AFCI offline kit - Part 7: Core arc-fault & appliance datasets
Offline development kit for an arc-fault detection / electrical-fire prevention system (current sensor -> STM32F407 ADC+DMA -> FFT/STFT features -> quantized DNN -> trip). Target offline PC: Windows 10/11 x64, CPU-only Python training.
Contents: core arc-fault & appliance datasets
Each file keeps its original publisher's license. download_log_stage7.json lists the exact source URL for every file.
| Folder | Dataset | Source | License | Use |
|---|---|---|---|---|
01_Lorraine_Ferracuti2025_preprocessed_windows/ |
Ferracuti et al. 2025, Real-time series arc fault detection and appliances classification in AC networks based on competing convolutional kernels: 230 V / 50 Hz series-arc current data (Univ. Lorraine test bench) already cut into 2-, 5- and 14-class windows at 10–100 kHz, plus MATLAB/Python code and results (6,969 files). | Zenodo 15279701 | CC-BY-4.0 | Best starting point for a 1D-CNN: arc vs normal + appliance class. The raw recordings are in Parts 8–10. |
02_TI_AC_arc_fault/ |
TI AC arc-fault datasets: ac_arc_fault.zip (current waveforms) and ac_arc_fault_log300.zip (LOG300 1–10 MHz envelope, ~100 kHz equivalent). Used by TIDA-010971 / MSPM0 ac_arc_fault_* examples (Part 4). |
software-dl.ti.com/C2000/esd/mcu_ai/datasets/ | "TI Internal License" label (publicly downloadable) | Arc/normal classification; HF-energy feature idea for your analog front end. |
03_TI_DC_arc_fault/ |
TI DC arc-fault datasets dsk/dsi (classification and anomaly-detection versions). | same | same | Autoencoder / anomaly-detection experiments (tinyml-tensorlab examples). |
04_DCADE_AC_arc_400Hz/ |
DCADE project: AC arc-fault signals on a terminal block, 115 V / 400 Hz (HIOKI and measurement-card CSVs + ReadMe). | Zenodo 10246364 | CC-BY-4.0 | Extra AC arc signatures (aircraft-type 400 Hz grid). |
05_Appliance_negatives_PLAID/ |
PLAID appliance current/voltage: TI's sub-metered NILM version (zip) + UEA/UCR time-series version (PLAID*.ts, sktime format). |
software-dl.ti.com; Zenodo 11197936 | TI label; CC-BY-4.0 | "Normal load" / nuisance-trip negatives; appliance recognition. |
06_DC_series_arc_IEEE_cleaned_Microchip/ |
Microchip's cleaned copy of the IEEE DataPort Low-voltage DC series arc dataset (doi:10.21227/kzr1-1w26). | github.com/MicrochipTech/ml-arc-fault-detection | no license file in repo | DC arc reference. |
07_PV_arc_self_extinction_small/ |
PV arc self-extinction modelling/validation CSVs + notebook (IEC 63027 electrode). | Mendeley Data fj5nb2vpn7 | CC-BY-4.0 | Arc physics / modelling reference (small). |
08_Synthetic_AC_arc_CC0/ |
Simulated AC arc with nonlinear loads. | github.com/mgiga23/dataset_arc_fault_nonlinear_AC | CC0-1.0 | Quick pipeline tests. |
Not included (links only, see Part 6 index): HIFDA household-appliance HF signals (Zenodo 14886758, 4.4 GB), PV DC arc datasets (Zenodo 14804124, 2.2 GB; Mendeley pd93m3c4rz, 3.8 GB) — off-target for residential AC; COOLL, LIT-Dataset, IEEE DataPort originals — need registration.
Tip: split train/test by measurement session or device, not randomly, to get honest generalization numbers.
Rebuilding split files
Files over 100 MB were split into *.part001, *.part002, ... pieces. After downloading the whole repo:
- Windows:
powershell -ExecutionPolicy Bypass -File .\join_parts.ps1 - Linux/macOS:
bash join_parts.sh
Both scripts rebuild the files and check them against SHA256SUMS.txt.
