SCADA (Supervisory Control and Data Acquisition) systems are widely used to monitor and control the behavior of electric power networks. As the data communication of SCADA systems is increasingly dependent on the Internet, the threat of a data attack becomes more imminent. For better protection it is vital to identify in the network the measurements vulnerable to data attack. An important vulnerability analysis problem can be posed as a constrained cardinality minimization problem, which in general is NP hard (i.e., not tractable). This cardinality minimization problem has the electric network interpretation that under certain constraints, the nodal voltages should be chosen to minimize the number of nonzero branch currents and nodal injection currents. This interpretation, with the additional assumptions such as all "branch currents" being measured, leads to an equivalent generalized minimum cut problem with costly nodes. This problem turns out to be solvable in polynomial time after an appropriate reformulation, and hence the original network vulnerabilities analysis problem is in fact tractable. We will demonstrate the computation results numerically on a realistic network.

Calendar of seminars