Comparison of Return to Launch Site Options for a Reusable Booster Stage

Transcription

1 Comparison of Return to Launh Site Options for a Reusable Booster Stage AE8900 MS Speial Problems Report Spae Systems Design Lab (SSDL) Shool of Aerospae Engineering Georgia Institute of Tehnology Atlanta, GA Author Barry Mark Hellman Advisor Dr. John R. Olds Spae Systems Design Lab (SSDL) August 1, 2005

2 Table of Contents Table of Contents Table of Contents... ii List of Figures...iii List of Tables... iv Nomenlature... v 1.0 Introdution Historial Bakground and Previous Work Early Historial Look at Reusable Boosters Previous RTLS Comparison Studies Glidebak Studies and Examples Flybak Studies and Examples Boostbak Studies and Examples Tehnial Approah Branhing Trajetory Problem Contributing Analyses Configuration Aerodynamis Roket Propulsion Airbreathing Propulsion Trajetory Simulation Weights and Sizing Design Struture Matrix and MDO Results Optimal Staging Points Trajetory plots Glidebak Trajetory Flybak Trajetory Boostbak Trajetory Mass Breakdown of Eah RTLS Method Disussion and Conlusions Aknowledgements Referenes Appendix A... A-1 ii

3 List of Figures List of Figures Figure 1. Reusable Booster with Inflatable Drag-Cone... 3 Figure 2. Reusable Booster with Paraglider Figure 3. Liquid Flybak Booster Configurations for the Spae Shuttle... 5 Figure 4. Trajetory of Tossbak RTLS method Figure 5. Asent and Boostbak Trajetory for Kistler K Figure 6. Notional Branhing Trajetory... 8 Figure 7. Notional 3-View of Baseline Configuration Figure 8. Isometri View of Total Configuration in APAS Figure 9. Isometri View of Booster in APAS Figure 10. Isometri View of Upper Stage in APAS Figure 11. Aerodynamis for Total Configuration Figure 12. Aerodynamis for Booster without Base Drag Figure 13. Aerodynamis for Booster with Base Drag Figure 14. Aerodynamis for Upper Stage Figure 15. Vauum Thrust to Weight for Booster Engine Figure 16. Exit Area for Booster Engine Figure 17. Vauum Thrust to Weight for Upper Stage Engine Figure 18. Exit Area for Upper Stage Engine Figure 19. Airbreathing Isp vs. Cruise Mah for Engine Parameters in Table III Figure 20. TPS Area Density as a Funtion of Maximum Re-entry Mah Number Figure 21. DSM Used for Modeling of the RTLS Methods in this Study Figure 22. Total Gross Mass of Booster and Upper Stage at Various Staging Points Figure 23. Total Dry Mass of Booster and Upper Stage at Various Staging Points Figure 24. Final Altitude Over Launh Site for Glidebak at Various Staging Points Figure 25. Comparison of RTLS Methods When Optimized for Lowest Gross Mass Figure 26. Comparison of RTLS Methods When Optimized for Lowest Dry Mass Figure 27. Altitude vs. Downrange for Glidebak Trajetory Figure 28. Ground Trak for Glidebak Trajetory Figure 29. Dynami Pressure, Normal Fore, and Overall Aeleration vs. Time for Glidebak Trajetory Figure 30. Mah Number and Aerodynami Angles for Glidebak Trajetory Figure 31. Altitude vs. Downrange for Flybak Trajetory Figure 32. Ground Trak for Flybak Trajetory Figure 33. Dynami Pressure and Mah Number vs. Time for Flybak Trajetory Figure 34. Normal Aeleration and Aerodynami Angles for Flybak Trajetory Figure 35. Altitude vs. Downrange for Boostbak Trajetory Figure 36. Ground Trak for Boostbak Trajetory Figure 37. Dynami Pressure and Normal Fore vs. Time for Boostbak Trajetory Figure 38. Mah Number and Aerodynami Angles for Boostbak Trajetory iii

4 List of Tables List of Tables Table I. Baseline Configuration Parameters for RTLS Booster Table II. Roket Engine Parameter Held Constant in this Study Table III. Ranges of Inputs for Turbofan DOE Table IV. Ranges for DOE Run in Flybak Dek in POST Table V. Inputs for SESAW Table VI. Parameters that Result from Optimal Staging Points Table VII. Mass Breakdown for Eah RTLS Booster Table VIII. Mass Breakdown for Upper Stage for Eah RTLS Method iv

5 Nomenlature Nomenlature APAS Aerodynami Preliminary Analysis Software BPR Bypass Ratio CAD Computer-Aided Drawing DOE Design of Experiments DOF Degree of Freedom DSM Design Struture Matrix ΔV Change in Veloity ELV Expendable Launh Vehile FPA Flight Path Angle FPI Fixed Point Iteration GUI Graphial User Interfae HABP Hypersoni Arbitrary Body Program ISP Speifi Impulse KSC Kennedy Spae Center LH2 Liquid Hydrogen LOX Liquid Oxygen MDO Multi-Disiplinary Design Optimization MER Mass Estimating Relationship MR Mass Ratio OBD Optimizer Based Deomposition POST Program to Optimize Simulated Trajetories RLV Reusable Launh Vehile RP Roket Grade Kerosene RSE Response Surfae Equation RTLS Return to Launh Site SESAW Simple Eletrial Systems and Avionis Wizard SSTO Single Stage To Orbit TSTO Two Stage To Orbit UDP Unified Distributed Panel USAF United States Air Fore v

