SP-4302 Adventures in Research: A History of Ames Research Center 1940-1965

 

PART II : A NEW WORLD OF SPEED : 1946-1958

1950-1953

 

 

8

Research

 

PATTERN

 

[199] AMES research during the 1950-1953 period was marked by a strong trend toward the more fundamental. A massive effort was made to develop the theory required for deeper understanding of transonic, supersonic, and hypersonic flows. The term "hypersonic," it should be noted, referred to a speed regime generally of Mach 5 and above, where linear theories, dependent on small disturbances, two-dimensionality, and constant gas properties, broke down. The trend toward fundamentality was also, and surprisingly, apparent in the field of flight research. The view being taken of the flight dynamics of an airplane and its relation to automatic electronic-guidance equipment was becoming increasingly scientific and sophisticated.

Of particular interest during this period was the dawning appreciation of the oneness of an airplane or a missile. In prewar days the wings of an airplane were regarded as assemblages of airfoil sections which required independent study. In the early postwar period it became clear that airfoil sections had lost much of their individual significance and that the wing had to be considered as a whole. Now, during this period, it was realized that the wings, body, and tail surfaces of an airplane or a missile were so powerfully interrelated that they could properly be dealt with only as a whole. Two factors were principally responsible for this situation. First, the fuselages of aircraft, and particularly of missiles, had become large relative to their wings; thus their mutual interference or interaction was great. Second, at transonic and supersonic speeds, the interferences between wings, bodies, and other components of an aircraft tended to be much more adverse than they were at subsonic speeds. In the higher ranges of speed, each component of the airplane produced a pressure wave which, depending on the arrangement of the components, would tend either to fortify or to cancel others; and when they fortified each other, the drag was usually much higher. The Area Rule developed by Whitcomb of Langley in 1951-1952 showed that, by properly [200] shaping and distributing the components of an airplane, the transonic drag could be greatly reduced. It was now clear that the airplane must be designed as a whole and that much greater care must be exercised in the placement of its various parts. Inasmuch as transonic drag was a major barrier to supersonic flight, the Area Rule discovery was regarded as an important breakthrough and was kept secret for several years.

Over the country, and the world, many people and agencies were now engaged in the ever-broadening field of aeronautical research. Thus it was often difficult to assign credit for the many overlapping accomplishments of the workers in this field. Research was in effect the game of trying to fill in the immensely complicated and endless jigsaw puzzle of nature with many players participating. The game would seem at one time to be approaching a standstill and then someone would lay down a key piece. This move would immediately inspire a great flurry of action. Players from all over the world would suddenly see how they could add to the pattern. A whole block would quickly be filled in and then the game would again slow down awaiting inspiration from the placing of some new key piece. Key pieces were obviously important but even they depended on the many pieces that had been laid before. Similarly the achievements of the Ames research men rested on contributions from many sources-sources too numerous and widespread to be properly credited in this volume.

 

BASIC CONFIGURATIONS AND AIRFLOWS

 

wings. In the early postwar period the general lack of transonic and supersonic theory made it necessary for experiment to proceed without the guidance that theory normally provided. But the development of theory was pressed with considerable vigor and during this period the output of theoretical papers reached rather impressive proportions, much of this work relating to wings. At Ames, wing theory was advanced by several different groups and individuals. A team composed of Harvard Lomax, Max Heaslet, Franklyn Fuller, and Loma Sluder produced TR 1077 (ref. B-19), an important work on two- and three-dimensional unsteady lift problems in high-speed flight. Max Heaslet and John Spreiter, in TR 1119, made interesting and useful additions to the reciprocity theorem, which states that the drag of a nonlifting wing at supersonic speed is the same whether the wing moves forward or backward. The reciprocity theorem had earlier been confirmed by Walter Vincenti through tests of a delta wing in the 1- by 3-foot wind tunnel.

In TR 1183 Milton Van Dyke added to the general theory of unsteady wing lift by including nonlinear thickness effects; and, in TR 1217, John Spreiter and Alberta Alksne provided an interesting method of predicting pressure distributions on nonlifting airfoils at high subsonic speeds. In the same period, Alfred Eggers, Clarence Syvertson, and others of the 10-by 14 [201] inch-tunnel staff derived a shock-expansion method by means of which some hypersonic flows could be calculated with an accuracy comparable to that obtainable with the generally more precise but far more tedious method of characteristics An example of the use of the shock-expansion method in calculating hypersonic airflows around airfoils is to be found in TR 1123 authored by Eggers, Syvertson, and Kraus.

It was not uncommon in the field of theoretical aerodynamics that flow patterns were expressed in terms of generalized mathematical formulas which could not readily be solved for any desired case. This situation prevailed in the field of transonic aerodynamics. Walter Vincenti and Cleo Wagoner found, however, that it was possible to use the equations of transonic small-disturbance theory to determine the aerodynamic characteristics of a double-wedge (elongated diamond) wing profile for the short range of supersonic speed that occurs just before the shock wave becomes attached to the leading edge.

