7.7.6. AC Arc Fault

Detect AC arc faults in electrical systems using current waveform analysis.

7.7.6.1. Overview

AC arc faults occur when electrical current flows through an unintended path, often caused by damaged insulation, loose connections, or worn conductors. This example demonstrates how to detect these dangerous conditions using machine learning on current sensor data.

Application: Electrical safety systems, circuit breakers, residential/commercial protection

Task Type: Time Series Classification

Data Type: Univariate (current waveform)

7.7.6.2. Configuration

common:
  target_module: 'timeseries'
  task_type: 'generic_timeseries_classification'
  target_device: 'F28P55'

dataset:
  dataset_name: 'ac_arc_fault'

training:
  model_name: 'ArcFault_model_700_t'
  training_epochs: 50
  batch_size: 32

testing: {}
compilation: {}

7.7.6.3. Running the Example

cd tinyml-modelzoo
./run_tinyml_modelzoo.sh examples/ac_arc_fault/config.yaml

7.7.6.4. Dataset Details

The AC arc fault dataset contains current waveforms sampled during normal operation and various arc fault conditions.

Classes:

  • Normal operation

  • Arc fault conditions

Input Features: Current waveform samples

7.7.6.6. See Also