Learning Forum

[bermetkerimkyzy]

Hi everyone,

I'm trying to simulate the creep behavior of a notched plate. I have a rectangular plate with a circular hole in the center. I'm modelling the quarter of the plate. 

When looking at the results of the stress near the notch tip (I'm looking at the max stress values), I noticed that after the initial loading step once the creep is activated, there's a sudden and very big drop in the stress values. Can somebody explain why so? According to the analytical solution, some stress relaxation is expected however not that big. Please for the stress values see the image below:

I put the model in the attached files.

Thank you!

 

BR Bermet 

[peteroznewman]

Don't look at stress.  The first few increments are to initialize the creep model and can be ignored. The stress is constant because the pressure is constant.

In a creep model, look for changes in deformation over a long time.  If I request 1000 substeps, request 1000 seconds of time and request the Y Deformation of the top edge of the part, there is a 0.168 mm initial displacement as the load is applied, then over the next 1000 seconds, the part creeps an additional 0.11 mm to get to a total deformation of 0.278 mm.

[bermetkerimkyzy]

Thank you for the answer, maybe I put the question wrongly, the initiation of creep is exactly what I'm talking about. I want the the stress during the creep to be around 30 MP. After the first step 10^-8 sec when there's no creep just to initialize the stress it goes to a desired ~30 MPa, but then it suddenly drops to ~25 MPa. My notch radius is 2mm, so stress concentration factor is 2.73 MPa. Therefore, with the 10 MPa applied on the edge of the specimen, I should have those 30 MPa. Why does the stress drop once it's initialized? And how can I initialize it correctly? 

 

thank you. 

[bermetkerimkyzy]

btw, I'm talking about the maximum stress values, since I want to check the phenomena only at the notch tip.

 

[peteroznewman]

1) Use smaller elements near the high stress gradient.

2) Use a single analysis step and turn on creep while the pressure is ramped up over .001 second.

[peteroznewman]

Isn't that how the pressure load is applied, as a Ramped load not a Step load?  That means it starts at zero at time zero and ramps up to its full value at the end of the time step. You can have an End Time of .01 seconds and get the same result.  If the End Time is too large, like 10 seconds, then significant creep occurs while the pressure ramps up to the full load.

[bermetkerimkyzy]

Yes, I found it. Thank you. I couldnt understand how can I apply pressure without having a first load step initializing it.