Detailed Film Cooling Measurements over a Gas Turbine Blade Using a Transient Liquid Crystal Image Technique

Detailed heat transfer coefficient and film effectiveness distributions over a gas turbine blade with film cooling are obtained using a transient liquid crystal image technique. The test blade has three rows of film holes on the leading edge and two rows each on the pressure and suction surfaces. A transient liquid crystal technique maps the entire blade midspan region, and helps provide detailed measurements, particularly near the film hole. Tests were performed on a five-blade linear cascade in a low-speed wind tunnel. The mainstream Reynolds number based on cascade eXit velocity is 5.3 l0s. Two different coolants (air and CO2) were used to simulate coolant density effect. Coolant blowing ratio was varied between 0.8 and 1.2 for air injection and 0.4-1.2 for CO2 injection. Results show that film injection promotes earlier laminar-turbulent boundary layer transition on the suction surface and also enhances local heat transfer coefficients (up to 80%) downstream of injection. An increase in coolant blowing ratio produces higher heat transfer coefficients for both coolants. This effect is stronger immediately downstream of injection holes. Film effectiveness is highest at a blowing ratio of 0.8 for air injection and at a blowing ratio of 1.2 for CO2 injection. Such detailed results will help provide insight into the film cooling phenomena on a gas turbine blade.


INTRODUCTION
In modern gas turbines, sophisticated cooling techniques are employed to cool the components from high temperature com- bustion gases.Some turbine blades are cooled by ejecting cooler air from within the blade through discrete holes to provide a pro- tective film on the surface exposed to hot gas path.Many studies have presented heat transfer measurements on turbine blades with film cooling.Nirmalan and Hylton (1990) and Camci and  Arts (1990) studied heat transfer coefficients on film-cooled tur- bine blades.Takeishi et al. (1992) compared the film effective- ness values for a stationary cascade with mainstream turbulence intensity under 4% and for a rotor blade using the heat-mass transfer analogy.Ito et al. (1978) and Haas et al. (1992)  stud-  ied the effect of coolant density on film effectiveness on turbine blades under low mainstream turbulence levels.Ito et al. (1978)   found that an increase in coolant-to-mainstream density ratio causes an increase in film effectiveness on both pressure and suction surfaces for a blowing ratio of 1.0.Haas et al. (1992) found that their results show the same trends as that of Ito et al. (1978).They reported that an increase in coolant density for a low blowing ratio of 0.5 causes a decrease in film effectiveness on the suction surface.However, an increase in coolant density causes an increase in film effectiveness at higher blowing ra- tios.Abhari and Epstein (1994) conducted heat transfer exper- iments on a film-cooled transonic turbine stage in a short du- ration turbine facility.They measured steady and time-resolved chord-wise heat flux distributions at three spanwise locations.They concluded that film cooling reduces the time-averaged heat transfer by about 60% on the suction surface compared to the uncooled rotor blade.However, the film cooling effect on time- averaged heat transfer is relatively low on the pressure surface.
Ou et al. (1994) and Mehendale et al. (1994) simulated un- steady wake conditions over a linear turbine blade cascade with film cooling.They studied the effect of unsteady wakes on a model turbine blade with multiple-row film cooling using air and CO2 as coolants.They measured heat transfer coefficients and film effectiveness at discrete locations using thin foil heat- ing and multiple thermocouples.Ou et al. (1994)  reported that   an increase in blowing ratio causes an increase in surface heat transfer coefficients, except in the transition region.They also reported the effect of coolant density on pressure and suction surface heat transfer coefficients.Mehendale et al. (1994) re- ported that a blowing ratio of 1.2 provides highest effectiveness for CO2 injection and a blowing ratio of 0.8 provides highest effectiveness for air injection.In the present study, detailed heat transfer coefficient and film effectiveness distributions are obtained in the near film hole region and over the entire blade surface using a transient liquid crystal technique.Such detailed distributions were not obtained by the previous studies using thin-foil-thermocouple measurement technique.
