1 Reply Latest reply on May 16, 2016 1:50 PM by Steven Barry

    Using a piecewise function within Equation Driven Curve?

    Mark Wood

      I have modelled a wing with anhedral arc and loft toward the wingtips. The cross-section is sketched using the 4 digit NACA formula within Equation Driven Curve. Using Eq Driven Curve as opposed to Curve through X,Y,Z, gives me the ability to use 'Driving Dimensions' as parameters within the equation, allowing a parametric shape optimisation within ANSYS DesignXplorer.

      My issue is that I can find no way of adjusting the camber of the airfoil cross-section. In order to account for the camber correctly, I must use a 'Piecewise Function' i.e. using a particular formula for the sketch line from x1=0% of chord length, to x2= position of max camber as % of chord length, and a separate formula for the line from x2 until x3= 100% of chord length ('x' being the horizontal coordinate).

      Any help/advise concerning how to do this would be really appreciated, as I seem to have hit a brick wall. Thanks


      I've attached the camber formula below for clarification:


      NACA_4 Digit Camber Formula.PNG

        • Re: Using a piecewise function within Equation Driven Curve?
          Steven Barry


          I can't say I've ever used the "Equation Driven Curve" feature before now.  I just tried it out, hoping to be able to use combinations of signum (step) functions to accomplish something piecewise.  It didn't quite work out, SolidWorks seems to be able to process sgn(n) where n is a constant, but it wont do sgn(x), with x being a variable.

          Also, I wasn't able to link any global variables in the Equation Driven Curve, so that seems quite problematic to your situation.

          Since your equations are both 2nd order polynomials, why don't you try using two parabolas for your profile.  That way you can control the shape by using Smart Dimensions (which CAN link to your global variables).  For instance, dimension the point that they meet as y = m.


          Just to add on to my answer, it seems like the second equation should have some dependence on "c", such that you would want y = 0 at x = c.  I would change that to y = m / (c-p)^2 * [ (c^2 - 2pc) + 2px - x^2 ]