As the flow around a wing (or fuselage) reaches the sonic speed, a shock wave forms across the flow just ahead of the wing. This normal shock wave moves closer as the speed increases and, if the wing has a sharp leading edge, it soon makes contact with the wing. With further speed increase, the shock inclines backward into the classical oblique wave, the angle of which is a function of Mach number. The flow condition which Vincenti and Wagoner chose to investigate, the one in which the normal wave had not quite made contact with the wing, was clearly of limited scope. But it was only because this condition offered the simplification of subsonic leading-edge flow that a solution was at all achievable.

The analysis produced by Vincenti and Wagoner is covered in TR 1095 (ref. B-20), dealing with nonlifting wings, and in TR 1180 (ref. B-21), which treats the case of a lifting wing operating at a small angle of attack. These reports, which delineated the useful scope of existing theory, are generally regarded as being of the high precision and quality that typified all of Vincenti's efforts. Vincenti's work was technically impeccable and his writings so lucid and unambiguous that editing could generally do only harm.

Solutions to transonic-flow equations had proved so difficult that alternative approaches of a simpler character were sought. Although exact solutions could not always be achieved, certain mathematical relationships were discovered which were common to all known exact solutions and presumably were basic to all solutions. These "similarity parameters," or "similarity rules,', were particularly useful in correlating and giving meaning to the diverse experimental data that had been obtained in wind tunnels and in flight. The transonic similarity rules had been defined originally by von Karman and others, but during this period their application was usefully expanded by John Spreiter in work published in TR 1153 (ref. B-22) . This [202] work not only contributed to similarity-rule knowledge but also provided a better interpretation of the few exact solutions that had been obtained by Vincenti and others.

An excellent example of the application of the transonic similarity rules was provided by John McDevitt in TR 1253 (ref. B-23). In this report, McDevitt was able to correlate a rather large amount of experimental transonic wing data which, as reported in TN 3501 and TN 3502, he together with Warren Nelson and Walter Krumm had obtained on the 16-foot-tunnel bump. The value of the transonic similarity rules for correlation purposes was clearly demonstrated in McDevitt's report.

In the meantime Dean Chapman's work on separated flows had led him into considerations of the optimum shape of airfoils for supersonic and hypersonic aircraft. There had been earlier indications that at high supersonic speeds the trailing edge of a minimum-drag airfoil should be blunt rather than sharp. This situation resulted from the fact that the suction losses behind a blunt trailing edge were more than offset by the lower pressure forces on the fore part of the airfoil made possible by the blunting. At hypersonic speeds the pressure forces on the fore part were predominant and, to minimize these forces, it was necessary to divert the oncoming air as little as possible. If it was assumed that some thickness of the airfoil was required for strength, minimum flow diversion would be accomplished if the airfoil was in the form of a thin wedge with an absolutely blunt, or bluff, trailing edge. At lower supersonic speeds, where the fore drag was somewhat less important, the trailing edge was rounded down (boat-tailed) a certain amount to reduce base suction drag. But how much should the airfoil be boat-tailed for any particular speed? This was the nature of the problem which Chapman attacked. His approach, of classic form, began with theory and ended with experiment. The theory is contained in TR 1063 (ref. B-24) . The confirming experimentation, in which he was aided by William Wimbrow and Robert Kester, is reported in TR 1109 (ref. B-25) .

Chapman's work had from the first been marked by breadth, originality, thoroughness, and genuinely scientific character. For his work on skin friction, base pressure, and heat transfer, he was in 1952 chosen to receive the Lawrence Sperry Award, one of the outstanding honors in aeronautics given annually by the Institute of the Aeronautical Sciences.

As noted, the theoretical research on wings during this period was extensive; the experimental work likewise was considerable. In the 7- by 10-foot tunnel, George McCullough and Don Gault continued their earlier work on the stalling mechanics of thin wings. This work was reported in TN 2502. Thin wings had an unfortunate habit of stalling suddenly as a result of separation of the flow at the leading edge. A possible solution to this problem was the application of suction through a porous surface installed over a portion of the leading edge. A study of porous materials suitable for....

 


[
203]

Dr. Dean R. Chapman.

Dr. Dean R. Chapman.


L. Frank Lawrence.

L. Frank Lawrence.

 

.....such applications was made by the 7-by-10 team of Dannenberg, Weiberg, and Gambucci and reported in TN 3093 and TN 3094. At the same time, a rather extensive series of tests on tail surfaces was carried out in the 7-by-10 and the 12-foot wind tunnels; the results of these tests were summarized by Jules Dods and Bruce Tinling in TN 3497.

Airfoil studies in the 1- by 3 1/2-foot tunnel were continuing and this work was aided by the development of some special instrumentation which made oscillating shock waves and vortical wake flows appear to stand still. This useful device, worked out by Frank Lawrence, Jeff Buck, Stanley Schmidt, and Floyd Looschen, is described in TN 2509 entitled "A Self Synchronizing Stroboscopic Schlieren System for the Study of Unsteady Air Flows."

The study of wing planforms begun so enthusiastically in the early postwar years carried over into the current period; and the results of a large part of this work, performed in several wind tunnels at Ames, are summarized in RM A53A30 entitled "Lift, Drag and Pitching Moment of Low Aspect Ratio Wings," by Charles F. Hall. This report gives particular attention to airplane configurations having delta, or triangular, wing planforms and likewise considers the matter of twisting and cambering delta wings to reduce the increment of wing drag caused by lift.

