Task 7 CFD simulation of baseline hot streakvane interaction

A number of computational simulations have been done during this reporting period, which include the simulation of the temperature profile that was measured downstream of the heater bank, a slot cooled endwall configuration, and a film-cooled endwall configuration. These computations were completed using the commercial software by FLUENT Inc. The passage was modeled to midspan and a symmetry condition was applied at midspan. Also, periodic conditions were applied in the stagnation plane and in the trailing plane. Table 1 provides the status of the computations being performed as well as those being planned in the near-term future.

Figures 4a-c show the effectiveness predictions for three different cases illustrating the effects of the coolant flow and of a non-uniform, spanwise (radially) varying temperature profile. Note that these computations were done using the two-dimensional upstream slot configuration. The temperature profile that was placed at the inlet to the turbine section for the predictions shown in Figure 4c was parabolic in the spanwise direction and uniform in the pitchwise (circumferential) direction. This is the same temperature profile that we are simulating in the wind tunnel and was shown in our October 2001 report. It is quite clear from Figure 4 that there is a warm ring around the turbine vane for all of the cases indicating the coolant flow does not exit the slot uniformly. In addition, most of the coolant is swept toward the suction side of the vane. These results do indicate, however, that a flush slot is more effective at cooling the downstream endwall than the upstream film-cooling holes as was discussed for our experimental results.

In comparing Figure 4a and 4b, it is clear that there is an added benefit to doubling the slot flow given the cooler temperatures predicted on the endwall. Although this benefit does exist, the warm ring around the vane still occurs for the doubled slot flow case. Figure 4c indicates the effect of the parabolic temperature profile on the endwall effectiveness levels for the 1% slot flow case. Note that the endwall temperature is much cooler due to the cooler temperatures of the parabolic profile as illustrated in Figure 5. In general the same effectiveness pattern exists in Figure 4a as compared with Figure 4c except there are still relatively cooler effectiveness levels for the profile case in Figure 4c. These cooler temperatures are a result of cooler fluid being present in the near-wall region that is diluting the slot flow as compared to the flat temperature profile results.

Figures 6-9 show the streamline patterns in the near wall region for a number of different cases. Note that for each of these figures the streamlines were seeded over the exact same region and in the exact same location to allow direct comparisons to be made. The pathlines are colored by the non-dimensional spanwise velocity magnitude such that red indicates a flow away from the endwall and blue indicates a flow towards the endwall.

Figure 6 represents the baseline case with no slot and a uniform temperature profile for comparison purposes. As can be seen from the streamlines in case 6, there is a leading edge and passage vortex that develops through the passage. In comparing Figures 6 and 7, the first noticeable effect of the slot flow is the extraneous streamlines shown in mid-passage in Figure 7. These streamlines originate from the main gas path, are ingested into the two-dimensional slot, and then exit from the slot back into the main gas path. This was also seen for the case with the flat temperature profile (Figure 8). Although the ingestion was not severe, it does provide less cooling capability for the slot flow due to the mix-out with the main gas path fluid. For the doubled coolant flow case shown in Figure 9 there appears to be no ingestion into the slot.

All of the cases shown in Figures 6-9 indicate the presence of secondary flows through the passage. Two effects were thought to minimize the secondary flows in the combustor-turbine junction. First, the upstream contraction has a tendency to flatten out the boundary layer along the approaching endwall, which would also flatten out the total pressure profile. Second, the slot flow present energizes the boundary layer and would also tend to give a flatter total pressure profile at the inlet to the turbine vane. More analysis of the computational predictions needs to be done to determine these various effects on the leading edge vortex. From the pathlines indicated in Figures 6-9 there is an indication, however, that there is an effect on the passage vortex. Figure 8 indicates the least severe passage vortex pattern as compared with the other cases. Although further analysis needs to be done on the CFD results, we believe the reason for the smaller passage vortex is because of the two reasons cited above. Without the blowing, Figure 6 indicates that the contraction is not enough to flatten out the profile. With too much blowing the boundary layer may be energized too much such that a strong peak is formed in the near-wall region (as was shown by the peaks in the film-cooling flow data in Figure 2b). Further work needs to be done to fully understand these effects on the passage vortex.

For the 2% slot flow there appears to be stronger secondary flows than in any of the other cases. At the leading edge, there appears to be a strong flow away from the endwall, which is caused by the stronger injection from the slot. This effect is also indicated in the adiabatic effectiveness contour given in Figure 4b in which some coolant is detected at the leading edge in comparison with Figure 4a in which there is no coolant detected at the leading edge. Further into the passage, the stronger secondary flow patterns are illustrated by the fact that the streamlines have moved further into the passage as can be seen by comparing Figures 8 and 9 for the 1% and 2% injection, respectively. The effectiveness contours for the 1% and 2% coolant flow cases (Figures 4a and 4b) indicate that even though more coolant is present for the latter case, it is still quickly swept off the endwall by the strong secondary flows.

Figures 10a-b show the effect of film-cooling the endwall. Note that these are preliminary results and that further grid refinements still need to be evaluated. As can be seen from these contours, there are hot streaks near the leading edge of the vane which are due to a lack of cooling in this region. As can be seen in Figure 10b, some of the coolant is also convecting up the suction side of the vane. Figures 10c-d show the streamline pattern in which the streamlines were seeded at the same location and over the same region as those in Figures 69. Note that Figures 10c-d should be directly compared with Figure 8, which includes 1% slot cooling and a flat temperature profile. In general, the largest differences that can be seen are that there is a stronger upward migration of the streamlines along the suction surface, as further illustrated in Figure 10d. These streamlines in the near-wall region are being pushed away from the endwall by the film-cooling injection.

0 0

Post a comment