According to the addon description, the 2 features it provides are self-collision of particles and spring-based interaction. (Disclaimer: i haven't used the addon yet nor looked at the code)
Collision may be suitable for some types of granular material, but i think that for true sand-like behavior you would need a tremendous amount of particles and a number of friction effects would simple not be captured by this approach. Already with the moderate amount of particles in your example the solver becomes somewhat unstable (this is simply a limitation of the method as far as i can tell, not a bug).
Spring-based interaction otoh leads to soft-body type motion and probably cannot produce sand material effects either. Connecting particles this way creates elastic forces that don't allow particles to flow past each other.
I would suggest to try and model the sand surface with keyframe animation instead, possibly with displacement and shape keys you could get enough control.
Non-Blender remarks:
In current research and other simulation packages material point methods (MPMs) have been developed in recent years, which allow a much wider range of material simulations, which don't fit well into the soft-body or pure fluid simulation categories. Unfortunately to my knowledge none of these have found their way into Blender at this point, but it might be possible to use external software in the meantime.
For reference: http://www.cs.ubc.ca/~rbridson/docs/zhu-siggraph05-sandfluid.pdf
The sample principle methods have been applied to snow and foam like materials:
https://www.youtube.com/watch?v=O0kyDKu8K-k
https://www.youtube.com/watch?v=3_DdAXV3IDw