AIAA 2001-0263INTEGRATED AERODYNAMIC DESIGNAND ANALYSIS OF LAUNCH VEHICLES

For permission to copy or republish, contact the American Institute of Aeronautics and Astronautics1801 Alexander Bell Drive, Suite 500, Reston, Virginia 20191-4344

39th AIAA Aerospace SciencesMeeting & Exhibit

8-11 January 2001 / Reno, NV

1

AIAA 2001-0263

INTEGRATED AERODYNAMIC DESIGN AND ANALYSIS OF LAUNCH VEHICLES

byMichael R. Mendenhall*, Martin C. Hegedus+,

Sarah T. Whitlock, Laura C. Rodman#, and Judy A. FaltzNielsen Engineering & Research

Mountain View, CA 94043

ABSTRACT

An integrated aerodynamic design and analysisprocedure for launch vehicles which incorporatesengineering prediction methods, historicalknowledge, and corporate memory into aknowledge-based system is described. Theobjective is to provide an economical means toconduct the aerodynamic design and analysis offuture launch vehicles to minimize the risk ofaerodynamic-induced failures. The method willprovide a guide to the aerodynamic-relatedinformation required during the conceptual andpreliminary design stages. In addition, the user willhave access to historical information on similarlaunch vehicles, the ability to set up geometriccharacteristics, and a capability to predictpreliminary aerodynamics of simple launchvehicles. The method includes a number ofsearchable databases containing technicalreferences on launch vehicle aerodynamics,engineering experiences and anecdotes on launchvehicle programs of the past, and informationabout various launch vehicles. The method can beused as a stand-alone tool for traditional designapproaches or as an integral part of amultidisciplinary design method. It has archivecapability, and it can be updated with newinformation as desired by the user.

_________________________

* Vice President, Associate Fellow+ Senior Research Engineer, Senior Member# Senior Research Scientist, Senior MemberCopyright 2001 by Nielsen Engineering &Research. Published by the American Institute ofAeronautics and Astronautics with permission.

INTRODUCTION

A primary mission goal for NASA and theaerospace industry is to design, develop, andmaintain high quality and reliable space launchsystems at less cost.1 Although aerodynamics isonly one of the many disciplines required tosuccessfully design a launch vehicle, if theaerodynamic design and analysis procedure canprovide accurate and reliable aerodynamiccharacteristics during all phases of the design cyclemore efficiently at less cost, there will be a positiveinfluence on the design process. Reduced risk,lower number of design cycles, more efficientdesign cycles, and ultimately, lower vehicle designcosts will result.

New entrepreneurial companies are being formedto design one- and two-stage-to-orbit reusablelaunch vehicles (RLV) and more traditionalexpendable launch vehicles (ELV) that use state-of-the-art and off-the-shelf technologies. Ofnecessity, these designs are unique and sometimesunusual, and most of these vehicles areconfigurations without a track record of successnor a backlog of experience in their flightcharacteristics.

The experienced designers and analysts wholearned from the successes and failures of the1960s and 1970s are leaving the industry as aresult of retirement and downsizing; thus much ofthis corporate memory and experience has thepotential to be lost.2 In addition, many of theyounger engineers who have experience with themodern computational methods do not haveexperience in evaluating the accuracy of the flowphysics provided by the advanced computational

2

fluid dynamics (CFD) methods. Without thenecessary mix of talent and experience, the newcompanies may find it difficult or impossible tobuild a launch vehicle able to perform the missionsneeded. One possible result is a number of newvehicles which simply cannot carry out the launchcapabilities promised. An even worse result is arepeat of past failures with the associatedincreased costs and lost confidence in the newvehicles. This latter result is not only bad for thecompany in question, it is bad for the launchindustry as a whole.

The aerodynamics design and analysis systemcalled LVX (Launch Vehicle eXpert) has beendeveloped to assist the launch vehicle designer inthe analysis and evaluation of the aerodynamics ofnew launch vehicles during all phases of the designcycle, but it has primary application duringconceptual and preliminary design.3 The objectiveis to increase engineering efficiency and reducedevelopment costs through a reduced number ofdesign cycles and more timely interaction withother disciplines needing aerodynamic information.Finally, LVX has as a primary goal to reduce riskin the aerodynamics of new launch vehicles bymaking past experience accessible for new designs.

