[Paper Review] Joint Grid Topology Reconfiguration and Design of Watt-VAR Curves for DERs

Manish K. Singh, Sina Taheri+3 more|arXiv (Cornell University)|Oct 7, 2019

Optimal Power Flow Distribution24 references4 citations

TL;DR

This paper proposes a mixed-integer linear programming (MILP) framework that jointly optimizes distribution grid topology reconfiguration, voltage regulator tap settings, and DER watt-var control curves under IEEE 1547.8 standards. It introduces a provably correct radiality constraint and models legacy regulators and DERs realistically, showing that optimal topologies and control settings vary dynamically with load-generation profiles, especially during steep net load transitions, to minimize losses and maintain voltage limits.

ABSTRACT

Operators can now remotely control switches and update the control settings for voltage regulators and distributed energy resources (DERs), thus unleashing the network reconfiguration opportunities to improve efficiency. Aligned to this direction, this work puts forth a comprehensive toolbox of optimization models leveraging the control capabilities of smart grid assets. We put forth detailed yet practical models to capture the operation of locally and remotely controlled regulators, and customize the watt-var DER control curves complying with the IEEE 1547.8 mandates. Maintaining radiality is a key requirement germane to various feeder optimization tasks. This requirement is accomplished here through an intuitive and provably correct formulation. The developed toolbox is put into action to reconfigure a grid for minimizing losses using real-world data on a benchmark feeder. The results corroborate that optimal topologies vary across the day and coordinating DERs and regulators is critical during periods of steep net load changes.

Motivation & Objective

• To develop a comprehensive optimization framework that jointly reconfigures grid topology and designs DER watt-var curves while accounting for legacy voltage regulators and radiality constraints.
• To address the gap in existing DNR models that either ignore regulators or assume fixed tap settings, by modeling their operation dynamically.
• To ensure radiality in the reconfigured network through a novel, intuitive, and provably correct MILP formulation that decouples radiality from power flow variables.
• To demonstrate the importance of coordinating DERs and regulators during periods of high net load volatility, using real-world load and solar data on the IEEE 37-bus system.
• To evaluate trade-offs between operating period length, scenario sampling, and system performance in terms of losses, voltage violations, and control action frequency.

Proposed method

• Formulates a mixed-integer quadratic program (MIQP) for distribution network reconfiguration (DNR) that jointly optimizes switch states, regulator tap positions, and DER watt-var curve parameters.
• Introduces a novel MILP model to enforce radiality by assigning a single parent to each bus (except the slack bus), ensuring connectedness and tree structure without loop elimination.
• Develops a detailed model of locally controlled voltage regulators that captures tap position dynamics and voltage regulation behavior under varying load and generation conditions.
• Incorporates IEEE 1547.8-compliant watt-var curves for DERs, allowing reactive power injection/absorption based on active power output, with piecewise-linear approximation for optimization.
• Uses a linearized three-phase power flow model (AC power flow approximation) to ensure computational scalability while maintaining accuracy for balanced distribution feeders.
• Employs scenario-based optimization over multiple time periods, with each period representing distinct load and solar generation levels, to capture time-varying system conditions.

Experimental results

Research questions

• RQ1 How can grid topology reconfiguration be optimally coordinated with DER watt-var control curves and voltage regulator tap settings to minimize losses while maintaining radiality?
• RQ2 What is the impact of modeling legacy voltage regulators dynamically (with variable taps) versus assuming fixed or known taps on the optimality and feasibility of DNR solutions?
• RQ3 How do optimal topologies and control settings change across different load-generation profiles throughout the day, particularly during periods of rapid net load changes?
• RQ4 What trade-offs exist between the length of operating periods and system performance in terms of total losses, voltage violations, and control action frequency?
• RQ5 To what extent does the proposed radiality constraint formulation outperform conventional methods in terms of correctness and computational efficiency, especially in the presence of DERs?

Key findings

• The optimal feeder topology varies across different time periods due to changing load and solar generation profiles, with three distinct topologies identified as optimal across five time intervals.
• During periods of high PV generation and low load (e.g., period 𝒯₂), DERs actively absorb reactive power only when overvoltages are imminent, as shown by watt-var curves that activate absorption only beyond 80% of active power capacity.
• In periods with high load and high generation (e.g., 𝒯₃), minimal reactive power absorption is required, resulting in watt-var curves that remain near the minimum absorption point.
• During steep net load decline (e.g., 𝒯₄), a high regulator tap setting (21) is required to prevent undervoltage, but this necessitates reactive power absorption from DERs to avoid overvoltage during peak PV output.
• Shorter operating periods increase control action frequency and communication overhead, while longer periods risk infeasibility or increased losses due to extreme load-generation variations, especially when merging high-volatility intervals.
• The solution time for each time period ranged from 50 to 920 seconds, with the most complex instance (𝒯₄) taking 800 seconds, indicating computational feasibility for real-time or economic dispatch applications.