X-33 Hypersonic Boundary-Layer Transition
Summary (4 min read)
Introduction
- The Access to Space Study 1 by NASA recommended the development of a heavy-lift fully reusable launch vehicle (RLV) 2,3 to provide a next-generation launch capability to serve National space transportation needs at greatly reduced cost.
- The goal of the RLV technology program is to enable significant reductions in the cost of access to space, and to promote the creation and delivery of new space services and other activities that will improve U.S. economic competitiveness.
- As part of the Single-Stage-To-Orbit (SSTO) RLV program, the X-33 was developed as a technology demonstrator.
- The purpose of this investigation was to experimentally examine issues affecting boundary layer transition and the effect of transition on the aeroheating characteristics of the X-33.
- The primary test technique that was utilized during these tests was the thermographic phosphor technique 9 , which provides global surface heating images that can be used to assess the state of the boundary layer.
Boundary Layer Transition in Flight
- From the perspective of boundary layer transition, the X-33 has many similarities to the Space Shuttle Orbiter.
- For a majority of these flights, boundary layer transition has been dominated by surface roughness (Ref. 10), in the form of launchinduced damage and/or protruding gap fillers.
- Early in the Shuttle program the ceramic TPS was recognized to be relatively fragile and studies were per-Figure 2 X-33 windward surface TPS.
- The sub-orbital trajectory used for the design of the TPS is shown in Fig. 3 and is designated as Old Malmstrom-4.
- The initial flights of the X-33 are lower Mach number trajectories to Michaels Air Force Base in Utah (preliminary Michaels trajectories are also shown for comparison).
Transition Prediction Approach
- A series of wind tunnel tests (see Table 1 ) were performed to investigate the X-33 aeroheating and boundary layer characteristics, while tracking changes to the configuration.
- The evolution of the X-33 configuration from the onset of Phase II has necessitated multiple entries into LaRC facilities to investigate the effects of outer mold line (OML) changes to X-33's aeroheating environment.
- The D-loft configuration emerged from the end of the Phase I competition and was heavily tested in the early part of Phase II.
- The Rev-F has the same forebody shape as Rev-C, but the dihedral of the canted fins was lowered from 37-deg to 20-deg (to improve pitch-trim characteristics across the speed range) and the size of the body flaps and vertical tails was increased.
- Then during test 6763, the effect of discrete roughness on the centerline of the Rev-F configuration was investigated.
Test Facility
- The present experiments were conducted in the LaRC 20-Inch MachÊ6 Air Tunnel.
- Miller (Ref. 17) provides a detailed description of this hypersonic blowdown facility, which uses heated, dried, and filtered air as the test gas.
- Typical operating conditions for the tunnel are stagnation pressures ranging from 30 to 500 psia, stagnation temperatures from 760 to 940-degR, and freestream unit Reynolds numbers from 0.5 to 8 million per foot.
- A bottom-mounted model injection system can insert models from a sheltered position to the tunnel centerline in less than 0.5-sec.
- Run times up to 15 minutes are possible with this facility, although for the current heat transfer and flow visualization tests, the model was exposed to the flow for only a few seconds.
Test Techniques
- The rapid advances in image processing technology which have occurred in recent years have made digital optical measurement techniques practical in the wind tunnel.
- Details of the phosphor thermography technique are provided in Refs. 9, 18, and 19, while Refs. 6, 14, 20, and 21 are recent examples of the application of the technique to wind tunnel testing.
- By acquiring fluorescence intensity images with a color video camera of an illuminated phosphor model exposed to flow in a wind tunnel, surface temperature mappings can be calculated on the portions of the model that are in the field of view of the camera.
- The LaRC 20-Inch Mach 6 Air Tunnel is equipped with a pulsed white-light, Z-pattern, single-pass schlieren system with a field of view encompassing the entire 20-in test core.
- Surface streamline patterns were obtained using the oil-flow technique.
Model Description
- Additional details about the various model configurations that were tested can be found in Ref. 16 .
- To meet these requirements, a unique, silica ceramic investment slip casting method has been developed and patented (Ref. 24) .
- The fiducial marks used for the present study are shown in Fig. 6 and the non-dimensional locations are listed in Table 2 .
- The roughness elements used in this study were similar to those used in Refs. 6 and 14, which were fabricated to simulate a raised Thermal Protection System (TPS) tile and were cut from 0.0025-inch thick Kapton tape.
Test Conditions
- The LaRC 20-Inch Mach 6 Air Tunnel provides a freestream unit Reynolds number variation of 0.5 to 8.0 million per foot.