Files (49 files, 1.2 GB total)
| File | Size | Split into |
|---|---|---|
download_log_stage7.json |
11.9 KB | - |
01_Lorraine_Ferracuti2025_preprocessed_windows/Ferracuti2025_realtime_series_arc_fault_detection_appliance_classification_competing_conv_kernels.zip |
429.2 MB | 5 |
02_TI_AC_arc_fault/ac_arc_fault.zip |
176.9 MB | 2 |
02_TI_AC_arc_fault/ac_arc_fault_log300.zip |
66.7 MB | - |
03_TI_DC_arc_fault/arc_fault_classification_dsi.zip |
12.7 MB | - |
03_TI_DC_arc_fault/arc_fault_classification_dsk.zip |
7.6 MB | - |
03_TI_DC_arc_fault/arc_fault_dsi_anomaly_detection.zip |
12.7 MB | - |
03_TI_DC_arc_fault/arc_fault_dsk_anomaly_detection.zip |
7.6 MB | - |
04_DCADE_AC_arc_400Hz/HIOKI_DCADE_Test1.csv |
37.5 MB | - |
04_DCADE_AC_arc_400Hz/HIOKI_DCADE_Test2.csv |
37.5 MB | - |
04_DCADE_AC_arc_400Hz/HIOKI_DCADE_Test3.csv |
96.4 MB | - |
04_DCADE_AC_arc_400Hz/HIOKI_DCADE_Test4.csv |
96.0 MB | - |
04_DCADE_AC_arc_400Hz/MCmeasurement-DCADE_Test3.csv |
39.7 MB | - |
04_DCADE_AC_arc_400Hz/MCmeasurement_DCADE_Test1.csv |
17.6 MB | - |
04_DCADE_AC_arc_400Hz/MCmeasurement_DCADE_Test2.csv |
18.1 MB | - |
04_DCADE_AC_arc_400Hz/MCmeasurement_DCADE_Test4.csv |
40.2 MB | - |
04_DCADE_AC_arc_400Hz/ReadMe.pdf |
708.7 KB | - |
05_Appliance_negatives_PLAID/PLAID_TEST.ts |
1.4 MB | - |
05_Appliance_negatives_PLAID/PLAID_TRAIN.ts |
1.4 MB | - |
05_Appliance_negatives_PLAID/PLAID_eq_TEST.ts |
11.8 MB | - |
05_Appliance_negatives_PLAID/PLAID_eq_TRAIN.ts |
11.9 MB | - |
05_Appliance_negatives_PLAID/plaid_nilm_submetered_dataset.zip |
23.1 MB | - |
06_DC_series_arc_IEEE_cleaned_Microchip/ieee-dataset-cleaned.zip |
94.4 MB | - |
07_PV_arc_self_extinction_small/Imported_data_Electric_Arc_IEC63027.ipynb |
2.1 MB | - |
07_PV_arc_self_extinction_small/Modelling - 0.8mm - 250V 16A - Tira 2A a cada 200ms.csv |
15.1 KB | - |
07_PV_arc_self_extinction_small/Modelling - 0.8mm 250V 5A - tira 0.1A a cada 70ms.csv |
37.5 KB | - |
07_PV_arc_self_extinction_small/Modelling - 1.1mm - 250V 16A - Tira 2A a cada 200ms.csv |
15.7 KB | - |
07_PV_arc_self_extinction_small/Modelling - 1.1mm 250V 5A - tira 0.1A a cada 100ms.csv |
41.9 KB | - |
07_PV_arc_self_extinction_small/Modelling - 1.6mm - 250V 16A - Tira 2A a cada 200ms.csv |
18.5 KB | - |
07_PV_arc_self_extinction_small/Modelling - 1.6mm 250V 5A - tira 0.1A a cada 100ms.csv |
47.8 KB | - |
07_PV_arc_self_extinction_small/Modelling - 2.5mm - 250V 16A - Tira 2A a cada 200ms.csv |
14.5 KB | - |
07_PV_arc_self_extinction_small/Modelling - 2.5mm 250V 5A - tira 0.1A a cada 100ms.csv |
77.3 KB | - |
07_PV_arc_self_extinction_small/Modelling - 5mm - 250V 16A - Tira 2A a cada 200ms.csv |
10.5 KB | - |
07_PV_arc_self_extinction_small/Modelling - 5mm 250V 5A - tira 0.1A a cada 100ms.csv |
51.6 KB | - |
07_PV_arc_self_extinction_small/Validation - Constant Current - 1.6mm - 5A 30Vmpp 5Impp.csv |
117.9 KB | - |
07_PV_arc_self_extinction_small/Validation - Constant Current - 1.6mm - 5A 60Vmpp 5Impp.csv |
114.3 KB | - |
07_PV_arc_self_extinction_small/Validation - Constant Power - 1.6mm - 220.32W 51Vmpp 10,8Impp.csv |
47.4 KB | - |
07_PV_arc_self_extinction_small/Validation - Constant Power - 1.6mm - 70W 35Vmpp 5Impp.csv |
46.4 KB | - |
07_PV_arc_self_extinction_small/Validation - Constant Resistance - 1.6mm - 4.17R 45Vmpp 10,8Impp.csv |
116.5 KB | - |
07_PV_arc_self_extinction_small/Validation - Constant Resistance - 1.6mm - 5.56R 60Vmpp 10,8Impp.csv |
114.8 KB | - |
07_PV_arc_self_extinction_small/Validation - Constant Voltage - 1.1mm - 200Vmpp 9,9Impp - Copia.csv |
115.5 KB | - |
07_PV_arc_self_extinction_small/Validation - Constant Voltage - 1.1mm - 200Vmpp 9,9Impp.csv |
57.8 KB | - |
07_PV_arc_self_extinction_small/Validation - Constant Voltage - 1.6mm - 180Vmpp 10,8Impp - Teste 2.csv |
114.6 KB | - |
07_PV_arc_self_extinction_small/Validation - Constant Voltage - 1.6mm - 180Vmpp 10,8Impp.csv |
116.8 KB | - |
07_PV_arc_self_extinction_small/Validation - Constant Voltage - 1.6mm - 195Vmpp 10,8Impp.csv |
116.5 KB | - |
07_PV_arc_self_extinction_small/Validation - Constant Voltage - 1.6mm - 210Vmpp 10,8Impp.csv |
114.5 KB | - |
07_PV_arc_self_extinction_small/Validation - Constant Voltage - 2.5mm - 250V 4A corrente constante.csv |
113.1 KB | - |
07_PV_arc_self_extinction_small/Validation - Constant Voltage - 2mm - 180Vmpp 10,8Impp.csv |
114.6 KB | - |
08_Synthetic_AC_arc_CC0/dataset_arc_fault_nonlinear_AC-main.zip |
1.7 MB | - |
Each file keeps the license of its original publisher (see the notes above).
Master index
This is one of 19 repos (AFCI_part1 … AFCI_part19). Overview of all parts: AFCI_part6 / 00_START_HERE_MASTER_INDEX.md · Installation: AFCI_part6 / 00_INSTALLATION_GUIDE.md
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support