Leonardo Torbidoni

e-mail: torbidoni@unige.it

Analytical Blade Row Cooling Model for Innovative Gas Turbine Cycle Evaluations Supported by Semi-Empirical Air-Cooled Blade Data
With the objective of performing reliable innovative gas turbine cycle calculations, a new procedure aimed at evaluating blade cooling performance is presented. This complete analytical (convective and lm) blade cooling modeling provides the coolant mass ow and pressure loss estimation, and is a useful tool in the eld of innovative gas turbine cycle analysis, mainly when alternative uids are considered. In this case, in fact, the conventional semi-empirical data based on the use of air as traditional coolant and working media are no longer suitable. So the analytical approach represents a way of properly investigating alternative cooling methods and uids. In the presented analysis the effects of internal blade geometry on cooling performance are summarized by the Z parameter, which also highly affects the coolant ow pressure losses. Since existing technology represents a natural starting point for the assessment of Z, the model is able to automatically estimate a proper value relying only on available semi-empirical data which were established for air-cooled gas turbine blades. When alternative uids are considered, the same estimated value of Z is still maintained for the calculation, with the result of investigating the performance of existing blade technology for novel operational conditions. This represents an example of how the analytical approach, supported by conventional air-cooled blade semi-empirical data, appears as an innovative tool in the analysis of novel gas turbine cycles. In fact, the simulation results for the cooled blade were easily employed on the whole system level (gas turbine). DOI: 10.1115/1.1707030

Aristide F. Massardo
e-mail: massardo@unige.it Thermochemical Power Group, Dipartimento di Macchine Sistemi Energetici e Trasporti, ` di Genova, Universita Via Montallegro 1, 16145 Genova, Italia

Introduction
In a society where economic and low-polluting power production has become of vital importance, the development of cheap, new, and clean power plants is important. In such a scenario the advent of the deregulated electricity market has reinforced the quest for cheap, clean, reliable and efcient means of heat and electric power generation. This ongoing trend, which is very likely to continue and even increase in the future, not only includes attempts to improve existing thermal cycles but also to develop new ones, most of which consider alternative uids and working media. In the eld of cycles based on the use of gas turbines, new solutions have been proposed, such as: the humid air turbine cycle, 1,2, the SOFC-GT cycle, 3,4, and the Matiant cycle, 5. Moreover, the realization of steam cooling in combined cycles is the main apparent example of innovative design at a cycle level, 6. Another important aspect is that the step from the theoretical analysis to the hardware demonstration of the new technologies should require signicant effort in research and technological development, 7,8 and could comprise considerable risk and high cost. This in turn means that it is uncertain how many and what innovative gas turbine cycles will actually be turned into operating power plants. Hence, it is necessary to identify what cycles are technically and economically interesting at an earlier stage, 2,9. A natural starting point for the hardware demonstration of an
Contributed by the International Gas Turbine Institute IGTI of THE AMERICAN SOCIETY OF MECHANICAL ENGINEERS for publication in the ASME JOURNAL OF ENGINEERING FOR GAS TURBINES AND POWER. Paper presented at the International Gas Turbine and Aeroengine Congress and Exhibition, Amsterdam, The Netherlands, June 3 6, 2002; Paper No. 2002-GT-30519. Manuscript received by IGTI, Dec. 2001, nal revision, Mar. 2002. Associate Editor: E. Benvenuti.

innovative cycle should be existing gas turbine technology. Therefore, it is important to assess how existing gas turbines will operate, adopting alternative coolants and working media which are often of interest in innovative cycles. The present paper deals with the problem of evaluating the inuence of changing uid composition on the analysis of the performance of innovative gas turbine-based cycles, with the purpose of increasing the reliability of the results at the gas turbine system level. A method for estimating the performance and the required technological development only for convectively cooled vanes which are operating in the presence of alternative media has been previously proposed by the authors, 10. In the present work, this method has been carefully revised and extended to include lm cooled blades too. The adopted analytical approach allows the blade cooling analysis in presence of alternative uids to be obtained. At the same time, the assessment of a few parameters through the use of the air cooled blade semi-empirical data not only should improve reliability, but also avoid the need of a large number of input data mainly geometrical data necessary to run such an analytical calculation. This makes the model a worthwhile and user-friendly tool when blade cooling performance has to be evaluated at the gas turbine system level for the investigation of traditional and innovative cycles.

