ACGT2014-052

Design and Performance Analysis of a 3-D Supersonic Impulse Turbine

with Optimization of Rotor Blade
Krupakaran R1, Abhilash S2, Allen P Doss3 and Ravichandran K S4
1
Department of Mechanical Engineering
PES Institute of Technology
100 Feet Ring Road, BSK 3rd Stage, Bangalore – 560085
Phone : 918050558903, E-mail : [email protected]
2
PES Institute of Technology
3
PES Institute of Technology
4
PES Institute of Technology

ABSTRACT conditions. A rotor design theory was developed by them

The losses in supersonic impulse turbines are mainly for constant passage width to reduce focusing of
attributed to shocks and flow separation. In the present compression shock waves on convex surface of the blade
work an effort is made to minimize losses by studying or minimize flow separation and thus decrease shock
the effects of variation in transition arc angle and passage losses. They proposed that the transition arc angle of
width defined by the blade profile on flow performance. rotor blade should be maintained between 100 and 170 to
CFD analyses of 2D cascade were performed for passage minimize shocks and flow separation. They also
width of 21.8%, 24.29%, 27.33%, and 29.14% of pitch, observed that the rotor performance was better when
with transition arc angle varying between 90 and 150. The passage width was reduced but at the cost of reduction in
computational results showed that the impulse blade with pitch.
transition arc angle of 120 and passage width of 24.29% Goldman (1968) in Lewis Research Center also
of pitch demonstrated optimum performance. conducted analysis on impulse turbine blade sections.
The 2-D blade profile with superior performance was His work focused on understanding the effects of lower
adapted for the design of a 3-D blade passage of an and upper surface Mach number distribution, and rotor
impulse turbine operating at 30000 RPM. CFD analysis inlet flow angle on blade shapes for Mach number range
of the performance of this turbine stage design shows of 1.5 to 5.0. He proposed a variety of blade shapes for
that the power output is about 80 kW. The flow field different Mach numbers and inlet flow angles. In a
computed by Fluent is discussed in terms of performance separate research work Goldman (1971) also has applied
parameters such as total pressure loss and entropy rise boundary layer theory to have turbine blade designs
across the stage. corrected for displacement thickness of the boundary
layer. However, shock formation and flow separation are
INTRODUCTION not accounted for in this design procedure.
Supersonic impulse turbines are widely used in Further, Deych (1964) developed their own design
applications which require a large power output in a theory to get supersonic rotor blades with weaker shocks.
single stage at high rotational speeds. In addition, these This design methodology proposed impulse blades of
turbines produce high specific impulse, have low weight varying passage width for any supersonic inlet Mach
to power ratio and are compact in nature. Some number.
applications are in propulsion of underwater missiles Later, the focus of researchers shifted to other
(Torpedoes), space rocket turbo pumps to rotate the parameters of the supersonic impulse turbine design like
pump feeding fuel or oxygen to the combustion chamber, axial gap effect, stator-rotor interaction (Rashid, 2007)
the Curtis stage of an industrial turbine (Aghaei, 2013), and partial admission design. The partial admission
turbochargers and in open cycle auxiliary power plants. turbines are generally preferred when available mass
However, the efficiency of these turbines is lesser than flow rate is too low for full admission. In addition, to
that of subsonic turbines. The efficiency may vary avoid losses due to low blade aspect ratio and
between 25% and 40% depending on the size of the manufacturing problems, partial admission is used
turbine. The losses are mainly attributed to the focusing (Dorney, 2002). However, the flow in partial admission
of compression waves on convex surface of the rotor is unsteady as the rotors periodically pass through flow
resulting in a shock and subsequent flow separation regions and regions of no flow (Dorney, 2002). Hence,
along the convex surface extending to the mean line of the rotor is subjected to undesired vibration and dynamic
the flow passage. loads. Also, the efficiency of partial admission turbines is
Preliminary analysis of an available impulse blade lesser than that of full admission ones (Fridh, 2012) and
design for torpedo application (Ravichandran, 2013) it has been found that performance of the partial
discussed later in this report showed high aerodynamic admission turbines is optimum only for a particular
losses due to large flow separation. One of the objectives degree of flow admission (Aghaei, 2013).
of this work was to explore whether an improved Hence, the present study focused on implementing a
performance could be achieved by redesigning the full admission turbine of relatively smaller size which is
impulse blade. Hence, a review of literature on the structurally subjected to weaker centrifugal loads in place
supersonic impulse blade design was conducted and is of an already existing relatively larger partial admission
discussed below: design. This would mitigate the disadvantages mentioned
Boxer et al (1952) applied free vortex flow theory to above for partial admission design. However, this
obtain supersonic rotor blades with vortex flow within approach also leads to blade torque that is
the passage. Later, another experiment was conducted by proportionately reduced and for the same rotor speed
Stratford (1962b) on rotor blades with supersonic inlet results in lesser shaft power output. This is compensated

