CMCthermalprotectionsystemforfuturereusablelaunchvehicles.pdf

Acta Astronautica 65 (2009) 165176/locate/actaastroCMCthermalprotectionsystemforfuturereusablelaunchvehicles:GenericshingletechnologicalmaturationandtestsT. Pichona, R. Barreteaua, P. Soyrisa, A. Foucaulta, J.M. Parenteaua,Y.Prelb,S. GuedronbaSnecma Propulsion Solide, SAFRAN Group, Le Haillan, FrancebCNES, Evry, FranceReceived 22 December 2006; accepted 13 January 2009Available online 4 March 2009AbstractExperimental re-entry demonstrators are currently being developed in Europe, with the objective of increasing the technologyreadiness level (TRL) of technologies applicable to future reusable launch vehicles. Among these are the Pre-X programme,currently funded by CNES, the French Space Agency, and which is about to enter into development phase B, and the IXV,within the future launcher preparatory programme (FLPP) funded by ESA. One of the major technologies necessary for suchvehicles is the thermal protection system (TPS), and in particular the ceramic matrix composites (CMC) based windward TPS.In support of this goal, technology maturation activities named “generic shingle” were initiated beginning of 2003 by SPS,under a CNES contract, with the objective of performing a test campaign of a complete shingle of generic design, in preparationof the development of a re-entry experimental vehicle decided in Europe. The activities performed to date include: the design,manufacturing of two C/SiC panels, finite element model (FEM) calculation of the design, testing of technological samplesextracted from a dedicated panel, mechanical pressure testing of a panel, and a complete study of the attachment system. Additionaltesting is currently under preparation on the panel equipped with its insulation, seal, attachment device, and representative portionof cold structure, to further assess its behaviour in environments relevant to its applicationThe paper will present the activities that will have been performed in 2006 on the prediction and preparation of these modalcharacterization, dynamic, acoustic as well as thermal and thermo-mechanical tests.Results of these tests will be presented and the lessons learned will be discussed. 2009 Elsevier Ltd. All rights reserved.Abbreviations: C/SiC, carbonsilicon carbide composite; CMC,ceramic matrix composites; FEM, finite element model; HMS,health monitoring system; FLPP, future launcher preparatory progra-mme; MMOD, micro-meteoroids and orbital debris; NDT, non-destructive testing; RCC, reinforced carbon/carbon; RLV, reusablelaunch vehicle; R&T, research & technology; TRL, technologyreadiness level; TPS, thermal protection systemCorresponding author.E-mail address: thierry.pichonsnecma.fr (T. Pichon).0094-5765/$-see front matter 2009 Elsevier Ltd. All rights reserved.doi:10.1016/j.actaastro.2009.01.0351. Shingle demonstration objectivesDuring the Pre-X preliminary study in 20012002,Snecma Propulsion Solide, entrusted with the respon-sibility of the high thermally loaded area of the wind-ward side of Pre-X, proposed to design and provide theTPS using C/SiC based shingles. The use of large sizeshingles has been given particular attention.In order to reduce the risks, a preparatory programmecalled “generic shingle” was initiated by CNES. It is166 T. Pichon et al. / Acta Astronautica 65 (2009) 165176External Aerodynamic shapeStructure shape Mechanical shellStand-offInternal InsulationFastenerSealInsulating WasherFig. 1. TPS shingle concept.basedonthedesign,manufacturingandtestingofalargeshingle element. The main goal of this programme isto demonstrate the applicability of the improved tech-nology to all the C/SiC shingle elements foreseen forthe windward side of experimental re-entry vehicles. Tosupport this demonstration, the following activities havebeen performed: design of the panel and attachments, analytical validations of the design, manufacturing of a complete large shingle, technological tests of the C/SiC panels particulari-ties, pressure and vibration test (sine, random, and acous-tic) of the C/SiC panel, thermal testing, and thermo-mechanical testing.2. Shingle concept descriptionThis concept also called “shingle” is divided into twosets of elements: theoneswithmechanicalfunctions(mechanicalshell,fasteners, and stand-offs) and the ones with thermal functions (inner insulation lay-ers, seals, and insulating washers).This is represented by Fig. 1.The material needed for the mechanical shell hasto be mechanically very efficient, and resist to highlyconstrained thermal environment; but its thermal con-ductivity characteristics are not the most important. Theinternal insulation and seals, not needing high mechani-cal properties, can be composed of low weight, flexible,and high performance insulating materials. The attach-ment system of the panel to the airframe structure mustresist to relatively high temperatures, enable the thermalFig. 