6 Introdution 1.0 Introdution Ever sine the United States beame involved in developing launh vehiles, there has been a desired to develop a vehile that ould plae an payload in orbit, return to its launh site, and be reused quikly similar to a onventional airplane. The major deterrent to developing a RLV has been the large upfront development ost without the launh frequeny neessary to justify that ost. The only partially reusable vehile developed is the Spae Shuttle whih has shown to have large operation expenses and a long turn around time. Its solid roket boosters must be extrated from the oean and undergo long maintenane proedures to lean out the salt water. All unmanned launhes around the world urrently make use of ompletely expendable launh vehiles whih an run up to 500 million a launh. Along with the urrent need in the USAF for responsive spae lift and for long range strike and the growing spae tourist market, there is now a growing desire to develop a partial or totally reusable TSTO launh vehile to replae the urrent ELV s and the Shuttle. A urrent trend in RLV development is to fous on reusing the booster while leaving the upper stage expendable. A vehile like this an serve as a stepping stone to a totally reusable system by developing operational tehniques, prove tehnologies, better determine eonomi benefits, and build publi support for reusable launh arhitetures. One of the main design deisions for a reusable booster is to determine the method it will use for RTLS. If the booster is unpowered after staging, then there is a limit as to how high the staging an our so the booster will have enough margin to return and land safely. If the booster uses airbreathing engines, hardware for those engines needs to be added along with another fuel tank thereby inreasing the omplexity, mass, and ost of the booster. The airbreathing engines, however, will allow for a higher staging point, sine it an return after turning around at a point farther downrange. Another possibility for a RTLS booster is to use the asent roket engines to provide the neessary aeleration for RTLS. This study will ompare the RTLS methods of boostbak, flybak with an airbreathing engine (referred to as flybak in this study), and glidebak and ompare them based on the optimal staging point for both gross mass and dry mass. In order to find this, a multi-disiplinary model of the vehiles inorporating many different design tools was made to be able to find an optimal sized vehile based on the staging point while keeping most other parameters onstant. 1

7 Historial Bakground and Previous Work 2.0 Historial Bakground and Previous Work 2.1 Early Historial Look at Reusable Boosters The desire to build a RLV has been around sine the days of the spae rae to the moon. As with today, they were onerned with the ost of spae aess. It was believed that eonomis would limit spae exploration and not tehnology 1. Many early studies foused on reusing the booster as the first method to lower the diret operating osts of plaing a payload in orbit. A study by Boeing 2 initially looked at the options of a RTLS booster and a booster that flies ballistially into the oean and is then piked up by a boat. The biggest differene in development ost was the lifting surfaes and the assoiated modifiations to a standard expendable booster, whih would inrease with staging altitude and Mah number. Also, adding wings would require flight testing for a manned system. Reovering the booster in the oean, however, would need small modifiations adding aerodynami stabilization and deeleration devies suh as parahutes. Overall, a higher researh and development ost of almost 30 to 1 to build a RTLS booster gave a muh more favorable outlook towards water reovery. The manufaturing ost to build eah RTLS booster would ost about 3.5 times that of the oean reovery. The advantages of the RTLS booster show up when looking at the reurring osts. It was estimated that an individual RTLS booster ould be used almost 5-6 times more than the oean reovery. In the early 1960 s, the Douglas Airraft Co. proposed using a inflatable drag-one, shown in Figure 1, system to land the booster in the oean that ould be towed bak to land 1. The one would be inflated by the residual hydrogen and helium pressurant. The gases would provide air buoyany at an altitude of about 2000 ft to help redue impat veloity. Additionally, the one would protet the booster from salt water, allowing for redued maintenane onerns. 2

8 Historial Bakground and Previous Work Figure 1. Reusable Booster with Inflatable Drag-Cone 1. A method for gliding bak was proposed at NASA Langley that would use a paraglider, shown in Figure 2, to glidebak to the launh site. The paraglider would ontrol the re-entry fores, provide neessary lateral turns, and land the booster at a very slow desent rate. The ost of implementing this type of system was expeted to be very high. Figure 2. Reusable Booster with Paraglider 1. Another method to allow for easier booster reovery from the oean was to inrease the amount of vertial rise of the booster before beginning to pith over for orbit insertion 3. This idea was based on the belief that it is advantageous to redue the distane from the launh site to the point of atmospheri entry. The studied showed that for a 50% inrease in vertial rise time of the booster allowed for a 15% redution in impat distane. The main motivation of this study of reduing the downrange distane of atmospheri entry is one motivation for the idea behind the boostbak RTLS method. 3

9 Historial Bakground and Previous Work 2.2 Previous RTLS Comparison Studies There are no widely known publily published omparisons in whih the boostbak RTLS method is ompared to other the other methods. The urrent omparison studies only fous on glidebak vs. flybak. One study onduted at Adelaide University in Australia and the University of Stuttgart looked at glidebak and flybak launhing out of the Woomera launh site in Australia 4. In this study both stages were assumed to be reusable with both stages using LOX/RP propellants. Along with analyzing the RTLS methods, the study also noted that it is not optimal to perform high g-loading turns as the extra wing weight would not make up for additional payload. This study allowed for 3.6 g turns during the RTLS branh. The optimal staging point for glidebak was at a veloity of about 3937 ft/se at and altitude of about 16.2 nmi, while flybak staged at about 10,498 ft/se at an altitude of about 34.5 nmi. This study did not fix a payload but fixed a gross weight at liftoff. So mass differenes between the two RTLS methods must fous on the payload. The flybak vehile was shown to lift almost 50% more payload into LEO than glidebak. Although adding extra engines for flybak, this study reommends using that method as it an lift signifiantly more payload to orbit. Another omparison study onduted at Tokyo University ompared flybak vs. glidebak and also looked at the differene between different fuels, methane, LH2, RP, and densified LH2 5. This study shows that in order for glidebak to be able to our, the mass ratio between the booster and the orbiter must be small. This study showed for the LOX/RP flybak booster, the dry weight is 17% less than LOX/LH2 but the gross mass is 48% higher. The LOX/Methane booster was heavier in both gross and dry mass than the other propellants. The only glidebak boosters looked at in this study were LOX/LH2 showing that glidebak had 92% more gross mass and 63% more dry mass, over flybak with LOX/LH Glidebak Studies and Examples Although studies have shown that the powered flybak RTLS method an be advantageous when onsidering the gross and dry weight, glidebak an make the booster a simpler vehile without needing to add in extra hardware for the turbine engine and extra fuel tank. Even going bak to the early days of spaeflight as shown above, there were many ideas to implement the simpliity of gliding bak to the launh site after releasing an upper stage(s). One example of fully reusable TSTO onept oneived at NASA Langley was the Future Spae Transportation System to replae the Spae Shuttle 6,7,8. This system was design to arry 150,000 lbs to LEO. The booster uses LOX/RP propellants and the upper stage uses LOX/LH2 propellants. Sine the both stages burn simultaneously, the booster arries 4

