Citation Context

...he geometry for the 15 deg forward expansion, 60 deg CA holes Journal of Turbomachinery OCTOBER 1996, Vol. 1 1 8 / 8 0 9 Downloaded From: https://turbomachinery.asmedigitalcollection.asme.org on 06/29/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use local effectiveness, laterally averaged effectiveness, and spatially averaged effectiveness being Srj = ±0.01, 657 = ±0.01, and Sfj = ±0.01, respectively. Based on the uncertainties for flow conditions and freestream, coolant, and surface temperatures, and the uncertainties due to frost effects, the sequential perturbation technique (Moffat, 1988) was used to estimate the resulting uncertainties for adiabatic effectiveness values. We also obtained a direct measure of the precision uncertainties for adiabatic effectiveness by performing a number of repeatability tests. Most of these were performed with CA = 0 deg round holes for which six sets of measurements were made in four separate experiments conducted over a span of three years. The momentum flux ratio for these tests was / = 0.16, and similar repeatability tests were conducted with / = 0.63 and / = 1.4. Statistical analysis of the variation of the local values of r\ gave an uncer...

Citation Context

...a similar technique to determine heat transfer results for a film cooling situation. Three narrow colour play liquid crystals with calibrated temperatures of 26.1°C, 30.6°C and 35.1°C were used in the current series of experiments. At the locations of the highest heat transfer coefficient, directly under the jet, the second and third crystals were used in the analysis whilst at regions of lower heat transfer, the first and second crystals were used. The 12mm perspex wall was thick enough to ensure that the semi-infinite assumption could be used in all data analysis. The perturbation method of Moffat (1988) was used to calculate the experimental uncertainty in h and effectiveness. Ireland and Jones (1987) showed that the double crystal technique is sensitive to measured uncertainties. An extensive rig validation exercise together with careful selection of plenum and crystal temperatures enabled an experimental uncertainty below 10% to be achieved. Superposition can be used to show that the local heat flux at the target plate can be described in terms of two heat transfer coefficients. q = h(7, -T)+hh(T, -T) The analysis is very similar to that detailed in Lucas et al. (1992). The double crystal ...

Citation Context

... +12% and was largest with the narrowest gap experiments due to their relatively high heat transfer coefficients (up to 964 w/m 2-°K). The heat transfer level (transient time) is directly related to the experimental uncertainty, as described by Hticker (1996) who investigated the optimization of the transient TLC technique on error estimation in detail. Considering the air supply system stability and pressure transducer accuracy, the uncertainty in friction factors was estimated to be within ±10%. Uncertainties were calculated using the single sample methods of Kline and McClintock (1953) and Moffat (1988). RESULTS AND DISCUSSION Figure 3 shows a typical air temperature profile within the channel gap at the middle of the twelfth and thirteenth rows of concavities. As expected, the air temperature is skewed toward the concavity side due to its high heat transfer rate. A typical Nusselt number distribution, which was reduced based on a local mixed mean temperature, is plotted in a projected view of the CCD camera as shown in Figure 4. In order to help the understanding of the relative position, the concavities were overlaid on the same plot. It is clear that the most of heat transfer enhancement ...

Citation Context

...resented by 255 by 206 pixels. The input heat flux was corrected for radiation losses, which amounted to between 7-23 % of the input power, and conduction losses, which amounted to 1.7 -3.5 % of the input power. No correction was necessary regarding heat losses from conduction to the turbine vane itself because the vane was constructed using Styrofoam. Using the measured temperatures and the remaining convective heat flux, the heat transfer coefficients were computed and reported as Stanton numbers. UNCERTAINTY ESTIMATES The partial derivative and sequential perturbation methods, described by Moffat (1988), were used to estimate the uncertainties of the measured values. Uncertainties were calculated based on a 95% confidence interval. For each velocity component 15,000 data points were used to compute the mean and turbulence quantities whereas when coincidence data was acquired 20,000 data points were acquired. The estimate of bias and precision uncertainties for the mean velocities were 1% while the precision of the rms velocities was 2.1% for u rms , 1.7% for the v rms , andCopyright (C) 2000 by ASME : http://www.asme.org/about-asme/terms-of-use Downlo3.2% for w rms . The precision uncertaint...

Citation Context

...nd static pressures correctly, independent total and static pressures were compared to those made using a pitot-static probe. The differential pressure measurements for each of the supply chambers were read from calibrated transducers with a maximum range of 12.7 mm H2O. As mentioned previously, the dilution flowrates were set by measuring the velocities at the exit of the dilution holes with a pitot-static probe. The pressure difference across the slot was measured using a transducer having a range of 127 mm H2O. The partial derivative and sequential perturbation methods, described by Moffat [18], were used to estimate the uncertainties of the measured values. Precision uncertainties were calculated based on a 95% confidence interval. For the streamwise velocities measured using the hot-wire, 100,000 samples were taken giving an uncertainty, including the bias and precision, of 1.8% for the mean velocity at a turbulence level of 14%. The precision uncertainty for the streamwise rms velocities was 2.6% at that same turbulence level. The integral length scales were calculated using eight samples of 80,000 data points taken at a sampling rate of 20,000 kHz giving a precision uncertainty ...

