Chapter 2: The High-Speed Airfoil Program
[55] In the late fifties, the first American jet transports became operational and the first concepts for supersonic transports began to appear. Langley aerodynamics researchers tended to regard the subsonic jet transport as a perfected accomplishment and devoted themselves to the problems of the supersonic concepts. Stack, and his successor after 1961, L. K. Loftin, Jr., strongly supported work on the first Mach 3 supersonic transport (SST) designs which emerged during this period. The British and French also became deeply involved at this time in the developments which led to the Concorde. Unlike NACA and NASA, however, they maintained a continuing program of high-speed airfoil research applicable to subsonic swept-wing aircraft (ref. 71). The cruising speeds of the more advanced subsonic jet transports were limited by the drag rise of the wing which started to occur in the vicinity of the critical Mach number, and Pearcey of the National Physical Laboratory undertook a study aimed at improving supercritical drag characteristics. He showed in 1962 (ref. 67) that the conditions for shock-free recompression previously suggested by Sinnott and others could be realized in airfoils whose curvature decreased abruptly downstream from the leading edge. For these airfoils a limited region of smooth isentropic recompression existed ahead of the terminal shock at supercritical speeds. Thus the shock which eventually occurred was weaker and the shock-induced drag rise was delayed by perhaps 0.03 in Mach number as compared to cases where there was no recompression. It was also apparent from experimental experience that this effect is present naturally to some degree for the thinner sections previously tested in which the critical Mach number could frequently be exceeded by as much as 0.2 without shock stall. Derivatives of the Pearcey sections were used on such second-generation jet aircraft as the 747, DC-10, and the A-300. Noting Pearcey's work, G. S. Schairer of Boeing suggested in his 1964 Wright Brothers Lecture [56] that additional research to evolve optimum airfoil shapes "when shocks are present would be timely."
One of the principal Langley investigators caught up in the SST program was Richard T. Whitcomb. He had evolved a configuration which enjoyed a higher lift-drag ratio than other competing Langley concepts. But when industry evaluations of these designs became available in mid-1963 it was evident that Whitcomb's design had the highest structural weight and poorer range performance than the others. All the designs had such high fuel and operating costs relative to subsonic transports that Whitcomb became quite disillusioned and rather dramatically declared that he was quitting the SST program.
For some months he cast about for a new challenge (ref. 72). Quite by accident he was asked by Loftin, his boss in the Langley front office, to comment on some high-speed model test data for a vertical takeoff (VTO) design under study by the Ling-Temco-Vought Company. The design incorporated an upper surface blowing slot supplied with engine air as a part of its VTO system. When the slot was operating it appeared to produce a substantial increase in the force break Mach number. Whitcomb reasoned the slot blowing effect was delaying shock-induced separation and he began to wonder if this mechanism might not be a way to increase the cruise speed of subsonic transports which in some cases were limited by the drag-rise Mach number. He became sufficiently interested to start experimenting, although there had been little pressure for work in this area.
The first tests were made on a conventional NACA 6-series section with a self-actuated slot in which air flowed from the high-pressure region under the wing. (Power blowing was ruled out from the start as being too costly in weight and complexity.) Whitcomb used Lachman's book for guidance in design of this slotted model (ref. 73). The slot action did delay and reduce the shock-induced separation losses. But in so doing the normal shock moved further aft and became so strong that the direct shock losses nullified much of the gain due to reduced separation. Thus the next step was to try to modify the upper surface shape so as to weaken the shock.
Whitcomb at this time was aware in a general way of the previous work of Lindsey's group which had culminated with Daley's systematic [57] studies of 6-series airfoils in the 4 x 9-inch tunnel extending to Mach I (ref. 55). He knew of the advantages of reduced camber for supercritical operation, but he was not aware of the special airfoil developed by Allen and recommended by Woersching to be used in the inverted attitude. He had, however, recently read Pearcey's paper (ref. 67) which utilized a flattening of the forward region of the upper surface. Whitcomb had, of course, realized for many years that reducing the curvature or flattening the upper surface would generally reduce the local Mach numbers and reduce the shock strength as he desired (see p. 39ff. and ref. 51). He therefore drafted a slotted airfoil with a flattened upper surface ahead of the slot, which naturally resulted in large negative camber. A large portion of the lift then had to be carried by a short, positively cambered portion aft of the slot. Tests of this arbitrary design showed a substantial increase in drag-rise Mach number (ref. 74). It was found a short time later that the slot could be eliminated for only a small penalty in the onset of drag rise and with considerable simplification in structure and ease of application in three-dimensional wings.
Continued development of these sections has taken place over the past decade. Flight demonstrations on the Navy's F-8 and T-2C airplanes and on an Air Force F-111 have verified the wind tunnel results (ref. 74).
In the course of developing the wings for these flight programs, it was learned that the supercritical airfoils had excellent high lift characteristics because of their large leading-edge radii. This important benefit tended to offset the fact that their subcritical profile drag is higher than for comparable 6-series sections.
Whitcomb's initial development of these supercritical sections was entirely experimental. By an Edisonian process of intelligent guesswork, intuitive reasoning, and cut-and-try testing-with the wind tunnel used in effect as a computer-successful profiles were achieved.
