect of Clearance on Shrouded
and Unshrouded Turbines at Two Levels of Reaction, 2010. ASME paper no.
GT2010e22541.