CHAPTER 6 LPV CLOCKING AERO-PERFORMANCE DATA
6.2.2 Summary of Integrated-clocking Effects
This research program did not start off as an attempt to measure the affect of clocking on the overall aeroperformance of the machine. The main goal was to investigate turbine aerodynamics and heat transfer, and the effect of clocking on the localized pressures on the airfoils (which is done in the next sections),. All of these measurements were done primarily for design code verification. However, to form a complete picture of clocking, one has to look at the overall efficiency changes across the machine, which is what is performed in this section. The work done in this section was undertaken to complement the other data analysis and also to see if the improvements made in instrumentation from previous rigs would allow us to discern what is happening at these very small levels of resolution. The improvements in the instrumentation quality have created improvements to the point where aeroperformance variations can be
measured and that the variations due to clocking are consistent between thermodynamic and mechanical measurements of efficiency. In addition, the improvements in the
109
techniques can be used to show what the requirements need to be for the instruments and the experimental set-up to better resolve these variations. If the experiment had
originally been designed as an efficiency experiment, some of these issues would have been addressed earlier.
It appears clear from looking at the speed variations, pressure variations and efficiency plots, that better resolution of small changes in efficiency work better when the individual time windows are used for each group, instead of taking an average time window. The fact that these different methods are close, confirms that the effect of the choke position is small, and future work could probably eliminate this step altogether by either not moving the choke, or providing stops for it. Both the speed variation and the pressure variations show an effect from clocking, but it becomes more easily understood in the context of efficiency. However, the speed changes and the pressure changes can be used as an indicator of the repeatability of the runs.
Improvements in the acceleration measurements are still needed, which translates into higher resolution speed measurements (and measurements without any 1/rev wobble in them). Until then, mechanical measurements over small time windows will be more susceptible to individual run errors, and the thermodynamic measurements will probably be better indicators of the true variation in efficiency that occurs due to clocking. For this situation the thermodynamic efficiency shown in Figure 6.2 probably represents the best measure currently of the effect of clocking on the overall machine efficiency. In this plot there is little change due to Reynolds number and the clocking area around point D and D-E seem to have the greatest change.
Before proceeding with the effects of clocking on the local pressures on the airfoil, it is worthwhile to layout exactly what can be done to improve these types of measurements in the future and to note some of the trade-offs.
1. When measuring clocking, one needs better resolution over the vane passage and over multiple vane passages. It is unclear why position A and E look different, and it was noted that A’ looks like E. Detailed work over this range should probably include at least two passages and
110
perhaps more. The differences between A and E may be real and may be due to passage differences. This would have a great deal of
mitigating effect on the interpretation of what could be achieved with clocking, if a lot of the effect could be dissipated due to hardware variation.
2. As mentioned earlier, the choke really needs to stay in one place to eliminate the effect that small area changes have on the pressure ratio (as seen in appendix D). This implies that vane placement variations should occur before speed variations are done.
3. Full and complete sets at each major condition need to be done if one is tying to compare different sets of data. In this case it was Reynolds number, and to do a better comparison, the higher Reynolds number case should have been done over the entire vane passage range (A-E instead of just C-E.
4. Repeat runs are needed at each condition with strict attention paid to the quality of the speed and pressure ratio data. Probably at least three runs should be done at each point (5 would be better).
5. The overall single most important improvement that can be made in instrumentation that will effect these measurements is the speed.
Current measurements of the speed are about 20 RPM change out of 7000-RPM average (or about 0.286 %). We realistically need to be trying to resolve in a time-resolved sense probably to about 0.2 RPM change to make these efficiency measurements less susceptible to experimental error. In addition, this data needs to be fully time- resolved, instead of having to be averaged over a rotor revolution to reduce manufacturing variations that occur in the encoder