After the first work began to produce impressive results, Loftin suggested in 1965 that a simple baptism similar in character to the area rule be found. Whitcomb proposed "supercritical," a more fortunate choice than the "peaky" appellation used for Pearcey's airfoils.
Loftin also instigated in 1969 the first program to apply transonic [58] theory to the supcritical airfoil problem, realizing that the Edisonian approach was hardly practical for producing the many different airfoils that would be needed to supply the increasing demands of designers for a variety of applications (ref. 75). Paul Garabedian of New York University was chosen to try to develop a practical theoretical program that could be used with large modern computers as a routine airfoil design tool. He describes the work of Murman and Cole (ref. 76) as the "breakthrough" which underlies the recent achievements of the transonic theory (ref. 70). The first theoretical results for Whitcomb's section did not account for the boundary layer displacement thickness and showed poor agreement with regard to shock location. Good agreement was obtained when the boundary layer was included (refs. 70, 74). The theory is now used routinely as a major tool in the program, saving an enormous amount of wind tunnel testing.
An important new development has also been contributed by the theoretical program: upper-surface shapes were found by Garabedian and Korn (ref. 77) which produced shockless supercritical flow for limited ranges of speed and angle of attack. The basic mechanism involved is the previously mentioned reflection of expansion waves back from the curved sonic line as compressions. The upper surface shapes which accomplish this recompression without a terminal shock arc remarkably similar to those of some of the Whitcomb sections. In fact, Whitcomb had noticed a drag reduction in certain tests of his sections which he attributed to the existence of local conditions approximating the requirements for shockless or near-shockless flow.
In von Karman's summary of compressibility effects in 1941 (ref. 61) he included a brief but significant review of the theoretical possibilities of exceeding the critical Mach number without the occurrence of shock. He cited the work of Taylor, Gortler, and Tollmein which suggested that local velocities as high as 1.6 times the speed of sound could be achieved with smooth shockless recompression, and concluded, "The mere fact that air passes over the wing with supersonic velocity does not necessarily involve energy losses by shock waves . . . or the compremibility burble," and "careful theoretical and experimental research might be able to push the velocity of [efficient] flying closer to the velocity of sound than is possible now." Coming as they did at the [59] threshold of the war, these wise words were lost, and a quarter century would pass before the theoretical supercritical airfoil program of the seventies would prove them correct. The general attitude of most airfoil researchers of the forties was that shockless flows were a curiosity of the theoretician not likely to exist in real viscous flows.
The extent to which the shockless designs will further improve supercritical airfoils is not clear at the early stage they are in at this writing. They were clearly of special interest at the recent airfoil conference (ref. 70), but it was too soon to expect a definitive perspective of their true potential. The Whitcomb airfoils have only weak shocks and thus complete elimination of the shocks would not be expected to make large improvements.
It is quite interesting that over the entire NACA history no attempt to develop superior high-speed airfoils by the Edisonian technique was ever made. A great deal of valuable experimentation was done to learn what was happening on particular airfoil shapes, and systematic testing of families such as the 16-series and 6-series was carried out from which the most effective members of these established families could be identified. But Whitcomb was the first to embark on a zealous crusade to develop an improved airfoil by intelligent cut-and-try procedures. This situation is even harder to explain when one notes that the Edisonian technique was often employed in other NACA programs-the cowling programs, for example. On problems of great complexity such as the supercritical airfoil this least sophisticated of all research techniques is likely to prove ineffective-unless the practitioner has truly unusual insights and intuitions.
Whitcomb's first successful supercritical sections contained the same type of camber distribution (negative camber over most of the forward portion followed by positive camber) recommended in 1951 by Woersching for maximum delay of the drag rise. Whitcomb, however, employed much more drastic profile changes leading to a radical new section. Woersching's airfoils (of which Whitcomb was not initially aware) looked more like slightly modified conventional sections.
[60] In spite of its doubtful credibility in the mid-sixties, the transonic theory in combination with the modern computer was actually on the verge of achieving shockless supercritical airfoils. Undoubtedly the focus and stimulation provided by the Whitcomb developments hastened the derivation of the shockless airfoils, which are very similar in their upper surface configurations to Whitcomb's designs. However, this achievement would certainly have come along eventually without the prior Whitcomb developments. Thus, in the long-term perspective, the Whitcomb contribution by the sheer accident of coming when it did, produced the supercritical airfoil perhaps some 10-20 years sooner than it might otherwise have emerged from the theoretical approach.
A final point of considerable interest centers on the fact that several important applications have appeared where the supercritical airfoil principle is used to achieve thicker wings rather than higher drag-rise Mach numbers, the thicker wings being lighter structurally, thus providing larger payload fractions and improved economics. Alternatively, thicker wings permit the use of higher aspect ratio with associated performance and economic benefits. The interesting fact here is that the existence of the new airfoils illuminated important needs and applications which were not clearly seen in the beginning of the development. This underscores once again the old, but often still not accepted axiom that it is impossible in advance to identify all the real applications and justifications for a research undertaking. Whitcomb's work has sparked a lively renaissance of high-speed airfoil R&D in which the new theoretical approaches, used in combination with experiments, are providing a degree of technical elegance that was lacking in the prior NACA programs. There can be little doubt that high-speed airfoil technology is now approaching its ultimate levels of sophistication and performance.