BACKGROUND

The development of LVX is based in part onNEAR’s experience in the computational andexperimental aerodynamic design and analysis ofseveral launch vehicles. This work has beenconducted for new companies without a history ofsuccess in this area. For example, NEAR providedthe initial aerodynamics for the Orbital SciencesCorporation’s Pegasus4 and Taurus launchvehicles before Orbital had accumulatedexperience and established a history as asuccessful launch provider. NEAR also developedthe initial aerodynamics for the concept studies forthe early X-34 vehicle. In an extensive integratedcomputational and experimental program, NEARdid the aerodynamic design and analysis for theKistler Aerospace Corporation’s K-1 reusablelaunch vehicle.5,6 In a similar combinedcomputational and experimental aerodynamicsprogram, the preliminary aerodynamics database

was developed on the BA-1 and BA-2 expendablelaunch vehicles for Beal Aerospace.5 Currently,NEAR is providing analytical aerodynamicanalysis for the conceptual design of the KellySpace and Technology 2nd Generation ReusableLaunch Vehicle.

Concurrently with the above aerodynamics designand analysis tasks, NEAR was working in theinformation technology and knowledge-basedsystem area to identify efficient ways to developand search information databases for manydifferent technology areas. For example, NEAR’sCFDexchange was developed to capture and storeCFD code usage and procedures and match CFDtechnologies with engineering applications.7 Thisinvolved database design to store problemdescriptions, available technology information, andcalculation results, and more importantly, itinvolved the retrieval of information throughsophisticated search capabilities. This databasestores an archive of the configurations and flowfields solved, the CFD technologies used, thecomputational and labor resources required, andthe results for previous calculations. In additionalwork, a scheme was devised for inserting new CFDcalculations into the database automatically, andfor compiling data from related runs which couldthen be viewed within a multidimensional designspace.8

The goal for LVX is to combine expertise inlaunch vehicle aerodynamics with capabilities ininformation management to develop a practicaltool for the aerodynamicist to employ in the designanalysis of new launch vehicles. Useful and hard-to-find information is identified and collected in thesystem to create an online knowledge base ofaerodynamics of launch vehicles. Informationtechnologies are used to assist the system user inretrieving applicable knowledge that will help inthe design. The method can be used as a stand-alone application for the aerodynamics discipline,to be run off-line to analyze and evaluate a numberof potential configurations during conceptual andpreliminary design. The method also has thepotential to be used as the aerodynamics module ina multidisciplinary design method as illustrated inthe simplified sketch in Fig. 1.

3

TECHNICAL APPROACH

LVX is made up of a collection of searchabledatabases with information linked together acrossthese databases. As illustrated in Fig. 2, LVX hasthree major paths available to the user; (1) designassistance, (2) engineering analysis capability, and(3) database maintenance. Each of theseapplications has differing requirements.

In the design assistance area, the objective is tocollect information in the form of publisheddocuments, personal experiences, aerodynamicdata, and data on previous vehicles into asearchable database. Much of this information hasbeen stored in disparate locations, but LVX makesit easily available to users in a single location. Theengineering analysis capability permits the user toset up simplified geometries and perform first-order aerodynamic analyses. At the same time,links are made to pertinent information on similarconfigurations stored in the databases. Thedatabase maintenance is necessary to permit theuser to update the various databases as designinformation advances over time.

The following paragraphs detail the variouscomponents of LVX shown in Fig. 2.

Keywords

It was decided early in LVX development that ahierarchical structure would be used to model theknowledge stored in the database. Each node inthe hierarchy is represented by a keyword thatdescribes a topic. This model allows theinformation stored to be linked in a logical fashion,which assists in both the knowledge acquisitionand the knowledge retrieval mechanism. Thekeyword topics were chosen from several sources:NASA-assigned descriptors used to documentrelevant technical reports and papers, terms fromthe NASA Thesaurus, and additional terms thatwere identified from conversations withaerodynamics experts. This resulted in astandardized LVX-specific list of keywords.

LVX contains a database of relevant publications,and these typically are assigned keyword

descriptors by NASA. However, since LVX’skeywords are not always the same as those used byNASA indexers, it was necessary to translatemany of NASA’s terms into LVX’s keywords.

Aerodynamics Experience

LVX is built around an aerodynamic knowledgebase which includes design rules and engineeringrules-of-thumb for launch vehicles. Thisinformation has been obtained from a number ofsources including historical data reports andinterviews with active and retired launch vehicledesign experts from NASA and industry. Thedesign rules and comments from these interviewsare included in a database which includes links toother design information associated with specificconfigurations, and where possible, to publisheddocuments.

The interview process was conducted at severalNASA centers, and it was accomplished with bothindividual and group interviews. Each knowledgeacquisition session was taped and subsequentlytranscribed. The transcriptions were later editedfor accuracy, grammar, and completeness. Aknowledge engineer worked from the editedtranscripts to glean the individual pieces ofinformation or knowledge data. In the originaldatabase, no specific piece of information wasattributed to any specific individual in an effort tokeep the interviewees open and unselfconscious.