Citation Context

...ions for an exit Reynolds number of 500,000 at low and high turbulence levels also compared well with the solid vane data of Ames (1997). The excellent replication between these two vanes at two turbulence conditions and Reynolds numbers gives confidence in the experimental method. The Journal of Turbomachinery OCTOBER 1998, Vol. 120 / 771 Downloaded From: https://turbomachinery.asmedigitalcollection.asme.org on 06/30/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use estimated uncertainty for the heat transfer coefficient is ±5 percent based on the root mean square deviation (see Moffat, 1988). Three separate tests were made to determine heat transfer with film cooling and adiabatic effectiveness. First, recovery wall temperatures were determined for the given condition without any heating or coolant injection. Second, a test was made for surface heat transfer with injection at a particular velocity ratio using the heated foil technique. Finally, a test for effectiveness was made with the foil heater off by heating the coolant and recording the vane and plenum temperatures. Initially, the procedure described for determining the base vane heat transfer was used to estimate the heat ...

Citation Context

... = 1.5 V H/B = 2.25 A H/B = 3.0 O H/B = 4.6 i i i i i i 1 - - _ - " 10 V. , (m/s) Fig. 7 Fully developed adiabatic heat transfer coefficient versus inlet velocity, parametric in channel height nr — T 1 ad,ij ~ •* meas,ij (gco#irf + Qrad + Qlw)i,j ™ad, iJ-A surf,i J (10) Uncertainty in Experimental Results. An Mh order uncertainty analysis was performed to determine the uncertainty ranges on the two main calculated quantities: the heat transfer coefficient and the kernel function. We used the standard single sample uncertainty analysis recommended by Kline and McClintock (1953) and extended by Moffat (1988). For 20:1 odds the uncertainty in the adiabatic heat transfer coefficient ranged from 3 percent at the smallest values of h to about 1 percent at the largest. The uncertainty in the superposition kernel function ranged from 11 percent at its low values to 5 percent at its high values. Experimental Results Data was taken for channel to module height ratios, H/B, of 1.5, 2.25, 3.0, and 4.6 for inlet velocities of V = 3.0, 5.6, 7.1 and 8.1 m/s on all eight rows of the array. Inlet velocities 5 - - — 1 '- ~7 - : i I 1 i s 1 I i 1 , 1 I ' V = 5.6 m/sec H/B = 2.25 • 5 , 1 \ - -_ -_ Z_ — \ Number of...

Citation Context

...tios are reported based on the inlet velocity, Uin, ininh c UA mM ρ =∞ (1) However, for each row of fan-shaped holes, blowing ratios are reported in terms of the local surface velocity, Ulocal, inlocalh c UA m M ρ = (2) Five blowing ratios were tested for the showerhead region, four blowing ratios were tested for each row on the pressure side, and three blowing ratios were tested for each row on the suction side. The range of blowing ratios was chosed to span typical engine operating conditions. Experimental Uncertainty The partial derivative and sequential perturbation method given by Moffat [17] were used to calculate uncertainties for the measured values. For a high reported value of St=0.0093 the uncertainty was ±3.23%, while the uncertainty for a low value of St=0.0023 was ±2.13%. The uncertainties for the adiabatic effectiveness measurements were ±0.012 for a high value of ηAW=0.9 and ±0.011 for a low value of ηAW=0.2. EXPERIMENTAL RESULTS Heat transfer results will be discussed first followed by adiabatic effectiveness results and a comparison to existing data from literature. Heat Transfer Results Heat transfer results are reported in terms of the Stanton number based on inlet ...

Citation Context

...assumption that the heat transfer coefficient is uniform over the entire surface of the heat sink/package. The fin efficiency, r\j, is a function of the fin geometry, its thermal conductivity, ks, and hs, t)f=fct(ks, ~hs, geometry) (5) The explicit form of Eq. (5) depends on the fin-tip boundary conditions employed and is given in most general references on heat transfer. Measurement of q and (Ts- T0), and simultaneous solution of Eqs. (3)-(5) gives hs, T/S, and -qf. Note, for bare packages, Af=0 and rjs= 1 in Eqs. (3) and (4). The analysis of error propagation is similar to that suggested by Moffat (1988). Estimates of the 95 percent-confidence level expected error of the measured quantities, and control sensitivity in data acquisition and conditioning, are used to generate sets of pseudo-data having random fluctuations about the nominal measurements. These are input into the data reduction algorithm, and the variance of the output about nominal results (for example: heat transfer coefficients, Reynolds numbers and heat sink efficiencies) is then used to establish the expected random error in the results. This is doubled in the reported estimate of total expected error because it is assumed th...