A model to predict the behaviour at part load operation of once-through heat

recovery steam generators working with water at supercritical pressure

Antonio Rovira3'*, Manuel Valdesb, Ma Dolores Duran c
"E.T.S. Ingenieros Industriales de la UNED, Departamento de Ingenieria Energetica, c/Juan del Rosal n° 12. 28040 Madrid, Spain
b
E.T.S. Ingenieros Industriales de la U.P.M., Departamento de Ingenieria Energetica y Fluidomecanica, c/Jose Gutierrez Abascal n" 2. 28006 Madrid, Sp
c
Facultad de Ingenieria de la UAEM, Division de Ingenieria Mecanica, Cerro de Coatepec, C.U. 50110 Toluca, Mexico

ARTICLE INFO ABSTRACT

This paper describes a one-dimensional mathematical model that allows simulating the heat exchange in
a steam generator working with water at supercritical pressure. The model has been developed in order
to simulate the full and part load behaviour of heat recovery steam generators (HRSGs) of combined cycle
gas turbine (CCGT) power plants. It takes into account the strong variation of some of the thermal and
Keywords: transport properties of fluids at supercritical pressure and discusses what parameters may be considered
Supercritical water as constant along the heat exchanger.
Heat recovery steam generator (HRSG) On the one hand, the model is useful because going supercritical is considered a way to further
Once-through boiler
Combined cycle gas turbine (CCGT), improve the efficiency of CCGT power plants and, on the other hand, because part load operation is the
exergetic losses most usual operation mode in power plants.

