Stability of Steel Trusses during lifting (Suspended Trusses) 1

There are a few schools of thought. The FHWA (NHI 15044) supports the idea of using the total length of the chord for girders, but allows that to be offset by using a very large Cb factor, partially in recognition of the stabilizing effect of having the load CG below the pick points.

One can always perform a FEA buckling analysis (with initial out-of-plane displacements) if the precision is warranted.

Hi @Lomarandil, this school of thought is about beams and girder (LTB), not about truss (flexural buckling).
I know some publications about lifting beams. The ASME BTH 01, in my opnion, is the best way to design lifting beams and spread bars, but any reference about lifting truss.

FEA buckling analysis could be used, but I have some doubt. Do you know any reference about it applied to lifting truss?

BR

The difference between beam LTB and truss flexural buckling is not so significant in a lifting application. After all, what is a beam but a truss with solid web members (or vice versa)?

Yes, an AISC erection webinar (I saw it back in 2012) describes the process for a plane truss.

ASME BTH is really not very applicable in this circumstance.

I also saw this webinar with Will Jacob and Clint Rex, but for me the method is too complicated (FEA buckling analysis + large displacemment in SAP2000).
I'm looking for some other reference that would make me comfortable. Thank you very much for your comments.

BR
Caio Marcon

So .... what are you going to use for loads during the lifting operation? just gravity loads? wind loads? dynamic "bouncing" loads?

FEA works quite well for buckling analysis (assuming you have sorted out the loading and boundary conditions). What are your "doubts"?

Dear @SWComposites I usually apply gravity load + a dynamic amplification factor (by GL Noble Denton or ASME BTH).

I have some question about boundary conditions in FEA buckling analysis. How do you build your model to solve the out of plane numerical inconsistency?

In reality, the lifting cables are not a support point in the structure, so there will be some numerical inconsistency in out of plane of the analysis.

How are the lift points not a support point? They are the support points during the lift. Maybe I don't get what you mean here but I don't see the issue nor the 'numerical inconsistency'.

If you are referring to the lack of global out of plane restraint at the support then that isn't how I personally would describe the scenario.

For a two point lift I consider the lift point as fixed in the out of plane direction because a simple coordinate transform from a global co-ordinate system to a local beam co-ordinate system results in that outcome.

*To be precise the coordinate system is defined by the plane defined by the two lift points and the hook point.

The lifting points is a support point to vertical loads but not for horizontal direction (out of plane or your plan of the analysis). This affect directly about unbraced length of the truss chords and the stability of the member.

A numerical inconsistency in the model can be defined when there is freedom to rotate or move in at least one direction and the assumptions of the stiffness matrix are not completely satisfied.

@caimarcon,
I may not understand this problem correct by I don't see why FEA shouldn't be a useful tool. I suspect that the primary issue is to describe the boundary conditions properly.

Based on the first picture you seem to have a truss and a (perhaps) complex lifting arrangement. If I assume that you have two points for vertical lifting supports, then that should be solvable. Then you have the horizontal issue, I assume that is your "numerical inconsistency" and that may be a software specific term. But if you don't have any horizontal support and you have some horizontal load effect, you have a problem . But based on the figure I think there are possibilities for horizontal support.

You mention the unbraced length of the chord and ask if the total length is a reasonable assumption. What do you intend to use that length for in your analysis? As a buckling length for the chord, I think it may be longer. But based on the figure I would model the truss and part of the lifting arrangement and see how they interact. But to get it correct you will probably need large displacement analysis since the displacements may be the reason for failure.

caiomarcon said:

My question is about design for stability, not practical field experience.

Click to expand...

Assume two lift points can be located anywhere along the top chord, not just at truss panel points.

Maximum stability is obtained when beam/truss bending stress is minimum.

To calculate minimum bending stress:
Magnitude of negative moment at Lift Point 1 = Magnitude of negative moment at Lift Point 2 = Magnitude of positive moment in Center Span.

For a uniformly symmetrical member with constant structural properties along it's length (e.g. steel "W" shape or square concrete pile) calculation of optimum lift points is straight forward (both cantilever lengths approximately 21% of total length)



A truss, such as the image in the original post, is not necessarily uniformly symmetrical and may not have constant structural properties along it's length. Calculating the optimum two lift point locations (for minimum bending moment in the entire truss) is more of a challenge.

Using any lift locations, except the two optimum ones, compromises the member's maximum stability.

All true SRE, but when we are talking about heavy lift/erection engineering, the question is not of determining maximum stability, it is one of sufficient stability.

There are many cases which preclude using the optimum lifting points. Crane rental fees (to have a machine on site with sufficient radius) being the typical limitation.

Lomarandil - The OP made it clear: "My question is about design for stability, not practical field experience."

All I am doing is holding him to that statement, and say so:
SRE: "Assume two lift points can be located anywhere along the top chord, not just at truss panel points."

Besides, in the OP's example photo the truss appears to be lifted at the second panel point (of eleven).
Would lifting at the first or third panel points be a better compromise?

He may not consider that location of the lift points influences stability... unless someone points it out... which is what I have done, and say so:
SRE: "Using any lift locations, except the two optimum ones, compromises the member's maximum stability."

Very true.

caiomarcon said:

The lifting points is a support point to vertical loads but not for horizontal direction (out of plane or your plan of the analysis).

Click to expand...

Shift your co-ordinate system and it IS a lateral support. This change of co-ordinate system thinking is a bit left field for structural engineers but it doesn't make it any less true. Think of it as a fixed boundary condition rather than at support if you will.

Either way you end up with lateral restraint at the point of vertical support. But you do lack twist restraint.

**This is my approach to analysis. I am open to having it refuted if it is accompanied by an explanation.

One of the keys to understanding the behavior of this setup, I think, is to recognize that a version of whole truss rotational restraint is required a the lift points, just as it normally is with a cantilevered beam. And that restraint is provided by the self weight of the bottom chord. Unfortunately, that feature of the problem makes it quite complex to evaluate in detail as it requires some rotation at the pick points prior to engaging the rotational resistance mechanism.

This may be the webinar that lomarandil referred to: Link. One is able to watch it and download the associated slide PDF's for free.

Agreed Kootk. Twist restraint is provided by self weight and that is critical to any analysis.

I'll have a look at the webinar on this. Thanks. I've designed several similar lifts but more through my own logic rather than through standing on the shoulder of giants.

This is the webinar that lomarandil referred to (please see the second speaker at 01:00): Link

This method was developed by Clint Rex based on the direct analysis method of the AISC. To solve the out-of-plane numerical inconsistencies, the speaker add small springs at all nodes for analytical stability while no interfering with buckling shape.

When I created this post, I was thinking of finding a more classic analysis method, but I already understood that this might not exist and the best way could be FEA Analysis Buckling.

Regarding the method developed by Clint, I found that the investigation of the chord strength capacity is too subjective (Track deflections vs.
applied loads to see softening of structure). What would be a reasonable limit?

I built an SAP2000 model based on this method to test but found a chord curve with another shape than the presented in the webinar.

Thanks guys, this problem is not elementary at all for me

BR