Wanted: Expereienced Flight Aerodynamicists for Winch Launch Investigation

[G Moore]

Wanted: Experienced Flight Dynamicist to Facilitate Development of Winching Pitch Axis-Only Glider Model

I am a pilot, instructor, and retired Electrical Engineer building a next generation active tension-controlled winch for improved safety, height attained, and pilot experience (among other innovations). All the development work is being done open source by myself and a few collaborators and all related technical documents are publicly available. I have no plans to make a business out of this work, i.e. building and selling winches or plans (but an enterprising entrepreneur would be welcome to do so). The investigation outlined below would be an opportunity for someone with the appropriate skills to make a significant contribution to the understanding of winch launch dynamics to our sport’s benefit.

A key aspect in closing a tension servo loop is the behavior of a coupled mass-spring-mass system formed by the winch equivalent mass (mostly due to the drum’s angular inertia), the cable’s elasticity, and the glider mass. This system’s dynamics are well understood during the ground roll and early rotation launch phases but how the glider aerodynamics affects this system following the rotation and into the climb phase is not. Some conceptual theories have been formulated but without a relevant pitch axis model they cannot be evaluated. To that effect, I am looking for help in developing a simple pitch axis-only glider model valid during winching. Conceptually, the thrust force in an airplane’s pitch model would be replaced by the cable tension force acting on a hook displaced from the CG whose pull angle varies throughout the launch. I am an experienced control systems engineer but I and my current collaborators’ limited aerodynamics knowledge is restricting our ability to develop a simple but representative model that encapsulates the fundamental behaviors exhibited by a glider during a winch launch. For full launch simulations, A simple non-linear (lift/drag proportional to airspeed squared but AoA solidly in the linear regime) model where the cable pull angle varies as the launch proceeds is desired. For the tension control systems analysis, linearized models for different launch portions (primarily cable pull angles) are needed. This model would encompass both the phugoid and short period modes. We have studied and basically understand Etkins on these subjects but struggle to decide what is important to include in such a model and what can be ignored while still capturing the essential behaviors. We simply lack the knowledge and experience a skilled aerodynamicist could bring to the investigation.

We would appreciate any assistance a member of this forum might provide. A contributor would likely find this a very fun and intellectually challenging project. Development of an appropriate model requires a strong familiarity with Laplace domain modeling of aerodynamic systems using pole-zero transfer functions. Experience in simulating such systems would be a plus but I have implemented full launch simulations using simpler models using Matlab and Simulink and incorporating a more sophisticated model in these simulations should be straightforward. Several options have been identified for how AoA/airspeed would be controlled, i.e., the pilot stick inputs, but further details are beyond this short introduction’s scope. If you should know of such a person who does not commonly participate in this forum that might be appropriate for this project, I would greatly appreciate your forwarding this solicitation to them.

If you think you might be interested, I would be happy to share the prior work and speculations on this subject (and in general on tension-controlled winching). If after reviewing this material, you decide this is not something you would like to become involved with there would be no obligation; we would be happy you took a look. But maybe you could point us at someone else that might be interested. The level of involvement is totally open. At a minimum, direction as to what to include/not include in an already outlined preliminary model and help in determining some representative values for some of the key parameters, e.g. glider rotational inertias about the pitch axis, would be extremely useful. But what would be ideal would be someone to discuss the problem and issues with and then they take the lead on determining what an appropriate model should be, how it should be parameterized, and evaluating is it providing the essential behaviors important to the larger winch mass/cable spring/glider system dynamics.

Should you be interested and a university educator, this could be a good project a Masters level Aeronautical Engineering student. The results would be publishable—again this is an open source project with everything in the public domain. This could be expanded to a prospective thesis topic.. How such an undertaking would be structured is open to discussion.

If you’d like to learn more, send me a note and we can figure out how to go from there. We would probably start with my sharing of some earlier reports, simulations, and logged data from actual launches for your review.

Thank you,

George Moore, CFIG, Ph.D.
gsmoore501 AT gmail DOT com

[G Moore]

