Air Bottoming Cycle: Use of Gas

Turbine Waste Heat for Power

Generation
0. Bolland
This paper presents a thermodynamic analysis of the Air Bottoming Cycle (ABC) as
well as the results of a feasibility study for using the Air Bottoming Cycle for gas
M. Forde turbine waste heat recovery/power generation on oil/gas platforms in the North Sea.
The basis for the feasibility study was to utilize the exhaust gas heat from an
Division of Thermal Energy LM2500PE gas turbine. Installation of the ABC on both a new and an existing
and Hydropower, platform have been considered. A design reference case is presented, and the recom-
Norwegian University of Science mended ABC is a two-shaft engine with two compressor intercoolers. The compression
and Technology, pressure ratio was found optimal at 8:1, The combined gas turbine and ABC shaft
Trondheim, Norway efficiency was calculated to 46.6 percent. The LM2500PE gas turbine contributes
with 36.1 percent while the ABC adds 10.5 percent points to the gas turbine efficiency.
The ABC shaft power output is 6.6 MW when utilizing the waste heat of an LM2500PE
B. Hande gas turbine. A preliminary thermal and hydraulic design of the ABC main components
Oil Engineering Consultants, (compressor, turbine, intercoolers, and recuperator) was carried out. The recupera-
Sandvika, Norway tor is the largest and heaviest component (45 tons). A weight and cost breakdown
of the ABC is presented. The total weight of the ABC package was calculated to 154
metric tons, and the ABC package cost to 9.4 million US$. An economical examination
for three different cases was carried out. The results show that the ABC alternative
(LM2500PE + ABC) is economical, with a rather good margin, compared to the
other alternatives. The conclusion is that the Air Bottoming Cycle is an economical
alternative for power generation on both new platforms and on existing platforms
with demand for more power.

Introduction bine/steam turbine cycle. It has in the past not been paid much
attention to, and the main reason for that is probably the rapid
The driving force in Norway for the interest in the Air Bot-
technological improvement during the last two decades and the
toming Cycle (ABC) is the C0 2 tax on the use of fuel in gas
commercial success of the combined gas turbine/steam turbine
turbines on oil platforms in the North Sea. This C 0 2 tax is
currently 11,8 cent/Sm 3 fuel gas (or 0.80 NOK/Sm 3 gas). The cycle.
annual payment of C0 2 tax from one LM2500PE gas turbine What Is the Air Bottoming Cycle? Like the steam bot-
is approximately 4.4-5.1 million US$ (30-35 million NOK). toming cycle in a combined cycle, the ABC may utilize heat
The gas consumed by gas turbines on the oil platforms could rejected from a gas turbine. A general flowsheet diagram of the
alternatively be sold, and the value would be to approximately ABC is shown in Fig. 1. Ambient air (1) is drawn through a
7.4-8.8 cent/Sm 3 (0.50-0.60 NOK/Sm 3 ). The C0 2 tax and filter and is compressed in the low-pressure compressor (LPC).
the sales value of the gas then amount to as much as approxi- The air is then cooled ( 3 - 4 ) before it is further compressed in
mately 20 cent/Sm 3 . an intermediate compressor (IPC). Again the air is cooled ( 5 -
Reduced fuel consumption can be achieved by using a com- 6) before the final stage of compression in the high-pressure
bined gas turbine/steam turbine cycle. This type of combined compressor (HPC). In Fig. 1 the process shown has the com-
cycle has gained widespread acceptance in the land-based power pression divided into three sections with two intercoolers. It is
generating industry and is definitely a "proven technology." of course possible instead of having two intercoolers to use one
On oil platforms in the North Sea, a number of studies on using or three or even more intercoolers. This is a question about
combined cycles have been conducted, but the typical result of what is practical and a balance between cost and benefit of
these has been that a combined gas turbine/steam turbine cycle having more intercoolers.
is not economical, despite the high C 0 2 tax. Even if such a The compressed air (7) is heated in a heat exchanger before
combined cycle offers a high fuel-to-power conversion effi- it enters (8) a turbine. In the turbine the air is expanded while
ciency, the cost due to the high weight of the equipment (espe- shaft work is generated. After the turbine the air is exhausted
cially the heat recovery steam generator) has been found to be to a stack. The work generated in the turbine is sufficient to
too high. The cost of weight of equipment is very high on an drive the compressor and a generator. A gear between the tur-
oil platform, and therefore implementation of fuel reduction bine shaft and the generator may be necessary.
initiatives should take place with lightweight equipment. The ABC offers the possibility to provide to an external
The ABC together with a topping gas turbine cycle is another process compressed air at moderate temperatures (la) and/or
type of combined cycle. It offers an efficiency close to that of compressed hot air (8a). The ABC may be a closed or partially
a combined gas turbine/steam turbine cycle, and offers the closed cycle by recycling the air (9a).
potential for lower weight compared to the combined gas tur- In the literature extensive surveys can be found on the use
and thermodynamic potential of a variety of power cycle work-
ing fluids where heat sources are latent and/or sensible heat
Contributed by the International Gas Turbine Institute and presented at Cogen- (Pettersen, 1987; Hornnes and Bolland, 1991). However, air
Turbo Power, Vienna, Austria, August 23-25, 1995. Manuscript received by the
International Gas Turbine Institute June 20, 1995. Associate Technical Editor: as a bottoming cycle working fluid has not been paid much
C. J. Russo. ASME Paper No. 95-CTP-50. attention to. Most notable are probably the patents by Farrell

Journal of Engineering for Gas Turbines and Power APRIL 1996, Vol. 1 1 8 / 3 5 9
Copyright © 1996 by ASME
Downloaded From: https://gasturbinespower.asmedigitalcollection.asme.org on 07/02/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use
 AIR RECYCLING OF AIR. . . _ 9a f TO STACK will be 21-22 MW and the efficiency approximately 42 percent
INTERCOOLERS at full load and with favorable part-load performance (Shepard
et al., 1995; Hawkins et al., 1994).

