Inambao since October 2007 and is part of an ongoing study in turbine blade heat transfer and aerodynamic analysis. Turbulence intensity generation grids were implemented to measure the effects of turbulence on blade heat transfer.
OVERVIEW
EXPERIMENTAL FACILITIES
Since only a limited amount of gas expands through the nozzle, stagnation conditions do not last long and the duration of the test is relatively short. A disadvantage of light piston compression tUIUlels is the transient establishment of the flow through the test section.
HEAT TRANSFER MEASUREMENT TECHNIQUES
THERMOCHROMIC LIQUID CRYSTALS
The color to provide the connection from the blade surface to the wire guides was gold. The method is statistical in nature, meaning that the effects of noise on the data can be reduced by using an iterative surface flux process.
NUMERICAL PREDICTION
Another key technique for modeling flow field and heat transfer characteristics in CFD is the use of the full suite of Navier-Stokes solvers, which are able to solve the main flow field as well as apply special fits to the near region. wall (boundary layer). FLUENT is capable of solving inviscid and Navier-Stokes equations, as well as RANS and LES approaches including the deployment of various turbulence models.
OVERVIEW
HEAT TRANSFER THEORY
It can also be assumed that the thickness of the TFG has a negligible effect on the thermal conduction. By using the analog printed circuit boards, the heat flux calculations are noise-reducing, as discussed earlier.
Calibration Parabola
Although this term contains the periodic, it is still constant, since the ratio ~ is constant with a parabolic signal. To calibrate the circuits, a parabolic signal is fed into the system, generating a step output voltage.
Calibration Step Response
OVERVIEW
Currently at UKZN, the facility for obtaining heat transfer research is a continuously running supersonic cascade device. Van der Steege (1990) initially designed and used the test facility to assess the performance of turbine blades with cooling.
HIGH SPEED RIG
- COOLING Box
The rear of the plenum provides access to the mechanical carbon seal, allowing the compressor shaft to pass into the plenum. By varying the operating pressure in the plenum, the Reynolds number can be varied based on the chord length. The loading valves on the pressure and suction sides of the blade are fixed, although Snedden (1995) allowed the loading valves to be adjustable.
The cooling circuit begins where air is drawn from the bottom drain of the plenum through a heat exchanger. A schematic diagram of the arrangement of the pneumatic cylinder and slide mechanism can be seen in Figure 4-7.
STATIC PRESSURE MEASUREMENT
- TEST BLADE
The data collection of the pressure measurements involves the use of the scanivalve and pressure transducers. A rotor in the scan valve steps between each measurement, connecting a common port on the wafers to each of the 24 ports on the stators. The remaining open ports on the scan valve were used to read the static pressure at the inlet and outlet of the cascade.
This is connected to the HP side of the Rosemount transducer and the total inlet pressure is monitored. The LabView software program allowed control of the stepper motor used for the scanner, along with reading the voltages from the pressure transducers and monitoring the position of the scanner rotor.
OVERVIEW
The heat transfer test blade needed restoration as there were an unsatisfactory number of broken sensors on the blade. With the limited resolution of sensors already on the blade, most of them had to work in such a way that an accurate analysis of the heat transfer coefficient distribution could be performed. By repainting the sensors on the blade, it was believed to have increased sensitivity during temperature changes, so the output would be clearer.
The test blade was repainted with platinum thin film thermometers, along with the gold paint connections, giving full resolution on the blade and increased sensitivity to temperature changes. Turbulence intensity generating grids were also used in tests to validate results obtained by Stieger (1998) as no related tests have been performed on the rig since then.
STATIC PRESSURE TESTS
A secondary reason for restoring the sensors was that the data acquisition system was generating too much electrical noise, an amount that could affect the results because sensitivity was relatively poor in this regard. For the Rosemount transducers, the total inlet pressure tap from the plenum was connected to the low pressure (LP) side, and the high pressure (HP) side was left open to atmospheric pressure. To observe the voltage signals, the PC 30 AID card was connected to the first two channels and read out via the WaveView software program.
The previous setting on the card was for a differential voltage output of -5 to +5V and using the Eagle Technologies PC 30 card manual it was found that the required jumper setting changes the output voltage to an absolute range of 0V to 10V While voltage changes were recorded in WaveView, the height of the mercury in the manometer was observed.
Rosemount B - Calibration Curve
The Kiel probe that measured the total cascade inlet pressure was also connected to free ports on one of the scanival wafers. As the compressor blade speed and air flow increased, so did the temperature inside the plenum. This temperature was monitored and adjusted to the desired value of 100 QC using the cooling water flow rate through the annular radiator.
