- The paper introduces nine AC power flow test cases formatted for MATPOWER and QCQP, significantly advancing optimal power flow research in large-scale grids.
- The methodology converts real grid data from iTesla, RTE snapshots, and PEGASE projects into large-scale sparse QCQP problems, enabling interdisciplinary applications.
- The study demonstrates computational feasibility with Knitro-driven OPF solutions while exploring global optimality and optimality gaps using DCOPF and SDP methods.
This paper presents the publication of nine new AC power flow test cases, formatted for MATPOWER and QCQP applications, aimed at improving optimal power flow (OPF) methods in the context of power systems research. The authors, with affiliations to the French transmission system operator RTE, focus on advancing grid data accessibility for both power systems and applied mathematics communities. This work includes data from iTesla, RTE snapshots, and PEGASE projects, capturing French high-voltage grids as well as a large-scale pan-European dataset.
Data Origin and Characteristics
The paper details three primary data sources:
- iTesla Project: As a part of a large EC-funded initiative, this data captures French very high-voltage (VHV) and high-voltage (HV) grid configurations. The iTesla project's dataset was generated using high-performance computing to evaluate grid stability with varying power loads and generation scenarios.
- RTE Snapshots: These snapshots offer an operational perspective, derived from real-time French national and regional SCADAs and Convergence software. They encompass more than 6000 nodes, reflecting an extensive view of France's VHV and HV grid structure.
- PEGASE Project: Encompassing a synthetic pan-European grid model, this dataset is invaluable for validating OPF methods. While fictitious, it mirrors the complexity and scale of European transmission networks, thus complementing previously published PEGASE test cases.
Data Processing and Conversion
The publication includes procedures to convert these datasets into the QCQP format, enabling applied mathematicians to address power flow problems without specific knowledge of power systems. This conversion process involves transforming grid data into a set of large-scale sparse QCQP problems, maintaining the non-convex nature of the OPF challenge.
Numerical Results and Feasibility
Through computational experiments, the paper verifies the validity of test data and provides initial insights into feasible OPF solutions. Using Knitro as the interior point solver, the authors achieve feasible solutions for OPF problems despite their inherent non-convexity. Noteworthy is the exploration of optimality gaps through the combination of DCOPF and semidefinite programming (SDP) methods, which underline SDP's potential for better lower bounds in OPF problems.
Discussion on Global Optimality
The study emphasizes the necessity for achieving global optimality in OPF solutions. The authors explore ways to improve both feasible solutions and lower bounds, thereby enhancing the overall accuracy of OPF solutions. They suggest that advancements in solving continuous OPF to global optimality could pave the way for tackling discrete OPF problems and bilevel programming, comprising a broader optimization framework for grid operations.
The paper's contributions lie not only in the dissemination of comprehensive data sets but also in fostering interdisciplinary collaboration by providing data in both MATPOWER and QCQP formats. The provision of MATLAB code to facilitate conversion further aids researchers in leveraging this data. Looking forward, the authors anticipate future iterations of their work with refined computational techniques and augmented datasets, supporting ongoing research endeavors in the optimization and stabilization of power systems.
