Power grids, which form the cornerstone our technical infrastructure,
will undergo a revolutionary change in the upcoming decades driven by
the advancement of renewable energy sources. Large centralized power 
plants are put out of operation in favor of many small, distributed
and potentially fluctuating generators. To face this enormous
challenge for the reliable operation of the power grid, both the
structure and control of the grid must undergo a radical change. This 
development must be supported by a great advancement of our theoretic 
understanding of complex networked systems.

In this talk I discuss the collective dynamics of power grids using an
idealized oscillator model that model grid dynamics on coarse
scales. This model bridges the gap between abstract large-scale,
basically structural network models from physics on the one hand and
detailed simulations of small electric engineering systems on the
other. I analyze self-organized synchronization of the grid depending
on the network topology and study the impact of dynamical
perturbations and the influence of topological changes. I find that, 
contrary  to  common  intuition, the addition of new transmission
lines may reduce the overall grid capacity and explain this phenomenon 
by geometric frustration in the oscillator model.