10 Historial Bakground and Previous Work extra LH2 in a separate tank for the upper stage. The booster would glidebak after staging at Mah 3. The overall payload fration estimated for this design is about 3%. The total dry weight is estimated to be about 456,000 pounds whih is just over 3 times the weight of the payload. The wings were designed for a 2.5g normal aeleration limit. Due to the low staging Mah number, no TPS is required. Another more reent study at NASA Langley looked at developing a small launh vehile to go along with the urrent desire for responsive spae lift. The vehile proposed is has three-stages with a reusable booster and expendable upper stages apable of plaing a 330 lb payload in polar orbit launhed out of Vandenberg. The staging point was set at Mah 3. The glidebak flight was limited to 2.5 g s of normal fore and a maximum dynami pressure of 1000 psf. The glidebak trajetory was analyzed in POST with the objetive funtion to maximize altitude over the launh site and to had be modified to avoid overflight of the oast.. The booster takes 356 seonds for RTLS. The estimated payload fration is about 2.8% and the dry is about 100 times larger than the payload weight. 2.4 Flybak Studies and Examples A major effort sine the spae shuttle began operating looked at replaing the solid roket boosters with liquid flybak boosters. The onfigurations of dual boosters and atamaran booster are shown in Figure 3. The boosters would stage at about Mah 5.2 and an altitude of about 163,000 ft 9. The boosters would oast to an apogee of about 270,000 ft, re-enter, perform a transoni turn and ruise bak to KSC at about 18,500 ft and Mah It was estimated that about 20,000 lb of fuel would be needed for the turbine engine 10. Figure 3. Liquid Flybak Booster Configurations for the Spae Shuttle 9. The Europeans also looked at using a flybak booster for their Ariane launh vehile. One study onduted by DLR in Germany 11 looked at different onfigurations inluding one booster, a double fuselage atamaran, two small boosters attahed to the single upper stage, and a single booster inline with the upper stage. The analysis of the single booster using LOX/RP showed that four turbofans were needed for flybak. This study allowed for a maximum normal aeleration of 3.5 g s. The staging Mah number for the various 5

11 Historial Bakground and Previous Work onfigurations range from 5.38 to Eah booster onfiguration reahes a peak altitude of about 328,000 ft. The maximum re-entry Mah number is about the same as the staging point. 2.5 Boostbak Studies and Examples In the mid 80 s MDonnell Douglas proposed a tossbak booster in whih the booster is ballistially thrown bak to the launh site 12. After staging and a 20 seond reorientation period, the booster pithes bak and exeutes a short roket engine burn. Figure 4 shows the trajetory for this type of maneuver. This plot shows how asent branhes up until the staging point is elevated when ompared to a normal optimized asent trajetory with no tossbak. This study attempted to maximize payload fration and showed that with tossbak a payload fration of about 5% an be ahieved with a staging point around 10,000 ft/se ΔV at an attitude of about 237,000 ft. The reovery of the booster was proposed to happen in a fresh water pond after a brief terminal desent burn from the booster engines. There are no wings for this method of RTLS giving the booster a max L/D of about This will still allow for 5-10 nmi rossrange ontrol. Figure 4. Trajetory of Tossbak RTLS method 12. A example of boostbak in whih hardware has atually been built but not yet flown is the Kistler K-1 13,14,15,16. This is a fully reusable launh vehile in whih the booster exeutes a burn similar to the tossbak method disussed above, shown in Figure 5. This staging ours at about Mah 4.4, around 135, ,000 ft., and 135 seonds into the flight. For the boostbak burn, the enter engine (out of 3) burns for 35 seonds. The booster uses parahutes to slowdown after re-entry and air bags for landing. The landing takes plae at about 12 minutes after liftoff. 6

12 Historial Bakground and Previous Work Figure 5. Asent and Boostbak Trajetory for Kistler K

13 Tehnial Approah 3.0 Tehnial Approah The main fous of this study is to ompare the vehile sizes required to omplete a ertain mission utilizing three different RTLS methods. The vehiles will be ompared by the total amount of gross weight and dry weight. The mission requirements for this omparison are to launh a 15,000 lb payload to a 100 nmi irular orbit and a 28.5º inlination. The launh site is Kennedy Spae Center (28.5ºN latitude and 80.5ºW longitude). To ensure enough margin to land bak at the launh site, the booster must return to KSC with at least an altitude 15,000 ft. Both stages use LOX/RP-1 for propellant. 3.1 Branhing Trajetory Problem The use of a reusable booster poses an multi-disiplinary problem known as a branhing trajetory 17. For a TSTO vehile, the overall flight path of both stages is broken up into three different branhes. The asent branh inludes both stages mated together starting at the launh site and flying to a ertain staging point. This staging point is seleted based on a ertain amount of ideal ΔV from the onditions at the launh site. The upper stage then flies along the orbital branh to orbit insertion. Figure 6 shows a notional branhing trajetory. In this study, both the asent and orbital branh were modeled in the same trajetory input file. The booster exeutes neessary maneuvers for the RTLS branh from the onditions at staging. A seven parameter state vetor is used to identify the staging onditions: latitude, longitude, veloity, flight path angle, altitude, azimuth diretion of the veloity vetor, and the after staging weight of the booster. Orbital Branh Staging Point RTLS Branh Asent Branh Figure 6. Notional Branhing Trajetory. 8