Film Cooling Model

In a previous work, 10, the authors have focused on the case of convective blade cooling analysis. In the present work this procedure has been carefully revised and extended to include lm Transactions of the ASME

498 Vol. 126, JULY 2004

Copyright 2004 by ASME

Downloaded From: http://gasturbinespower.asmedigitalcollection.asme.org/ on 10/18/2013 Terms of Use: http://asme.org/terms

cooled blade rows too. A brief description restricted to the lm cooling analysis is presented below. A wider explanation of the procedure is available in the cited work. The model provides the blade cooling analysis with an analytical approach, and it can be used to evaluate, among other things: i the blade row coolant mass ow rate; ii the blade internal geometry data, necessary for the evaluation of iii the pressure drop of the coolant that ows through the blade; iv the main ow hot gas side Stanton number and v the adiabatic wall lm effectiveness; vi the temperature of the coolant as it leaves the blade at the lm feeding outlet section, and vii the ratio of cooling mass ow used to feed the lm to the total cooling mass ow. It should be emphasized that since the model is analytical, it can take blade row geometry, cooling methods and changing alternative uids composition inuence on blade cooling performance into account in a detailed way, and it can also calculate the pressure drop of the coolant as it passes through the cooling channels inside the blade. On the other hand analytical models generally require a large number of input data mainly geometrical data, which are rarely available at a cycle level analysis. Concerning the internal geometry of the blade, i.e., the conguration of the cooling channels inside the blade, a number of geometrical input data could be worth summarising in the value of the cooling system technology level parameter Z, dened as
0.8 1.2 Z i 0.2 h n p E h c / d .

replaced by the lm average temperature T F , and the average external heat transfer coefcient h g by the one for the lm condition, when available. This last is a very important aspect as the external convective heat transfer coefcient varies in the presence of lm cooling. The disturbance caused by the hole ribs and the injected air can cause transition in a laminar boundary layer, so that it is possible for lms to increase, rather than decrease, the convective heat ux to the blade wall, 12. Since it was far easier to measure steady state temperatures produced by the lm on an adiabatic wall, lm cooling investigations have produced correlation of measured lm performance in terms of adiabatic wall effectiveness, and only rarely the more onerous wall heat ux measurements were made. This meant that such adiabatic wall measurements produced no data on the effect of lm injection on the convective heat transfer coefcients. This lack of knowledge forces the analyst to assume the same correlation as in the case of purely convective cooling to evaluate the external heat transfer coefcient in the presence of lm cooling, 12. As a result the adiabatic wall lm temperature T F was obtained from the adiabatic wall lm effectiveness ad , which is dened:

ad

T g T F . T g T co

(2)

(1)

When set, this parameter permits the full determination of the inuence of the internal blade geometry on blade cooling performance. In this way, the mass ow requirements and pressure losses can be carefully determined, not only for conventional uids but also for new coolants and new media, such as steam or carbon dioxide. Increasing Z means increasing the cooling system performance, and as a consequence the costs of the blade row itself, i.e., Z is also of importance from an economic point of view. Values of Z are usually calculated through Eq. 1, which requires knowledge of the detailed internal blade geometry. Unfortunately such information is not easily available when the investigation is carried out at a cycle level of analysis. As described below, the alternative way adopted in the model was to estimate Z by relying on knowledge based on existing technology. This represents a point of innovation introduced by the analytical model, which allows the blade cooling system technology level parameter to be automatically determined on the basis of the air cooling semi-empirical data established for air-cooled conventional gas turbines. The simultaneous use of semi-empirical data permits a proper evaluation of the inuence of blade geometry, which is tted to actual engine performance. In this way, not only is the reliability of the results expected to improve, but also the required input to run the analytical calculation is reduced. The analytical procedure considers the blade row geometry and the inuence of alternative uids and cooling methods on the heat transfer through the cooled blade wall in detail. The pressure loss of the coolant that passes through the cooling channels of the blade is also estimated. The thermal analysis of the lm cooled blade row is almost the same as the one presented by the authors for the convective cooled blade row, 10,11 except that the hot gas temperature T g must be

The value of ad was calculated by means of the semiempirical correlation proposed by Goldstein and Haji-Sheikh 13, which suits the injection of lm through a continuous slot over at plates:

ad
1 0.329 where

c p,g g c p,c c

1.9 Prg 2/3

0.2

(3) 0.8

x Rew 0.25 wM

1 0.00015 Rew
and where

c Wg sin g Wc

x is the distance in downstream chordwise direction from the injection location; w is the injection slot width; M ( c U c )/( g U g ), which is the blowing rate and U c is the coolant velocity as it is injected to feed the lm; Rew(Ucw)/c , which is the Reynolds number referred to the injection velocity. The term takes into account the injection angle different from zero for the coolant lm. When is equal to zero, is equal to one, and the coolant is injected in the direction parallel to the blade surface. Taking the injection geometry into account it is then possible to translate the semi-empirical correlation Eq. 3 into the more useful form

ad

c pg x 0.2 1 0.329 Re g c pc c

0.8

1.9Pr2/3 g

2 a t g

mg mc

1 Wg 0.00015 Reg sin rfc Wc

(4)

