Network-Constrained Robust Unit Commitment for Hybrid AC/DC Transmission Grids

L. P. Mohasha Isuru Sampath, Matthias Hotz, Hoay Beng Gooi, Wolfgang UtschickarXiv 2018 · DOI · arXiv

Abstract

The day-ahead energy and reserve management with transmission restrictions and voltage security limits is a challenging task for large-scale power systems in the presence of real-time variations caused by the uncertain demand and the fluctuating power output of renewable energy sources (RESs). The proposed formulation in this work supports joint scheduling of energy and reserve to promote an economic and reliable operation. To improve its scalability, a two-stage iterative optimization algorithm is proposed based on Bender's decomposition framework. Therewith, the optimal schedule is computed subject to the feasibility of AC network constraints (AC-NCs) at predetermined uncertain realizations. A convex relaxation is applied to AC-NCs to support the convergence of the algorithm. Moreover, the integration of RESs is often limited by transmission congestion issues in existing grids. For such grids, AC to DC conversion schemes can be viable and attractive options for capacity expansion. In this study, we adopt a hybrid AC/DC transmission grid (HTG) architecture with a specific topology to improve the effective utilization of the grid capacity and accommodate more demand and RESs, which is showcased via simulations on a large-scale network. Moreover, HTG induces exactness of the convex relaxation of AC-NCs, supporting the validity of the optimal schedule.

Introduction

Modern power grid management requires a complex balancing of Unit Commitment (UC) and reserve capacity allocation. System operators must make day-ahead decisions while accounting for the inherent uncertainty of renewable energy sources (RESs) like wind and solar. Traditional UC models frequently employ "DC power flow" approximations, which neglect critical factors such as voltage limits and reactive power. This simplification can lead to transmission congestion or voltage violations when schedules are implemented in real-world AC networks.

Furthermore, existing AC grids are reaching their physical limits under high renewable penetration. While constructing new transmission lines is often prohibitively slow and expensive, converting existing AC lines to DC offers a promising alternative. However, optimizing these Hybrid AC/DC Transmission Grids (HTGs) introduces significant mathematical complexity.

Methodology

This research proposes a scalable, robust optimization framework for joint energy and reserve scheduling. The core of the approach is a two-stage iterative optimization algorithm utilizing Benders' Decomposition Framework (BDF).

  1. Master Problem (MP): Determines the 24-hour commitment of units and reserve allocations, considering inter-temporal constraints such as generator ramp rates and minimum up/down times.
  2. Subproblems (SPs): Function as feasibility checkers, verifying if the AC network can accommodate power delivery under specific uncertain scenarios (e.g., sudden drops in wind generation) without violating thermal or voltage limits.

To ensure computational tractability, the authors utilize Taguchi’s Orthogonal Array Testing (TOAT) to select a representative subset of worst-case scenarios. Additionally, Second-Order Cone (SOC) relaxation is applied to the non-convex AC network equations. A key finding is that the specific topology of an HTG—comprising radial AC subgrids linked by DC lines—promotes the "exactness" of this SOC relaxation, ensuring that the mathematical model accurately reflects physical grid behavior.

Results and Discussion

The framework was tested on a large-scale benchmark: the Polish Transmission Grid (2383-bus) during winter peak conditions, integrated with 15 wind farms.

Economic Efficiency and Loadability

The hybrid architecture demonstrated significant advantages over traditional AC grids. The HTG was able to accommodate up to 14.5% more demand. As shown in Table I, the HTG consistently achieved lower total operating costs across various load scales compared to the AC grid (ACG).

Table I: Comparison of Economic Efficiency and Loadability

Convergence Performance

The HTG architecture also improved algorithmic performance. The Benders' decomposition converged significantly faster for the HTG (8 iterations) than for the ACG (over 30 iterations). This is attributed to the HTG's increased flexibility, which results in fewer constraint violations during the iterative search process.

Figure 1: Convergence trajectories for (a) AC Grid and (b) Hybrid Grid

Conclusion

The proposed robust unit commitment framework for HTGs offers a mathematically sound and computationally efficient solution for modern grid management. By leveraging the physical properties of hybrid AC/DC architectures, the model ensures the exactness of convex relaxations and improves both economic efficiency and renewable integration. Future work will focus on addressing potential "pathological" power injection profiles and exploring adaptive robust optimization to further refine the balance between operational cost and system risk.

Summary updated

Abstract

The day-ahead energy and reserve management with transmission restrictions and voltage security limits is a challenging task for large-scale power systems in the presence of real-time variations caused by the uncertain demand and the fluctuating power output of renewable energy sources (RESs). The proposed formulation in this work supports joint scheduling of energy and reserve to promote an economic and reliable operation. To improve its scalability, a two-stage iterative optimization algorithm is proposed based on Bender's decomposition framework. Therewith, the optimal schedule is computed subject to the feasibility of AC network constraints (AC-NCs) at predetermined uncertain realizations. A convex relaxation is applied to AC-NCs to support the convergence of the algorithm. Moreover, the integration of RESs is often limited by transmission congestion issues in existing grids. For such grids, AC to DC conversion schemes can be viable and attractive options for capacity expansion. In this study, we adopt a hybrid AC/DC transmission grid (HTG) architecture with a specific topology to improve the effective utilization of the grid capacity and accommodate more demand and RESs, which is showcased via simulations on a large-scale network. Moreover, HTG induces exactness of the convex relaxation of AC-NCs, supporting the validity of the optimal schedule.
BibTeX
@misc{sampath2018network,
  author = {L. P. Mohasha Isuru Sampath and Matthias Hotz and Hoay Beng Gooi and Wolfgang Utschick},
  title = {{Network-Constrained Robust Unit Commitment for Hybrid AC/DC Transmission Grids}},
  year = {2018},
  eprint = {1811.08754},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/1811.08754},
}
Figure