Applying Computational Structural Analysis to Flight Research

We now turn to an area of activity that provides, for aviation, the ulti­mate proof of design techniques and predictive capabilities: flight-test­ing. While there are many fascinating projects that could be discussed, we will consider only five that had particular relevance to the subject at hand, either because they collected data that were specifically intended to provide validation of computational predictions of structural behav­ior, or because they demonstrated unique structural design approaches.

Two of these are the YF-12 Thermal Loads project and the Rotor Aerodynamic Limits survey, both of which collected data for validat­ing and improving predictive methods. The remaining three are the Highly Maneuverable Aircraft Technology (HiMAT) digital fly-by-wire (DFBW) enhanced agility composite-structured canard demonstrator, the AD-1 oblique wing demonstrator, and the Grumman X-29 forward- swept wing (FSW) research aircraft. These three projects exercised, in progressively more challenging ways, the concept of aeroelastic tailor­ing: that is, predicting airframe flexibility and having enough confidence in those predictions to design an airplane that takes advantage of elas­tic deformation, rather than just trying to minimize it. In all of these, NASA-rooted computational structural prediction proved of great, and even occasionally, critical, significance.

The investigation of aircraft structural mechanics or, indeed, of almost any discipline, can be considered to include the following activ­ities: investigation by basic theory, computational analysis or simula­tion, laboratory test, and flight test (or, more generally, any test of the final product in its actual operating environment). Many arguments have been had over which is the most valuable. This author is of the opinion—based on his experience in the practice of engineering, on a certain amount of historical research, and on the teaching and example of mentors and peers—that theory, computation, laboratory test, and flight test all constitute imperfect but complementary views of reality. Thus, until someone comes up with a way to know the exact state of stress and deflection in every part of a vehicle under actual operating conditions, we must form our understanding of reality as a composite image, using what information we can gain from each available source:

• Flight test, obviously, is the best representation we have of an aircraft in actual operational conditions. However, our ability to interrogate the system is most severely compromised in this activity. Many data parameters are not available unless special instrumentation is installed, if at all, and this is the most difficult environment in which to obtain stable, high-quality data.

• Laboratory test offers better visibility into the opera­tion of specific parts of the system and better control of experimental parameters, at the price of some separa­tion from true operational conditions.

• Computation offers even greater opportunity to inter­rogate the value of any data parameter at any time(s) and to simulate conditions that might be impossible, difficult, or dangerous to test. Computation also elimi­nates all physical complications of running the experi­ment and all physical sources of noise and uncertainty.

But in stepping out of the physical world and into the analytical world, the researcher also becomes subject to the limited fidelity of his computational method: what effects are and are not included in the computation and how well the computation represents physical reality.

• Theory is sometimes the best source of insight and of understanding what parameters might be changed to obtain some desired effect, but it does not provide the detailed quantitative data necessary to implement the solution.

In this light, the following flight programs are discussed. Much more could be said about each of them. The present discussion is necessarily confined to their significance to the development or validation of loads and structural computation methods.