On Wednesday, October 2, 2019 at 11:16:16 AM UTC-7, G Moore wrote:
> Wanted: Experienced Flight Dynamicist to Facilitate Development of Winching Pitch Axis-Only Glider Model
>
> I am a pilot, instructor, and retired Electrical Engineer building a next generation active tension-controlled winch for improved safety, height attained, and pilot experience (among other innovations). All the development work is being done open source by myself and a few collaborators and all related technical documents are publicly available. I have no plans to make a business out of this work, i.e. building and selling winches or plans (but an enterprising entrepreneur would be welcome to do so). The investigation outlined below would be an opportunity for someone with the appropriate skills to make a significant contribution to the understanding of winch launch dynamics to our sport’s benefit.
>
> A key aspect in closing a tension servo loop is the behavior of a coupled mass-spring-mass system formed by the winch equivalent mass (mostly due to the drum’s angular inertia), the cable’s elasticity, and the glider mass. This system’s dynamics are well understood during the ground roll and early rotation launch phases but how the glider aerodynamics affects this system following the rotation and into the climb phase is not. Some conceptual theories have been formulated but without a relevant pitch axis model they cannot be evaluated. To that effect, I am looking for help in developing a simple pitch axis-only glider model valid during winching. Conceptually, the thrust force in an airplane’s pitch model would be replaced by the cable tension force acting on a hook displaced from the CG whose pull angle varies throughout the launch. I am an experienced control systems engineer but I and my current collaborators’ limited aerodynamics knowledge is restricting our ability to develop a simple but representative model that encapsulates the fundamental behaviors exhibited by a glider during a winch launch. For full launch simulations, A simple non-linear (lift/drag proportional to airspeed squared but AoA solidly in the linear regime) model where the cable pull angle varies as the launch proceeds is desired. For the tension control systems analysis, linearized models for different launch portions (primarily cable pull angles) are needed. This model would encompass both the phugoid and short period modes. We have studied and basically understand Etkins on these subjects but struggle to decide what is important to include in such a model and what can be ignored while still capturing the essential behaviors. We simply lack the knowledge and experience a skilled aerodynamicist could bring to the investigation.
>
> We would appreciate any assistance a member of this forum might provide. A contributor would likely find this a very fun and intellectually challenging project. Development of an appropriate model requires a strong familiarity with Laplace domain modeling of aerodynamic systems using pole-zero transfer functions. Experience in simulating such systems would be a plus but I have implemented full launch simulations using simpler models using Matlab and Simulink and incorporating a more sophisticated model in these simulations should be straightforward. Several options have been identified for how AoA/airspeed would be controlled, i.e., the pilot stick inputs, but further details are beyond this short introduction’s scope. If you should know of such a person who does not commonly participate in this forum that might be appropriate for this project, I would greatly appreciate your forwarding this solicitation to them.
>
> If you think you might be interested, I would be happy to share the prior work and speculations on this subject (and in general on tension-controlled winching). If after reviewing this material, you decide this is not something you would like to become involved with there would be no obligation; we would be happy you took a look. But maybe you could point us at someone else that might be interested. The level of involvement is totally open. At a minimum, direction as to what to include/not include in an already outlined preliminary model and help in determining some representative values for some of the key parameters, e.g. glider rotational inertias about the pitch axis, would be extremely useful. But what would be ideal would be someone to discuss the problem and issues with and then they take the lead on determining what an appropriate model should be, how it should be parameterized, and evaluating is it providing the essential behaviors important to the larger winch mass/cable spring/glider system dynamics.
>
> Should you be interested and a university educator, this could be a good project a Masters level Aeronautical Engineering student. The results would be publishable—again this is an open source project with everything in the public domain. This could be expanded to a prospective thesis topic. How such an undertaking would be structured is open to discussion.
>
> If you’d like to learn more, send me a note and we can figure out how to go from there. We would probably start with my sharing of some earlier reports, simulations, and logged data from actual launches for your review.
>
> Thank you,
>
> George Moore, CFIG, Ph.D.
> gsmoore501 AT gmail DOT com

Still looking. I would ask members of the forum to review the project and think about other club members or acquaintances that might have the necessary skills and might be interested in helping with this problem.

-George

[SF]

You might have better luck searching fo someone that is looking for a doctoral thesis subject at one of the aeronautical schools.

[[email protected]]

George, I will show/send your posting to some of the Embry Riddle kids we have flying at our winch only club here in Prescott AZ. Not sure if any, about 5 currently, are Aerodynamicists, but they may be interested.

Cliff Hilty CFIG Prescott Area Soaring

CH ASW27

[G Moore]

On Saturday, October 5, 2019 at 7:53:50 AM UTC-7, wrote:
> George, I will show/send your posting to some of the Embry Riddle kids we have flying at our winch only club here in Prescott AZ. Not sure if any, about 5 currently, are Aerodynamicists, but they may be interested.
>
> Cliff Hilty CFIG Prescott Area Soaring
>
> CH ASW27

Cliff,

Thank you very much. Even if they don't have the needed skills or interest, they may know other students that do/are or may run it by a facility member.

[G Moore]

On Saturday, October 5, 2019 at 6:19:32 AM UTC-7, SF wrote:
> You might have better luck searching fo someone that is looking for a doctoral thesis subject at one of the aeronautical schools.

