	Product(s):	RAM Structural System; RAM Frame
	Version(s):	Any
	Environment:	N/A
	Area:	Analysis

How can I assign a member to be tension only?

Any vertical or horizontal brace member may be assigned to be tension-only. The assignment is made in RAM Frame using the Assign menu - Braces - Tension-Only command. Assigning a brace to be tension only does not automatically release the brace for moments or torsion, however. That is controlled through the Assign menu - Braces - Fixity command.

If no assignment is made, then the braces will resist tension and compression elastically. Beams, columns and walls are always capable of resisting compression.

How are reverse direction lateral loads handled when using tension-only members?

When defining program generated lateral loads, like wind or seismic loads, there is an option to "Generate Additional Load Cases for Analysis with Tension-Only Members" that should be checked when the model includes any tension-only braces. Using this option leads to a load case in the positive direction, say west to east, and another load case in the negative direction, east to west. For each of these load cases, different members may be active in resisting the load so it's important to perform a separate finite element analysis of each direction.

If the model includes tension-only braces, but this option is not checked then a warning is generated.

When all the braces are tension and compression braces, then the whole model in linear-elastic and the analysis can be simplified by only analyzing the positive direction loads and then using both positive and negative terms on lateral loads in load combinations to account for reversibility.

Note, for models with user defined loads or nodal loads the same rules apply; if there are any tension-only braces in the model, then user loads should be created for both directions.

How can I resolve an error, "tension-only member in compression"?

In the steel design mode, you might see the warning "T-O in comp" on the View/Update dialog box or "Member with Tension Only Shape is in Compression" on the design report. There are some cross sections that the RAM Frame Steel postprocessor can only check when they are in tension. If a single angle, rod or flat bar goes into compression, then the design will always give these warnings. You can manually check the design of such using the member force envelope results.

The same error message, "T-O in comp", can also occur for a member that is defined as a tension-only brace if there are custom load combinations utilizing negative terms on lateral load cases. It's important for models with any tension-only braces that the lateral loads be applied in all directions so that the load combos can use positive terms exclusively.

Why is my tension-only brace failing in compression?

If a brace that was specifically assigned to be tension only is going into compression, that usually indicates a problem with the load combinations. For models with tension only members, the load combinations should never include negative terms in any load combinations (generated or custom). If there is a combo with a negative term, the resulting brace force could be compression and the warnings above will occur.

Consequently, any model with tension only members requires one lateral load case in the positive direction and another load case in the opposite direction if all the braces are to be correctly evaluated. For program generated loads, there is a check box for "Generate Additional Load Cases for Analysis with Tension-Only Members" that can be selected to do this automatically, but for user defined story forces or nodal loads, an additional case for the reverse direction must be manually input.

How is a tension only brace considered during dynamic analysis?

For any dynamic, response spectrum load case, all braces are assumed to be 50% effective. This generally accounts for an appropriate amount of stiffness for a symmetrically braced frame, but does not prevent the braces from showing compression results. This approach is also used for the eigensolution.

Why am I getting large deflections under gravity loads?

In the RAM Frame analysis, braces that are designated to be tension-only are completely ignored during the analysis of the gravity load cases (dead, live, roof and snow). For the evaluation of lateral loads the program uses an iterative approach to determine which braces are in tension under each lateral load case. Then the program superimposes the results based on the defined combinations in order to code check the braces.

In structures where all the braces are tension-only, a global instability can often result for gravity loads. If you get an instability warning or if the deflections are significantly large under gravity load cases there are several things you can do.

Simply ignore the gravity load deflections. For the purpose of evaluating deflections, use the lateral load cases alone (uncombined). So long as there are no significant shears or moments developing in the other lateral members due to the drift, there is really no harm in this.
Provide a secondary system to give the structure some lateral stiffness even when the tension-only braces are ignored. Typically, this is a matter of fixing the columns all the way to the base. If a fixed base condition is too stiff, try adding a rotational foundation spring at the base of the columns to reduce the lateral stiffness.
Remove the tension-only assignment from the braces altogether. In this case, the braces can typically be hand-designed for tension equal to two times the maximum (or minimum) axial force from any combination (obtained from a member force envelope report). This approach can alter the load path affecting column and foundation design however.

An enhancement was added in RAM Structural System v14.06, which should resolve the problem with large displacements and instability errors. Beginning with this version, P-Delta effects will not be included when analyzing gravity load cases with the special tension-only analysis. This is a reasonable implementation since in virtually all cases the P-delta moments due to lateral sway caused by gravity loads is extremely small. P-Delta effects will still be considered with the lateral load cases. More on this can be found in the v14.06 release notes, which are located here: V14.6 Release Notes.

Starting in version 14.07 another option was added to the RAM Frame General criteria for Buckling Restrained Braced Frames (BRB). If the option to "Exclude Buckling Restrained Braces from the Gravity Load Case Analysis" is checked, the same problem noted above for tension-only braces may occur.