CONCLUDING REMARKS

I have attempted, using a case study of wings at supersonic speed, to show how research engineers built their knowledge of aerodynamics in the days before electronic computers. As in other advanced fields of engineering, the process involved the comparative, mutually illuminating use of experiment and theory, neither of which could reproduce exactly the actual problem. In theoretical work, limited ability at direct numerical computation required physical approximations and assumptions – in the present case, the customary inviscid gas, plus thin wings at small angles of attack for three-dimensional problems – to bring the calculations within the scope of analytical techniques then available. For the two­dimensional problem of airfoils, the linear, thin-wing approximation could be improved upon; even then, however, the inviscid assumption could be circumvented only by means of qualitative concepts and quantitative estimates from the independent boundary-layer theory. On the experimental side, the effect of the inevitable support-body interference could be estimated in some aspects, but then only roughly. The inaccuracies accompanying experimental measurement, which I have not gone into, also had to be considered. As in much engineering research, experimental ingenuity, theoretical capability, and analytical insight formed essential parts of the total process.

The material here illustrates clearly two parts of the threefold makeup of modem engineering research (and much design and development) pointed out in the introduction. All three parts – theory, experiment, and use – appeared varying degree in my recent paper on the early development of transonic airfoil theory, research in which I participated later at Ames.17 To quote Constant again in regard to still another example from aeronautics, “the approaches were synergistic: discovery or design progressed faster when the three modes interacted.”18 Use in flight could not be involved in the wing research here, since supersonic aircraft were not yet available; the principle still appeared in the motivation for the study, however. It is this kind of synergism, as much as anything else, I believe, that provides the power of modern engineering generally.

My transonic paper contains a further concept pertinent here: the view of an engineering theory as an artifact – more precisely, a tool – to use in testing the performance of another artifact.19 Here the linear theory was used to test wings on paper in much the same way as the wind tunnel served to test them in a physical environment. In the present application, both tools were put to use in research for knowledge that might someday be employed in design of aircraft. They are also employed regularly side by side in the typical design process. This view – of theory and experiment as analogous artifacts for both research and design – I find useful in thinking analytically about modem engineering. It helps me to focus on theory and experiment in a parallel way in sorting out the synergistic interaction pointed to above. (Use might be looked at similarly, though I have not thought the matter through.)

As pointed out earlier, the wing research did not achieve anything broad and reliable enough to be included under the “theoretical design methods” mentioned in the introduction. Whether and to what extent the research contributed to the accompanying “general understanding” and “ways of thinking” is difficult to know; this would depend on what the audience took away from my New York talk and on how widely and thoroughly our reports were read and thought about. The story does exemplify, however, the kinds of things that make up those categories.20

“Ways of thinking” in my view comprises more or less structured procedures, short of complete calculative methods, for thinking about and analyzing engineering problems. As appears here, aeronautical engineers had long found it useful to regard the aerodynamic force on a wing in terms of lift, center of lift, and drag. This division has the virtue, among other things, of relegating the influence of viscosity, and hence the need to take it into account as a major factor, primarily to drag. Such division has been the practice for so long that engineers dealing with wings take it as almost natural. Designers of axial turbines and compressors, however, because the airfoil-like blades of their machines operate in close proximity to one another, think of the forces on them rather differently. A second example in the present work is the manner of accounting for the various performance characteristics of a wing in terms of the interplay of the inviscid pressure distribution and the viscous boundary layer, both of which can be analyzed, to a first approximation, independently. This procedure too had been around for some time, but the present example, by being fairly clear, may have added something. I have found it useful, at least, in teaching.

“General understanding” consists of the shared, less structured understandings and notions – the basic mental equipment – that engineers carry around to deal with their design and research problems. This and ways of thinking are perhaps best seen as separated, indistinctly bounded portions of a continuum rather than discrete categories. At the time of the present work, the difference in propagation of pressure signals between subsonic and supersonic flow had been understood for many years, and the concept of the Mach cone was becoming well known. Its consequences for the flow over wings were being explored, and the benefits and problems of sweepback were topics of widespread research and discussion. A feeling was developing in at least the research portion of the aerodynamic community that in some semiconscious way we “understood” something of the realities of supersonic flow. In our work at Ames, we contributed to this understanding in a small degree and advanced the knowledge of the powers and limitations of linear theory. After three years of living with supersonic wing problems, our group had acquired some of the mental equipment needed to understand and deal with such problems; the necessary ideas had become incorporated into our technical intuition. Indications later materialized that some of this was picked up from our reports by the aerodynamic community, but how much is anyone’s guess. It is these kinds of knowledge that I see under the rubric “general understanding.”