As the angle of attack of a wing is increased to produce lift, the resultant force vector tilts backward, thus increasing its component in the drag [204] direction. This additional increment of drag, due to lift, is reduced by any leading-edge suction that may be generated by the wing. Although theory indicates that a plane wing with leading edges swept within the Mach cone will develop leading-edge suction, in actuality the predicted suction does not appear. It was realized that the higher drag-due-to-lift resulting from this unfortunate circumstance could significantly reduce the performance of supersonic airplanes in climb and cruising and could curtail their range. In his investigation of this matter, Charlie Hall found that a substantial reduction of drag-due-to-lift could be achieved with a delta-wing airplane if the wing was twisted and suitably cambered throughout its span. For practical reasons it was desirable, he found, to incorporate the camber only in the forward portion of the wing. In the selected arrangement the cambered part of the wing appears as a segment of a cone varying linearly in extent from zero at the root to a maximum at the tip. Ames engineers felt that conical camber had the potential for significantly improving the performance of future delta-wing airplanes.

Bodies. Al Eggers and his staff in the 10- by 14-inch tunnel were in charge of one of the first hypersonic test facilities at Ames and thus eagerly sought, by both theoretical and experimental means, to advance the science of hypersonic aerodynamics. As A1 was a very able and energetic leader with a very good staff, progress was quite rapid. The theoretical attack was headed by Eggers himself, with effective support provided by Clarence Syvertson, Raymond Savin, Frank Hamaker, Stanford Neice, Thomas Wong, and others. In their theoretical work these individuals made use of the existing method of characteristics to develop the more amenable shock-expansion method for computing hypersonic airflow patterns. The shock-expansion method accounted for the entropy rise in oblique shocks and also, to some extent, allowed for changes in gas properties arising from the high temperatures produced by shock waves.

 


Charles Hall and conical-camber wing

Charles Hall and conical-camber wing.

 

[205] In TR 1249 (ref. B-26), Eggers and Savin applied shock-expansion methods to the flow around bodies of revolution. In this work they made use of an earlier determined hypersonic similarity parameter represented by the ratio of the free-stream Mach number to the fineness ratio of the body. Hypersonic similarity rules had themselves been studied (TR 1147 by Hamaker, Neice, and Wong, as well as TN 2250 by Ehret, Rossow, and Stevens) with a resulting extension of their range of application.

In the meantime, Milton Van Dyke had become interested in second order supersonic-flow theory. He made a number of contributions in this field, beginning in 1949 with his doctoral thesis at Caltech (later published as TR 1081) which dealt in particular with flow over cones. Van Dyke's work continued in TR 1194 entitled "A Study of Hypersonic Small-Disturbance Theory" (ref. B-27).

Whenever possible, theory was checked by experiment and, for the experimental study of hypersonic flows over bodies, the supersonic free-flight tunnel was very useful. One of the early reports from that facility was RM A52A14b entitled "Experimental Investigation of the Drag of 30°, 60°, and 90° Cone Cylinders at Mach Numbers between 1.5 and 8.2", by Alvin Seiff and Simon C. Sommer. There was at this time considerable interest in the problem of determining the optimum shape of a body for supersonic and hypersonic flight. Because the pressure forces on the forebody represented such an important part of hypersonic body drag, the determination of the shape of the forebody was considered a most important phase of the overall problem. The noses of many of the bodies tested were in the form of sharp pointed cones. This choice of nose was convenient since, in most theories, sharp-nosed bodies had been assumed, but there was considerable interest also in the effect of rounding the nose. A coordinated program of body tests was conducted in the 1- by 3-foot and the supersonic free-flight tunnels. This program, concerned mainly with nose shapes and covering a Mach-number range of 1.24 to 7.4, is reported in TR 1386 (ref. B-28) by Edward Perkins, Leland Jorgensen, and Simon Sommer. Jorgensen had earlier derived optimum nose shapes for flight at various Mach numbers and had noted that, as far as pressure drag was concerned, the optimum shapes for supersonic and hypersonic flight were much the same.

Wing-Body Interference. For a number of years the trend in the design of supersonic airplanes and missiles had been toward vehicles with larger, longer bodies and smaller, shorter wings. This trend was accentuated in missile design where it was common practice to use small wings in cruciform configuration and where considerable dependence was placed on the lift of the body itself. A separate set of control surfaces was often mounted on the forebody. The aerodynamic interference and interaction between the various surfaces and the body became a matter of great importance to designers. The lifting body produced a pair of vortices, as Allen and Perkins [206] had shown; and these, combined with the vortices, shock waves, and down wash fields generated by the lifting surfaces, contributed to an extremely complex, mixed-up flow pattern. It would be supposed that no one would have the courage to attempt a theoretical study of such complicated interference phenomena but that is exactly what Jack Nielsen of the 1- by 3-foot tunnel staff undertook to do. He spent several years studying various phases of the wing-body interference problem and in confirming his theories through wind-tunnel tests. Representative of the work that Jack did during this period was the theoretical-experimental study reported in TR 1252 (ref. B-29). Another example of his work is TR 1307 (ref. B-30), a study in which he collaborated with William Pitts and George Kaattari. Confirmation of Nielsen's theories was provided by Tom Canning and Pat Denardo in RM A52C24, which covered a program of tests in the supersonic free-flight tunnel on the lift and the center of pressure of low-aspect-ratio cruciform and rectangular wings in combination with a slender fuselage at Mach numbers up to 6.2.

