GREMLIN (GReen Energy Monitoring for Large INfrastructure) is a non-intrusive, facility-scale
diagnostic layer that turns the electromagnetic-interference (EMI) signatures electrical devices
emit into actionable information — disaggregating where energy is used and flagging which
equipment is degrading — from a few high-bandwidth measurement points rather than per-device
instrumentation. It couples single-point sensing with non-intrusive load monitoring (NILM) and a
learned classifier benchmarked against established classical spectral methods, on the
GNU Radio 4.0 streaming runtime, with models exchanged in
the portable ONNX format.
This is the first public release — a documentation-first milestone that sets out the approach,
its physical basis, and the prior art, and provides a citable, archived snapshot of the project at
the start of its iRIS phase. It is an early (0.x) release: the concept is validated at small scale
today, and the EU Horizon Europe iRIS project funds maturing it toward higher technology
readiness at larger device counts under real operating conditions.
What's included
- Documentation site (https://fair-acc.github.io/gremlin/):
- How it works — EMI signatures, single-point disaggregation and NILM, the classical-baseline
benchmark, and ageing as an RF signature. - EMI signatures & compliance — conducted-versus-radiated propagation, the unavoidable-residual
argument, and an overview of conducted-emission limit envelopes (CISPR 11, FCC Part 15,
MIL-STD-461 CE102). - What it provides — energy footprint, availability and predictive maintenance, unaccounted-for
loads, and grid compliance. - References — NILM, EMI-ageing, and historical prior-art bibliography.
- How it works — EMI signatures, single-point disaggregation and NILM, the classical-baseline
- Citation and archival metadata —
CITATION.cffand.zenodo.json. - Open licensing and the EU Horizon Europe iRIS funding acknowledgement.
Status and scope
- Research → engineering. An engineered model is validated at small scale; quantified energy and
cost benefits are an iRIS deliverable set by detection performance — projected mechanisms,
not yet measured values. - Documentation-first. Signal-processing blocks, trained models, and reproducible demonstrations
build on the FAIR-ACC open stack and will arrive in subsequent releases and linked repositories.
Ecosystem
| Repository | Purpose |
|---|---|
| gremlin (this repo) | documentation, models, demonstrations |
| pulsed-power-ml | pulsed-power proof-of-concept and electrical-network-compliance work |
| gnuradio4 | GNU Radio 4 signal-processing runtime |
| gr-digitizers | DAQ / digitiser integration |
| opendigitizer | UI / UX |
Licensing
- Code:
LGPL-3.0-or-later WITH LGPL-3.0-linking-exception - Documentation:
CC-BY-SA-4.0
Copyright © GSI Helmholtz Centre for Heavy Ion Research and FAIR — Facility for Antiproton and Ion
Research in Europe, Darmstadt, Germany, and contributors.
Funding
This project has received funding from the European Union's Horizon Europe research and innovation
programme under grant agreement No. 101275935 (iRIS — Intelligent Research Infrastructure
Sustainability). Views and opinions expressed are however those of the author(s) only and do not
necessarily reflect those of the European Union or the granting authority. Neither the European
Union nor the granting authority can be held responsible for them.
How to cite
This release is archived on Zenodo under the EU Horizon Europe iRIS community; the version DOI is
minted automatically on publication and will be added to the DOI badge and CITATION.cff shortly
afterwards.
Authors: R. J. Steinhagen, A. Krimm, S. Lebedev, H. Welker, and the GREMLIN contributors (GSI/FAIR).
Full author and citation metadata are in CITATION.cff.
Documentation: https://fair-acc.github.io/gremlin/