Service Communautaire d'Information sur la Recherche et le Développement - CORDIS

FP7

TRANSHYBERIAN Résumé de rapport

Project ID: 263182
Financé au titre de: FP7-SPACE
Pays: Belgium

Final Report Summary - TRANSHYBERIAN (Characterization of Wall Temperature Effect during Transition of Hypersonic flow over a Cone By Experiments And Numerical Simulations)


Executive Summary:

The development of secure and re-usable re-entry vehicle requires the complete control of the heat distribution on its Thermal Protection System (TPS).

During the most critical re-entry phase, the hypersonic flow along the vehicle initiates a laminar boundary layer inside of which most of the transfer phenomena take place (heat, momentum and mass transfer). If at one position of the vehicle, this boundary layer experiences a transition from the laminar to the turbulent regime then at the corresponding position the TPS will receive a sharp increase of the incoming heat flux (minimum 3 times higher). If the vehicle aims to be re-usable, it is mandatory to protect it adequately against this overheat. Therefore, aerospace designer needs to posses the proper information about the heat transfer phenomenon anticipated around the vehicle and tools allowing a better prediction and ultimately a better control of the transition in hypersonic regime.

This activity proposes a detailed characterization of the influence of local thermal control on boundary layer transition. It is possible, by modifying locally the wall temperature over a vehicle, to implement an active control of the boundary layer mechanisms, thus anticipating or delaying the boundary layer transition onset.

Both an experimental and a numerical investigation have been conducted on a simple conical geometry at different Reynolds numbers and Mach numbers to identify the effects of the thermal control on the boundary layer. Because of the sensitivity of transition studies on the wind tunnel freestream disturbances, a detailed characterization has been performed showing a comparable level of freestream fluctuations in all the facilities involved. The results obtained using the local thermal control confirm the effectiveness of this potential technology able to modify the boundary layer transition behaviour on a vehicle in the high speed regime and thus lower the head load on the re-entry vehicle.

The experiments conducted in the ground test facilities, validated by specific linear stability computations (LST) and direct numerical simulations (DNS), they show that a localized strong cooling at the wall is able to visibly delay the boundary transition onset whereas a strong heating is anticipating it.

The numerical computations show indeed that on one hand because of the local surface cooling the second modes are initially destabilized and subsequently strongly stabilized downstream the cooling region, whereas the first modes are in general not sufficiently destabilized to take part in the transition process. On the other hand a local surface heating strongly destabilizes both first and second modes, as shown in the DNS computations, in a scenario where the first modes critical N-factor rapidly increases approaching the N-factor of the second modes in a competition for the strongest transition mechanism.

Project Context and Objectives:

The development of secure and re-usable re-entry vehicle requires an expert design of its shape that impact on its future stability, maneuverability and distribution of the heat dissipated by the high speed movement in a dense atmosphere. This last parameter is particularly important as the integrity of the Thermal Protection System (TPS) is very often the integrity of the vehicle itself.

Indeed, during the most critical re-entry phase, the hypersonic flow along the vehicle initiates a laminar boundary layer inside of which most of the transfer phenomena take place (heat, momentum and mass transfer). If at one position of the vehicle, this boundary layer experiences a transition from the laminar to the turbulent regime then at the corresponding position the TPS will receive a sharp increase of the incoming heat flux (minimum 3 times higher). If the vehicle aims to be re-usable, it is mandatory to protect it adequately against this overheat. Therefore, aerospace designer needs to anticipate properly the heat transfer phenomenon around the vehicle. This may be achieved by correlations and/or by numerical tools allowing a better prediction and ultimately a better control of the transition in hypersonic regime.

This activity proposes a detailed characterization of the influence of local thermal control on boundary layer transition. It is theoretically possible, by modifying locally the wall temperature over a vehicle, to implement an active control of the boundary layer mechanisms, thus anticipating or delaying the boundary layer transition onset.

Both an experimental and a numerical investigation have been conducted on a simple 7 deg half-cone geometry at different Reynolds numbers and Mach numbers measuring the effects of the thermal control on the boundary layer. Because of the transition sensitivity on the freestream disturbances, a detailed characterization of the wind tunnels from the different partners has been performed. Measurements obtained both with intrusive and non-intrusive techniques show comparable level of freestream fluctuations (generally 1-3% of normalized pressure fluctuations) in all the facilities involved. Therefore the experimental results can be compared among them with a reasonable confidence. This comparison is often desired in the literature but only few attempts had been done in the World or in Europe. It is the first time that this EU-Russia comparison of perturbation in hypersonic facility is achieved.

