MILP-Based Dispatch and Network Topology Optimization Through Bus Splitting: a Preliminary Analysis for Grid Congestion Management
Pee Jay N. Gealone, Adonis Emmanuel DC. Tio · 2025
The UN Sustainable Development Goals emphasize the necessity for sustainable, affordable, clean energy sources. This call has prompted many countries to set more ambitious objectives for transforming their energy mix to be primarily, if not entirely, composed of renewable energy sources. Integrating high levels of variable non-dispatchable renewable energy (VRE) with increased variable demand poses significant challenges in grid congestion management. The traditional approach to congestion management necessitates infrastructure investments to enhance line and substation capacities. Emerging technologies, such as grid-enhancing technologies and dynamic line rating, show promise but require time and investment. Because of this, there is growing interest in exploring topology reconfiguration, which utilizes existing line-switching and/or bus-splitting options as a tool for congestion management. This paper shows the preliminary analysis of the optimization model proposed that co-optimizes generation dispatch and network topology by incorporating the substation bus-splitting model into a mixed-integer linear programming problem. To achieve this, a DC power flow optimization model is developed to address congestion by (1) topology reconfiguration, (2) dispatch optimization, and (3) their combination. The paper used a simple IEEE 6-bus operated with 72-hour dispatch scenarios representing maximum VRE, average VRE, and minimum VRE output to demonstrate how the model works. These case studies demonstrate scenarios where congestion can be mitigated by reconfiguring the network through bus-splitting while minimizing VRE curtailment. The simulation indicated that by using bus-splitting in topology optimization, up to $50 \%$ of congestion scenarios can be resolved, and by co-optimizing it with optimal dispatch, all congestion scenarios can be addressed without load shedding. This work showed that treating static facilities such as the topology as dispatchable assets can be a powerful tool for congestion management and reducing VRE curtailment.