1
by an increase in the mass flow rate for the same
stagnation conditions upstream.
Though the research on impulse blade profile with
minimum losses have been proposed from late 1960’s as
seen above, there has been no research lately to either
improve or understand the effects of changes in certain
parameters of the impulse blade. In this regard, this
report attempts to utilize and improve the Stratford and
Sansome design methodology for rotor profile design.
The effect of the parameters - transition arc angle and
passage width was studied to obtain optimum rotor
profile. The study is carried out in 2 stages: Firstly, Fig. 1 Stator blade to have convergent-divergent nozzle
different impulse 2-D blade profiles were studied for profile (Stratford, 1962a); all dimensions in mm
their relative performance. Secondly, full admission 3-D
analysis was carried out with the impulse blade profile Impulse Blade
shortlisted for better performance. The design procedure elaborately explained in
Stratford (1962b) was adopted to design the impulse
blade profile. According to the design procedure, for
NOMENCLATURE
Cp specific heat capacity in kJ/kgK rotor inlet Mach number M0, is defined as
follows:
M Mach number
µo= Sin-1 (1)
Ptot total pressure in Pascal

Pabs absolute pressure in Pascal (2)

radius of curvature of convex surface in mm From the results acquired through tunnel test
experiments, Stratford concluded that the values obtained
radius of curvature of concave surface in mm from Eq. (2) should be reduced by 20% to 30% for
practical applications. In this paper, M0 is taken as 2.4
radius of curvature of passage centre line, mm which is also the exit flow Mach number of the stator
used. Hence the value obtained from Eq. (2) for this case
w passage width in mm is 0.9091. Including the reduction of 20% to 30% as
mentioned by Stratford and Sansome, should be
α Nozzle angle in deg
maintained between 0.72728 and 0.63637 respectively.
β Impulse blade angle in deg
Table. 1 Parameters of impulse blades modelled
µ0 Mach angle in deg
w as w
% of (mm) (mm) (mm) (mm)
transition arc angle in deg pitch
21.80 1.9674 4.128 6.096 5.112 0.6771 0.3848
k turbulent kinetic energy 24.29 2.1860 3.916 6.102 5.009 0.6417 0.4364
27.33 2.4600 3.649 6.109 4.879 0.5973 0.5042
ω turbulent dissipation rate 29.14 2.6232 3.490 6.113 4.802 0.5709 0.5463

ρ density in kg/m3

Subscripts
w
0 stator outlet, rotor inlet
1 Nozzle exit
2 Impulse blade inlet

in rotor inlet
out rotor outlet

DESIGN OF BLADES
The stator was designed referring to Stratford Fig. 2: Impulse blade profile; all dimensions in mm
(1962a) design method to obtain a shock free supersonic
flow with Mach number of 2.4 at the stator exit. The For every passage width shown in Table 1, a series of 2-
pitch and chord of stator passage is 14mm and 12mm D impulse blade profiles were obtained for transition arc
respectively. A Convergent–Divergent nozzle passage angle of 100, 110, 120, 130 and 150 which resulted in
designed with throat width of 1.105 mm at mean twenty impulse blade profiles. The chord of all these
diameter, exit to throat area ratio of 2.962 and nozzle exit impulse blades was maintained as 11.979 mm. The pitch
angle of 140 was obtained for 2-D cascade. Figure 1 and total turning angle of these impulse blade passages
shows the layout of the nozzle passage. was fixed at 9 mm and 1440 respectively.