A few previous studies have measured detailed heat transfer coefficients on turbine blades without film cooling.Martinez-  Botas et al. (1995) presented detailed heat transfer coefficient distributions on a non-film cooled blade in an annular transonic cascade using a transient liquid crystal technique.Hoffs et al.  (1997) measured heat transfer coefficients in a linear cascade without film cooling using a transient liquid crystal technique and compared the results with measurements using the naphthalene sublimation mass transfer technique.Duet al. (1997) pre- sented detailed heat transfer coefficient distributions for a turbine blade without film cooling using the transient liquid crystal tech- nique.They studied the effects of upstream unsteady wakes with trailing edge coolant ejection at various free-stream turbulence conditions.Other studies such as Vedula and Metzger (1991)  and Ekkad et al. (1997a, b) presented detailed heat transfer co- efficient and film effectiveness distributions for film cooling on a flat surface with a single row of simple-angle or compoundangle holes.They used a transient liquid crystal technique for detailed heat transfer coefficient and film effectiveness measurement.
The present study uses a similar transient liquid crystal tech-  nique (Vedula and Metzger, 1991; Ekkad et al., 1997a, b; Du   et al., 1997) to measure detailed heat transfer coefficients and film effectiveness on a film-cooled turbine blade with air and CO2 injection.The present study presents detailed heat trans- fer coefficient and film effectiveness measurements on a blade under steady, low-turbulence flow conditions.This will help isolate the film cooling effect on the blade surface heat transfer distributions.

EXPERIMENTAL APPARATUS
Figure 1 shows the schematic of the test section and cam- era locations.The test apparatus consists of a low-speed wind tunnel with a suction-type blower.The five-blade linear cascade FIGURE 1 Schematic of test section and camera arrangement.is shown in the figure.The mainstream turns 107.49 and the flow is accelerated 2.5 times from inlet to exit of the cascade.
The test apparatus is described in detail by Ou et al. (1994).A heater box, shown in the figure, is used to preheat the middle test blade prior to the transient test.The same blade is coated with a thin layer of thermochromic liquid crystals.The blade surface color changes during the transient test are analyzed using a high-speed, high-resolution image processing system.The image processing system consists of four cameras indi- vidually connected to a color frame grabber board inside the PC.The cameras are focused using a color monitor.Commer- cial software is used to digitize the liquid crystal color changes.During a transient test, only one camera is operational.Since the color changes are processed real time and no frames are stored in the PC, the frame speed will be reduced if all the four cam- eras are operated at the same time.Hence, four separate tests are required to map the entire blade surface using four different camera locations as shown in the figure.Details on the image processing system are presented by Duet al. (1996).
Each blade in the linear cascade has an axial chord length of 17 cm and radial span of 25.2 cm.The blade-to-blade spacing is 7.01 cm at the cascade inlet and the blade span-to-throat ratio is 5.The blade configuration, scaled up five times, produces a velocity distribution typical of an advanced high-pressure tur- bine blade row. Figure 2 presents a section of the film-cooled turbine blade model.The coolant is supplied to various locations on the blade surface through five cavities.The first cavity sup- plies coolant to the three leading edge film hole rows and each of the other four cavities supply coolant to each row on the pres- sure and suction surfaces.Coolant is fed into each cavity from the bottom of the blade and flow rate into each cavity is con- trolled using a flowmeter.The coolant flow from each flowmeter is passed through a solenoid-controlled three-way diverter valve before the flow enters the coolant cavity inside the blade.Each solenoid-controlled valve is connected to a switch that triggers the coolant flow into the cavities at the instant the transient test is initiated.The blade film hole row geometry and configuration are shown in the figure.Figure 2 also presents a three-dimensional view of the pressure and suction surfaces of the test blade.The liquid crystal coated surface area is 15.2 cm wide and the data acquiring area is 7.6 cm wide along the midspan region of the test blade.
