RAM Steel Beam vs Frame

Question

Can someone please answer this question for me because I am currently at a loss. I am looking at a situation where two columns support a cantilevered floor. I have two columns with "cantilevers" coming off of them at the floor level. At the end of the cantilevers I have a diagonal brace which attaches back to the top of these columns respectively. And at the end there are two posts up and simple "lateral" roof beams. So I just have a simple cantilevered cube with two diagonal braces to take the end of the cantilever in tension. 
 In Steel beam these two floor beams are failing miserably. I have seen that Steel beam considers everything pinned but this is clearly not true. It is calculating the moments as though one end is fixed and the other free. Based on the loads in the "View Loads" button I can duplicate these moments. When I go into Frame, the moments are reduced by about 40% and there is an axial load in the floor beams (as would be expected). But the brace has 0 load in it so I know that the brace is being calculated incorrectly. So I cannot understand why there is a discrepancy between Steel Beam and Frame. I cannot duplicate the loading on these beams from the Post Processing Frame Loads Report, however I can duplicate the loading on one of the infill beams that frames between the two. ANY help would be greatly appreciated.

Jeffjr02 · Accepted Answer

Thanks for your help. Just to sum up one final thought. 
 So any beams and columns that are lateral may not have correct designs when just ran in the Steel Beam/ Steel Column modules correct? If a lateral beam/column fails in the respective modules, but works in Frame that Frame is the one to be trusted. The idea being that while beam and column are a quick easy way to design most columns, the Frame module in fact takes into consideration ALL effects.

Seth Guthrie · Answer

The braces are not considered in the Ram Steel analysis. The comment from the Analysis Notes wiki that everything is pinned is an oversimplification. Cantilever beams and Stub cantilevers are still treated as fixed. I'll edit that wiki in a moment to make it more clear. 
 So, in any case where a brace is necessary to resist gravity loads, Ram Steel cannot give the accurate answer. That's why we hide the frame beam results normally. Designing the beam in Ram Frame - Steel mode is the way to go. 
 This wiki on columns has some nice pictures about something similar.

Seth Guthrie · Answer

Very strange indeed. Clearly the braces in your file are elongating under the gravity loads, yet no force is reported. I tried several things and eventually deleted and remodeled the braces using the add-special method and zero offset from the nodes , now they are behaving properly. I've just sent that version of the file to you in a private message, but my copy is in v14.06.01.00 format. 
 Did you import the braces from Revit or maybe model then using the standard knee-brace option?

Seth Guthrie · Answer

Ah-ha, tension-only braces were used initially. In Ram Frame analysis, tension-only braces are always ignored in the gravity load case analysis. We do this to prevent the pair of braces in an X braced frame from both going into compression and therefore getting thrown out of the stiffness matrix. I think it's also common practice to not consider such braces for gravity analysis. In cases where the braces are necessary to resist the gravity loads, they must be tension and compression members . 
 More on our tension only member wiki .