I understand you want to model the part of the bridge deck, right? It depends on how good an estimation you need. I mean, if the value will be used mainly to estimate deflections and check how far from the limit stresses you are, you do not need to go into too much detail, as it won't change much regarding deflections. Generally, keep in mind that near locations with concentrated loads or small openings you'll need a more dense mesh than in the remaining body.

I'd suggest you tried with a simple mesh (OK, not too wide elements) and then start increasing its density, until the change in the deflections/stresses you get in the new model is not that significant compared to the previous one.

I understand you want to model the part of the bridge deck, right? It depends on how good an estimation you need. I mean, if the value will be used mainly to estimate deflections and check how far from the limit stresses you are, you do not need to go into too much detail, as it won't change much regarding deflections. Generally, keep in mind that near locations with concentrated loads or small openings you'll need a more dense mesh than in the remaining body.

I'd suggest you tried with a simple mesh (OK, not too wide elements) and then start increasing its density, until the change in the deflections/stresses you get in the new model is not that significant compared to the previous one.

Dear Ranjith,

Level of detail actually is a wide field and it is up to the research community how to deal with definitions. I just can speak for CityGML, which is a geometrical, semantic 3D Representation for Cities and standardized by the Open Geopspatial Consortium. It is able to model bridges as semantic objects as well. In the actual version of CityGML LoD0 represents just 2D shape of a building or bridge feature. LoD1 would represent the same feature as a block model. LoD2 is a representation of a feature with some more details of the outer hull. If you need to get some a geometric representation of some bridge installations, CityGML’s LoD 3 would be sufficient. LoD4, however, is to represent the interior of a building or bridge, respectively.

Ranjith,

The following links should give you additional guidance to arrive at the solution your looking:

http://www.biomesh.org/fea-basics

(info. is related to FEA in Biological models, however the 3-d concepts are transferable to Cvil Engineering)

http://www.dot.ca.gov/hq/esc/techpubs/manual/bridgemanuals/bridge-design-practice/pdf/bdp_4.pdf

(this are excerpt from a cvil engineering handbook for bridge design practices)

http://www.engr.sjsu.edu/ragarwal/ME165/ME165_Lecture_Notes_files/FEA_Lectures/Chapter_1_%20Introduction.pdf

(the section 11.5 Convergence, Assuring Optimum Mesh Size provides

valuable insight in regards to your problem)

http://home.cogeco.ca/~pbudgell/ANSYS_tips.html

http://home.cogeco.ca/~pbudgell/Modeling_issues.html#C2

(Tips on using ANSYS for FEA)

http://www.steelconstruction.info/Modelling_and_analysis

Basically, you have to refine the mesh until you reach a desirable mesh density.

May want toexplore the engineering analysis software from LUSAS as they specialize in software products for bridge analysis, design and assesmsnt. http://www.lusas.com/index.html

Its relative. But roughly speaking, models for analysis have less details. Typically, model simplification happens when a detailed CAD model is taken for analysis in CAE. If your aim is Analysis, then you can avoid such features while doing modeling itself.

Small-irrelevant features like holes, fillets chamfers are removed in CAE model as they complicate the mesh unnecessarily. But this needs some engineering judgment as such features, if are present near Boundary condition or are in load path, they are not removed.

If part has symmetry in it (mirror or pattern, revolve features) you can just analyse sub portions of it. If there are thin plate like portions, they are modelled as mid surfaces for shell elements in CAE. If there are slender portions, they are modeled as skeletons for beam elements in CAE.

Such simplified-abstracted-idealized model gives reasonably accurate results in lesser resources.

What are dimensions of your girder?

Dear Ranjith,

Generally, I do a sensibility analysis to observe the dipendence of the results by the mesh. You can leave with a coarse mesh and after you can refine it until when you view an negligible difference (

For wind turbines - a generalized, first and second eyegen mode analysis can provide a descent approximation. Having the modal displacements and forces you can easily find the stress and strain at any point. Very convenient for dynamic simulations.

Other wise, fem convergence analysis has to be carried out. Parts you find more interesting can be "cut out" together with the displacement or force state and simulated with a lot finer mesh if necessary.

Hi,

Its very simple..just simuate with corse mesh first, by this u will come to know what are the stress loaction your are getting make sure out of this which are the key imp locations. Then run gain with very fine mesh..at this time you can eliminate fillets and sharp cor's.I hope it will reduce your itarations and time.

Hi,

The 3D modelling made by SolidWorks or Catia may be meshed automate by Cosmos or Abaqus that are integrated as solution. You may change the FEM analyse and input easy the new material properties, change the restrictions in deflections, etc.,etc. That way of modelling used in mechanical and aviation engineering represent the easy and quickly way for that type of problems.

the 3D MODEL is considered the best for cable stayed bridges in which static and dynamic responses such as deformation and stresses are required. also soil structure interaction are easy to be implemented in such model