Small disturbances can trigger cascading failures in power grids

With the sharp increase of power demand, large-scale blackouts in power grids occur frequently around the world. Cascading failures are the main causes of network outages. Therefore, revealing the complicated cascade mechanism in grids is conducive to design the efficient policy to restrain the failures and further ensure a stable supply of power to users. Motivated by the recent advances of network dynamics, we proposed a framework based Lyapunov stability to analyze the dynamically induced cascading failures in complex networks. We abandoned the assumption that the network is stable in traditional static failure models and then detected that small disturbances actually can trigger cascading failures in unstable networks. What's worse, such failure usually accompanied the overload failure of lines during the equilibriums conversion process. Through the simulation analysis of the Spanish grid, we summarized that the features of this new failure mode include low incidence, large destructiveness, and fast propagation speed. And it usually tends to occur first in edge nodes and subsequently propagate to the central nodes. These features are consistent with the empirical observation of outages in practice.