Area Rule. One of the most significant developments of this period was the discovery, by Richard Whitcomb of Langley, of a method, called the Transonic Area Rule, for reducing the transonic drag of aircraft.1 Actually the theoretical basis for the Area Rule had been established a little earlier, but it was not until Whitcomb, using the new transonic tunnel at Langley, made his independent discovery that the significance of the method was fully appreciated. The Transonic Area Rule expresses the concept that the transonic drag of an airplane is strongly dependent on the distribution of the cross-sectional area of the airplane, including the wing and all other components, and that as far as pressure drag is concerned the airplane could be represented by a body of revolution having the same longitudinal distribution of cross-sectional area as the airplane.

The optimum distribution of area was not precisely determined by Whitcomb, but it was concluded that the area distribution curve should be nicely rounded and free of sharp peaks or humps. The fuselage of a transonic airplane usually had an area distribution of this kind, but the addition of a wing, an engine nacelle, or a wing-tip fuel tank produced a hump in the area curve which caused the pressure drag of the airplane in the transonic range to rise to a high peak. Airplanes that designers hoped would be supersonic were thus sometimes limited to sonic speed as a result of unexpectedly high transonic drag. Whitcomb showed that the hump in the area curve' and thus the high transonic drag, could be eliminated if the added area contributed by the wing was balanced by a deliberate reduction in the cross-sectional area of the fuselage at the wing juncture. The resulting fuselage shape had a constriction like a Coca-Cola bottle, and the terms "coke bottle" or "Marilyn Monroe" fuselage were often heard.

 


[
207]

B-58 model in tunnel, which incorporates conical camber and Coke-bottle fuselage.

B-58 model in tunnel, which incorporates conical camber and Coke-bottle fuselage.

 

As it turned out, necking the fuselage was only one of several measures that could be taken to improve the area distribution. Although it was generally desirable to reduce the cross-sectional area as much as possible, a nicely rounded area curve could sometimes only be achieved by adding area to the fuselage at strategic points. Also it became clear that, on airplanes carrying externally mounted bodies such as engine nacelles, rocket pods, and fuel tanks, an opportunity was provided the designer to arrange these various items in such a way as to obtain a desirable area curve and a low transonic drag.

The Area Rule discovery precipitated a surge of research-and-development activity throughout the country that clearly would last for several years. At Ames the Area Rule was confirmed by tests in available transonic facilities and also by means of the drop-test technique which the Ames Flight Engineering Branch had developed. Representative of the Area Rule drop-test work was a study reported in RM A54F22 entitled "An Experimental Investigation of Reduction in Transonic Drag Rise at Zero lift by the Addition of Volume to the Fuselage of a Wing-Body-Tail Configuration and a Comparison with Theory," by George H. Holdaway.

The Transonic Area Rule at the time of its discovery had a very limited theoretical basis, a condition which certain members of the Ames staff took immediate steps to correct. One of the early contributors to Area Rule theory at Ames was R. T. Jones, who reviewed the whole subject in TR [208] 1284 (ref. B-31) and introduced the concept of the Supersonic Area Rule. Through measurement of the area along Mach cones rather than along transverse cross sections, the Supersonic Area Rule made it possible to minimize the drag of an airplane for any chosen supersonic Mach number. TR 1284 was a substantial contribution to the theory and application of the general Area Rule.

Another group at Ames, comprising Barrett Baldwin and Robert Dickey, approached the Area Rule problem with a slightly different viewpoint. They observed that the Transonic Area Rule minimized the drag at sonic speed and the Supersonic Area Rule minimized it at some specific supersonic speed; but what was needed, they felt, was a method of minimizing the drag over a range of transonic and low supersonic speeds. In RM A54J19, Baldwin and Dickey developed a theory, known as the Moment of Area Rule, which, as confirmed by tests in the 2- by 2-foot tunnel, accomplished the desired objective to a useful degree.

 

PROPELLERS AND INLETS

 

Although at this time the use of jet propulsion was well established, there was still considerable interest in propellers. The turboprop engine was felt to be applicable where takeoff thrust and cruising efficiency were important; such applications included cargo, transport, and reconnaissance types of aircraft, some having high-speed capabilities. Propeller requirements were now substantially different from what they had been when reciprocating engines were in vogue. The power to be absorbed was much greater than for a reciprocating engine and airplane speeds were generally higher. On some of the faster airplanes, propellers would have to operate efficiently at supercritical, if not supersonic, tip speeds.

To satisfy these requirements, a new class of propellers was being developed, the performance of which remained to be evaluated. To perform this service, and to learn more about the subject in general, the 12-foot tunnel in 1951 began an extensive series of propeller investigations. Some of the first work in this program was reported by Robert Reynolds, Robert Sammonds, and John Walker in TR 1336 (ref. B-32). In this phase of the program, advanced types of blades were tested at both forward and reverse thrust in configurations representing four-blade, single-rotation and eight-blade, dual-rotation propellers. The propellers were tested at wind-tunnel speeds up to Mach 0.84, but at tip Mach numbers up to 1.4.

