Voltage positioning using co-optimization of controllable grid assets in radial networks

TL;DR

This paper proposes a convex optimization framework for scheduling controllable grid assets in radial distribution networks to optimize voltage positioning amid high solar PV variability, balancing mechanical and inverter-based resources.

Contribution

It introduces a convex, mixed-integer linear programming approach for holistic, robust voltage control that accounts for forecasted PV and grid constraints.

Findings

Validated on IEEE feeders showing effective voltage regulation.

Minimized inverter usage for better resource availability.

Achieved robust voltage positioning despite uncertainties.

Abstract

With increasing penetration of solar PV, some distribution feeders are experiencing highly variable net-load flows and even reverse flows. To optimize distribution systems under such conditions, the scheduling of mechanical devices, such as OLTCs and capacitor banks, needs to take into account forecasted solar PV and actual grid conditions. However, these legacy switching assets are operated on a daily or hourly timescale, due to the wear and tear associated with mechanical switching, which makes them unsuitable for real-time control. Therefore, there is a natural timescale-separation between these slower mechanical assets and the responsive nature of inverter-based resources. In this paper, we present a network admissible convex formulation for holistically scheduling controllable grid assets to position voltage optimally against solarPV. An optimal hourly schedule is presented that utilizes mechanical resources to position the predicted voltages close to nominal values, while minimizing the use of inverter-based resources (i.e., DERs),making them available for control at a faster time-scale (after the uncertainty reveals itself). A convex, inner approximation of the OPF problem is adapted to a mixed-integer linear program that minimizes voltage deviations from nominal (i.e., maximizes voltage margins). The resulting OPF solution respects all the network constraints and is, hence, robust against modeling simplifications. Simulation based analysis on IEEE distribution feeders validates the approach.