Chordal Conversion Based Convex Iteration Algorithm for Three-Phase Optimal Power Flow Problems

The three-phase optimal power flow (OPF) problem has recently attracted a lot of research interests due to the need to coordinate the operations of large-scale and heterogeneous distributed energy resources in unbalanced electric power distribution systems. The nonconvexity of the three-phase OPF pr...

Full description

Saved in:
Bibliographic Details
Published in:IEEE transactions on power systems Vol. 33; no. 2; pp. 1603 - 1613
Main Authors: Wang, Wei, Yu, Nanpeng
Format: Journal Article
Language:English
Published: IEEE 01.03.2018
Subjects:
Admittance
Chordal conversion
Convex functions
convex iteration
distribution system operator
Matrix decomposition
Partitioning algorithms
Reactive power
Relaxation methods
three-phase optimal power flow
ISSN:0885-8950, 1558-0679
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The three-phase optimal power flow (OPF) problem has recently attracted a lot of research interests due to the need to coordinate the operations of large-scale and heterogeneous distributed energy resources in unbalanced electric power distribution systems. The nonconvexity of the three-phase OPF problem is much stronger than that of the single-phase OPF problem. Instead of applying the semidefinite programming relaxation technique, this paper advocates a convex iteration algorithm to solve the nonconvex three-phase OPF problem. To make the convex iteration algorithm computationally efficient for large-scale distribution networks, the chordal conversion based technique is embedded in the convex iteration framework. By synergistically combining the convex iteration method and the chordal based conversion technique, the proposed three-phase OPF algorithm is not only computationally efficient but also guarantees global optimality when the trace of the regularization term becomes zero. At last, to further improve the computational performance, a greedy grid partitioning algorithm is proposed to decompose a single large matrix representing a distribution network to many smaller matrices. The simulation results using standard IEEE test feeders show that the proposed algorithm is computationally efficient, scalable, and yields global optimal solutions while resolving the rank conundrum.
ISSN:0885-8950
1558-0679
DOI:10.1109/TPWRS.2017.2735942