9 MECHANICAL
Computational Models and Design Methods
WORK In the following, models and design philosophy are presented
for the present work. These include models for both component
design and cycle analysis.
Cycle Analysis. The ABC uses a gaseous working fluid,
COMPRESSED HOT which implies both gas compression and expansion (a Brayton
AIR TO PROCESS cycle). In connection with recovery of sensible heat from a gas
turbine topping cycle, the disadvantage with the ABC is that
the compression of the working fluid restricts how much heat
can be recovered. In the ABC the heat is picked up between
Fig. 1 A schematic diagram of a possible configuration of the air bot- the compressor and the turbine, and it is actually the compressor
toming cycle discharge temperature that determines how much heat can be
recovered. To overcome this restriction, the compressor may
be intercooled whereby the compressor discharge temperature
(1988, 1992). Also others like Pettersen (1987), Wicks et al. is lowered.
(1992), Walker et al. (1992), and Weston (1993) have pro- In the present work the flow rate of the ABC is determined
posed and discussed bottoming cycles with air as the working by fixing the mcP ratio between the cold and hot sides to unity.
fluid. The reason for this is that for a given power output from the
ABC, the heat transfer area required is at a minimum when this
Alternatives to Using the ABC. On oil platforms there ratio is unity.
will in the next years be a few options to using simple-cycle
gas turbines for power generation. One of them is the ABC as Traditional first-law analysis based upon component perfor-
described in this paper, but there are also other alternatives that mance characteristics coupled with energy balances invariably
may be competitive with the ABC. These are: leads to a correct final answer. However, such analysis cannot
locate and quantify the sources of loss that lead to that result.
High-temperature, high-pressure-ratio simple-cycle gas tur- This is because the first law embodies no distinction between
bines: The development of gas turbines has now come to the work and heat, no provision for quantifying the quality of en-
point where efficiencies in excess of 40 percent are achievable ergy. These limitations are not a serious drawback when dealing
in the near future. Examples are the LM6000PA (s=»40 MW), with familiar systems, since an intuitive understanding of the
the FT8 (~26 MW), and the Trent («51 MW). different parametric influences on system performance and sec-
ond-law qualitative appreciation of "grade-of-heat" and effect
Combined gas turbine/steam turbine cycles: Current de- of pressure loss can be developed. When analyzing novel and
sign practice is not suitable for being applied on oil platforms. complex thermal systems, however, such an understanding
However, there is a potential toward a more compact light- should be supplemented by more rigorous quantitative methods.
weight design, especially for the heat recovery steam generator Second-law analysis, or exergy analysis, provides these tools.
(Saether, 1993). The use of the "once-through" concept, to- Second-law analysis is no substitute forfirst-lawanalysis; rather
gether with close packing of short finned tubes (serrated fins) it is a supplement. In the present work results using both first
with a small diameter, enables a significant weight reduction. and second-law calculations are presented.
The combined gas turbine/steam turbine cycle offers the The exergy of a flow stream for a given pressure (pt) and
highest efficiency potential for the power generating alternatives temperature (Tx) is given by Eq. (1):
on oil platforms, which at present would be in the range 47-
49 percent. e(.P\, Ti) = hi - ha- Ta(sx - sa) (1)
Intercooled recuperated gas turbine: Rolls-Royce and where the index a denotes the ambient condition. The term (st
Westinghouse are now developing an intercooled and recupera- - sa) can for an ideal gas, which is the case for the ABC, be
tive gas turbine (WR-21) for the US Navy. The power output written

Nomenclature
c = chord, m Rg = gas constant, kJ/(kg K) tp = flow coefficient
CL = lift coefficient s = entropy, kJ/(kg K) U) = pressure loss coefficient
CM = Mach number correction factor t = tip clearance, m Ap2/pV2
cP = specific heat capacity, kJ/(kg K) T = temperature, K
Deq = equivalent diffusion ratio U = tangential velocity at mean line, Subscripts and Superscripts
e = exergy, kJ/kg m/s 0 = stagnation
h = enthalpy, kJ/kg V = absolute velocity, m/s 1 = inlet
H = wake form factor W = relative velocity, m/s 2 = outlet
hB = blade height, m Y = specific work for a turbine stage, a = ambient
i = incidence, deg J/kg p = profile
M = Mach number f3 = flow angle, deg S = dynamic
m = flow rate, kg/s 6 = momentum thickness, m tip = tip clearance
ns - specific speed, rad/s p = density, kg/m 3 w = annulus
p = pressure, Pa a = solidity (chord/spacing) Z = axial direction
Q = volumetric flow rate, m3/s \fj = pressure coefficient oo = vector mean velocity
R = radius, m to = rotational speed, rad/s = averaged

360 / Vol. 118, APRIL 1996 Transactions of the ASME

Downloaded From: https://gasturbinespower.asmedigitalcollection.asme.org on 07/02/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 C7> dT „ f"> dp neglected here. The profile losses are corrected for Mach num-
J
ra T JPa p ber effects by using the following:

To quantify the loss of exergy for a component, an exergy CM = 1 + aM2 (6)

balance around the component is applied. The difference in where CM = correction factor for Mach number effects, a =
exergy between the ingoing and outgoing streams will then be 0.1 (stator) and a = 0.3 (rotor). Equation (6) is a rough curve
the loss of exergy. The loss of exergy is caused by heat ex- fit of Koch and Smith's method (1976) for solidity of 1.35,
changer temperature differences, "mismatched" heat ex- maximum thickness-to-chord ratio of 0.06, trailing-edge-thick-
changer temperature profiles, heat losses, pressure losses, mix- ness-to-chord ratio of 0.01, diffusion factor of 0.45, and Reyn-
ing losses, mechanical losses, generation of entropy in compres- olds number of 106.
sion and expansion processes, and rejection of exergy to the The correlations used for annulus u>„ (Eq. (7)), secondary
surroundings (Bolland, 1991). u5sec (Eq. (8)), and tip clearance G3lip (Eq. (9)) losses are based
on the recommendations of Grieb et al. (1975). The loss factors
Compressor. A mean line prediction method (MLPM) was are assumed to be independent, and the total loss factor is the
used to model the flow through the compressor. NACA 65 sum of individual loss factors (Eq. (10)).
profiles are assumed at this preliminary design stage. Each blade
row forms a control volume, which has to satisfy conservation
of mass, momentum, and energy. Viscous effects are included U)w = 0.01 — \[~ (7)
by the use of empirical correlations. The method demands that
every blade row be described with inlet/outlet angles, blade
height, spacing, chord, and outer/inner radius. The weakness 1 W
of the method is that three-dimensional effects (including hub = 0.016Q — (8)
hBJ cos(P) \W2
and tip effects) are only taken into account by using rough
empirical laws. Uniform axial velocity in each cross section
was assumed. This can be achieved with a free vortex tangential 1 w
u>tip = 0.60CL (9)
velocity distribution. The intention here is to list the empirical hB)\hBj cas(P) \W2
correlations used in the compressor flow model. Further details
concerning the correlations can be found from Hande (1992). U> - W„ + ljjw + UJ..C + LJ„, (10)
The losses in a compressor blade row can be categorized into
profile losses, annulus losses, secondary losses, and tip clear- Surge inception is determined by the use of the effective static
ance losses. pressure rise coefficient as described by Sovran and Klomp
The profile losses are caused by boundary layer development (1967), Koch (1981), and Hande (1992).
on the blade surfaces. A highly recognized method for pre- The effective flow area is the geometric area times a
dicting these losses at design point operation was developed by blockage factor. This blockage factor accounts for the effects
Lieblein (1959). The profile total loss coefficient depends on of blade thickness, boundary layer, and wakes. This factor was
the momentum thickness growth on the blade surface, which set to 0.9.
in turn is closely related to the diffusion of the blade. The Turbine. The ABC turbine is designed by using a mean
model of Lieblein was modified by assuming that zero incidence line prediction method base on one-dimensional parameters.
corresponds to the minimum loss condition, and by using the The parameters are; the pressure coefficient i/' (or blade loading
axial velocity correction suggested by Klapproth (1959), the coefficient) (Eq. (11)), the flow coefficient <j> (Eq. (12)), and
corrected equivalent diffusion ratio with the assumptions of the specific speed ns (Eq. (13)).
NACA 65(A10) blades becomes:
Cp{Toi — TQ2)
(ii)
1.12 + 0.0117(()' 43
^W^ if/ 2

