Degrees of freedom at the supports (boundary conditions)

[rozsasarpi]

See this table for an overview. The current degrees of freedom are not valid for temperature load (they were good for traffic and pier settlement).

https://docs.google.com/spreadsheets/d/1q6h4RVGnkSxBDYe_FPl1PU1LNNTpeeRcqGQNGCORfj0/edit?usp=sharing

[jerbaroo]

@rozsasarpi a few clarifications:

y: is this the vertical or transverse direction? My guess is vertical.

applies only to one one node of the support, the rest of the nodes have the opposite

So e.g. the central node of the abutment fixed in z translation, and the remaining abutment nodes free in z translation?

applies only to all nodes of one of the piers in transverse direction, the rest of the piers and their nodes have the opposite freedom

So e.g. fix all nodes of one pier in z translation. All other piers at the same x position, are free in z?

[rozsasarpi]

y: is this the vertical or transverse direction? My guess is vertical.

You are right.

So e.g. the central node of the abutment fixed in z translation, and the remaining abutment nodes free in z translation?

Correct. Although I would fix one of the side nodes (that's a more typical solution). Note that this is just a guess from my side, I do not know how the actual boundary conditions are in reality, Diana might help, though it is probably not decisive as no temperature load is applied to the structure there.

So e.g. fix all nodes of one pier in z translation. All other piers at the same x position, are free in z?

Exactly.

The reasoning behind the second and third is that this way uniform temperature loads will not induce additional stresses in the structure (the structure can move), preferred solution whenever it is possible.

[jerbaroo]

I have sent a reply by email.

[jerbaroo]

Hi @rozsasarpi , I have a few questions if you have 10mins before your holiday, hopefully last for the year. And I am sharing my plan for remainder of December. Best, Jeremy

thermal.pdf

[rozsasarpi]

My comments are in the pdf. Let me know if something is not answered adequately or unclear.

thermal_ar.pdf

[jerbaroo]

Thermal loading

yes the moments should be applied at rx and ry arguments

Just to clarify, I am using notation of x (longitudinal), y (vertical), z (transverse), .

So therefore rx and rz

what do you mean by 'in meters'?

typo, newton meters

the degrees of freedom should be the same for all loadings

Currently I am using two settings for fixing nodes, A, and then B for thermal. A (from Diana):

• all y (vertical) translation fixed
• all z (transverse) translation fixed
• x (longitudinal) translation fixed for "red" piers in the image you sent B:
• all y (vertical) translation fixed
• z (transverse) translation fixed for: 1 pier at each x position, and one (central) node per abutment
• x (longitudinal) translation fixed for "red" piers in the image you sent

(except pier displacement, the pier being displaced gets to have y translation freed)

So should setting B be used in all loadings, including pier displacement and concentrated load?

moment_bottom?

Now included.

Updated plots

Do the plots at the bottom look more correct to you? Now it's like a wave. Issue was a mixup between rz and rx loading calculations.

Strain

Note: this section is not related to thermal loading, this is checking strain under a concentrated load.

are these strain convergence plots coming from a concentrated load? min and max (an in turn mean) might be considerably affected by the loading.

Yes they are.

How would they be affected?

Is there an alternative to check convergence apart from applying the concentrated load. Only concentrated load makes sense to me, because that is what the vehicle loading is based on.

Is this max strain location below the concentrated force?

I will have to check if the max strain recorded corresponds to the location below the load.

For these convergence plots I was recording min and max of all nodes in the model.

I will also explicitly record the strain under the loading position, and plot that.

Thermal

thermal-deck-moment-axial_load-Displacement.pdf thermal-deck-moment-axial_load-X translation.pdf thermal-deck-moment-axial_load-Z translation.pdf

[jerbaroo]

A question about Axis when selecting a temperature load. To simulate an axial load this is setting Treference to 1 in my case? And then for the moment load in a separate simulation it is like setting T1 to 1 and T2 to -1?

[rozsasarpi]

A question about Axis when selecting a temperature load.

Just to be sure:

• axial (uniform temperature): T1=1; T2=1; Tref=0
• pure bending (linear temperature): T1=0.5; T2=-0.5; Tref=0 (it is a bit nicer to have deltaT=1 for both temperature loads)

[rozsasarpi]

Just to clarify, I am using notation of x (longitudinal), y (vertical), z (transverse), .

Ok, I think it is good.

So should setting B be used in all loadings, including pier displacement and concentrated load?

Using B for traffic and pier settlement loadings should not make a difference compared to A (please check it). So I would stick to B.

Updated plots

thermal-deck-moment-axial_load-Displacement.pdf I would not expect the side spans to go downwards.

How would they be affected?

There is a large strain gradient close to concentrated loads, to capture it and the corresponding peak you need finer mesh in that region so convergence can be slower (especially if you have an even/homogeneous mesh.)

Is there an alternative to check convergence apart from applying the concentrated load. Only concentrated load makes sense to me, because that is what the vehicle loading is based on.

Yeah, it is better to check for the loading type you're going to use. One important question is whether you will have sensors right below the concentrated loads or not. If yes then you need to have convergence there as well.

[jerbaroo]

Muchas gracias por el feedback, enjoy Florence!

[rozsasarpi]

Haha, thanks. Enjoy your holiday as well!

[jerbaroo]

This is closed, thanks to Henco