FEA of castings

[gurmeet2003]

Hello greetings to everyone,

I am trying to model hydrostatic testing to destruction of some Ductile Iron castings. All castings have some defects. But my FEA model does not model these defects. My question is: How much discrepency should one expect in the predicted burst pressrue because of this discrepency?

Thanks

Gurmeet

 

[gelu]

Try to model the defects too.
Small cracks by unconnected nodes, empty spaces in the volumes by deleted elements and so on.
Maybe you will feel necessary to do non-linear and large displacement analysis.
gelu

 

[corus]

Castings can have a significant number of defects that can not be modelled by deleting elements or by inserting cracks arbitarily. The material can have a high level of pososity that, in effect, greatly reduces the Young's Modulus of the material. You would need to measure the material strength in some way before being able to make any prediction on the burst pressure.

corus

 

[MToft]

It might help to get some idea of the defect size and number. Two possibilities are:

1) Model the casting process with software like Magma

2) Measure and count defects post-manufacture with non-destructive testing or by cutting slices

You could then maybe take some account of their effect by reducing the yield or strength in proportion to the area or volume reduction. If you're not sure of the defect size and number then possibly try a sensitivity study to see just how significant the problem would be.

Hope this helps + good luck!

M Toft

 

[magicme]

hi
what you are trying to do is extremely difficult if the casting is a complex shape. the gradients of stresses may be extreme and high stress areas may be very localized....and may or may not be near defects.....and the burst pressure of course would change depending on where the defect was.

you would have to make some educated guesses of where the defect are (get the casting people to agree to these probable locations)then reduce the local element modulus around each location. run a static stress case and compute a "failure pressure" based on the local stress around each "defect" location. that would give you a matrix of defect locations and "failure pressure".

i think that will work.

daveleo

 

[gurmeet2003]

Gelu, Corus, MToft, daveleo

Thanks for your help.

Gurmeet

 

[tonireneirene]

Gurmeet:
It is not physically possible to the effect of certain casting defects especially spongy shrinkage, gas holes or microporosity on FEA , since they affect
properties very diffently.

Programs like Magma are useless when it comes to predicting any of these defect types they are only fit for gross shrinkage.

Tony