2
 A 2-D cascade CFD analysis was carried out for all Grid
the configurations mentioned above. Fig. 3 illustrates the
2D cascade model used for the analysis. 2-D grid. Quadrilateral elements were used
Pressure inlet throughout the 2D cascade flow domain. Successive ratio
of 1.05 was maintained for the meshes normal to the
stator and rotor walls to accurately capture the viscous
effects and flow separation as indicated in enlarged view
of rotor passage of Fig. 4. The 2D cascade flow domain
had a total of 49650 cells.

Periodic boundary

Pressure outlet

Fig. 3 2D cascade model with two stator passages and

three rotor passages; all dimensions in mm

3D DESIGN
The 2D CFD analysis (discussed under 2D results)
Fig. 4 Meshed 2D cascade with enlarged view of rotor
showed that the profile with a passage width of 2.186
passage having a successive ratio of 1.05
mm and transition arc angle of 120 was a suitable choice
for optimum performance. This impulse blade profile 3-D grid. Unstructured tetrahedral elements were
was therefore used to design a 3-D impulse turbine of used in the stator and rotor domain of the 3D model,
mean diameter 60 mm. The design parameters are while hexahedral elements were used in the duct domain.
tabulated in Table. 2. The 3D mesh contains a total of 4.147 million cells. This
was reduced to around 1.45 million cells after the
Table. 2 Specifications of stator and impulse rotor blades tetrahedral cells were converted to polyhedral cells in
for 3D model Fluent because the latter reduces the computational time
and cell skewness while increasing orthogonality.
Profile Stator Rotor
Height (mm) 8 8
Chord (mm) 12 11.979
Pitch (mm) 14.03 9

α1 14 0
-
β2 - 18.40
Throat (mm) 1.105 2.186
Quantity 13 21

NUMERICAL ANALYSIS
The numerical analysis was carried out using the Fig. 5 Meshed 3D turbine model with stator, rotor and
commercially available software Ansys Fluent. duct domains from left to right.

Governing Equations Boundary Conditions

The RANS (Reynolds averaged Navier-Stokes) The boundary condition at the inlet was set to a constant
equations were solved for both the 2D and 3D stagnation pressure and temperature of 40e5 Pa and
computational analysis. 7230K respectively. Similarly, the outlet was set to a
pressure of 1.75e5 Pa. Steam under these conditions is
+ superheated and remains so throughout the flow domain.
Translational periodicity was applied to the outer
boundaries other than wall, inlet and outlet in the
Where U = (ρ, ρu, ρv, ρw, ρE) and F,G,H are the circumferential direction for the 2-D cascade analysis to
respective flux vectors. The shear-stress transport (SST) accommodate periodicity of fluid flow in the regions
k-ω model in Fluent developed by Menter (1994) is used between stator and rotor, upstream of stator and
for modeling the effects of turbulence. k-ω SST model is downstream of rotor.
known for its good behaviour in adverse pressure For the case of 3-D calculations, mesh interface
gradients and separating flows. To accommodate viscous boundary condition was used in the interface between
effects near the wall where the Reynolds number is low, rotor-stator and in between rotor and duct as this
this model mimics k-ω equation and away from the wall condition ensures the continuity in solution to flow field
where the Reynolds number is high, it reduces to k-ϵ variables between the rotating and non rotating domains
equation applicable to free stream conditions. (non-conformal mesh).
Water vapour is used as the working fluid in the
analysis. Density of water vapour was set to behave as

3
that of an ideal gas, Cp is defined as a piece-wise
polynomial function of temperature defined between
temperatures 250K and 1400 K. The viscosity co-
efficient and thermal conductivity of water vapour are
also defined as functions of temperature as available in
Fluent.