4> = ^ (12)
+ 0.61 " ^ (tan A - ^ t a n / 3 2 ) (3)
U
Vzi 7.
.Qx (13)
The relation between momentum thickness 6 and Deq is:
where the indices 1 and 2 denote in and out of a stage, respec-
- = 0.005 + 0.01 (Dtq - 1.3) + 0.025(De, - 1.3)6 (4) tively.
c The values for isentropic efficiency as a function of flow and
pressure coefficients are taken from Smith's diagram (Sayers,
The high power of the last term gives the sudden rise in momen- 1990). This diagram gives the efficiency at the design point.
tum thickness caused by boundary layer separation at high Deq. For this specific study Smith's diagram is applied in a rather
The relation between momentum thickness and profile loss coef- conservative manner. The turbine polytropic efficiency is esti-
ficient is: mated to 90 percent (total-to-total), which is 1-2 percentage
points below that of the diagram for each stage.
2ft For the three-dimensional design of the turbine stage, the
cos (Pi) 3H2 - 1 free-vortex method is applied, which gives constant work from
o>„ = 2C«,
c J cos (Pi) \cos (P2) hub to tip over the blade. Furthermore, at this stage of the
preliminary design it is assumed that the blade height is equally
c) cos (P2) distributed on each side of the mean diameter of the blade. This
may have to be redistributed in the final design analysis.
(5)
Intercoolers. The intercoolers are supposed to be of the
The form factor effect, the last fraction in Eq. (5), contributes "compact heat exchanger" type. The cold and hotfluidsflow
in most cases less than 10 percent of the profile losses and is in finned channels arranged next to each other, and they are

Journal of Engineering for Gas Turbines and Power APRIL 1996, Vol. 1 1 8 / 3 6 1

Downloaded From: https://gasturbinespower.asmedigitalcollection.asme.org on 07/02/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 separated by thin plates. The heat transfer surface geometry and 8000
design method are based on the work of Kays and London 48.5-i Z' A
\d&
7600
(1984). 48 0- One intercoo er :
7200
47.5: n
Exhaust Gas to Air Heat Exchanger. The heat of the gas 47.0-; V LJ 6800
turbine exhaust is recovered and transferred to the compressed
air between the compressor and turbine of the ABC. For this
application there are two different types of heat exchanger that
tj
46.5:
46.0: •4 6400

can be used: regenerators and recuperators. In a recuperator, 1 45.5 i • — &

;
6000
5600
I
1
heat is transferred through walls that separate the flows. In a 45.0:
regenerator, the heat transfer surface is alternately exposed to 44 5 ^
i ;
;
5200
the two flows. Kays and London (1984) list three advantages 44.0 H
00 4800
that regenerators have over recuperators: 43.5 -:
1 —J. 4400
1 A much more compact heat transfer surface can be em-
ployed.
i 43.0: *
42.5 H~-&
J
/ Lu
;
4000
42.0 3600
2 The heat transfer surface is substantially less expensive
per unit of transfer area. 2 3 4 5 6 7 8 9 10 11 12 13
Compressor pressure ratio
3 Because of the periodic flow reversals, there are no per-
manent flow-stagnation regions and, consequently, the Fig. 2 Efficiency and ABC power output as function of compressor pres-
surface tends to be self-cleaning. The self-cleaning feature sure ratio for no intercooling, one and two intercoolers
has been demonstrated by Ljungstrom-type air preheaters
used in central station plants employing lowest grade of
coal. As for the intercoolers, the heat transfer surface geometry
and design-method for the recuperator are based on the work
A fourth advantage is that regenerators require only light- of Kays and London (1984).
gage containment for the low-pressure exhaust gas over most
of their volumes. Recuperators require heavy-gage containment ABC Cycle Analysis
over their entire volumes. Kays and London (1984) also list
two disadvantages of regenerators compared to recuperators: Number of Intercoolers. From a thermodynamic point of
view the ABC should have an isothermal compression. This is
1 There is some mixing of the hot and the cold fluids be- of course difficult from a practical point of view. However, a
cause of leakage and carryover. larger number of intercoolers is wanted, up to a point where
2 If the fluids are at different pressures, as in the ABC, the the benefit of an extra intercooler is balanced with the parasitic
sealing problem of a regenerator is a difficult one. losses. When considering practical design and construction of
the ABC, the question reduces to whether one or two intercool-
Additionally one can say that recuperator design for this type ers should be used. A comparison between different numbers
of application and size range is a much more proven technology of intercoolers and compressor pressure ratios is shown in Fig.
compared to that of regenerators. 2. In the present work two intercoolers are chosen as well as a
Some fluid mixing is acceptable for regenerators in gas tur- total compressor pressure ratio of 8:1 (including the intercooler
bine systems; however, leakages have been unacceptably pressure losses), which are based on the results presented in
high—over 4 percent of the compressor flow in many cases. Fig. 2. With one intercooler the ABC power output peaks at a
This is because the seal designs have been poor. This is why pressure ratio of approximately 6:1. With no compressor in-
recuperators have been favored over regenerators for gas turbine tercooling, the combined gas turbine/ABC efficiency peaks at
systems. With a proper seal design it is possible to have leakages a pressure ratio of approximately 4:1. The difference in com-
at about 3 percent. An air leakage of 3 percent reduces the bined gas turbine/ABC efficiency between using one or two
combined gas turbine and ABC efficiency approximately 1 per- intercoolers is slightly above 1 percent.
centage point.
Number of Spools. The ABC engine can be designed as a
The core of the regenerator is smaller and weighs less than single-shaft or two-shaft engine. In that respect an important
the core of a recuperator. The slats in a regenerator are not point is how the topping gas turbine will be operated and what
required to support any pressure difference, and may be as thin part-load performance is required. In order to fit in as an advan-
as 25 /jm. The plates of a recuperator are typically about 200 tageous option on oil platforms, it is unquestionable that the
/im thick (Wilson and Kupratis, 1994). The fact that the regen- ABC should be able to operate well at gas turbine part-load.
erator's slats are eight times thinner makes the core at least
In the following a general qualitative comparison is made
eight times lighter (Beck, 1994). Second, the thinner slats offer
between a steam bottoming cycle and the ABC at part-load
a smaller area for core heat conduction in the flow direction.
operation of the topping gas turbine. At part-load operation of
Conduction reduces heat exchanger effectiveness, so the flow
the topping gas turbine (assuming a multispool engine) both
length in a plate-fin recuperator must be longer to compensate
exhaust gas flow rate and temperature are decreased compared
for the effects of conduction. This means that to achieve the
to full-load operation. Both these types of bottoming cycle con-
same pressure drop across the core, the recuperator's plates
sist of three major process steps that influence the performance:
must be spaced farther apart than the regenerator's slats, and
compression, heat exchange from the topping to the bottoming
this will increase the size of the recuperator.
cycle, and expansion.
In the present work it was chosen to go for a recuperator and
not a regenerator. The most important reasons for this decision Compression: The work of compression (feed water
are the efficiency reduction caused by air leakage from a regen- pumps) is small in a steam bottoming cycle and may be regarded
erator, and that large recuperators are a more proven technology as unchanged at part load because of its small influence on plant
compared to regenerators for this application. When or if regen- efficiency. In a gas turbine cycle, the work of compression
erators become a more mature technology in this size range, it typically constitutes half of the turbine work (as high as 70
is indeed an interesting option for the ABC because of the high percent for the ABC). At part load it is highly important for
NTU number required in order to obtain a high cycle efficiency. an axial compressor to have the optimal combination of speed,