Journal of Engineering for Gas Turbines and Power

JULY 2004, Vol. 126 499

Downloaded From: http://gasturbinespower.asmedigitalcollection.asme.org/ on 10/18/2013 Terms of Use: http://asme.org/terms

where r f c is the ratio between the coolant mass ow that feeds the lm and the total coolant mass ow entering the blade row m c i.e., a measure of the feeding rate of the lm, while the ratio x / c was xed equal to 0.5 in the present calculation. The analytical procedure starts producing the lm cooling analysis applied at an elementary blade surface in a radial direction, and evaluates i. the thermophysical properties of the uids which are exchanging the heat ow; ii. the parameters which dene the effects of the lm on the boundary layer; iii. the parameters which govern the convective heat ow on the gas side of the blade; iv. the conductive heat ow through the blade wall; v. the convective heat ow on the coolant side of the blade; vi. the heat ow rate entering the blade. Then integrating along the blade height, the procedure can estimate among other things: i. the distribution of the metal blade temperatures on the gas and on the coolant side of the blade; ii. the temperature of the coolant itself that ows through the cooling channels inside the blade; iii. the maximum blade wall metal temperature achieved; iv. the pressure loss of the coolant owing through the cooling channels, basing on the knowledge of the internal blade geometry Z parameter and under the assumption of the incompressibility of the coolant gas. Therefore, an iterative procedure is necessary in order to estimate the minimum coolant mass ow m c which allows the blade row metal temperature to be maintained below its maximum permissible value T bmx i.e., a constraint restricted by material and structural requirements. The analytical approach of the algorithm makes it a useful tool for the analysis of cooling in the presence of different cooling methods closed-loop, open-loop, TBC, and lm cooling and different working means and coolants e.g., dry and humid air, steam, carbon dioxide. In this case the changes in uid properties are not only considered through a variation of the specic heat c p value for both the hot gas and coolant which are assumed as semi-perfect gasses, but also by varying the viscosity , density and all the related parameters, such as the average Stanton and Prandtl numbers. The Stanton number value for the hot gas was determined by the relationship suggested by Louis, 14:
0.37 2/3 Stg 0.5 Re Pr g g

Fig. 1 Flow chart of the blade row cooling ow and pressure losses procedure conventional and alternative uid

(5)

while for the coolant, Colburns one was adopted:

0.2 2/3 Stc E h 0.023 Re Pr . c c

(6)

Instead, the Mach number of the mainstream was iteratively calculated on the basis of the knowledge of the expansion inlet section and the mainstream mass ow. As already stated, the key parameter for the modeling is the parameter Z. This parameter should be evaluated from the knowledge of the internal blade geometry Eq. 1, data that is rarely available in open literature. On the contrary, Z is the free parameter used in the present calculation to obtain concordance between the output of the analytical procedure and the semi-empirical data, which were established for air-cooled gas turbines: as a result Z is evaluated in accordance with the actual engine performance. Furthermore, Z determines among other things: i the number of cooling channels n ch inside the blade, ii the number of passages n p for each channel through the blade 1, 2, or 3, and iii the cooling channel hydraulic diameter d, which must be superior to an assumed minimum value given as input. The proper Z parameter setting based on air-cooled blade semiempirical data is possible as the analytical code is supported by an 500 Vol. 126, JULY 2004

iterative procedure which considers a calculation unit called a semi-empirical assessment module, Fig. 1. This module allows the estimation of i the air mass ow m c required to cool the blade row and ii the temperature T co of the coolant as it leaves the blade to feed the lm layer, simply relying on information derived from available semi-empirical curves a detailed explanation is presented in the Appendix. In Table 1, a list of input and output variables for the code is presented in a manner that distinguishes between the cited assessment module and the main analytical calculation branch: that is the reason why a few outputs are considered as input for the semi-empirical module see also Fig. 1. When the external blade geometry parameters in Table 1 are known, together with the thermodynamic properties of the coolant and of the gas surrounding the blade, the main analytical calculation output m c , T co and the coolant ow pressure loss depend on the value assigned to the parameter Z. Also, the assessment module estimation of m c and T co depends on the free parameter T bu , which is the uniform blade temperature in a standard blade, 10. Since both the procedures describe the same air-cooled blade, the calculated values of the two outputs m c and T co should be equal, i.e., it is possible to write schematically: m c , A Z m c , B T bu T co , A Z T co , B T bu (7) (8)