Meanwhile in the 40-by-80, a research team led by Vernon Rogallo and Paul Yaggy was investigating the unsteady airloads to which propellers are subjected while operating in nonuniform flow fields such as might be encountered in front of straight or swept wings that are producing lift. Of principal concern to the investigators was the structural integrity of propellers as [209] affected by the stresses arising from such unsteady airloads. Examples of work performed in this program are contained in TN 2308 and TN 2957.

Inlet work continued during this period but at a fairly moderate pace. studies in this field were carried out in the 7- by 10-foot and the 1- by 3-foot tunnels as well as by Flight Engineering through the use of drop-test models. One of the more significant contributions during this period was TR 1141, entitled Method and Graphs for the Evaluation of Air Induction Systems," by George B. Brajnikoff.

 

DYNAMIC STABILITY AND LOADS

 

Problems concerned with dynamic stability and loads received a limited amount of attention from Ames research groups from 1950 to 1953. As indicated in TR 1088, the theory of missile dynamics was advanced through the efforts of Gaynor Adams and Duane Dugan of the 6 by 6-foot tunnel. Contributions were also made in this field by Robert Chubb and Dave Reese of the same Branch.

The Laboratory's experimental efforts in the study of aircraft dynamics included some interesting drop tests by Norman Bergrun of the Flight Engineering Branch. These tests, reported in TN 2985, confirmed the existence of a peculiar type of inertia-coupling instability that William Phillips of Langley had predicted. The instability was characteristic of aircraft of advanced type having long fuselages and short wings.

While Bergrun was making his drop tests, Ben Beam, in the 12-foot tunnel, was giving further attention to the pitch damping of triangular wings; and Alfred Boissevain, in the SSFF tunnel, was measuring the damping m roll of triangular wing-body arrangements at Mach numbers up to 6.0. The SSFF tunnel was particularly good for this sort of work as the test model was completely free.

In addition to the tests just mentioned, a number of transonic wing buffeting and flutter studies were made in the 16-foot tunnel using the transonic bump for some and, for others, a demountable, two-dimensional throat installed in the test section. Of the buffeting studies, one was made by Charles Coe and Jack Mellenthin, and another by Andrew Martin and James Reed. Data from the latter study were confirmed by flight tests of the F8F-1 and the X-1 airplanes. The flutter studies were performed by Robert Barnes, Raymond Herrera, John Wyss, and James Monfort and are reported n RM A51I25, RM A54A29, and RM A54C24.

The theoretical studies of static airloads on wings of arbitrary planform, studies far advanced in the early postwar years, were continued during this period. In TR 1071, John DeYoung extended his span-load analysis to include the effects of flap deflection. Lynn Hunton and Harry James, in TN 3040, used the DeYoung method to predict loads on swept wings with both leading-edge slats and flaps.

 

[210] FLIGHT RESEARCH

 

Guidance and Control. A large part of the work of the Flight Research Branch falls naturally into the field of guidance and control. Under the leadership of Harry Goett, Larry Clousing, and Steve Belsley, the work of this Branch was advancing along new and productive lines. With the advent of automatic control for airplanes and missiles, the dynamic behavior of aircraft was being looked at in a much more sophisticated and scientific light than ever before. The pilot with his enormous ability to compromise and adapt had been able to compensate for the imprecision of existing design knowledge regarding airplane dynamics; but the electronic and mechanical servo systems that would now guide our aircraft did not have these rare abilities. If we were to let electromechanical systems fly our aircraft, we must be able to describe with great precision, and in a language meaningful to such systems, exactly how the aircraft might be expected to respond to a deflection of its control surfaces. Thus any successful attack on the problem would require the marrying of two rather diverse disciplines-one being flight aerodynamics and the other electronics and servomechanisms.

Also to be considered were the cases in which a combination of human and electronic guidance would be required. These cases would include fighters with computing gunsights and interceptors with automatic or semi-automatic missile fire-control systems. Here a compatibility between human and electronic elements would have to be achieved. The resulting system would be expected to satisfy essential requirements of both pilot and electromechanical system. The whole picture clearly indicated that flight research had taken a quantum jump in complexity and sophistication.

Swept wing Problems. By no means was all of the flight research during this period concerned with exotic new trends. Investigations were made, for example, of the special stability and control problems of sweptwing airplanes. For this work the F-86 proved useful. In sweptwing airplanes, the boundary layer forming along the span tends to run out toward the tips and cause early stalling of the outboard sections of the wings. Such stalling often causes one wing to drop abruptly, a phenomenon called "wing dropping," and if both wing tips stall at the same time, the center of lift suddenly moves forward, causing a disconcerting and possibly dangerous "pitchup." To overcome this fault, a number of cures were attempted. As earlier noted, chordwise barriers called "fences" were installed at various points on the upper surface of the wing. Leading-edge extensions on the outboard part of the wing were also tried as were vortex generators installed on the outboard upper surface of the wing. The vortex generators were simply a row of small, airfoil-like projections which, by promoting turbulent mixing, reenergized the sluggish boundary layer, thus rendering it less susceptible to separation. All of these devices were helpful in some degree.

[211] A number of studies of sweptwing control problems of the kind just mentioned were made during this period. One of these is described in TN 3523 by Norman McFadden, George Rathert, and Richard Bray. A second study, published in RM A54F21, was made by Norman McFadden and Donovan Heinle. Heinle was an Ames test pilot as well, of course, as an engineer.

