Hi Qiqin,

I think that the article below can answer your question:

Best regards...

http://www.mul2.polito.it/materiale/ArticleJournal/2015/CarreraPaganiZangallo_ComparisonVarious1D2Dand3DFEmodelsForAnalysisThinWalledBoxWithTransverseRibsSubjectedToLoadFactors.pdf

Hi Qiqin, 

For static analysis  it's obvious to consider the bridge as a beam based on the boundary  conditions.

For modal analysis it's not the case, the degree of freedom affects the analysis significantly. 

For shell elements the eigenvalues are calculated over a plane, so there will be more modes in both directions over a frequency range compared to that of a beam.

So when you decide the mode numbers for the plate you have see the mode shapes and name them based on the axis of displacement and number of half waves in each direction.

If you have already taken care it by now, then it is possible that for modal analysis over the breadth of the beam the degree of freedom is restricted which results in more resistance to the modal displacements in the plane of the beam. Which results in higher frequencies for the same mode.

Dear 

Before you use beam elements or shell element. You must know that the beam elements have 1D by mean 1 translation and 1 rotation for x direction (beam theories) . Shell element used for 2D in x,y direction ( plate theories). And also the thickness. 

See the book of J N Reddy

Dear,

You should also take a look at the mode shapes to compare the two analysis you conducted. There is a chance that you are comparing modes that are not equivalent just because the two elements have different degrees of freedom as mentioned above.

I think that the problem is that you are not comparing like with like: the 2D element mesh has lots of nodes while I cannot see any on the beam. Break the beam up into as many parts as the mesh and you should get comparable results.

Peter is right: you should divide the beam in several discrete elements in order to compute more eigenvalues. Moreover, the effect of the soil at on the walls should be accounted for by means of springs to be calibrated. In fact, you should expect high frequency modesowing to the limited bridge span.

Just to complement, the shell element carries forces in biaxial bending + membrane action (in-plane shear+axial forces), whereas the beam element has uniaxial behavior only, and if it´s an Euler-Bernoulli beam formulation (i.e. not Timoshenko), shear deformations are neglected, too. Based on your bridge geometry these issues could be relevant, so don´t expect the same frequencies.

Hello!

did you checked the mass details of the structure??.(for both beam and shell by assigning the cross sectional properties).

I guess the mode behavior is purely based on the inertial forces acting on a body.

also the DOF and BC's are play role in creating a mode.

hope this will help you to resolve your problem.