Micro Gas Turbines - A Short Survey of Design Problems

Full text

RTO-EN-AVT-131 1 - 1
Micro Gas Turbines – A Short Survey of Design Problems
R.A. Van den Braembussche
von Kármán Institute for Fluid Dynamics
Waterloose steenweg, 72
B-1640, Rhode-St-Genèse
BELGIUM
vdb@vki.ac.be
INTRODUCTION
Micro gas turbines have experienced a growing interest during the last decade. Their large energy density
(Whr/kg) makes them attractive for portable power units as well as for propulsion of small airplanes
(UAV). They are also of interest for distributed power generation in applications where heat and power
generation can be combined. The need for high performance in both applications is at the origin of a
worldwide interest and research on micro gas turbines and the motivation for the present lecture series.
Scaling is a common technique to define larger or smaller geometries with similar characteristics.
However a simple scaling of a high performance large gas turbine will not result in a good micro gas
turbine. The main factors perturbing such a scaling are:
• The large change in Reynolds number.
• Massive heat transfer between the hot and cold components (negligible in large machines).
• Geometrical restrictions related to material and manufacturing of miniaturized components.
The purpose of the present lecture is to provide a first insight into the aero-thermal problems of micro gas
turbines.
1.0 GAS TURBINE SCALING
1.1 Scaling Model
Conservation of the characteristics of the thermodynamic cycle is a first requirement when scaling a gas
turbine. This means the same exchange of the energy between fluid and rotating components but at a
smaller mass flow. Hence the enthalpy change in the compressor and turbine should be conserved.
∆H = ω
2
.R
2
2 (1)
This provides a relation between RPM and impeller diameter that can easily be satisfied. It also means that
the power in each component,
∆E = Q.∆H
(2)
and hence the gas turbine power will be proportional to the mass flow Q. Another consequence of
maintaining the same velocities and pressures is that the mass flow and power depend on the cross section
area and scale with R
2
2
.
Van den Braembussche, R.A. (2005) Micro Gas Turbines – A Short Survey of Design Problems. In Micro Gas Turbines (pp. 1-1 – 1-18).
Educational Notes RTO-EN-AVT-131, Paper 1. Neuilly-sur-Seine, France: RTO. Available from: http://www.rto.nato.int/abstracts.asp.

Micro Gas Turbines – A Short Survey of Design Problems
1 - 2 RTO-EN-AVT-131
The main conditions to maintain the level of efficiency of compressors and turbines is the conservation of
velocity triangles, Reynolds number and Mach number.
The conservation of Mach number
TR
R
M
G
..
2
γ
ω
=
(3)
does not create problems except for applications with big changes of inlet temperature (T) or gas
characteristics (γ, R
G
).
Conservation of Reynolds number
HRR ∆== ..Re
2
2
µ
ρ
µ
ρ
ω
(4)
conflicts with the conservation of enthalpy change when changing the dimensions of the rotor. A reduction
of Reynolds number with decreasing dimensions is unavoidable unless also the viscosity and/or pressure
level are modified. The impact of reducing dimensions on compressor performance can be obtained taken
from Fig. 1 where the change in efficiency with decreasing mass flow is indicated by the vertical dashed
line.
Figure 1: Variation of Compressor Performance with Pressure Ratio and Mass Flow (acc. [1]).
The “Specific speed”
4/3
/.
H
Q
Ns
∆
=
ρω
(5)
is a non-dimensional parameter characterizing the shape of a radial compressor and can be used to
estimate the achievable performance. Experience has shown that the highest efficiencies can be obtained

