plasticDeformation

Forum for OpenSees users to post questions, comments, etc. on the use of the OpenSees interpreter, OpenSees.exe

Moderators: silvia, selimgunay, Moderators

Post Reply
arsepas
Posts: 12
Joined: Wed Aug 31, 2016 8:56 am
Location: urmia university

plasticDeformation

Post by arsepas » Sun Dec 31, 2017 10:49 pm

hello
I've modeled a rc frame using forcebeamcolumn element and fiber section and load it under a push over horizontal load, now i need to read plasticDeformation of plastic hinges. i wrote recorder command as below:

recorder Element -file beams.out -time -ele 1section 1 plasticDeformation

but there was nothing in out put file.
i cant find the problem.
i think maybe plasticDeformation doesn't work for forcebeamcolumn element. is that right? or what is the problem?

selimgunay
Posts: 913
Joined: Mon Sep 09, 2013 8:50 pm
Location: University of California, Berkeley

Re: plasticDeformation

Post by selimgunay » Mon Jan 01, 2018 3:37 pm

Are you interested in the curvatures? If so, you can directly record the curvatures and subtract the yield curvature to get the plastic curvatures. You can determine the yield curvature from the moment-curvature relationship.

arsepas
Posts: 12
Joined: Wed Aug 31, 2016 8:56 am
Location: urmia university

Re: plasticDeformation

Post by arsepas » Mon Jan 01, 2018 8:57 pm

thanks for your kind attention.
i just want to have plastic rotation of a cross section along a reinforced concrete beam or column.

Jhno
Posts: 214
Joined: Sat May 05, 2012 2:55 pm
Location: Laval

Re: plasticDeformation

Post by Jhno » Tue Jan 02, 2018 11:35 am

You should put a space between ele # and section :
Also, in an element recorder you can ask a FiberSection for its 'deformations', 'forces', 'forceAndDeformation', 'fiber $fiberNum $matArg1 ..', 'fiber $yLoc $zLoc $matTag $matArg1 '
Page 10 : http://opensees.berkeley.edu/wiki/image ... Output.pdf

selimgunay
Posts: 913
Joined: Mon Sep 09, 2013 8:50 pm
Location: University of California, Berkeley

Re: plasticDeformation

Post by selimgunay » Wed Jan 03, 2018 5:23 pm

You cannot get the plastic rotations directly for the forcebeamcolumn element with fiber sections. You should multiply the plastic curvatures with an assumed plastic hinge length to get the plastic rotations

EricsonEncinaZ
Posts: 108
Joined: Mon Sep 16, 2013 1:14 pm
Location: University of Auckland

Re: plasticDeformation

Post by EricsonEncinaZ » Wed Jan 03, 2018 8:35 pm

From the fibre section you can determine the strains (eleResponse $Ele section $sec fibre $y $z strain) and curvature (eleResponse $Ele section $sec deformations, or from the strains) of each integration point(IP). You can also get the IP locations (eleResponse $elemento integrationPoints) and weights (eleResponse $Ele integrationWeights). Note that the commands used will return values during runtime directly into OpenSees so you can evaluate everything in runtime, and so they will not be saved in output files. If you want to post-process you can transform the commands here into recorders.

The rotation of an IP will be its curvature times its weight (not mass·g); and if you need the rotation of an element, then it will be the sum of all the IPs' rotations of that element.

Now answering your question, by definition the yield rotation is the rotation achieved when the first bar yielded, and you can determine it based on what is written above, thus the plastic rotation will be the total rotation detemined as above minus the yield rotation you already got.

EDIT 29 Aug 2018:
This portion of the answer "...if you need the rotation of an element, then it will be the sum of all the IPs' rotations of that element." is wrong. Lets think in an element with 3 IPs, in which each IP sustains a curvature phi_j and has a length L_j, with j = 1..3. Rotations are determined based on an integration scheme (think in a gaussian quadrature), therefore rot_j = phi_j · L_j. Then the total displacement at the end of the element produced by bending is the contribution of each IP rotation times the distance to the end of the element, disp_flex = rot_1·(L_1+L_2+L_3) + rot_2·(L_2+L_3) + rot_3*L_3. So then if one would like to determine an "overall" (or may be better called a weighted average) rotation, that flexural rotation may be determined by rot_ele = disp_flex/(L_1 + L_2 + L_3), this is assuming that all the rotation occur at the initial node.

EDIT 14 Mar 2019:
I recently tried to calculate the flexural deflection in a cantilever column using what I mentioned in the "Edit 19 Aug 2018". The model was a fibre-based FBE with tip load and elastic section, so I could compare it with hand calcs. I tried from 2 to 10 IP Lobatto and the procedure I described above did not work ... well ... just worked when I had 2 IPs. So I guess, it is better to obtain the total rotation of the element using ¿ node_displacement / length_of_the_element ?
Last edited by EricsonEncinaZ on Thu Mar 14, 2019 1:23 am, edited 3 times in total.

selimgunay
Posts: 913
Joined: Mon Sep 09, 2013 8:50 pm
Location: University of California, Berkeley

Re: plasticDeformation

Post by selimgunay » Sun Jan 07, 2018 7:06 pm

This is not correct. Rotation difference between two points along a beam or column is equal to the area under the curvature diagram. So, with a rough assumption of constant curvature along the plastic hinge region, the plastic rotation can be computed as the plastic curvature multiplied by the plastic hinge length. To make a more accurate computation, curvatures at different integration points can be considered and the curvature distribution can be assumed to be constant between the integration points. Then the plastic rotation is calculated as the area of the plastic curvature distribution.

EricsonEncinaZ
Posts: 108
Joined: Mon Sep 16, 2013 1:14 pm
Location: University of Auckland

Re: plasticDeformation

Post by EricsonEncinaZ » Mon Jan 08, 2018 4:43 am

Hello Selim,

Isn't what you wrote the same as I wrote?

The IP's weight is the length associated to the IP, and for a beam-column (BC) element the curvature is constant WITHIN each IP length, therefore the area under the curvature diagram - as you said - is the length of the IP (i.e. its integration weight, not mass time g) times the curvature, so we are saying the same :-D. Then if the element has n IPs, the total rotation of the element is the sum of the rotations determined at each IP that belong to that element. The number of IPs in a BC element reflects the samplig that the user wants along that element to calculate its response.
And we know that the total rotation of an IP = yield rotation of the IP + plastic rotation of the IP. From the analysis we can get directly the total rotation or curvature by recording the section deformations and we can save the yield curvature when this occurs, so we have everything to determine the plastic rotation.

EDIT. see notes addeded on my previous post.

Post Reply