Skip to content
#

simbench

Here are 6 public repositories matching this topic...

Do topology-aware grid foundation models survive feeder reconfiguration? SimBench MV study: the topology encoding is what transfers (2-8% voltage error on unseen feeders, flat across switching states) — with zero-shot, DC-baseline, and scale gaps named plainly.

  • Updated Jul 21, 2026
  • Jupyter Notebook

Stackelberg demand response on a real feeder: recreation of Maharjan et al. (2013), AC-power-flow-audited equilibria, DLMP-lite repricing (+89% hosting capacity), model-free learning agents, and Shapley-audited fairness — with negative results logged.

  • Updated Jul 21, 2026
  • Jupyter Notebook

Generative models of distribution-feeder operating data with power-flow guarantees: physics by representation beats physics by penalty. Full-year SimBench AC-PF datasets, graph-conditioned flows, calibrated pseudo-measurement generation — eight evidenced findings incl. three negative results.

  • Updated Jul 18, 2026
  • Jupyter Notebook

Add this topic to your repo

To associate your repository with the simbench topic, visit your repo's landing page and select "manage topics."

Learn more