14 Tehnial Approah The RTLS branh brings in a modeling problem of the booster when it requires additional propellant. The booster and upper stage without RTLS an easily be sized so the proper payload an be inserted into a desired orbit. Adding in extra propellant for the booster for RTLS, requires the booster to grow in size, thereby inreasing the amount propellant needed to reah staging, and the problem beomes ylial. Feedbak is thus required in the vehile model to ensure proper sizing. In order to model the TSTO vehile with a reusable booster, MDO methods are needed to handle the feedbak aspet of the problem. 3.2 Contributing Analyses Modeling the TSTO vehile with and RTLS booster is broken down into many different design disiplines that rely on information from other disiplines and assumptions to be made. This setion desribes eah of the different disiplines or ontributing analyses, what assumptions are made and what design tool is used Configuration The shape of the booster is based off the shape of the SSTO RLV using the ACRE- 92 designed for the NASA HRST study and further modeled by Spaeworks Engineering, In. 18. The baseline booster fuselage for this study is ylindrial with a drooped nose setion and a blunt end. The propellant tanks make up most of the fuselage with the forward dome of the fuel tank loated inside the nose setion. There is a 2 foot spaing between the propellant tanks and a 5 foot setion in the rear. For the airbreathing RTLS booster, the rear setion is assumed to be 10 feet to allow for a flybak fuel tank and extra hardware for the turbine engines, thereby lowering the pakaging effiieny. The wing is a delta shape with vertial fins loated at the wing tips. The rear of the booster has a body flap underneath the engines. The upper stage is ylindrial body mounted on top of the booster. The diameter of the upper stage is set to 6 ft. The propellant tanks have domes with a height to radius ratio and a 2 foot intertank adapter. The thrust struture is loated on the rear dome of the oxidizer tank. The fairing has a 10 foot onial setion on top of a ylindrial housing for the payload. A 2-D 3-View of the booster is shown in Figure 7. The baseline onfiguration parameters are listed in Table I. 9

15 Tehnial Approah Table I. Baseline Configuration Parameters for RTLS Booster. Fuselage Length 165 ft Total Fuselage Volume 94,676 ft 3 Pakaging Effiieny (Glidebak and Boostbak) % Pakaging Effiieny (Flybak) % Strutural Span 110 ft Wing Referene Area 6,000 ft 2 Maximum Wing Root Thikness 9 ft Total Vertials Planform Area 425 ft 2 Body Flap Length 10 ft Primary Struture Area 2758 ft 2 Seondary Struture Area ft 2 Body Flap Planform Area 280 ft 2 TPS Planform Area 9,512.7 ft 2 TPS Total Wetted Area 23,579 ft 2 LOX Tank RP Tank LOX/RP Engines 110 ft 28 ft LOX Tank RP Tank 165 ft Figure 7. Notional 3-View of Baseline Configuration. 10

16 Tehnial Approah Aerodynamis The aerodynami model for the booster was alulated using the APAS ode 19. The vehile inputted into the ode had dimensions based on a 0.5 sale fator of the baseline onfiguration dimensions for the booster shown above. The upper stage was inputted with a 6 ft radius and 40 ft length. The vehile shape was analyzed in UDP for low speed and transoni flight (Mah 0 Mah 2) and HABP for the hypersoni flight (Mah 3+). Sreen shots of the booster with the upper stage attahed, booster alone, and upper stage alone are shown in Figure 8 through Figure 10. The total onfiguration and upper stage alone were analyzed without base drag. Sine the booster will at times be gliding, it was analyzed with and without base drag. Figure 11 through Figure 14 shows the drag polars and lift to drag trends for the aerodynami model. Figure 8. Isometri View of Total Configuration in APAS. 11

17 Tehnial Approah Figure 9. Isometri View of Booster in APAS. Figure 10. Isometri View of Upper Stage in APAS. 12

18 Tehnial Approah Lift Coeffiient Mah 1.5 Mah 3.0 Mah 5.0 Mah Drag Coeffiient 3 Lift to Drag Ratio Mah 1.5 Mah 3.0 Mah 5.0 Mah Angle of Attak (deg) Figure 11. Aerodynamis for Total Configuration. 13

19 Tehnial Approah Lift Coeffiient Mah 1.5 Mah 3.0 Mah 5.0 Mah Drag Coeffiient 3 2 Lift to Drag Ratio Mah 1.5 Mah 3.0 Mah 5.0 Mah Angle of Attak (deg) Figure 12. Aerodynamis for Booster without Base Drag. 14

20 Tehnial Approah Lift Coeffiient Mah 0.5 Mah 0.8 Mah 1.5 Mah 3.0 Mah 5.0 Mah Drag Coeffiient 6 4 Lift to Drag Ratio Mah 0.5 Mah 0.8 Mah 1.5 Mah 3.0 Mah 5.0 Mah Angle of Attak (deg) Figure 13. Aerodynamis for Booster with Base Drag. 15

21 Tehnial Approah Lift Coeffiient Mah 3.0 Mah 5.0 Mah 10.0 Mah 20.0 Mah Drag Coeffiient 16

22 Tehnial Approah Roket Propulsion Two different rubberized engine types were assumed for the booster and upper stage. Both engines types are single hamber, LOX/RP, gas-generator yle, with a series flow sequene. The basi engine parameters, listed in Table II, were assumed to be onstant for both stages. Eah engine has Engine Health Monitoring, double ontroller redundany, gimbals, and no engine naelle. Table II. Roket Engine Parameter Held Constant in this Study. Oxidizer to Fuel Ratio 2.7 Booster Engine Expansion Ratio 30 Upper Stage Engine Expansion Ratio 60 Chamber Pressure 2500 psi Gas Generator Oxidizer to Fuel Ratio 1.5 The only parameter that was allowed to vary for the engines in this study was the amount of vauum thrust per engine. The vauum thrust to weight (T va /W e ), exit area, and vauum Isp were alulated using REDTOP-2 version The 1-D urve fits used for T va /W e and exit area in the overall vehile model are shown in Figure 15 through Figure 18. The Vauum Isp output did not show any trend to fit a urve so the average outputted Isp for engine type was used, se for the booster, and se for the upper stage. The number of engines an be seleted so the engines are appropriately sized. The amount thrust needed for the engines was taken from a given sea-level thrust to weight at liftoff of 1.2 and vauum thrust to weight at staging of 1.0. The reason that the thrust to weight at staging was set to 1.0 was beause for the Glidebak RTLS option, the upper stage required more thrust at staging to reah orbit beause of the low staging point. 17