- For a 0.0132-scale model, this corresponds to a length Reynolds number of approximately 0.41 to 6.7 million.
- For the baseline data, the model angle of attack (a) was varied from 20-deg to 40-deg in 10-deg increments and the sideslip was maintained at zero for all the runs presented herein.
- For each model configuration, the unit Reynolds number was varied between 1 and 8 million per foot to obtain the smooth baseline data for comparison to the tripped data.
- The Reynolds number where significant nonlaminar flow first appears downstream of the roughness element identifies the critical value.
Data Reduction
- Heating rates were calculated from the global surface temperature measurements using one-dimensional semiinfinite solid heat-conduction equations, as discussed in detail in Refs. 9 and 19.
- Based on considerations pre-sented in Ref. 9, phosphor system measurement error is believed to be better than ±8%, with overall experimental uncertainty of ±15%.
- Heating distributions are presented in terms of the ratio of heat-transfer coefficient h/h FR , where h FR corresponds to the Fay and Riddell 26 stagnation-point heating to a sphere with radius 0.629-in (the nose radius of the Rev-F configuration scaled to the model size).
- Repeatability of the centerline heat transfer distributions was found to be generally better than ±4%.
Computational Methods
- Computational predictions for comparison to the wind tunnel aeroheating test results were generated at select angles-of-attack and test conditions using the General Aerodynamic Simulation Program (GASP) code, 27 and the Langley Approximate Three-Dimensional Heating Analysis code.
- 28 The GASP computations were used to assess the state of the boundary layer for the wind tunnel cases, while the LATCH results were used to generate the boundary layer transition correlation parameters.
- GASP is a threeÐdimensional, finitevolume Navier-Stokes solver that incorporates numerous options for flux-splitting methods, thermochemical and turbulence models, and time-integration schemes.
- A perfect gas air model was employed and both fully laminar and fully turbulent solutions were obtained.
- An integral heating method is used to compute the heating rates along three-dimensional inviscid streamlines.
Smooth Body
- As will be discussed in more detail subsequently, the increased heating region towards the aft-end of the model identifies the onset of boundary layer transition.
- Note the changing shape of the transition front as a increases.
- At the lower angles of attack the forebody generates surface streamlines which curve predominately in towards the centerline .
- The effect of varying Reynolds number on extracted heat transfer profiles along the model centerline for each a is shown in Fig. 10 with comparison of the smooth body results to tripped cases and to laminar and turbulent heating predictions from Ref. 8.
- The LATCH results do not compare as favorably along the centerline; however, as shown in Ref. 8, this disparity is only along the centerline as off-centerline the results compare within the experimental uncertainty.
Discrete Roughness along Model Centerline
- As the turbulent wedge does not start immediately behind the trip, this case would be classified as Òcritical.
- From the sample results presented in Fig. 11 , one might be tempted to conclude that the X-33 vehicle would be more sensitive to discrete trips at higher angles of attack.
- Figure 13 illustrates the function in relation to the TPS-design trajectory.
- Using this value of k/d and the calculated boundary layer thickness at the point on the trajectory that corresponds to the onset of transition, the allowable roughness heights over the windward surface are inferred.
Discrete Roughness Off-Centerline
- Once the roughness criterion was established for the X-33 centerline, the next step was to verify if the discrete centerline criterion was applicable to off-centerline locations as well.
- Once the attachment line locations were verified for each angle of attack, the roughness effects were examined.
- This conclusion is also not supported when the attachment line discrete trip results are compared to the results for the centerline discrete trip correlation (Fig. 12 ).
- Hollis, et al. (Ref. 8) has shown that the LATCH computations are in very good agreement with Navier-Stokes calculations in the chine region where the attachment lines are located.
- This information can then be used as input to a swept cylinder boundary layer code (Ref. 30) to compute the momentum thickness Reynolds number for an ÒequivalentÓ swept cylinder.
Distributed Roughness
- The final transition issue that was investigated for the X-33 was the effect of the bowed metallic TPS panels.
- The series of tests on the various bowed panel configurations has only recently been completed.
- The surface streamlines are seen to serpentine around the various bowed panels.
- The onset of transition results (at x/L = 0.8) for the smooth, discrete, and bowed-panels data for each a was located along these three curves.
Concluding Remarks
- A series of experimental investigations into several issues affecting boundary layer transition on the X-33 vehicle has been performed in the LaRC 20-Inch Mach 6 Tunnel.
- These investigations examined natural transition on a smooth body, transition due to discrete roughness on the centerline and attachment lines, and transition due to distributed roughness in the form of Òwavy-wallÓ bowed panels.
