Loading ...
Sorry, an error occurred while loading the content.
 

Re: [AWECS] Re: Reliable Kite Auto-Pilots

Expand Messages
  • dave santos
    Massimo,   Allow me to assert that we are not actually debating, just clarifying details, as i greatly admire your work & consider us to share core AWE
    Message 1 of 19 , Jan 10, 2011
      Massimo,
       
      Allow me to assert that we are not actually debating, just clarifying details, as i greatly admire your work & consider us to share core AWE design philosophy.
       
      "Autopilot" is traditional nomanclature & still covers the subject we are discussing (fully autonomous flight, including take-off & landing). "Flight controls" is the superset (including engine controls, control surfaces & linkages, etc.). I come from both a robotics & aviation background & will continue to use "autopilot" as a proper term for any autonomous flight controller. Perhaps you mean that a kite is not an aircraft, which was the FAA default postion, but we know it is, & the FAA is now ready to agree-
       

      Aircraft flight control system - Wikipedia, the free encyclopedia

      Turbulence is not a trivial aviation issue. It continues to drive flight procedures & pilot training, & cause accidents. In particular CAT (Clear Air Turbulence) is an issue for visual piloting, as it can occur unexpectedly & invisibly. Avoidance is the primary strategy & few are aware how often modern aircraft still change airspeed, altitude, or course, to stay in smoother air. Our challenge is to fly 7/24 in the inherently turbulent Boundary Layer that most aircraft avoid. A dedicated kite flyer soon learns turbulence is a real challenge.
       
      A note to Stefano about Big Dog on ice: Yes, ice is a non-linear trigger (stick & slip physics). Marc's early robots where single legged pogo-sticks & would fail the ice test, but a four-legged robot is far more robust, due to the redundant legs. This effect is not a breakthrough in advanced controls; a centipede, after all, would do even better on ice. Similarly, kites teamed together in trains & arches exhibit what KiteLab calls Aggregate Stability, where the local chaos of a single kite is cancelled out by the overall stability. This is a great illustration-
       
       
      daveS
       
       
      PS let me know if these links are breaking when pasted
       

    • christopher carlin
      I don t disagree that very difficult control challenges can be fixed with sophisticated control systems. However the control systems designer has a
      Message 2 of 19 , Jan 10, 2011
        I don't disagree that very difficult control challenges can be fixed with sophisticated control systems. However the control systems designer has a responsibility to push back  hard against aerodynamics and performance people who want to squeeze the last small percentage of performance out of a system and in the process create something that requires a sophisticated control system to operate. Developing the sophisticated control system including certification and providing redundancy if it's flight critical is almost certain to be expensive, time consuming, and heavier than might otherwise be the case. Bottom line select a kite system concept which is simple and stable. Figure out its performance. Then compare other more complex concepts to it. In order to select a more complex concept you must demonstrate that the net economic benefit including realistic control system development and hardware costs is really there. 

        It is the nature of the of the aerospace business to create complex high performance systems. As evidenced by the A 380 and the Boeing 787 they tend to have schedule delays because each piece of the system is stretching the state of the art to the maximum to gain a competitive advantage. I could site many examples of military programs with similar problems and also some wave power generation systems which have had inordinate delays. I can also site two of the greatest development efforts in terms of getting a new technology into production quickly. One is George Stephenson's invention of the locomotive and railway system. The other is Bill Gate's introduction of software for PCs. Both initial deployments were far less than optimal but they got to market first and became commercially successful as a result. In Stephenson's case Brunel's competitive broad gauge system sank only to reborn almost 150 years later in the TGV. In Gate's case Apple's superior product was almost sunk by Microsofts lower acquisition cost although it is now doing well.

        My point is forget fancy concepts. Go with something easy and relatively inexpensive to build, deploy and control. Wherever possible use existing hardware. I agree with the definition of autopilots and flight control/stability augmentation systems. Ideally you want a kite system which if it needs control at all only needs an autopilot. The basic system needs to be inherently stable and able to ride through turbulence on it's on. 

        Chris    
        On Jan 9, 2011, at 9:45 PM, mmarchitti wrote:

        Dave, a traditional aircraft can fligh also inside high turbulence conditions: I think we all have experienced the shaking effect that produces the phenomenon inside the aircraft cabin, and in small aircraft it is more remarkable. The weight and geometry configuration of a traditional aircraft stabilize the flight after it has been perturbated, without touching the command.

        Very different are the unstable aircraft, that show the instability characteristic in pitch control. However, imagine for the yaw control to move the fin from the rear fuselage to the front: in that case you will have to continuosly and very rapidly operate with the rudder, also in very quiet weather condition, impossible to do with human control... well it depend on the degree of instability.