Validation
A validation exercise was conducted in order to
determine the effectiveness of Fluent in calculating the
flow field parameters in high pressure gradient
supersonic flows within the impulse blade passages. For
this purpose, a 2-D cascade CFD analysis was carried out
for Stratford’s (1962b) impulse blade geometry and inlet
Mach number conditions. These results were compared
with the annular cascade experimental data from
Stratford and Sansome (1962b).
Stratford and Sansome conducted experimental study
for different impulse blade designs. For this validation,
the CFD analysis was performed for one specific blade
called 1st blade design in Stratford (1962b). Fig. 7 Total to total pressure ratio of rotor outlet and
Fig 6 shows the Mach number contour plot for the rotor inlet across w at mid span
same. It can be observed that there is mild flow
separation near the entrance and exit of the rotor which is 2D RESULTS
consistent with the results of Stratford and Sansome The CFD analysis of the impulse blade used in the
(1962b-pp26). available partial admission design for the boundary
conditions mentioned above showed significant flow
separation as indicated in Fig. 8 (Ravichandran, 2013).
Suitable modification to the impulse blade profile was
clearly necessary. To reduce the shock losses, flow
separation extent had to be minimized and the separation
point needed to be moved further downstream of leading
edge of convex surface.

Fig. 6 Mach number contour plot for first impulse blade

Region of flow
design of Stratford and Sansome
separation
In addition to this, the ratio of total pressure at the
rotor outlet to that at the rotor inlet across the passage
width, calculated from 2-D cascade CFD results was
compared with that of the experimental data from
Stratford and Sansome (1962b). Fig 7 shows the
normalized total pressure blade-to-blade wake profile Fig. 8 Contour plot of Mach number for impulse blade
compared with experimental data. It is seen that the used in partial admission design
nature of graph is similar for both CFD results and that of
the experiment. The CFD curve agrees well with that of The 2D cascade models were analysed
experiment beyond 65% of passage width from the computationally for all the impulse blade designs
convex surface. However, near the convex surface the mentioned under heading Impulse Blades. The contour
total pressure from CFD is slightly higher than that of the plot of Mach number was plotted for all the variants to
experiment indicating reduced losses in this region. This understand the occurrence of shocks and flow separation
may be related to some deficiency in the grid used and in the rotor passage.
consequent inability of the SST model to accurately
capture the thickness and extent of the separation bubble Results For Impulse Blade With Passage Width of
mentioned earlier. Nevertheless it may be surmised that 24.29% (2.186 Mm) of Pitch
the CFD model can successfully capture the essential According to Stratford (1962b), flow separation
characteristics of the flow through a supersonic impulse could be considerably reduced if the radius of curvature
blade. of transition arc is twice that of the circular arc and when
θT is retained between 100 and 170.

4
 1

Fig. 9a Contour plot of Mach number for w of 24.29% of

pitch and transition arc angle of 110 Fig. 10 Total pressure change across the rotor passage for
passage width of 24.29% pitch against θT

The better performance of transition arc angle of 120 can

be further explained by observing the pressure
distribution along the convex and concave surface of the
blade. The use of single circular arc or constant passage
curvature for the impulse blade would lead to zig-zag
pressure gradients (Fig. 11, left) along its surface which
is believed to affect the boundary layer and cause flow
separation as explained by Stratford and Sansome
(1962b). However, when a combination of transition and
circular arcs are used to satisfy the flow curvature
condition of Stratford and Sansome (1962b), ideal
Fig. 9b Contour plot of Mach number for w of 24.29% of absolute static pressure distribution along convex and
pitch and transition arc angle of 120 concave surface of impulse blade will be as shown in
figure 11 (right).

