Thermal stress analysis 1

[schdan65]

Hello forum,

I work with ABAQUS 6.4-1 and have the following problem: I’d like to simulate residual stresses which appear in case of a welding process of two cylinders (one of steel, one of cemented carbide, welded frontal) during cooling down.

Now I have the following questions:

1.Which material data are necessary? Are E, ?, ? and the thermal conductivity enough?
2.Is it reasonable to modelling both components as one part and define over different sections the different properties? Or is it better to model two parts and assemble them in the Assembly Module and define in the Interaction Module a “weld” connection?
3.I’d like to simulate the cooling down process with a given time (cooling down from 1400 °C to 400°C within x seconds).
4.I’d try to do this with a Coupled temp-displacement – is this the right way? Or there are alternatives?
5.In the Load Module I have defined a start temperature under BC of 1673 °K (1400 °C). In the Step Module (as well Coupled temp-displacement) I’m not really sure what I have to define under Basic and Incrementation. How can I define in the initial Step a start temperature of 1673 °K (1400 °C)?
6.Furthermore I have selected in the Load Module under Load the type Surface Heat Flux to define the cooling down process. Here I’m as well unsure what I have to define under Magnitude (end temperature or area-related value). Have I now my under BC defined start temperature set on inactive? When yes, how does this work?

I’m an ABAQUS-beginner and thankful for every kind of help. Thanks in advance for all who tortured themselves trough my novel.

Greetings,

Daniel

 

[corus]

As its a thermal transient you need conductivity, density and specific heat. In your case the properties will be highly temperature dependent and you'll also have to include the effects of latent heat, and even phase changes in the materials. Very difficult to obtain.

You can either use two parts and tie them together or make a single part and partition them to assign the different material properties.

You don't need coupled temperature displacement, unless one is dependent upon the other. In this case I doubt that.

For cooling I'd be looking at natural cooling rather than a surface heat flux, so you need natural convection plus radiation. I'd refer to a text book or look at previous topics in the forum.

corus