The new addition in this experiment compared to previous studies (Ou et al., 1994; Mehendale et al., 1994) is the heater box.The transient test requires that the blade be heated to a temperature higher than the liquid crystal color range (37.2C).During the transient test, the hot blade surface is suddenly cooled by exposing it to a cooler mainstream flow.The heater box has the blade profile and is slightly larger than the test blade.A gap of 10 mm exists between the blade outside surface and the heater box inner surface.The insides of the heater box are in- strumented with thin foil heaters and controlled using several variacs to provide a near uniform blade surface temperature.The heater box is lowered to completely cover the test blade during heating.The blade surface temperature is monitored us- ing embedded thermocouples during heating.The uniformity of surface temperature with heating is within -t-1.2C.An interpo- lation scheme was used to further reduce the temperature vari- ation in the initial surface temperature to within +0.2C.When the surface is heated to the required uniform temperature, the suction type blower is switched on.The mainstream reaches full flow within 20-30 seconds.Once the mainstream has reached required flow, the heater box is lifted up completely to expose the test blade to the mainstream.The coolant flow and the im- age processing system are automatically triggered at the same instant the test surface is exposed to the mainstream.The color changes during the transient test are monitored by the system.The times of color change on the blade surface from colorless to Row S I Row red at every pixel location is measured.Wide band liquid crystals (Hallcrest: R32C5W) were used for surface temperature mea- surements.However, in the present experiment, only the green- to-red color transition point (32.7 C) in the liquid crystal band is analyzed.

DATA ANALYSIS
A transient liquid crystal technique was used to measure the detailed heat transfer coefficients and film effectiveness on the blade surface.The technique is similar to the one described by Vedula and Metzger (1991) and Ekkad et al. (1995a, b).A one-dimensional transient conduction model into a semi-infinite solid with convective boundary condition is assumed.The solu- tion for surface temperature is obtained as T Ti where Tw is the wall temperature when liquid crystals change to red from green (32.7 C) at time t, Ti is the initial surface tem- perature, Tm is the oncoming mainstream flow temperature, and c and k are the thermal diffusivity and conductivity of the blade material respectively.The heat transfer coefficient is obtained from Equation (1).For film cooling tests, the local film temper- ature (Tu) which is a mixture of the coolant (To) and mainstream temperatures replace the mainstream temperature (Tm) in Equation (1).The film temperature is defined in terms of r/,, .which is the film effectiveness: For the film-cooling test, we obtain an equation similar to Equation [31 Two similar transient tests are run to obtain the heat transfer coefficient (h) and film effectiveness (r/).In the first test, the blade surface is heated and the coolant and mainstream tem- peratures are nearly the same.In this case, there is only one unknown, h, in the equation.For the second test, the coolant is heated to a temperature close to blade initial temperature.The calculated local heat transfer coefficient from the first test is substituted in the equation to obtain the local film effectiveness.
The above equation is solved at each point (57,760 points) on the blade surface to obtain the detailed heat transfer coefficient and film effectiveness distributions.
The average uncertainty in the film effectiveness measure- ment includes the additional uncertainty in heat transfer coef- ficient measurement and was estimated to be about 4-5.9%.It should be noted that the worst case uncertainty is in the imme- diate vicinity of the hole (less than 1-diameter around the hole) and close to the blade trailing edge where it could be as high as 4-17% due to invalidation of the semi-infinite model assumption.However, the semi-infinite solid assumption can be applied where thickness of material is higher than 0.51 cm.

RESULTS AND DISCUSSION
Experiments were performed at a cascade exit Reynolds num- ber of 5.3 x 105.The corresponding flow velocity at the cascade exit was 50 m/s.Two different coolants, air and CO2, are used to simulate coolant-to-mainstream density ratios of DR 1.0 and DR 1.5, respectively.Air as coolant was tested at blowing ratios of 0.8 and 1.2 and CO2 was tested at blowing ratios of 0.4, 0.8, and 1.2.The flow conditions for the film cooling tests are summarized in Table I.For all the results presented in this study, the oncoming free-stream turbulence intensity was measured to be about 0.75% at the cascade inlet.

