The shape of Critical Infrastructures (CIs) has changed drastically in recent years, leading them to become interconnected systems with complex interactions. This will be especially true for future power grids, known as “smart grids”. In such complex systems, one of the main challenges consists in understanding the possible impact of failures on the overall system, and avoiding cascading or escalading effects. In this paper we present an extensible framework for the analysis of propagations in power networks, based on model templates developed using the Stochastic Activity Networks formalism. The approach is applied to a case study of a power network, derived from a real system. The obtained results demonstrate the effectiveness of the approach in assessing the impact of failure on the network, both when considering random failures and when performing what-if analyses on specific nodes.