Below follows corrections by Doug McLean to my attempt to collect Key Statements by Aerodynamics Expert Doug McLean. As before my comments are in italics in parenthesis.
Do the statements in your summary give correct expression of my views? For the most part, no:
"1. Flight is perfectly understood because fluid flow obeys the RANS equations."
Your comment: "1 is empty: Perfect understanding does not come from looking at RANS numbers. "
This isn't how I would put it. I'd say lift is perfectly understood in a scientific sense because the RANS equations with a no-slip BC and a good turbulence model correctly model the physics of the flow point by point, and solutions to the equations yield the right global flow patterns and flow details. The good agreement with reality verifies that the principles embodied in the equations are the right ones. Thus I say lift is perfectly understood in a scientific sense because we know what principles apply point-by-point in the flow, and we know that to make global predictions, all we have to do is solve a set of PDEs. So the understanding provided by RANS is based on a lot more than "looking at RANS numbers".
"2. Qualitative explanation(s) of lift (are) not essential to the science, and their faults don't contradict my assertion that the science is well understood."
This one is true.
"3. No-slip is enforced in physical terms because the required force arises naturally from the solution to the viscous-flow equations."
C"3 is misunderstanding of the difference between real physics and mathematics."
To conclude that I misunderstand this you have to misrepresent what I said. I was responding to your claim that the no-slip BC is "unphysical" because it contains no explicit reference to a force, and a force would be required to enforce the zero velocity. I replied "Of course forcing the fluid to have zero velocity at the wall requires some applied force, but the required force arises naturally from the solution to the viscous-flow equations." Any reasonable person would agree that both of us, in referring to "force" and "velocity", were referring to the theoretical world modeled by the equations, not to actual physical flows. To accuse me of saying that a mathematical equation can cause a force to arise in the real world is absurd.
"4. New Theory and RANS are equally "deficient" in the sense of failing to provide to a "qualitative" understanding of flow patterns."
Your comment: "4 is misunderstanding of New Theory."
I stand by this one. Your counterargument: "The New Theory comes with an explanation of the generation of large lift at small drag as a consequence of the general explanation of slightly viscous incompressible bluff body flow as potential flow modified by 3d rotational slip separation." is gobbledygook that explains nothing and fails to answer qualitative questions such as why pressure differences and changes in flow velocity appear in the field, and how "3d rotational slip separation" actually modifies the flow.
(Doug admits that New Theory of Flight to him is "gobbledygook", which means that he has not understood. To dismiss something on the only ground that one does not understand, is not the way science is to be performed. The New Theory is not hard to understand since it builds on (i) potential flow modified by (ii) 3d rotational slip separation. I ask Doug: What is it that is "gobbledygook" with (i) + (ii)?)
"5. Headline of Scientific American 2020 (is) sensationalistic nonsense."
This one is true. (This is a most remarkable statement.)
"6. There is no need of New Theory, even if it happens to be correct, because there are already many theories (for different audiences) which are (even if not correct) by the aeronautical/scientific community agreed in broad consensus to model lifting flow correctly to their respective levels of physical fidelity."
No. There is no need of New Theory, even if it happens to be correct, because RANS and all the simplified theories related to it (in retrospect derivable from RANS with simplifying assumptions) already provide our understanding in a scientific sense. The qualitative explanations for different audiences aren't really part of our scientific understanding, as I've explained to you more than once. (Also most remarkable)
"7. New Theory (might) make a contribution for (massively) separated flow."
Perhaps. We'll see.
The no-slip condition also came up in a previous note where you said:
"You say that slip at the wall is fiction, yet you present in your book a turbulent bl in Fig 4.1.14 which meets the wall with effectively slip. Contradiction."
There's no contradiction. On an airliner wing in cruise, the thickness of the viscous sublayer is typically less than 0.2% of the thickness of the TBL, over most of the chord. So the sublayer is too thin to be seen on the scale of plots like fig 4.1.14. But the sublayer is real, and it obeys no-slip at the wall. Modeling it realistically with no slip, as most RANS codes do these days, is preferable taking the shortcut of a slip BC.
"You say that no-slip is a physical boundary but you do not answer my question how in physical terms you can control fluid particles to have zero velocity."
I did answer the question, but you twisted my answer beyond recognition in your response to item 3 above. In an actual physical flow, forces internal to the fluid arise naturally (pressure forces, viscous forces, and Reynolds stresses), and these forces can produce the accelerations needed to meet the no-slip condition. When we model this mathematically, these forces are modeled in the equations that apply throughout the field. This is the basis for my statement in item 3 above. I'm not saying that the equations "control" the flow; I'm saying that actual physical forces control the flow. No-slip is a valid BC for the equations. There's no need for force to appear explicitly in the BCs, as you seem to believe.
