Topology identification in smart grid with limited measurements via convex optimization

With the growing penetration of renewable and demand response programs which lead to frequent flow reversals and substation reconfigurations, correct identification of the topology becomes an imperative task in future power grid management. However, due to low measurement redundancy especially on distribution networks, the aforementioned task is inevitably challenging. In this paper, we are thus motivated to propose a maximum a posterior based mechanism, which is capable of embedding prior information on the breaker status, to enhance the identification accuracy. Building upon semidefinite programming, our goal is converted to solving a relaxed convex optimization problem. Within the optimization problem, the sparsity in prior knowledge is also promoted using compressed sensing technique to further alleviate the effect of insufficient measurements. Numerical tests on the IEEE 14-bus model corroborate the effectiveness of our proposed scheme.