Industrial NILM datasets · v2.4

Your NILM models are starving for data.
Public datasets aren't enough anymore.

Real multi-site, multi-sector industrial energy datasets. Machine-level ground truth, high sampling rate, three-phase measurements. The scientific baseline for load disaggregation and predictive maintenance.

sample_rate 1 Hz
phase 3φ · L1L2L3
meters 42 · gap 0.02s
uptime 99.98%
load_dataset.py
python 3.11 · nilmtk 0.4
# Load a dataset in three lines — NILMTK-compatible out of the box.
from nilmtk import DataSet

ds = DataSet("watt_industrial_v2.h5")
ds.set_window(start="2025-01-01")
elec = ds.buildings[3].elec

# > MeterGroup(meters=42, sample_period=0.02s, sectors=['mechanical'])

// The gap

Industrial NILM data is scarce, short, and mono-context.

HIPE, IMDELD, SIDED — you can count public industrial datasets on one hand. Three months of data, a single site, a single sector. Models trained on them don't generalize. Collecting proprietary data costs months and tens of thousands of euros in hardware, installation and labeling.

≤ 5
Public industrial NILM datasets available
~3 mo
Typical temporal coverage
1 site
Almost all public datasets are single-site
€ 40k+
Cost to bootstrap a proprietary collection

// The dataset

Engineered for load disaggregation research.

Every spec is measurable. No marketing padding.

1 Hz · 3φ
High-rate three-phase acquisition

V, I, P, Q, S, harmonics, THD, power factor per phase.

1 meter / machine
Machine-level ground truth

Submetering per single machine, not per circuit.

12 sites · 5 sectors
Multi-site, multi-sector

Mechanical, food & beverage, plastics, packaging, HVAC.

≥ 12 months
Long temporal coverage

Seasonal variation and full production cycles captured.

> 2M labeled events
Labeled events

On/off transitions, state changes, anomalies, standby.

HDF5 · CSV · YAML
NILMTK-compatible

HDF5 + CSV. Documented metadata schema. Ready for your pipeline.

GDPR · commercial OK
GDPR-compliant

Anonymized. Clear license for commercial model training.

40+ households
Domestic set included

Residential appliance-level data to complement industrial.

// Comparison

Us vs. public datasets

Feature parity with HIPE, IMDELD, SIDED — plus everything they lack.

OursHIPEIMDELDSIDED
Duration≥ 12 months3 months6 months4 months
Sites12111
Machines180+10811
Sectors5111
Sampling rate1 Hz5 kHz1 Hz1 Hz
Machine-level labelsYesPartialYesPartial
Event labelingYes, > 2M eventsNoLimitedNo
Commercial licenseYesNoNoNo
Technical supportIncludedNoneNoneNone

// Use cases

Built for the workloads NILM teams actually run.

01
Disaggregation model training

Train FHMM, seq2seq, CNN, transformer architectures on realistic multi-machine mixtures.

02
Algorithm benchmarking

Reproducible baselines with known ground truth. Compare against SOTA on the same conditions.

03
Predictive maintenance

Electrical signatures across full production cycles for anomaly detection and failure precursors.

04
Transfer learning

Cross-site and cross-sector data for domain adaptation to unseen machinery families.

// Methodology

Every measurement is documented, reviewed, and traceable.

We ship a datasheet with each dataset. QA is not optional.

Instrumentation

Class 0.5 revenue-grade meters, PT/CT chain calibrated, timestamped via PTP/NTP with drift monitoring.

Labeling protocol

Dual-review labeling with operator interviews. Event schema versioned in Git.

Quality assurance

Automated coverage checks, gap analysis, energy-balance reconciliation, anomaly flagging.

Custom collections

On-demand acquisition on your machinery, your protocol, your license terms.

// Pricing

Three tiers.

Sample and academic tiers are self-serve. Commercial is a real conversation.

Sample
€0
free

For evaluation. Enough data to validate your pipeline before you commit.

  • 1 site · 5 machines · 7 days
  • HDF5 + CSV export
  • NILMTK loader
  • Community support
Download sample
Recommended
Research
€10,000
per dataset, academic license

For universities and public research labs. Full dataset access, non-commercial.

  • Full multi-site coverage
  • Machine-level ground truth
  • Event labels + metadata
  • Email support · 48h SLA
  • Citation guidelines
Request research license
Commercial
On request
commercial license + custom collections

For product teams shipping NILM in production. Includes custom acquisition on request.

  • Everything in Research
  • Commercial training license
  • Custom sites & sectors
  • Dedicated technical contact
  • Update stream (quarterly)
Talk to the team

// Get in touch

Request a sample or a technical call.

We reply within one business day. No sales funnel, you talk to engineers.

// call · 30 min
Book a technical call

30 min with an engineer. Pick a slot that works for you.

Book a technical call
Or send us the details