On Wednesday, October 29, 2014 1:38:45 PM UTC-7, jfitch wrote:

One problem complicating this is that thermals are normally accompanied by local horizontal "gusts" which are actually sustained flow field (as mentioned in some of the above posts).

Signal to noise can always be a problem - the good news should be that if the air is really going up you ought to be able to pick that out, but I agree if there is turbulence on the thermal entry that has greater velocity than the thermal itself and/or goes on for a long way a human observer won't be able to integrate the net effects for long enough to figure out what is going on. A computer might have a better shot at it.

However, if the phugoid response difference is your only signal, that is going to be a challenge.

I was saying something a bit different. In the case of a horizontal gust the main dynamic reaction from the glider is some modest horizontal deceleration and a slow pitch up followed by a slow pitch down (post-gust) from the phugoid response. In a thermal entry you get mostly a vertical surge plus some downward pitching moment from the short period response if the glider has static stability. Most modern gliders don't benefit from being flown at the aft limit but even if you do the response should be different. If you have a glider that generates a nose-up pitching moment from an increase in angle of attack, that would be a real handful to fly even under benign conditions. This is at least some of the reason why thermals "feel" different. That surge you feel has a different linear acceleration vector and a different (opposite) pitch response.

If the Butterfly uses Kalman filters to separate out the air mass movement that is exactly what I was suggesting (and attempting to explain why) - you ought to be able to pick out the air movement vector IF you have the right onboard sensors AND you have an accurate enough dynamic model for the glider. The better the model represents all the aerodynamic and inertial coefficients the more accurate the answer should be. I could also imagine intelligently looking not just at the instantaneous airmass velocity, but also the profile of thermals for a given day to help identify good ones from bad ones, though that is a much more complex matter.

Thermals also have temperature gradients - though the experiments I participated in recently were not conclusive.

9B