362 / Vol. 118, APRIL 1996 Transactions of the ASME

Downloaded From: https://gasturbinespower.asmedigitalcollection.asme.org on 07/02/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 COOLING WATER flow rate and pressure ratio with respect to efficiency. If a
INTERFACE single-shaft engine is used, meaning more or less constant flow
rate and low air temperature, the pressure ratio at part load is
reduced significantly. If using a single-shaft engine, the differ-
ence between the compressor discharge temperature and the
turbine inlet temperature will be small at part load and thus
result in a small shaft work. The pressure ratio is also reduced
significantly at part load for a two-spool engine because the
flow rate reduction is more important than the more or less
maintained turbine inlet temperature compared to using the sin-
gle-shaft engine (according to the choked nozzle equation, the
turbine inlet pressure is proportional to flow rate and propor-
tional only to the square root of temperature). Concerning the
part-load performance of the turbine, it is advantageous with a
reduction in both flow rate and temperature (two-spool engine),
instead of only temperature reduction at constant flow rate (sin-
gle-spool engine).
It can be concluded that a combined gas turbine/steam tur-
bine plant will operate at higher part-load efficiencies than the
combined gas turbine/ABC. In order to reduce this difference,
a possibility would be to use variable first row stator vanes in
the turbine. Next it can be concluded that the part-load perfor-
mance of a two-spool ABC engine will clearly be better than
a single-shaft engine.
From a more general point of view, one can say that if the
topping gas turbine is a two-spool (or three-spool) engine, the
LM2500PE
ABC should be a two-spool engine. On the other hand, if the
topping gas turbine is a single-spool engine, the ABC also
should be single-spool in order to match the flow rate of the
ABC to that of the topping cycle.
Fig. 3 Flowsheet diagram for the LM2500PE gas turbine and ABC
Results From the Cycle Calculation. In the following the
final cycle design is presented. The assumptions as well as site
flow rate, and pressure ratio. Otherwise the efficiency drops off conditions for the cycle analysis are given in Appendices A.l-
significantly. A two-shaft engine will in principle be better than 3. For the cycle calculations it was emphasized to use accurate
a single-shaft engine in this respect. A single-shaft compressor models for the thermophysical properties of the air (Olikara
would have to run at constant speed, and this is of course a and Borman, 1975). The referendum temperature for enthalpy
disadvantage at part load. The part-load performance of the of air was set to 25°C. A cycle flowsheet diagram is shown in
ABC compressor was estimated and it was found that the com- Fig. 3.
pressor has very steep constant-speed lines in a compressor map A summary of the plant performance is given in Table 1.
diagram. At part load the pressure ratio will decrease, but if The pressure ratios for the different compressor modules were
constant speed is assumed, the compressor flow rate is more or chosen after an optimization with respect to cycle efficiency.
less maintained. If the flow rate is maintained at constant speed The shaft power output from the ABC was calculated to 6.6
and reduced pressure ratio, the compressor efficiency decreases MW. This extra power increases the power generation efficiency
significantly. The steep constant-speed line in the compressor from the simple cycle gas turbine efficiency of 36.1 percent to
map is caused by the intercoolers, which reduces the cumulative a combined cycle shaft efficiency of 46.6 percent. The ABC
effects of "disturbances" through the compressor. Even if an adds 10.5 percent points to the simple-cycle gas turbine effi-
axial compressor is operated at the most favorable conditions ciency.
at different loads, the compressor efficiency is reduced to some
extent and thus accordingly reducing the plant efficiency.

Heat transfer from the topping to the bottoming cycle: For

Table 1 A summary of the plant performance data
a steam bottoming cycle, the steam generation is reduced as
well as steam pressure (assuming sliding pressure operation)
Gas turbine net power output 22643 kW
and steam temperature. The steam generation is mainly a func-
ABC shaft power output 6605 kW
tion of heat transfer. The heat exchanger in the ABC acts in
ABC power output at generator terminals 6407 kW
principle different from a steam generator in the sense that it
is not the heat transfer that decides the flow rate on the cold ABC auxiliary power requirements 67 kW
side but rather the compressor. If using a single-shaft ABC ABC net power output 6340 kW
engine with a more or less a constant flow rate at part load, the Combined gas turbine/ABC power output 28983 kW
air exit temperature from the heat exchanger is reduced a lot Gas turbine fuel consumption 1.3132 kg/s
more than the gas turbine exhaust temperature. For a two-shaft Gas turbine efficiency 36.10 %
ABC engine where the flow rate is reduced at part load, the ABC cycle efficiency1 23.23 %
heat exchanger air exit temperature is maintained closer to the ABC cycle efficiency2 10.53 %
exhaust gas temperature. Combined gas turbine/ABC shaft efficiency 46.63 %
Net combined gas turbine/ABC efficiency 46.21 %
Expansion: A steam turbine operates over a wide range of 1
Ratio between ABC shaft power output and the heat transferred from
flow rates at rather high part-load efficiencies (sliding pressure the gas turbine exhaust
2
operation). The ABC turbine with its low pressure ratio (com- Ratio between ABC shaft power output and the gas turbine fuel
pared to a steam turbine) is much more sensitive to changes in energy input