This value varied, along with that of the compressor speed, and both had to be continuously monitored and adjusted throughout the test procedure. Since no turbulence intensity grid was used, the value of the free stream turbulence is 3%, as measured by Snedden (1995).
Static Pressure Distribution
HEAT TRANSFER TESTS
- INITIAL TESTS
The calibration was performed independently of the power source and the knife stack was disconnected from the circuit. Once the blade temperature reached 40°C, instrumentation was set up to record heat transfer data just before it collapsed. The value of the heat transfer coefficient at the leading edge of the blade for Tu= 25% was 532 W/m2K.
The distribution on the suction side of the blade showed a similar trend to that of the Tu = 15%. The largest step of increased heat transfer visible is that of the case for Tu= 25.5%.
OVERVIEW
SOLVERS AND TURBULENCE MODELS
As the advantages and disadvantages of the standard k-c model have become known, improvements have been made to the model to improve its performance in the form of variants of the RNG k-& model and the Realizable k-& model . These features make the RNG k-& model more accurate and reliable for a broader class of flows than the standard k-& model. The feasible k-& model is a relatively recent development and differs from the standard k-& model in two important ways.
The SST k-w model was developed by Menter (1994) to effectively combine the reliable and accurate formulation of the k-w model in the near-wall region with the free-stream independence of the k-& model in the far field. The standard k-w model and the transformed k-& model are both multiplied by a mixture function and both models are added together.
MODELLING AND RESULTS
The conversion of the heat flux to a heat transfer coefficient was carried out in the same way as in the experimental calculation. It can be seen that the general trend of the k-G models followed the experimental distribution with respect to the magnitude of the heat transfer coefficient on the pressure side. There was a large overprediction in the magnitude of the heat transfer coefficient compared to the experimental curve, and it was found to be somewhat constant throughout the distribution.
The Spalart-Allmaras model again accurately predicted the heat transfer coefficient at the leading edge of the blade for both turbulence conditions when compared to the free inlet condition. There was a large over-prediction of the magnitude of heat transfer on the suction side for all turbulence models.
SUMMARY
CONCLUSIONS
However, every gauge was operational and this full resolution would aid in a full understanding of the results generated. The heat transfer distribution results at higher levels of turbulence intensity again compare well with those of Stieger (1998) . The k-B turbulence models of the Realisablek-Band RNGk-B models (including the Spalart-Allmaras model) were found to be accurate in predicting the heat transfer around the leading edge of the blade when compared to experimental data.
Although the trend of the distribution curve did not match well, the magnitude of the heat transfer was favorable, with the Spalart-Allmaras model accurately predicting the heat transfer coefficient at the leading edge. The SST k-w turbulence model was shown to accurately predict the trend of the heat transfer coefficient distribution compared directly to the experimental set, but the model overpredicted the magnitude of heat transfer greatly along the entire blade surface.
RECOMMENDATIONS FOR FUTURE WORK
Short duration measurements of heat transfer rate in a gas turbine blade.ASME Journal ofEngineeringfor Power. Theory of Advanced Multi-Layer Thin Film Heat Transfer Gauges.Journal of Heat Mass Transfer. An experimental study of endwall and airfoil surface heat transfer in a large-scale turbine blade cascade.
1994. Measurement of tip scour pressure distribution and heat transfer in a turbine blade at realistic blade speeds. Transient liquid crystal technique used for sub- and transonic heat transfer and film cooling measurements in a linear cascade.
ApPENDICES
The positions of the thin film gauges on the SMR-95 blade are shown in Table Al-2 and the pressure points in Table Al-3. Connect the cooling pipes from the fan to the corresponding inlets of the cooling box. i) Connect the corresponding hoses of the compressed air line to the pneumatic cylinder. Note that it is vital to repeat the process as adjusting the spacer screw affects the zero setting.. a) Leave the HP side of the Rosemount transducer open to atmosphere and connect the LP side to the total inlet pressure.. b) Combine the total pressure into a V-tube manometer.
While the vacuum pump is starting, record the output signal corresponding to the total pressure of the Kulite, via channel 3 on the PC 71 board with the WaveView software. e). Record the readings of all thermocouples, as well as the oven display and the average of the final temperature. g) Record the resistances of all associated sensors by swapping the plugs and changing the selection box. h) Increase the oven temperature by approximately 20°C and repeat the steps. a) Provide ±15 V power to the heat transfer analog and ensure that the blade stack is not connected, creating an open circuit to the heat transfer analog Cards. Post-processing of the data sampled in the experimental heat transfer tests was performed using a program designed by Snedden (1995) with LabView.
The details and code of Isotemp 1 are given in Appendix 4 of the thesis of Snedden (1995).