"What is the physics on a microscope level that realises no-slip? "
See p. 15 of my book for a simplified explanation. (Why simplified? What is a real explanation?)
Conclusion: Your blog page "Key Statements by Aerodynamics Expert Doug McLean" misrepresents most of my statements. Please revise it to reflect more accurately what I've actually said.
(Let me here try with a new short summary of McLean's views):
- Lift is perfectly understood in a scientific sense because the RANS equations with a no-slip BC and a good turbulence model correctly model the physics of the flow point by point, and solutions to the equations yield the right global flow patterns and flow details.
- Qualitative explanation(s) of lift (are) not essential to the science, and their faults don't contradict my assertion that the science is well understood.
- Forcing the fluid to have zero velocity at the wall requires some applied force, but the required force arises naturally from the solution to the viscous-flow equations.
- New Theory is gobbledygook.
- Headline of Scientific American 2020 (is) sensationalistic nonsense.
- There is no need of New Theory, even if it happens to be correct, because RANS and all the simplified theories related to it (in retrospect derivable from RANS with simplifying assumptions) already provide our understanding in a scientific sense.
- New Theory might perhaps make a contribution for (massively) separated flow.
(Let's see if McLean accepts this as a summary of his views. On 7: All bluff body flow (wings, airplanes) is (massively) separated flow.)
Question to Doug:
You admit that New Theory of Flight is "gobbledygook" to you, which means that you have not understood. To dismiss something on the only ground that one does not understand, is not the way science is to be performed. The New Theory is not hard to understand since it builds on (i) potential flow modified by (ii) 3d rotational slip separation. What is it that is "gobbledygook" with (i) + (ii)?)
Answer by Doug:
You've taken what I said our of context. This comes from my response to your item 4, which specifically referred to the fact that CFD calculations, including both RANS and your New Theory, don't provide any "qualitative" explanation of flow patterns, the kind of explanation that would be understandable to a non-technical person.
Your response was gobbledygook in the sense that a non-specialist would have no idea what you're talking about. I also stand by my statement that, even to a specialist, it doesn't explain anything about lift in a qualitative flow pattern sense.
No, I understand your New Theory just fine. I just contend that it hasn't added anything to our understanding of lift.
My new questions:
- So the theory is not "gobbledygook" to you, only to a "non-technical person"?
- Ok Doug, you say you understand New Theory. I then ask you: Do you think New Theory is correct or not correct? If not, what is wrong?
- What qualitative flow pattern would you like to see explained, which you do not find in the New Theory? Attachment on top? Separation without pressure rise at trailing edge? Large lift at small drag?
(Truth is that experiments for NACA0012 at zero angle of attack shows C_D = 0.006 without tripping and C_D = 0.008 with artificial tripping, while Euler CFD with zero skin friction gives C_D = 0.006 in close agreement with experiments without tripping. Now, a wing on an airplane does not have a tripping device (ribbon) on the leading edge, and so Euler CFD matches the real case. What is then the role of tripping in the experiment, when the real case does not have tripping? (Q1)
The motivation given in the standard literature is that the the ribbon is there to guarantee that the boundary layer is turbulent and so does not separate until the trailing edge, which follows from an expectation that a laminar boundary separates early. Is this also your answer? (Q2)
With tripping C_D increases with 0.002 which can be seen as an effect of changed geometry with a turbulent wake after the ribbon as additional form drag, which is identified as skin friction drag. Do you see that it is not so clear how to distinguish between form drag and skin friction drag in tripped experiments? (Q3)
If now the effect of tripping is to increase skin friction contribution to C_D with 0.002 as an effect of a turbulent boundary layer, what does that say about skin friction contribution without tripping? Still dominating? If tripping adds 0.002 skin friction, how much skin friction is there without tripping, smaller or bigger than 0.002? (Q4))
The motivation given in the standard literature is that the the ribbon is there to guarantee that the boundary layer is turbulent and so does not separate until the trailing edge, which follows from an expectation that a laminar boundary separates early. Is this also your answer? (Q2)
With tripping C_D increases with 0.002 which can be seen as an effect of changed geometry with a turbulent wake after the ribbon as additional form drag, which is identified as skin friction drag. Do you see that it is not so clear how to distinguish between form drag and skin friction drag in tripped experiments? (Q3)
If now the effect of tripping is to increase skin friction contribution to C_D with 0.002 as an effect of a turbulent boundary layer, what does that say about skin friction contribution without tripping? Still dominating? If tripping adds 0.002 skin friction, how much skin friction is there without tripping, smaller or bigger than 0.002? (Q4))
And here I'm talking about conventional calculations for which the turbulence model was calibrated to flat-plate TBL data and used unaltered for everything else, on whatever portion of the surface that the BL was predicted to be turbulent. Your claim that conventional methods can't predict drag correctly for wings, etc. unless the turbulence model is tuned on a "case-by-case" basis is false.