1. Introduction Fig. lb shows a triple pressure HRSG with a supercritical high

pressure level. Such behaviour could remain at part load operation
Combined cycle gas turbines (CCGTs) power plants are one of if the exhaust temperature of the gas turbine is controlled (for
the most efficient energy conversion systems. Owing to the good example, using compressors with variable inlet guide vanes) as the
thermodynamic, economical and environmental performances that drop of temperature and pressure of the steam is lower than using
this kind of power plants may reach, nowadays they are undergoing other regulation systems [3,4]. Other advantages are the simplicity
widespread installation and the research in this field has notably of the HRSG — once-through HRSGs can be employed — which
increased. should lead to cheaper designs [5,6] and faster start-up times [7[.
A possible way to further improve the efficiency of these Owing to these reasons and to the advances in experimental
systems is to minimise the exergetic losses in the heat recovery research — for example, Dechamps and Galopin [8] and Dumont
steam generator (HRSG) by means of the use of water at super- and Heyen [9] — , working with water at supercritical pressure is
critical pressure. The difference of temperatures in the heat near to be technically and economically feasible in CCGT power
exchange from a hot fluid to a cold one leads to exergetic losses. The plants (Najjar [5], Dechamps [10] and Galopin [11]) as well as in coal
exergetic losses imply a diminution of the steam turbine power fired power plants (Beer [12]).
and, consequently, a decrease in the CCGT efficiency [1,2]. The main The simulation of the heat exchange in the HRSG when the
advantage of working at supercritical pressure is that there is not power plant operates at part load conditions is interesting in order
a saturation temperature, unlike at subcritical pressure. Instead, the to predict the power plant performances when the demanded
plain zone in the enthalpy-temperature diagram does not exist any power is lower than the reached at full load operation or when the
more and any heat input towards the working fluid will increase its ambient conditions change. Part load calculations are usually much
temperature, so the mentioned exergetic losses owing to the heat more time-consuming than the design condition (full load opera-
exchange decrease. This effect is observed in Fig. 1: Fig. la shows tion) ones. For example, the calculation time of a subcritical CCGT
the thermal energy—temperature diagram of a subcritical triple from full load until a load of 50% may be over one thousand times
pressure HRSG (with three drums at different pressures) while greater than that spent in calculating the design condition. Due to
the nature of the water at supercritical pressure, which is intro-
duced in Section 2, the time of calculation of a HRSG working with
* Corresponding author. Tel.: +34 913988224; fax: +34 913987615. water at supercritical pressure would become even larger. Since
E-mail addresses: [email protected] (A. Rovira), [email protected]
(M. Valdes), [email protected] (MaD. Duran). this point of view, a complete three-dimensional model that
 2.1. Heat exchange equations in HRSG
Nomenclature
Two equations are needed to predict the behaviour of heat
A heat exchange surface, m 2 exchangers working with fluids at subcritical pressure: the energy
Cp specific heat, J/(kg K) balance and a heat exchange equation. When the HRSG of a CCGT
D diameter, m power plant is analysed at the design condition, the energy balance
F log mean temperature difference correction factor (together with the knowledge of a given number of design
h convective heat-transfer coefficient, W/(m 2 K) parameters) provides the value of the design temperatures in every