23 Tehnial Approah Tva / We y = Ln(x) R 2 = Vauum Thrust (klbf) Figure 15. Vauum Thrust to Weight for Booster Engine Exit Area (ft 2 ) y = 0.037x R 2 = Vauum Thrust (klbf) Figure 16. Exit Area for Booster Engine. 18

24 Tehnial Approah Tva / We y = Ln(x) R 2 = Vauum Thrust (klbf) Figure 17. Vauum Thrust to Weight for Upper Stage Engine Exit Area (ft 2 ) y = x R 2 = Vauum Thrust (klbf) Figure 18. Exit Area for Upper Stage Engine. 19

25 Tehnial Approah Airbreathing Propulsion The flybak RTLS option makes use of a turbofan engine to ruise bak to the launh site after re-entering the atmosphere and turning around. The atual amount of thrust required for the engine at any given point during the ruise is alulated by POST to maintain a onstant Mah number and altitude. The other input required for POST is the Isp of the engine at the ruising altitude. Sea-Level thrust, used for sizing the engine mass, was alulated based on the density ratio between ruising altitude and sea-level and the required ruising thrust at the beginning of the flybak ruise alulated by POST. The installed thrust to weight for the turbofan was estimated to be 3.0. The engine performane ode PARA version was used to alulate the SFC (then onverted to Isp) of the turbofan based on five inputs, ruise altitude, ompressor pressure ratio, fan pressure ratio, BPR, and ruise Mah number. Sine this program is stritly a GUI interfae and ould not be integrated into ModelCenter very easily, a RSE based on the five inputs with ranges shown in Table III is used to rapidly alulate the SFC and then onverted to Isp. The R 2 value for this RSE is The RSE was reated using a 43 run DOE that is based on a Central Composite Design 22. The airbreathing engine Isp at 15,000 ft alulated from the RSE is shown in Figure 19 Table III. Ranges of Inputs for Turbofan DOE. Input Parameter Lower Bound Upper Bound Value Used Cruise Altitude 10,000 ft 35,000 ft 15,000 ft Compressor Pressure Ratio Fan Pressure Ratio BPR Mah Number Output From POST 20

26 Tehnial Approah Turbine Engine Isp (se) Cruise Mah Number Figure 19. Airbreathing Isp vs. Cruise Mah for Engine Parameters in Table III Trajetory Simulation The trajetory simulations were modeled using NASA s 3 DOF Program to Optimize Simulated Trajetories. This is a FORTRAN 77 legay ode very highly used in the spaeraft and launh vehile industries. The input dek of the trajetory is broken down into different events that begin at user defined triggers and inorporates all aspets of flight inluding aerodynamis, the urvature and oblateness of the earth, the orientation of the vehile, veloity losses, et. The program will integrate along the equations of motion for a given set of user defined input variables. On top of the trajetory analysis, POST an use 1 of 3 methods to optimize the input variables based on a set of user defined dependant variables to ahieve a user defined objetive funtion 23. For the asent branh and the three different RTLS branhes, POST s aelerated projeted gradient optimizer was used. Projeted Gradient is based on the Method of Feasible Diretions optimization sheme in whih the optimizer will find a feasible design point and follow the onstraints to find the optimized value 24. The four POST deks used in this study are in Appendix A. The Asent Branh dek is ommon to all three RTLS options. Its objetive funtion is to maximize the burnout weight at a irular 100nmi orbit at 28.5º inlination. The first event is the total onfiguration taking off vertially from KSC and beginning to pith over. The pith angles along the path are optimized based on a linearly interpolated 21

27 Tehnial Approah table of time steps. The staging point ours at an inputted value of ideal hange in veloity (ΔV). At this point, the weight of the booster is dropped and the upper stage oasts for 5 seonds before lighting the upper stage engines. The upper stage then flies along optimized pith angles until reahing orbit veloity at 100 nmi and 0 flight path angle. The engine thrust is throttled to ensure a maximum aeleration of 6 g s. The vetor of onditions at staging are wrapped into ModelCenter and passed onto the RTLS branh. The glidebak RTLS branh starts at the staging onditions from the asent branh with the weight of the booster after staging. The booster flies at zero bank and angle of attak of 45º. The booster oasts to the peak of its trajetory and then re-enter the atmosphere. When the wind loading reahes 2.5 g s, the angle of attak is lowered using POST s generalized aeleration steering, until the dynami pressure reahes its peak (hange in dynami pressure equals zero). At this point, the booster begins to turn at optimized angles of attak and bank angles. The booster turns until it reahes an optimzied azimuth angle and then has 20 seonds to pull out of the turn to a zero bank and an angle of attak for best L/D based on the given aerodynami model. The POST dek ends when the booster reahes the longitude of the launh site. The independent variables are the angles off attak, turning bank angles, and the azimuth angle to turn to. The onstraints are the final latitude, max wing loading of 2.5 g s, and max dynami pressure of 500 psf. The flybak RTLS branh flies a onstant angle of attak of 50 degrees from the staging point. When it reahes a wing loading of 2.3 g s, the angle of attak is lowered using generalized aeleration steering so as to keep the wing loading onstant. Sine the booster has not used any RTLS fuel and is heavier than the landing weight, the redue g-load is used to aount for the fat that normal fore should be 2.5 times the landing weight and not the weight after staging. The total aeleration inreases due to the drag in the axial diretion. After the total aeleration drops bak to 2.3 g s, whih is after the booster phugoid mode altitude, the booster will begin to bank at optimized angles of attak and bank angles. It will keep turning until it reahes the appropriate optimized azimuth angle to fly bak to the launh site, take 25 seonds to pull out of the turn, and then will fly at 10º angle of attak (the best L/D ratio for this aero model) and zero bank until it reahes about 16,000 ft. It will then start the engine and pullout to about 15,000 ft. The angle of attak, throttle, and bank angle are then varied using generalized aeleration steering to fly straight and level until reahing the longitude of the launh site. The independent variables are the angles of attak and bank angles during the turn and the azimuth angle to turn to. The onstraints are the final longitude, wing loading, max angle of attak, and dynami pressure during the turn. The asent trajetory for the flybak method was modified to set a staging flight path angle to ontrol the maximum dynami pressure during re-entry. The flight path angle at staging used to produe the results shown here was 20º. 22

