Siming Bayer
—
Siming Bayer: machine learning for power grid protection — offline RL relaying and fault-detection benchmarks
2
papers here
2
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
—
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-14) and cover the author's entire public record, not just the papers listed here. Reproducibility signal comes from Evolve's own reviews.
Research profile
Output—
Output metrics such as total publication count, venue types, and annual productivity are not yet available; the radar lacks sufficient bibliometric coverage to score research output beyond the two cataloged 2026 preprints.
Impact—
Impact measures — citations, h-index, and uptake by later work — cannot yet be computed from the available snapshot; this axis would capture how widely the two 2026 preprints are cited and adopted by the grid-protection community.
Focus0/100
Focus scores 0/100 because the cataloged output splits evenly — 50% in power system protection (machine learning / offline RL) and 50% in electrical power systems / applied machine learning — across two overlapping but distinct field labels. In practice both papers address machine learning for grid fault detection and protection, so the low score reflects labeling breadth rather than scattered interests; the topical core remains protective relaying with learned models.
Collaboration51/100
The collaboration dimension scores 51/100, derived from 12 distinct coauthors recorded across the two cataloged preprints. Recurring names include Julian Oelhaf, Christian Bergler, and Andreas Maier, with additional collaborators Alexander Luce, Georg Kordowich, Paula Andrea Pérez-Toro, Tomás Arias-Vergara, and Johann Jäger. Bayer appears as last author on both papers without corresponding-author status, indicating a senior-contributor role in moderate-sized rather than sprawling teams.
Activity—
Activity — publication cadence and sustained engagement over time — cannot be scored, since only two same-year preprints are recorded; this axis would track annual output rhythms and ongoing research participation across multiple years.
Reproducibility60/100
Reproducibility scores 60/100, the mean across two assessed papers. The offline RL protection study is rated high for reproducibility readiness, while the systematic fault-detection benchmark is rated low, owing to a single-topology PowerFactory setup, missing deep-learning baselines, and possible cross-validation leakage. This internal split signals that code and benchmark transparency varies considerably across the cataloged work.
Research summary
Siming Bayer is an applied machine learning researcher working at the intersection of power engineering and data-driven grid protection; this profile is name-resolved from public bibliographic data. The Evolve catalog holds two arXiv preprints from 2026, both with Bayer as last author. The first applies offline reinforcement learning — conservative Q-learning — to distribution-grid protection, learning line-selective tripping on a CIGRE benchmark network with F1 0.9738, while candidly reporting poor non-fault security and single-seed training; Evolve rates its reproducibility readiness high. The second systematically benchmarks machine learning methods for fault detection and fault line identification, reporting F1 0.991 within 10 ms windows, though in a single-topology PowerFactory setting without deep-learning baselines and with possible cross-validation leakage; its reproducibility readiness is rated low. Radar scores put collaboration at 51 and reproducibility at 60 on a 0–100 scale, with 12 distinct coauthors, while output, impact, and activity metrics are not yet available. The picture is an early-stage, simulation-centered profile focused on machine learning for protective relaying.
Collaboration pattern
Across the two cataloged 2026 preprints, Bayer worked with 12 distinct coauthors and sat in the last-author position on both, without corresponding authorship. Recurring collaborators include Julian Oelhaf, Christian Bergler, and Andreas Maier, alongside Alexander Luce, Georg Kordowich, Paula Andrea Pérez-Toro, Tomás Arias-Vergara, and Johann Jäger. The pattern suggests a stable, moderately sized team around applied machine learning for grid protection rather than a broad, loosely affiliated network.
Research keywords
Highlights
Offline RL learns line-selective tripping for distribution grids (F1 0.9738)
Systematic ML benchmark for grid fault detection and line identification
Timeline
- 2026First cataloged arXiv preprint: offline reinforcement learning (conservative Q-learning) for distribution-grid protection, with Bayer as last author
- 2026Systematic machine learning evaluation of fault detection and line identification in electrical power grids published to arXiv, again as last author
- 2026Two preprints now sit in the Evolve catalog, both on machine learning for power grid protection, spanning 12 distinct coauthors
Papers in the Evolve catalog
Frequent co-authors
About this profile
This profile is AI-structured by Evolve.Pub from public bibliographic data (OpenAlex) and Evolve's own paper reviews. It is not written or endorsed by the researcher unless the profile is claimed. Radar dimensions are computed from real bibliometrics with fixed formulas — raw figures are shown alongside; missing dimensions mean insufficient data, never zero. Errors? Claim the profile to correct or hide it.
- Bibliometric radar scores are computed from a public bibliometric snapshot only; no institution, country, or ORCID is provided, so this profile is name-resolved from public bibliographic data.
- Only two arXiv preprints (both 2026) form the cataloged evidence base; peer-review status, venues, and citation counts are unknown.
- Output, impact, and activity axes lack sufficient data and are reported as missing rather than estimated.
- The frequent-coauthor list contains a duplicate entry for Siming Bayer, indicating some name-disambiguation noise in the underlying data.