Steel truss - joint eccentricities

[drbrainsol]

Hi,

Newbie here so please don't be too harsh!

I am designing some modification to an existing steel truss. Due to the sizes of the structural elements used (welded CHSs), the joints between the braces have have what I believe to be significant eccentricities at the joints with the top / bottom chords (see sketch attached).

BS EN 1993-1-8 states that bending moments arising from join eccentricity should be considered in the analysis of the chords (distributed equally on either side of the joint). My question is...could these eccentricities be taken into account in the global analysis model and then "ignored" in the design of the joint? Would this be good enough to satisfy the BS EN requirement to consider the joint eccentricities?

The main reason for doing the above is that I am using MIDAS Gen to model and design the steel structure to BS EN 1993 and hand calcs to design the joints. If modelling the eccentricities spares me from having to do an additional hand calc to prove that the chords can take the moment from the joint I would prefer to do the analysis this way.

Hope this makes sense?

 

[CANPRO]

Not familiar with the your governing code, but it I think your approach sounds reasonable. Your model should include rigid offsets as shown below. I believe from that model, you could pull your web forces (which would reflect your eccentricities) and then design the connection by hand.

 

[drbrainsol]

Thanks CANPRO.

The relevant general clause from BS EN 1993-1-8 is Cl. 2.7(1):
Where there is eccentricity at intersections, the joints and members should be designed for the
resulting moments and forces, except in the case of particular types of structures where it has been
demonstrated that it is not necessary

Are the vertical rigid offsets required though? Would it not be sufficient to model the gaps between the braces accurately relative to the centerline of the chord? i.e. 54.5mm in the sketch above.

The way I am modelling this truss is by using beam elements for the chords and truss elements for the braces.

 

[CANPRO]

I've thought about that before - rigid offset vs offset nodes. Might end up being the same results...you could always model a small simple truss and check the difference (if you do this, please report back, I'd love to know if they're the same!).

For ease of modelling, I think the rigid offset would be the way to go. Especially if you anticipate having to make changes to the geometry - to change a slight slope of your webs/eccentricity you just need to adjust one node/member length per joint instead of figuring your new joint locations for two different nodes.

 

[Settingsun]

I'd model it the same way as Canpro. Your other method would be fine too as you just need that small additional stress applied to the chords. Given the offset is smaller than the section dimensions, neither method using line elements is totally accurate in this local connection zone.

Why I'd go Canpro's way is because most (all?) connection design guides assume that's how you've analysed the truss. Maybe no difference but maybe there is in some cases, so why not follow their method exactly when it's equally easy?

 

[Istructeuk]

"BS EN 1993-1-8 states that bending moments arising from join eccentricity should be considered in the analysis of the chords (distributed equally on either side of the joint)."

-> It’s correct only if the eccentricity is within the limit.

"My question is...could these eccentricities be taken into account in the global analysis model and then "ignored" in the design of the joint? Would this be good enough to satisfy the BS EN requirement to consider the joint eccentricities?£

-> You would need to consider both of them because the eccentricity is significant in your case (e>do).

"The way I am modelling this truss is by using beam elements for the chords and truss elements for the braces."

-> It’s better if you also consider the braces as beams. In your case, since the eccentricity is out of the limit, the eccentricity moment should be distributed for all the members of a joint based on their relative stiffness coefficient, so the brace may have interaction between axial load & bending moment.

Regards,