Search found 586 matches

by mhscott
Mon Jun 08, 2009 6:41 am
Forum: OpenSees.exe Users
Topic: Convergence problem with nonlinearBeamColumn element
Replies: 22
Views: 7741

I understand you need to use a trilinear model, but can you try a simple bilinear model *just to see if it works* and if you get the same convergence problems? Most convergence problems arise from the constitutive models, not the elements. You may be running in to objectivity problems with the combi...
by mhscott
Fri Jun 05, 2009 3:44 pm
Forum: OpenSees.exe Users
Topic: Convergence problem with nonlinearBeamColumn element
Replies: 22
Views: 7741

Have you tried using a simpler moment-curvature model, e.g., bilinear? I doubt it's an element problem, but rather a problem with the input parameters for the hysteretic material. Also, you may want to try using a beamWithHinges element, especially if you use a negative slope on your moment-curvatur...
by mhscott
Thu May 21, 2009 11:34 am
Forum: OpenSees.exe Users
Topic: nonlinearBeamcolumn Converging
Replies: 2
Views: 1023

Assuming your model is defined correctly, you may be using too small of a tolerance, or you may need to use load incrementation to "ramp up" the gravity loads. These, rather than the element or whatever I have to fix, are likely the reasons your analysis is not converging.
by mhscott
Thu Oct 30, 2008 10:30 am
Forum: OpenSees.exe Users
Topic: SectionAggregator
Replies: 1
Views: 933

It means one of your aggregated materials has reached a point of zero stiffness in the analysis. The stiffness matrix of the section aggregator is block-diagonal, so when one of the aggregated materials has zero stiffness, the section stiffness matrix is non-invertible (singular).
by mhscott
Wed Oct 29, 2008 6:44 pm
Forum: Feature Requests/Future Directions
Topic: Material and Element Wrappers
Replies: 1
Views: 3528

There are no such element wrappers; however, there is a plane stress material wrapper for 3D models. No plane strain wrapper, but it would be very easy to implement if needed.
by mhscott
Tue Oct 28, 2008 8:37 am
Forum: OpenSees.exe Users
Topic: ON THE EQUILIBRIUM FOR DB ELEMENTS - CONTINUING....
Replies: 3
Views: 1809

By "it is what it is" I meant that as long as you are seeing linear curvature distribution in the DB elements, the bending moment is computed accordingly. The bending moment may satisfy strong equilibrium, but it will not in general, as you see when you change the number of integration points.
by mhscott
Mon Oct 27, 2008 8:32 am
Forum: OpenSees.exe Users
Topic: Fragility Curves and OpenSees Reliability Module
Replies: 1
Views: 1919

You will have to modify the source code for those material models and provide implementations for the setParameter and updateParameter methods.
by mhscott
Mon Oct 27, 2008 8:08 am
Forum: OpenSees.exe Users
Topic: ON THE EQUILIBRIUM FOR DB ELEMENTS - CONTINUING....
Replies: 3
Views: 1809

Nothing special is done in the OpenSees implementation to scale any kind of section response. The elements simply use Gauss-based numerical integration (Gauss-Legendre for the DB elements and Gauss-Lobatto for the FB elements). You can look up numerical integration in any numerical analysis textbook...
by mhscott
Fri Oct 24, 2008 1:37 pm
Forum: OpenSees.exe Users
Topic: IS THERE A BUG IN SECTION RECORDER WHEN USING A DB ELEMENT?!
Replies: 9
Views: 4041

Riccardo, Force-based elements satisfy internal equilibrium in strong form over the element domain, i.e., for all values of x. You see this result, where the section axial forces are equal to the applied axial load. However, equilibrium is only satisfied in weak form in displacement-based elements, ...
by mhscott
Tue Oct 21, 2008 1:37 pm
Forum: OpenSees.exe Users
Topic: Reliability Examples may not work
Replies: 1
Views: 2149

In the latest source code of CVS, any finite element reliability analysis now requires parameter objects that point to the finite element domain. RandomVariablePositioners are still required to associate parameters with random variables. In addition, the tags assigned to random variables can now be ...
by mhscott
Tue Oct 21, 2008 10:36 am
Forum: OpenSees.exe Users
Topic: Reliability Examples may not work
Replies: 1
Views: 2149

Reliability Examples may not work

Some reliability examples, e.g., those with filenames formatted as reliability_x_y_z.tcl, may not work with the current version of OpenSees. In addition, just about every reliability example involving a discretized random process will not work with the current OpenSees. While there is no specific re...
by mhscott
Tue Oct 21, 2008 9:04 am
Forum: OpenSees.exe Users
Topic: Dynamic reliability
Replies: 3
Views: 1435

Can you e-mail me your input file so I can reproduce the error and track down the problem?

michael.scott at oregonstate dot edu
by mhscott
Wed Oct 01, 2008 7:50 am
Forum: OpenSees.exe Users
Topic: eleload as a time dependent uniform load
Replies: 8
Views: 2933

The nonlinearBeamColumn is assuming a 1.0 load factor from all element loads. Frank is working to fix it so that variable load factors are used.
by mhscott
Thu Sep 18, 2008 2:21 pm
Forum: OpenSees.exe Users
Topic: BeamWithHinhes
Replies: 8
Views: 3813

I can't think of any off the top of my head, but you may want to search for articles from the following authors:

C. W. Roeder
D. W. White
W. F. Chen
by mhscott
Wed Sep 17, 2008 6:06 pm
Forum: OpenSees.exe Users
Topic: I don't think rotational spring works properly
Replies: 2
Views: 1097

You may have the wrong dir in the zeroLength command. Although you have a 2D problem, according to the documentation, the dir values are 3D. So, for rotation about the z-axis, you should use 6.