The experiments performed in different facilities show a strong dominance of second mode disturbance in the transition mechanism over the cone without localized thermal control. The so-called second mode waves are measured along the wind tunnel model with a frequency included between 100 kHz and 500 kHz. At this Mach number, this second mode is known to be predominantly important in the process of transition from a laminar flow to turbulent flow (i.e. this second mode is formed of perturbations at high frequency and they are amplified faster than the perturbations at lower frequencies known as the first mode). When the Reynolds number increases, this second mode is modified with respectively an increase of its frequency and of its amplitude. If the measurement point is located further downstream i.e. closer to the probable transition point, it is noticed that the characteristic frequency associated to the second mode is shifted to the lower frequencies proportionally to the growth of the boundary layer thickness. It is also noticed these measurements done further downstream are also characterized by higher amplitudes. For most of the actors, it was the first time that second mode measurements were succeeded in these conditions thanks to high fidelity pressure transducers.

Using locally, a thermal control at the wall of the cone model, a clear modification of the second mode instabilities has been observed with repeatability in all the facilities. In particular the boundary layer downstream the position of a localized cooling is found thinner compared to the same boundary layer without thermal control. As a consequence, the measured second mode is characterized by higher frequencies. Surprisingly, as the boundary layer passes respectively from a hot wall, to a cold insert and to a hot wall again, the second mode is found globally stabilized. In other words, the measurements of the second mode amplitudes, downstream of the cooling location, show a general decrease whereas the infrared thermography confirms a downstream shift of the natural transition front with respect to the non-cooling experiment.

On the other hand in the case of heating these experimental observations are inverted. The second mode frequencies decrease because of a thicker boundary layer downstream the thermal insert whereas the same second modes are found destabilized by the successive passage from a relatively cold wall to a hot insert before passing to a relatively cold wall again. As a result, an upstream shift of the natural transition location can be clearly measured by the infrared thermography.

For all the actors, it was the first time that second mode sensitivity to local heating was revealed.

Configurations with different tip bluntness were also investigated but a clear conclusion of the combined effect with local thermal control is asking for further investigations.

In the numerical counterpart of the project, different temperature distributions were numerically simulated. Respectively; some baseline flows without heating or cooling, some strong and weak cooling, and some strong and weak heating. Spatial Linear Stability Analysis (LST) was performed accounting for the second mode (2D disturbances) and for the first mode (3D oblique disturbances) using the mean flow provided thanks to axisymmetric Navier-Stokes solvers.

Computations confirmed that the second mode instability is a dominant component of the disturbance field in the boundary layer over the cone for both the baseline and heating/cooling cases. Nevertheless, in the case of strong heating, an interesting competition between the first mode disturbances and the second mode disturbances has been underlined for the first time.

Moreover, it was noticed that the choice of the initial point location for the LST analysis is only slightly influencing the results obtained for the baseline and heating cases, whereas it strongly affects the LST analysis in the case of wall cooling. In particular, if the N factors are computed from the neutral points, the presence of the cold strip leads to an upstream shift of the N-factor envelope resulting in an earlier transition. On the other hand, if the N factors are computed from the fixed point located upstream of the strip, the computations predict a later transition. Results suggest that the location of the cooling strip has a critical influence on the LST computation.

As expected, for higher Reynolds number the instabilities growth even faster and predicted N factors are higher when compared with the lower Reynolds number cases.

To account for the nonparallel effects, which are enhanced by the cooling/heating strip, Direct Numerical Simulations (DNS) are performed for 2D disturbances excited by a suction-blowing slot located upstream of the heating/cooling strip. The DNS results demonstrated that the hot strip leads to an increase of the instability amplitude, whereas the cold strip produces an opposite effect. A comparison of the DNS solutions and the experimentally measured instability growth shows a very good agreement. Moreover, the LST computations of the second-mode amplification starting from the suction-blowing locus agree satisfactory with the DNS solutions.

A number of engineering correlations and CFD RANS transition models were used to predict the effect of local heating/cooling on boundary layer transition under the conditions of the experiments performed.

Most of the collected engineering correlations are built upon a uniform wall temperature and not on a localized temperature difference. Nevertheless, these correlations may be divided onto three groups: (1) totally wall temperature independent, (2) correlations dependent on wall temperature only through the momentum thickness and (3) correlations explicitly dependent on wall temperature. The second group of correlations demonstrates only a minor variation of the transition location with different wall temperature because of the weak dependence of the momentum thickness on Tw. Instead the third group has a more pronounced effect of the transition location in the presence of a wall temperature variation. In general all these correlations (except the one used for lower Mach numbers) well agree with the behavior of boundary layer transition dominated by second mode disturbances and they show a later transition location with wall cooling and an earlier location with wall heating. Among all these correlations, the most adequate results can be obtained with the algebraic ONERA-CERT transition zone model.

Regarding the RANS simulations, the γ-Reθ transition model of Langtry-Menter has been modified to be used in hypersonic regime. This modified model assumes that the momentum thickness is a function of the boundary layer shape factor as well as the Mach number and the wall temperature. The corresponding modifications of the model are proposed and validated during this work. Without modifications the γ-Reθ transition model demonstrated arbitrary and unpredictable behavior when a localized temperature variation was applied. Instead the modified transition model is in agreement with the trends shown by the correlations.