28 Tehnial Approah A RSE was made from the POST model of the flybak branh. This was done is response to the diffiulty of the total flybak RTLS model to onverge to a point using FPI. The reason believed for the onvergene diffiulty is the oasionally failure of POST s projeted gradient optimizer to find a true optimum eah time it was run thereby ausing the RTLS required MR to flutuate. The outputs of the RSE s are flybak propellant required (R 2 of ), maximum re-entry Mah number (R 2 of 1.0), ruise Mah number (R 2 of ), and ruise thrust (R 2 of ). The DOE used was an 8 variable Central Composite design 22. The ranges for the DOE are listed in Table IV. Table IV. Ranges for DOE Run in Flybak Dek in POST. Staging Latitude Staging Longitude Staging Relative Veloity Staging Relative Flight Path Angle Staging Azimuth Angle Staging Altitude Staging Weight Airbreathing Engine ISP deg deg ft/se deg deg 155, ,000 ft 120, ,000 lb se The boostbak RTLS branh begins with the booster pithing around to an angle of attak of 180º so that the thrust is opposite the veloity vetor with the roket engine off with in 10 seonds. One engine is then turned on and the booster flies to an optimal pith angle and that. The riteria for shutting off the engine was seleted to be the horizontal veloity towards the base (POST variable vxf). After shutting off the engine, the booster went to a 40º angle of attak and held that angle until the wing loading reahed 2.5 g s, sine the booster at this point has undergone the RTLS burn and is now at the landing weight. The angle of attak was then adjusted using generalized aeleration steering to maintain that normal fore until the altitude leveled off before the phugoid mode altitude rise. Then the booster glided at optimal pith and bank angles until reahing the longitude of KSC. The independent variables are inertial pith angles while the engine is running, and the angle of attak and bank angles during the gliding portion of flight. The onstraints are the maximum number of normal g s while gliding, maximum dynami pressure, a minimum altitude of 15,000 ft at the end of the flight, the final latitude, and a maximum angle of attak of 40º during gliding flight Weights and Sizing Weights and Sizing was set up in Mirosoft Exel and was done by breaking up the vehile into MERs for the booster and upper stage and sizing eah stage separately 25. The MR for eah of the three branhes is used to properly size the two stages. The upper stage is first sized by varying the total amount of propellant in the vehile. The booster is then sized

29 Tehnial Approah adding in the gross weight of the upper stage by varying the saling parameter of the baseline dimensions and the total amount of propellant required for the RTLS branh. The sizing parameters are set using Mirosoft Exel s Solver so that all required MR and alulated MR are equal. There are no tehnology redution fators added to these MERs so that the mass estimation is based on the urrent state of the tehnology. The Simple Eletrial Systems and Avionis Wizard 26 is used to estimate the weight of the avionis required for the booster and the upper stage. Table V shows the settings used in SESAW for eah stage. An Exel spreadsheet was made to give the same results from the same relationships 27 in the SESAW program so that it ould be integrated into ModelCenter. Table V. Inputs for SESAW. Input Parameter Booster Value Upper Stage Value System Redundany 2 2 Program Development Year Stage Length Current Value Current Value Guided Landing? Yes No Takeoff Vertially? Yes Yes Crew on Board? No No Piloted by Crew? No No At Least 1 Other Stage Reusable? No Yes Altitude Other Reusable Stages are Dropped N/A 50 nmi Is This Stage Expendable? No Yes Upper or Only Stage? No Yes Highest Altitude for this Stage Suborbital LEO Time in Orbit (days) The propulsion weight sizing is based on thrust to weight ratios disussed above. The roket engines thrust to weight ratios omes from the REDTOP-2 design tool. The airbreathing engine weight was alulated based on an assumed thrust to weight of 3 based on the sea level thrust. The thermal protetion MER used for the booster requires an input of surfae area density. The hart shown in Figure 20 shows the surfae area density used as a funtion of maximum re-entry Mah number that ours after the booster reahes the peak of its trajetory. Below Mah 3, no TPS is needed. A 15% propellant reserve was used flybak and 1.5% propellant was reserve was used for boostbak. The residuals for RTLS were 0.5%. There were 1% propellant reserves and residuals for asent on eah stage. Eah stage assumed a 1% startup amount of propellant. 24

