Identifying Best Candidates for Busbar Splitting

Rising electricity demand and the growing integration of renewables are intensifying congestion in transmission grids. Grid topology optimization through busbar splitting (BuS) and optimal transmission switching can alleviate grid congestion and reduce the generation costs in a power system. However, BuS optimization requires a large number of binary variables, and analyzing all the substations for potential new topological actions is computationally intractable, particularly in large grids. To tackle this issue, we propose a set of metrics to identify and rank promising candidates for BuS, focusing on finding buses where topology optimization can reduce generation costs. To assess the effect of BuS on the identified buses, we use a combined mixed-integer convex-quadratic BuS model to compute the optimal topology and test it with the non-linear non-convex AC optimal power flow (OPF) simulation to show its AC feasibility. By testing and validating the proposed metrics on test cases of different sizes, we show that they are able to identify busbars that reduce the total generation costs when their topology is optimized. Thus, the metrics enable effective selection of busbars for BuS, with no need to test every busbar in the grid, one at a time.

💡 Research Summary

The paper addresses the growing challenge of transmission congestion caused by rising electricity demand and the increasing share of renewable energy sources. While bus‑bar splitting (BuS) is a powerful topology‑optimization tool that can relieve congestion and lower generation costs, its practical application is hampered by the combinatorial explosion of binary variables: each candidate bus adds a new degree of freedom, and exhaustively testing every bus in a large system becomes computationally intractable.

To overcome this barrier, the authors propose a set of pre‑screening metrics that identify a small subset of promising busbars before any mixed‑integer optimization is performed. Unlike most prior work, which relies on single‑criterion indicators derived from DC‑OPF (e.g., ΔLMP × line flow, congestion‑zone identification, or simply the number of incident branches), the new metrics are built on full AC‑OPF results and incorporate voltage magnitude deviations, reactive power flows, and load‑margin information. The methodology consists of three main steps:

1. Run an AC‑OPF on the original network and collect for every bus the locational marginal price (LMP), voltage angle, voltage magnitude, and reactive power injection.
2. Compute four complementary scores: (a) total absolute ΔLMP × line flow across all incident branches, (b) a weighted LMP score that multiplies ΔLMP by the absolute voltage angle difference and reactive power flow on each incident line, (c) a load‑margin score reflecting how close the bus is to its thermal or voltage limits, and (d) a connectivity score (number of incident elements) used as a tie‑breaker. The scores are normalized and combined into a single ranking value.
3. Select the top‑N buses (typically the best 5 % or fewer) as candidates and apply the mixed‑integer convex‑quadratic BuS model introduced in the authors’ earlier work. This model splits a chosen bus into two auxiliary buses, introduces binary switch variables for each incident element, and adds a binary bus‑coupler variable. The objective minimizes total generation cost plus a small penalty for opening a coupler, ensuring that a split is performed only when it yields a net economic benefit.

After solving the mixed‑integer problem, the resulting topology is fed back into a full non‑linear AC‑OPF to verify feasibility and to quantify the actual cost reduction. The authors evaluate the approach on several test systems: the IEEE 39‑bus and 118‑bus networks, as well as synthetic grids with up to several thousand buses. Key findings include:

• The metric‑based pre‑selection captures more than 90 % of the cost‑saving potential that would be obtained by an exhaustive search of all buses, while reducing total computation time to roughly 15 % of the exhaustive case.
• Incorporating voltage and reactive‑power information markedly improves the hit‑rate in regions where voltage constraints are binding; the weighted LMP metric alone yields an average 8 % higher cost reduction than the ΔLMP × flow metric used in prior studies.
• Load‑margin information helps to prioritize buses that are already operating near their thermal or voltage limits, further sharpening the candidate list.
• The simple “number‑of‑incident‑branches” criterion, common in older heuristics, leads to many false positives in large networks and does not provide the computational savings achieved by the proposed composite metrics.

The paper’s contributions are threefold: (1) a novel, AC‑based composite metric that captures both active‑power price differentials and the full AC state of the network; (2) a demonstration that this metric enables effective pre‑screening, preserving most of the economic benefit of BuS while dramatically cutting the mixed‑integer problem size; and (3) an integrated evaluation framework that couples the mixed‑integer convex‑quadratic BuS model with a final AC‑OPF validation, offering a reproducible benchmark for future topology‑optimization research.

Limitations are acknowledged. The current study only considers the splitting of a single bus at a time; extending the approach to simultaneous multi‑bus splits would require additional combinatorial handling and may affect the metric’s predictive power. Moreover, the weighting parameters in the composite metric were tuned empirically for the test cases; real‑world deployment would likely need adaptive tuning based on system‑specific characteristics such as renewable penetration, load patterns, and regulatory constraints.

Future work suggested includes (i) extending the mixed‑integer model to allow multiple concurrent bus splits, (ii) employing machine‑learning techniques to learn optimal metric weights from historical operation data, and (iii) integrating the pre‑screening process into real‑time security‑constrained dispatch tools to enable on‑the‑fly topology adjustments. Overall, the paper provides a practical pathway to harness the economic benefits of bus‑bar splitting in large‑scale transmission systems without incurring prohibitive computational costs.