Micro Gas Turbines – A Short Survey of Design Problems
RTO-EN-AVT-131 1 - 3
for Ns values between .6 and .8 (Fig. 2). Once the enthalpy rise (∆H) and mass flow (Q) are known it
defines the required radial compressor rotational speed. Optimum Ns values result in a higher RPM for a
required mass flow. Designing for a lower Ns value allows a lower RPM for the same mass flow
(or power) but penalizes on efficiency. The higher values of Ns lead to RPM that may not be reachable
(rotor-dynamic and bearing problems) and a lower value of Ns may be a design option to alleviate the
bearing problems. The corresponding penalty on efficiency can be obtained from figure 2.
Figure 2: Variation of Achievable Compressor Efficiency with Specific Speed (acc. [2]).
1.2 Performance Predictions
Previous model has been used to predict the compressor performance at different values of impeller RPM.
The results for high and low values of Ns and high and low values of the turbine inlet temperature are
summarized in table 1 to 4. Each table shows the variation of the gas turbine characteristics and performance
when decreasing the dimensions to finger tip size impellers. The pressure ratio, compressor inlet
temperature T1, Turbine inlet temperature Tit, Specific speed Ns, radial compressor slip factor, Turbine
efficiency etaT, specific heat Cp and Gas constant R
G
, at which the calculations are made, are listed in the
first rows. Calculations are made for a simple cycle without recuperator.
The compressor efficiency is predicted in two steps. The variation of Eta-Comp (column 8) accounts for
the reduction in efficiency with decreasing Reynolds number (Fig. 1). The second step corrects for non-
optimum specific speed (Fig. 2). The global result is listed under Eta-C(corr) (column 9).
The turbine efficiency, needed to define the cycle efficiency, is assumed to be independent of the size and
fixed at a low (70.%) and high value (80.%). The lower value stands for the worst case when the
efficiency is limited by geometrical or manufacturing restrictions (2D instead of 3D impellers).
All predictions are made for pressure ratio = 3. Losses in the piping and combustor are neglected.
An increase in ∆H and hence peripheral velocity U2 is required to compensate for the decrease in
compressor efficiency with decreasing size. The important decrease in cycle efficiency with decreasing
size, shown on Table 1, illustrates the large impact of compressor and turbine performance at moderate
values (1200.
o
K) if Tit.

Micro Gas Turbines – A Short Survey of Design Problems
1 - 4 RTO-EN-AVT-131
Table 1: Variation of Gas Turbine Performance with Size (Tit = 1200.
o
K, Ns = .3)
Table 2 illustrates the favorable influence of an increase in specific speed (from .3 to .5) on compressor
efficiency. Together with an increase of turbine efficiency (etaT = 80% instead of 70.%) it results in a
considerable increase of cycle efficiency. At the smallest dimensions the cycle efficiency has increased to
10% and the gas turbine power has changed from 0. to a small positive value. The mass flow is doubled
and the power output is increased with a factor 5. This comparison illustrates the large impact of compressor
and turbine efficiency on overall performance and emphasizes the need for optimized components [3].
Table 2: Variation of Gas Turbine Performance with Size (Tit = 1200.
o
K, Ns = .5)
Comparing the results on Table 3 with those on Table 1 illustrates the performance improvements when
increasing the Tit from 1200
o
K to 1500
o
K, without any change in compressor and turbine characteristics.
It also results in a considerable increase of the power output without change in mass flow.

Micro Gas Turbines – A Short Survey of Design Problems
RTO-EN-AVT-131 1 - 5
Table 3: Variation of Gas Turbine Performance with Size (Tit = 1500.
o
K, Ns = .3)
The best results are obtained with increased Tit and high compressor and turbine performance as shown on
Table 4. However it is questionable if this performance level is realistic in small gas turbines where the
small dimensions prohibit any blade cooling. Even the use of new materials as SiN will not allow such
temperatures and the achievable cycle efficiency may be closer to the one listed in Table 1. The use of
these unconventional materials for gas turbine rotors requires new manufacturing techniques and may
impose restrictions on the geometry (i.e. two-dimensional rotors with low Ns) and have a negative effect
on performance.
Table 4: Variation of Gas Turbine Performance with Size (Tit = 1500.
o
K, Ns = .5)
The extreme high RPM that are required together with the high temperatures, impose the use of oil free
bearings. Pushing the air-bearings and electrical generators beyond the present limits is a challenging task.
The most advanced knowledge on small high temperature turbomachines is in the field of turbochargers,
operating up to 200.000 RPM, and Auxiliary Power Units (APU). All operations above this limit require
new developments as well in terms of material and manufacturing techniques as in terms of aero-
thermodynamics [4]. There is also no guarantee that the existing flow solvers and turbulence models are
still accurate also for these extremely low Reynolds numbers.
Previous study shows that the design of a small gas turbine with positive power output is a challenging
task.