Flow Measurement
Figure 3 presents the local-to-exit velocity ratio (V/V2) dis- tribution around the blade.A pressure tap instrumented blade was used to measure the surface static pressure distributions which were then converted to local mainstream velocity distri- bution around the blade (Ou et al., 1994).The flow is accelerated significantly on the suction surface from leading edge to about 50% of the suction surface, after which the velocity decreases a little toward the trailing edge.On the pressure surface, the flow accelerates a little with a sudden increase in velocity just down- stream of the leading edge to about X/PL 0.05.The flow then stays constant till X/PL 0.15 and then accelerates gradually toward the trailing edge.The pressure-tap-instrumented blade is of same profile as the heat transfer blade but does not have film holes.The prediction shown in Fig. 3 was calculated by GE's in-house code.

Effect of Blowing Ratio
Figure 4 presents the detailed Nusselt number distributions on the blade suction and pressure surfaces for CO2 injection and blowing ratios of 0.4, 0.8, and 1.2.Case is for a no-film hole  blade; case 2 is for M 0.4; case 3 is for M 0.8; case 4 is for M 1.2 for a blade with film holes.The blades with and without film cooling holes have the same shape and flow angles.The blades are also made from the same material (Ren Shape).
EffEct on suction surface.The no-film hole surface Nusselt number (case 1) levels drop significantly from the leading edge with increasing streamwise distance on the suction surface.Nusselt numbers are lowest around X/SL 0.5, after which the Nusselt numbers increase again.This is due to boundary layer transition to turbulence.Nusselt numbers are higher toward the trailing edge, as the transition is not complete.For a film cooled blade (case 2) with M 0.4, Nusselt number show streaks due to film cooling jets downstream of leading edge holes.High Nusselt numbers immediately downstream of injection decrease rapidly and the jet effect is nonexistent upstream of the first film hole row on the suction surface (S 1).Downstream of the row S 1, jet streaks of higher Nusselt numbers are obtained along the holes.The streaks extend all the way up to the next film hole row $2.However, the jets do not cause any Nusselt number enhancement between the holes for row S 1. Downstream of film hole row $2, the Nusselt numbers are significantly higher than for case 1.Some streaks of high Nusselt numbers are obtained along the holes.Such high levels of Nusselt number downstream of row $2 may be explained.Film injection may cause boundary layer in- stabilities, which cause earlier laminar-turbulent boundary layer transition.This may produce higher heat transfer coefficients of up to 80% higher downstream of injection.After completion of transition, the effect of coolant jets disappears and the Nusselt numbers are not affected by blowing ratio.Nusselt numbers de- crease after transition with the growth of the turbulent boundary layer.The streaks downstream of hole row S become stronger and appear to mix downstream with hole row $2.The effect of blowing ratio after hole row $2 is to increase the Nusselt number slightly.With an increase in blowing ratio, the jet-mainstream Effect of blowing ratio on detailed Nusselt number distributions for CO2 injection.
interaction increases causing more turbulence downstream.This may be the cause for higher Nusselt numbers with an increase in blowing ratio immediately downstream of injection.
Effect on pressure surface.For case 1, the Nusselt num- bers drop rapidly till X/PL -0.15 and then increase a little bit over the entire surface.For case 2, film injection has a very small effect immediately downstream of leading edge row injection.However, Nusselt numbers are enhanced between the leading edge row and first row P1 compared to case 1. Downstream of rows P1 and P2, film injection enhances Nusselt numbers up to 60% over case 1.The effect of film injection on the pressure sur- face is more in the region immediately downstream of injection.
