Category AERONAUTICS

Prior to the X-15 flight-test program, there were several theories pre­dicting the amount of friction heat that would transfer to the surface of a winged aircraft, with substantial differences in the theories. Wind tunnels, ballistic ranges, and high-temperature facilities such as arc-jets were unable to adequately duplicate the flight environment necessitating

full-scale flight test to determine which theory was correct. The X-15 was that test aircraft. The design needed to be robust in order to survive the worst-case heating predictions if the theories proved to be correct.

A heat-sink structure was selected as the safest and simplest method for providing thermal protection. Inconel X, a nickel alloy normally used for jet engine exhaust pipes, was selected as the primary structural mate­rial. It maintained adequate structural strength to about 1,200 °F. The design proceeded by first defining the size of each structural member based on the air loads anticipated during entry, then increasing the size of each member to absorb the expected heat load that would occur dur­ing the short exposure time of an X-15 flight.

As with most of the first missile and aircraft explorations, early hyper­sonic flights in the X-15 showed that none of the prediction methods was completely accurate, although each method showed some validity in a cer­tain Mach range. In general, the measured heat transfer was less than pre­dicted. Thus, one of the most significant flight-test results from the X-15 program was development of more accurate prediction methods based upon real-world data for the thermal protection of future hypersonic and entry vehicles.[750] The majority of aerodynamic heating issues that required attention during the X-15 flight-test program were associated with local­ized heating: that is, unexpected hot spots that required modification. Some typical examples included loss of cockpit pressurization because of a burned canopy seal (resolved by installing a protective shield in front of the canopy gap), cockpit glass cracked because of deformation of the glass retainer ring (resolved by increasing the clearance around the glass), wing skin buckling behind the slot in the leading edge expansion joint (resolved by installing a thin cover over the expansion joint), thermal expansion of the fuselage triggering nose gear door deployment with resulting damage to internal instrumentation (resolved by increasing the slack in the deploy­ment cable), and buckling of skin on side tunnel fairings because of large temperature difference between outer skin and liquid oxygen (LOX) tank (resolved by adding expansion joints along the side tunnels).

Most of these issues were discovered and resolved fairly easily since the flight envelope was expanded gradually on successive flights with small increases in Mach number on each flight. Had the airplane been exposed to the design entry environment on its very first flight, the

combined results of these local heating problems would probably have been catastrophic.

Not only has NASA played a strong role in the development of new CFD algorithms, it has delivered these contributions to the technical pub­lic in the form of highly developed computer codes for the user. In the context of this survey, it would be remiss not to underscore the impor­tance of these codes, three in particular, which this author (and his stu­dents) have used as numerical tools for carrying out research: LAURA, OVERFLOW, and CFL3D.

The LAURA code was developed principally by Dr. Peter Gnoffo at the NASA Langley Research Laboratory.[783] This code solves the three­dimensional Euler or Navier-Stokes equations for high-speed super­sonic and hypersonic flow fields. It is particularly noteworthy because
it deals with very detailed nonequilibrium and equilibrium chemically reacting flows pertaining to hypersonic reentry vehicles in Earth’s and foreign planetary atmospheres. Some applications involve flow-field temperatures so high that radiation becomes a dominant physical fea­ture. The LAURA program readily handles radiative gas dynamics, and, to this author’s knowledge, it is the only existing standard code to do so. The LAURA code has been used for the design and analysis of all NASA entry bodies in recent experience and is the most powerful and useful code in existence for high-temperature flow fields.

Of particular use for computing lower speed subsonic and transonic flows is OVERFLOW. This code was developed in the early 1990s by Pieter Burning and Dennis Jesperson as a collaborative effort between NASA Johnson Space Center and the NASA Ames Research Center. It solves the compressible three-dimensional Reynolds-averaged Navier – Stokes equations by means of a time-marching algorithm. OVERFLOW is widely used for the calculation of three-dimensional subsonic and tran­sonic flows, and it proved particularly valuable for computing subsonic viscous flows over airfoils in a recent graduate study of innovative new airfoil shapes for high lift undertaken at the University of Maryland at College Park’s Department of Aerospace Engineering.[784]

In the mid-1980s, Dr. Jim Thomas and his colleagues at the NASA Langley Research Center recognized the need for a code that contained the latest advancements in CFD methodology being developed by the applied mathematics community. Out of their interest sprang CFL3D, one of the earliest (yet still most powerful) CFD codes developed by NASA.[785]

This code is applicable across the whole flight spectrum, from low-speed subsonic flow to hypersonic flow. Not only does it handle steady flows, but it calculates time-accurate unsteady flows as well. Much effort was invested in the development of detailed grids so that it readily handles flows over complex three-dimensional bodies. An appreciation of the power, usefulness, and widespread acceptance of CFL3D can be gained by noting that it is used by over 100 researchers in 22 companies, 13 universities, NASA, and the military services.

LAURA, OVERFLOW, and CFL3D are just three of the CFD codes NASA researchers have generated. Most importantly, because they are the product of taxpayer-supported research, all are readily available, free of charge, to the general public, making NASA unique among other organi­zations working in the field of CFD. NASA’s commitment to making sci­entific and technical information of the highest quality available to the public—a legacy of its predecessor, the National Advisory Committee for Aeronautics—has influenced its approach to CFD code development and may be counted one of the Agency’s most valuable contributions to the whole discipline of computational fluid dynamics. When students and professional practitioners alike need viable computer codes for complex fluid dynamic applications, they have ready access to such codes and the extremely competent individuals who develop them. This is perhaps the highest accolade one can pronounce upon NASA’s computational fluid dynamics efforts.