Micro Gas Turbines – A Short Survey of Design Problems
1 - 6 RTO-EN-AVT-131
2.0 HEAT TRANSFER
Internal heat transfer has an important impact on the performance of the very small turbomachines, used in
micro- and nano- gas turbines. The heat flux from the hot turbine to the colder compressor results in a
cooling of the flow in the turbine and a heating of the flow in the compressor. The performance changes
and can no longer be evaluated by the flow conditions measured at inlet and outlet of the components.
This problem has first been recognized and studied for small turbochargers where it was shown that the
distance between the hot turbine and the cold compressor might have a considerable impact on the flow
conditions [5]. The impact on micro-turbomachinery performance has been discussed by Gong et al. [6]
and Ribaud [7]. Procedures to correct for this internal heat transfer have been proposed [8].
2.1 Prediction Model
The diabatic compressor and turbine performance prediction is based on the assumption that the impeller
friction losses are not changed by the heat transfer. However, heating the flow during the compression
results in a higher outlet temperature, hence lower density at the compressor outlet. The main
consequences are less diffusion than with adiabatic flow and, as can easily be evaluated from the outlet
velocity triangles, lower work input and pressure rise. The latter ones result in a further decrease of the
density at impeller outlet.
In what follows one will assume that the corresponding velocity increase at the outlet is compensated by a
proportional increase of the impeller exit width. As a consequence the velocity is unchanged along the
flow path. The original velocity triangles are re-established and friction losses can be evaluated from the
polytropic efficiency of an adiabatic compression on an equivalent geometry. This can eventually be
verified by a diabatic Navier-Stokes calculation.
Heating the flow during compression has a negative effect on the efficiency because the enthalpy dh
needed for an elementary isentropic compression dP increases with the temperature.
TR
P
dPdP
dh
G
.==
ρ
(6)
Cooling the flow during the expansion in a turbine has also a negative effect on the efficiency because the
energy dh obtained from an isentropic pressure drop dP decreases with decreasing temperature.
The first law of thermodynamics provides the relation for non isentropic diabatic compression:
dQdL
dP
TdS
dP
dh ++=+=
ρρ
(7)
where dL is the heat produced by the internal friction losses and dQ is the amount of heat transmitted
trough the walls (Fig. 3).

Micro Gas Turbines – A Short Survey of Design Problems
RTO-EN-AVT-131 1 - 7
Figure 3: H,S Diagram for a Diabatic Compression (left) or Expansion (right).
Expressing the losses and heat addition as a linear function of the enthalpy rise
dh
hh
L
dL
12
2,1
−
=
dh
hh
Q
dQ
12
2,1
−
=
(8)
and substituting it into (7) provides following relation.
ρ
dP
dh
hh
Q
hh
L
=








−
−
−
−
12
12
12
12
1
(9)
Using the relations for perfect gas to calculate the density and expressing the enthalpy in function of the
temperature change and constant specific heat coefficient Cp, one obtains:
P
dP
TdT
TTCp
Q
TTCp
L
γ
γ
1
).().(
1
12
12
12
12
−
=








−
−
−
−
(10)
Integration from T
1
to T
2
(compressor) results in
1
2
12
1212
1
2
ln
1
).(
1ln
P
P
TTCp
QL
T
T
γ
γ
−
=








−
+
−
(11)
or
µ
1
2
1
2
P
P
T
T
=
(12)
where

Micro Gas Turbines – A Short Survey of Design Problems
1 - 8 RTO-EN-AVT-131
).(
1
1
12
1212
TTC
QL
p
−
+
−
−
=
γ
γ
µ
(13)
Equation (12) is similar to the definition of polytropic efficiency of an adiabatic compression where µ stands
for
p
η
γ
γ
µ
1−
=
(14)
Hence
).(
1
12
1212
TTC
QL
p
p
−
+
−=
η
(15)
η
p
decreases with positive values of Q
12
because this increases both the nominator and denominator.
Equ. (15) shows that any heat addition will have the same effect as a reduction of polytropic efficiency.
The value of L
12
can directly be derived from (15) and the known polytropic efficiency for an adiabatic
compression (Q
12
=0).
(
)
).(1
1212
TTCL
adippadi
−−=
η
(16)
The same procedure provides the non adiabatic efficiency and outlet temperature T
2dia
by substituting the
previously defined value of L
12
and a given value of Q
12
≠0.
Similar equations are easily derived for turbines.
The diabatic outlet temperature (Q
12
≠0) cannot directly be used to calculate the power transmitted by the
compressor and turbine. The mechanical- or shaft power is obtained by subtracting the heat flux from the
total energy transfer, defined by inlet and outlet temperature.
E
C,dia
=
m
&
.C
p
(T
2dia,comp
-T
1
)-Q
12
E
T,dia
=
m
&
.C
p
(T
1
-T
2dia,turb
)-Q
12
(17)
For a given pressure ratio, the real compressor and turbine efficiencies can also be defined from by
substituting following corrected exit temperatures for compressor and turbine in (12):
mCQTT
pdiacorr
&
//
12,2,2
−= mCQTT
pdiacorr
&
//
12,2,2
+
=
(18)
The decrease of efficiency of both components is shown on Fig. 4 and 5 as a function of adiabatic efficiency
and Q
12
between and turbine compressor. Q
12
is expressed as a % of the adiabatic compressor power input
at η
p
= .7 . It depends on the impeller size and fluid temperatures and is evaluated in next section.