Since the boundary layer on the pressure surface is thicker than on the suction surface, the effect on Nusselt numbers due to film injection is reduced.Jet streaks are not evident on the pressure surface.With further increases in blowing ratio from M 0.4 to M 1.2, Nusselt numbers are not significantly affected.The increase in injectant mass into a thicker boundary layer does not appear to disturb the boundary layer as significantly as in the case of the thinner boundary layer on the suction surface.Nusselt numbers with film injection are significantly en- hanced due to the boundary layer disturbance cause by injection.Earlier studies on film cooling have shown that film injection can create local turbulence intensities as high as 15-20% depending on the blowing ratio.With such high local turbulence, heat transfer coefficients downstream of injection are significantly enhanced as seen in the figure .Figure 5 presents the span-averaged Nusselt number distribu- tion for cases 1-4.The Nusselt number distributions for case 1 are the lowest on both pressure and suction surfaces.Laminar- turbulent boundary layer transition on suction surface occurs at about X/SL 0.55.Nusselt numbers are enhanced significantly with addition of film injection (case 2-4).As explained earlier, film injection disturbs the boundary layer and causes earlier transition to a turbulent boundary layer.Higher Nusselt numbers are obtained downstream of injection row S 1.Further increase in Nusselt numbers occurs with transition and addition of coolant at row $2. On the pressure surface, film injection produces higher heat transfer coefficients from leading edge in- 1600 Case M DR [e SmthSurf'(Ouetal',1994) jection location to downstream of hole row P2.Higher blowing ratio produces higher Nusselt numbers downstream of injection for both pressure and suction surfaces.
The present results are compared to results for the same cases from Ou et al. (1994).The results for the no-film holes blade are in good agreement with the present data.Ou et al. (1994) show much higher Nusselt numbers downstream of injection holes on the suction surface.However, the pressure surface Nusselt numbers with film injection are in good agreement with present results.There may be two reasons for the differences in the re- sults.The study by Ou et al. (1994) had the same coolant cavity feeding the hole rows P2 and $2.This may cause more coolant to exit out of hole row $2 and causing higher heat transfer coef- ficients.The pressure surface results are closer for both studies, as the blowing ratio does not have a significant effect on Nusselt numbers as compared to the suction surface.However, in the present study, two different cavities feed coolant to the hole rows.This may explain the lower Nusselt numbers obtained on the suction surface.Also, it is difficult to estimate the heat loss to the coolant during the steady state test by Ou et al.Since   the heat loss to coolant was not estimated, that could add to the error.Ou et al. ( 1994) measured four thermocouple locations in the spanwise direction at every axial location.
Figure 6 presents the detailed film effectiveness distributions for cases 2, 3, and 4.
Effect on suction surface.For case 2, film effectiveness im- mediately downstream of leading edge holes is as high as 0.5 but drops rapidly.The coolant protection dissipates rapidly in this high-curvature region as the jets do not adhere to the surface.High effectiveness is observed downstream of row S along the holes.The high film effectiveness streaks extend up to the next hole row $2.However, the film effectiveness between the holes is lower due to lack of spanwise mixing of jets.Effectiveness downstream of injection from row $2 shows shorter streaks with the jets coalescing downstream.The high curvature of the blade and the boundary layer transition to turbulence in this region (Fig. 4) may be the reason for the shorter streaks.As the blow- ing ratio increases from M 0.4 to M 1.2, film effectiveness downstream of each injection hole row increases.Effectiveness is higher downstream of leading edge hole rows with short jet streaks.The effectiveness is significantly higher downstream of row S along the hole centerline.The streaks of high effective- ness are stronger and appear to mix with downstream row $2. Downstream of $2 also, the effectiveness is higher.As blow- ing ratio increases, more coolant is injected into the mainstream providing more protection to the surface.The effectiveness is as high as 0.2 at about X/SL 0.6 for M 1.2.
Effect on pressure surface.Effectiveness distributions on the pressure surface do not show as strong jet-like streaks as on the suction surface.Effectiveness levels are also not very high downstream of injection holes.As the blowing ratio increases, film effectiveness in the entire injection region increases due to increased amounts of coolant injected into the boundary layer.