k thermal conductivity, W/(m K) point of the HRSG (Valdes et al. [13] and [15]). If a counter-flow
m mass flow, kg/s exchanger is used, the heat exchange equation allows calculating
V pressure, Pa the UA product (product of the overall heat-transfer coefficient U
0. heat-transfer rate, W and the exchange surface A) by means of the logarithmic mean
T temperature, K temperature difference of the exchanger3:
U overall heat-transfer coefficient, W/(m 2 K)
V velocity, m/s
Q = UA (Tg ftg out -
(1)
' w out
Greek symbols In
'•g OUt "
H dynamic viscosity, Pa s
p density, kg/m 3 The UA product is the base to predict later the performance of
the heat exchanger at part load operation.
Subscripts The expressions of the mathematical model that describe the
des design heat exchanger working with a fluid at supercritical pressure are
gas similar to the subcritical case. Nevertheless, they should take into
g
i inside account that most of the properties, for example the specific heat,
0 outside cannot be assumed as constant — with the exception of the overall
w water heat-transfer coefficient (U) in a first approach, as it will be
explained in Section 2.2. In such a case, equation (1) is not valid any
more. To avoid this problem, when a flow works at supercritical
simulates the heat exchangers without any simplification is not pressure the following methodology could be used:
advisable if the purpose is, for example, systematically analysing
wide part load ranges or performing annual simulations of CCGT. 2.2.2. Full load operation
The objective of this paper is to propose a simplified mathe- The equations that govern the heat exchange (Fig. 2), again in
matical model, similar to the already existent one-dimensional the case of a counter-flow heat exchanger3 applied to the design
subcritical ones (Valdes et al., [13]), that allows simulating the heat point are:
exchange in once-through HRSGs working with water at supercrit-
ical pressure in a broad load range of the CCGT with enough accuracy f d Q = -riigdes'Cpg-dTg
and calculation times comparable to the subcritical model ones. S dQ = - r h w d e s c p w d 7 w (2)
Finally, the model — that has been used in Rovira [4] — is vali- I dQ = Udes-dA- {Tg - 7W) = (LW)des- (Tg - r w •dx
dated with experimental data obtained from ref. [8].
where x varies between 0 and 1 and represents a dimensionless
2. Heat transfer to water at supercritical pressure length of the exchanger.
Combining the first and the second equations, the following
The behaviour of fluids at supercritical pressure is mainly energy balance is obtained:
characterised by a strong variation of their thermal and transport
properties 1 with the temperature. This behaviour makes impos-
sible to assume that the properties are constant in the mathemat- rhg des • cP g • dig = m'w d e s ' CP w ' d J w (3)
ical models, unlike what is usual in the subcritical mathematical 'g in des v out des
models. For that reason, numerical methods are required instead of
the usual symbolic solutions.
Furthermore, the strong variation of the fluid properties near m
w des' C P \
the pseudocritical condition2 causes the phenomena called 'heat- * g ~~ g in des •dTvi (4)
mg des • Cp g
transfer enhancement' and 'heat-transfer deterioration', whose J
w out des
influence on the HRSG heat transfer has to be considered.
On the other hand, combining the last two equations of (2):
The expressions that describe the heat exchange of the different
HRSG heat exchangers (economisers, evaporators and super- m
j„ _ w d e s ' CP w J j
heaters) are shown in the following sections. The well known (5)
UAdes • (Tg - r w )
subcritical models will be adapted to simulate the heat exchange
towards water at supercritical pressure in the cases of full load Equation (5) may be integrated assuming U as constant along
operation and part load operation of the power plant. the exchanger (see Section 2.2):