Journal of Engineering for Gas Turbines and Power APRIL 1996, Vol. 1 1 8 / 3 6 3

Downloaded From: https://gasturbinespower.asmedigitalcollection.asme.org on 07/02/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Table 2 The balance of energy flows going in and out of the ABC "con- • Reasonable constant head per stage
trol volume"
• First-stage inlet relative Mach number between 0.75 and
Gas turbine exhaust gas heat (in)1 37825.1 kW 0.8
• Equivalent diffusion ratio (A.,) below 1.7 for all blades
ABC net power output 6340.5 kW
• Constant or slightly decreasing axial velocity through
Generator losses 198.2 kW each unit,
Mechanical losses 46.6 kW
Auxiliary power requirements 66.5 kW Chord and spacing have been selected by careful study of
Compressor inlet air flow 1231.4 kW gas turbine axial compressor design. Stator blade angles have
LP intercooler heat rejection 3510.6 kW been determined to give zero incidence for both stator and
HP intercooler heat rejection 4942.0 kW rotor. No use of variable geometry has been evaluated. The
Heat loss from recuperator & ducts 56.9 kW computational results are given in Appendix A.4.
ABC turbine stack heat rejection 12045.1 kW As an alternative to the current design, it would be possible
Gas turbine exhaust stack heat rejection 9387.6 kW to use a centrifugal HPC instead of axial type compressor. This
Sum of outgoing flows (out) 37825.2 kW would be a simpler and less expensive design, though the perfor-
1 mance would be somewhat poorer compared to an axial HPC.
The referendum for enthalpy by which the heat is calculated, is 25 °C
Turbine Design. The preliminary design for the turbine
covered two different options; a one-spool engine running at
A complete energy flow breakdown for the ABC "control 8500 rpm, and a two-spool engine with a HP turbine at 8500
volume" is given in Table 2. It is shown how the heat trans- rpm and a LP turbine (or power turbine) at 3000/3600 rpm.
ferred in the recuperator is converted to other energy flows. For the second option the HP turbine has to be designed to
In Table 3 an exergy loss breakdown is shown. The compres- match the compressor work, which reduces the freedom in
sors constitute 14 percent of the total losses, the intercoolers the design of the different turbine stages. As discussed earlier,
13 percent, the recuperator with ducting 11.1 percent, the tur- it was concluded that the two-spool alternative is the pre-
bine 11.2 percent, and the stack losses (both from the ABC and ferred one.
the gas turbine) constitute 47.3 percent of the total losses. The The pressure coefficient ("J) is set between 2.0 and 3.0 and
exergy efficiency of the ABC is 36.8 percent when using the the flow coefficient (4>) target is approximately 0.6. Keeping
gas turbine exhaust exergy as a reference. the pressure coefficient as high as possible reduces the blade
When comparing the numbers in Tables 2 and 3 one can tip diameter but increases the blade loading and thereby reduces
easily distinguish between what appears to be the losses given the polytropic efficiency. The flow coefficient has to be set to
by a traditional first-law analysis, and the true losses resulting an optimum between blade height and meridional flow velocity.
from a second-law analysis. Lowering the flow coefficient gives a lower meridional velocity
A sensitivity analysis was carried out, and the main conclu- and thereby lower losses related to velocity, but on the other
sion is that for the final design of the ABC engine, engineering hand the blade height will be larger and imply a more compli-
efforts should be put into designing for highest possible turbine cated blade design.
and compressor efficiencies. For the 3000 rpm power turbine the maximum exit turbine
diameter is approximately 1.5 m and including housing and
Component Design dressing the maximum diameter can be as much as 1.8 m,
For the 3600 rpm power turbine the exit turbine diameter is
Compressor Design. The compressor design process ap- approximately 1.3 m and including housing and dressing the
plied consists of two phases, determination of main dimensions maximum diameter will be 1.6 m. The 3000 rpm two-spool
and the more detailed preliminary design. In the design of the option is about 0.6 m larger in diameter than a one-spool turbine
main dimensions, one determines number of stages, speed, and (at 8500 rpm) would be, and the 3600 rpm two-spool option is
diameter by using dimensionless design parameters. The deter- 0.4 m larger. The main geometric parameters for the ABC tur-
mination of these parameters is based on: recommended design bine are shown in Fig, 4.
rules as found in the literature (0verli, 1992) and predicted
values from actual compressor designs. The number of stages
is determined by selecting a maximum pressure ratio (or head) Table 3 Exergy balance of the ABC
per stage. In this case a maximum pressure ratio of 1,25 per
stage was used. This resulted in three stages each for the LPC Gas turbine exhaust gas exergy (in)1 17222.4 kW
and IPC and four stages for the HPC. Inner and outer blade ABC net power output 6340.5 kW
radius is determined by selecting the inlet hub/tip ratio 0.6 and Generator losses 198.2 kW
the inlet axial Mach number 0.45. Mechanical losses 46.6 kW
The rotational speed can be found by selecting a flow coeffi- Auxiliary power requirements 66.5 kW
cient (axial velocity/rotational velocity). Recommended values Compressor inlet air flow 0 kW
0.4-0.5 for the low-pressure stages and 0.4-0.3 for the high- Air filter 57.7 kW
pressure stages. The ABC compressor consists of three serial LP compressor 494.9 kW
coupled units. These units will be coupled on the same shaft, LP intercooler2 564.1 kW
running with the same speed, 8500 rpm. As a compromise the IP compressor 430.6 kW
selection of design speed was based on the IPC unit. As a HP intercooler2 863.3 kW
result the design speed for the LP unit is somewhat higher than HP compressor 595.5 kW
optimum and the design speed for the HP unit is slightly lower Recuperator & ducts3 1206.4 kW
than optimum. The main dimensions of the flow path of the Turbine 1215.8 kW
compressor is shown in Fig. 4. ABC turbine stack loss 3068.3 kW
The preliminary design is based on the main dimensions, and Gas turbine exhaust stack loss 2074.1 kW
is obtained with the mean line prediction method as previously Sum exergy losses & net work (out) 17222.4 kW
described. The purpose is to determine rotor blade angles, stator
blade angles, blade height, blade chord, and blade spacing. 2 Ambient condition: 7=281.15 K (8 °C) and p=1.013 bar.
Includes both rejection of heat and pressure loss
The following design criteria have been used to determine ' Includes the heat transfer loss and pressure losses for both air and
the blade geometry: exhaust gas.

364 / Vol. 118, APRIL 1996 Transactions of the ASME

Downloaded From: https://gasturbinespower.asmedigitalcollection.asme.org on 07/02/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 2 STAGES 2 STAGES
H

Fig. 4 A sketch of the ABC engine

Intercooler Design. A multipass crossflow configuration mcP of the hot and the cold fluids. For a gas turbine with air
was chosen for both intercoolers. The compressor air makes cooling, a certain percentage of the compressor air (the cooling
one pass and the cooling water makes two passes in each in- air) does not go through the recuperator. The specific heat ca-
tercooler. An alumina alloy, AA 6082 T6, was chosen as the pacity of exhaust gas is about 4 - 5 percent higher than that of
core material (Solheim, 1994). This alloy is often used in off- the air at relevant temperatures. The mcP ratio will probably be
shore applications. The density is 2710 kg/m 3 and the thermal somewhere between 0.75 and 0.85 for a gas turbine recuperator,
conductivity is 170 W/(m K). The intercoolers are designed while for the ABC this ratio was found optimal at unity. When
as modules; where 12 modules form a ring around the compres- comparing these two recuperators for the same effectiveness
sor body for each of the two intercoolers (see Fig. 5). The and same heat transfer, the NTU number will be significantly
main design data for the intercoolers (1 out of 12 modules) are higher for the ABC recuperator. This means that the ABC recu-
shown in Table 4. The air side pressure drop was allowed to perator will be notably larger than a gas turbine recuperator.
be 2 percent for each of the intercoolers. Nearly half of this The details from the present recuperator design are shown in
pressure drop is related to the heat exchanger core, while the Table 5. Surface type P9 or 9.03 was chosen for the exhaust
remaining pressure drop is due to diffuser, bends, and headers. gas side and P13 or 19.86 for the air side (Kays and London,
1984). For the current design an alloy steel was chosen instead
Recuperator Design. In the present study a design of a of ordinary steel (conductivity of 20 W/(mK)). The WR-21
recuperator for the ABC was made. A comparable design is recuperator is going to be made of 14Cr4MO stainless steel,
that of the recuperator, which is going to be used in the WR- which was preferred to the more expensive Incoloy 800 (Shep-
21 intercooled and recuperated gas turbine from Rolls-Royce ard et al., 1995). If there will be more corrosion than expected,
and Westinghouse (Ricket, 1994; Shepard et al, 1995; Hawkins the rather expensive INCO 625 is the preferred material. The
et al., 1994). The design made in the present study is to a large recuperator core weight and volume were in the present study
extent similar to that of the WR-21. The WR-21 recuperator is calculated to 42.4 metric tons and 53.6 m3, respectively. Efforts
designed as a plate-fin heat exchanger with an effectiveness of were made to design a compact recuperator, which is evident
approximately 90 percent and weighing approximately 18 met- when considering the relatively high number for heat transfer
ric tons. The main difference between the WR-21 recuperator area per volume.
and the ABC recuperator is the ratio between the heat capacities
In Fig. 6 the influence of recuperator effectiveness on cycle
efficiency and recuperator weight is shown, where an mcP ratio
of 1 is assumed. There are two curves for the recuperator weight;
ONE one for alloy steel and one for ordinary steel. The difference in
MODULE weight between these two options is around 6 metric tons for
a recuperator effectiveness of 90 percent. As the recuperator
effectiveness is increased, the cycle efficiency increases lin-
early. The recuperator weight increases exponentially as the

