Neutral Point and stability analysis

[Märten Mikk]

Hi!

I've run into a weird problem. Running the steady stability analysis gives me a x_np = ~144. When I put that as the reference point and run VSP without the stability analysis from alpha 0 to 10 degrees, the Cmy vs alpha, then the slope is non-zero. Why might that be? 

If I try to find the NP manually by adjusting the reference point in VSPAero GUI, then it comes around ~150.

I've attached my model. My length units are in mm and mass in grams.

My second question is about the P stability analysis. When should one use it? The Q and R analyses return stability derivatives wrt to alpha and beta dot, but what extra information does the P analysis give? 

Finally, a third question. My final objective is to model an already existent design to get some more estimates for the stability derivatives. These derivatives are then plugged into a 6DOF simulator to develop some control laws. How would you recommend to model the mass properties? Set mass and inertia values for each component separately? Could I only model one component and set other components to have 0 mass? Should I try to make some estimates about the density of the part so that the model scales better? How does VSP handle 0 mass?

All the best,

Märten

[Rob McDonald]

Here is the CMx vs CL result I get for your case -- with the Xcg = 144.

This is for a bunch of alphas from [-10,10]

If you look at the viewer, you'll see that the dip at -5 deg is a point where the wake is directly impinging on the horizontal stabilizer -- I would say the wake has not fully converged in that case.  Increasing the spanwise resolution on the tail surfaces might help some here.

The pitch stability mode runs cases at 0 and 1 degree alpha, it uses those two points to calculate Xnp.  You'll see that in the range of [0,1], the chart is flat as expected.

The other important point is the magnitude of the numbers -- these results are all very close to zero.

Your reference quantities are set very large -- this is OK, but by setting them in a non-conventional manner, you lose any intuition you would have based on the magnitude of the results.  For example, your cref is set to 1000, but the actual average chord is 80.  Your Sref is 1e6, but the actual wing area is ~47000.  So, I would expect your moments to be about a factor of about 200 too small.

The P stability analysis gives results with respect to roll rate.  The results should be largely similar to those resulting from the steady roll analysis.

Rob

[Märten Mikk]

Hi Rob,

You've nicely cleared things up for me. Thanks for taking the time!

But, I'd still inquire a bit about my third question. What would your advice be on how to model mass properties for getting an estimate for stability derivatives? I apologize if this is too trivial a question.

[Rob McDonald]

Sorry I missed that one.

VSPAERO does not care about the mass or inertias.  It only calculates the aerodynamic derivatives, it is up to you to turn those into a meaningful dynamic analysis of the aircraft -- which will require the mass and inertias.

VSPAERO only uses the CG information as the reference point for the moments.  You could also just use (0,0,0) and as long as your tools are set up to use that as the reference point for moments, everything will work out fine.

If you want to use OpenVSP to estimate the mass properties for your dynamic analysis, you can do that -- but VSPAERO doesn't care.

OpenVSP's mass properties calculation is a bottoms-up approach.  It simply does the math on the provided information.

In conceptual design, we usually take a more top-down approach.  I.e. we use historical regressions for mass and inertia of aircraft or components.  This means that the mass estimate of a wing (for example) includes everything -- lightening holes, fasteners, sealant, hydraulic/pneumatic/electric/fuel lines and systems, control surfaces, access panels, actuators, paint, etc.

OpenVSP only knows about the OML of the wing -- you can tell it a surface area based density and a volumetric density, but that is it.  OpenVSP doesn't know about all that mess that gets captured in the top-down approach.

So, if you're modeling an EPP foam hand-launched glider, then OpenVSP's mass properties analysis is going to be very good.

OpenVSP's mass properties methods can be used, but you have to really think about what you're calculating.  They are often most useful in finding deltas -- what is the pitch inertia change caused by moving the engines from the wing to the tail etc.

If you are modeling an aircraft that really exists for flight simulation purposes, then you might not need  OpenVSP's mass property calculations at all -- just use the numbers you already have.

Or perhaps you have mass numbers, but not good inertia numbers.  In that case, I would work through the aircraft components one-by-one and calibrate the density values to reflect the results of the real components.  Use point-masses for all the discrete components you can and come up with smeared density values for the surface and volume that captures all of the other stuff not explicitly in the OpenVSP model.

With some care, you should be able to come up with a model that gives you a reasonable estimate of the inertias.

Rob

[Märten Mikk]

Thank you for the extensive answer and clarifying these things!