1
The IAPWS Industrial Formulation 1997 for the Thermodynamic Properties of
Water and Steam [14] has been used.
2
The pseudocritical condition is reached when the specific heat has a maximum
value. That condition may be considered as the frontier between the liquid and In this section, the mathematical model is developed for a counter-flow heat
vapour state. exchanger. It may be also developed for any other one using the correction factor.
 Thermal energy transfer Thermal energy transfer

Fig. 1. Thermal energy—temperature diagram of a subcritical HRSG (a) and a supercritical HRSG (b).

d ( 7 g - r w ) = r-rhwCpw-dTv; (10)
m C
w des P\
•dlvi (6)
UA,des (TS where expressions (9) and (10) derive from equation (2). Inte-
grating equation (10):
ft-7"-) T„
C
LMdes = mw Pw / d(Tg-Tw) = rhw- / r-cPw-drw (11)
des' •dlw (7)
(Tg - r w
(Tg i„-T w o m ) ?w out

For heat exchangers different from counter-flow ones, the

equation (7) should be corrected with a factor (F) which is less [Tg — Tw)j — [Tg m — T w out) + Tiw' r-Cp w - d r w (12)
than 1:

C The integral may be numerically solved decreasing Tw at each

P w
r
des' LMdes = til w des' •dr u (8) step of the process. Thus, the value of (r g -T w ) is known for each
intermediate value of Tw.
At the same time, considering equation (5) and using its
Equation (8) may be numerically integrated at the design point respective UA value and the correction factor F:
of the power plant taking into account equation (4) (to calculate the
gas temperature) and the value of cpw, which varies at each step of Cp,
mw •dTvi (13)
the numerical integration. Therefore the Fdes-UAdes product is
FUA' J (Tg-Tw)
calculated. Tv
Equations (4) and (8) are the energy balance and the heat
exchange expressions respectively, and the F- UA product may be With equations (12) and (13) it is possible to calculate the value of x
obtained. for each Tw. The numerical integrations conclude when x = 1 and
then, Tg out and Tw ;n are the values corresponding to this final step
2.2.2. Part load operation of the process.
Once the F- UA product is obtained at the design point, it may be Equations (12) and (13) are again the energy balance and heat-
calculated at any other part load of the CCGT as it will be shown in transfer equations.
Section 2.2. The calculation of any two variables, for example Tg out
and Tw in, may be done assuming that the others, Ts ;n, Tw 0ut. niw 2.2. Calculation of the overall heat-transfer coefficient of
and riig, are known and using the corresponding UA product in the supercritical fluids at part load operation
part load operation point:
In order to evaluate the expressions (12) and (13), the knowl-
d(7 g - 7W) = dig - d7w edge of the F- UA product at any part load operation is needed. To
this end it is enough to find the (F(i)/(Fdesl-'des) ratio since the
1 1
•dQ = - r d Q (9) previously calculated heat exchange surface A is constant.
Cp g • /Tig Cp w • /Tlvv
As it is known, the overall heat transfer coefficient is defined as
the resistance that exists in the heat exchange between two
different fluids. In the particular case of a heat exchanger in which
the fluids flow inside and outside a tube, U may be calculated [16]:
1
r gi „ U = (14)
_ r0-ln(r0/ri) r0
he k n • hv