2. Aerodynamic external shape of the panel of generic shingle.0501001502002503000100 200 300 400 500 600 700 800 9001000 1100 1200time (s)Flux (kW/m2)Fig. 3. Pre-X generic shingle reference flux.expansion of the panel, and transmit out-of-plane me-chanical loads between the panel and the cold structure.3. Main shingle requirementsThe panel is to be manufactured with the same gen-eral requirements, same C/SiC material and same pro-cess as the ones foreseen for the Pre-X windward sideshingles. As such, the requirements are derived from thespecifications of the Pre-X TPS system and are summa-rized here after. Size of panel no smaller than an equivalent aerody-namic surface of 800400mm2. Ability to demonstrate the capacity to manufacturenon-rectangular patterns as a minimum angle of 15is necessary between the aerodynamic flow lines andedges of shingles. The approximate geometry is givenin Fig. 2. Areal mass goal of 15kg/m2, the panel itself shouldweigh less than 2.3kg. Ability to withstand the most severe heat flux andmechanical loads encountered on the windward sidewhen fixed in a representative way to a rigid struc-ture. The reference heat flux considered is provided inFig. 3. Evolution of the flux on the generic shingle surfacederived as a percentage of the reference flux. Thisevolution is provided in Fig. 4. Under this flux, thestructure should be maintained under 150C.T. Pichon et al. / Acta Astronautica 65 (2009) 165176 167100%100%100%110%110%300mm500mmparoiT41 case1zxyFig. 4. Flux evolution on the aerodynamic surface of the genericshingle panel.Fig. 5. Exterior view of the panel design, with the two skin sewnparts. Mechanical loads specified for the panel: Pressure difference between the inside and outsideof the shingle element is +100mbar during take-off and 100mbar during re-entry, accelerations 10g in the panel plane and 5g normalto the panel plane, dynamic loads considered as static accelerations of15g in the panel plane and 10g normal to the panelplane, acoustic spectrum derived from Ariane 5, and flight loads, such as thermo-mechanical loads anddeformation of the cold structure.4. Panel designIn order to withstand the above requirements and totake into account the manufacturing process, the designof the panel is represented in Figs. 5 and 6.The need for three stiffeners has been determinedconsidering simple mechanical formulas. However, theneed for nine attachment points has been establishedusingFEMthermo-mechanicalanalysis.Inaddition,theFig. 6. Interior view of the panel design, with the edge and insidestiffeners. All links are textile sewing except for the stitched blueinterior stiffener. (For interpretation of the references to the colourin this figure legend, the reader is referred to the web version ofthis article.)interfaces between two adjacent panels is designed totake into account: the reduction of the steps and gaps, in particulardue to relative thermo-mechanical displacementsand the ability to allow external access for integrationand dismounting.This led to a conceptual design of two-shingle interfacerepresented in Fig. 7.Finally, the design has been made in order to addas many manufacturing process validations as possible.That is why the skin of the panel was manufacturedin two separate textile parts sewn together. This is notnecessary for this specific panel size, but if larger partsarerequired,itmaybeofinterest.Inaddition,twotextilelinking techniques were used to assemble the stiffenersto the skin: weaving and stitching.5. Attachment system designIn order to attach the panel to the vehicle structure,a flexible attachment system has been designed. Thissystem is able to fulfil the following functions: mechanically attach the panel to the structure, enable expansion differences between panel andstructure by adapted flexibility of the stand-off, prevent large outer mould line deformation, throughsufficient stiffness, participate to thermal protection of the structure, be mountable and dismountable without visibilitythrough a small hole and without losing