Fig. 11 Absolute static pressure distribution across

Fig. 9c Contour plot of Mach number for w of 24.29% of
impulse blade (left) constant passage curvature (right)
pitch and transition arc angle of 130
varying passage curvature; (Stratford, 1962b)
The Mach number contour plots in Figs9a-c show that
the variation of transition arc angle could significantly
affect the flow separation even if it is maintained within
the range prescribed above.
Figure 9b shows that for a constant passage width of
24.29% of pitch, the flow separation was highly reduced
for transition arc angle of 120 and the separation point
was also moved downstream of convex surface to mid-
chord region. The total pressure drop across the blade
passage is also the least for transition arc angle of 120
which is in good agreement with the contour plots.

5
Fig. 12 Absolute static pressure distribution on concave for θT = 150 for passage width of 2.46 mm but the
and convex surface of impulse blades with w of 24.29% entropy increment are almost the same for both 120 and
of pitch for of 100, 120 and 150 150.

The comparison of the pressure distribution on the

concave and convex surface for different transition arcs
for w of 24.29% of pitch is as shown in the Fig. 12. The
fluctuation in pressure is high for of 100 and 150 with
3 suction peaks in comparison with of 120 with two
peaks. According to Stratford and Sansome (1962b),
when the transition arc curvature is relatively high as for
13 and 15 deg, the compression waves from leading edge
of concave surface will not fall on convex surface due to
it being straight rather than taking a curved path and also
because of upstream flow separation in leading edge of
convex surface (Fig. 9c). This may be the reason for
higher flow separation in the above said . On the other Fig. 14 Contour plot of Mach number for w of 21.8% of
hand, the flow separation in leading edge for of 100 pitch and θT = 120
and 110 (Fig. 9a) is believed to be because of transition
arc curvature being much smaller than the threshold
curvature required in preventing the upstream flow
separation on leading edge of convex surface. As for the
convex side pressure distribution, no great deal of
difference in pattern is observed visibly. This explains
the better performance of the rotor with 120 transition
angle.

Comparison of Results For Impulse Blades With

Different Passage Widths
Here, an effort is made to understand the effects of
variation in passage width of impulse blade on flow
performance. The steps performed for w of 24.29% are Fig. 15 Contour plot of Mach number for w of 27.33% of
extended to every other passage width analyzed. i.e. for pitch and θT = 120
21.8% (1.967mm), 27.33% (2.46mm), and 29.14%
(2.6232mm) of pitch. Comparison has been made Furthermore, for every transition arc angle the flow
between impulse blades with better performance among separation increases with increase in passage width. This
every passage width variant and is presented here. behaviour can be deduced from amount of entropy
Contour and velocity vector plot colored by Mach increment across the impulse blade passage in Fig.13 and
number was captured for every case mentioned above to can be explained as follow: As passage width increases,
precisely understand their behaviour. the compression waves from leading edge of impulse
blade concave surface would be unable to touch its
convex surface and hence leads to flow separation. This
effect amplifies as passage width increases and is clearly
visible in the Mach number plots shown in Fig. 14 to Fig.
16.

Fig. 13 Entropy increment across the impulse blades of

different passage width against θT

It was observed that the impulse blade with θT = 120 Fig. 16 Contour plot of Mach number for w of 29.14% of
performed better compared to other transition arc angle pitch and θT = 120
for every passage width analyzed. The entropy increment
as a measure of efficiency across the impulse blade Thus the CFD analyses of the 2D cascade model for
passage plotted for various passage widths against various impulse blades designed showed an optimum
transition arc angle (Fig. 13) concurs well with the above performance for a transition arc angle of 12 0 and passage
observations. Only exception is with passage width of width of 24.29% of pitch. It was observed that the
1.967 mm where the performance is almost the same for impulse blades with transition arc angle of 120 performed
100, 110 and 120. In addition, flow separation was least better when compared to the transition arc angles 10 0,