Other concerns appear here that are treated at some length in my transonic paper.21 I mention them briefly for the reader who may wish to look into them further:

(1) Our story provides examples of both the experimental and theoretical aspects of what scientist-cum-philosopher Michael Polanyi called “systematic technology,” which I take to be the same as what scholars and engineers currently speak of as engineering science. This, in Polanyi’s words, “lies between science and technology” and “can be cultivated in the same way as pure science” (Polanyi’s emphasis) but “would lose all interest and fall into oblivion” if the device or process to which it applies should for some reason cease to be found useful by society.22

(2) Our research at Ames, by requiring as many people as it did, illustrates how engineering advance is characteristically a community activity. I subscribe wholeheartedly to Edward Constant’s contention that communities committed to a given practical problem or problem area form “the central locus of technological cognition” and hence a community of practitioners provides “a primary unit of historical analysis.”23 Here the community we have examined existed entirely at Ames, where our wind-tunnel group depended critically also on the laboratory’s machine-shop, instrumentation, and electrical – maintenance sections. Externally, however, through our reports, my New York talk, and visitors who came to consult us, we were at the same time becoming increasingly a part of the international supersonic research-and-design community that was then forming.

(3) The personal motivation for some in our group came from the fact that the work was part of the job from which they earned a living. For people with greater responsibility, the work offered intellectual and experimental challenge and excitement, heightened by the potential utility of the results – typical research-engineering incentives. (The necessary administrative motive had come from discussions about our section’s overall program with our research superior, the chief of the laboratory’s High-Speed Research Division; the choice was fairly obvious, however, given our new wind tunnel and the existing state of knowledge.) Motivation of the NACA as a governmental institution flowed presumably from its desire to maintain its competitive position vis-a-vis other countries in supplying knowledge to the aircraft industry for design of supersonic airplanes, should they prove practical. Motivation overall was thus a complex mix.

(4) The laboratory’s institutional context for research could scarcely have been improved. Supervision, which was by engineers who had done (or, in the case of our section’s division chief, was still doing) research, was informed, supportive, and free of pressure; interaction with other research sections of the laboratory was encouraged. Skilled service groups provided support when called upon. My fellow research engineers and I didn’t realize how fortunate we were.

In closing, I would make one more point. The process we have seen is now in one

respect a thing of the past. Thanks to large-scale digital computers, designers and

research engineers today can calculate the flow over complicated shapes in detail without either the inviscid assumption or mathematical linearization or other approximation of any sort. To the categories of physical experiment, analytical theory, and actual use, we can thus add a kind of direct “numerical experiment” as a fourth instrument in both our search for aerodynamic design knowledge and in design itself24 The resort to mathematical analysis in the way seen here is thus no longer essential. Our ability to incorporate turbulence and turbulent boundary layers into direct calculations, however, still leaves something to be desired. Where such phenomena are important, which includes most practical problems in aerodynamics, comparison between numerical and physical experiment still plays a role. Such comparison can also be important in instances of great geometrical complexity, which computers encourage aerodynamic designers to attempt.

The foregoing statements, I must emphasize, have been entirely about aerodynamics; it should not be assumed that they apply to engineering generally. In fields where analytical and numerical methods are not so advanced, experiment and use may still predominate. Overall, the situation is still very mixed. Other fields and details of the present case aside, however, the point I would emphasize for the topic of the workshop is this: To understand the evolution of flight in the twentieth century, tracing the nature and evolution of research and knowledge may be as necessary as is the study of aircraft and the people and circumstances behind them.

EPILOGUE

The work we have followed found echo years later at the renowned Lockheed Skunk Works in southern California. In 1975, Richard Scherrer, one of the test engineers in the Ames research, headed a Skunk Works group engaged in preliminary design that would lead to the F-l 17A “Stealth Fighter.”25 Mathematical studies by one of Scherrer’s group had suggested that the military goal of negligible radar reflection might best be attained by a shape made up of a small number of suitably oriented flat panels. Scherrer’s memory of the Ames tests encouraged him to believe that such a startlingly unorthodox shape might in fact have acceptable aerodynamic performance. His faceted flying wing, laid out along the lines of the double-wedge triangular wings of figure 10, became known to his skeptical Skunk Works colleagues as the “Hopeless Diamond.” The idea, however, proved sound. The largely forgotten research from Ames thus contributed 30 years later to cutting – edge technology that could not have been imagined when the research was done. As in human affairs generally, serendipity plays a role in engineering.