Load distribution in girders: FEM vs traditional methods

[BridgeDOR]

While analysing for monolithic RCC slab-on-girders bridges by SAP2000, I got some strange results. For simplicity Lets take an example as follows:

Bridge span = 25 m
Number of longitudinal girders = 3
Number of cross girders (diaphragms) = 6 including end diaphragms
Spacing of logitudinals = 2.5 m
Spacing of diaphragms = 5 m
Let's ignore the deck slab for simplicity.

Let the CG of an arbitrary moving load be applied at 0.25 m outwards from the central longitudinal girder.

Now after analysis, the moments on the middle longitudinal girder was higher than the end (outermost) girder. If we analyze by traditional method (e.g. Courbon's method) in such cases the end girder has larger moment than the middle girders.

Is my model wrong? Or is it natural when we analyze by FEM? I would greatly appreciate if someone takes the trouble of modelling the above structure and see what the result is?
Thank you!

 

[BridgeDOR]

Sorry, i missed some information. The response of the longitudinal girders varies according to the stiffness of the diaphragm and spacing of the girders.