where the subscripts A and B refer to the main analytical procedure and to the semi-empirical assessment module, respectively. If no other free parameter has been left other than the two above mentioned, it follows that just one pair of values exists Z * and T * bu , which simultaneously veries the two equations. The integrated analytical procedure is illustrated in Fig. 1. It starts with the knowledge of the component level e.g., external blade geometry that provides, together with the assumption of dry air as coolant, the required input data for the model. Values of Stg and ad are provided to the semi-empirical assessment module Transactions of the ASME

Downloaded From: http://gasturbinespower.asmedigitalcollection.asme.org/ on 10/18/2013 Terms of Use: http://asme.org/terms

Table 1 Input and output for the main analytical procedure and the semi-empirical assessment module Main Analytical and Semiempirical Procedure T g , p g , m g , T ci , p ci , coolant and gas composition m c , T co , c pc , c pg Semi-empirical Assessment Module Stg , ad , A b / A g , Semiemp curve Fig. A1 T bu

Main Analytical Procedure Input T bmx , r f c , Bibw , 10 geometry parameters Z p c / p ci , d , Stg , ad , Stc , h c / h g , n p

Free parameter Output

by means of an iterative procedure, which allows the analytical model to be supported by the knowledge based on existing technology. In this way the unique combination of values Z * and T * bu is found, guaranteeing the agreement with the air-cooled blade semi-empirical data refer to Eqs. 7 and 8. In this way the effects of the internal blade geometry on blade cooling are assessed, which provides the following advantages: i. arbitrary setting of important parameters is avoided; ii. internal blade geometrical data Z parameter useful for cooling analysis are derived even if the detailed knowledge of the blade geometry is not available; iii. the acquired knowledge of the internal blade geometry allows the calculation of the coolant ow pressure losses and gives some information on blade design requirements; iv. after the Z parameter has been set, based on available semiempirical data, which have been established for conventional air cooled blades, investigation of alternative coolants and working media can be carried out for a given internal blade geometry that should increase the reliability of the results see Fig. 1; v. information can be obtained about the required changes in internal blade geometry, as thermodynamic properties around the studied gas turbine blade row are altered e.g., for different values of T g , T ci , etc.. Moreover, the presence of temperature uctuations i.e., hot spots in the ow leaving the combustion chamber is taken into account by using an estimated maximum value for the gas temperature: T MAX T g K comb T comb g (9) where T comb is the temperature rise in the combustion chamber 12,15,16. The combustion chamber pattern factor K comb is an empirical constant which depends on the type of combustion chamber aero or industrial and the position of the blade row with respect to the combustion chamber outlet section; adopted values are: 0.1 in the rst nozzle, 0.05 in the rst rotor, and 0.025 in the rest of the cooled downstream.

described, for given values of T g and T bmx , the procedure produces the sole existing couple of values of Z and T bu that satisfy Eqs. 7 and 8. The identied operating point ( m c / m g 0.041; T co 983 K), that is dened as the reference point, is characterized by T * bu 1199 K and Z * 82. Figure 2 shows the reference point in a diagram where the coolant ow rate is plotted versus the hot gas temperature. Varying T g in the range 14002100 K, when the rest of the input data are considered as xed and coincident with the reference point, the coolant mass ow rate is evaluated for different values of Z and the maximum permissible material temperature T bmx . The points on the lines in Fig. 2 verify the conditions i m c , A m c , B , and ii T co , A T co , B as described above. The model is an efcient tool for determining not only the coolant mass ow rate, but also the appropriate value of Z for a novel blade, of the same type lm cooling as the reference blade. The determined new value of Z gives information on the required internal blade geometry, which could be useful for initial studies on how to develop internal blade conguration. As an example, starting from the reference point, if the hot gas temperature rises, the cooling system can be considered at xed Z xed geometry and technology while the material performance must be increased iso-Z curves. On the contrary, if the material performance is xed, the improvement of the cooling system can be obtained through a quantied variation of the Z value iso-T bmx curves. When varying T g , the variation of the Z parameter along the iso-T bmx curves appears to have a point of maximum value. Z rises when the gas temperatures are low and vice versa. The reason for this has to be found in the shape of the adopted semiempirical curve, that is shown in Fig. A2. It appears that with the

Table 2 Reference point input data Hot Gas Gas composition Tg pg mg Coolant Coolant composition T ci p ci Model Cooling method External geometry Bibw T bmx rfc Film cooling 10 input 0.5 1273 K 0.35 Dry air 687 K 1.64 MPa Products from dry-air combustion 1723 K 1.6 MPa 143 kg/s

Model Implementation
The model was originally created as a Fortran code and then loaded as a dynamic link library dll into the commercial software IPSEpro, 17, as described by Jordal and Torbidoni 18. Specic heat and viscosity are calculated on the basis of data reported by Perry and Green 19, while the rest of the thermodynamic properties are calculated from the IPSEpro data base.