(No, Doug Standard CFD with wall and turbulence models contain parameters which are determined on a case-by-case basis. Do you really claim that there is one choice of parameters for all cases? If so, where can I find the values of these parameters? (Q5))
Your reasons for rejecting the conventional theories at other levels (circulation theory, lifting line, BL theory, etc.) are also false, in my opinion.
I'm not alone in this assessment. You've had some years now to make your case, and it seems you've convinced very few, if any, of the cognoscenti in the field. As a counter to this general line of argument, you could cite the example of plate tectonics, which was widely rejected in its field for many years before finally becoming the dominant paradigm. Given the evidence I've seen, however, I'd say the New Theory has practically zero chance of becoming the plate tectonics of high-Re aerodynamics.
What qualitative flow pattern would I like to see explained, which I do not find in the New Theory? Just about any qualitative feature of the flow: flow turning, changes in flow speed, differences in pressure, etc. You can observe these things in your numerical solutions, but that doesn't explain in a qualitative cause-and-effect sense how they come about. The New Theory is similar to conventional CFD in this regard.
(No Doug, if you do not see that New Theory explains large lift at small drag in physical terms as (i) attachment because the flow is potential before separation, and (ii) 3d rotational slip separation without pressure rise at trailing edge, then you have not understood the physics of New Theory. What is it with (i) and (ii) that you do not understand? And if you understand (i)+(ii), what is it that you do not understand concerning consequences of (i)+(ii) on lift and drag? (Q6))
You know where I stand. I think the other Wikipedia editors have already made a valid case for not including the New Theory in the lift article. I hope that their arguments prevail and that I won't have to take part in that discussion.
Answers by Doug:
(Euler CFD with zero skin friction gets untripped C_D correct (=0.006) within measurement accuracy, with thus form/pressure drag = 0.006. Standard CFD claims 50-80% skin friction drag which with the same total drag means that form drag is max 0.003 = max 50 percent of form drag from Euler CFD. The reason is probably that Standard CFD does not catch the correct separation (too high pressure at the trailing edge) and so gets too small form drag. Standard CFD in this respect resembles potential flow separation with pressure rise at separation and zero form drag.)
And you're wrong about turbulence models needing to be adjusted "case-by-case". Turbulence models vary widely in quality. There are bad ones that probably do need case-by-case adjustment. Perhaps this is what misled you to think that this is generally true of turbulence models. But it isn't. Good turbulence models don't require case-by-case adjustment. I know this from decades of experience using several different turbulence models.
And you're wrong about turbulence models needing to be adjusted "case-by-case". Turbulence models vary widely in quality. There are bad ones that probably do need case-by-case adjustment. Perhaps this is what misled you to think that this is generally true of turbulence models. But it isn't. Good turbulence models don't require case-by-case adjustment. I know this from decades of experience using several different turbulence models.
(How do you distinguish good turbulence models from bad?)
Your new questions Q1-Q7 have easy answers, none of which would support your arguments. But, given that you don't seem to actually read my answers, I think continuing to answer your questions would be pointless, so I'm stopping here. My joining a technical discussion on Wikipedia would be equally pointless.
Your new questions Q1-Q7 have easy answers, none of which would support your arguments. But, given that you don't seem to actually read my answers, I think continuing to answer your questions would be pointless, so I'm stopping here. My joining a technical discussion on Wikipedia would be equally pointless.
(Ok, it seems that our discussion has come to an end since you do not want to answer my Questions Q1-Q7. But the question "what keeps planes in the air" remains to be answered and this will form next headlines in Science and Nature. I want to thank you for your participation displaying the state of modern fluid mechanics, which will serve an important role as reference in the scientific discussion, which will continue until basic questions have found answers. This is the way science works, or is supposed to work. Best regards, Claes.)
Inlägg
Inga kommentarer:
Skicka en kommentar