Ram Elements 2D Truss Analysis

What version are you using? There was a change at one point that was made to report moments based in the stronger principle axis rather than the numbered axis and it causes some confusion. The way to confirm is if you change the size of the angle so that the legs that are together are a bit longer does the reported moment switch to M33?

Seth's response is correct.

The member forces are referenced with respect to the principal axes and not the local member axes. For double angles with the short leg back-to-back, the moment of inertia about the local 2-2 member axis is larger than the moment of inertia about the local 3-3 member axis. As a result, the principal axis is rotated degrees from the local member axes. The local 3-3 axis coincides with the principal 2'-2' axis. The local 2-2 axis coincides with the principal 3'-3' axis.

The following web page may also have helpful information:

http://communities.bentley.com/products/structural/structural_analysis___design/w/structural_analysis_and_design__wiki/23793.ram-elements-local-versus-principal-axis-in-unsymmetrical-shapes

Seth and Karl, I am using 13.4.0.177 (latest). It is confusing to know when Ram is using Local Axis (2-2, 3-3) or major/minor principal axes. Is the major principal axis always 3-3 and the minor principal always 2-2? The local axis 3-3 could be the strong or the weak axis for double angles. Using the local axis for reporting analysis results makes more sense to me because it is independent of the type of section chosen. During the design process, different sections may be substituted freely until a final one is chosen. This all pertains to analysis and design reporting.

To add more confusion to the mix - when modeling the structure and defining the bracing points for flexure or unbraced length segments for compression, should I be using local or principal axes? I assume local since that would be independent of the section used.

For flexure, sections are always designed with respect to the principal axes unless the option for "Laterally Restrained for Torsion" is used (see discussion on the web page in the previous reply for more on this option). The bending moments associated with these members are reported with respect to the principal axes and not the local axes. If "Laterally Restrained for Torsion" is checked, the section is designed with respect to the local member axes. The bending moments associated with these members are reported with respect to the local axes and not the principal axes.

For compression, sections are also designed with respect to the principal axes. In the Section titled "Compression in the Major Axis 33," the KL/r value uses the radius of gyration associated with the principal axis and not the local axis and the parameter L33 is used for the unbraced length. L22 and L33, then, correspond to the principal axes and not the local axes. If "Laterally Restrained for Torsion" is checked, the section is designed with respect to the local member axes. In this case, the KL/r value used in the 'Compression in Major Axis 33" check uses the radius of gyration associated with the local axes and L33 corresponds to the local axis and not the principal axis.

I see a potential issue with flexural-torsional buckling of members with a principal axis rotated 90 degrees relative to the local axis that may be causing problems with the compression check. I will need more time to review and test this.

In summary, forces and unbraced lengths are associated with the principal axis and not the local axis unless the "Laterally Retrained for Torsion" option is used. Forces and unbraced lengths for sections that use the "Laterally Restrained for Torsion" are associated with the local axis and not the principal axis.

Thanks for the clarification. I like to model the "physical" members instead of splitting the members at each node. So, defining the lateral brace points and unbraced lengths is crucial to the steel design. Working with double angles for trusses or bracing can cause pain because depending on which section you may choose, the principal axes may or may not match the local axes. This is true for Equal Leg, SLBB and LLBB angles.

Since design is often an iterative process, you may be switching dbl-angle sections a couple of times and must check which way the principal axes are turned so that the lateral brace points and unbraced lengths get changed with each iteration. I presume that this would be true also if you use the Ram Elements "Optimize" design feature. My current truss model is certainly a victim of this confusion. I can send it to you, if you would like an example.

To keep things simple for the engineer instead of the computer programmer, I would like to suggest defining the bracing in terms of the "physical" model (i.e. local axes) similar to defining the "physical" members to streamline the design process. And let Ram Elements figure out which way the principal axes are turned and which principal axis gets braced at what points.