For a given problem, how should i decide on the grid resolution? Like for eg, how should i decide on the no.of elements or cells, type of cells, grid density at high gradient regions etc etc.... Plz help me

Decide on a decent size grid. Run your simulation and look at your variable of interest. Double your grid and see how it changes. Continue doing this until results do not change. If you have anlytical/experimental data compare simulation results to it.

Venkat,

You can probably look at Verification and Validation at the following website. It might have what you require. As agg suggested, you can run on different grids and compare it with existing computations/experiments.

http://www.grc.nasa.gov/WWW/wind/val.../tutorial.html

Also, This one :

Roache, P.J., Verification and Validation in Computational Science and Engineering, Hermosa Publishers, Albuquerque, New Mexico, 1998.