Testing any all-wing configuration presents challenges for supporting the model in the wind tunnel and obtaining force and moment measurements from it. The oblique all-wing concept further aggravates these problems because it has the potential to produce large untrimmed roll and yaw moments, and the lack of flow symmetry causes unusual support system interference, even at zero sideslip angle.
The NASA program of the early 1990’s provides one approach to overcoming these testing issues, despite the fact that the results remain unpublished. Certainly other solutions may also be feasible. The following discussion describes the model support and force balance used in those tests, and addresses various issues and experiences that arose during the tests.
The model was mounted in the NASA Ames 9 ft x 7 ft Supersonic Wind Tunnel using a blade strut that attached to the underside of the wing at mid-span. A notional sketch of the support is shown in the figure below. Because of the asymmetrical flow in the region below the wing, the blade was cambered and twisted in an attempt to align it with the local flow, minimizing the lateral load on the blade and resulting influence on the model.
A procedure for designing the strut shape was developed using the TRANAIR full potential CFD code. The oblique all-wing configuration was modeled in the nominal cruise condition with a survey grid of off-body sampling points arranged in the location where the model support blade would be. Local velocity information at these points was used to compute a “warped” stream-surface shape that was everywhere tangent to the local flow. This surface was treated as the mean surface to define the blade support, with a supersonic airfoil (wedge-plate-wedge) thickness distribution added. Once this was done, a second iteration TRANAIR model was run with the blade included. The results of this computation showed that the blade was producing some lateral force, so the local incidence (twist) of the blade was adjusted at several stations and the configuration re-run until there was essentially no lift on the blade. Of course since the blade must have finite thickness, there remained the pressure influence on the lower surface of the wing related to the displacement of the flow around the strut.
It was found during the wind tunnel test that there was a moderate untrimmed roll moment on the model that was suspected of originating from interference from the blade strut. An OVERFLOW Navier Stokes CFD solution of the wing with and without the blade later confirmed that the blade was in fact inducing the rolling moment on the model. These CFD solutions were to be used to generate strut corrections for the wind tunnel data.
One conclusion from these results is that the simple method employed to warp the blade was not adequate. Apparently the pressure field caused by the thickness of the blade influenced the wing differently on the left and right sides. This would be expected given the large wing sweep. Although it is not possible to design a strut that produces no influence on the model, it would be desirable to use an optimization procedure to minimize the induced moments (predominantly roll moment) from the model support system.
The oblique all-wing configuration tested by NASA in 1993 was designed as a passenger transport, so the center region of the wing was fairly thick (about 16% t/c). Even with this rather thick wing, it was impossible to install a conventional 6-component internal force balance inside the wing. Several arrangements of external balances were studied, but all suffer from requiring extremely large pitching moment capacity that precluded high resolution drag measurement. Ultimately it was decided to design a special force balance that was arranged as a flat box with a center metric region attached to the non-metric surrounding frame by appropriate gaged flexures. More information on this balance is available from Unitary Wind Tunnel testing organization at NASA Ames Research Center.
Three different boundary layer conditions were tested during the NASA OAW tests. The oil-flow interferometry shear-stress measurement technique and traditional sublimation technique were used to determine the presence and location of boundary layer transition. With no forced boundary layer transition, the wing demonstrated a substantial amount of laminar flow. In addition to free transition, two different heights of cylindrical “trip-dots” were tested. The trip dots were 0.040 inch diameter epoxy disks, spaced 0.125 inch apart, following the methodology adopted within the NASA High Speed Research (HSR) program that was proceeding concurrently. Based on traditional boundary-layer trip methodology, an initial trip height of 0.0075 inches was selected. Oil-flow shear-stress measurements found that the 0.007 inch trip height was not sufficient to establish prompt boundary layer transition, and the trip height was increased to 0.011 inches for the remainder of the testing. Unfortunately, some of the test configurations tested with the first trip height were not repeated with the second trip height, complicating efforts to extract useful force and moment increments for those configurations.
It is worthwhile noting here that the same under-estimation of required trip height occurred in the NASA HSR program, where the highly-swept, inboard portion of the wing with a subsonic, rounded leading-edge demonstrated remarkable resistance to boundary layer transition. This may seem counter-intuitive given degree of cross-flow produced by the high-sweep, rounded leading edge. Extensive testing of various trip heights and resulting transition location, as well as an improved method of determining the drag increment from the forced transition is documented in a AIAA Journal article by Aga Goodsell and Robert Kennelly8. The lesson from these studies is to always verify the desired boundary layer transition location as the first objective of testing.
Figure 5.1. Oblique Flying Wing
In order to better understand some of the stability and control issues associated with the oblique flying wing concept an actively controlled low speed testbed was built and flown in 1990 by researchers at Stanford University led by Stephen Morris. This work is described in detail in Dr. Morris' thesis (Ref. 6) as well as in conference papers (Refs. 7-8) available through the links here. A video clip of the flight test at NASA Ames Research Center is also available.
In addition to some of the basic oblique wing design concepts described in Chapter 2 of this paper, the testbed highlighted the importance of thrust/pitch coupling, the influence of vertical surfaces on (long axis) pitch control, the variation in trailing edge control surface effectiveness with sweep (different on forward and aft wings), the non-trivial landing gear design challenges, and the importance of digital lags and latency on a vehicle with short time constants in the long axis dynamic modes.