This information can be updated to include newand corrected information.

References Database

The system also includes a database which containreferences to published reports. This extensivedatabase contains more than 2,100 citations totechnical documents on launch vehicleaerodynamics and data. Keywords are ascribed toeach document, and these keywords are used in thesearch and retrieval process.

The documents were identified through a numberof sources, including (1) a list of relevant papersand reports collected by the authors in the course

4

of past work on launch vehicles, (2) citations fromsurvey papers and other work of prominentresearchers in the field, (3) a search for applicablereports in the NEAR Technical Library, and (4) asearch in the NASA Reconplus database usingselected keywords.

In an effort to make this database as useful andrelevant as possible, each citation was checked byan engineer to be sure only appropriate technicalinformation was included. When in doubt aboutthe quality of a citation, it was included in theinterest of completeness.

Geometry Description

In the design portion of LVX, the user can set up asimplified model of a new configuration using acomponent buildup method. As the user specifiesthe various components, nose shape, body shape,wing and/or tail geometry, and other items, thecode automatically assigns appropriate keywordsto a list. This list can be used to search for similardesigns and key references from the database. Theuser is free to add or delete the keywordsassociated with a particular geometry.

The user can also import a more detailed geometryfrom a geometry package outside LVX. This canbe used to compare with the simplified internalmodel, and it can provide some guidance as to howthe simple component model can be improved.

Analysis Methods

The system includes analysis codes of varyingfidelity that the user can call upon to analyze thedesign under consideration. Several codes can berun directly from the system, and the results areimmediately available for analysis and archivepurposes. Every aerodynamics code cannot beincluded in the system, but an interface can beprovided so that the user can run any desired codeoutside the boundaries of the system, and theresults can be imported into LVX for analysis andarchive purposes.

LVX can create input files for two codes, R13079

and Missile DATCOM10,11 and work is in progress

to add APAS12,13 to the list. R1307 is a linearaerodynamics, slender body theory method used tocalculate aerodynamic characteristics of wing-body-tail combinations. Missile DATCOM usesboth analytic and empirical methods to calculateaerodynamics of simple configurations. APAS isbased on linear potential theory at subsonic andsupersonic speeds, and at hypersonic speeds it usesimpact solutions. Both R1307 and MissileDATCOM are easy to set up and run, but they areconstrained to address conventional geometries,such as geometries with axisymmetric bodies andwings without winglets or camber. APAS canhandle arbitrary geometries but requires moredetail in the geometry input and added user effortin running the code.

The user creates the input for a code by specifyingthe geometry with components and setting the flowcondition and reference conditions. The user thensaves the input to disk, and the analysis program isrun external to the LVX system. The computedresults are brought back into the LVX system.The archiving of results in this mannerautomatically builds a history of the projectaerodynamic results.

Aerodynamics Database

The aerodynamic database can consist of bothexperimental data and computational analysisresults. Currently, LVX has experimental dataincluded for the X34 and Shuttle. The user mayexpand this database by adding data and results.When adding aerodynamic information to thesystem, the user tags the data with the appropriatedescriptor and content keywords. The descriptorkeywords describe or identify the data; forexample, supersonic X-34 CFD analysis. Thecontent keywords describe the characteristics ofthe data; for example, normal force and pitchingmoment vs angle of attack and Mach number.ASCII data files may be imported into the systemas either formatted or unformatted files. Theformatted input organizes the data into 3-Ddatabases delimited by zones, rows, and columns.An example would be an experimental data setorganized so the zones represent Mach numbersand the rows and columns represent angles of

5

attack and aerodynamic coefficients, respectively.With the data in this format, LVX can manipulateit. The only manipulation requirement currentlyplanned is for LVX to export the user requireddata in a spreadsheet format. This may beexpanded in future work. When the data areunformatted, the user may view it in LVX andexport the information unchanged.

The aerodynamic databases in the system can bemodified by the user to include both experimentaldata and computational analysis results as theybecome available during the configuration design;thus the system can be used to archiveaerodynamic results as the new design evolvesthrough the design process.

Design Iteration

A design iteration involves defining thecharacteristics of the proposed design, accessingexpert conclusions about the viability of thedesign, viewing comparable vehicles in thedatabase, and perhaps running an analysis code.The user may decide to iterate on the design toremedy potential problems and/or improve on theoriginal design. In some cases LVX may suggestcourses of action to improve on the design, or theexplanatory facility may indicate the source ofproblem areas. The user has the option to repeatthe design process with modifications to theoriginal design. The results of each designiteration can be stored in the database.

