Cascading failures in smart grid: Joint effect of load propagation and interdependence

The smart grid mainly suffers from two types of cascading failures: 1) interdependence cascading failure and 2) load propagation cascading failure. The former one happens due to the interdependence between power grid and communication networks. The latter one is caused by the load propagation in the single power grid. A tiny failure leads to the simultaneous occurrence of these two cascading failures. In this paper, we study the system robustness by considering the interdependence and load propagation. First, we develop a mathematical tool to analyze the load propagation in single network. Then, a percolation-based method is proposed to calculate the remaining fractions of survivals after the cascading failures stop. We estimate the node tolerance parameter T (the ratio of capacity to initial workload) threshold T´_c, below which the entire system may suffer from the cascading failure. The effect of interdependence on T_c´ is also studied, where lower Tc´ is required for the less compact interdependence. We prove that the system performance approaches to the upper bound once the tolerance parameter T → ∞. Our analysis indicates that the fraction of survivals in the power grid is always greater than that in communication network, although the initial failure occurs in the power grid. The extensive simulations validate our mathematical analysis, and demonstrate that the relation between the number of initial failures and tolerance parameter threshold is super-linear.