where r0 and n are, respectively, the outside and inside radius of the
tube, k is its thermal conductivity and hg and h w are the gas and the
water-side convective heat-transfer coefficients. Typical values for
Fig. 2. Schematic of a counter-flow heat exchanger. the terms in subcritical HRSGs of CCGT are shown in Table 1. In this
Table 1 The subscript des makes reference to the nominal or design
Typical values of the terms of equation (13) in HRSGsa. condition. The relations hg/hg des and h w /h w des may be calculated as
h g [W/(m 2 K)l k/r 0 -ln(r 0 /n)[W/(m 2 K)l n-/i w /r„ [W/(m 2 K)] it is indicated in Section 2.3.
Economizers 50—100 2000-6000 2000-20000 For economisers and evaporators expression (18) may be
Evaporators ~ 50 5000-10000 2000-20000 simplified as follows:
Superheaters 50-100 2000-10000 1000-5000
a
Usual geometrical data were obtained from ref. [17]. Heat exchanger designs h«
t/ = h e = h,g d e s ' u (19)
and convective heat-transfer coefficients were calculated as explained in ref. [18].
Thermal conductivity of the materials was extracted from ref. [16]. "g des
and the FUA product may be calculated:
table it may be observed that, especially in economisers and
evaporators, the term corresponding to the gas side (hg) is much FU
lower than the thermal conductivity and the water-side terms (k FUA = Fdes-UA,
"•des = ^ e s ^ Ades (20)
'"des' ^ d e s t[ies fig [i e s
and h w ). Therefore, in economisers and evaporators expression (14)
may be simplified: The coefficients hg des and hw des depend on the geometry of the
heat exchanger. Thus, equations (18) and (19) have the disadvan-
U tage of requiring the knowledge of the geometric design of the
~V** = hg (15)
exchanger. Equation (20) suggests that the heat exchange is mainly
governed by the gas flow. For that reason U may be assumed as
In flows at subcritical pressure, the convective heat-transfer
constant along the heat exchanger, which should be taken into
coefficients (h) are usually calculated by means of empirical
account when integrating equation (5).
dimensionless relations. For flows inside a tube, a commonly used
expression is the Dittus—Boelter equation [19]: The influence of the F/Fdes ratio is assessed in Section 3 by means
of a sensitivity analysis.
Nu = 0.023-Re 08 Pr 0 4 (16)
2.3. Calculation of the convective heat-transfer coefficient of flows
where Nu is the Nusselt, Re the Reynolds (both based on the at part load operation
hydraulic diameter of the tube) and Pr the Prandtl number. All of
them, at subcritical pressures, are evaluated at the mean bulk
The convective heat-transfer coefficient (h) is a variable that
temperature.
depends on the temperature, the pressure, the velocity of the
At supercritical pressure, the heat-transfer coefficient may stream and the geometry of the exchanger. As it was said before, for
deviate from the Dittus—Boelter equation near the pseudocritical a stream flowing through the outside of a finned tube the equation
condition, as reported by Petuhkov [20]. This phenomenon is called (17) may be used. Replacing the non-dimensional numbers by their
heat-transfer enhancement or deterioration (as it increases or definitions:
decreases the coefficient respectively) and it should be taken into
account to correct the Dittus—Boelter equation [21—24]. Although
these phenomena might be important in some applications, in this hgDg = Q fPgVgD, MgCpg, . £ l - m (21)
study they has been neglected, mainly because the heat-transfer kg y /xg
coefficient of the water is much higher than the gas one (as it is
where Dg is the characteristic length of the flow. Solving for hg
shown in the Appendix). In such a situation, the gas establishes the
strongest resistance to the heat exchange and governs the heat „m,,mr)m ,.nrn
exchange. Also, the nearby of the pseudocritical point is only (22)
reached in a very small region of the heat exchange domain. Da Mg< k"
For the flows outside of the tubes (gas side in the HRSG), the Reorganising the terms and taking into account that
expression proposed by Shmith [25] and used in HRSG by Weir [26] nig = pg-vg-Ag (where Ag is the cross sectional surface):
may be used:
n.nm-l i,nrn r,
u
Nu = a - R e m P r n < ^ (17) ha = " % ^gLPg K
g • m (j>l-m
Ag ' kg 'fig'ms' a

where $ a is a parameter that depends on the geometry of the n.nm-l r

n

exchanger (it takes into account the number and size of the fins and (23)
the bank tube geometry). For a heat recovery boiler, a, m and n are Ag jig- k"-x
n s

respectively 0.3, 0.625 and 1/3 (Weir [26]). Nu and Re are based on Rapun [27] and Duran [28] expand and reorganise expression
the hydraulic diameter (four times the cross sectional area divided (23) as below in order to express the convective heat-transfer
by the wetted perimeter) and Nu, Re and Pr are evaluated at the coefficient as a function of three terms. One of them is constant and
mean bulk temperature. only depends of the geometry (F), the second variable depends on
Once the convective heat-transfer coefficient has been calcu- the thermal state of the fluid ((3) and the last one depends on the
lated at the design condition, U could be obtained at every part load mass flow of the stream:
condition using the following equation:

