Facet-Defining Inequalities for the Angle-Based DC Optimal Transmission Switching Formulation

TL;DR

This paper introduces new inequalities for the angle-based DC-OTS formulation, improving the bounds on voltage angle differences and enhancing the efficiency of power system reconfiguration.

Contribution

It provides a novel polyhedral analysis and constructs an extended formulation with facet-defining inequalities for the DC-OTS problem.

Findings

Derived facet-defining inequalities tighten angle-difference bounds.

Extended formulation improves the convex hull representation.

Challenged the need for solving the intractable longest path problem.

Abstract

The switching of transmission lines can significantly improve the economic and operational efficiency of power systems. The Direct-Current Optimal Transmission Switching (DC-OTS) problem provides a formal framework for minimizing power generation costs by reconfiguring the transmission network topology under a linearized power flow model. DC-OTS is typically formulated as a mixed-integer linear program that incorporates disjunctive constraints to capture the required relationships between certain variables via big-M parameters. More specifically, these parameters represent upper bounds on voltage angle differences across non-operational transmission lines. In practice, overly conservative (and arbitrary) bounds tend to be used. The belief is that tightening these values requires the solution of the computationally intractable longest path problem. This work challenges that view through a novel polyhedral analysis of the angle-based DC-OTS formulation. We construct an extended formulation for the convex hull of an angle-based relaxation and derive facet-defining inequalities that tighten angle-difference bounds.