Need Information on "Moment Releases" for Structural Analysis 1

[ga30224]

Hi everyone, I have been in the civil engineering field for a long time but relatively new to modeling with structural analysis software such as STRAP, STAAD-Pro, etc. One of the issues that confuses me most is "Moment Releases" in modeling your structures.

For example, when I am designing a triangle truss, say, with a few web members, I need to "release the moments" so that the software knows that my web members are not designed to take any moments, i.e. they do not become "moment connections". I release the moments at both ends of the webs, but only one end for the top chord on the right side, and only one end for the bottom chord.

1. Could someone knowledgeable in this area please explain in plain English the rationale and principles behind "moment releases"? For example, we never "release moments" when manually working out the forces in a truss.

2. How do you know which member to "release moment", and whether we relase it at one end of the member, or both ends?

3. Would a catastrophe occur, if we "release the wrong member", for instance?

4. How does the structure in the field knows that the structural engineer has released moment for the web member at both ends, the right top chord on one end, but not the left top chord (of a triangular truss, say) and for it to behave accordingly in the field?

5. Could you direct me to a website, textbook, publication, or whatever that has a good and clear discussion on "Moment Release"?

I hope someone would help me as I am completely stumped by this issue. Thank you so much.

 

[JAE]

To fully understand moment releases, you should properly understand how a computer program models and calculates the relationship between points in your modeled structure. This would take a semester in Matrix Analysis of Structures to fully comprehend.

But in a few short paragraphs, I'll try to give the general concept.

First, when modeling a structure with two-node finite elements (i.e. stick elements), each end of the stick (or beam) has considered a number of degrees of freedom such as axial translation, vertical translation, horiz. trans., axial moment (i.e. torsion), and moment about the two orthoganal axes. Some two-dimensional programs will use only three DOF - axial, vert. shear, and in-plane moment.

When you put the sticks together to form a frame, say a simple bent with two columns supporting a single horizontal beam, the initial condition is that the computer will model the members as fully tied together....in other words, any bending moment at the ends of the beams will be transferred into the tops of the columns.

This also applies to the axial and shear as well, but the moment DOF is the most widely used in terms of pins.

Now if you add a pin, say at one end of the beam, then the model will behave such that the pinned end of the beam is free to rotate relative to the top of the column. No moment occurs at this end of the beam (M = 0). The other end is still fully "fixed" to the column.

Your statement above - that "the software knows that my web members are not designed to take any moments, i.e. they do not become "moment connections"" is entirely correct.

So a pinned end represents, in reality, a condition where the connection is:

1. Not designed to transfer moment from one member to the next, and,
2. Is relatively free to rotate.

Item 2 means that your connection has some freedom to allow rotation in the joint. It most likely has some resistance to rotation and this is ignored in your model and is a conservative assumption for the member designs. For the connection design, it is not necessarily conservative, but most standard double angle, or plate connections have the ability to allow this rotation without compromising their capacity to resist the shear.

In your truss, the end connections of diagonal members (web members) is usually modeled in a computer program as pinned-pinned so that the diagonals only take axial load. The chord members are, in reality, continuous, so adding pins at the joints is not necessary and would incorrectly model the real condition. In years previous, where trusses were calculated by hand, adding pins everywhere was a conservative way of allowing the truss to become a non-redundant structure so the hand calculations were made much simpler.

The choice to add a pin to your model is up to you. If you place a pin at a particular member end, you should detail the resulting connection such that it is not a specifically designed "moment" connection. The pin forces the moments to accumulate elsewhere making the member design a bit heavier. If you leave out the pin, then you WILL have moments there and you must design the connection to take those moments.

There's much more to be said -but this is getting long - hope it helps.