Lifting a structure

[Anonymous]

I need to model the lifting of a structure using cables and a spreader bar.

I have read this topic:

https://forums.autodesk.com/t5/robot-structural-analysis-forum/lifting-the-construction/td-p/3420819...

I modeled the cables as solid round bars, d = 90 mm, and a new material I created that has E = 9 x E Steel and specific weight = 0.

I have applied Pinned-Pinned, and Pinned-Fixed releases to the cables, as you can see in the picture.

I have 4 static load cases including self weight, and a combination (2 x all four). I have added all four load cases as "global mass" in load to mass conversion, then I placed the lifting point above the point where the calculation note says "Coordinates of structure centroid with dynamic global masses considered".

I am not sure if my model is correct, I have big displacements (75 cm) even if I place spring supports in the four corners of the structure, UX and UY with 10 KN/m.

[Rafacascudo]

Why are you using "global mass"? This is only useful if you have any kind of dynamic analisys.

Also use truss bars for the "cables" to avoid using bar releases.

[Anonymous]

Hello, 

I used "global mass", to obtain the CG of the structure considering the 4 static load  cases, so I can get an approximate idea where to place the lifting point.

What I am not sure about is the entire process, it seems that if I vary the thickness of the slings, or just move the support point a little I get very different results for the structure.

There are three ways I identified until now about how to do it:

- thick slings (90 mm) with E = 9 x E steel, and releases. This is the initial model, it gives me large displacements.

- slings modeled as truss bars with no releases

- thin slings (5 mm, to avoid moments on slings) with a very high E, and no releases

I am sure there are other factors I could also modify.

Or should I use a model with directional supports? And should I have spring supports at corners?

What rigidity should these spring supports have? DNV code recommends spring supports at corners with 10 KN/m.

http://mgsicestorm.com/wp-content/uploads/2017/12/DNVGL-ST-E271.pdf page 71

Every time I get different results. I attach the model, with slings modeled as truss bars. Thank you.

[Anonymous]

I created a new model, I placed some short and very rigid bars under the lifting points, these bars point directly to the former points where the slings were attached to the spreader bar.

Bars are articulated at base. This model produces compression in the lower longerons of the structure between the lifting points - as it should.

Can anyone confirm this is a correct lifting model, or is there a problem with it? I attach a picture.

[Rafacascudo · Accepted solution]

There is a mistake on the CG considered because you also converted the SW case into global mass .It is not necessary because SW is already automatically taken in account in the global mass calculation.

So checking the mass in the calculation note and comparing to the sw combination reactions , you will easily see that sw was taken twice .

So , considering the correct CG which is

"Coordenadas do centroide da estrutura considerando as massas estáticas globais:
X = 9.971 (m)
Y = 2.202 (m)
Z = 3.434 (m)"

you will get the below axial forces on the main beams and cables

model attached

[Anonymous]

Thank you.

I was interested about the efforts in the structure when it is lifted, not about the forces in the slings.

The "Self Weight" case, is there a reason I should ever add it as Global Mass or Dynamic Mass?

Also I see you used "Skyline" method of solving. If I use Automatic I get an error, if I use SparseM I get some higher bar ratios compared to Skyline.

[Rafacascudo]

   I don´t know why different solvers are giving different results. Probably because the structure is too unstable . I would trust the analysis only if I could get reasonable results on "automatic" solver .

But , having a deeper look at your model,

1- You are using claddings for the model floor ,which is not very good to distribute loads with the ultimate precision you need for this type of structure. You should use Shell(meshed) calculation model for the floor.

2- Also , I don´t see the Self weight for all the objects assigned as claddings( floor and the 4 walls)

3- If possible , place your 4 lifting points on the structure placed symetrically to the structure CG .

Only then , we can think about the solvers .

[Anonymous]

If I transform cladding into panels, for example 2 mm thick steel, I add some rigidity to the structure. I don't want to count on the panels for this. How can I add shells without adding panels? Maybe I should place the loads directly on bars?

The self weight of the steel panels that are modeled as cladding is included in the "Insulation" load case.

The location of the loads in the structure is approximate, I am not sure about benefit of placing the lifting points symmetrically to the CG.

[Rafacascudo · Accepted solution]

Hi, @Anonymous 

I would adopt your model on msg 4 with DNV springs on bottom slab corners nodes. You can use directional supports on bottom lifting nodes(pointing dir.x to lifting beam nodes) instead of rigid bars . They should give the same result.

Still , I don´t understand why it is giving different results for different solvers.

Maybe @Rafal.Gaweda  @Artur.Kosakowski can clarify why

test model attached

[Simau]

@Anonymous @Rafacascudo 

in order to avoid instabilities, why not focus only on the lifting process.
In this test model the structure is simplified to the maximum and the lifting is done by cables.

[Simau]

you can of course place the cables wherever you want

[Anonymous]

Thank you Rafacascudo

The model with directional supports and springs at corners gives the same results as the model placed on rigid bars from message 4, if those bars are Fixed at the base and have Fixed-Pinned releases.

Also the same results as the model with slings modeled as Truss bars, with CG computed without Self weight as global mass, and using Skyline solving method. If I use Automatic or SparseM I get totally different results.

Simau - the structure will be lifted from the bottom, I believe that is why the model is unstable. Also -  I do not have those bracings in the walls.

[Anonymous]

Hello,

I have a problem with the reactions for lifting a structure. The model is almost identic to the first one posted here. 

I used truss bars for the slings and elastic supports at the corners. The CG I used is computed without the self weight case in the Load to Mass conversion.

The FZ reactions at the lifting point are incorrect (even 2 times smaller) than what I calculate. Sometimes I even get negative reactions.

Also the reaction values differ if I chose Skyline instead of Automatic for solving.