        I mean that control law techniques today are very elaborate and accurate, and I think they can solve the kite flght.

        Just to be precise: the "autopilot", in the aeronautical jargon, is the system that authomatically follows a route or keeps some flight parameters as the speed, the altitude ecc. Whereas, in our case, the flight control system is the correct name. The difference is remakable: in the automotive situation the autopilot could be the constant speed device, whereas a sort of drive control system could have been the active shock absorbers.

        --- In AirborneWindEnergy@yahoogroups.com, dave santos <santos137@...> wrote:
        >
        > Massimo, 
        > 
        > Boston Dynamics' Big Dog is a great example of complex expensive & ultimately still brittle autonomy. I have been a fan of Marc Raibert's legged robots for thirty years & we attended many of the same AI & robotics conferences. His state machine designs are maximally elegant & strongly influenced my own high-DOF robots. While rough terrain is a challenging robot environment, windfield chaos is far more dynamic & complex, & the CFD problems are far less tractable, so we need to rethink control (see Chris Carlin's bottom note).
        > 
        > You wrote-
        > 
        > 
        > ========================
        > Re: [AWECS] Control Engineering for AWE      
        > Mon Jun 14, 2010 6:15 pm 
        > 
        > Dave, 
        > 
        > Don't know if you're a controls guy by trade or have just read a lot. Anyway I spent 40 years in the business and from what I understand of kites of the size and complexity we're talking about I think the controls problem is challenging to say the least. Not only are the kites not easy to model but the turbulent atmosphere is very difficult to deal with analytically particularly if you're trying to guarantee crash free operation over a long period of time. And of course the whole state space thing is based on a linear world. NLMPC is a fine idea conceptually but very challenging for the devices we're talking about. Should you seriously start to develop something I'd be interesting in kibitzing it a bit. My basic advice is to develop a kite which for the most part flies itself and only needs modest effort by the control system - for example reefing in strong winds.
        > 
        > Chris
        > 
        > ===================================
        > 
        > PS Your English is quite good, still, i am sorry about the language divide & do not wish to contend tech issues with unfair advantage. Many thanks to Carlo, Mario, & Ugo for making the great Italian AWE work more accessible.
        > 
        > daveS
        >


      • mmarchitti
        The aeronautical industry has always streched the limit of the available technology, starting with the Wright brothers. In my opinion the actual delays of
        Message 3 of 19 , Jan 11, 2011
          The aeronautical industry has always streched the limit of the available technology, starting with the Wright brothers. In my opinion the actual delays of these two big projects (you should add also the A400) are determined by the industrial organization that has become excessively large, bureaucratic.

          KiteGen can be fancy if you think that a kite based system could be more powerful than a huge and heavy wind tower. But
          KiteGen is also a simple concept, when it is understoot: in the case of the Stem version, a cross wind motion that produce lift, combined with an ascensional motion that produce power. However the present technology that is used has to be judiciously assembled, a very demanding task.


