Welding dimensioning

[Ignicolist]

Hi all:

I'm designing a steel structure. This structure consist on welded steel plates (with different, but similar thicknes).
My main issue is that I find a lot of probles while dimensioning the thicknes of weldings.

I made a "hand calculation" of the forces in the position of the weldings, but comparing with FEM calculation, I find that in the point where deformation is higher, FEM stress is higher. This is clear, as in hand calculation all plates are considered rigid. In addition, the calculations are simplified because the structure is quite complicated.

I would like to know how to dimension the weldings. First of all regarding static force and then considering fatigue. I would like to dimension them to achieve "infinite" life, or at least more than 30.000 cicles.

I can easily calculate the structure with FEM, but I don't know how to dimension the weldings with these results.

I would appreciate any help and if additional information is needed, just ask and I can provide it.

Thanks in advance for your help

 

[desertfox]

Hi

I assume your talking about fillet welds? or do you mean butt welds?
Assuming the former dimensioning the weld is either done by throat thickness or by leg length.
On the site you will see a and z dimensions where (a) is the throat dimension and z is the leg length.

see this site:-

https://www.google.co.uk/search?q=d...=-BSeRI673-U2DM:;geCEiuih7TsB8M;http%3A%2F%2F

http://www.roymech.co.uk%2Fimages%2...2FUseful_Tables%2FDrawing%2FWeld.html;401;393

Another site for dimensioning:-

http://www.ddu.ac.in/academics/fot/mech/wp-content/uploads/2012/07/Chapter-3B_MDID.pdf

for static and fatigue calcs of welds go here:-

http://www.gowelding.com/

 

[Ignicolist]

Thanks for the quick answer.

As you said, I would like to make fillet welds and dimension the throat of all welds.

I check the last link you send about this calculation. I see that I need to find the stress in the most critical point of the welding. This stress must be splited in two directions (perpendicular to weld throat - Normal stress) and (paralel to weld throat - shear stress).
However, as my structure is no simple, stress can't be hand-calculated. In addition, stress is highly influenced by the deformations. Is is possible to use any FEM calculation to obtain these stress? (If I try to simulate welding with FEM to obtain these stress, I only find huge stress due to stress concentration.

Finally, regarding Fatigue. I have to find the range of stress in the "critical point" of the welding. (In my structure I have, from 0MPa (No force) to Von_Mises_max=80MPa)
With this data, I can find about 312.500 cycles. This works for any steel? And, it is the same to have an stress range of 0-80 as a range 100-180?

 

[desertfox]

hi

Can you upload a drawing of your structure I cannot imagine what its like from your description, here is another site for weld calculations, best I can do :-

http://www.roymech.co.uk/Useful_Tables/Form/Weld_strength.html#Guidance

 

[SnTMan]

For shop drawings, leg dimension is much preferred.

Regards,

Mike

 

[tbuelna]

Ignicolist-

The situation you are experiencing is common with almost every structural weld joint design. Areas in a structure like those created by fillet welds tend to produce local stress concentrations due to abrupt changes in section properties. Additionally, it is general practice to apply knock down factors to an analysis of a weld joint for weld defects such as porosity, undercutting, etc. If cyclic fatigue life is an issue, using the appropriate analysis knock down factors for weld defects is especially important.

Ideally, due to inferior mechanical properties that may be present within a weld joint, you want to design your weld joint and the adjacent local structures such that there is no stress concentration at the weld joint itself. This may involve adding doublers, or gradually increasing the section thickness/stiffness of parts around the weld joint. A well designed weldment has a nice even distribution of stress between the weld joint and the adjacent structures, with stress levels at the weld joint within acceptable limits that take into account all appropriate analysis factors.

In short, how you dimension the weld itself is only one part of proper weld joint design. Your weldment design should also include optimization of adjacent structures to eliminate stress concentrations at the weld, and should even consider how manufacturing and quality control might effect weld properties.

Hope that helps.
Terry

 

[PeteDB]

The link below is for an article which explains how to size welds using data derived from FEA.
See if it helps you.

http://www.eng.uwo.ca/designcentre/FEA resources/116_Welds.pdf

 

[Ignicolist]

Thanks for all provided information! I'm checking all and trying to use it in my structure.

Anyway, I enclose pictures and .stp file from one of my structures. I simulate the weldings with non-calculated dimensiones (random).
The force is applied direct in the marked 4 holes. The point of application of the force (64.000N) is separated from the center of these 4 holes by 127mm and 95mm (check the pictures). All this can be sumarized saying that there are 1 force and two torques.

Thanks a best regards to all!

 

[desertfox]

Hi

Not sure about others but I can't open a step file any chance you can make it a PDF file.

 

[Tmoose]

I think European or Asian standards may reference the opposite of US standard RE: leg vs throat dimension

 

[Ignicolist]

Now you can find some pictures and a PDF file that you can move. I guess that this gives a good overview of the structure.

 

[desertfox]

Hi again

The file extension says PDF.rar so again it won't open

 

[Ignicolist]

I think you only need winrar. Anyway, here you have the file

 

[desertfox]

Thanks for the PDF, I am not on windows software so win rar is of no use.
So what you have is a bracket and presumably loaded perpendicular to the four holes in the bracket, I think we can get some good approximations of the stresses using hand calculations which I will try to post later.

Desertfox

 

[Ignicolist]

Thanks for your support!
You are right, the force is applied in the holes, but the point of application is separated. In these pictures I show the position of this point where 64kN force is aplied:

 

[desertfox]

hi
I haven't any dimensions for your bracket so I have laid out roughly how you can analyse the welds in bending once the resultant force as been resolved into horizontal and vertical components, in my uploaded files the force in the x axis only imparts shear stress on the weld in the transverse direction which I believe is fairly straight forward and you can handle that without input from me.
Another consideration for the bracket is the bolt holes, or more importantly the size and number of bolts that fix the bracket, these might turn out to be your weakest link and not the welds.