parts, and transfer loads from the panel to the structure.168 T. Pichon et al. / Acta Astronautica 65 (2009) 165176Fig. 7. Junction between two adjacent shingles on a standard area (left) and at an attachment point (right).Fig. 8. Attachment configuration.In order to fulfil these needs, the stand-off solution pro-vides a two axis flexibility feature. Then, this stand-offis mechanically fixed to the structure while minimizingheat transfer. This implies the need of thermal washers.Elastic washers allow to maintain a correct tighteningwhile the parts expand differently due to temperature.In order to comply with the accessibility needs evenwith no visibility, a system where all the washers arefixed to one of the parts (panel or stand-off and struc-ture) before assembly has been designed. The resultingconfiguration is shown in Fig. 8.6. Panel-components manufacturingTwo panelsbased onthedesign described above weremanufactured. One was cut up to perform tests on par-ticular local areas, to validate these particularities, andfurther refine the design of the shingle. The second onewas kept intact and is currently used for testing. Themanufacturing process is described below.The production begins with the manufacturing of acarbon fibre carbon precursor reinforced preform. ThisFig. 9. The two manufactured generic shingle panels.preform is based on one ply multi-layer woven fabricwhich can be assembled by weaving to result in a selfstiffened panel. The moulding of these panels is per-formed on a carbon/epoxy composite material mouldable to withstand at least the temperature needed for thehardening process. This allows to reduce the thermalexpansion mismatch between mould and part that canappear with metallic tooling. It also enables to increasethe easiness of use by its low weight.Then the preform is hardened by a liquid route. Thesilicon carbide matrix is added by CVI densification onasimplegraphiteframeholdingthepart.Aftergrinding,de-burring and drilling of holes, the SiC CVI densifica-tion process is performed once again. It is also possi-ble to add an oxidation protection coating if necessary(Fig. 9).An infra-red NDI technique was applied to detect anypotential defects, as shown in Fig. 10.Apart from local lack of compacting of the preform,no additional defects were encountered. Overall den-sity of the part is correct although slightly less thanexpected.T. Pichon et al. / Acta Astronautica 65 (2009) 165176 169Fig. 10. NDI infra-red inspection of shingle panel.Improvement of the compacting efficiency of themould and higher overall density will be implementedon future productions.The attachment system was procured, based on thedesign presented here above. The insulation is a com-mercially available off-the-shell silica/alumina felt. Theseal around the panel is based on a silica/alumina feltinside a Nextel envelope, and was manufactured specif-ically to the geometry of the shingle.Finally, the necessary metallic support structure forthe various tests were procured, in order to be able tocorrectly interface with the test facilities, while remain-ing representative in terms of heat sink of the real coldstructure for the thermal tests, and with a comparablerigidity for the dynamic tests.7. Panel mechanical testA mechanical pressure test has been carried out onthepanelwiththeothercomponents.Theaimofthistestwas to validate the behaviour of the panel alone underrepresentative loading. The flexible attachment is notincluded but an infinitely rigid condition is considered.A FEM prediction of the test has been performed inorder to take into consideration the specificities of thetest and a factor of 1.3 on the load application has beenconsidered necessary in order to be representative offlight mechanical loads.It was foreseen to calibrate to 50mbar the test appli-cability and hence make a preliminary settling of thepanel material. Then the loading was to be slowly andincrementally increased up to a 130mbar pressure dif-ference representative of the 100mbar in flight.The observed airflow through the material porosityduring calibration led to the conclusion that it wouldnot be possible to reach the 130mbar during test. Asignificant part of the leakage came from the local lackFig. 