Micro Gas Turbines – A Short Survey of Design Problems
RTO-EN-AVT-131 1 - 9
Figure 4: Decrease of Compressor Efficiency in Function of Q12 and η
p
.
Figure 5: Decrease of Turbine Efficiency in Function of Q12 and η
p
.
Input to this model is the internal heat transfer obtained from diabatic flow calculations on compressors of
different size at different impeller surface temperatures.
2.2 Heat Flux Calculations
The amount of heat flux through the compressor walls (Q
12
>0) depends on the compressor wall temperature
which in turn is a function of the turbine inlet gas temperature, the blade and disk surface, the distance
between turbine and compressor impeller, the conductivity of the material, the heat transfer coefficient,
etc. A complete calculation requires a combination of a heat transfer calculation in the whole rotor with a
diabatic flow calculation in the compressor and turbine (Fig. 6).

Micro Gas Turbines – A Short Survey of Design Problems
1 - 10 RTO-EN-AVT-131
Figure 6: Rotor Geometry and View on Compressor Section.
Following is an estimation of typical values of the heat flux in radial compressor impellers of different size
with different wall temperatures (Table 5). The flow and heat flux is calculated by the TRAF3D Navier
Stokes solver developed by Arnone [9] on a grid with 400 000 cells using the Baldwin Lomax turbulence
model. This may not be the most appropriate model but an experimental study of low Reynolds number
flow in a rotating channel with wall roughness is presently under way to evaluate it.
Table 5: Navier Stokes Test Cases
Outlet diameter (mm) 8 20.
RPM
1050000 420000
Mass flow (gr/sec) ~1. ~6.7
P
0
2
/P
0
1
2.1 2.1
T
wall
(
o
K) 600 500/600/700
The 20 mm diameter 2D compressor impeller is a geometrically scaled version of the 8 mm diameter one.
It has four blades and four splitters (Fig. 6).
The model, explained in section 2, assumes the same velocity distribution inside the impeller for adiabatic
and diabatic flows. Hence the inlet and outlet velocity triangles should be conserved during the diabatic
calculations. This is achieved by:
• Adjusting the pressure ratio proportional to the inlet density (dh=dP/ρ) to obtain the same volume
flow and hence the same inlet velocity triangles; and
• Increasing the impeller outlet width b
2
to compensate for the decrease in outlet flow density with
increasing fluid temperature.
The heat flux is defined by integrating the flux on all surface cells (i.e. on the blades, hub and shroud).
Fig. 7 shows the dependence of the calculated heat flux on compressor size and wall temperature. It varies
between 27% of the large compressor power at 500
o
K wall temperature to 62% at 700
o
K wall
temperature. It increases by almost 10% at the smaller impeller size.

Micro Gas Turbines – A Short Survey of Design Problems
RTO-EN-AVT-131 1 - 11
Figure 7: Variation of Heat Flux with Wall Temperature for 8 mm. Ø (◊) and 20 mm Ø (o) Impeller.
The Navier Stokes calculations predict a reduction of adiabatic impeller efficiency, based on shaft power
and pressure ratio, from 69.6% to 63.5% for the diabatic compression. This value is in good agreement
with the one defined by the analytical model.
2.3 Cycle Efficiency
The total energy needed for an adiabatic compression, or obtained from an adiabatic expansion from to
P
2
to P
1
is
).(
12,
TTCmE
adipadiC
−=
&
).(
21, adipadiT
TTCmE
−
=
&
(19)
The mechanical energy required to compress the gas in a diabatic process is defined in (17).
Although the adiabatic and diabatic compression processes suffer from the same friction losses, E
dia
is
larger than E
adi
because the diabatic compression takes place at a higher temperature. For negative values
of Q
12
(cooling of the compressor)
the required energy would be lower than the one specified in (19)
because of the lower fluid temperature during compression. The process is then closer to the more efficient
isothermal compression.
The opposite phenomenon occurs in the turbine. Expanding the flow in the turbine at a lower than adiabatic
temperature, reduces the power output because the available energy, corresponding to a given pressure
drop, decreases with decreasing temperature (6)
diaTadiT
EE
,,
> .
Heating the turbine would result in an increased power output in a way similar to what is expressed by the
reheat factor for turbines.
Less power output from the turbine in combination with more power required by the compressor results in
a lower gas turbine power output
adiGTadiCadiTdiaCdiaTdiaGT
EEEEEE
,,,,,,
=
−<−= (20)
and lower cycle efficiency