There is lesser interaction between coolant and mainstream on the pressure surface due to the thicker boundary layer.The effect is significant downstream of LE holes.Effectiveness also increases around injection hole rows P1 and P2 with values up to 0.4.Higher blowing ratio for CO2 injection produces higher effectiveness on the blade surface.
Figure 7 presents the span-averaged film effectiveness distri- butions for cases 2-4.Effectiveness increases with an increase in the blowing ratio on the suction surface.Effectiveness drops rapidly downstream of leading edge holes on the suction sur- face, then increases immediately downstream of hole row S1, drops again, and increases immediately downstream of hole row X/PL X/SL FIGURE 7 Effect of blowing ratio on span-averaged film effectiveness distributions for CO2 injection.
$2 and then decreases gradually toward the trailing edge for all three blowing ratios.High effectiveness is obtained downstream of each row of holes.However, the effect of coolant injection decreases rapidly further downstream of each row.Results from Mehendale et al. (1994) for the same coolant and flow condi- tions are also presented at M 0.8.The data are comparable away from the injection holes.However, in the injection region, there are differences in the measured effectiveness levels between the studies.The reason for this could be similar to the reason explained for heat transfer coefficient data. 2-4,5-6).Air injection simulates a density ratio of 1.0 whereas CO2 injection simulates a density ratio of 1.5.Film injection causes earlier boundary layer transition on the suction surface and enhances the Nusselt numbers over the entire surface, as in- dicated earlier.Effects of coolant density are limited to regions downstream of injection holes.Higher density coolant produces higher Nusselt numbers downstream of injection at the same blowing ratio.This effect is stronger on the suction surface.On the pressure surface, the density ratio effect vanishes immedi- ately downstream of injection.Change in coolant density ratio has only a small effect on the already high Nusselt numbers produced by film injection.

E.ff'ect of Coolant Density
Figure 9 presents the span-averaged film effectiveness distri- butions for the same cases as for Fig. 8. Film effectiveness on the suction surface is higher for CO2 injection at M 1.2.At a lower blowing ratio ofM 0.8, air provides higher effectiveness than CO2 injection.At low blowing ratio of 0.8, air has higher momentum (I 0.64) compared to CO2 (I 0.42) and protects the surface better.Higher momentum injection does not diffuse into the mainstream as rapidly as that for lower momentum in- jection and thus provides increased effectiveness.At M 1.2, air possesses very high momentum (I 1.44) and coolant jets blow into the mainstream penetrating the boundary layer and do not provide good protection compared to CO2 injection (I  0.96).Overall, CO2 injection at M 1.2 provides highest ef- fectiveness downstream of injection.But in the injection hole region, it is difficult to distinguish the blowing ratio and density ratio effects.
Figure 10 presents the effect of coolant-to-mainstream mo- mentum flux ratio (I) on span-averaged Nusselt number ratios and film effectiveness values at three streamwise locations.All the locations shown in the figure are in the region between hole rows S and $2.The hole row S is at X/SL 1.2 and hole row FIGURE 10 Effect of coolant-to-mainstream momentum flux ratio (I) on spanwise averaged Nusselt number ratio (Nu/Nuo) and film effectiveness values at three axial locations between hole rows S1 and $2.
$2 is at X/SL 0.28.The overall averaged values for the en- tire pressure or suction surface distributions are not considered here due to the presence of multi-row injections.This figure is an attempt to capture the behavior of the heat transfer co- efficient and film effectiveness under varying film cooling flow conditions.The Nusselt number ratio is obtained by dividing the film-cooled blade Nusselt number with a no-holes blade Nusselt number.Results show that Nusselt number ratio increases with increasing I values (with highest values at I 1.0) and then be- gin to decrease.It is difficult to make any conclusions for high I values as this study only investigates one I value over 1.0.It appears that the location effect is significant for lower I values, with the location closest to the hole providing lowest Nusselt number ratio.The coolant at lower momentum flux ratios tends to stay closer to the surface after injection and mixes less with the mainstream, providing lower heat transfer coefficients.However, farther away from the hole, the mainstream breaks down the jet structures and thus causes higher heat transfer coefficients.
