Hardfire: Physics of 9/11

Niclas said:
Perhaps i am misunderstanding you, but i see no mention of scale in
Ryans experiment, it seems to be independent of the lenght of the steel,
"h" in the model, is it not the relationship between "M" and "h" that
matters?
Click to expand...

Yes,misunderstanding.(From when second is measurment of lenght?)
I reffered to impulse of force and time.

And cannot find "M".Where is it?Is it in video or in downloadable presentation?
 
GregoryUrich said:
Hi Ryan,

Thanks for your video. I am looking forward to parts II - ?

If I understood correctly you stated that the pressures involved in the simple fuel impact model were approaching what is needed to fail the columns. I'm concerned that this will be misunderstood. The pressures you calculated were around 4000-9000 psi but the yield stress of the weakest columns was around 32,000 psi. A column may indeed fail due to the force caused by pressure over a large enough area but the pressure itself is not nearly sufficient.
Click to expand...

That is correct. That is why it takes both pressure and impulse to fail a steel column member -- both to sustain the displacement, and to do so over a sufficient area that local stresses exceed the ultimate stress in the member.

This is also one of the ways in which pressure and stress are not the same. They have the same units, but they behave somewhat differently.

To everyone else, what Gregory is referring to is the yield stress -- the ultimate stress -- in the steel. What that means is if you measure the stress in the steel, at any point, and it exceeds this value, the steel is likely to rupture.

In the simplest case, suppose we have a block of steel that is completely supported by an immovable, unbendable plate on one side, and we apply a constant pressure on the other side. The stress in this case is equal to the pressure, so it will take 32,000 psi (referred to in NIST as 32 KSI, for "kilopounds per square inch") to cause the steel to fail.

What really happens, however, is geometry in the problem can create stresses much higher than the actual pressure. For instance, the perimeter columns were often made from even stronger steel, in places over 100 KSI yield strength -- but they were often very thin, as well. You should intuitively understand that a paper-thin column of 100 KSI steel is weaker than a stout, inches-thick column of 32 KSI steel, and this is absolutely correct. The yield strength of the material is just one of many factors that contributes to the overall strength of a given element, just like the strength of the aircraft is only one contributor to the severity of impact.

In our problem, what we have is a length of column that is supported at top and bottom, and we're hitting it with pressure in the middle. This won't cause the steel itself to totally rupture, but what it does instead is bends the steel column. This bending creates kinks, and at those kinks the stress is considerably higher -- that is where and how the steel will fail. Think of a lever effect if it helps to visualize this process.

You will note in NIST's calculations in Chapter 10, NCSTAR1-2B, there are several different computed values depending on how much area the given input pressure affects. If we have, say, 4800 psi but it only covers a tiny spot in the column, the result is much less severe than if we apply that same pressure over the entire length. That's why they get different values -- the different dots match different parts of the wing hitting a given column, and thus a larger or smaller area of impact.

As a result, there is no conflict between our prediction that 4800+ psi is enough to blaze through steel columns, and the fact that the yield stress is higher than the incident pressure. This is an apples-to-oranges comparison. One needs to track the impulse as well, and one needs to compute the actual failure point of the steel with its geometry in mind (something I do not do in this talk because it would take far too long, but is elementary if you already know solid mechanics).

For sake of comparison, a bomb or a tornado that produces a 15 psi overpressure will destroy virtually any civilian structure, despite being vastly lower than the yield strength of any building material. This is because the pressure affects a huge area and delivers a lot of impulse. Again, pressure and yield strength are simply not directly comparable except in the purely axial strain case.

Niclas said:
If i way, i would also like to pose a question about part 2. Modeling, from
Ryans supplement to the Hirdfire presentation: "some basic physics of 9/11"

I am looking at "Model process", part 2 and 3, and i wonder, how come the impact is only transfered to the columns of the first floor, and not the second? ( and the third and the forth and the fifth, and so on, had there been more floors)

You say: "colums absorb impact until they fail and buckle", how come the first columns are the only ones to absorb the impact?


Or are they not?
Click to expand...

This is a simplification, but one that is actually fairly close to reality.

