Question:
Hi,
I have tried to model my bridge using grillage method and plate method. I have three query regarding my model's output.
1. I've noticed that the bending moment envelope due to "Shrinkage Secondary" in both model differs significantly (in terms of shape and value) as shown below.
Shrinkage Secondary BM Envelope for Grillage
Shrinkage Secondary BM Envelope for Plate
2. I have also checked the envelope for Creep Secondary. The two model doesn't have big difference but there is
positive sagging moment for the
grillage model, but
not for the
plate model.
Creep Secondary BM Envelope for Grillage
Creep Secondary BM Envelope for Plate
3. As for the other static and moving load cases, it can be noticed that the results obtained from the
grillage method is
more critical (higher moment) as compared to
plate method.
SW BM Envelope for Grillage
SW BM Envelope for Plate
Since the difference is only on the method used in modelling, I assumed that results should be more or less the same. I have tried to diagnose the problem, however, I've failed to find the cause for this. I am unsure whether it's my loading definition issue or other issues, which causes this discrepancies. I hope to hear some comments from MIDAS team on this.
Attached together is the model using grillage method for your reference. I will attach my model with plate in the subsequent reply. Hope to hear from you soon.
Thank you.
Regards,
Answer:
Hi Alvin,
For all beam elements
If we want to consider composite action, (where Girder + Deck properties need to taken together for calculation of design forces), then we select a
composite section in a
grillage analysis. In this case, the effect of the concrete deck is taken into consideration via composite stiffness values as shown below:
Now if we want to model deck via plate elements, then
then plate elements stiffness will also be considered additional to the composite section. Hence we need to alter the stiffness via stiffness scale factors. Such that double stiffness of deck should not be considered.
kindly let me know if I missed anything.
Thanks,
Suman
Hi,
In the case of the deck to be completely modeled as plate elements, which can be used for design, we need to ensure that girders are not composite type but instead should be steel girder sections.
In the below snapshot, I have highlighted the plate elements which represent the deck slab, and the girders are steel sections. PFA the example model
If we consider the composite section for girders and model the deck slab with plate elements, then we will consider the double stiffness of the slab in the analysis calculations.
kindly let me know if I missed anything.
Thanks,
Suman
Dear Suman,
Sorry for the delayed response. I have tried the following :
a) Replace composite with normal beam element.
b) Applied concrete properties to the plate element (slab).
c) Reanalyzed the structure.
d) Reanalyzed the structure with addition of "Change Element Dependent Material Property" to whole structure.
e) Reanalyzed the structure with addition of "Change Element Dependent Material Property" to beam elements only.
Basically, it seems that plate model of option "e)" gave a more closer "Shrinkage Secondary" result as compared to the grillage method (mid span having less moment).
As for the "Creep Secondary" result, the plate model showed very different result than the grillage model.
As can be seen in my first ticket submission, the two model doesn't have big difference in the result for Creep initially.
Can I know why is it so?
Regards,
Hi,
I am DK Lee. Your ticket has been reassigned to me.
First of all, when you see the beam diagram for the composite section in your grillage modal, you need to select one among Total, Part 1 and Part 2. Total = composite section, Part 1 = beam only, Part 2 = slab only. In the plate model, the beam diagram shows the moment for the beam only. Thus, in the grillage model you should select Part 1 to compare the moments between grillage model and plate model.
Secondly, in the grillage model, the ages of beam and slab are 28 days and 7 days, respectively. The h value for the slab is 0.179.
But, in the plate model, the ages are 14 days for the beam and 0 days for the slab as shown below.
In the plate model, the slab is modeled with plate elements for which the h value defined from the Creep/Shrinkage dialog is applied. It is 0.01 as shown below.
In order to apply h = 0.179 m for the slab, you need to create separate material ID and Creep/Shrinkage ID for slab as shown below and assign to the plate elements only.
Those are major errors in you modeling. I have attached the revised files for your reference. Note that I deleted some data including moving load analysis to reduce the analysis time. Also, I created two simple beam models, one for composite beam model and one for beam+plate model. You can use these models for your further comparison.
Regards,
DK Lee