_ 0.3 -D0--625-1 \ / cgf5 0.625

U = 2
1 •ln(r0/n r0 ~ Lo.625. $0,625-1 I ' I pj.0.625-1/3 fe0 6 25_! I ' mS
+ r° >+•
k
1 0.625
(24)
(18) = rg-/V"i!
i r •ln(;roM)
+ ° where the a, m and n coefficient have been replaced by the values
h,g d e s ' / , r h Jha
' i " w des ft . proposed in ref. [26].
 Table 3
Experimental data and model estimation for the HRSG variables.

Variable Condition 1 Condition 2

""~©>— Real Simulation Real Simulation
m g (kg/s) 22.2 22.1 14.7 14.8
EvSCl
m w (kg/s) 3.83 3.85 2.6 2.6
Tgi (K) 923 923 923 923
Tg2 (K) 822 823.7 808 810.0
Tg3 (K) 493 488.2 479 474.6
Tw3 (K) 793 793 793 793
Tw2 (K) 665 667.1 667 662.8
EvSC2 378 378 378 378
Twi (K)
p (bar) 240 240 240 240
©-*• F-IMEVSCI (kW/K) 17.4 16.9 12.1 12.7
F-lMEvSc2(kW/K) 102.9 110.2 91.2 84.5

Fig. 3. Once-through HRSG.

(F-LM)des product of their heat exchangers were calculated as
Finally, the expression for hg/hg des is obtained: described in Section 2.1.1. Once the (FiA4)des products were
obtained, the value of the air mass flow was gradually reduced in
0.625 0.375 0.625 the simulation program until condition 2 was reached. As said
h _ rg-Ps-™ Cp g
before, Tgl, Twl, TW3 and the pressure remained constant and Tg2, Tg3,
•0.625 kg
r g -/5 e TW2 and the steam mass flow were the results of the simulation,
0.625-1/3 0.625 which followed the steps described in Section 2.1.2.
Pr ff mg Table 3 compares the results obtained in the mentioned work
(25)
Pr ff "igdes together with those obtained in the simulation. It can be observed
that the results obtained using the proposed model for mw, Tg2, Tgj
The above expression depends on the thermodynamic state and and TW2 fit well with the experimental values, and that the trends
the gas mass flow but not on the geometric design of the exchanger. obtained predict accurately the actual HRSG performance. The
Therefore, the disadvantage that arose in equation (18) disappears value of the F- UA product of the heat exchangers are also compared,
in equation (20). both for the real and the simulated HRSG. They were obtained using
equation (8). Differences are below 10% for both exchangers at the
two conditions. It should be noted that the correction factor F may
3. Validation of the model affect the results at condition 2 because their values may vary from
the full load condition to the part load one.
In order to validate the proposed model, data from an experi- In Table 3, the results have been obtained for a value of F/Fdes of 1
mental HRSG [8] are used. The schematic of the HRSG is shown in in equation (20). When the load of a CCGT decreases, the generation
Fig. 3 and its main geometric data are shown in Table 2. It has two of steam in the HRSG also does. Therefore, at part load operation
heat exchangers: the first one (EvSCl) works as an economizer and the HRSG is oversized and its efficiency slightly increases. For that
an evaporator and the second one (EvSC2) works as a superheater. reason, the heat transfer along the exchangers should vary and the
The gas flows outside the tubes of the heat exchangers. ratio F/Fdes may differ from 1 [16]. Actually, the factor Ftakes a value
In the cited work, the performance of the HRSG is provided in between that corresponding to a cross-flow exchanger and 1, cor-
two different operating conditions. These conditions, shown in responding to a counter-flow exchanger. The greater the amount of
Table 3, differ in the mass flow of the gas, but the inlet gas bank of tubes is, the higher the value of F is, and its value is close to
temperature (Tgi), the feed water temperature (Twi), the steam 1 for well-designed exchangers [29]. Since HRSG consists on several
temperature (TW3) and the pressure are the same in both cases. banks of tubes, F/Fdes should not vary at a large extent.
They were the input data of the simulation, so the variables to be In order to quantify the uncertainty that this ratio introduce,
validated were the gas outlet temperature of each exchanger (Tg2 sensitivity analysis was done. Table 4 shows how the steam
and Tg3), the intermediate water temperature (TW2) and the steam production and the temperatures vary when the ratio F/Fdes is
mass flow. altered. The other parameters do not vary. Calculations were done
It has been assumed that condition 1 of Table 3 corresponds to only for condition 2 because condition 1 corresponds to the design
the full load condition of the experimental HRSG. Hence, an HRSG point, and the results are shown as percentage changes. When
similar to the experimental one was simulated and, afterwards, the F/Fdes is modified a 20% toward higher or lower values, variations of
the results are below 4.5%. If F/Fdes is modified a 10%, the variations
Table 2
are even below 2%. Lower variations of F/Fdes, expected in large
Main heat exchange surface design data (extracted from ref. [8]). HRSGs, lead to small uncertainties.
Heat exchanger Econo -evap Superheater
Tube diameter (mm) 25 26.5
Tube thickness (mm) 2.9 4.2 Table 4
Number of rows 18 6 Sensitivity analysis of F/Fdes.
Useful length (m) 6 6
Fins per meter (m _ 1 ) 200 200 FIFdes 0.8 0.9 1 1.1 1.2
Fin diameter (mm) 49 49 Amw (%) -4.38 -1.87 0.00 1.70 3.30
Tube layout Staggered Staggered ATg2 (%) 2.37 0.87 0.00 -1.48 -2.91
Transverse pitch (mm) 83 83 ATg3 (%) 4.14 1.83 0.00 -1.72 -3.32
Longitudinal pitch (mm) 73 73 ATw2 (%) 1.03 0.30 0.00 -0.47 -0.80
4. Conclusions