Help Features

Help features for the system currently include codeand keyword descriptions. The aerodynamics codedescriptions provide the user with guidance inselecting the proper code for an analysis. Thecodes are categorized by their analysis fidelity; forexample, engineering, intermediate, and CFD.Strengths and weaknesses, advice for applications,and code availability are given.

A manual and tutorial for the use of LVX will beincluded.

DISCUSSION

LVX has been distributed in a beta version forreview and comment. The following figures showscreen shots of this version of LVX.

Fig. 3 shows the search window. The keyword X-34 has been selected from the hierarchical keywordlist on the left. The text boxes at the bottom of thewindow are provided for title and/or authorsearches. The scope of this search includes allfour databases; References, Vehicles, Data, andFacts as indicated by the checked boxes at the top.

Figs. 4, 5, and 6 show the results of the search. InFig. 4, information in all four databases was foundin this search as indicated by the tabs at the toplabeled References, Vehicles, Data, and Facts. TheParameters tab is simply a list of the searchparameters. Fig. 5, the Vehicles tab, provides adescription of the vehicle with a link to aphotograph. Fig. 6, the Data tab, shows a list ofthe experimental datasets on the X-34 vehiclewhich are currently stored in LVX.

The user may save the results of a search to anASCII file, and the information may behyperlinked to related information.

Figs. 7 and 8 show the design window with asimple configuration with an ogive nose and astraked wing. The window’s lower left pane is usedfor the keywords associated with the geometry.Listed in the right window pane are the modes thatcan be selected to specify keywords; ReferenceConditions, Geometry, Flight Conditions, MissionCharacteristics, and Keywords. These multipleselection modes can be used simultaneously. Atthe top of the window, just below the menu bar,are the various components that can be selected tobuild a geometry. In the upper left window of Fig.8 is a list of the current geometry buildup, andselecting a component, such as the nose, willprovide an edit window for changing theconfiguration parameters or the nose type.

Similar windows are available for creating theinput files for the engineering analysis codes.These can be saved to disk for use during the

6

running of the code external to LVX. Two codes,R1307 and Missile DATCOM, are set up forexecution by LVX. For other codes, such asAPAS or other higher level codes, the user mustperform the action to execute the codes. Theresults can be imported back into LVX through thearchive interface provided by LVX.

LESSONS LEARNED

There were a number of lessons learned in eachpart of LVX development. Some were expected,but a number were unexpected.

Some of the interesting problems which occurredin the knowledge acquisition process involveddirecting the interviews and keeping thediscussions on topic. The interviews were tapedand transcribed for use in sorting out the relevantfacts. One of the interesting aspects of theinterview process was the difficulty in obtainingrelevant information from expert interviews. Oftenthe pieces of information were unconnected andincomplete, and some follow-up was necessary toinsure the accuracy of the data. It was alsonecessary to sort out proprietary information fromthe interviews.

Another problem was creating appropriate links toother information in the database. Rules and factsare included without attribution; however, thiscreates a problem evaluating the quality of theinformation. The user should have some means toassess the accuracy of the data. This approachmay need to change to include some sourceinformation attached to each item in the knowledgedatabase.

The generation of the technical documentsdatabase, although seemingly straightforward,provided a number of unanticipated problems.One search produced more than 2,600 citationswhich needed to be examined to eliminateinappropriate documents. It was during thisprocess that one of the major problems wasdiscovered, and that involved incorrect,inconsistent, and missing information in thecitations. In addition, it was surprising that thereference lists in a number of published reports

were found to have many errors and inaccuracies.In fact, in some cases it was not possible to locatecited references from the information provided inthe report/paper reference list; that is, truncatedtitles, misspelled and missing authors’ names, andincorrect dates.

Because of the various problems with the citations,it was necessary to make the corrections to theindividual citations by looking at them one at atime. This was a very labor intensive process andextremely time consuming. Unfortunately, noother approach could produce the quality ofdatabase required for LVX.

Similarly, for the aerodynamic analysis codesportion of LVX, it was necessary to be aware ofproprietary, licensing, and ITAR requirements.This is the reason that more codes are not includedand that provision was made to run codes outsideof LVX. The same problem occurred with theinclusion of experimental data.

CONCLUSIONS

A prototype system has been developed todemonstrate the aerodynamic design and analysisprocedure. The final LVX system described is stillin the testing phase, although most of the databasesand interfaces are complete.