Micro Gas Turbines – A Short Survey of Design Problems
1 - 12 RTO-EN-AVT-131
)(
iTiTp
CT
cycle
CTmC
EE
−
−
=
&
η
(21)
C
iT
is the combustion inlet temperature.
Following figures, showing the consequences for a typical small gas turbine rotor, are obtained at following
operating conditions:
P
2
/P
1
= 3. T
1C
=293.
o
K η
pC
= .7 T
It
= 1600.
o
K η
pT
= .8
Fig. 8 shows the ratio of the diabatic over adiabatic power output as a function of the compressor polytropic
efficiency for the typical cases of heat transfer. Q
12
is 25.% 50.% or 75. % of the compressor adiabatic
input power at polytropic efficiency of .7 . The central value is close to the 52% predicted in a micro gas
turbine with rotor diameter 8 mm at 600
o
K compressor wall temperature.
Figure 8: Ratio between Diabatic over Adiabatic Gas Turbine Power
Output as Function of Compressor Polytropic Efficiency.
The heat-transfer from the turbine to the compressor as well as the lower compressor efficiency increase
the compressor exit temperature and hence combustion chamber inlet temperature (Fig. 9). The effect is
comparable to a recuperator and the corresponding increase of the combustion chamber inlet temperature
means that less fuel will be needed to reach the prescribed turbine inlet temperature. This partially
compensates the decreased compressor and turbine efficiency and explains the relatively small impact of
the internal heat transfer on the cycle efficiency (Fig. 10).

Micro Gas Turbines – A Short Survey of Design Problems
RTO-EN-AVT-131 1 - 13
Figure 9: Effect of Internal Heat Transfer
on Cycles (without recuperator).
Figure 10: Effect of Internal Heat Transfer
on Cycles (without recuperator).
This compensation does not occur if a recuperator is used (Fig. 11). The smaller difference between the
turbine- and compressor exit temperature results in a smaller heat recuperation. Results shown on Fig. 12
indicate a non-negligible impact of heat transfer on cycle efficiency assuming a constant (.70) effectiveness
of the recuperator.
Figure 11: Effect of Internal Heat Transfer
on Cycles (with recuperator).
Figure 12: Impact of Heat Transfer on
Cycle Efficiency (with recuperator).
3.0 CONSTRAINTS
Previous rules for turbomachinery scaling are valid only in combination with an exact geometrical scaling,
i.e. scaling all dimensions, including blade thickness and roughness, with the same factor. This may not
always be possible when going to extreme small dimensions. Limitations may occur depending on:
different manufacturing techniques that are required to make the smaller geometries, new materials that
are used or change of lay-out for operational or mechanical reasons.

Micro Gas Turbines – A Short Survey of Design Problems
1 - 14 RTO-EN-AVT-131
Scaling the blade thickness proportional to the impeller diameter will lead to extreme thin blades that have
an insufficient mechanical resistance to shocks or any other mechanical solicitations. Finite element
calculations have also shown that a thickening of the blades and a larger fillet radius at the blade root may
reduce stresses. However CFD evaluations show that a thickening the blades by a factor 4, as shown on
Fig. 13, results in an efficiency drop of more than 6%.
Figure 13: Effect of Blade Thickness on Stresses (different scales).
It is generally accepted that Roughness has no impact on friction losses as long as the Reynolds number
based on the equivalent sand grain size ks is smaller than 100.
100
..
Re
2
<=
µ
ω
ρ
ksR
ks
(22)
The etching techniques, used for nano-gas turbine rotors, do not allow very smooth surfaces which may
limit the validity of Reynolds number influence on efficiency (Fig. 1). It is further not verified that this
rule of thumb (22) is still valid for the very low Reynolds numbers.

