Date(s) - 27 Mar 2014
10:00 AM - 10:50 AM
3043 ECpE Building Addition
Title: A distributed continuous-time gradient dynamics approach for optimal power flow problems
Speaker: Xu Ma, ECpE Graduate Student
Adviser: Nicola Elia, Associate Professor
Abstract: In this talk, we consider the non-convex optimal power flow problems. We are interested in solving these problems by applying the continuous-time gradient dynamics. First of all, we start our work with a simple class of OPFs, the equality constrained active power loss minimizations. We show that an optimal solution can be achieved by running the associated gradient dynamical system. An important feature of this approach is that it is naturally distributed, i.e., each bus in the network only uses the local information for computation. Next, we move to the general optimal power flow problems. We adopt a three-bus counterexample to numerically show that there exists a strong Lagrange dual for a general OPF problem. Although its explicit form is still not clear, this Lagrange dual is different from the known SDP dual. Finally, we derive an sufficient condition to characterize the saddle points for the Lagrangian associated with the OPF problem. Simulation results are also provided to illustrate our work.