In closing, a proper history of CFD would require a lengthy book and a greater perspective of the past: something yet impossible, for the history of this rather young discipline is still evolving. The challenge is akin to what one might have expected trying to write a history of the balloon in the early 1800s, or a history of flight in 1914. In this case, I have tried to share my perspective in an accessible format, based in part on my own experiences and on my familiarity with the work of many colleagues, especially those within NASA. I have had to leave out so many others and so much great work in CFD just to tell a short story in a limited amount of pages that I feel compelled to apologize to those many others that I have not men­tioned. To them I would say that their absence from this case certainly does not mean their contributions were any less important. But this has been an effort to paint a broad-stroke picture, and, like any such picture, it is somewhat subjective. My best wishes go out to all those researchers, pres­ent and future, who have and will continue to make computational fluid dynamics a vital, essential, and lasting tool for the study of fluid dynamics.

Langley was the first NACA laboratory, established in 1917. As such, it is the oldest and most distinguished of NASA aeronautics Centers, with a pedigree that dates to meetings held prior to the First World War to determine the future aeronautical laboratory structure of the Nation. Since the earliest days of American aviation, Langley has constantly anticipated, reacted, and adapted as necessary to meet the Nation’s aeronautical research needs, reflecting its broad technical capabilities and expertise in areas such as aerodynamics, aircraft and spacecraft structures, flight dynamics, crew systems, space environmental phys­ics, and life sciences.

Among the very earliest NACA technical reports were several con­cerning loads calculation and structural analysis, some of which are cited in the introduction to this paper. These papers, and others that fol­lowed throughout the era of the NACA, were widely used in the aircraft industry. By the time NASA was founded, Langley had become a major Center for all forms of aeronautics research, engineering, and analysis.

Through the 1980s and1990s, Langley had approximately 150 tech­nical professionals in the structural disciplines (not including Materials), covering both aircraft and spacecraft applications. This work was orga­nized primarily in two divisions, Structural Mechanics (static prob­lems) and Structural Dynamics, plus a separate Optimization Methods group of approximately 15 members.[908] Structural Mechanics included Composites, Computational Structural Mechanics, Thermal Structures, Structural Concepts, and AeroThermal Loads.[909] Structural Dynamics included Aeroelasticity, Unsteady Aerodynamics, Aeroservoelasticity, Landing and Impact Dynamics, Spacecraft Dynamics, and Interdisciplinary Research.[910] (Reorganizations sometimes changed the specific delineation of responsibilities.) Langley researchers pur-

sued many separate computational structural analysis studies and efforts, but overall, the Center was particularly (and intimately) involved with NASTRAN, the Design Analysis Methods for Vibration (DAMVIBS) rotorcraft structural dynamics modeling program, and efforts at integration and optimization.

After NASTRAN was developed during the period from 1965 to 1970, management of it was transferred from Goddard to Langley. Accordingly, a major emphasis at Langley through the 1970s was the maintenance and continuing improvement of NASTRAN. The first four Users’ Colloquia were held at Langley. While COSMIC handled the administrative aspects of NASTRAN distribution, the NSMO was responsible for technical management and coordinating NASTRAN development efforts across all Centers and many contractors. The program itself is discussed in greater detail elsewhere in this case.

The DAMVIBS research program, conducted from 1984 to 1991, reflected Langley’s long-standing heritage of research on rotorcraft struc­tural dynamics. DAMVIBS achieved concrete advances in the industry state of the art in helicopter structural dynamic modeling, analysis-to-test matching, and, perhaps most importantly, acceptance of and confidence in modeling as a useful tool in designing helicopter rotor-airframe systems for low vibration. Key NASA program personnel were William C. Walton, Jr., who spearheaded program concept and initial direction (he retired in 1984); Raymond G. Kvaternik, who furnished program direction after 1984; and Eugene C. Naumann, who supplied critical technical guidance. The industry participants were Bell Helicopter Textron, Boeing Helicopters, McDonnell-Douglas Helicopter Company, and Sikorsky Aircraft. The participants developed rotor-airframe finite element models, conducted ground vibration tests, made test/analysis comparisons, improved their models, and conducted further study into the "difficult components” that current state of the art rotorcraft analysis could not adequately model.[911]

Modeling "guides”—documented procedures—were identified from the start as a key element to the program:

This program emphasized the planning of the modeling. . .

the NASA Technical Monitor insisted on a well thought out

plan of attack, accompanied by detailed preplanned instruc­tions. . . . The plan was reviewed by other industry representa­tives prior to undertaking the actual modeling. Another unique feature was that at the end of the modeling, deviations from the planned guides due to cause were reported.[912]

All of the participants reported that finite element modeling could predict vibrations more accurately than previously realized but required more attention to detail in the modeling, with finer meshes and the inclu­sion of secondary components not normally modeled for static strength and stiffness analysis. The participants further reported on specific improvements to dynamic modeling practice resulting from the exer­cise and on the increased use and acceptance of such modeling in the design phase at each respective company.[913] As a result of DAMVIBS:

• Bell and Boeing incorporated DAMVIBS lessons into the modeling of their respective portions of the V-22.[914]

• Boeing made improvements the NASTRAN dynamic model of the CH47D, which was still in production, achieving greatly improved correlation to test data. Boeing

credited Eugene Naumann of Langley with identifying many of the needed changes.[915]

• McDonnell-Douglas improved its dynamic models of exist­ing and newly developed products, achieving improved correlation with test results.[916]

• Sikorsky developed an FEM model of the UH60A air­frame "having a marked improvement in vibration-pre-

8 dicting ability.”[917]