New Model Analysis of a Film Cooled Blade Row

The new model has been applied to evaluate the performance of an existing air lm-cooled blade row, external geometry of which was known. The input data are presented in Table 2. As previously Journal of Engineering for Gas Turbines and Power

JULY 2004, Vol. 126 501

Downloaded From: http://gasturbinespower.asmedigitalcollection.asme.org/ on 10/18/2013 Terms of Use: http://asme.org/terms

Fig. 2 Coolant mass ow, blade maximum temperature and corresponding values of Z , obtained through a variation of gas temperature T g for the air lm cooled blade row

purpose of keeping the required coolant mass ow as low as possible and with it the gas turbine expansion losses connected with the mixing of the lm layer with the main stream Z increases with increasing T g as long as the gas temperatures and the heat ow entering the blade are low. On the contrary, when the gas temperatures and the entering heat ow become too high, Z decreases with increasing T g to provide lower convective cooling efciency c Eq. A1. Therefore a higher amount of coolant feeds the lm i.e., higher m c required with respect to a constant lm feeding rate r f c ) with lower T co , this allowing higher lm effectiveness ad Eqs. 2 and A4 and enhances the effect of the barrier for the heat ow entering the blade.

Coolant Pressure Losses. Considering the reference point in Fig. 2, it is possible to evaluate the coolant ow pressure loss based on the known value of Z ( Z * 82). In a wider approach it is possible to evaluate the relative pressure losses p c / p ci for all the points presented in the diagram of Fig. 2. The results are shown in Fig. 3, where coolant pressure losses are plotted versus hot gas temperature T g with Z and T bmx as parameters for the air lm-cooled blade row. Even if they deal with two completely different gas turbine blade rows, it is worthwhile comparing the diagram from lm cooling presented here with the one for purely convective cooling

Fig. 3 Coolant pressure losses, blade maximum temperature, and corresponding values of Z , obtained through a variation of gas temperature T g for air lm cooled blade row

502 Vol. 126, JULY 2004

Transactions of the ASME

Downloaded From: http://gasturbinespower.asmedigitalcollection.asme.org/ on 10/18/2013 Terms of Use: http://asme.org/terms

Fig. 4 Coolant mass ow required by the blade row when lm cooled with air dotted lines or carbon dioxide solid line, for varying gas temperature T g , maximum blade temperature T bmx and technology level parameter Z

presented by the authors in their previous work i.e., Fig. 6 in 10. It becomes evident that the relative pressure loss in convective cooling rises more rapidly with T g than in lm cooling. This is determined by the increasing value of Z, which is the only way to deal with the effects of the rising gas temperature in a convectively cooled blade: In fact, in this case Z can reach very high values. In accordance with Eq. 1, higher values of Z correspond to more cooling channels with lower hydraulic diameter d, which in turn means a higher pressure loss for the coolant ow. A dif-

ferent situation is presented in the diagram in Fig. 3, where the adopted lm cooling method allows the Z parameter values to be kept low and so to keep pressure losses low.

Film Cooling With Alternative Coolants

A sample of the analysis for lm cooling performed with alternative coolants is presented here.

Fig. 5 Relative coolant pressure drop for the blade row when lm cooled with air dotted lines or carbon dioxide solid line, for varying gas temperature T g , maximum blade temperature T bmx and technology level parameter Z

Journal of Engineering for Gas Turbines and Power

JULY 2004, Vol. 126 503

Downloaded From: http://gasturbinespower.asmedigitalcollection.asme.org/ on 10/18/2013 Terms of Use: http://asme.org/terms