Blade curvature effect may also be significant at lower I values.
For higher I values, the location's effect on Nusselt number ratio is reduced.Film effectiveness values show considerably lower effect of location from the hole.Highest effectiveness occurs around I 0.5.Effectiveness decreases at larger I values due to jet lift-off.
CONCLUSIONS kair Detailed Nusselt number and film effectiveness distributions L on a turbine blade with film cooling were presented under the M effects of coolant blowing ratio and density ratios.Results were obtained at a cascade exit Reynolds number of Nu 5.3 x 105.Film injection was provided through three rows on Nu the leading edge and two rows each on the pressure and suction P surfaces.PL 1. Detailed Nusselt number and film effectiveness distributions using the transient liquid crystal technique were obtained on SL the entire blade midspan region.The strong spanwise and axial variations due to film injection are clearly evident in Tc the detailed distributions that also provide valuable insight Ti into the film cooling process.Tm 2. Film injection causes earlier laminar-to-turbulent boundary layer transition on the suction surface.Film injection also in-V creases spanwise and axial variations in Nusselt number dis- tributions on the blade mid-span.Nusselt numbers on the suc-V tion surface are significantly enhanced for a blade with film V2 injection compared to a blade without film holes.Pressure surface Nusselt numbers are also enhanced with film injec-VR tion, but are not as significantly high as on the suction surface.X 3. Higher film effectiveness values are observed downstream of injection holes on the suction surface.However, such high ef- fectiveness values do not extend far downstream of injection on the pressure surface, v 4. Nusselt numbers increase with an increase in blowing ratio, particularly in the region immediately downstream of holes, tom Film effectiveness significantly increases with an increase in blowing ratio for CO2 injection.Air produces higher effec- tiveness at a blowing ratio of 0.8.
5. An increase in coolant density has little effect except in the regions immediately downstream of injection, where CO2 in- jection provides higher Nusselt numbers than air injection.
Higher density coolant (CO2) produces higher effectiveness at higher blowing ratios (M 1.2).However, lower density coolant (air) provides higher effectiveness at lower blowing ratios (M 0.8).
The local heat transfer coefficient and film effectiveness re- sults can be combined to obtain the overall film cooling performance for each case.Such information was provided in an earlier study by Mehendale et al. (1994).thermal conductivity of mainstream air length of film injection hole coolant-to-mainstream mass flux ratio or blowing ratio, pcVc/pmV local Nusselt number based on axial chord, hCx/kair span-averaged Nusselt number film hole pitch streamwise length on the pressure surface (25.6 cm) Re Reynolds number based on exit velocity and axial chord, VzCx/v streamwise length on the suction surface (33.1 cm) liquid crystal color change time coolant temperature initial temperature of blade surface mainstream temperature liquid crystal color change temperature from green to red local mainstream velocity along the blade pressure or suc- tion surface cascade inlet velocity cascade exit velocity coolant hole exit velocity coolant-to-mainstream velocity ratio, Vc/V streamwise distance from blade leading edge thermal diffusivity of blade material (0.135 x 10 -6 mZ/s) local film effectiveness span-averaged film effectiveness kinematic viscosity of mainstream air coolant density mainstream flow density

FIGURE 3
FIGURE 3Local-to-exit velocity ratio (V/V2) distributions on the test blade surface.
FIGURE 4 FIGURE 5 Effect of blowing ratio on span-averaged Nusselt number distributions for CO2 injection.
FIGURE 6Effect of blowing ratio on detailed film effectiveness distributions for CO2 injection.

Figure 8
Figure8presents effect of coolant density on span-averaged Nusselt number distributions for M 0.8 and M 1.2 (cases -mainstream momentum flux ratio, Pc V, /[Om V 2, MZ/DR thermal conductivity of blade material (0.159 W/m-C) $2