A methodology to simulate the heat exchange process between

a gas and a fluid at supercritical pressure has been proposed. It
takes into account that the UA product cannot be calculated using
the logarithmic mean temperature difference due to the strong
variation of the thermal and transport properties of the fluid.
Instead, the calculation is done using a numeric integration along
the heat exchanger length. Furthermore, the model allows simu-
lating heat exchangers at the design point and at off-design oper-
ation, what is useful in many engineering fields.
The model developed in this paper has been applied to once-
through HRSGs and successfully validated with experimental data
[8]. Furthermore, the sensitivity analysis done shows that the
uncertainties in the heat exchange calculation will not lead to high
inaccuracies, especially in large HRSGs.
Fig. 5. Usual convective heat-transfer coefficient of steam.
Finally, the simulation method does not need a thorough
knowledge of the geometric design of the heat exchanger. In that
way, the number of input data required for the off-design simula- pressure far from the saturation line or the pseudocritical point and
tion is remarkably small. that the value increases with the pressure near the pseudocritical
point. The heat-transfer coefficient at the water side is always more
Acknowledgements than two orders of magnitude higher than the gas side one.
On the other hand, Fig. 5 is similar to Fig. 4 but calculating the
The authors wish to acknowledge the financial support of UNED coefficient of the steam in a tube of 30 mm of diameter where the
to this work. velocity is 25 m/s (typical in superheaters). In this case it can be
observed that, at subcritical pressure (below of 221 bar), the value
of the convective coefficient is an order of magnitude lower than
Appendix. Order of magnitude of the convective heat- the obtained before for the liquid water case (Fig. 4). However, it is
transfer coefficient of water at supercritical pressure also observed that the value reached at supercritical pressures is
similar to the obtained for the liquid water.
In subcritical economisers and evaporators of HRSGs, the water- Likewise, Fig. 6 shows the evolution of the convective heat-
side convective heat-transfer coefficient may be considered as transfer coefficient of the experimental HRSG used by Dechamps
much higher than the gas-side ones. For that reason, in these and Galopin [8]. The similarity between both coefficients — at liquid
subcritical exchangers expression (19) may be used (see Table 1). and vapour state — is shown again. Similar results were obtained by
This appendix shows, firstly by means of the Dittus—Boelter Dumont and Heyen [9].
equation and later considering the heat-transfer deterioration, that Finally, as it was said in Section 2.2, at supercritical pressure the
the convective heat-transfer coefficient of supercritical liquid and Dittus—Boelter equation is not valid to predict accurately the value
vapour water (below and above of the pseudocritical point of the convective heat-transfer coefficient. In order to solve this
respectively) are similar to the subcritical liquid water when they problem several corrections has been given by many authors.
flow through economisers and superheaters respectively. Conse- Cheng and Schulenberg [30] made a literature review about the
quently they are much higher than the gas-side ones. matter. In their work, they show a certain lack of agreement
Fig. 4 shows the convective heat-transfer coefficient of the liquid between the different authors and point out that the corrections to
water versus the temperature taking the pressure (subcritical and the Dittus—Boelter equation differ in each particular case. Ana-
supercritical) as a parameter. It was calculated by means of the lysing the results shown in ref. [30], which do not take into account
Dittus—Boelter correlation applied to a tube of 30 mm of diameter the heat-transfer deterioration phenomenon, it may be concluded
and a velocity of 2.5 m/s (typical values in economisers of HRSGs). that the order of magnitude of the convective heat-transfer coef-
The results show that the coefficient is almost independent of the ficient is the same than the obtained by means of the Dit-
tus—Boelter equation. The effect of the heat-transfer deterioration
50
30Q

