What are the limitations of grillage analysis?

In designing the number of cells for concrete box girder bridges, in case the depth of a box girder bridge exceeds 1/6 or 1/5 of the bridge width, then it is recommended to be designed as a single cell box girder bridge. However, if the bridge depth is smaller than 1/6 of the bridge width, then a twin-cell or multiple cell is a better. However, one should note that even for wider bridges with small depths, the number of cells should be minimized because there is not much improvement in transverse load distribution when the number of cells of box girder is increased to three or more.

For structural analysis of bridges, grillage analysis, which involves the structure to be modeled as a series of longitudinal and transverse elements which are interconnected at nodes, is normally adopted.

Grillage analysis suffers from the following shortcomings based on E. C. Hambly:

(i) For coarse mesh, torques may not be identical in orthogonal directions. Similarly, twists may differ in orthogonal directions.

(ii) Moment in any beams is mainly proportional to its curvature only. However, moment in an element depends on the curvatures in the beam’s direction and its orthogonal direction.

Grillage analysis cannot be used to determine the effect of distortion and warping. Moreover, the effect of shear lag can hardly be assessed by using grillage analysis. By using fine mesh of elements, local effects can be determined with a grillage. Alternatively, the local effects can be assessed separately and put in the results of grillage analysis.

This question is taken from book named – A Self Learning Manual – Mastering Different Fields of Civil Engineering Works (VC-Q-A-Method) by Vincent T. H. CHU.