Pilot Opinion. The basic stability and control characteristics of the new types of fighter aircraft being developed were much different from those of the older propeller-driven types. The differences resulted for the most part from the elimination of the propeller, the change in configuration and mass distribution, and the higher operational speeds and altitudes. Much was yet to be learned about the stability and control of these new aircraft, and also about their handling qualities as appreciated by the pilot. Quantitative evaluation of both the mechanical-aerodynamic and the human aspects of the control problem for such aircraft was required. For studies of this kind, the variable-stability airplane, earlier described, proved to be particularly useful.

When the F6F variable-stability airplane was first used, its ailerons were driven to allow simulation of various degrees of dihedral. Arrangements had subsequently been made to drive the rudder as well as the ailerons so that a wider range of lateral-directional stability variables might be investigated. A study in which the modified variable-stability airplane was used is reported in RM A51E16 by Charles Liddell, Brent Creer, and Rudolph Van Dyke. In this investigation lateral-directional handling qualities were evaluated by 12 different pilots including two from the Air Force, four from the Navy, five from NACA, and one from the Cornell Aeronautical Laboratory, which had begun work with variable-stability airplanes at about the same time as Ames. The pilots were asked to assign numerical ratings to each of the configurations simulated by the airplane and were given specific questions to answer with the aim of obtaining the reasons for their ratings. This was an interesting attempt to obtain pilot opinion in quantitative form and to correlate it with airplane design parameters.

Tracking. A critical measure of the controllability of a fighter airplane was the accuracy with which its pilot could follow a moving and maneuvering target such as another airplane. Such tracking studies were accomplished through the use of gunsights and gun cameras using a second airplane as a target. Through the statistical analysis of the tracking errors revealed by the gun camera, much could be, and was, learned about the dynamic response and control characteristics of airplanes. Particularly instructive was the comparison of the performance characteristics of the older types of airplanes, about which much was known, with the characteristics of the new high-speed and high-altitude fighters about which little was known.

A program of tracking tests was undertaken at Ames in 1952. In an [212] early phase of this program, a comparison was made of the tracking performance of the F-51H, the F8F-1, the F-86A, and the F-86E airplanes representing two straight-wing and two swept wing types. The tests, reported in RM A53H12 by George Rathert, Burnett Gadeberg, and Howard Ziff, were made with a fixed gunsight; in other words, there were no electronic elements in the guidance circuits. The next step in fighter development was to install electronically controlled and computer-equipped gunsights that would "lead" the target by an appropriate amount.

Automatic Control. Owing to uncertainties in existing knowledge of aerodynamic loads at high speeds, the proof testing of new airplanes to design limits subjected the pilots of such airplanes to great hazard. To avoid such hazards, the Naval Air Experiment Station developed a radio-operated control system by means of which the flight of a test airplane could be remotely controlled from another aircraft or, for takeoff, from a ground-control station. Such a system was installed in a Curtiss SB2C-5 dive bomber while a Grumman F6F-5 was equipped to serve as the base of control. The airplanes were turned over to Ames for investigation and they became the Laboratory's entry into the new and important field of automatic control.

The first investigation undertaken with these aircraft, reported in TN 3496 by Howard Turner, John White, and Rudolph Van Dyke, was essentially an evaluation of the system, the measuring of responses, and the establishment of control settings. In these determinations use was made of the opinion of "check" pilots who rode as passengers in the drone airplane or who sometimes operated the "remote" control equipment from within the airplane.

The Navy apparently lost interest in the remote-control system, at least for the use originally intended, and thus the SB2C-5 drone became available to Ames for research on automatic control systems. One of the notable studies made with the airplane is reported by Howard Turner, William Triplett, and John White in RM A54J14 (ref. B-33) . For this study, additional equipment was installed in the airplane to allow it to simulate a radar-controlled interceptor. The radar was represented by an optical system-a manually controlled periscope. The periscope was pointed at a maneuvering target (another airplane) and the SB2C-5 was then flown directly toward the target. Photographs of the target taken simultaneously through the periscope and through a gun camera rigidly mounted on the airplane were later compared to determine tracking accuracy. To obtain the most desirable control settings, a detailed investigation of the responses of the airplane was made both in flight and on the ground. The ground studies were simulations using the Reeves Analog Computer. Much knowledge and know-how useful in later work on automatic control systems was obtained during these tests.

An essential ingredient in the design of an automatic control system for [213] a high-performance airplane is a set of quantitative expressions representing the dynamic response of the airplane to movement of its controls. These response characteristics can be predicted from wind-tunnel test data but can be determined more reliably, perhaps, by flight-test methods. The latter methods of obtaining the dynamic response characteristics of an airplane, specifically of the F-86A airplane, are illustrated in TR 1250 (ref. B-34) by the flight research team of William Triplett, Stuart Brown, and G. Allan Smith. Smith was a member of the Research Instrumentation Division which, under the quiet but effective direction of Jim White, gave valuable support to much of the work at the Laboratory.

