A comparison of Integer Linear Programming methods for AC power grid design

Abstract

Due to the rapid development of the number of heat pumps, charging stations, new companies and sustainability initiatives, the power transmission grid is reaching its current limits. To design a new grid, we must choose the optimal locations of the generators and power lines given the locations of power demand. To find a good solution for this computationally hard problem we approximate is using Integer Linear Programming. Thereto, we must linearize the non-linear AC power flow equations. In this thesis, we compare and experimentally test four different linearization methods. We conclude that of these four linearizations, the AC approximation method by Koster and Lemkens (2011) performs best.