• Sikorsky also developed a new program (PAREDYM, programmed in NASTRAN DMAP language) that could automatically adjust an FEM model so that its modal characteristics would match test values.[918] PAREDYM then found use as a design tool: having the ability to modify a model of an existing design to better match test data, it also had the ability to modify a model of a new design not yet tested, to a set of desired modal char­acteristics. Designers could now specify a target (low) level of vibration response and let PAREDYM tune its model—essentially designing the airframe—to meet the goal. (The improvements would not be "free,” however, as the program could add weight in the process.) After discovering this usage mode, the developers then added facilities for minimizing the weight impact to achieve a desired level of vibration improvement.[919]

DAMVIBS ended in 1991, though this did not mark an end to Langley’s work on rotorcraft structural dynamics.[920] Rather, it reflected a shift in emphasis away from the traditional helicopter to other aeronautics and

astronautics research ventures as well.[921] As basic analysis capability had become relatively mature by around 1990, attention turned toward the integration of design, analysis, and optimization; to the integration of structural analysis with other disciplines; and to nondeterministic meth­ods and the modeling of uncertainty.[922] Projects included further work on rotorcraft, aircraft aerostructural optimization, control-structural optimi­zation for space structures, and nondeterministic or "fuzzy” structures, to name a few.[923] Many optimization projects at Langley used the CONMIN constrained function minimization program, developed at Ames, as the optimization driver, interfaced with various discipline-specific analysis codes developed at Langley or elsewhere.

In the 1970s, NASA Langley began what would prove to be some very significant multidisciplinary optimization (MDO) studies. Jaroslaw Sobieszczanski-Sobieski pioneered the Bi-Level Integrated System Synthesis (BLISS), a general approach that is applicable to design opti­mization in any set of disciplines and of any system, aircraft, or otherwise. His work at Langley, spanning from the 1970s to the present, is recognized throughout the aerospace industry and the MDO community. BLISS and related methods constitute one of the major classes of MDO techniques in widespread use today. Some of the early work on BLISS was concerned with improving the structural design process and addressing aerodynamic and structural problems concurrently. For example, in the late 1970s, Sobieszczanski-Sobieski developed methods for designing metal and/or composite wing structures of supersonic transports for minimum weight, including the effect of structural deformations on aeroelastic loads.[924]

This Langley work continued into the 1980s, when Langley research­ers moved forward to apply the knowledge gained with BLISS to space­craft, generating two other systems: the Integrated Design and Evaluation of Advanced Spacecraft (IDEAS) and Programming Structural Synthesis (PROSS). IDEAS did not perform optimization per se, but it did pro­vide integration of design with analysis in multiple disciplines, includ-

ing structures and structural dynamics.[925] PROSS combined the Ames CONMIN optimizer with the SPAR structural analysis program (developed at NASA Lewis). PROSS was publicly released in 1983.[926] Several subsequent releases incorporated either new optimization strategies and/or improved finite element analysis.[927]

One of these was ST-SIZE, which started as a hypersonic vehicle structural-thermal design code. In 1996, Collier Research Corporation obtained an exclusive license from Langley for the ST-SIZE program. Under a new model for NASA technology transfer, Collier agreed to pay NASA royalties from sales of Collier’s commercialized version of the code. This version, called HyperSizer (trademark of Collier Research Corporation), was intended to be applicable to a wide variety of uses, including office design and construction, marine systems, cargo contain­ers, aircraft, and railcars. The program performed design, weight buildup, system-level performance assessments, structural analysis, and struc­tural design optimization.[928] In 2003, Spinoff reported that this model had worked well and that Collier and NASA were still working together to enhance the program, specifically by incorporating further analysis codes from NASA Glenn Research Center: Micromechanics Analysis Code with Generalized Method Cells (MAC/GMC) and higher-order theory for functionally graded materials (HOTGFM). Both of these were developed collaboratively between Glenn, University of Virginia, Ohio Aerospace Institute, and Tel Aviv University.[929]

JPL operates both ground-based and space-borne antennas. These pose their own array (no pun intended) of structural challenges because of their large size, the need to maintain precision alignment, and, in the case of space-borne antennas, the need for extremely light weight. JPL was one of the early users of NASTRAN, using it to calculate gravity defor-

mation effects on ground-based 210-foot-diameter antennas in 1971.[993] In 1976, JPL developed a simplified stiffness formulation (translational degrees of freedom only at the nodes) coupled with a structural member sizing capability for design optimization: "Computation times to exe­cute several design/analysis cycles are comparable to the times required by general-purpose programs for a single analysis cycle.”[994] In the late 1980s, JPL upgraded a set of 64-meter antennas to a new diameter of 70 meters. This project afforded "the rare opportunity to collect field data to compare with predictions of the finite-element analytical models.” Static and dynamic tests were performed while the antenna structures were in a stripped-down configuration during the retrofit process. The data provided insight into the accuracy of the models that were used to optimize the original structural designs.[995]

3) TRASYS Radiative Heat Transfer (Johnson, 1980s-1990s)

TRASYS was developed by Martin Marietta for Johnson to calculate inter­node radiative heat transfer as well as heat transfer to a model from the surroundings. It was used extensively through the 1990s. Applications included propulsion analysis at Glenn Research Center and Structural – Thermal-Optical analysis (when integrated with NASTRAN for structural calculations and MACOS/IMOS and POPOS optical codes) at Marshall Space Center and JPL.[996] BUCKY was developed in-house. BUCKY was initially introduced as a basic plane stress and plate buckling program but was extensively developed during the 1990s to include plate bend­ing, varying element thickness, varying edge and pressure loads, edge moments, plasticity, output formatting for visualization in I-DEAS (a CAD program developed by Structural Dynamics Research Corporation), three-dimensional axisymmetric capability, and improvements in execution time.[997]

