torsdag 13 september 2012

AIAA Defends Old Incorrect Theory of Flight 3

Let us continue the analysis of the message from associate editor Greg Blaisdell when rejecting our article New Theory of Flight submitted to AIAA Journal.

Blaisdell formulates his goal to be to re-educate (or brain-wash) us and teach us the Old Theory, which is exactly the theory which we have studied very carefully and found to be incorrect:
  • I can see several misconceptions you have concerning basic concepts in fluid mechanics.
  • My goal below is ... to help you see some of the areas where you need to improve your understanding.
  • ...there are basic concepts in fluid mechanics that you do not understand.
  • What you seem to be missing ... is that this (our) view of the classical theory is “truncated”; it is not a correct view.... it is the truncated form of the classical theory that you object 
  • Unfortunately, the complete theory is not always taught, with the result that many students have misunderstandings concerning flow over airfoils and wings.
  • Another set of basic concepts that you misunderstand deal with circulation and vorticity. They are related, but they are not the same, and both are important to understanding the fluid mechanics of airfoils and wings.
In his re-education mission Blaisdell seeks to force us to accept precisely what we have shown to be incorrect physics:  
  • At small angles of attack the skin friction contribution can be much larger than the contribution from the pressure; while pressure or form drag is more important at higher angles of attack where the airfoil acts more like a bluff body.
  • The trailing edge vortices in your simulations form in counter-rotating pairs. Their net circulation is zero, and their presence does not alter the circulation or the lift on the airfoil. 
  • What reviewer 1 stated about vorticity not being generated in the interior of the flow under the assumptions of (i) incompressible flow and (ii) a conservative body force, is correct.
  • Kelvin’s theorem, that circulation does not change for a closed loop moving as a material curve (moving with the fluid), only depends on the assumptions of (i) inviscid flow, (ii) incompressible (low-speed) flow, and (iii) a conservative body force (e.g., gravity).
However, Blaisdell is struck by disbelief from the mounting difficulties in his presentation and seeks to encourage himself by twisting lack of evidence into its opposite:   
  • Many of the concepts associated with viscous boundary layers, viscous-inviscid interactions, circulation, vorticity and the generation of lift, are not intuitively obvious. 
  • The generation of vorticity on solid surfaces is one of the more difficult subjects included in my graduate introductory fluid mechanics course. 
  • The fact that some of the concepts are difficult to understand does not make them wrong.
Blaisdell's long lecture is essentially a repetition of the long lecture by Reviewer 2 analyzed in a previous post. 

Like Blaisdell, Reviewer 2 not only lectures but also admits that aerodynamics education of today is a mess:
  • Aerodynamics today is therefore almost always taught in a truncated version that retains all of the utility, but has lost much of the profundity. 
  • Even the truncated version is no longer as highly respected as it used to be, because Computational Fluid Dynamics delivers, with no requirement for deep thought, most of the practical answers that are needed. In consequence, there are many employed today in the aerospace industry, and even in academia, whose grasp of the basic theory of flight contains many gaps. 
  • These gaps are apparent to thoughtful students, who frequently attempt to fill them in for themselves, although the remedy is usually worse than the disease. 
  • ... simplified or discarded in modern presentations to create a pragmatic treatment focusing on utility. 
  • Undergraduate textbooks these days all too often simply omit anything that students find difficult.
  • I believe that the authors ... are right to quarrel with the truncated version that they, like others, have apparently received. 
Reviewer 2 cannot but acknowledge important merits of our article, which Blaisdell ignores:
  • The authors have put their fingers accurately on many of the defects in the truncated versions of aerodynamic theory that are now current.
  • However, the authors are correct that separation might be fundamentally different in 3D than in 2D. 
  • The authors have then sought their own explanations, stimulated by interesting results from their Navier-Stokes code.
Reviewer 1 acknowledges one of our key points:
  • At most, perhaps they present a numerical model of the governing equations which avoids the need to discretize the boundary layer.
But Reviewer 1 does not understand that this suddenly makes the Navier-Stokes equations computable from being uncomputable by Prandtl's dictate of boundary layer resolution, and thus fundamentally changes both theory and practice of aerodynamics.

Summary: What is specially remarkable is the fervor of the re-education the editor and reviewers want us to undergo (like e.g. the Emperor in Red China): They want to be sure that we come out of the process fully re-educated to the correct belief of basic concepts.  No effort should be spared to reach this goal, as if our ideas somehow are dangerous to AIAA and thus have to be suppressed.  But why spend so much energy on cranks?

