Desktop version

Home arrow Engineering arrow Small Unmanned Fixed-Wing Aircraft Design. A Practical Approach

Panel Method Solvers - XFoil and XFLR5

XFoil[1] is a panel-based method that uses full potential methods to calculate the bulk flows around two-dimensional sections which can then be corrected for the presence of the boundary layer. The code is aimed at dealing with viscous analysis of airfoils, allowing for forced or free transition, transitional separation bubbles, limited trailing edge separation, and lift and drag predictions just beyond CLmax. Since the corrections used are based on well-established experimental data, very good approximations of lift and drag can be predicted as long as the boundary layer remains attached to the airfoil. Once separation has begun, leading on to stall, panel codes are less reliable, which is a fundamental limitation of such methods. Figure 13.1 shows the kinds of data provided by XFoil, here at a high AoA where separation has begun. To obtain this result, the following XFoil steps are used:

  • • The desired foil is selected with the “NACA 0012” command.
  • • The panel nodes are set directly from the foil with the “PCOP” command.
  • • The direct operating point menu is entered with the “OPER” command.
  • • Viscous analysis is selected with the “Visc” command.
  • • The Reynold’s number (here 4.4 million) is entered with the “Re4400 000” command.
  • • The airfoil behavior at the desired AoA (here 16°) is calculated with the “ALFA 16” command.

By dividing the lifting surfaces into stream-wise strips, it is possible to take the results from XFoil and compute the behavior of whole wings - this is what XFLR5 does,[2] while also allowing for simple streamlined fuselage elements and multiple combinations of wings. Again, it will not deal so well with heavily separated or stalled flows and, although interference

C and streamline plot for the NACA0012 foil at 16° angle of attack as computed with XFoil

Figure 13.1 Cp and streamline plot for the NACA0012 foil at 16° angle of attack as computed with XFoil.

effects between widely spaced lifting elements (such as wings and tails) can be dealt with, such methods cannot accurately predict the benefits of slotted flaps and other boundary layer control systems. XFLR5 does, however, readily permit calculation of the aircraft’s dynamic stability, which can be used to check whether elevator and fin sizes and positions are acceptable given the likely mass, inertias, and flight speeds. To analyze a simple wing with XFLR5, the following steps are used:

  • • The desired foil section to be used in the wing design is chosen along with the operational Mach number for the wing; the section is loaded into XFLR5 via the “Direct Foil Design” menu. AirCONICS provides a good selection of basic foils, the UIUC database has many more.
  • • The foil section is then analyzed with the “XFoil Direct Analysis” menu in batch mode; analysis type 1 is used to specify a range of section operational Reynold’s numbers and an increment in these along with a sweep through likely section angles of attack. This builds an internal database of results using XFoil from which the wing analysis can be generated (typical values are Reynold’s numbers from 100 000 to 6 million in steps of 100 000 with AoA sweep from -4° to 16°. The system simply aborts any combinations that cannot be converged by XFoil; such a sweep typically takes 2 or 3 min. Figure 13.2 shows the results for the NACA 64-210 section at Mach 0.17).
  • • The wing is loaded under the “Wing and Plane Design” menu by reading an XML file as created by the AirCONICS system for the wing of interest; this comprises the locations and sizes of sections along the wing and references the previously chosen airfoil section. (Note: if this step is carried out before the chosen section has been loaded, a simple planform diagram is created instead.)
  • • Alternatively, the wing is created directly in XFLR5 with manual inputs, in the “Plane Define a new plane” area, where a title and description can be set before defining the main wing; each section requires values for y, chord, (leading edge) offset, dihedral, twist, section name (from the drop-down of sections built earlier), and the number and distribution of panels (typically between 10 and 30 in cosine or sine pattern).
  • • The desired analysis is defined by using type 1 again but now with the chosen operational speed and one of the four available analysis types (Lifting Line Theory, Horseshoe Vortex, Ring Vortex, or 3D Panels; the first of these does not allow viscous affects to be included, so we prefer to use the 3D Panel approach as the computational costs are still slight, but this is possible only if the tail and fin are not included. 3D panels have to be used to generate full pressure maps for structural analysis, however).
  • • The computed reference area, span, and chord length should be checked against what is expected.
  • • If desired, inertia terms, nonstandard atmospheric properties, and extra drag items can next be added.
  • • The sequence of angles of attack can then be specified and analyzed and checked for any errors; the most likely error is that one of the sections along the wing goes outside the previously computed set of polars for a given wing AoA (a local section lift coefficient is requested that is above what can be achieved for the section at the given operating point). If this happens, modify the chosen sweep of angles to ensure that all angles can be fully
Results of XFoil analysis sweep for the NACA 64-201 foil at Mach 0.17 as computed with XFLR5

Figure 13.2 Results of XFoil analysis sweep for the NACA 64-201 foil at Mach 0.17 as computed with XFLR5.

dealt with.[3] Figure 13.3 shows the results of a typical sweep through the available angles of attack. These results can be exported to a text file for subsequent analysis or comparison with other results.

XFLR5 also allows tail fins and elevators to be added and also a second wing (typically for bi-plane configurations). These are all defined as for the main wing and, in the case of fins, can be either single- or double-sided. All the surfaces can have dihedral, sweep, and twist as required, and the code allows for the flow fields around the individual items to impact on each other in terms of downwash but not wake effects. It is also possible to add fuselage-like elements to link the lifting surfaces together but these are not correctly modeled aerodynamically. Dynamic stability analysis can readily be carried out to give the natural frequencies, damping of modes, and so on. To demonstrate what these codes can achieve during design, they are used in the subsequent subsections to predict the behavior of simple 2D airfoil sections, simple wings, and a built-up airframe.

  • [1]
  • [2]
  • [3] The length of time needed to carry out the sweep will be directly related to the number of angles analyzed and thenumber of sections used to define the wing. If a complex wing is being chosen with as many as 100 sections, thiscan take up to an hour. If this is not acceptable, remove some of the intermediate sections in the XML file to simplifythings.
< Prev   CONTENTS   Source   Next >

Related topics