30 Tehnial Approah TPS Area Density (lb / ft 2) Maximum Mah Number Figure 20. TPS Area Density as a Funtion of Maximum Re-entry Mah Number. 3.3 Design Struture Matrix and MDO Eah of the ontributing analyses fit into a design struture matrix shown in Figure 21. The links on the top right of the CA s are known as the feed forwards and the links on the bottom left are known as the feedbaks. Sine it is very diffiult, often impossible, to make a omputer model with the feedbak loops diretly, a MDO tehnique must be utilized to break the feedbak links. While this DSM enompasses any possible aspet of an launh system with a RTLS booster, not all of the links are needed for eah method, i.e. airbreathing engine is only needed for the flybak method. The MDO tehnique known as Fixed Point Iteration was used in ModelCenter to handle the feedbak. In FPI, guesses are made for the variables that are feedbak. All CA and feed forward links are then ran. The outputs from the model are then logged and put bak into the model as the new guesses. When the differene between the guesses and the outputs are within some appropriate tolerane, then system is assumed to be onverged. The only system variable in this study was the staging ideal ΔV. Relaxation was used were appropriate to dampen out onvergene osillations. The glidebak RTLS branh was taken outside of the overall system loop beause it does not give any output to the rest of the model beause it only tries to maximize the final altitude over the landing site based on the staging onditions. For glidebak, the booster and upper stage are sized based on the staging point and the glidebak RTLS branh is then run to determine what altitude it will reah over the launh site. 25

31 Tehnial Approah Configuration Aerodynamis (APAS) Roket Propulsion (REDTOP-2) Airbreathing Propulsion (PARA) (If Needed) Asent Trajetory (3D POST) RTLS Trajetory (3D POST) Avionis Weights (SESAW) Weights & Sizing (EXCEL and Solver) Figure 21. DSM Used for Modeling of the RTLS Methods in this Study. 26

32 Results 4.0 Results 4.1 Optimal Staging Points For the boostbak and flybak RTLS methods, a sweep of staging points was run to show where the staging point ours for lowest gross mass and for lowest dry mass. For glidebak, a sweep of staging points was done to find the final altitude after gliding bak to the launh site. The sweeps of staging points for gross and dry mass are shown in Figure 22and Figure 23. Figure 24 shows the final altitude at various staging points of the booster after gliding bak to the launh site. For both boostbak and flybak, the sweep of staging points show a buket shape for gross and dry mass show that there is an optimal staging point for minimum mass. For glidebak, no optimal point an be seleted beause there is a utoff point where the booster an t make it bak with the required altitude margin. Table VI shows the onditions at the optimal staging point for eah of the RTLS methods and the vehile parameters that results from that staging point. Figure 22 and Figure 26 show diret omparisons for the optimal gross and dry mass, for eah RTLS method. 1,000, ,000 Total Gross Weight (lbm) 900, , ,000 Flybak Glidebak Boostbak 750, Staging Mah Figure 22. Total Gross Mass of Booster and Upper Stage at Various Staging Points. 27

33 Results 105, ,000 Total Dry Weight (lbm) 95,000 90,000 85,000 Flybak Glidebak Boostbak 80,000 75, Staging Mah Figure 23. Total Dry Mass of Booster and Upper Stage at Various Staging Points. 50,000 45,000 Final Altitude Over Launh Site (ft) 40,000 35,000 30,000 25,000 20,000 15,000 10,000 5, Staging Mah Number Figure 24. Final Altitude Over Launh Site for Glidebak at Various Staging Points. 28

34 Results Table VI. Parameters that Result from Optimal Staging Points. Parameters from Optimal Staging Point Glidebak Flybak (Lowest Gross Mass) Boostbak (Lowest Gross Mass) Flybak (Lowest Dry Mass) Boostbak (Lowest Dry Mass) 29

35 Results 120, ,000 Flybak Total Dry Mass (lbm) 80,000 60,000 40,000 Glidebak Boostbak 20,000 0 Figure 26. Comparison of RTLS Methods When Optimized for Lowest Dry Mass. 4.2 Trajetory plots Glidebak Trajetory As shown above, the staging ondition that allows for the booster to have a final altitude of 15,000 ft above the launh site ours at about Mah 3.1. The altitude vs. downrange plot of the trajetory is shown in Figure 27 with the ground shown in Figure 28. The booster reahes a peak altitude of about 31 nmi, re-enters the atmosphere, and then turns south for returning. The turn begins when the booster is supersoni and ends after it has dropped below Mah 1. During its turn, it reahes a maximum downrange distane of about 52 nmi. It begins the straight glide under Mah 0.8 and deelerates until it reahes the launh site. Sine the peak re-entry Mah number of 1.86 (the peak in Figure 30) is well below Mah 3, no TPS is required for this method. The ground trak for glidebak and the other RTLS methods below show that when launh out of KSC, there is no need to modify the trajetory to avoid over-flight issues. The booster returns about 14 minutes after staging. 30

36 Results Altitude (nmi) Downrange (nmi) Figure 27. Altitude vs. Downrange for Glidebak Trajetory. Figure 28. Ground Trak for Glidebak Trajetory. 31

37 Results Dynami Pressure Wing Normal Fore Overall Aeleration 5 Dynami Pressure (psf) Wing Normal Fore (g's) Time (se) 0 Figure 29. Dynami Pressure, Normal Fore, and Overall Aeleration vs. Time for Glidebak Trajetory Mah Number Mah Number Angle of Attak Bank Angle Aerodynami Angle (deg) Time (se) Figure 30. Mah Number and Aerodynami Angles for Glidebak Trajetory. 32

38 Results Flybak Trajetory Figure 31 through Figure 34 shows the booster s trajetory for the optimal dry weight flybak method. The booster holds a onstant angle of attak reahing a peak altitude of about 48 nmi. As it re-enters the atmosphere, the angle of attak is dropped rapidly and then is inreased to hold the required normal fore until the phugoid mode altitude rise ours. At the peak of the altitude rise, the booster begins its gliding turn in whih the booster s speed is subsoni. The booster reahes a maximum downrange of about 350 nmi. After gliding about 30 nmi (whih is about the max downrange distane for the glidebak booster), the airbreathing engine must be turned on to hold an altitude of about 15,000 ft. The engine s Isp is about 3610 se requiring about 8,000 lbf of thrust for eah engine at the beginning of the ruise. The RTLS branh takes about 77 minutes. Also, the ground trak out of KSC reahes beyond the Bahamas whih ould possibly be used as an emergeny landing site. The peak re-entry Mah number is about 6.93, requiring TPS Altitude (nmi) Downrange (nmi) Figure 31. Altitude vs. Downrange for Flybak Trajetory. 33