250

g 150-1

2 100

50

0 578 678 778

478
378 T (K)
500 BOO -EvSC 1 - Ev SC 2
T (K)
Fig. 6. Evolution of the water-side convective heat-transfer coefficient within the
Fig. 4. Usual convective heat-transfer coefficient of liquid water. analysed once-through HRSG.
could be analysed from ref. [31], where it could be observed that [14 Staff report, Release on the IAPWS industrial formulation 1997 for the ther-
modynamic properties of water and steam. The International Association for
despite of this phenomenon, the order of magnitude of the the Properties of Water and Steam, 1997.
convective heat-transfer coefficient still remains. In ref. [31] it is [15 M. Valdes, M.D. Duran, A. Rovira, Thermoeconomic optimization of combined
also discussed that the deteriorated heat transfer can be suppressed cycle gas turbine using genetic algorithms. Appl. Therm. Eng. 23 (2003)
by means of flow obstructions and other heat-transfer enhancing 2169-2182.
[16 A.J. Chapman, Transmision del Calor, third ed. Bellisco, Madrid, 1984.
devices. For these reasons equation (19) could be used since the [17 P.J. Dechamps, Incremental Cost Optimization of Heat Recovery Steam
gas-side heat-transfer coefficient for usual HRSG designs is much Generators. ASME COGEN-TURBO, Viena, Austria, 1995.
lower than the water one. [18 D. Duran, A. Rovira, M. Valdes, Methodology for the thermoeconomic and
design optimization of a heat recovery steam generator. In: Proc of the 19th
ECOS Int. conference, Aghia Pelagia, Greece, 2006. p. 577—86.
[19 F.W. Dittus, LM.K. Boelte vol. 2, Univ. Calif., Berkley, Publ. Eng., 1930, 443.
References [20 B.S. Petuhkov, Heat transfer and friction in turbulent pipe flow. Adv. Heat
Transfer 6 (1970) 511-564.
[1] J.H. Horlock, Combined Power Plants, first ed. Pergamon Press, Oxford, 1992. K. Yamagata, K. Nishikawa, S. Hasegawa, T. Fujii, S. Yoshida, Forced convective
[21
[2] M. Mufioz, A. Rovira, Ingenieria Termica. Universidad National de Education heat transfer to supercritical water flowing in tubes. Int. J. Heat Mass Transfer
a Distancia, Madrid, 2006. 15(1972)2575-2593.
[3] R. Kehlhofer, F. Hannemann, F. Stirnimann, B. Rukes, Combined-Cycle Gas & G. Scalabrin, L Piazza, Analysis of forced convection heat transfer to super-
[22
Steam Turbine Power Plants, third ed. Pennwell, Tulsa, Oklahoma, USA, 2009. critical carbon dioxide inside tubes using neural networks. Int. J. Heat Mass
[4] A. Rovira, Desarrollo de un modelo para la caracterization termoeconomica de Transfer 46 (7) (2003) 1139-1154.
ciclos combinados de turbinas de gas y de vapor en condiciones de carga D. Squarer, T. Schulenberg, D. Struwe, Y. Okab, D. Bittermann, N. Aksan,
[23
variable, PhD thesis. Universidad Politecnica de Madrid, Madrid, 2004. C. Maraczy, R. Kyrki-Rajamaki, A. Souyri, P. Dumazh, High performance light
[5] Y.S.H. Najjar, Efficient use of energy by utilizing gas turbine combined water reactor. Nucl. Eng. Des. 221 (2003) 167-180.
systems. Appl. Therm. Eng. 21 (2001) 407-438. X. Cheng, T. Schulenberg, D. Bitterman, P. Rau, Design analysis of core
[24
[6] Staff report, Agawam merchant GT24 combined cycle plant uses once-through assemblies for supercritical pressure conditions. Nucl. Eng. Des. 223 (3) (2003)
HRSG. Mod. Power Syst (September 1999) 31-38. 279-294.
[7] Staff report, Fast work for the Benson once-through HRSG. Mod. Power Syst. [25 D. Shmith, Mean temperature difference in cross-flows. Engineering 138
(February 2004) 2 3 - 2 5 . (1934) 479.
[8] P.J. Dechamps, J.F. Galopin, Once through heat recovery steam generators [26 CD. Weir, Estimating the performance of gas turbine heat-recovery boilers
working with sub-and supercritical steam conditions for combined cycle. off-design. Proc. Inst. Mech. Eng. 202 (1988) 269-277.
Power Gen. 3 (1997) 237-249. [27 J.L. Rapun, Modelo matematico del comportamiento de ciclos combinados de
[9] M.N. Dumont, G. Heyen, Mathematical modelling and design of an advanced turbinas de gas y vapor, PhD thesis. Universidad Politecnica de Madrid,
once-through heat recovery steam generator. Comput. Chem. Eng. 28 (5) Madrid, 1999.
(2004) 651-660. [28 D. Duran, Estudio de Calderas de Recuperation de Calor de Ciclo Combinados
[10] P.J. Dechamps, Advanced combined cycle alternatives with the latest gas de Turbinas de Gas y Vapor Empleando la Tecnica de Algoritmos Geneticos,
turbines, Trans. ASME J. Eng. Gas Turbine and Power 120 (2) (1998) 350-357. PhD thesis. Universidad Politecnica de Madrid, Madrid, 2004.
[11] J.F. Galopin, Going supercritical: once-through is the key. Mod. Power Syst. [29 RK. Shah, D.P. Sekulic, Fundamentals of Heat Exchanger Design. John Wiley
(December 1998) 39-42. and Sons, Inc, Hoboken, New Jersey, 2003.
[ 12] J.M. Beer, High efficiency electric power generation: the environmental role. [30 X. Cheng, T. Schulenberg, Heat Transfer at Supercritical Pressures-Literature
Prog. Energy and Combust. Sri. 33 (2007) 107-134. Review and Application to an HPLWR FZKA 6609. Wissenschaftliche Berichte,
] 13] M. Valdes, A. Rovira, M.D. Duran, Influence of the heat recovery steam generator Forschungszentrum Karlsruhe, 2001.
design parameters on the thermoeconomic performances of combined cycle [31 I.L. Pioro, R.B. Duffey, Experimental heat transfer in supercritical water
gas turbine power plants. Int. J. Energy Res. 28 (2004) 1255—1267. flowing inside channels (survey). Nucl. Eng. Des. 235 (2005) 2407-2430.