In conclusion, many results obtained within this project are the first of their kind in Europe for the hypersonic community. The experiments conducted in the ground test facilities, validated by specific linear stability computations (LST) and direct numerical simulations (DNS), demonstrate that a localized strong cooling at the wall is able to visibly delay the boundary layer transition onset whereas a strong heating is anticipating it. The implementation of such technology on a gliding or a re-entry vehicle could possibly influence and/or control the boundary layer transition of the vehicle. This interesting non intrusive way of control may have large technological application on stability of the vehicle and on its heat load distribution.

Project Results:

Both an experimental and a numerical investigation have been conducted on a simple conical geometry at different Reynolds numbers and Mach numbers to identify the effects of local thermal control on a hypersonic boundary layer. Because of the sensitivity of transition studies on the wind tunnel freestream disturbances, a detailed characterization has been performed showing a comparable level of freestream fluctuations in all the facilities involved.

Freestream pressure fluctuations (around 1%) have been measured in the ground test facilities using different measurement techniques and data reduction methodologies. This result confirms that in a noisy facility, like the ones involved in the project, the noise in the freestream is generated mainly from the turbulent boundary layer on the nozzle walls of the facility itself. Therefore the level of pressure fluctuations in the freestream is in general already saturated, that is the reason why the measured values in the different facilities are equivalent.

Using different solutions to control the local temperature on the wall of a 7deg half-cone, it was the first time that Europe and Russia measured the effect of such thermal control on boundary layer transition in hypersonic regime. Respectively, Europe adopted a closed oil circuit with an external unit to control the surface temperature whereas Russia used a simpler approach with electrical heaters and liquid nitrogen.

Steady state results using infrared thermography are showing a clear effect of the thermal control on the boundary layer transition onset. Experiments show that in general a local cooling is delaying transition and a local heating is anticipating it. Moreover, it is shown that the local temperature ratio (wall temperature against recovery temperature) is a key parameter. The amount of shift of the transition onset it is dependent on the local temperature ratio, therefore the shift can be adjusted simply by increasing or decreasing the level of heating/cooling.

To better understand the mechanism involved in the thermal control of the boundary layer, several ultra high frequency pressure sensors were used in the streamwise direction along the surface of the conical wind tunnel model. The spectra of the fluctuations revealed in general a shift of the second mode instabilities toward higher values of frequency and lower values of amplitude when cooling is applied; whereas these trends are reversed when wall heating is used.

To better explain and validate the effect of the thermal control on the boundary layer mechanisms numerical computations using linear stability (LST) and direct numerical simulations (DNS) where used.

The numerical computations show indeed that on one hand because of the local surface cooling the second modes are initially destabilized and subsequently strongly stabilized downstream the cooling region, whereas the first modes are in general not sufficiently destabilized to take part in the transition process. On the other hand a local surface heating strongly destabilizes both first and second modes, as shown in the DNS computations, in a scenario where the first modes critical N-factor rapidly increases approaching the N-factor of the second modes in a competition for the strongest transition mechanism.

The results obtained using the local thermal control confirm the effectiveness and the potential impact of this technology able to modify the boundary layer transition behaviour on a vehicle in the high speed regime and thus lowering the head load on a re-entry vehicle for example. Summarizing, the experiments conducted in the ground test facilities and successively validated by specific linear stability computations (LST) and direct numerical simulations (DNS), they show that a localized strong cooling at the wall is able to visibly delay the boundary transition onset whereas a strong heating can anticipating it.

Potential Impact:

The project aimed an investigation of a new technique for boundary layer transition control for flight vehicle in high speed regime. The technique, based on local thermal variation, is a “non intrusive” way for an active control of the boundary layer transition around a flight vehicle. The results showed that with a local wall temperature variation it is possible to modify the mechanisms taking part into the boundary layer transition physics. In general both wind tunnel experiments and numerical tools (i.e. RANS and DNS) demonstrate that indeed such technology it is applicable to a simple wind tunnel model at hypersonic speed with large success.

The applicability of such newly-proof technology is extremely wide, nevertheless more effort has to be spent in the deep analysis of some unknown mechanism discovered during the experiments and confirmed by the direct numerical simulation (DNS).

The results obtained are already an important contribution to the aerospace community (ESA and ROSCOSMOS) to maintain the scientific competiveness with other countries. Moreover the results can provide the aerospace engineers with new tools to support the design of re-usable re-entry vehicles (see for example IXV, ARV and EXPERT) and to increase the safety in the civil hypersonic aircrafts of the future (see the LAPCAT project).

List of Websites:

www.transhyberian.eu

Informations connexes

Contact

Patrick Rambaud, (Assistant Professor.)
Tél.: +3223599622
Fax: +3223599600