SF,

Yes, I am hoping some forum members may know professors in some of the aeronautical schools with an interest in soaring that they might forward this solicitation on to. As noted in the original posting, this could be expanded to a Master level thesis topic. Not sure if it rises to the level of a PhD thesis.

[[email protected]]

Wanted: Expereienced speel chekr. Univercity levl cumunikashuns skils is sho'nuff perfured. If'n intristid, call BR-549...

[Patrick Grady]

I know how to measure a glider's pitch moment of inertia. In our hangar, we use a hanging cradle to lift a glider for more floor space. This means that we essentially have a pendulum with a glider at the bottom.

You could measure the period of oscillation with a glider and without. The extra mass of the glider wouldn't change the period of oscillation, but the moment of inertia would. I've used this technique to measure rotational inertia of bicycle wheels, and I think it's possible to get within about 5% accuracy.

https://en.wikipedia.org/wiki/Moment_of_inertia#Compound_pendulum

[G Moore]

On Saturday, October 5, 2019 at 1:23:56 PM UTC-7, Patrick Grady wrote:
> I know how to measure a glider's pitch moment of inertia. In our hangar, we use a hanging cradle to lift a glider for more floor space. This means that we essentially have a pendulum with a glider at the bottom.
>
> You could measure the period of oscillation with a glider and without. The extra mass of the glider wouldn't change the period of oscillation, but the moment of inertia would. I've used this technique to measure rotational inertia of bicycle wheels, and I think it's possible to get within about 5% accuracy.
>
> https://en.wikipedia.org/wiki/Moment_of_inertia#Compound_pendulum

Patrick,

Interesting. I had not thought about a pendulum's bob's moment of angular inertia affecting itsperiod. I would assume the glider would have to be rigidly mounted to the pendulum arm for this effect to be exhibited and that might be problematic with your sling.

The key thing I am looking for is the pitch-axis model. Most importantly, as simple a model as possible. I want to understand the fundamentals of what the aerodynamics do the the larger mass-spring-mass model. I often paraphrase a quote that was attributed to Einstein, "Every model should be as simple as possible, but no simpler." Only after that model has been established do I need to parameterize it. I suspect X-Plane, which I have never worked with, would have the parameters for some gliders but I believe the models it uses are more sophisticated than what I am looking for. That is where I would hope an experienced aerodynamicist would be able to determine a proper model for the investigation and then help parameterize it; possibly extracting the needed parameters from simulators like X-Plane.

George

[[email protected]]

It might be worth reading a bit on X-plane. It works starting with a model of the plane's shape and calculates the aerodynamic forces on the fly. There are tools for making models of new planes and a fair library of existing ones.

It is made to talk to other programs in a motion simulator environment. It has an external interface to get out the state vectors and probably put in the controls.

Does anybody know if it has a way to put in the tow rope forces?

[[email protected]]

I looked a little more into x-plane

The X-plugin data interface data items are listed at.
http://www.xsquawkbox.net/xpsdk/docs/DataRefs.html

search for fside_plug_acf
Appears to be 6 data items for extra plugin provided 3d force and 3d moment.

The simulator also include a builtin winch launch system.
Not sure what the model for the winch power curve is, but the parameters controlling it appear to be at /sim/world/winch in the same file.

[G Moore]

On Sunday, October 6, 2019 at 7:30:51 AM UTC-7, wrote:
> I looked a little more into x-plane
>
> The X-plugin data interface data items are listed at.
> http://www.xsquawkbox.net/xpsdk/docs/DataRefs.html
>
> search for fside_plug_acf
> Appears to be 6 data items for extra plugin provided 3d force and 3d moment.

[G Moore]

On Sunday, October 6, 2019 at 12:16:43 PM UTC-7, G Moore wrote:
> On Sunday, October 6, 2019 at 7:30:51 AM UTC-7, wrote:
> > I looked a little more into x-plane
> >
> > The X-plugin data interface data items are listed at.
> > http://www.xsquawkbox.net/xpsdk/docs/DataRefs.html
> >
> > search for fside_plug_acf
> > Appears to be 6 data items for extra plugin provided 3d force and 3d moment.
> >
> > The simulator also include a builtin winch launch system.
> > Not sure what the model for the winch power curve is, but the parameters controlling it appear to be at /sim/world/winch in the same file.
>
> Stu857xx,
>
> Thanks for looking into X-Plane a bit more. The primary motivation for this investigation is to help in determining the tension servo control law characteristics and does it need to change for different phases of the launch.. For that, I need a simple representative model that will capture the glider's aerodynamics effects on the mass-spring-mass system. A suitable pitch model can be linearized for the various phases of the launch for this servo system analysis. The X-Plane model is overkill for the problem to be analyzed. Regardless, I would need a skilled flight dynamicist to determine what elements in the X-Plane simulator are needed and which could be eliminated for the simple model so that brings me back to my original need for a flight dynamicist.
>
> A secondary purpose for this model would be for full launch simulations that will eventually include the full active tension servo system. One place where the X-Plane simulation engine might be very valuable is as a check on the simple model's validity. If full launch simulations were done with both the simplified model and the more sophisticated X-Plane model and there were significant differences in observed behaviors, that could signify there are issues with the simple model, i.e., it is too simple and not representing the behaviors of interest. Again, having someone with experience with such models and simulations would be a tremendous asset.
>
> -George