Early aviation pioneers gradually came to realize that an effective flight control system was necessary to control the forces and moments act­ing on an aircraft. The creation of such a system of flight control was one of the great accomplishments of the Wright brothers, who used a
combination of elevator, rudder, and wing warping to achieve effective three-axis flight control in their 1903 Flyer. As aircraft became larger and faster, wing warping was replaced by movable ailerons to control motion around the roll axis. This basic three-axis flight control system is still used today. It enables the pilot to maneuver the aircraft about its pitch, yaw, and roll axes and, in conjunction with engine power adjust­ments, to control velocity and acceleration as well. For many genera­tions after the dawn of manned, controllable powered flight, a system of direct mechanical linkages between the pilot’s cockpit controls and the aircraft’s control surfaces was used to both assist in stabilizing the aircraft as well as to change its flight path or maneuver. The pilot’s abil­ity to "feel” the forces being transmitted to his flight controls, especially during rapid maneuvering, was critically important, because many early aircraft were statically unstable in pitch with the pilot having to exert a constant stabilizing influence with his elevator control. Well into the First World War, many aircraft on both sides of the conflict had poor stability and control characteristics, issues that would continue to chal­lenge aircraft designers well into the jet age. Wartime experience showed that adequate stability and positive aircraft handling qualities, coupled with high performance (as exemplified by parameters such as low wing loading, high power-to-weight ratio, and good speed, turn, and climb rates) played a major role in success in combat between fighters.

By the Second World War, electrically operated trim tabs located on aerodynamic control surfaces and other applications of electrical con­trol and actuation were emerging.[1101] However, as aircraft performance increased and airframes grew larger and heavier, it became increas­ingly harder for pilots to maneuver their aircraft because of high aero-

dynamic forces on the control surfaces. World War II piston engine fighters were extremely difficult to maneuver in pitch and roll and often became uncontrollable as compressibility effects were encountered as speeds approached about Mach 0.8. The introduction of jet propulsion toward the later stages of the Second World War further exacerbated this controllability problem. Hydraulically actuated fight control sur­faces were introduced to assist the pilot in moving the control surfaces at higher speeds. These "boosted” control surface actuators were con­nected to the pilot’s flight controls through a system consisting of cables, pulleys, and cranks, with hydraulic lines now also being routed through the airframe to power the control surface actuators.[1102]

With a boosted flight control system, the pilot’s movements of the cockpit flight controls opens or closes servo valves in the hydraulic sys­tem, increasing or decreasing the hydraulic pressure powering the actu­ators that move the aircraft control surfaces. Initially, hydraulic boost augmented the force transmitted to the control surfaces by the pilot; such an approach is referred to as partial boost. However, fully boosted flight controls quickly became standard on larger aircraft as well as on those aircraft requiring high maneuverability at high indicated airspeeds and Mach numbers. The first operational U. S. jet fighter, the Lockheed P-80 Shooting Star, used electric pitch and roll trim and had hydraulically boosted ailerons to provide roll effectiveness at higher airspeeds. The first jet-powered U. S. bomber to enter production, the four-engine North American B-45, flew for the first time in March 1947. It had hydraulically boosted flight control surfaces and an electrically actuated trim tab on the elevator that was used to maintain longitudinal trim. Despite their undeniable benefits, boosted flight control systems could also produce unanticipated hazards. The chief test pilot for the Langley Aeronautical Laboratory of what was then the National Advisory Committee for Aeronautics (NACA), Herbert "Herb” Hoover, was killed in the crash of a B-45 on August 14, 1952, when the aircraft disintegrated during a test mission near Barrowsville, VA.[1103]

As a NASA report noted: "The aerodynamic power of the trim-tab – elevator combination [on the B-45] was so great that, in the event of an inadvertent maximum tab deflection, the pilot’s strength was insuf­ficient to overcome the resulting large elevator hinge moments if the hydraulic boost system failed or was turned off. Total in-flight destruc­tion of at least one B-45, the aircraft operated by NACA, was probably caused by this combination of circumstances that resulted in a normal load factor far greater than the design value.”[1104]

The Air Force/NACA Bell XS-1 rocket-powered research aircraft was equipped with an electrically trimmed adjustable horizontal stabilizer. It enabled the pilot to maintain pitch control as the conventional eleva­tor lost effectiveness as the speed of sound was approached.[1105] Using this capability, U. S. Air Force (USAF) Capt. Chuck Yeager exceeded Mach 1 in level flight in the XS-1 in October 1947, followed soon after by a North American XP-86 Sabre (although in a dive). Hydraulically boosted controls and fully movable horizontal tails were rapidly implemented on operational high-performance jet aircraft, an early example being the North American F-100.[1106] To compensate for the loss of natural feed­back to the pilot with fully boosted flight controls, various devices such as springs and bob weights were integrated into the flight control sys­tem. These "artificial feel” devices provided force feedback to the pilot’s controls that was proportional to changes in airspeed and acceleration. Industry efforts to develop boosted fight control surfaces directly ben­efited from NACA flight-test efforts of the immediate postwar period.

The Air Force Boeing YC-14 Short Take-Off and Landing (STOL) jet transport technology demonstrator flew for the first time on August 9, 1976, from Boeing Field in Seattle, WA, during the period between Phase I and Phase II of the NASA DFBW F-8 program. Two proto­types were built with the second aircraft flying in October 1976. The YC-14 is noteworthy in that it was the first aircraft to fly with a fault – tolerant multichannel redundant digital fly-by-wire flight control system.

