EURO 2025 Leeds
Abstract Submission

2166. Solving non-linear optimal power flows with Artelys Knitro in PowSyBl open-source toolbox

Invited abstract in session MD-46: Planning and operation of power networks, stream Energy Economics & Management.

Monday, 14:30-16:00
Room: Newlyn 1.07

Authors (first author is the speaker)

1. Marvin Stanczak
Artelys
2. Hugo Chareyre
Artelys
3. Pierre Arvy
Artelys
4. Jeanne Archambault
Artelys
5. Hadrien Godard
Artelys
6. Nicolas Omont
Artelys

Abstract

Electrical networks have grown significantly, requiring advanced tools like Optimal Power Flow (OPF) to ensure reliability and efficiency. OPF problems are crucial for determining the most efficient operating conditions of power systems, minimizing costs or emissions while meeting system constraints. These problems are complex due to large-scale optimization, numerous constraints, and mixed variables. The power systems community has focused on enhancing AC modeling of OPF, which involves non-linear and non-convex constraints, making real-world scenarios challenging to solve.
PowSyBl is an open-source Java framework for grid modeling and simulation. It simplifies the development of complex software for power system analysis and includes a reusable code base for solving non-linear AC OPF problems. The framework uses the AMPL modeling language and the Artelys Knitro solver for optimization. It allows flexible configuration and the use of slack variables to facilitate convergence in power-imbalanced grids. The PowSyBl community continuously improves the framework, which has supported various research projects, including a divergence analyzer for detecting network parameter inconsistencies.
Future perspectives include enhancing and broadening the use of the Artelys Knitro solver within PowSyBl. Recent efforts aim to enable its use in additional processes that benefit from OPF solving. This aligns with Artelys' long-standing research and development on OPF problems.

Keywords

Status: accepted

Back to the list of papers