First, understand that my model is not intended to be a very good model -- in fact, it's as simple as I can make it while still being mostly valid. You will see this when the third show is finished. I'm not claiming my model is the best one, instead I'm using this model to motivate scaling. So the transmission from impact floor to the lower structure is neglected as an assumption. I'm not claiming the model is a perfect reconstruction of reality.

Having said that, the actual force transmission will be low. As Dr. Bazant explained in a reply to Dr. Jones and others, the columns of a given floor simply cannot transmit a force greater than their own yield strength. It's impossible. The columns will buckle before this force ever develops.

Next, we reason that the first column to yield will be the one at the top of the stack, almost without exception. This is for two reasons. First, the columns grow progressively stronger as you move down the structure, since their static loads are higher. Second, the columns on top are the least supported, having lost their bracing as the upper floors are damaged and have come loose.

Because of this, the force required to buckle the uppermost column -- which is the maximum force transmitted -- is insufficient to buckle any column further down. Since we are treating buckling, any column that doesn't actually fail is likely to remain elastic, and thus whatever energy is lost is only lost for an instant, but will be returned as soon as the top column buckles.

There are still other reasons why this mechanism is fairly accurate, principal among them the action of impact and fracture. We aren't applying a steady, ideal axial load to the column stack. Instead we're hitting the column with an eccentric and shearing load, and one moving quite quickly. This favors fracture and shearing of connections leading to further weakening, and this again favors damage near the point of impact rather than loading or damaging the column distributively. A point made here long ago is that the actual contact between upper mass and lower columns is actually likely to be moment rather than axial strike -- twisting the tops of the columns inward or sideways instead of simply hammering straight down on them -- and this type of stress wave, either Rayleigh or Love waves, will not transmit into lower columns efficiently at all owing to the splicing method.

So, in summary, my model is incomplete and I freely admit it. You are invited to come up with a better one. But for the purposes of my discussion, it is sufficient. The actual error introduced by this simplification is extremely minor, indeed smaller than uncertainties in the actual mass of the Towers or understanding about precisely when or how connections would fail.

Good questions. Much better than the only other comment so far from the Truth Movement, namely a claim that "No research organization within the 9/11 truth movement supports the 'no-plane' theories." As I explained before, this misses the point, we only used it as an example. Also I wonder if that applies at the Pentagon as well? :D
 
Last edited:
R.Mackey said:
.....snip for brevity :D
Click to expand...
I think the fact that Pressure and stress have the same units was the confusion. Stress IS pressure, but it is the result of a load.
Applied pressure is a different bag of cats. Pressure (psi) * area=force.
so, a 4000psi pressure applied over 10 square inches of surface area=40,000 lb force.
As Mackey pointed out, this gets applied as a crippling load to an end-loade column, making failure more likely (If the load itself doesnt locally fail the column)
in the end--Stress is the result of an applied load.
You cannot apply stress to anything.
 
Homeland Insurgency said:
No. You compared a ship hit with a bomb to the WTC Mackey. If it's stupid then why did you do it?
Click to expand...


Given that you suggested aircraft might have their engines replaced with bombs in their use as kamikazes, I don't think what you consider as stupid will be given much credence.

You still don't seem to have grasped the basic fact that the aircraft penetrates first, then the bomb goes off.

And you still haven't grasped the fact that the kinetic energy alone involved in the WTC jet impacts was the energy equivalent of a 2,000 lb. bomb.
 
And when will the second Hardfire interview/presentation take place? I look fwd to the video! :)
 
R.Mackey said:
the transmission from impact floor to the lower structure is neglected as an assumption. I'm not claiming the model is a perfect reconstruction of reality
Click to expand...

Thank you for your reply.

Can we make a scaled down model that DOES take this into account?
 
Last edited:
Niclas said:
Thank you for your reply.

Can we make a scaled down model that DOES take this into account?
Click to expand...

Absolutely yes, we can!

This will become clearer once the third show is released, but the purpose of my mathematical model is to motivate construction of a physical model. The mathematics don't capture everything that happens in real life, and they never will -- we're just doing the best we can. But the physical model will have all the features of reality. It is reality.