- The experimental heating levels were compared to predictions to determine the onset location of transition and fully turbulent flow.
- The discrete roughness results on centerline were used to provide a transition correlation for the X-33 flight vehicle that was applicable across the range of reentry angles of attack.
- Finally, the effect of bowed panels was qualitatively shown to be less effective than the discrete trips, however, the distributed nature of the bowed panels affected a larger percent of the aft-body than a single discrete trip.
Did you find this useful? Give us your feedback
Citations
257 citations
224 citations
210 citations
Cites background from "X-33 Hypersonic Boundary-Layer Tran..."
...And one article by Berry et al. (2001b) related to objective 5: “Obtain data that can be used to develop empirical correlations for phenomena that resist analytical and/or numerical modeling, such as boundary-layer transition and turbulence modeling.”...
[...]
203 citations
186 citations
References
1,422 citations
203 citations
186 citations
139 citations
118 citations
Related Papers (5)
Frequently Asked Questions (17)
Q2. What is the way to obtain accurate heat transfer data?
In order to obtain accurate heat transfer data using the one-dimensional heat conduction equation, models need to be made of a material with low thermal diffusivity and well-defined, uniform, isotropic thermal properties.
Q3. How can the velocity thickness be calculated?
By iterating on this velocity gradient, the momentum thickness Reynolds number from the swept cylinder calculation can be matched to the approximate three-dimensional momentum thickness Reynolds computed by LATCH.
Q4. How long did the model have to be exposed to the flow?
Run times up to 15 minutes are possible with this facility, although for the current heat transfer and flow visualization tests, the model was exposed to the flow for only a few seconds.
Q5. What was the distribution of the roughness?
The distributed roughness was in the form of a wavywall that simulates the expected metallic TPS Panel bowing in flight due to temperature gradients across the panel.
Q6. What is the effect of the thermal gradients within the metallic TPS panels?
During a hypersonic entry, thermal gradients within the metallic TPS panels will produce an outward bowing of the panels on the order of 0.25-in.
Q7. How many layers of Kapton tape were used to obtain the roughness heights?
Variations on the roughness heights (k) were obtained by stacking multiple layers of Kapton tape (k = 0.0025, 0.0050, and 0.0075-inch).
Q8. What is the common form of surface roughness in the Shuttle?
For a majority of these flights, boundary layer transition has been dominated by surface roughness (Ref. 10), in the form of launchinduced damage and/or protruding gap fillers.
Q9. What is the main advantage of phosphor thermography?
Phosphor thermography is routinely used in Langley's hypersonic facilities as quantitative global surface heating information is obtained from models that can be fabricated quickly (within a few weeks) and economically (cost an order of magnitude less than the thin-film technique).
Q10. What was used to predict the location of the attachment lines for the X-33?
The LATCH code (Ref. 5) was used to predict the location of the attachment lines for angles of attack of 20, 30, and 40-deg (with nominal tunnel flow conditions as inputs).
Q11. What are the examples of typical heating images?
Examples of typical heating images, which illustrate flow symmetry, and extracted heating profiles along the attachment line are provided in Figs. 16 and 17.
Q12. Who was critical to the successful completion of this work?
The following individuals were critical to the successful completion of this work: Mark Cagle, Joe Powers, Mark Griffin, Mike Powers, Rhonda Manis, Grace Gleason, Johnny Ellis, Bert Senter, Sheila Wright, Glenn Bittner, Steve Alter, Matt Kowalkowski, Derek Liechty, and Richard Wheless.
Q13. What is the effect of bowed panels on the boundary layer?
Further analysis of the distributed bowed panel results is required to determine if bowed panels in the vicinity of the chine region might influence the crossflow dominated flow field at lower angles of attack.
Q14. What was the effect of the Rev-C configuration on the forebody?
The Rev-C configuration instituted small modifications to the nose shape (to simplify the construction of the metallic TPS panels) and to the base region (in the vicinity of the engine).
Q15. What was the effect of bowed panels on the aft body?
the effect of bowed panels was qualitatively shown to be less effective than the discrete trips, however, the distributed nature of the bowed panels affected a larger percent of the aft-body than a single discrete trip.
Q16. How is the X-33 expected to transition back to a turbulent state?
Based on earlier results published in Ref. 6, the X-33 boundary layer is expected to transition back to a turbulent state at a Mach number near 9.
Q17. How many units of kapton tape were used to obtain the smooth baseline data?
For each model configuration, the unit Reynolds number was varied between 1 and 8 million per foot to obtain the smooth baseline data for comparison to the tripped data.