EURO 2024 Copenhagen
Abstract Submission

EURO-Online login

2074. Mixed-integer non-linear programming approach for identifying parameters inconsistencies in load flow calculations

Invited abstract in session WB-19: OR in Energy, stream OR in Energy.

Wednesday, 10:30-12:00
Room: 44 (building: 116)

Authors (first author is the speaker)

1. Martin Debouté
Artelys
2. Pierre Arvy
Artelys

Abstract

For several decades, electrical networks have experienced continuous expansion in size and scale, facilitating the transmission of electrical power from sources to end-users. The growing complexity of these networks necessitates the use of tools like Load Flow calculations, which are crucial for power systems engineering, playing a critical role in ensuring grid reliability and efficiency. This numerical analysis of electric power within interconnected systems determines steady-state voltages, currents, and power flows across network components. Given the significance of LF calculations, the power systems community has invested substantial efforts in enhancing the convergence of the methods used to solve them. Despite these efforts, real-world scenarios often involve imperfect or erroneous network parameters, leading to challenging convergence issues for existing methods. The physical problem studied may be mathematically infeasible due to complex inconsistencies in the parameters. This research project explores a novel approach centered around Mixed-Integer Non-Linear Programming to detect inconsistencies or errors in the parameters of a power network. The study extends with the implementation of an open-source tool using the AMPL modeling language, and conducts a solution to the problem using the commercial solver Artelys Knitro. Finally, the tool has been tested to ensure that it successfully identifies inconsistencies in a real-world use case within an acceptable time.

Keywords

Status: accepted

Back to the list of papers