A mechanical backup flight control capability was retained. The full authority triply redundant digital fly-by-wire flight control system, designed by the British Marconi Company, performed computational commands for pitch, roll, and yaw that were used to control the eleva­tor, aileron, and rudder actuation systems. The reconfigurable computer architecture divided the basic control path into three subfunctional ele­ments with these elements replicated to provide fault tolerance. The inter­nal element redundancy management function was intended to detect and isolate faulty elements and perform the necessary reconfiguration. The input signal selection methodology was intended to guarantee that all three computers used the same numbers and thus produced identi­cal output values. During normal operation, the overall system output value was selected as the midvalue of the three individual values. The system would continue to operate in the event of a failure of one com­puter by taking the average of the output of the two remaining comput­ers. If they disagreed, both were disabled and the aircraft reverted to the backup manual control system.[1179]

The YC-14 was also noteworthy in that it used optical data links to exchange data between the triply redundant computers. The optical com­munications medium was chosen to eliminate electromagnetic inter­ference effects, electrical grounding loop problems, and the potential propagation of electrical malfunctions between channels. Optical cou­pling was used to maintain interchannel integrity. Each sensor’s out­put was coupled to the other channels so that each computer had data from each of the other sensors. Identical algorithms in each computer were used. They consolidated the data, enabling equalization and fault detection/isolation of the inputs. The computers were synchronized to avoid sampling time differences and to assure that all computers were receiving identical data inputs.[1180]

An important observation involving redundant computer-controlled fly-by-wire flight control systems was derived from the YC-14 flight-test experience. As noted above, the system was designed to ensure that all computers used the same sensor input values and should therefore produce identical outputs. However, a significant fault in the digital
flight control software was encountered during flight-testing that had not been detected during ground laboratory testing. The software fault resulted in incorrect tracking of control law computations in each of the three flight control channels, with each channel performing signal selections on a different set of values. This resulted in different input data for the three channels. Although the discrepancies between each channel’s inputs were small, the cumulative effect led to large tracking errors between flight control channels when airborne.[1181]

Following cancellation of the Air Force YC-14A program in 1979, the two prototypes were placed in storage at Air Force’s Aerospace Maintenance and Regeneration Group (AMARC) at Davis-Monthan AFB, AZ, in April 1980. The first prototype is now displayed at the Pima Air and Space Museum in Tucson, AZ.

The X-31 was the first international experimental aircraft development program in which the U. S. participated. Two X-31 Enhanced Fighter Maneuverability (EFM) demonstrator aircraft were designed and con­structed by Rockwell International Corporation’s North American Aircraft Division and Deutsche Aerospace. Assigned U. S. Navy bureau Nos. 164584 and 164585, the aircraft would be used to obtain data that could be applied to the design of highly maneuverable next-generation fighters. During the conceptual phase of the program, the personnel examined the application of EFM technologies and defined the require­ments for the demonstrator aircraft. Next, the preliminary design of the demonstrator and the manufacturing approach were defined. Technical experts from the U. S. Navy, German Federal Ministry of Defense, and

NASA evaluated all aspects of the design. Detail design and fabrication followed, with the two aircraft being assembled at the Rockwell International (now Boeing) facility at Palmdale, CA. Both aircraft were required to fly a limited flight-test program at Rockwell. The first aircraft flew its first flight on October 11, 1990, piloted by Rockwell chief test pilot Ken Dyson. The second aircraft made its first flight on January 19, 1991, with Deutsche Aerospace chief test pilot Dietrich Seeck at the controls.[1235]

The X-31 had a digital fly-by-wire flight control system that included four digital flight control computers with no analog or mechanical backup. Three synchronous main computers drove the flight control surfaces. The fourth computer served as a tiebreaker in case the three main computers produced conflicting commands. Three thrust vectoring paddles were mounted on the X-31’s aft fuselage adjacent to the engine nozzle. Directed by the DFBW flight control system, the paddles were moved in and out of the exhaust flow with the resultant thrust vectoring augmenting the aerodynamic control surfaces in pitch and yaw control to improve maneuverability. Made of an advanced carbon-fiber-rein­forced composite material, the paddles could sustain temperatures of up to 1,500 degrees Celsius. The X-31 also had movable forward canards for pitch control. As a result of controllability issues identified during the X-31 flight-test program, fixed strakes between the trailing edge of the wing and the engine exhaust were incorporated. They provided additional nose-down pitch control at very high angles of attack. Another fix that was found neces­sary was the addition of small fixed-nose strakes to help control sideslip.[1236]

During flight-test operations at the Rockwell Aerospace facility, the two X-31s flew 108 test missions, validating the use of thrust vectoring to compensate for loss of aerodynamic control at high angles of attack and expanding the poststall envelope up to 40 degrees angle of attack. The poststall envelope refers to the region in which the aircraft dem­onstrated an ability to maintain controlled flight beyond the normal X-31 stall angle of attack of 30 degrees. X-31 flight operations moved to NASA Dryden in February 1992, with the first flight under International Test Organization (ITO) management occurring in April 1992. The ITO initially included about 110 people from NASA, the U. S. Navy, the U. S. Air Force, Rockwell Aerospace, the Federal Republic of Germany, and

Daimler-Benz. The ITO staff was eventually reduced to approximately 60 people. Overall management of the X-31 program came under by the Defense Advanced Research Projects Agency, with NASA responsi­ble for flight-test operations, aircraft maintenance, and research engi­neering after the project moved to Dryden. The ITO director and NASA’s X-31 project manager at Dryden was Gary Trippensee. Pilots included NASA pilot Rogers Smith, U. S. Navy Cdr. Al Groves, German pilots Karl Lang and Dietrich Seeck, Rockwell International pilot Fred Knox, and Air Force Flight Test Center pilot Lt. Col. Jim Wisneski. By July 1992 the X-31 flight envelope was being expanded in preparation for mili­tary utility evaluations that would fly the aircraft against nonthrust vec­tored fighters to evaluate effectiveness in simulated air combat. Thrust vectoring effectiveness at supersonic speed was evaluated out to Mach

1. 28 at an altitude of 35,000 feet.

