-
Filter by Topic
Most Popular
All Topics
- List all discussions
Facebook
Twitter
Delicious
Digg
Stumbleupon
More services...point displacement in z direction only
|
Thread index | Previous thread | Next thread | Start a new discussion |
RSS feed |
Turn on email notifications |
3 Replies
Last post: September 4, 2010 7:29am UTC
September 3, 2010 4:54pm UTC
point displacement in z direction only
This should be an easy answer but for some reason I cannot find it in Comsol 4.0a. I have a fiber that i am heating up and I want to pick points on the fiber so that I can measure the displacement but only in the fiber direction (the z direction in this case). I know how to do a total displacement after the solution is solved, but not a z direction dependent displacement for a point. My problem is that I have this fiber sitting on a large metal tube and when it heats up, it expands in all directions. Thus the total displacement of the fiber is not realistic. If I can get the z direction displacement I can find out how much the fiber is stretching as a result of the heating up process. My overall goal is to calculate the strain (delta L) /(initial L) in units of microstrains ( 1 per million) by using these displacement numbers.
Thanks in advance,
-Neal
Reply | Reply with Quote | Send private message | Report Abuse
September 3, 2010 5:16pm UTC in response to Neal Pfeiffenberger
Re: point displacement in z direction only
Hi
I'm not sure I understand you, acause for me your question seems too trivial, I may have missed something. But the variables u,v,w are the three components of the displacement, in COMSOL and the total displacement isonly the norm sqrt(u^2+v^2+w^2). Now youcan get u,v,w fore a point, but often its more precise to use the average over an area. In V4 now you have the min,max and average operators you can define over given areas.
Is this what you want ?
When you talk about the strain, you also have this variable defined, check your doc, or use the CNTRL space in a probbe or postprocessing data field
--
Good luck
Ivar
Reply | Reply with Quote | Send private message | Report Abuse
September 3, 2010 9:49pm UTC in response to Ivar Kjelberg
Re: point displacement in z direction only
Hi Ivar, Thanks for your response. Yes the u, v, and w sounds like what I want. Where can you find these and how could I apply them for the 4 blue points in the picture I have attached? Also you mentioned the min, max, and average operators. Where can I find these to apply them in my model? Also would the first principle strain be the same as (delta L) /(initial L)? If not what should I be looking for in the postprocessing to get this type of output. I am unfamiliar with working with the heat transfer side of things since I mainly do research in the RF package so please forgive me.
Thanks again,
-Neal
Reply | Reply with Quote | Send private message | Report Abuse
September 4, 2010 7:29am UTC in response to Neal Pfeiffenberger
Re: point displacement in z direction only
Hi
really you should read through the structural (ST) and HT doc to get used to the variables, otherwise you will get probably easily wrong conclusions ;), know your tool, but for that start with very simple cases that you might analyse by simple analytical means to compare.
So u,v,w are the variables (default for displacements in COMSOL) and you need to define yourt points (set them in a selection Definition) (right click the "Definition" node and look carefully through all items, you will find the min, max, average, integration etc oprerator there too (there are mosty several layers of subnodes to pass, the "gems" are often hidden rather low in the three structure ;) then you apply the operator on these points. Once solved in the postprocessing you might call up the operator, or just look indicidually (unfortunately postprocessing does not accept an array :( would be nice ;) and you get the values in a table.
One thing, ormally point values are averges from the surrounding lines/surfaces and can be strongly influenced by their neighbours. Therefore I always check that I have a sifficient mesh density, and I average over an area covering enough elements (approaching St Venants principle) rather than picking points (or I define a metrology probe point in the middle of a surface, so it gets an extrapolated value but not directly the one of the adjacent physics border
--
Good luck
Ivar
Reply | Reply with Quote | Send private message | Report Abuse
Rules and guidelines
