Start with the Correct Static Resisting Force

Nonlinear Analysis 2 Minutes

Thanks to my involvement with OpenSees, I have been asked to review several manuscripts on nonlinear dynamic structural analysis. It’s always wonderful to see what other researchers are up to and how they use OpenSees or other software to achieve their research objectives.

A brief overview of the governing dynamic equilibrium equation, i.e., the equation of motion, is standard in the manuscripts I’ve reviewed–and in this post. For the common case of linear damping and inertial forces, nonlinearity comes from only the static resisting force vector, {\bf P}_r, which is expressed as a function of the displacement vector, {\bf U}:

{\bf M}\ddot{\bf U} + {\bf C}\dot{\bf U} + {\bf P}_r({\bf U}) = {\bf P}(t) (CORRECT)

Unfortunately, a few manuscripts I’ve reviewed get the equation of motion wrong by expressing the static resisting force vector as {\bf K}_T{\bf U}, the product of the tangent stiffness matrix, {\bf K}_T=\partial{\bf P}_r/\partial{\bf U}, with the displacement vector:

{\bf M}\ddot{\bf U} + {\bf C}\dot{\bf U} + {\bf K}_T{\bf U} = {\bf P}(t) (INCORRECT)

This equation tries to fit a changing stiffness matrix into the framework of linear dynamic structural analysis. Yes, nonlinear analysis can be viewed as a sequence of linear analyses, but not in that way. The difference between these two expressions for the static resisting force vector can be significant.

Consider case (a) in the figure below, where the static resisting force is in a state of loading. The solid circle is the correct static resisting force, {\bf P}_r({\bf U}), while the incorrect {\bf K}_T{\bf U} is the height of a triangle made with displacement as its base and the tangent stiffness (moved to the origin) as its slope. The incorrect static resisting force is much smaller than the correct value.

Now consider case (b), where the static resisting force is unloading with its initial stiffness. After forming a triangle from the displacement and the tangent stiffness at the origin, the incorrect {\bf K}_T{\bf U} is much larger than the correct value. Furthermore, as unloading continues, {\bf P}_r({\bf U}) and {\bf K}_T{\bf U} will have different signs.

The figure is a single degree of freedom representation of a multi-dimensional problem. So, instead of magnitudes, signs, and triangles, let visions of norms, vectors, and hyperplanes dance in your head.

Although they use software that computes the correct static resisting force, it’s possible the authors don’t realize their error. If I see the incorrect equation of motion when reviewing a manuscript, my review decision is easy–I try to be constructive and not too snarky in my review comments. But would you trust the interpretation of results made by authors who do not express the equation of motion correctly?

Unknown's avatar

Published by Michael H. Scott

I have been involved in the development, maintenance, and growth of OpenSees since its early days. I am interested in learning Python and improving my academic writing.

Published

9 thoughts on “Start with the Correct Static Resisting Force”

  1. Pingback: Norms and Tolerance – Portwood Digital
  2. Pingback: Off the Hooke – Portwood Digital
  3. Pingback: Building Blocks – Portwood Digital

  4. If what you tell here is not correct, you may be misleading people. In example, did you try dividing the external load (incrementation) in the loading example?

    Like

    Reply

      1. If you do not use stiffness times displacement and use a resisting force vector, then you do not calculate a stiffness matrix for elements, right? You do not calculate the stiffnesses of beams, columns, plates, shells, you calculate force vectors. Then, in example you do not invert a stiffness matrix in implicit time integration, since you do not have a stiffness. Also, this page should be corrected, it states “eqilibrium equation for the finite element system in motion at time t:

        <math>M \ddot U_t + C \dot U_t + K U_t = R_t</math>".  https://opensees.berkeley.edu/wiki/index.php/Central_Difference       Do you use Newton Raphson, or something different? So you mean we do not use stiffness matrix in dynamic analysis, they are for static analysis... Do you mean something different? I guess you mean the resisting force at Newton Raphson algorithm not the stiffness matrix times displacement by writing resisting force vector...

        Like

  5. Pingback: OpenSees Mandela Effect – Portwood Digital
  7. Pingback: The misleadingly few variables in the nonlinear equation of motion – Vibrations and stuff

Leave a comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.