13 kommentarer:

  1. If this new theory works, then it could be used to calculate something that old/others modells fails to calculate/predict or?
    Is there no experimental data that will prove or disprove this new modell?
    /Jocke F

    SvaraRadera
  2. We solve the Navier-Stokes equations with slip boundary condition (without turbulence models beyond numerics), and we thus can accurately compute e.g. lift and drag of arbitrary objects. With conventional theory this is simply impossible since resolution of thin boundary layers asking for quadrillions of mesh points, or unknown turbulence modeling, is required. The new theory opens to a vast area of applications within computability.

    SvaraRadera
  3. so why don't you present your theory as an alternative model, show an example in which it gives the same result as the other (that is an example that lies in the realm of validity of the other theory) and one example where you calculate something that the other cannot calculate instead of focussing so much on saying that the old theory is incorrect ? (which means offering your side to the attack of those who are the paladins of the old theory and who have the power)

    PS
    i expect an answer talking about contradictions that are unstable and will lead to a catastrophe or a conspiracy theory about the entire world being against you. but maybe this time you will surprise me!

    SvaraRadera
  4. We have already done what you ask for. Read and you will see. As concerns conspirracy I say what Gore Vidal says: of course there is a lot of conspiracy around; what is a political party but a conspiracy to take power?

    SvaraRadera
  5. lorenzo, your analysis is brilliant! As we say in Sweden: "som man bäddar får man ligga" which here implies: if you attack people by saying "You are wrong!" it is obvious they will defend themselves with all their weapons they have.

    Claes: do you have any plans of the comparison lorenzo suggests?

    Anders

    SvaraRadera
  6. We compute lift and drag in accordance with observation for all angles of attack by solving the NS eq with a posteriori error estimation and without turbulence model. This is new and opens a whole new field of computational exploration. Science is not like war or love, where all weapons can be used, but should be a fair game, where arguments can meet. For example, refusal to speak out is a tactic which should not be allowed, but is used by my collegues at KTH.

    SvaraRadera
  7. You answer as a politician... Isn't it possible to answer yes or no to my question?

    I agree with you that you collegues at KTH should answer your earlier questions. Especially in the light that AIAA recommends this communication (Greg Blaisdell: "... your colleagues in the Mechanics Department at KTH, and I have a lot of respect for their knowledge of fundamental fluid mechanics and applied mathematics. I strongly suggest you talk with them..."). Is it really so that your collegues haven't answered at all?

    Anders

    SvaraRadera
  8. Yes, it is so, they refuse to communicate with me.

    We have already given example where our theory coincides with the old theory (lift) for small angle of attack while it differs for larger angles of attack (lift and drag). This is not a politician's answer. It is straight and clear if you only read what we write.

    SvaraRadera
  9. PS They have good reasons to refuse comparison of arguments. Ours are very strong.

    SvaraRadera
  10. Anonym
    If this new theory works, then it could be used to calculate something that old/others modells fails to calculate/predict or?
    Is there no experimental data that will prove or disprove this new modell?
    /Jocke F
    Claes
    We solve the Navier-Stokes equations with slip boundary condition (without turbulence models beyond numerics), and we thus can accurately compute e.g. lift and drag of arbitrary objects.
    Lorenzo
    so why don't you present your theory as an alternative model, show an example in which it gives the same result as the other (that is an example that lies in the realm of validity of the other theory) and one example where you calculate something that the other cannot calculate instead of focussing so much on saying that the old theory is incorrect ?
    Claes
    We have already done what you ask for. Read and you will see.
    Anonym
    Claes: do you have any plans of the comparison lorenzo suggests?
    Claes
    We compute lift and drag in accordance with observation for all angles of attack by solving the NS eq with a posteriori error estimation and without turbulence model. This is new and opens a whole new field of computational exploration.
    Anonym
    You answer as a politician... Isn't it possible to answer yes or no to my question?
    Claes
    We have already given example where our theory coincides with the old theory (lift) for small angle of attack while it differs for larger angles of attack (lift and drag). This is not a politician's answer. It is straight and clear if you only read what we write.

    I am confused, y'all ask for proof and Claes says he has already provided it. Claes, could you point to particulars in your book/paper/ wherever with your proofs??

    SvaraRadera
  11. We compute lift and drag of wing for all angles of attack (or arbitrary body like car) by direct solution of Navier-Stokes equations without user input of extra information such as turbulence model, wall model, Kutta condition et cet. This is new.

    SvaraRadera
  12. OK, thanks for yor reply.
    But I guess it is difficult to convince the academic world. Since they in 95% is not interested in the "thruth". Then you have to go to the people who "need" accuarate results.
    What do industry say?, i guess you must have cooperation with the industry (automotive producers and others that are interested in simulation of fluid flows)
    //Jocke F

    SvaraRadera
  13. Yes, we have, but science/fluid mechanics is like the Catholic Church with the truth given by cardinals at the university following the dictate of Prandtl the Pope.

    SvaraRadera