The parameter Z having been determined, the internal geometry of the blade is sufciently known to evaluate the cooling of the same blade row when alternative coolants are adopted in place of dry air see Fig. 1. The calculations were carried out using exactly the same input data as for Figs. 2 and 3, except for the different coolant composition which was carbon dioxide. For comparison, the results for air cooling shown in Figs. 2 and 3 are reproduced in dotted lines in Figs. 4 and 5, respectively. The diagram in Fig. 4 shows the coolant mass ow required versus hot gas temperature and Fig. 5 shows coolant relative pressure drops. It can be seen that carbon dioxide cooling substantially reduces the coolant mass ow requirement with respect to that of air. The pressure losses inside the blade are also greatly reduced. Considering that the cooling system geometry, and with it also the cooling channel diameter, are equal for the same values of Z, the reduction of the pressure loss for the carbon dioxide cooling depends on i the different thermophysical properties of the coolant and ii the different rate of the coolant mass ow required i.e., a different coolant velocity through the constant cooling channel geometry. Since the pressure losses depend also on the square of the velocity, their reduction becomes more evident for high values of the hot gas temperature, due to the increasing coolant velocity. It is possible to conclude that the carbon dioxide cooling performs better than that with dry air, and also that the improvement is much more evident for high values of the hot gas temperature. In that range, in fact, the alternative coolant allows for a sensitive reduction of both the coolant pressure loss and the coolant mass ow rate requirement this last being also connected with the gas turbine expansion losses caused by the mixing of the lm layer with the hot gas stream.

method for the analysis of the cooled gas turbine on a cycle level with increased reliability for alternative coolants and working media.

Appendix
The Semi-Empirical Assessment Module. The semiempirical assessment module relies on knowledge based on existing technology in order to allow the presented analytical code to set a proper value for the cooling system technology level parameter Z. It uses conventional air cooled blade semi-empirical curves to calculate the amount of coolant m c required to cool one gas turbine blade row and the temperature T co of the coolant as it leaves the blade and mixes with the hot gas stream. The fundamental equations for the module can be found in Holland and Thawke 20 and the procedure is based on the standard blade assumption presented by Halls 21. A standard blade has a uniform blade temperature T bu equal to the average temperature of that in the actual blade, the same geometry as the actual blade, and the cooling air is warmed up to the uniform blade temperature before leaving the blade. The cooling efciency c describes the ratio between the temperature rise of the coolant ow in the actual blade and in the standard blade:

T co T ci . T bu T ci

(A1)

Conclusions
In order to perform reliable innovative gas-turbine cycle calculations, a new computational tool for the analysis of lm cooled blades and vanes is proposed. It consists of an analytical procedure for the investigation of the lm cooling starting with the knowledge of the blade row geometry required to run the simulation. The effect of the internal blade geometry i.e., cooling channels geometry on the cooling performance is considered through the cooling system technology level parameter Z. The procedure is able to automatically set a proper value for Z by means of a semi-empirical assessment module. As the assessment module relies on semi-empirical data, the Z setting is determined with respect to existing technology. Once Z is determined, the internal blade conguration is determined, which makes it possible to calculate the pressure losses of the coolant and the behavior of a given blade conguration in an environment consisting of a new cooling medium and/or a new hot gas composition. It is also possible to study how Z should be altered for, e.g., by varying hot gas temperature, in order to avoid excessive coolant mass ow or a pressure drop. This means that valuable information on blade design requirements for alternative uids can be obtained. Therefore the following main conclusions may be stated: i. the method proposed here allows a very detailed analysis of blade cooling inuence on innovative gas turbine cycle performance, and it is a noteworthy step in this eld; ii. when traditional cooling uids are investigated, the model represents a very detailed method to evaluate not only the mass ow rate m c , but also the coolant pressure loss; iii. when alternative uids are investigated and the semiempirical data become not anymore available, the proposed analytical model overcomes the limits of the semi-empirical approach, thanks to the preliminary Z analysis carried out with air as coolant. Planned future work is to extend the integration approach to include the analytical expansion model. This should provide a 504 Vol. 126, JULY 2004

For lm cooled blades, T co is the temperature to which the coolant ow is warmed before emerging from the blade to feed the lm. Instead, T ci is the temperature of the coolant entering the blade row at the inlet section. When the lm cooling method is adopted, the value of the cooling efciency is determined by the equation

c ad
B ad 1 c

(A2)

where the cooling effectiveness c is dened as c T g T bu T g T ci (A3)

and for the adiabatic wall lm effectiveness ad Eq. 2 is adopted. The cooling parameter B, referred to as w by Holland and Thawke 20, is included in the expression mc Ab c pg B St . mg c pc g A g (A4)

