• Quick note - the problem with Youtube videos not embedding on the forum appears to have been fixed, thanks to ZiprHead. If you do still see problems let me know.

term question

grandmastershek

Graduate Poster
grandmastershek
Joined
Nov 2, 2009
Messages
1,461
user-defined break element. its found in NCSTAR 1-6. i can't find a definition for it. but i got a twoof telling me it means that it means they had to manually input values to make the collapse mechanism work because the computer wouldn;t recreate the collapse.
 
I guess "user-defined break element" is the same as "pull it"
 
grandmastershek, can you quote the exact line or paragraph from NCSTAR? Maybe seeing it in context would help with a definition. (I'm sure it isn't what twoofer thinks/claims it means.)
 
From NCSTAR 1-6

"Break elements (ANSYS user defined element) were developed for use in the full floor model that captured the temperature dependant failure modes of the truss seat connections, strap anchors and spandrel studs. these elements replaced the detailed 3-D solid finite element model used for the truss connection analysis, allowing a reduction in model size.

A break element is a unidirectional spring element that can simulate connection failure by disconnecting two "active nodes" when the relative displacement between two "control nodes" exceeds a specified threshold "
Click to expand...


So my understanding is that they make a smaller model that works the same way as the more detailed model so that when you model an entire skyscraper it doesn't use more space than you have on your computers.
 
I don't know ANSYS, but this sounds pretty straightforward: ANSYS probably comes with a large set of predefined structural elements, each with predefined properties. Since ANSYS does not know, how you will use them (which properties will be significant for your particular simulation), they model each element rather detailledly and complexely. However, in many simulation programs, you can define your own elements. Often, because none of the predefined elements model your real world element well enough (user-defined is better than pre-defined), or (in this case) to reduce complexity and help computing time and resources along.

There is no reason to think user-defined elements are any less realistic or applicable than pre-defined elements. In fact, they may be better suited, and my best guess is, that's why NIST defined some: To make a better model.
 
Last edited:
From my limited knowledge of finite element modeling: a 3-D finite element model of the truss connection, which uses a mesh or grid of nodes to represent discrete points in 3 D space and takes a lot of computation, was replaced by a simpler model where the meshes representing the trusses and columns were connected by "springs" which would disconnect at preset changes in their separations. So each connection is represented by 1 entity rather than a bunch of individual points.

Since even the finite element model needs parameters on which to determine failure, having to input failure values for the simpler model is not some big deal. Truther needs to show that the values used were unreasonably low, or were tweaked until they gave the desired result.
 
Last edited:
You ought to ask the mods to change the title to "ANSYS terminology question" or something that'll catch the attention of engineering folks here, like Newton's Bit, Ryan Mackey, rwguinn, etc. I think Newton's Bit actually uses it regularly, and while I don't know how much Ryan has used it, he's displayed much more than just a passing familiarity with it.
 
Reactor drone said:
From NCSTAR 1-6

Quote:
"Break elements (ANSYS user defined element) were developed for use in the full floor model that captured the temperature dependant failure modes of the truss seat connections, strap anchors and spandrel studs. these elements replaced the detailed 3-D solid finite element model used for the truss connection analysis, allowing a reduction in model size.

A break element is a unidirectional spring element that can simulate connection failure by disconnecting two "active nodes" when the relative displacement between two "control nodes" exceeds a specified threshold "
Click to expand...

So my understanding is that they make a smaller model that works the same way as the more detailed model so that when you model an entire skyscraper it doesn't use more space than you have on your computers.
Click to expand...
Exactly.
I'm not an ANSYS guy-I use IDEAS and NASTRAN/PATRAN mostly, so this sounds like a contact-type element.
Each node in a FEM has 6 degrees of freedom. A solid element has a minimum of 8 nodes, and it takes a minimum of 3 elements through the thickness of the material to get a goo representation. The aspect ratio of the elements is important, also.
So for a chunk of metal 0.6 inch thick, 4 inches wide, and 4 inches long, you need 400 elements and something like 2000 nodes time 6 dof, for an array size of 12000 or so-just for that little chunk. (Do an inversion on a 12000 x 12000 matrix for me...)
So you define an element that has the same properties as the chunk--in this case, it takes load in one direction, just like real life (NB or one of the structures guys can answer that) so that it only has 2 nodes, or 12 DOF. 12 is a lot less than 12000. If it can't take bending, you end up with only 6 dof, even better.
ETA: Uniaxial reduces it to 2 dof, and was likely a decision made based on the physical representation of the joint. Newtons bit or one of those guys can better answer that.
 
