Estimation of smart grid topology using SCADA measurements

Power grid topology is essential for various aspects of smart grid monitoring and operations. Recent studies show that by using the grid topology, an adversary can construct stealthy attacks that can cause significant disruption to power delivery and the critical infrastructure. This paper shows that the power grid topology can be approximately estimated simply by observing multiple power injection measurement data. We formulate the topology estimation problem as a blind convex optimization problem and solve using an efficient alternating direction method of multipliers (ADMM). To evaluate the performance, eigenvalue analysis and centrality measures from the complex network theory are used. Experimental results using IEEE test system demonstrate that the estimated grid topology is very close to the original topology and exhibits almost similar patterns and characteristics.