6
110, 130 and 150 for all the passage widths analysed. The appropriate approach. The density based coupled solver
design of the 3D stage was accordingly based on the was used and the rotational speed was set to 30000 RPM.
shortlisted profile from 2D analysis. Numerical divergence encountered in the initial attempts
were overcome by ramping up the inlet stagnation
pressure gradually from 3 bar (the outlet pressure was
Grid Independence kept fixed at 1.75 bar) to 40 bar through 5, 8, 15, 25 and
The 2D cascade results were checked for grid 35 bar in sequence. For each of these inlet conditions the
independence. The profile chosen for this was the one solver was allowed to run until the convergence or for at
with maximum passage width of 29.14% of pitch (2.6232 least 200 iterations.
mm). This ensured that the ones with lesser passage
width had finer elements than this case. The grids were
varied in the direction perpendicular to the flow. The
velocity magnitude profile at the exit of the rotor was
chosen as the parameter to compare the different mesh
results. This was used because all the variations inside
the impulse blade passage would affect the profile of
velocity at the rotor exit and hence would give us the
idea of the reliability of the mesh to capture the shocks
and flow separation.
The mesh sizes of 40, 60 and 80 elements
perpendicular to flow were analyzed for the case of
impulse blade with 2.6232 mm passage width and
transition arc angle of 120. Figure 17 shows the profile Fig. 18 Contour plot of Mach number for the turbine at
of velocity magnitude at the rotor exit for all these 50% blade height (mean diameter of 60 mm) for 1st
variations in mesh sizes. design
The maximum deviation in the magnitude of velocity
was found to be 1.826% between grid of 80 elements and The 3-D full admission simulation showed that there
that of 60 elements. The maximum deviation of 2.2% was occurrence of unforeseen oblique shocks at the stator
was found between grids with 60 and 40 elements. The exit as shown in Fig. 18. This led to a subsonic flow for
60 grid mesh was found to be satisfactory in capturing most part the rotor passage and these shocks became
the characteristics of flow with modest computational predominant when approached closer to the stator hub.
requirements. Hence the rest of the analysis was carried
out for 60 elements mesh.

Fig. 19 Mach number at different stations of 3-D impulse

turbine
As shown in Fig. 19, the average Mach number at the
exit of the stator plane is 1.36 which is substantially
lower than what it was in the 2D cascade case (M = 2.3).
This is due to the unforeseen oblique shocks at the stator
exit. The reason for exit shocks may be attributed to the
tendency of the flow to expand near the tip of stator and
impulse blades.

Stator
Fig. 17 Deviation of velocity magnitude at rotor exit for blade tip
different grids

3D RESULTS
The objective of 3-D analysis was to establish the
ability of the stage design to deliver the rated power Impulse
operating at 30000 RPM and for conditions of full blade tip
admission. Under the time constraints of academic
deadlines, it was decided that an inviscid analysis would Fig. 20a Velocity angle contour plot at the mean blade
serve the stated objective. A steady state analysis in the height when the blade tips are closer
MRF framework made available by Ansys Fluent was an

7
 pitch. Thus, it can be concluded that the impulse blade
with passage width of 2.186 mm and transition arc angle
of 120 showed optimum performance among the various
passage widths analyzed. The 3D aerodynamic analysis
of this impulse blade under inviscid, full admission
conditions showed weaker oblique shocks near the
Stator trailing edge of stator, which was not observed in 2D
blade tip
cascade results. This turbine gives a torque output of
Impulse 25.95 Nm and power output of 81.49 kW. It is believed
blade tip that the power output would increase if the trailing edge
shocks are eliminated.

