This page contains frequently asked questions related to modal analysis.

General FAQ

How should I setup a modal analysis for a structure with cables?

Extended Question: I am doing a modal analysis of a model that has cable elements in it. The model is intrinsically non-linear since it has tension-only cable elements. But the modal analysis is theoretically an elastic analysis. If you force SAP to do such an analysis, what will happen? In another words, what is the algorithm involved?

Answer: To obtain meaningful results for modal analysis of a structure with cables, the modal analysis should be using stiffness at the end of nonlinear load case in which the dead loads (and possibly other types of loads of interest) are applied. Then the modes will be based on the stiffness of the cables under the applied loads. If you need to consider complete dynamic nonlinear behavior, you should consider running nonlinear direct-integration time-history analysis.

What is the meaning of the "Eigenvalue was found out of sequence" warning message?

The "Eigenvalue was found out of sequence" warning message indicates that the program encounters some numerical problems due to the structure being unstable when solving for eigenvalues.

What are reasons for and consequences of a warning message that loads are applied to massless degrees of freedom in Ritz-vector modal analysis?

Extended Question: When I run analysis, I get a warning message saying: Loads are applied to one or more massless degrees of freedom, corresponding Ritz modes may be unavailable or inaccurate for dynamics. When does this type of warning indicate? How can I modify my model to get rid of this message? Does that have to do with boundary conditions? If nothing is done to avoid this warning, what are the possible consequences?

Answer: This message means that you are applying load in that load case to a joint that has no mass , possibly at a joint that is restrained.

In a structure with frame or shell elements, each joint has up to 6 degrees of freedom (DOF): translations UX, UY, and UZ; and rotations RX, RY, and RZ.

Masses which come from the elements (which, in turn, come from the material mass density) are always assigned only to the translational DOF. You can additionally assign joint masses at any of the 6 DOF of a joint.

Gravity load on a frame element creates a set of fixed-end forces at the joints which include moments as well as shears. Hence gravity (self-weight) load in most structures will impose load (moments) on massless DOF.

Ritz vectors used for dynamic loading work best if all applied loads act on degrees of freedom with mass. Earthquake loads automatically act on mass DOF. If you are applying other types of loads or are using nonlinear NLLink elements, which SAP2000 treats as external force generators, mass should be assigned to the loaded DOF.

If you are generating Ritz vectors for the purpose of applying gravity and other static loads in a time-history analysis, then it is not necessary to have mass on the loaded DOF, since the load is applied so slowly that dynamical behavior is not important. So for this type of problems, this warning can be ignored. Alternatively, you could specify a very small nonzero mass for the DOFs in question.

Also, please see the CSI Analysis Reference Manual, chapter "Modal Analysis", section "Starting Load Vectors".

After assigning shell joint offsets, why I am getting warning messages that loads are applied to massless degrees of freedom in Ritz-vector modal analysis?

When joint offsets are used for the shell element, translational forces may cause moments at the joints. However, the masses are still lumped at the joints, and do not include rotational inertia.

These messages are to be expected. Unless the offsets are very large, they are not very important. To eliminate them, you can assign small rotational masses at the joints. A reasonable value would be equal to the translational mass as the joint, multiplied by the square of the offset distance, divided by 100 or 1000.

Why am I getting warning message about excessively large effective-damping ratios for modal time history analysis based on Ritz vectors?

Extended Question: My model gives "Excessively Large Effective-Damping Ratios Have Been Set To 0.999950 ..." warning when I run time history using modes from Ritz Vector modal load case. When it uses Eigenvectors with all other parameters being the same, there is no such a warning. Is there any explanation for this warning message?

Answer: This could be caused by using proportional damping, for which the damping increases with large frequency, and Ritz vectors generate high-frequency modes that, hence, have large damping.