FEA for Bridge Analysis 1

[Andy10]

I would apprechiate some assistance in finding some good resources and or examples on how to best model and perform bridge analysis utilizing Finite Element Analysis.

As an example I would be interested in is performing a rigorous analysis on a Prestressed Channel beam structure(inverted U) discretely connected by either welded angles or bolted with no transverse PT utilizing AASHTO LRFD requriements. Also FEA on Channel girder bridge with Lateral PT at 1/3 points which developed cracks in the diaphragms near the PT ducts and into the webs of the channles.

Any help or guidance on how to find relavent information that would assist in modeling this and evaluating the model would be of great bennifit.

Thanks,
Andy

 

[Qshake]

I have often thought that a standard method of analysis should exist for modeling of regular bridges and have yet to find anything of the kind on the subject.

I have always perused the technical journals and recommend that as a place to start. ASCE's Journal of Structural Engineering and Journal of Bridge Engineering have in the past demonstrated the many techniques used by researchers. So, if your so inclined, you can read and draw some conclusions on what methods are best. Please keep in mind that researchers often have software that practioners don't and so the practiioner is left without a robust library of elements.

A big problem (probably more tedious) lies in the connection of various elements via centerlines, which are required mathematically but simply don't occur in the constructed environment. Some programs have special rigid links for these connectors and others do not. Those that do not have special connectors the user needs to create and insert them and that creates problems for models requiring finer meshes.

Another problem that exists for the analyst is that there is little testing done to verify the bridge FEA. So you really don't have much to baseline your results too. The code equations are not very good since even the most recent code equations don't capture all the parameters and are regressed. They are better now than the earlier AASHTO code equations. Another fallback is that the US code is generally focused toward a single line girder analysis, which can't accurately depict the behavior of a "real" bridge as well as some would like. The UK does much more FEA of bridges than the US engineers. This is my perspective only but I've been in the business for way too many years!

In addition to the journals a good book on practical FEA will help to understand generally which elements are better for what reasons than others.

So the answer lies in many references not a single one.

Good Luck.



Regards,
Qshake

Eng-Tips Forums:Real Solutions for Real Problems Really Quick.

 

[johnhors]

look at this perhaps?

http://www.lusas.com/products/bridge_tour_overview.html

http://www.Roshaz.com

quality, cost effective FEA solutions

 

[Andy10]

Thanks Qshake, I was kind of hoping I would get a response from you! I have really apprechiated your input on eng-tips! Thanks for sharing and your input!

As for the post, I was thinking along the line of doing a grillage type of analysis to obtain the loading / reactions at the transverse diaphragms and then attempting to use shell elements to model the actions at the diaphragms where cracking was occuring. I am not sure if this is the right way to go.

Andy

 

[Qshake]

Andy -

Just so I understand this configuration, you have concrete tub girders (channel beams) that are spaced at some distance away from each other but not immediately adjacent to each other? And the beams/girders have transverse elements for load sharing and lateral stability. Then you indicate you'd like to introduce some transverse P/T via a concrete diaphragm between the channels. Are the channel beams set so the "U" is up as in the traditional "U" beam or set downward? If downward, are the beams set against each other to form the superstructure with a concrete infill?

Regards,
Qshake

Eng-Tips Forums:Real Solutions for Real Problems Really Quick.

 

[Andy10]

Qshake,

No, the bridge configuration is adjacent channel girders (ie like adjacent box girders minus the bottom slab).

One system involves having diaphragms at the third points with a duct for transverse post-tensioning These diaphragms are cast with the precast section (part of the mold).

A second system consists of connecting the adjacent channel girders with angles embedded in the top slab of the channel section at the 1/3 points (including the diaphragms but no post-tesioning duct). These angles are then bolted or welded to provide a connection. From a design perspective no load distribution is assumed with this connection => one wheel load per channel is assumed.

Attached is a current detail utilizing the bolted angle connection.

I hope this clears this up. If not I could make a scan of the details and post them as well