Last edited:
thanks everyone for your responses. i squashed his argument by asking him to validate his claim with some sort of source rather than his baseless personal conjecture.

to give you a heads up this guys is a diesel mechanic and thinks that being one makes him qualified to claim that the WTC buildings emitted no noticeable explosions outside the buildings because the tube in tube design acted like a muffler. essentially his entire argument was mufflers are tubes and the WTC design was tube in tube, therefore the WTC acts like a muffler.

he also argued that the NIST reports are not science since they used estimates in their models. according to him a hypothesis is an estimate and since a hypothesis is just an educated guess and an estimate is a hypothesis the NIST model is not actual science based on evidence; or some sort of nonsense like that. he then topped it off by claiming i play word games. :boggled:
 
grandmastershek said:
... claim that the WTC buildings emitted no noticeable explosions outside the buildings because the tube in tube design acted like a muffler. essentially his entire argument was mufflers are tubes and the WTC design was tube in tube, therefore the WTC acts like a muffler.
Click to expand...
doh.gif


grandmastershek said:
... he also argued that the NIST reports are not science since they used estimates in their models. according to him a hypothesis is an estimate and since a hypothesis is just an educated guess and an estimate is a hypothesis the NIST model is not actual science based on evidence; or some sort of nonsense like that. he then topped it off by claiming i play word games. :boggled:
Click to expand...
:mgduh
 
grandmastershek said:
thanks everyone for your responses. i squashed his argument by asking him to validate his claim with some sort of source rather than his baseless personal conjecture.

to give you a heads up this guys is a diesel mechanic and thinks that being one makes him qualified to claim that the WTC buildings emitted no noticeable explosions outside the buildings because the tube in tube design acted like a muffler. essentially his entire argument was mufflers are tubes and the WTC design was tube in tube, therefore the WTC acts like a muffler.

he also argued that the NIST reports are not science since they used estimates in their models. according to him a hypothesis is an estimate and since a hypothesis is just an educated guess and an estimate is a hypothesis the NIST model is not actual science based on evidence; or some sort of nonsense like that. he then topped it off by claiming i play word games. :boggled:
Click to expand...
One of the problems lay people have is , they don't understand that we have a LOT of actual test data on materials and joints in the real world. We also have a LOT of analysis correlation data for the same. most of any error is a function of the manufacturing process-joints, material variations, etc. We get around that by analysing to Minimum allowable, maximum slop, etc.
NASA, for example, requires that your structural mathematical model correlate within 5% of reality--and you do tests to prove it!
The behavior of steel, aluminum, and glass can be accurately predicted under every condition anyone can think up and run a test on, within reasonable error margins. Usually, the actual hardware is stouter than the analysis.
 
rwguinn said:
One of the problems lay people have is , they don't understand that we have a LOT of actual test data on materials and joints in the real world. We also have a LOT of analysis correlation data for the same. most of any error is a function of the manufacturing process-joints, material variations, etc. We get around that by analysing to Minimum allowable, maximum slop, etc.
NASA, for example, requires that your structural mathematical model correlate within 5% of reality--and you do tests to prove it!
The behavior of steel, aluminum, and glass can be accurately predicted under every condition anyone can think up and run a test on, within reasonable error margins. Usually, the actual hardware is stouter than the analysis.
Click to expand...

well this jackhole refuses to believe you can reasonably estimate the damage to the fireproofing w/o actually seeing the damage yourself.
 
As I understand, it's an element that breaks at a specific point defined by the user.
 
You must log in or register to reply here.

Back
Top Bottom