Table 4 Intercooler design data: The numbers are for one module out
of 12 parallel ones for both the LP and HP intercooler

LPI HPI
air water air water
DIMENSIONS (mm) Flow rate [kg/s] 5.994 4.208 5.994 6.046
LH HE! Inlet pressure [bar] 1.897 2.0 3.561 2.0
A = 383 A = 486 Inlet temperature [°C] 72.68 13.0 93.48 13.0
B = 541 B = 370 Film coeff. [W/(m2 K)] 149.9 2383 225.3 2447
C = 650 C = 650 Fin efficiency [-] 0.905 0.693 0.865 0.687
D = 520 D = 730 Mass flux [kg/(m2 s)] 37.98 121.1 59.45 112.7
E = 220 E = 210
F = 1300
Reynolds number [-] 6981 338.1 10495 314.7
F = 1323
G = 1450 G = 1420 Total film coeff. [W/(mz K)] 121.0 166.7
Resistance ratio air/water-side 7.4 5.22
Fig. 5 A view of the intercooler modular design in the axial direction of
the ABC engine
Core volume [m3] 0.145 0.119

Journal of Engineering for Gas Turbines and Power APRIL 1996, Vol. 1 1 8 / 3 6 5

Downloaded From: https://gasturbinespower.asmedigitalcollection.asme.org on 07/02/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Table 5 Recuperator design data Table 6 Cost estimates for the ABC components

Exhaust gas Air Component Cost Weight

Flow rate [kg/s] 69.24 71.93 Mill. US$ Metric tons
Inlet pressure [bar] 1.033 8.012 Air inlet system 0.294 10.6
Inlet temperature [°C] 527 114.8 ABC engine (without intercoolers) 4.4121 30.0
Film coeff. [W/(m2 K)] 49.8 121 Intercoolers 0.206 3.0
Fin efficiency [-] 0.567 0.798 Recuperator 2.941 45.0
Mass flux [kg/(m2 s)] 4.31 15.90 Generator including enclosure 0.809 20.3
Reynolds number [-] 662.5 1002 Ducting, dampers and stack 0.403 37.4
Fin frequency [m1] 355.3 781.9 Cooling water pumps 0.011 0.7
Plate spacing [mm] 20.9 6.35 Sea water heat exchangers 0.077 5.2
Fin thickness [mm] 0.203 0.152 Other 0.206 2.0
Heat transfer area/volume [m2/m3] 800.5 1841 Total 9.359 154.2
Total film coeff. [W/(m2 K)] 21.3
The production cost may be approximately 1/3 of this number
Resistance ratio exhaust/air-side 2.28
NTU [-] 9.12
Heat transfer effectiveness [%] 90.4
Plate thickness [mm] 0.2 other cases must give the same power output as for the com-
Core volume [m3] 53.6 bined LM2500PE and ABC (29 MW). On this basis two cases
Core weight [kg] 42400 (1-2) were defined for an installation on a new oil platform,
and one case was defined for a retrofit case on an existing
platform (3). The cost of installing equipment of this type on
a new platform was estimated to 47 US$/kg (Rasmussen,
recuperator effectiveness goes up. If the recuperator weight is 1994). The cases are:
to be reduced from the present value of 42 metric tons down Reference case: 1 X LM2500PE + ABC. The weight and
to, say, 20 metric tons, the recuperator effectiveness must be cost of the LM2500PE are estimated to 125 metric tons and
reduced to approximately 82 percent. The combined gas turbine 11.3 MUS$, respectively. The weight and cost of the ABC are
and ABC efficiency is then reduced from 46.6 to approximately estimated to 154 metric tons and 9.4 MUS$. The power output
44.7 percent. and efficiency are 29 MW and 46.2 percent, respectively.
When comparing the weight of the present recuperator design
to the weight of a regenerator, there is a difference by a factor 1 The reference case compared to using 2 X LM2500PE.
of approximately 4 - 5 . These two gas turbines are assumed to operate on part load
such that the power output equals that of the reference case.
Cost and Economics of the A B C The gas turbines will be operating at an efficiency of 32.7
percent. The weight of these two gas turbines including all
A weight and cost estimation of the components in the ABC
equipment is estimated to 250 metric tons. The total cost is
was carried out. The results are summarized in Table 6. The
estimated to 22.6 MUS$.
net present value of the added efficiency/power by the ABC at
constant fuel consumption was calculated to 9 million US$. 2 The reference case compared to using an LM2500+. It
This is approximately the same as the predicted cost of the is assumed that this gas turbine will operate at full load with
ABC (9.4 million US$). an efficiency of 37.0 percent. The cost and weight of an
LM2500+ are estimated to 13.5 MUS$ and 150 metric tons,
An economic analysis was carried out. Different cases were
respectively. The LM2500+ is an upgraded and partly rede-
defined and comparisons were made between these. Some gen-
signed version of the LM2500PE. The main difference between
eral parameters for the economic calculations are given in Ap-
these two is that LM2500+ has a new stage added in front of
pendix A.4. The reference case for comparison is the ABC
the compressor, which increases the air flow approximately 20
together with an LM2500PE gas turbine. It is assumed that the
percent. The pressure ratio is raised from 18 to 23, and the
efficiency is increased 1-2 percent compared to the LM2500PE
(all these numbers are preliminary).
/f,.,.'X, r .. 110 3 Assuming that on an existing platform is an LM2500PE.
,. , -.-:—^
48.5 - 100 The two alternatives to be considered consist of: (a) ABC is
^ 48.0- — -0— - Recuperator weight - alloy steel 1 installed beside an existing LM2500PE, assuming that there is
47.5-
ta — • Recuperator weight - ordinary steel y ,
90 enough space available and that the platform structure can take
47.0-
46.5^
46.0-i
n
/ r
/ /
80 f