In early flight-testing, the X-31 flight control system went into a reversionary mode four times in the first nine flights because of dis­agreement between the two air data sources.[1237] The X-31 was very sen­sitive to sideslip. This caused difficulties for the flight control system at higher angles of attack. Below 30 degrees, the nose boom updated the inertial navigation unit with air data. Above angles of attack of 30 degrees, the inertial navigation unit began calculating erroneous side­slip angles as a result of changes in the relative wind vector. To resolve this problem, a so-called Kiel probe replaced the standard NASA Pitot tube to calculate airflow. The Kiel probe was bent 10 degrees downward from the standard pitot configuration. In addition, the sideslip vane was rotated downward 20 degrees relative to the nose boom to compensate for a yawing oscillation that occurred at an angle of attack of 62 degrees. These changes resulted in accurate air data being provided to the iner­tial navigation unit throughout the X-31 flight envelope with false side­slip readings at high angles of attack eliminated.[1238]

Throughout the process of envelope expansion, many modifications to the flight control laws were required because actual aerodynamics of the aircraft were somewhat different from wind tunnel predictions. When the pilots started flying at angles of attack above 50 degrees, they
encountered erratic lateral lurching movements. In an attempt to coun­ter this phenomenon, narrow, 1/4-inch-wide strips of grit were attached to the sides of the nose boom and the radome. These effectively changed the vortex flow across the forward fuselage of the aircraft, reducing the randomness of the lurches and enabling expansion of the flight envelope to the design angle of attack limit of 70 degrees at 1 g. However, pilots encountered unintentional departures from controlled flight as the air­craft approached poststall angles of attack of 60 degrees during Split-S maneuvers.[1239] The asymmetric yawing moment encountered during this maneuver was beyond the capability of the thrust vectoring system to maintain adequate control.[1240] Testing in the Langley full-scale wind tunnel resulted in nose strakes and a modified slightly blunter nose tip design that were fitted to the two aircraft, allowing resumption of the flight-test program. The nose strakes were 6/10 of an inch wide and 20 inches long and forced more symmetric transition of forebody vortexes. The blunted nose tip reduced yaw asymmetries.[1241]

Poststall pitch control effectiveness, especially with the X-31 center of gravity at the aft allowable design location, was initially marginal.[1242] In these high-angle-of-attack conditions, test pilots rated aircraft response as unsatisfactory. NASA Langley conducted wind tunnel tests of various approaches intended to provide increased nose-down pitch control at high angles of attack. Sixteen different modifications were rapidly tested in the full-scale wind tunnel, with Langley recommending that a pair of strakes 6 inches wide and 65 inches long be mounted along the sides of the aft fuselage to assist in nose-down recovery. These were incorporated on the X-31, with subsequent flight-testing confirming greatly improved nose-down pitch control.[1243] Positive control at 70 degrees angle of attack with a controlled roll around the aircraft velocity vector was demonstrated November 6, 1992. On April 29, 1993, a minimum radius 180-degree post-

stall "Herbst Maneuver” was accomplished for the first time.[1244] During the final phase of evaluation, the X-31s engaged in simulated air combat sce­narios against F/A-18s. During these scenarios, the X-31s were able to outmaneuver the F/A-18s purely through use of poststall maneuvers and without use of thrust vectoring. X-31 test pilots did not support trading off basic fighter characteristics to acquire poststall maneuvering capa­bilities but concluded that improved pitch pointing and velocity vector maneuvering possible with thrust vector control did provide additional options during close-in combat. Thrust vectoring, combined with fully controllable poststall maneuvering, enabled X-31 pilots to position their aircraft in ways that adversary pilots could not counter, but it had to be used selectively and rapidly to be effective.[1245]

In 1994, software was installed in the X-31 to simulate the feasi­bility of stabilizing a tailless aircraft at both subsonic and supersonic speed using thrust vectoring. The aircraft was modified to enable the pilot to destabilize the aircraft with the rudder to lower stability levels to those that would have been encountered if the aircraft had a reduced – size vertical tail. For this purpose, the rudder control surface was used to cancel the stabilizing effects of the vertical tail, and yaw thrust vec­tor commands were applied by the flight control system to restabilize and control the aircraft. The X-31 was flown in the quasi-tailless mode supersonically at 38,000 feet at Mach 1.2, and maneuvers involving roll and yaw doublets, 30-degree bank-to-bank rolls, and windup turns to 2 g were flown. During subsonic testing, simulated precision carrier land­ing approaches and ground attack profiles were successfully evaluated. The quasi-tailless flight-test experiment demonstrated the feasibility of tailless and reduced-tail highly maneuverable fighter/attack aircraft designs. Such designs could have reduced drag and lower weight as well as reduced radar and visual detectability. It determined that thrust vec­toring is a viable flight control effector that can replace the functions provided by a vertical tail and rudder control surface. Potential disad-

vantages include the added weight, complexity, and reliability issues associated with a thrust vectoring system. Additionally, flight condi­tions that require lower engine thrust settings (such as approach and landing) may necessitate provision of additional aerodynamic high – drag devices to enable high-thrust settings to be maintained, ensuring adequate thrust vectoring control. Early integration of such consider­ations into the overall design process, along with an increased level of interaction between propulsion and flight control systems, is required in order to derive the maximum benefit from reduced or tailless aircraft that rely on thrust vectoring for stability and control.[1246]

