Online Algorithms and Optimal Offline Algorithms for Dynamic Optimal Power Flow

Date
2014-09-30
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Optimal Power Flow (OPF) studies the optimal routing of electrical power within a power grid, between supply and demand locations. Along with the Unit Commitment Problem (UCP) that studies generation unit scheduling, OPF is one of the two key problems at the core of power network optimization. Dynamic OPF (DOPF) is the OPF problem in power grids with battery energy storage explicitly considered. This thesis further incorporates renewable energy generation into the DOPF problem. Recent studies suggest the important role of energy storage systems and devices in balancing intermittent power supply and fluctuating demand, and reducing operational cost. DOPF is essentially a quadratically constrained problem (QCP), which has a non-convex nature, and is NP hard to solve in general. We provide a sufficient condition under which DOPF is equivalent to a Semidefinite Programming (SDP) problem, which is computationally efficient to solve. An optimal offline optimization algorithm is designed for DOPF by solving the SDP and then recovering an optimal offline solution to DOPF. Addressing the dynamic nature of DOPF, our work further proposes the first online algorithm for DOPF. Partially motivated by our offline algorithm, we first transform DOPF to a Linear Programming (LP) problem and then a packing problem. An online algorithm is then tailored based on the primal-dual method. Trace driven simulation studies verify the efficacy of the proposed offline and online algorithms.
Description
Keywords
Computer Science
Citation
Fu, H. (2014). Online Algorithms and Optimal Offline Algorithms for Dynamic Optimal Power Flow (Master's thesis, University of Calgary, Calgary, Canada). Retrieved from https://prism.ucalgary.ca. doi:10.11575/PRISM/27178