In the method illustrated in TR 1250, the transient response (motion) of the airplane arising from a pulsed input to the controls is recorded as a function of time. By means of a Fourier transformation, these time histories are converted to a basic frequency-response form which is not only more amenable than the time history to detailed analysis but also more compatible with servomechanism analysis methods. The frequency-response data thus obtained are in the form of phase angle and amplitude ratio plotted against frequency. From these graphical data, simple analytical expressions for airplane response, called "transfer functions," are obtained by a semi-empirical method using templates and, as a final refinement, an analog computer.

Hazards. Flight research, though now more sophisticated, was nonetheless hazardous. On June 1, 1953, while putting an F8F through some step-control maneuvers, Rudolph Van Dyke crashed into San Francisco Bay and was killed. The cause of the crash was never determined, but it was conjectured that, since the crash had been preceded by a long steep glide in which no evidence of recovery attempt was apparent, Van Dyke had perhaps failed to connect his oxygen mask and had lost consciousness from a lack of oxygen. Rudy had served as a research pilot at Ames since 1947 and had made important contributions to many of the Laboratory's flight-research projects. His passing was a serious loss to the Laboratory and caused much sorrow.

 

BOUNDARY LAYER, SKIN FRICTION, AND AERODYNAMIC HEATING

 

Aerodynamic heating was becoming of prime importance in the 1950's to the designers of high-speed aircraft. The heating to be experienced by hypersonic aircraft would not only affect the skin friction and thus the overall drag of such craft but ultimately would melt the leading edges of the vehicle unless some form of thermal protection was provided. The optimum form of such protective means was still unknown.

Aerodynamic heating of aircraft is directly related to the compression and viscous shear (skin friction) that occurs in the boundary layer as well as to the heat-transfer efficiency of the boundary layer-its ability to transfer its heat to the body. As the skin friction and the heat-transfer efficiency were [214] known to be much higher for turbulent than for laminar boundary layers, a turbulent boundary layer was to be avoided if at all possible. The boundary layer at the nose of a body was always laminar but, as it flowed backward along the body, it would at some point become turbulent. The position of this transition point was affected by a number of factors such as surface roughness, pressure gradient, and Reynolds number and it also seemed to be affected by the temperature of the body relative to that of the boundary layer. It appeared that transition was delayed, that it occurred farther downstream, if the body was cool and heat from the boundary layer was flowing into it. This fortunately was the normal condition encountered both in flight and in hypersonic test facilities, but in the test facilities the temperature difference was usually much less than in flight.

Analysis of the heat-transfer problem was considerably simplified as a result of a direct relationship, called the Reynolds analogy, that existed between skin friction and heat transfer. The analogy was known to apply at moderate speeds where aerodynamic heating was caused largely by viscous shear, but there was a question as to whether it would also apply at higher speeds where the heat of compression became important. In the investigation of these matters, there was clearly much to be done both theoretically and experimentally. Facilities for experimental studies were still, however, somewhat limited.

One of the early reports from the small heat-transfer wind tunnel at Ames was TN 2740 by Randall Maydew and Constantine Pappas. The work covered by this TN was an investigation of the skin friction in the laminar boundary layer of a flat plate at Mach 2.4 and a comparison with existing theory. On the theoretical side, Morris Rubesin, in TN 2917, provided a modified Reynolds analogy which allowed for compressibility effects in the boundary layer. In the supersonic free-flight tunnel, James Jedlicka, Max Wilkins, and Alvin Seiff made an experimental determination of boundary-layer transition on bodies of revolution, with both rough and smooth surfaces over the nose portion. This study, reported in TN 3342, revealed that under some conditions the boundary-layer transition, instead of occurring all at once, developed as a series of bursts of turbulent flow separated by short periods of laminar flow. This phenomenon was revealed very clearly in shadowgraph photographs of the test body in flight.

In TN 3284, Al Seiff added to the growing store of knowledge on heat transfer by analyzing a collection of test data from several wind tunnels and from flight to confirm the validity of the Reynolds analogy and, more particularly, the modified analogy proposed by Rubesin. Although the analyzed data represented a diversity of sources and test conditions, the results of the analysis supported the modified analogy to a remarkable degree.

A growing concern over the difficulties experienced in cooling pointed bodies led to a number of studies of the feasibility and benefit of using bodies [215] with bluff or blunt noses. Pursuing this matter, Howard Stine and Kent Wanlass analyzed the effect on boundary-layer heat transfer of a strong negative pressure gradient (rapidly accelerating flow), such as would occur on a hemispherically shaped body, and followed up their analysis with tests in the 1- by 3-foot tunnel of round-nosed bodies at a Mach number of 1.97 The results of their work are presented in TN 3344.

Aerodynamic heating of the exterior surfaces of high-speed aircraft had become a major problem. Without thermal protection, these surfaces particularly the leading edges, could reach structurally intolerable temperatures at Mach numbers as low as 2 or 3. Among the more promising methods for protecting wing leading edges that had appeared was one called transpiration or "sweat" cooling. In this method a cooling fluid is introduced between the hot boundary layer and the wing surface by forcing the fluid through a porous wing skin. One of the early studies of transpiration cooling at Ames was made by Morris Rubesin and published in TN 3341. In this report Rubesin provided an analytical method for estimating the effect of transpiration cooling, assuming air as the cooling fluid, on the heat-transfer and skin-friction characteristics of a compressible turbulent boundary layer. Most of the earlier analyses had been restricted to considerations of a laminar boundary layer; thus Rubesin's report covered new and interesting ground.