The No. 1 X-31 aircraft was lost on its 292nd flight on January 19, 1995. German test pilot Karl Lang had just finished a series of test maneu­vers and was in the process of recovering back to a landing at Edwards. At an altitude of 20,000 feet, he observed discrepancies in the air data dis­plays along with a master caution light. The aircraft then began a series of diverging pitch oscillations and became uncontrollable. Lang ejected safely at an altitude of 18,000 feet, and the aircraft crashed in an unpop­ulated desert area just north of Edwards. The crash was determined to have resulted from an unanticipated single-point failure in the nose – mounted Kiel probe that provided critical airspeed and altitude data to the aircraft flight control system computers. These data were critical to safe flight, yet the Kiel probe did not include provision for electrical de­icing, presumably because the aircraft would only be flown in clear desert weather conditions. However, during descent to recovery back to Edwards, ice accumulated in the unheated X-31 pitot tube, resulting in the flight control system automatically configuring the aircraft control surfaces for what it assumed were lower airspeed conditions. Unanticipated move­ments of the flight control surfaces caused the aircraft to begin oscillating about all axes followed by an uncontrolled pitch-up to an angle of attack of over 90 degrees.[1247] The subsequent X-31 accident investigation board recommended that training be conducted on the system safety analy­sis process, that procedures be implemented to ensure that all test team members receive configuration change notices, and that improvements be made in the remaining X-31 to prevent similar single-point failures.[1248]

A panel that included former Dryden Research Center director Ken Szalai met at Dryden in early 2004 to review the X-31 accident. The panel noted that the primary contributing factor was the installation of the unheated Kiel probe in place of the original heated Pitot tube. The lack of electrical de-icing capability on the Kiel probe had not been considered a safety risk because X-31 mission rules prohibited flight in precipitation or clouds. However, there was no stipulation specifically restricting flight during potential icing conditions, despite simulations that showed icing of the Pitot static system could lead to loss of con­trol.[1249] Information had been distributed among the X-31’s test pilots and flight-test engineers explaining the Pitot tube change, but a formal process was not in place to ensure that everyone fully understood the implications of the change. Test pilot Lang had noticed anomalies on his cockpit instrumentation and, assuming the presence of icing, told the control room that he was switching on Pitot heat. Shortly afterward, he advised that he was leaving the Pitot heat on for descent and approach to landing. The ground controller then told Lang that the pitot heat switch in the cockpit was not functional. Discrepancies between the X-31’s air­speed and altitude readouts were being observed in the control room, but that information was not shared with the entire control room staff. There was a redundant source of air data and a pilot-selectable alterna­tive control mode that could have saved the aircraft if better commu­nications had existed. Dryden X-31 project manager Gary Trippensee noted that complacency is the enemy of success in flight research; prior to the accident, 523 successful X-31 research missions had been flown.[1250]

In 2000, the remaining X-31 was brought back from long-term storage at NASA Dryden, where it had been since 1995, and reconfig­ured for another round of flight-testing for the Vectoring, Extremely Short Takeoff and Landing Control and Tailless Operation Research (VECTOR) program. This program would explore the use of thrust vec­toring for extremely short takeoff and landing (ESTOL), with a focus on the aircraft carrier environment. An international Cooperative Test Organization was created for the VECTOR program. U. S. partic- ipants/partners were the Navy, Boeing, General Electric, and NASA.[1251]

The Swedish government was represented by Volvo and SAAB, with the German Ministry of Defense and DASA (Daimler-Benz consor­tium) from Germany. The X-31 aircraft was modified to incorporate a Swedish RM-1 engine, the same powerplant used in the Saab JAS-39 Gripen fighter.[1252] On February 24, 2001, flown by U. S. Navy Cdr. Vivian Ragusa, the upgraded X-31 took to the air for the first time from Naval Air Station (NAS) Patuxent River.[1253] German test pilot Rudiger "Rudy” Knopfel, U. S. Marine Corps Maj. Cody Allee, and Navy Lt. J. R. Hansen would fly most of the subsequent ESTOL test program.[1254] The VECTOR X-31 went on to accomplish over 2 years’ of flight-testing, culminating in the final ESTOL flight by Maj. Allee on April 29, 2003.

From April 22 to 29, 2003, the VECTOR X-31 flew 11 test flights, during which fully automated, high-angle-of-attack approaches to land­ing were conducted. The automated flight control system utilized inputs from a special Global Positioning System (GPS)-based navigation sys­tem to maneuver the aircraft to a precise spot above the runway. Known as the Integrity Beacon Landing System (IBLS), it was supplemented by two virtual satellites, or "pseudolites,” on both sides of the runway. Precise spatial position and flight attitude data were inputs for the auto­matic approach control and landing system used in the VECTOR X-31. An ESTOL approach began with the pilot flying into the area covered by the pseudolites; after entering an engagement box, the automatic approach and landing system was activated. The aircraft then assumed a high-angle-of-attack approach attitude and followed a curvilinear path to the touchdown point. Just before touchdown, with the thrust vectoring paddles less than 2 feet above the runway, the X-31A auto­matically reduced its attitude back down to the normal 12-degree angle of attack for landing. An autothrottle system from an F/A-18 and a special autopilot developed by the VECTOR team were coupled with the flight control system to provide the integrated flight and propul­sion control capability used to automatically derotate the aircraft from its steep final approach attitude to touchdown attitude at 2 feet above the runway.