Now, what might happen is that the effect we're concerned about -- say the transmission of momentum through intact columns, or flexibility of the descending mass -- doesn't scale the same way as the other properties. Without a good analytical explanation of those effects it's hard to say. So the model will contain these effects, but they might be much stronger or much weaker in the model than they were in full scale. This takes some follow-up, but it's a straightforward process.

There are three basic approaches to refining the model:

1. Observe the effect with a subscale model, and try to figure out the equations through curve-fitting and similar data analysis.

This could take a number of trials, high-speed photography, very careful measurement of the model before the test, etc. Or we could set up a more focused experiment just to study this one mechanism, and later apply what we've learned to the original model. But it is possible. More often, we can make some guesses about different mechanisms, and use even rough data to choose between them. Once we have this, see how it scales to model size, adjust the model, and run again for the payoff.

2. Consider a limiting case rather than an accurate model.

Suppose we really don't know how strong the effect is, but we can guess its rough order of magnitude. Then we can adjust the model in some other way such that the effect is definitely negligible, and see what happens. In this case, suppose I double the (model-adjusted) strength of columns? That should easily more than compensate for this effect possibly missing. But if I still predict a progressive collapse, then it definitely doesn't matter.

In this approach you lose resolution -- you may never know whether the collapse should have taken 12 seconds or 15 seconds or 18 or whatever -- but you will know that the collapse makes sense.

3. Combine the effect into one of the other variables of the problem.

In my model, I ultimately conclude that I cannot scale the columns on the basis of these equations alone -- I also need to consider static strength, which scales differently, and there's no easy way to scale a column in two different ways in the same dimension. But what I'm really after is energy absorption above and beyond what the static load inflicts. So I advocate testing model-sized columns individually until they produce the right energy absorption numbers.

No reason we can't fold this effect into that testing as well. Instead of destructively failing just one column on top of an anvil, let's destructively fail a whole stack of columns, and scale accordingly.​

That all applies if and only if you want to be careful.

If you want to be sloppy, when you scale down these effects will be stronger than in the real system. This is for two reasons. First, connections will not shear as easily because there is less leverage in the smaller model. Second, while we've made pieces smaller, we've used the same materials, and so the wave speed in the model is the same as the original; since the model is smaller, this means that stress waves travel farther in the model than they would in real life.

So throwing a WAG at it, it makes no difference. But you can, if you're clever, model it down to the last detail. These effects are no different, and in fact will be present to some degree in any physical model of the system.
 
At about time 24.47 into the show Anders B. Bjorkman (aka Heiwa at JREF) is introduced in some confusing way that has to be clarified at the next Hardfire show next week. It's about not testing models or whatever?? Someone cycling. Whatever. FYI info Anders B. Bjorkman is hitting the ski slopes tomorrow, a quite physical thing, to test his new skies. Anders B. Bjorkman knows that he + skis are non-rigid and will try to avoid crushing anything in the slopes to test Mackey physics. I look forward to next week's show.
 
Last edited:
Robertson's design point of a 707 at 180 mph matches, and you could figure out a plane can be resisted at 180 mph back in the 60s when the WTC was built; Robertson did that.
 
Heiwa said:
At about time 24.47 into the show Anders B. Bjorkman (aka Heiwa at JREF) is introduced in some confusing way that has to be clarified at the next Hardfire show next week. It's about not testing models or whatever?? Someone cycling. Whatever. FYI info Anders B. Bjorkman is hitting the ski slopes tomorrow, a quite physical thing, to test his new skies. Anders B. Bjorkman knows that he + skis are non-rigid and will try to avoid crushing anything in the slopes to test Mackey physics. I look forward to next weeks show.
Click to expand...
They said your failed work was next. If your work was not a failure, you would have a Pulitzer Prize instead of being petitioner of the month for AE nut case idea believers on 911.

Yep, my daughter went skiing last week, Your scary quite physics thing is debunked too.

Bragging about skiing, this is your big thing after failing to get anyone in the real world to see your paper you go skiing. At least you don’t let your utter failure in engineering on 911 issues affect your activities and it is good to see ignorance on 911 does not keep you from physical activity.
 
Last edited:
You must log in or register to reply here.

Back
Top Bottom