 

BALLISTIC MISSILE PROBLEM

 

The development of the intercontinental ballistic missile was a tremendously complicated task and an undertaking of the utmost importance to the Nation. It required the solution of many difficult technical problems most of which were beyond current experience and some beyond current knowledge. The missile would have to be flown at speeds of 15,000 miles per hour or more and to altitudes of perhaps 500 miles; it would then have to be accurately guided to a target 5000 miles away by a guidance system that could not be jammed by enemy action; and finally, and most importantly, it would have to deliver its nuclear warhead, in operating condition to the target.

The necessity of minimizing the structural weight of the missile was a problem in itself. Any unnecessary weight in the structure would either greatly increase the weight and cost of the whole missile or, more likely, reduce the weight and power of the warhead. Thus the slightest unnecessary Weight in any part of the missile was intolerable. The explosive charge itself had a structural container the weight of which had to be minimized. Sitting atop the missile, this warhead structure had been of pointed configuration m early designs. The use of a pointed nose for rifle shells and high-speed missiles was traditional. The concept was firmly implanted in the thinking....

 


[
216]

Shock waves produced by blunt bodies (above) and pointed bodies (right).

Shock waves produced by blunt bodies (above) and pointed bodies (right).

 

....of the aerodynamicists and aeronautical engineers who were designing the missile.

It was clear that the objective of missile launching would be completely frustrated if the warhead burned up owing to aerodynamic heating before it reached its target. This danger was very real for, as it reentered the atmosphere, the warhead would generate a temperature rise in the surrounding air of many thousands of degrees. Some means of protecting the warhead against reentry heating was required, but every scheme proposed for this purpose involved painful weight penalties. It was at this stage that Harvey Allen bent his mind to the problem.

Harvey knew that the kinetic energy lost by a missile, or a warhead, as it enters the earth's atmosphere is totally converted into heat. The heat comes from two sources and, significantly, is generated in two places: i.e., inside and outside the boundary layer. The heat appearing outside the boundary layer is generated by shock-wave compression. Of the heat generated in the boundary layer, some results from compression but much more arises from viscous shear or skin friction. The heat generated by the shock wave outside the boundary layer is for the most part well removed from the body and cannot reach it by convection through the insulating blanket of boundary layer. On the other hand, the heat generated in the boundary layer has ready access to the body and even readier access if the boundary layer is turbulent than if it is laminar. It thus appeared to Harvey that for any overall rate of heat generation-which would be determined by the [217] degree of deceleration-the body would be heated less if a larger fraction of the heat were generated by the shock wave (try pressure drag, as Harvey put it) and a smaller fraction by the viscous skin friction. How could this juggling of heat inputs be accomplished? The answer seemed quite obvious to Harvey: by making the nose blunt in order to strengthen the bow shock wave and increase the pressure drag.

The bow wave standing out in front of a bluff body is much more powerful, and generates much more heat, than the oblique waves leaning back from a pointed nose. Very significantly also, the bow wave, unlike the oblique wave, does not touch the body at any point. The oblique waves, though cooler than the bow wave, touch the pointed body at a spot where the insulating boundary layer is very thin or nonexistent. They are thus able to transmit their heat to the tip of a missile with great ease and so fast that it cannot be carried away by conduction. A hot spot develops and the point generally melts off.

It appeared then that much less heat would be transmitted to a bluntnosed warhead than to a pointed one and that the heat reaching the blunt warhead would be more evenly distributed and less likely to produce hot spots. The desirability of reducing skin friction by maintaining laminar flow over the heated area was also evident. But there was more to the problem....

 


H. Julian Allen explaining blunt-nose principle.

H. Julian Allen explaining blunt-nose principle.

 

[218] ....than appeared in this analysis, which assumed that the maximum rate of heat generation would be the same for a blunt body as for a pointed one. Inasmuch as heat generation depends on rate of deceleration, this assumption seemed rather broad. The dynamics and trajectory of the warhead would have to be examined, and this Harvey Allen and Al Eggers undertook to do. In due time the two men were able to demonstrate mathematically that the maximum deceleration of a body entering the atmosphere- and thus the maximum rate at which it generates heat-is determined by its entrance angle and velocity and is independent of its physical characteristics: i.e., mass, size, shape, or drag coefficient. This astonishing conclusion was based on the assumption, appearing reasonable for warheads, that the body in all cases will reach the point of maximum deceleration before hitting the ground. It was also regarded as a reasonable assumption in the analysis that, regardless of their shape, bodies with the usual densities of warheads would convert most of their kinetic energy into heat before hitting the ground. The trajectory analysis thus supported Harvey's blunt-body heating theory by showing that the maximum rate of heat generation by a warhead, and the total amount of heat generated, would be the same regardless of whether it had a blunt or a sharp nose The important difference between the two body shapes was in the amount of the generated heat that actually entered the body.

Allen's blunt-body theory was conceived in late 1951. It was published as a classified document and later as TN 4047. Finally, it was published as TR 1381 (ref. B-35). Al Eggers was a contributor to the analysis and a coauthor of the report. Pointed-body tradition was still strong in 1952. Harvey's idea was not quickly picked up by industry; it just lay there and steamed for a while.

 


1 See preceding chapter.


previousindexnext