REFERENCES

Fig. 20b Velocity angle contour plot at the mean blade Aghaei, T and Tousi, A.M., 2013, “Experimental and
height when the blade tips are farther numerical investigation of design optimization of a
partial admitted supersonic turbine,” Propulsion and
From Fig. 20a it is clear that when the stator and Power Research, Vol. 2(1), pp. 70–83
impulse blade tips are closer, the angle of flow deviates Boxer, E, Sterrett, J. R., and Wlodarski, J., 1952,
more to interfere with the flow from the nearby stator “Application of Supersonic Vortex-Flow Theory to the
passage. In Fig. 20b, where the blade tips are farther Design of Supersonic Impulse Compressor or Turbine -
apart, the deviation persist but the extent is lesser. Blade Sections,” NACA RM, L52B06. unexpected
Demuren, H. O., 1976 “Aerodynamic performance
and heat transfer characteristics of high pressure ratio
transonic turbines,” MS Thesis in Aeronautics,
Massachusetts Institute of Technology, USA
Deych, M. Ye. and Troyanovskiy, b. M., 1964,
"Investigation And calculation of axial-turbine stages,”.
Dorney, D. J., Griffin, L. W., and Sondak, D. L.,
2002, “Full- and Partial Admission Performance of the
Simplex Turbine,” AIAA-3638
Fridh, J., “Experimental Investigation of
Performance Flow Interactions and Rotor Forcing in
Axial Partial Admission Turbines,” PhD Thesis, KTH.
Goldman, L.J., 1968, “Analytical investigation of
Supersonic turbomachinery blading I1 - analysis of
impulse turbine-blade sections,” NASA TN D-4422.
Goldman, L.J., 1971, “Supersonic turbine design and
performance”, NASA N7212783.
Fig. 21 Comparison of total pressure at different stations Jeong, E., Lee H., , Park, P., and Kim, J., 2008, “Tip
of 2-D cascade and 3-D impulse turbine. Clearance Effect on the Performance of a Shrouded
Supersonic Impulse Turbine,” Journal of Propulsion and
It can be seen from the Fig. 21 that the total pressure power, vol.24, No.6.
drop across the stator passage is larger for the 3-D Rashid, S., Tremmel, M., Waggot, J., and Moll, R.,
turbine than that of the 2-D cascade. This can be 2007, “Curtis stage nozzle / rotor aerodynamic
attributed to the dissipation of energy across the oblique interaction and the effect on stage performance,” Journal
shocks at the stator passage exit which did not exist in of Turbomachinery 129, (3) 551–562.
the 2-D cascade results. However, the total pressure in 3- Ravichandran, K.S., Pravesh,Y.P., 2013,
D turbine is found to follow the same trend as that of 2-D “Aerodynamic Design and CFD validation of a Single
cascade beyond stator exit. Stage Supersonic Impulse Turbine,” Technical Internal
The resulting torque for this turbine is 25.95 Nm with Report, CFD-2013, PES Institute of Technology,
a mass flow rate of 0.56 kg/s. The corresponding power Bangalore.
output is 81.49 kW. This is lesser than the extrapolated Stratford, B. S. and Sansome, G. E., 1962a, “The
power output of 100 kW from the 2-D cascade results. Performance of Supersonic Turbine Nozzles,” ARC
The reason for the drop in power output for the 3-D R&M 3273,
analysis can be attributed to the 3-D losses such as end Stratford, B. S. and Sansome, G. E., 1962b, “Theory
wall losses and trailing edge losses occurring due to and Tunnel Tests of Rotor Blades for Supersonic
finite thickness of the blade trailing edge which causes Turbines,” ARC R&M 3275.
flow separation and shock-expansion wave interaction WEI, N., 2000, “Significance of Loss Models in
(WEI, 2000). Aerothermodynamic Simulation for Axial Turbines,”
Ph.D Thesis, Royal Institute of Technology, Stockholm,
CONCLUSION Sweden.
From the 2D cascade results, it can be inferred that
the losses are minimum for transition arc angle of 120 for
all the passage widths analyzed and the flow performance
of the impulse blade was superior when the passage
width was maintained between 21.8% and 25% of rotor