Thermal analysis of composite slab

[bothul]

Dear All,

I am trying to model a fiber reinforced concrete slab on steel beams. I have applied the load in the first step and applied thermal conditions (convection and radiation) in the second step. The analysis reached 400seconds (tops) ( I was expecting 7000s). I have modeled the structure with solids. Slab tied to the beam.
the analysis was run with Abaqus Standard. I have tried the same model using Explicit but the second step won't even start due to convergence. I tried contact between slab and beam.
I have tried the same structure with shell elements but still "no go".

Does anyone have references related to the topic? I would be glad to have someone to discuss to.

Regards everyone,

 

[corus]

Presumably it failed in the second step when you applied temperatures.
Again, presumably it's a coupled temperature displacement model so are there significant displacements that could cause convergence problems during the thermal loads? Have you tried just running a thermal model alone assuming tied surfaces and then loading the temperatures into a general static model? What was the error message when the job failed?

 

[bothul]

Thank you Corus for reply!

For both Standard and Explicit I have used: "static, general" analysis for the first step and "coupled temperature-displacement" analysis for the second step.

For the Standard part the error is: "Time increment required is less than the minimum specified"
For the Explicit (by mass scaling) after 20h, it was still running and it reached 0.8s from the 7000s of the step time while the OUT file is 665GB,,,,so I stopped it.

Your message gave me an idea. I tried the analysis without structural load (pressure on the slab) and I just want to see how it behaves due to expansion. Maybe the expansion in the steel beam is cracking the concrete slab. The concrete model I have tried is Concrete Damaged Plasticity.
Now the analysis is running. I will be back with the results.

Best!

 

[bothul]

Hello,

So, after running the Standard model again without structural loading, only with thermal conditions, the model stops at 600s. I obseerved that the beam is expanding (at the bottom flange)more than 12mm while the slab expands 1mm. Do you think this can be the cause of "in-convergence".
The message is : ***ERROR: TIME INCREMENT REQUIRED IS LESS THAN THE MINIMUM SPECIFIED

I tried the explicit model but I am pretty sure I don't do the scaling correctly. Can you guide me to some tutorials, examples or materials on scaling? Do I have to scale mass or time. Due to the fact that I have also temperature in the elements I think the mass scaling is not advised.

Thank you!

Best!

 

[corus]

If you run the thermal transient separately from the structural model then you can scale the mass in the structural model without any problems. A couple temperature displacement model takes a lot longer to run and is ill advised to carry out unless your temperatures are significantly dependent upon the displaced shape. My advice would be to keep the model as simple as possible and to introduce more complexity after you've got some results. It could be your model fails due to contact 'chattering' causing smaller and smaller time steps. If so you'd have to review your mesh and how you apply the contact definition.

 

[bothul]

Hello everybody,

Dear Corus,

Thank you for your advice. I works faster now with sequentially analysis.
I have modeled the composite slab using shell elements. I have tied the beam to the slab. The analysis stopped due to over stress in the beam due to compression ( the beam is expanding more than the slab). I defined a contact between the beam and the slab: Tangential= Frictionless and Normal= hard contact . Contact: Surface to surface, Kinematic contact method, finite sliding.
I run the analysis and it can be observed that the slab and the top flange of the beam are distanced at the beginning of the step. also the temperatures in the slab after 250s, are less than the predefined field of 20C in initial step and the temperature are not increasing according to the heat transfer analysis. The temperatures in the slab are OK.

The warnings I got are:

The option *BOUNDARY,TYPE=DISPLACEMENT has been used with a jump in displacements at the nodes in node set WarnNodeDispBCJump-Step1 at the beginning of the next step; all jumps in displacement across steps are ignored. See the status file for further details.

Both the file option and the data line format option are used for the *temperature in the same step. For a given node the values given on the data line will take precedence if op=mod parameter is specified or if op=new parameter is used to specify new values for the removed field

Do you have any advice regarding these warnings?

Thanks!
Best

 

[corus]

You can check that the temperatures have been read in correctly by requesting the temperature as an output variable in the structural analysis, and then viewing them from the structural odb file. If your mesh isn't the same between the thermal and structural model then you can get errors in the results unless you used the incompatible mesh option. I also choose the same time interval in the structural model as the thermal model though I think that's not strictly necessary. Also check your bstep, estep, etc. values are correct when reading in the temperatures.

 

[bothul]

Dear Corus,dear all,

I am analyzing the same structure described above, concrete slab with steel beams.
I have performed a sequentially coupled thermal analysis of the slab (with 7 integration points) and then a mechanical analysis.
I was checking the temperatures given by the thermal analysis and the temperatures in the mechanical analysis. The temperatures in the thermal analysis varies from 200(NT11) - 600(NT17) degrees across the thickness of the slab and in the mechanical analysis the temperatures between the two faces are close (differ only 50 degrees) and are about 600 C. I have used the same mesh for both analysis, the same integration points, I only changed the element type. The thermal analysis has one step with 9000s and for bstep I used 1 and for estep also 1.I didn't defined the beggining and end increment

Do you know what might be the cause of this temperature difference between the two analysis when comparing NT11 and NT17?

Thank you

Best wishes!