The rules database generated from numerousinterviews with active and retired launch vehicleaerodynamicists is not complete, and this databasecan be expanded with current information as itbecomes available.

The LVX system is intended to captureaerodynamics design data, yet several factorsprevent the databases from being complete. Oneproblem is the issue of proprietary data. The otherissue is that collective knowledge on aerodynamicsdesign increases over time. For these two reasons,LVX has the capability for the users of the systemto add their own knowledge to the databases.Thus, users can add proprietary information thatpertains to their projects, and they can add newinformation as it is generated over time. Thisfacility allows LVX to remain current and relevant.

7

ACKNOWLEDGMENTS

The development of the LVX system has beenconducted under a Phase II SBIR contract from theNASA Marshall Space Flight Center. The authorsappreciate the constructive comments from thebeta users at the MSFC and the assistance of themany engineers who participated in the interviewand knowledge acquisition process to share theirknowledge and experiences from the United Statesspace program.

REFERENCES

1. Ryan, R. and Townsend, J.,“Fundamentals and Issues in LaunchVehicles Design,” AIAA 96-1194, Apr.1996.

2. Velocci, A. L., Jr., “Brain Drain ThreatensAerospace Vitality,” Aviation Week &Space Technology, Apr. 24, 2000, pp. 24-26.

3. Mendenhall, M. R., Whitlock, S. T., andRodman, L. C., “An IntegratedAerodynamic Design and Analysis Methodfor Advanced Launch Vehicles,” NEARTR 539, Nielsen Engineering & Research,Mountain View, CA, 1998.

4. Mendenhall, M. R., Lesieutre, D. J.,Caruso, S. C., Dillenius, M. F. E., andKuhn, G.D., “Aerodynamic Design ofPegasus - Concept to Flight with CFD,”AIAA 91-0190, Jan. 1991.

5. Mendenhall, M. R., Chou, H. S. Y., andLove, J. F., “Computational AerodynamicDesign and Analysis of Launch Vehicles,”AIAA 2000-0385, Jan. 2000.

6. Mendenhall, M. R., Chou, H. S. Y., andLove, J. F., “Aerodynamic Design andAnalysis of a Reusable Launch Vehicle,”ICAS 2000-0322, 22nd InternationalCongress of Aeronautical Sciences,Harrogate, United Kingdom, Aug. 2000.

7. Rodman, L. C., Tzou, K. T. S., andTreidler, E. B., "A CFD TechnologyExchange System for Code Validation andDocumentation," NEAR TR 540, NielsenEngineering & Research, Mountain View,CA, Nov. 1998.

8. Rodman, L. C., Reisenthel, P. H., andChilds, R. E., "An AutomatedDocumentation and Reporting Tool forAerospace Design Data," NEAR TR 555,Nielsen Engineering & Research,Mountain View, CA, Jun. 2000.

9. Pitts, W. C., Nielsen, J. N., and Kaattari,G. E., “Lift and Center of Pressure ofWing-Body-Tail Combinations atSubsonic, Transonic, and SupersonicSpeeds,” NACA Report 1307, 1957.

10. Vukelich, S. R., Stoy, S. L., Burns, K. A.,Castillo, J. A., and Moore, M. E., “MissileDATCOM Volume I - Final Report,”AFWAL-TR-86-3091, Dec. 1988.

11. Blake, W. B., “Missile DATCOM, User’sManual - 1997 Fortran 90 Revision,” WL-TR-1998-3009, Feb. 1998.

12. Bonner, E., Clever, W., and Dunn, K.,“Aerodynamic Preliminary AnalysisSystem II, Part I - Theory,” NASA CR182076, Apr. 1991.

13. Sova, G. and Divan, P., “AerodynamicPreliminary Analysis System II, Part II -User’s Manual,” NASA CR 182077, Apr.1991.

8

Fig. 1.- Sample integrated design process.

Typical User OptionsDesign Assistance

GeometryRef. & FlightConditionsKeywords

Submit Design to Expert

OutputFactsReferencesVehiclesData

Help Utility

Code HelpKeyword GlossaryAbout...

EngineeringAnalysis Options

Build Input File

Run Code

Import

Advanced User OptionsDatabase Maintenance

AddKeywords

AddVehicles

AddReferences

InputData

Search Options

References

Vehicles

Data

FactsDatabase

AddFacts

Input Design:

Fig. 2.- Launch Vehicle eXpert System

9

Fig. 3.- LVX search window. Fig. 4.- Search results; References.

Fig. 5.- Search results; Vehicle Description. Fig. 6.- Search results; Experimental Database.

10

Fig. 7.- Design window.

Fig. 8.- Design window; configuration components.