39 Results Figure 32. Ground Trak for Flybak Trajetory Dynami Pressure Wing Normal Fore Total Sensed Aeleration 2.50 Dynami Pressure (psf) Wing Normal Fore (g's) Time (se) Figure 33. Dynami Pressure and Mah Number vs. Time for Flybak Trajetory. 34

40 Results Sensed Aelration (g's) Mah Number Angle of Attak Bank Angle Aerodynami Angle (deg) Time (se) Figure 34. Normal Aeleration and Aerodynami Angles for Flybak Trajetory Boostbak Trajetory The trajetory of the optimal dry weight booster using the boostbak RTLS method is shown in Figure 35 through Figure 38. The booster maximum downrange is about 24 nmi whih is less than the glidebak trajetory. This effet is due mostly to the reverse engine thrust providing more deeleration than and aerodynami turn. Also the deeleration an start ourring above the atmosphere, whih a glidebak maneuver annot due. The peak of the trajetory is at 28 nmi. During the engine burn, the angle of attak is very high to allow for the thrust vetor to be opposite the veloity vetor. After the engine burn, the booster is a dive like state when re-entering, and levels off after the peak deeleration. The maximum re-entry Mah number is below 3 so no TPS is required. After the phugoid mode altitude rise, a mild amount of banking is required to line up the booster with the launh site. The time of flight for the boostbak RTLS branh is about 3 minutes. When the booster reahes the launh site, it is well over 15,000 ft in altitude giving plenty of margin to land. 35

41 Results Altitude (nmi) Downrange (nmi) Figure 35. Altitude vs. Downrange for Boostbak Trajetory. Figure 36. Ground Trak for Boostbak Trajetory. 36

42 Results Dynami Pressure Wing Normal Fore Total Sensed Aeleration 6.00 Dynami Pressure (psf) Wing Normal Fore (g's) Time (se) Figure 37. Dynami Pressure and Normal Fore vs. Time for Boostbak Trajetory Mah Number Mah Number Angle of Attak Bank Angle Aerodynami Angle Time (se) Figure 38. Mah Number and Aerodynami Angles for Boostbak Trajetory. 37

43 Results 4.3 Mass Breakdown of Eah RTLS Method Table VII. Mass Breakdown for Eah RTLS Booster. Mass Component Glidebak Flybak Boostbak Wing Group 4,848 7,600 5,950 Tail Group Body Group 9,541 12,680 12,880 Primary Struture 1,906 2,530 2,480 Seondary Struture Crew Cabin (w/ esape system) Body Flap Thrust Struture 2,361 2,520 2,640 Oxidizer Tank 3,654 5,100 5,430 Fuel Tank 1,192 1,800 1,770 Airbreathing Flybak Fuel Tank Thermal Protetion 0 6,910 0 Landing Gear 1,982 2,760 2,180 Propulsion Hardware 14,483 19,800 15,790 Roket 8,974 9,270 9,670 Roket Feedlines, Pressurization, et ,840 6,120 Airbreathing 0.0 4,700 0 RCS Propulsion 850 1, Primary Power Eletrial Conversion & Dist 3,595 4,870 3,870 Hydrauli Systems Surfae Control Atuation Avionis 1,150 1,260 1,250 Environmental Control 2,652 2,650 2,650 Personnel Equipment Piggybak Struture 8,480 4,830 5,610 Thrust Struture 7,950 4,530 5,260 Link Struture Dry Weight Margin (15%) 8,697 11,790 9,350 Dry Weight 57,983 78,610 62,350 Crew and Gear Payload Provisions Residual Propellants 2,087 2,980 3,110 Asent Residuals 2,087 2,910 2,900 RTLS Residuals RTLS Reserve Propellants 0 2, Landed Weight 60,069 83,690 66,100 RTLS Propellant Weight 0 14,040 42,700 RCS Propellant Weight 632 1,030 1,150 After Staging Weight (Entry Weight) 60,702 98, ,950 Upper Stage 353, , ,710 Asent Reserve Propellants 2,087 2,910 2,900 In-flight Losses and Vents 4,203 3,060 3,500 Before Staging Weight 420, , ,060 Asent Propellants 417, , ,250 Oxidizer 304, , ,430 Fuel 112, , ,830 Gross Liftoff Weight 837, , ,320 Startup Losses 4,173 5,820 5,800 Maximum Pre-launh Weight 841, , ,120 38

44 Results Table VIII. Mass Breakdown for Upper Stage for Eah RTLS Method. Mass Component Glidebak Flybak Boostbak Struture 7,300 7,980 8,430 Primary Struture (intertank) 650 1,060 1,080 Fuel Tank (RP) 1, Oxidizer Tank (LOX) 2,680 1,420 1,680 Thrust Struture Payload Support 1,500 1,500 1,500 Seondary Struture (Payload Fairing) 630 2,990 2,980 Aero Surfaes Propulsion Hardware 5,510 3,340 3,810 Main Engines 3,720 2,290 2,600 Plumbing, Feed, and Pressurization 1,790 1,050 1,210 Power Reation Control 1, Avionis and Controls Piggybak Struture 2,120 1,210 1,400 Shear Struture 1, ,050 Link Struture Growth Margin (15%) 3,020 2,580 2,800 Dry Weight 20,130 17,230 18,630 Residuals and Reserves 3,180 1,690 2,000 Cirulization Propellant Payload 15,000 15,000 15,000 Burnout Weight 38,320 33,920 35,630 Main Propellants 313, , ,080 Fuel 84,710 44,970 53,260 Oxidizer 228, , ,810 Gross Weight 351, , ,710 Startup Losses 1, ,000 Maximum Weight 353, , ,710 39