11. Mechanical test of the C/SiC panel with all strain anddisplacement gauges.00.511.522.533.540 20 40 60 80 100 120 140 160pressure (mbar)displacement (mm)LVDT4Fig. 12. Displacement gauges in the middle of the larger skincompartment for example (initial prediction in black dotted line,large displacement post analysis in black plain line, and differenttest runs in colour).of compacting discussed before. Hence this leakage canbe reduced for future panels. However, the total surfaceof the normal material also participates in the leakage,due to the inherent permeability of the material. Thispermeability contributes to the reduction of the in-flightre-entry pressure difference, and thus measurement ofits value can help in evaluating the increase in designmargins with regards to the current design.After artificially air-tightening the system with an in-ternal flexible membrane an incremental load applica-tion was performed up to 130mbar representative of the100mbar in flight (see Fig. 11).After the test no failure of the panel was observed.A local damage is observed at approximately 90mbar,which is consistent with the tests performed in previousphases. However, the analyses are conservative com-pared to the global behaviour during the test.The displacements measured are lower than the onespredicted and the measured strain is also lower, asshown in Figs. 12 and 13.170 T. Pichon et al. / Acta Astronautica 65 (2009) 165176050010001500200025000 20 40 60 80 100 120 140 160pression (mbar)dformation (def)J1602000160J19200400600800100012001400160018000 20 40 60 80 100 120 140pression (mbar)J19dformation (def)Fig. 13. Strain gauges on the longitudinal edge stiffener for example.For example, the local damage, which was expectedto occur at approximately 35mbar, was observed at90mbar, and general damage, which was expected atapproximately 107mbar, did not occur.In addition, the test results show that the analysisoverestimates the deformation in the skin and in theangles, even the linear behaviour range of the material.With the post-test data analysis performed, the max-imum skin deflection expected in flight is no more than3.7mm with the current geometry, corresponding to awaviness of approximately 2%. The resulting marginof safety in the panel skin is greater than 0.86. In theframe angle regions, with a conservative approach, themargin of safety is expected to be positive. Only theattachment zone angle on the small frame very locallyshows a negative margin of safety: as identified afterdesign activities, an additional attachment point wouldbe necessary if pressure levels specified are confirmed.Refinedevaluationofthebehaviourwillbeperformedduring the thermo-mechanical post-test analysis.8. Dynamic testsAll the dynamic tests have been performed at IABGtest centre in Germany. These tests are based on thedynamic loading of the generic shingle, and the testcampaign had the following objectives: provide information on the modal characteristicsof the panel assembly with and without the in-sulation, validate the assembly to a given dynamic frequencyspectrum, composed of different dynamic frequen-cies. The loading is oriented, on one hand in z direc-tion, and in the other hand in x direction, and validate the assembly to a given acoustic spectrum.T. Pichon et al. / Acta Astronautica 65 (2009) 165176 171Fig. 14. Modal test set-up (credit IABG).8.1. Modal testWith regards to the objectives of the modal test, thefirstmodesoftheassemblyforfrequenciesrangingfromlow frequency (5Hz) to 350Hz have been investigatedwith and without the insulation material. The modalcharacteristics to be determined were the following: the eigen frequencies, the modal shapes, and the modal damping factors.The test article (with attachments and seals) was placedon a rigid support, and excited at two different locations(see Fig. 14).The measurements were well correlated with the pre-dictions made from FEM analyses for the configurationwithout internal insulation, up to 200Hz. The dampingwas slightly inferior to the arbitrary value taken in themodel (0.41% instead of 1%). With the insulation, thebehaviour of the equipped shingle was significantly al-tered: damping increased to 2.2% for overall displace-mentmodesandto5%forskinmodes.Suchhighdamp-ing is generally considered beneficial as it ought to re-duce the peak stress level under flight dynamic loads.The frequencies were also slightly modified by thepresence of the insulation, the first mode frequencyshifting from 95 to 91Hz. In both cases, the first modesarein-planeoveralldisplacementmodeslinkedtostand-offs, while the first out-of-plane mode occurs at 115Hzwith the insulation and 122Hz without the insulation asshown in Fig. 15.8.2. Vibration testThe a