The relationship between c and B is determined through the use of semi-empirical data, a principle that was previously used n 22. Semi-empirical diagrams of c as a function of B by Rose can be found in works by Haselbacher 23 and Mukherjee 24. The data from these two references are plotted in Fig. A1. In the present work, the data comes from the diagram presented by Haselbacher, and the curve adopted in the calculation is the average one between the lower and upper limit curves that describe the lm cooling Fig. A2. Thermal analysis for the lm cooled blade is the same as that for the convective cooled blade, 10, except that the hot gas temperature T g must be replaced by the average lm temperature T F . Another very important aspect is how the external convective heat transfer coefcient varies in the presence of lm cooling. The lack of knowledge forces the analyst to assume the same correlation as in the case of purely convective cooling to evaluate the external heat transfer coefcient also in the presence of lm cooling, 12. Thus Eq. A2 produces the expression for the effect of lm cooling, expressed in terms of adiabatic wall lm effectiveness ad , on the cooling efciency of a convective plus lm cooled blade c . The same Eq. A2 can be transposed in such a way as to give for a purely convective cooled blade: Transactions of the ASME

Downloaded From: http://gasturbinespower.asmedigitalcollection.asme.org/ on 10/18/2013 Terms of Use: http://asme.org/terms

Fig. A1

c as a function of B , reproduced with data from Haselbacher dotted lines, and Mukherjee solid lines

Fig. A2

c as a function of B , reproduced with data from Haselbacher for lm cooling i.e., average curve

cFA

c . B 1c

(A5)

By substituting the appropriate values for the lm cooled blades on the right-hand side of Eq. A5, the apparent convective cooling efciency of a lm cooled blade cFA is obtained. The efciency value deriving from Eq. A5 is the value to be read when the lm cooled blade is plotted on the conventional correlation graph such as in Figs. A1 and A2, 12. Apart from selecting the appropriate curve from Fig. A1, the parameters to be set in order to run a solution are the geometry parameter A b / A g i.e., wetted blade and adjacent surfaces to be cooled over average gas cross-sectional area, the hot gas side Journal of Engineering for Gas Turbines and Power

Stanton number Stg , the adiabatic wall lm effectiveness ad , and the uniform blade temperature T bu . The value of A b / A g was set equal to ve in the work presented by Jordal 25, a value that was previously proposed by Bolland and Stadaas 26. The same value has been adopted based on the external geometry of the blade row investigated here. The value of Stg was provided by correlation with the main analytical code and so passed to the assessment module as shown in Fig. 1. The adiabatic wall lm effectiveness ad is also adopted as it is evaluated by the code. The input and output for the semi-empirical module are listed in Table 1. The main parameter that remains to be set in order to complete the input for one blade row is the uniform blade temperature T bu . JULY 2004, Vol. 126 505

Downloaded From: http://gasturbinespower.asmedigitalcollection.asme.org/ on 10/18/2013 Terms of Use: http://asme.org/terms

Nomenclature
A b wetted surface of the blade row bladesshrouds A g actual gas cross-sectional area a t ratio between bladeshroud surface and blade surface B dimensionless coolant parameter Bi Biot number c chord c p specic heat at constant pressure d cooling channel hydraulic diameter dll dynamic link library E h parameter that considers increased surface of cooling channels due to turbulators h convective heat transfer coefcient K comb combustion pattern factor M lm blowing rate m mass ow rate n ch number of the cooling channels inside the blade n p number of cooling channel passages p pressure Pr Prandtl number Re Reynolds number r f c fraction of the coolant ow used for lm cooling St Stanton number T temperature t b average blade thickness, blade cross-sectional area/c TBC thermal barrier coating U velocity W molar weight w injection slot width x distance in downstream chordwise direction from the lm injection location Z blade cooling system technology level parameter lm injection angle h ratio, empty blade cross-sectional area/c 2 arithmetic difference c cooling effectiveness c cooling efciency ad adiabatic wall lm cooling effectiveness cFA apparent convective cooling efciency of the lm cooled blade dynamic viscosity kinematics viscosity density solidity g parameter that accounts for the reduction of the gas cross-sectional area due to the blade thickness, 1 *t b / c i parameter that considers inuence of heat transfer between adjacent cooling channels Subscripts b bw c comb F g i blade blade wall coolant, cooling, coolant side combustion chamber lm layer gas, gas side inlet