Stu857,

I read further on X-Plane and learned that its models employ what is termed Blade Element Theory, a theory originally developed in 1879 to study the behavior of propellers, by breaking up the propeller blade into multiple small parts and integrating their effects (according to Wikipedia). This goes to your earlier comments about the simulator building a model based on "the plane's shape and calculates the aerodynamic forces on the fly." I believe such a model that integrates the effects of all the glider's shapes would be very difficult for anyone, even a skilled flight dynamicist, to convert to the kind of model that I need for these control system analyses. I am hoping that a simple pair of coupled systems with the glider's mass and angular inertia comprising the energy storage elements will suffice for my needs. (This goes to the classic short period and phugoid modes of simple pitch axis aircraft models.)

-George

[SoaringXCellence]

George,

I don't see him on RAS anymore, but you might check with Bill Daniels. He has been promoting automated winches for many years and may have details to share.

I'd check the SSA member locator for his contact information.

Mike B.

[G Moore]

On Monday, October 7, 2019 at 11:26:36 AM UTC-7, SoaringXCellence wrote:
> George,
>
> I don't see him on RAS anymore, but you might check with Bill Daniels. He has been promoting automated winches for many years and may have details to share.
>
> I'd check the SSA member locator for his contact information.
>
> Mike B.

Mike,

Thanks. I have worked with Bill for over 15 years on advanced winching. I originated many of the winch tension control, semi-automation, and battery-electric drive concepts employed in his designs. I have discussed the need for this model with him previously and he does not have the needed skills for this analysis. I have been active in the WinchDesign and WinchEngineer winching forums and have made similar solicitations for help there over the years with no success. I decided to try RAS to expand the audience that might contain, or help identify, someone with the right skills that would be interested in this problem. Again, if anyone would read the problem statement at the beginning of this thread and might know of someone that would be a good fit for this problem I would greatly appreciate their forwarding the solicitation to them for their review.

-George

[SF]

If you need a plan B. Instrument the cable with a tension logger. Instrument the winch with throttle position indicator, cable angle & Drum RPM. Take it to a winch clinic. Compare the graphs of the instructor's demo's with the students flights. In my opinion you could do it with speed control, cable angle, ramp rates, and tension limiter overrides. All the winch driver has is the throttle, his process seems pretty repeatable from launch to launch once he gets dialed in on speed. The students actions will be all over the place. Reliable control based on glider pitch angle, and Launch phase combined with the lag in the system with possible signal loss might be a tough go when the pilot does something stupid or the coms break down. These days you can't neglect the inherent liability of designing a fool proof system that encounters a master level fool with a competent lawyer. Good luck, this is interesting stuff, sorry I can't help with the math.

SF

[G Moore]

On Monday, October 7, 2019 at 5:26:40 PM UTC-7, SF wrote:
> If you need a plan B. Instrument the cable with a tension logger. Instrument the winch with throttle position indicator, cable angle & Drum RPM. Take it to a winch clinic. Compare the graphs of the instructor's demo's with the students flights. In my opinion you could do it with speed control, cable angle, ramp rates, and tension limiter overrides. All the winch driver has is the throttle, his process seems pretty repeatable from launch to launch once he gets dialed in on speed. The students actions will be all over the place. Reliable control based on glider pitch angle, and Launch phase combined with the lag in the system with possible signal loss might be a tough go when the pilot does something stupid or the coms break down. These days you can't neglect the inherent liability of designing a fool proof system that encounters a master level fool with a competent lawyer. Good luck, this is interesting stuff, sorry I can't help with the math.
>
> SF

SF,

We have inline instrumentation pods with tension sensing, GPS tracking, and accelerometers. We also have instrumentation packages for monitoring the winch during the launch. We have logged a number of launches on several classic winches. Those logs have validated many of our theoretical models for winch launching. In particular, they show the oscillatory behavior predicted by the mass-spring-mass model during the ground roll and early rotation. But during the climb, the observed behaviors are not so clear. That is why we are looking for a flight dynamicist to help us improve that model to include aerodynamic effects. Knowing how/if the aerodynamics affect this mass-spring-mass oscillatory system will greatly simplify the design of the control law for the tension servo system on our prototype.

-George