To look for methods for improving the robustness of power-communication coupled networks and guiding the construction of the smart grid

we proposed a cascading model for the power-communication networks with failure probability based on DC power flow optimization model and the scheduling function of communication network. We conducted edge attacks to study the effects of coupled patterns and coupling strength on the robustness of power-communication coupled networks

and analyzed the effects of failure probability p0 of nodes between the two networks. Simulation results show that the inner inter-similarity coupled links can significantly improve the robustness of power-communication coupled networks

and reduce the failure scales

especially the betweenness-degree（B-D） coupled pattern. Meanwhile

there is no effect of p0 on the cascading propagation of small-scale power-communication networks with scale-free property. Furthermore

results also show that a similar phenomenon can be found that the average load loss of different coupling strength has a downward trend after rising first with the increasing failure ratio of communication edge. High coupling strength can effectively reduce the scales of failures when the failure ratio of communication edges is small.