Micro Gas Turbines – A Short Survey of Design Problems
RTO-EN-AVT-131 1 - 15
The decrease in efficiency with decreasing size requires higher peripheral velocities at impeller exit to
reach the required pressure ratio. Hence the stress problems increase at reduced dimensions. It has been
shown that shrouding the impellers has a favourable impact on the blade bending stresses but results in
high stresses at the shroud eye (Fig. 14).
Figure 14: von Mises Stresses in Unshrouded and Shrouded Impeller.
Adding a shroud to the impeller has a small favourable effect on efficiency. The friction of the fluid on the
non-rotating shroud wall (with velocity
V) and the losses by the clearance flow, are replaced by a friction
of the fluid on the rotating shroud (with velocity
W) and friction on the rotating shroud wall (with velocity
U) (Fig. 15). The net effect is a small increase of efficiency of the order of 1 to 2% depending on the
thickness of the outer rim.

Micro Gas Turbines – A Short Survey of Design Problems
1 - 16 RTO-EN-AVT-131
Figure 15: Friction on Unshrouded (left) and Shrouded (right) Impeller.
One of the micro gas turbine design targets is a compact machine with high power density. Integrating the
rotor of the electrical power generator into the compressor shroud may help to reach this goal. It also
alleviates vibration problems resulting from long rotors.
4.0 CONCLUSIONS
Good micro gas turbines are not miniaturized copies of large ones.
• The non-adiabatic compression and expansion perturbs the whole thermodynamic cycle.
• Combustion chambers, not yet discussed in this paper, are very different from their larger version
because of unscalable characteristics like combustion time.
The very high rotational speed, that is needed to obtain the enthalpy and pressure changes prescribed by
the gas turbine cycle, is the major mechanical problem.
New materials and new manufacturing techniques are needed. They should allow low cost production
because this small power output devises remain in competition with heavier but cheap batteries.
One of the major problems in micro gas turbines is the decrease of compressor and turbine efficiency with
decreasing dimensions. This is further enhanced by the effect of larger roughness resulting from materials
and manufacturing techniques. The resulting decrease in cycle efficiency does not make micro gas
turbines to ecological devices. Performance more than any other criterion might define the lower limit for
these devises.
ACKNOWLEDGMENT
Part of the results presented in this lecture are obtained in the context of the project SBO 030288
“PowerMEMS” financed by IWT (Institute for Promotion and Innovation of Science and Technology in
Flanders).
The contribution by Z. Alsalihi and T. Verstraete to this lecture is gratefully acknowledged.
REFERENCES
[1] Japikse D. and Baynes N., Introduction to Turbomachines, Concepts ETI and Oxford University
Press, 1994.
[2] Rodgers C., Specific Speed and Efficiency of Centrifugal Impellers, proc. ASME 25
th
Gas Turbine
Conference, 1980.

Micro Gas Turbines – A Short Survey of Design Problems
RTO-EN-AVT-131 1 - 17
[3] Van den Braembussche R.A., Islek A.A. and Alsalihi Z., Aero-thermal Optimization of Micro-Gas
Turbine Compressors Including Heat Transfer, Proceedings of the IGTC2003Tokyo, International
Gas Turbine Conference 2003, Tokyo, November 2-7, 2003.
[4] Nagashima T., Ribaud Y., Ivanov M.J. and Van den Braembussche R.A., Collaborative Research
about Thermo- Fluid-dynamic Design of Ultra –micro Gas Turbine, Journal of the Gas Turbine
Society of Japan, Vol. 30, No 4, 2002, pp. 42-49.
[5] Rautenberg M., Mobarak A. and Malobabic M., “Influence of Heat Transfer between Turbine and
Compressor on the Performance of small Turbochargers”, GTSJ Intl. Gas Turbine Congress, Tokyo,
23-29 November, 1983.
[6] Gong Y., Sirakov B.T., Epstein A.H and Tan C.S., “Aerothermodynamics of micro-turbomachinery”,
ASME-GT2004-53877.
[7] Ribaud Y., “Internal Heat Mixing and External Heat losses in an Ultra Micro Turbine” Proceedings
International Gas Turbine Congress 2003 Tokyo, November 2-7, 2003. IGTC2003Tokyo OS-109.
[8] Rautenberg M. and Kammer N., “On the thermodynamics of non-adiabatic compression and
Expansion Processes in Turbomachines”, 5
th
Intl. Conference for Mechanical Power Engineering,
October 13-15, 1984.
[9] Arnone A., “Viscous analysis of Three-Dimensional Rotor Flow Using a Multigrid Method” ASME
Journal of Turbomachinery, Vol. 116.

Micro Gas Turbines – A Short Survey of Design Problems
1 - 18 RTO-EN-AVT-131