Impact of Training Dataset Size for ML Load Flow Surrogates

Conrad T, Kim C, Jäger J, Maier A, Bayer S (2025)

Publication Language: English

Publication Type: Conference contribution, Conference Contribution

Publication year: 2025

Event location: Zittau

Abstract

Efficient and accurate load flow calculations are a bedrock of modern power system operation. Classical
numerical methods such as the Newton–Raphson algorithm provide highly precise results but are
computationally demanding, which limits their applicability in large-scale scenario studies and optimization
with time-critical contexts. Research has shown that machine learning approaches can approximate
load flow results with high accuracy while substantially reducing computation time. Sample efficiency,
i.e. their ability to achieve high accuracy with limited training dataset size, is still insufficiently researched,
especially in grids with a fixed topology. This paper considers a systematic investigation of the sample
efficiency of a Multilayer Perceptron and two Graph Neural Networks variants on a dataset based on a
modified IEEE 5-bus system. The results for this grid size show that Graph Neural Networks achieve
the lowest losses. However, the availability of large training datasets remains the dominant factor for
performance over architecture.

Authors with CRIS profile

How to cite

APA:

Conrad, T., Kim, C., Jäger, J., Maier, A., & Bayer, S. (2025). Impact of Training Dataset Size for ML Load Flow Surrogates. In Proceedings of the Oberlausitzer Energiesymposium 2025 & Zittauer Energieseminar. Zittau, DE.

MLA:

Conrad, Timon, et al. "Impact of Training Dataset Size for ML Load Flow Surrogates." Proceedings of the Oberlausitzer Energiesymposium 2025 & Zittauer Energieseminar, Zittau 2025.

BibTeX: Download