          --- In AirborneWindEnergy@yahoogroups.com, christopher carlin <christopher.m.carlin@...> wrote:
          >
          >
          > It is the nature of the of the aerospace business to create complex
          > high performance systems. As evidenced by the A 380 and the Boeing 787
          > they tend to have schedule delays because each piece of the system is
          > stretching the state of the art to the maximum to gain a competitive
          > advantage. I could site many examples of military programs with
          > similar problems and also some wave power generation systems which
          > have had inordinate delays. I can also site two of the greatest
          > development efforts in terms of getting a new technology into
          > production quickly. One is George Stephenson's invention of the
          > locomotive and railway system. The other is Bill Gate's introduction
          > of software for PCs. Both initial deployments were far less than
          > optimal but they got to market first and became commercially
          > successful as a result. In Stephenson's case Brunel's competitive
          > broad gauge system sank only to reborn almost 150 years later in the
          > TGV. In Gate's case Apple's superior product was almost sunk by
          > Microsofts lower acquisition cost although it is now doing well.
          >
          > My point is forget fancy concepts. Go with something easy and
          > relatively inexpensive to build, deploy and control. Wherever possible
          > use existing hardware. I agree with the definition of autopilots and
          > flight control/stability augmentation systems. Ideally you want a kite
          > system which if it needs control at all only needs an autopilot. The
          > basic system needs to be inherently stable and able to ride through
          > turbulence on it's on.
          >
          > Chris
          > On Jan 9, 2011, at 9:45 PM, mmarchitti wrote:
          >
          > > Dave, a traditional aircraft can fligh also inside high turbulence
          > > conditions: I think we all have experienced the shaking effect that
          > > produces the phenomenon inside the aircraft cabin, and in small
          > > aircraft it is more remarkable. The weight and geometry
          > > configuration of a traditional aircraft stabilize the flight after
          > > it has been perturbated, without touching the command.
          > >
          > > Very different are the unstable aircraft, that show the instability
          > > characteristic in pitch control. However, imagine for the yaw
          > > control to move the fin from the rear fuselage to the front: in that
          > > case you will have to continuosly and very rapidly operate with the
          > > rudder, also in very quiet weather condition, impossible to do with
          > > human control... well it depend on the degree of instability.
          > >
          > > I mean that control law techniques today are very elaborate and
          > > accurate, and I think they can solve the kite flght.
          > >
          > > Just to be precise: the "autopilot", in the aeronautical jargon, is
          > > the system that authomatically follows a route or keeps some flight
          > > parameters as the speed, the altitude ecc. Whereas, in our case, the
          > > flight control system is the correct name. The difference is
          > > remakable: in the automotive situation the autopilot could be the
          > > constant speed device, whereas a sort of drive control system could
          > > have been the active shock absorbers.
          > >
          > > --- In AirborneWindEnergy@yahoogroups.com, dave santos
          > > <santos137@> wrote:
          > > >
          > > > Massimo,
          > > >
          > > > Boston Dynamics' Big Dog is a great example of complex expensive &
          > > ultimately still brittle autonomy. I have been a fan of Marc
          > > Raibert's legged robots for thirty years & we attended many of the
          > > same AI & robotics conferences. His state machine designs are
          > > maximally elegant & strongly influenced my own high-DOF robots.
          > > While rough terrain is a challenging robot environment, windfield
          > > chaos is far more dynamic & complex, & the CFD problems are far less
          > > tractable, so we need to rethink control (see Chris Carlin's bottom
          > > note).
          > > >
          > > > You wrote-
          > > >
          > > >
          > > > ========================
          > > > Re: [AWECS] Control Engineering for AWE
          > > > Mon Jun 14, 2010 6:15 pm
          > > >
          > > > Dave,
          > > >
          > > > Don't know if you're a controls guy by trade or have just read a
          > > lot. Anyway I spent 40 years in the business and from what I
          > > understand of kites of the size and complexity we're talking about I
          > > think the controls problem is challenging to say the least. Not only
          > > are the kites not easy to model but the turbulent atmosphere is very
          > > difficult to deal with analytically particularly if you're trying to
          > > guarantee crash free operation over a long period of time. And of
          > > course the whole state space thing is based on a linear world. NLMPC
          > > is a fine idea conceptually but very challenging for the devices
          > > we're talking about. Should you seriously start to develop something
          > > I'd be interesting in kibitzing it a bit. My basic advice is to
          > > develop a kite which for the most part flies itself and only needs
          > > modest effort by the control system - for example reefing in strong
          > > winds.
          > > >
          > > > Chris
          > > >
          > > > ===================================
          > > >
          > > > PS Your English is quite good, still, i am sorry about the
          > > language divide & do not wish to contend tech issues with unfair
          > > advantage. Many thanks to Carlo, Mario, & Ugo for making the great
          > > Italian AWE work more accessible.
          > > >
          > > > daveS
          > > >
          > >
          > >
          > >
          >
        • Doug
          Slowly all the concepts drift toward propellers, just as they have for 3000 years. Sure, it s a kite . Then make it longer. Then make it stronger. Then
          Message 4 of 19 , Jan 12, 2011
            Slowly all the concepts drift toward propellers, just as they have for 3000 years.
            Sure, it's a "kite". Then make it longer. Then make it stronger. Then chase a circular path. At some point you might just break down and call it a blade and admit you're trying to fly a wind turbine as an autogyro, the rotor of which IS a wind turbine rotor. Little difference, if any. Noticing that a central attachment point magically controls pitch while insuring that circular path, and handling centrifugal force without wasting aerodynamic lift for steering, and you've re-"discovered" the propeller. Congratulations! Wow, a "hub". Yup, that IS the way to do it in the actual field of wind energy. Yup a circular path and a hub. Who knew?
            By the time they eliminate the secondary propellers, take their power from the center, and add several levels, they will have a Superturbine(R)!
            :)
            All roads lead to Superturbine(R)
            Doug S.

            --- In AirborneWindEnergy@yahoogroups.com, "mmarchitti" <marchitti@...> wrote: SEE
          Your message has been successfully submitted and would be delivered to recipients shortly.