On the final flight of the VECTOR program, the angle of attack during landing approach was 24 degrees (twice the angle of attack on a normal landing approach). Approach airspeed was 121 knots, or about 30 per­cent lower than the normal 175 knots, and the resultant landing distance was only 1,700 feet, compared to the normal landing distance of nearly 8,000 feet. Maj. Allee commented on the experience of riding along on a VECTOR X-31 automatic approach and landing: "There are no g forces and you sit leaning somewhat backwards in the ejection seat while the nose is pointing sharply upwards. . . . At angle of attacks greater than 15 degrees the pilot cannot see the runway except on the screen on the right – hand side of the instrument panel. . . . Whereas on a normal landing the landscape flashes by, now everything takes place as if in slow motion.”[1255]

Another technical accomplishment demonstrated during the VECTOR X-31 program was the successful test of an advanced Flush Air Data System (FADS). Based on data collected by a dozen sensors located around the nose of the aircraft, the FADS provided accurate air data, including airspeed, altitude, angle of attack, and yaw angle, to the flight control system at angles of attack up to 70 degrees all the way out to supersonic speed.[1256]

The two X-31 aircraft completed a total of 580 flights, a record for an X-plane program. Of these, 559 were research missions and 21 were flown in Europe in support of the 1995 Paris Air Show. Fourteen pilots from NASA, the U. S. Navy, the U. S. Marine Corps, the U. S. Air Force, the German Air Force, Rockwell International, and Deutsche Aerospace flew the aircraft during the original program at Palmdale and Dryden, with two U. S. pilots (one Navy and one Marine Corps) and a German pilot flying the VECTOR X-31 test program at Patuxent River. The surviv­ing X-31, U. S. Navy Bureau No. 164585, flew 288 times, making its last flight on April 29, 2003. This aircraft is now on display at the Deutsches Museum annex at Oberschleifiheim, near Munich, and it will eventually be returned to the United States. The other X-31, bureau No. 164584, had flown 292 times before it was lost on January 19, 1995.

The Engine Component Improvement project was tasked with enhanc­ing performance and lowering fuel consumption of several existing com­mercial aircraft jet engines, in particular Pratt & Whitney’s JT8D and JT9D engines and GE’s CF6. The specific goals included:

• Improving the current versions of the engines without requiring a brand-new design or engine replacement.

• Reducing the amount of fuel a typical jet engine would use on any given flight by 5 to 6 percent.

• Significantly slowing the pace at which the engine’s components would naturally degrade and cause a loss of performance over time.

To do this, researchers tried and tested several ideas, including reduc­ing the clearance between rotating parts, lowering the amount of cool­ing air that is passed through the engine, and making refinements to the aerodynamic design of certain engine parts to raise their efficiency. All together, engineers identified 16 concepts to incorporate into the engines.[1309]

Ultimately, as a result of the Engine Component Improvement efforts, engine parts were incorporated that could resist erosion and warping, better seals were introduced, an improved compressor design was used, and ceramic coatings were added to the gas turbine blades to increase their performance. Tests of the improvements were so promising that many were put into production before the program ended, benefiting the

workhorse airliners at the time, namely the McDonnell-Douglas DC-9 and DC-10, as well as the Boeing 727, 737, and 747.[1310]

Observing the success of fuel cells in space flight, NASA researchers in the late 1990s began to explore the potential for fuel cells to power air­craft. The cells were attractive to NASA’s aeronautics directorate because they have near-zero emissions, are quiet and efficient, and can work in any environment where the temperature is lower than the cell’s oper­ating temperature. Valerie Lyons, NASA’s Division Chief of Power and In-Space Propulsion, said she began pushing for NASA aeronautics to pursue fuel cell research about 10 years ago. "I would venture to say they hadn’t really looked at it before that,” she said. "When I looked at the fuel cell area, I said, ‘This is pretty interesting, can we use this?’”[1478]

One of NASA’s main fuel cell initiatives was the Next Generation Clean Aircraft Power (NEXCAP) program, which sought to use the cells as auxiliary power units (APUs) for aircraft.[1479] The APUs could be used in onboard electrical systems to power a grounded aircraft, providing an alternative to wasting fuel and producing emissions by drawing on power from an idling engine. NASA researchers hoped that the APU research would eventually lead to the design and test flight in 2030 of an electric airplane, which would rely on fuel cells for propulsion.[1480] While NASA’s electric airplane never came to fruition, Boeing flew the first piloted airplane powered by hydrogen fuel cells in 2008.[1481] NASA researchers maintain a strong interest in the potentiality of fuel cells to meet the future energy needs of the Nation.

NASA and AeroVironment upgraded the Pathfinder UAV to an improved configuration known as the Pathfinder Plus, which had a longer wing­span, two additional electric motors for a total of eight motors, and improved solar cells. The two additional motors were more efficient than the six carried forward from Pathfinder. The Pathfinder Plus UAV, which was intended to serve as a "transitional” aircraft between Pathfinder and the next-generation Centurion UAV, had a wingspan of 121 feet that

included use of four of the five wing sections from the original Pathfinder. The new center wing section contained more efficient solar cells that converted approximately 19 percent of the solar energy into electrical power for the vehicle’s motors, avionics, and communications system. This compared with an efficiency rating of only 14 percent for the solar cells on the remaining four wing sections from the original Pathfinder. The addition of the fifth wing section enabled Pathfinder Plus to gen­erate 12,500 watts, as compared with the original Pathfinder’s 7,500 watts. Pathfinder Plus had a gross weight of 700 pounds, up to a 150- pound payload capacity, an aspect ratio of 15 to 1, a wing chord of 8 feet, and a power-off glide ratio of 21 to 1. Pathfinder Plus enabled higher – altitude flights and was used to qualify the next-generation Centurion wing panel structural design, airfoil, and SunPower Corporation’s solar array. Several flight tests were conducted in Hawaii. On its last flight, on August 6, 1998, Pathfinder Plus set a world altitude record of 80,201 feet for solar-power – and propeller-driven aircraft. These flight tests demon­strated the power, aerodynamic, and systems technologies needed for the Centurion and confirmed the practical utility of using high-altitude remotely controlled solar powered aircraft for commercial purposes.[1542]