but as far as I know, Smooth Particle Hydrodynamics (SPH) may be the best technique to simulate shock-waves, Eulerian or lattice models seem to fail at it.
The topic of simulating explosions has an extensive history, and there is a large literature on the subject, as well as open-source simulation codes that are widely available. If you are interested in the details of the shock propagation, SPH is not at all the ideal tool; instead, there is a long history of 'shock-capturing' schemes that have been developed, such as flux-corrected transport (FCT) and - perhaps most prominently - the various methods based on the Godunov scheme and the idea of approximate Riemann solvers (such as Piece-wise Parabolic Method, or PPM). All this is well-documented in the literature and easy to look up. If you want access to one of the open-source codes that can do problems of this sort from the get-go, one I know well (because I was involved in its creation) is the Flash Code (http://flash.uchicago.edu/). Have fun!
It all depends on what you intend to simulate. I suggest you go for a CFD code FLACS ( FLame ACcelerator Simulator) particularly if congestions (obstacles) are present.
LS-DYNA is reasonably accurate if you are including structural response and need to couple this to the fluid pressure. Although not considering FSI, I have published a recent paper which has information on explosive modelling using LS-DYNA, including experimental validation.
Dr Sam Rigby
Article A Numerical Investigation of Blast Loading and Clearing on S...