Truss Conveyor Design

[mrMikee]

I'm reviewing the design of several truss conveyors constructed from angles. These have been modeled and solved using Multiframe. There are no gusset plates in these trusses, the angles are welded together by overlapping the legs and welding, creating an eccentrically loaded connection.

Generally when I have a single angle strut in a structure I model it as a line element concentrically loaded and introduce bending in my code checking spreadsheet using the axial force and corresponding x & y offsets. From this geometry I calculate the resulting z & w bending and stresses. I have considered these acceptable designs in the past when framing into heavier W shaped columns and beams. With a structure comprised completely from angles I'm concerned that I could be overdesigning in some cases.

The top and bottom chords being continuous members may require a different approach. Maybe the angles should be modeled with offset connections but that would involve a considerable amount of extra work creating the models. What I'm trying to do now is come up with a rational procedure to be used for analysis of about 10 to 12 conveyors totaling about 1000 feet of trusses, so an efficient method would be nice.

Before I get too far with this project I'm curious if anyone here has some words of advice or warnings. I probably should mention that I still use the ASD spec.

 

[rb1957]

1st, continuous members ...
when you analyze a truss by hand you're assuming each member is discrete. i think this is unconservative locally, that the continuous chords allow much more moment at their ends (than a discrete analysis). but it's probably way conservative for the member (as a column, pinned vs fixed ends); and if you're mainly concerned with axial loads maybe it isn't much of a problem. i'll gladly deferr to a more experienced analyst.

2nd, offset attachments ...
yeah, you're analyzing the members as though they're loaded on the centroid, but i wouldn't think that the welds react the endload in the continuous members, but they are responsible to transfer the difference in member loads, including the effects of the offsets. clear as mud ?

 

[mrMikee]

For the diagonal bracing I solve as if they were concentrically loaded members. Since the CGs don't line up the members obviously aren't concentrically loaded so I use the technique in the AISC on page 3-55 for Single Angle Strut design and calculate the z and w axis bending stresses. As I mentioned I think this gives a reliable design.

In the case of the top chords however I think I may be overdesigning possibly by a lot. The axial force in each segment of the top chord does not necessarily come from loading at the segment end points, but also from loads farther down the truss. Do these loads at points away from the segment under question cause x & y bending (and therefore z & w bending) in a way consistent with the assumptions of the ASD method? I'm not sure.

Perhaps I'm just not seeing something here clearly. I have a sense that what I normally do would be an overdesign, but on the other hand I can't ignore the eccentric axial loads and resulting moments either.

I agree, clear as mud.

 

[cessna98j]

The approach I took when modeling a conveyor truss with eccentric vertical webs, was to model the section with the highest brace forces as eccentric while leaving the rest of the model concentrically braced. I rationalized that the eccentric bracing would have the most impact (highest induced moment) at the highest brace forces. By modeling the eccentricity of the braces at the highest loaded brace points I was able to see the induced bending stresses locally in the chords.

 

[mrMikee]

Thanks for the comments. I'm going to modify my design calcs to use the eccentricity between the top chord angle cg and the surface where the two angles are welded together. Since in my opinion this will overestimate the bending moments in the top and bottom chords I'll have to come up with a reduction factor to better predict end moments.

At this point I'm not sure what that will be however. Still working on that.

Regards,
-Mike