70 r
S
the extra weight of the ABC. (b) The existing LM2500PE is
exchanged with a new LM2500+. According to GE it will be
possible to exchange only the engine itself, and other compo-
45.5- 60 'S nents like the air intake system do not have to be changed.
</ / 50 $ These two alternatives would give the same power output. The
45.0-i cost and weight of the ABC is assumed equal to that of the ABC
445 J { S
,s in the reference case. The cost of the LM2500+ is estimated
44.0^ ..-- ^ 40 g. to 6.2 MUSS. The platform-related cost for exchanging the
1 43.5 i
-&" k3 3- 30 I LM2500PE with the LM2500+ is estimated to be equal to that
of a new installation, meaning that the weight (150 metric tons)
! 43.0 s 33-'
u 125*
W- 20
10
is multiplied with the specific cost of weight used in the present
study.
42.0-i
0 One should note that the assumption of equal power output
80 82 84 86 88 90 92 94 96 means that for case 1 there will be invested in surplus power
Recuperator effectiveness [%] that is not given any value in the economical analysis.
Fig. 6 Combined gas turbine/ABC efficiency and recuperator weight The results of the economic comparison of the cases are
as function of recuperator effectiveness presented in Table 7. It should be noted that the numbers for

366 / Vol. 118, APRIL 1996 Transactions of the ASME

Downloaded From: https://gasturbinespower.asmedigitalcollection.asme.org on 07/02/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Table 7 Results from the economic comparison between the Grieb, H., Schill, D. G., and Gumucio, R., 1975, " A Semi-empirical Method
LM2500PE/ABC and the other cases for the Determination of Multistage Axial Compressor Efficiency," ASME Paper
No. 75-GT-ll.
Case 1 2 3 GTPRO v.6.5, 1994, commercially available gas turbine database and power
cycle design program, Thermoflow Inc., Wellesley MA, USA.
2xLM250OPE lxLM2500+ lxLM2500 Hande, B. M., 1992, "Simulation of Steam Injected Gas Turbines," Doctoral
new new existing Thesis, University of Trondheim, ISBN 82-7119-401-1,
platform platform platform Hawkins, W. J., Mathieson, D., Bruce, C. J., and Socoloski, P., 1994, "System
NPV capital costs [MUS$] 0.6 -13.3 -3.4 Development Test Program for the WR-21 Intercooled Recuperated (ICR) Gas
Turbine Engine System," ASME Paper No. 94-GT-186.
NPV CCy-tax [MUSS] 21.8 13.1 13.1 Hornnes, A., and Bolland, O., 1991, "Power Cycle Working Fluids," SINTEF-
NPV fuel gas cost [MUSS] 15.0 9.0 9.0 report STF15 A91041, Trondheim, Norway.
NPV Total [MUSS] 37.4 8.9 18.8 Kays, W. M., and London, A. L., 1984, Compact Heat Exchangers, 3rd ed.,
Annuity [MUSS/year] 4.1 1.0 2.1 McGraw-Hill.
Klapproth, J. F., 1959, Discussion of paper by Lieblein listed elsewhere in this
Simple payback [years] - 5.5 1.4 reference list, ASME Journal of Basic Engineering, Vol. 81, p. 398.
Koch, C. C , and Smith, L. H., 1976, "Loss Sources and Magnitudes in Axial-
Flow Compressors," ASME JOURNAL OF ENGINEERING FOR POWER, Vol. 98,
No. 3.
net present value (NPV) and annuity in Table 7 are calculated Koch, C. C , 1981, "Stalling Pressure Rise Capability of Axial Flow Compres-
by subtracting the key number values of the present case from sor Stages," ASME JOURNAL OF ENGINEERING FOR POWER, Vol. 103, No. 4.
those of the reference case. The net present value of the capital Lieblein, S„ 1959, "Loss and Stall Analysis of Compressor Cascades," ASME
cost is negative when the investment cost of the reference cost Journal of Basic Engineering, Vol. 81, No. 3.
Olikara, C , and Borman, G. L., 1975, " A Computer Program for Calculating
is higher than that of the present case (as for case 2 and 3). Properties of Equilibrium Combustion Products With Some Applications to I.C.
The results in Table 7 show that all cases gave a positive net Engines," SAE Paper No. 750468, Feb.
present value. The simple payback time was calculated to 5.5 0verli, J. M., 1992, Str0mningsmaskiner (Turbomachinery), Vol. 3, 2nd ed.,
Tapir, Trondheim.
and 1.4 years for cases 2 and 3, respectively. For case 1 the
Pettersen, J., 1987, "Bunnprosesser for gassturbin-kraftverk (Bottoming Cycles
investment cost for the reference case was found less, and thus for Gas Turbine Power Plants)," Diploma Thesis, Division of Refrigeration,
no payback time was calculated. Norwegian Institute of Technology.
Rasmussen, J. B., 1994, personal communication, Norsk Hydro T&U.
Ricket, R., 1994, personal communication, Westinghouse Electric Corporation,
Conclusions ICR Design Engineering.
Sayers, A. T„ 1990, Hydraulic and Compressible Flow Turbomachines,
The conclusion is that the Air Bottoming Cycle is an econom- McGraw-Hill, New York, p. 236.
ical alternative for power generation on both new platforms and Shepard, S. B., Bowen, T. L., and Chiprich, J. M., 1995, "Design and Develop-
on existing platforms with demand for more power. ment of the WR-21 Intercooled Recuperated (ICR) Marine Gas Turbine," ASME
The recommended ABC is a two-shaft engine with two com- JOURNAL OF ENGINEERING FOR GAS TURBINES AND POWER, Vol. 117, pp. 5 5 7 -
562.
pressor intercoolers. The pressure ratio was found optimal at Solheim, K. S., 1994, personal communication, Norsk Hydro Aluminium Struc-
8:1. A recuperator is recommended instead of a regenerator tures.
for the heat transfer from the gas turbine exhaust gas to the Sovran, G., and Klomp, E. D., 1967, "Experimentally Determined Optimum
compressed air of the ABC. Geometries for Rectilinear Diffusers With Rectangular, Conical or Annular Cross-
Section," Fluid Mech. of Internal Flow, Elsevier Publishing, Amsterdam.
The combined gas turbine and ABC shaft efficiency was Sacther, S., 1993, "Combined Cycles With Compact Heat Recovery Steam
calculated to 46.6 percent. The LM2500PE gas turbine con- Generators," SINTEF Report STF15 A93111, Trondheim, Norway.
tributes with 36.1 percent while the ABC adds 10.5 percent Walker, G., Kremer, J., Fauvel, R., Reader, G., and Bingham, E. R., 1992,
points to the gas turbine efficiency. The ABC shaft power output "Stirling Bottoming Cycle for the Gas Turbine Exhaust Streams of Pipeline
is 6.6 MW when utilizing the waste heat of an LM2500PE gas Compressor Stations," Proceedings of the 27th Intersociety Energy Conversion
Engineering Conference, San Diego, CA, Aug.
turbine. For a final design of the ABC, the compressor and Wicks, F., Berven, G., and Marchionne, D„ 1992, "A Combined Cycle With
turbine efficiencies are very important to get as high as possible Gas Turbine Topping and Thermodynamically Ideal Gas Turbine Bottoming,"
with respect to cycle power output and efficiency. Proceedings of the 27th Intersociety Energy Conversion Engineering Conference,
San Diego, CA, Aug.
The preliminary compressor and turbine design has shown Wilson, R. A., and Kupratis, D. B., 1994, "Future Vehicular Recuperator
that an axial intercooled compressor and an axial turbine, which Technology Projections," ASME Paper 94-GT-395; accepted for publication in
are adapted to the air bottoming cycle for an GE LM2500PE, the ASME Journal of Turbomachinery,
can be built using conventional design practice. Weston, K. C , 1993, "Dual Gas Turbine Combined Cycles," Proceedings of
the 28th Intersociety Energy Conversion Engineering Conference, Atlanta, GA,
At present it is suggested to use a recuperator and not a
regenerator for the ABC.
A weight and cost breakdown of the ABC is presented. The APPENDIX
total weight of the ABC package was calculated to 154 metric A . l Site Conditions
tons, and the ABC package cost to 9.4 mill. US$. In this study a specific location on an oil platform in the
An economic examination for three different cases was car- North Sea was chosen for a possible implementation of the
ried out. The results show that the ABC alternative (LM2500PE ABC. Site conditions typical on an oil platform were chosen
+ ABC) is economical, with a rather good margin, compared as: ambient temperature = 8°C, ambient relative humidity =
to the other alternatives. 60 percent, ambient air pressure = 1.013 bar, cooling water
temperature = 8°C, generator speed = 60 Hz.
Acknowledgments
Norsk Hydro, Kvaerner Energy, and the Research Council of A.2 Topping Gas Turbine
Norway supported this work. The LM2500PE gas turbine is the most commonly used gas
turbine in the Norwegian sector of the North Sea. This gas
turbine was chosen as the topping cycle for the ABC. Perfor-
References
mance and exhaust gas data are given in Table 8. A typical fuel
Beck, S., 1994, "Create Baleen Regenerator," Memorandum, Creare Inc.
Bolland, 0 „ 1991, "A Comparative Evaluation of Advanced Combined Cycle
gas composition was used. The source of the LM2500PE data
Alternatives," ASME JOURNAL OF ENGINEERING FOR GAS TURBINES AND POWER, is GTPRO (1994).
Vol. 113, pp. 190-197.
Farrell, W. M., 1988, "Air Cycle Thermodynamic Conversion System," United A.3 Computational Parameters
States Patent No. 4,751,814.
Farrell, W. M., 1992, "Air Cycle Thermodynamic Conversion System," Euro- Ambient temperature = 8°C, ambient air pressure = 1.013
pean Patent Specification 0 208 162 B l . bar, compressor pressure ratio: LPC = 1.91, IPC = 1.91, HPC