mx o u w

maximum outlet uniform related to the lm injection slot

Superscripts
MAX

* reference point
maximum

References
1 Rao, A. D., 1989, Process for Producing Power, U.S. Patent No. 4,829,763. 2 Traverso, A., and Massardo, A. F., 2002, Thermoeconomic Analysis of Mixed Gas-Steam Cycles, Appl. Therm. Eng., 22, pp. 121, Elsevier, New York. 3 Harvey, S. P., and Richter, H. J., 1994, Gas Turbine Cycles With Solid Oxide Fuel Cells, ASME J. Energy Resour. Technol., 116. 4 Massardo, A. F., and Lubelli, F., 2000, Internal Reforming Solid Oxide Fuel CellGas Turbine Combined CyclesPart I: Cell Model and Cycle Thermodynamic Analysis, ASME J. Eng. Gas Turbines Power, 122, pp. 2735. 5 Mathieu, P., and Nihart, R., 1998, Zero Emission Matiant Cycle, ASME, Paper No. 98-GT-383. 6 Stambler, I., 2000, Steam-Cooled H-Technology Machines Set for Commercial Operation in 2002, Gas Turb. World, JanFeb. 7 Layne, A. W., and Hoffman, P. A., 1999, Gas Turbine Systems for the 21st Century, ASME, Paper No. 99-GT-367. 8 Layne, A. W., 2000, Developing the Next Generation Gas Turbine SystemsA National Partnership, ASME Paper No. 2000-GT-176. ` , M., 2000, Thermoeconomic Analysis of Gas 9 Massardo, A. F., and Scialo Turbine Based Cycles, ASME J. Eng. Gas Turbines Power, 122. 10 Jordal, K., Torbidoni, L., and Massardo, A. F., 2001, Convective Blade Cooling Modeling for the Analysis of Innovative Gas Turbine Cycles, ASME Paper No. 2001-GT-0390. 11 Torbidoni, L., 2000, Espansione Refrigerata nel Turbogas al Variare delle Tecniche e dei Fluidi di Raffreddamento, Tesi di Laurea, DIMSET, Italy. 12 Holland, M. J., 1992, Craneld Lecture NotesRotor Blade Cooling in HP Turbines, Technical Report, Department of Turbine Technology, Ref. GN30884, Rolls Royce plc. 13 Goldstein, R. J., and Haji-Sheikh, A., 1967, Prediction of Film Cooling Effectiveness, Proc. 1967, Semi-International Symposium (Tokyo), Japan Society of Mechanical Engineers, Tokyo, pp. 213218. 14 Louis, J. F., 1977, Systematic Studies of Heat Transfer and Film Cooling Effectiveness, AGARD CP 229. 15 Young, J. B., and Wilcock, R. C., 2001, Modeling the Air-Cooled Gas TurbinePart II: Coolant Flows and Losses, ASME Paper No. 2001-GT0392. 16 Consonni, S., 1990, Performance Prediction of Gas/Steam Cycles for Power Production, MAE Dept., Ph.D. thesis No. 1893-T, Princeton University, Princeton, NJ. 17 Anon, 1998, IPSEpro User Documentation, Simtech Simulation Technology, Graz, Austria. 18 Jordal, K., and Torbidoni, L., 2000, Linking of a Fortran Code to IPSEpro, DIMSET, Internal report TN 4-2000, Italy. 19 Perry, R. H., and Green, D., 1984, Perrys Chemical Engineers Handbook, 6th Ed., McGraw-Hill, New York. 20 Holland, M. J., and Thawke, T. F., 1980, Rotor Blade Cooling in High Pressure Turbines, Int. J. of Aircraft Eng., 176. 21 Halls, G. A., 1969, Air Cooling of Turbine Blades and Vanes, AGARDograph No. 120. n, P. M., 1993, Evaporative Gas TurbinesA Thermodynamic Evalua22 Rose tion of Their Potential, Report No. ISRN/LUTMDN/TMVK-7010-SE, Lund University. 23 Haselbacher, H., 1989, Gas Turbines, Standard Handbook of Power Plant Engineering, Thomas C. Elliot, ed., McGraw-Hill, New York. r Schaufelku hlung der Gasturbine, Brown Boweri 24 Mukherjee, D. K., 1977, Zu Mitteilungen. 25 Jordal, K., 1999, Gas Turbine Cooling ModelingThermodynamic Analysis and Cycle Simulations, Report No. ISRN LUTMDN/TMVK-7034-SE, Lund University. 26 Bolland, O., and Stadaas, J. F., 1995, Comparative Evaluation of Combined Cycles and Gas Turbine Systems With Water Injection, Steam Injection and Recuperation, ASME J. Eng. Gas Turbines Power, 117.

506 Vol. 126, JULY 2004

Transactions of the ASME

Downloaded From: http://gasturbinespower.asmedigitalcollection.asme.org/ on 10/18/2013 Terms of Use: http://asme.org/terms