Journal of Engineering for Gas Turbines and Power APRIL 1996, Vol. 1 1 8 / 3 6 7

Downloaded From: https://gasturbinespower.asmedigitalcollection.asme.org on 07/02/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Table 8 LM2500PE performance and exhaust gas data (GTPRO, 1994) 324.8 302.9 285.4 1.591 1633
246.5 325.9 309.1 1.916 1577
Fuel consumption 1.3132 kg/s
Exhaust gas flow rate 69.24 kg/s
Exhaust gas temperature 527 °C LPC polytropic efficiency (total-to-total): 1.3, work: 4742 kW, exit
Compressor inlet pressure loss 10 mbar pressure: 1.916 bar.
Turbine exhaust pressure loss 20 mbar
Generator terminal power output 22935 kW Intermediate-Pressure Compressor—IPC
Gas turbine auxiliary power 111 kW
Miscellaneous plant auxiliaries 181 kW Inlet Outlet Blade
Net power output 22643 kW rotor rotor height Spacing Chord -•Mnner
Net efficiency 36.1 % angle angle [m] [m] [m] [m] [m]
Exhaust gas: Nitrogen - N2 75.15 vol-%
Argon - Ar 0.90 vol-% 47 27 0.132 0.045 0.04 0.352 0.220
48 26 0.116 0.042 0.038 0.340 0.224
Oxygen - 0 2 13.74 vol-% 50 27 0.100 0.040 0.035 0.330 0.230
Carbondioxide - C0 2 3.36 vol-%
Water - H2Q 6.85 vol-% M)2 Ts2 Ts, Po2 Work
Stage [K] [K] [K] [bar] [kW]
298.1 282.6 282.6 1.883
= 2.28, compressor inlet temperature: LPC = 8°C, IPC = 25°C, 321.0 298.6 280.2 2.403 1658
HPC = 25°C, turbine inlet temperature = 490°C, turbine poly- 343.2 321.4 304.5 2.971 1612
tropic efficiency = 90.0 percent, turbine exit pressure drop = 366.2 345.3 327.6 3.596 1669
4 mbar, compressor inlet pressure drop = 1 0 mbar, recuperator
air side pressure drop = 1 6 0 mbar, recuperator heat loss = IPC polytropic efficiency (total-to-total): 89.7, work: 4938 kW, exit
0.2 percent, recuperator air loss = 0.0 percent, pressure drop pressure: 3.596 bar.
intercoolers: LP = 2.0 percent, HP = 2.0 percent, mechanical
efficiency = 99.3 percent, generator efficiency = 97.0 percent, High-Pressure Compressor—HPC
auxiliary power ([percent] of net power) = 1.0 percent; compu-
tational interest rate = 7 percent, economical life time = 1 5
years, operational hours per year = 8000 h, fuel gas C0 2 tax Inlet Outlet Blade
= 11.76 cents/Sm 3 , fuel gas sales value = 8.09 cents/Sm 3 . rotor rotor height Spacing Chord Romer
angle angle [m] [m] [m] [m] [m]
A.4 Compressor Design Details 48 20 0.0820 0.036 0.033 0.292 0.210
48 19 0.0700 0.034 0.032 0.284 0.214
Low-Pressure Compressure—LPC 48 19 0.0610 0.031 0.030 0.278 0.217
48 20 0.0540 0.028 0.028 0.273 0.219
Inlet Outlet Blade
rotor rotor height Spacing Chord ** inner T02 1S2 Ts, P02 Work
angle angle [m] [m] [m] [m] [m] Stage [K] [K] [K] [bar] [kW]
50 37 .220 0.047 0.065 0.440 0.220 298.1 284.0 284.0 3.536
53 37.7 0.190 0.046 0.06 0.420 0.230 322.5 302.7 283.4 4.574 1768
55 39 0.165 0.044 0.05 0.405 0.240 345.5 325.7 308.3 5.678 1672
367.5 347.6 331.5 6.883 1594
*02 Ts2 Work
388.1 368.4 353.5 8.062 1498
Stage [K] [K] [K] [bar] [kW]
281.1 267.4 267.4 1.003 HPC polytropic efficiency (total-to-total): 89.0, work: 6530 kW,
302.3 278.1 261.4 1.271 1532 exit pressure: 8.062 bar.

368 / Vol. 118, APRIL 1996 Transactions of the ASME

Downloaded From: https://gasturbinespower.asmedigitalcollection.asme.org on 07/02/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use