7.7.10. Electrical Fault

Classify transmission line faults using voltage and current measurements.

7.7.10.1. Overview

This example demonstrates classification of electrical faults in power transmission lines. It uses voltage and current measurements to identify different fault types including line-to-ground, line-to-line, and three-phase faults.

Application: Power grid protection, substation automation, fault localization

Task Type: Time Series Classification

Data Type: Multivariate (voltage and current signals)

7.7.10.2. Dataset Variants

Two dataset scripts are provided for different classification tasks:

Script	Output Dataset	Classes	Use Case
electrical_fault.py	electrical_fault_dataset.zip	2 (fault / no fault)	Binary fault detection
electrical_fault_6class.py	electrical_fault_6class_dataset.zip	6 (G, C, B, A combinations)	Fault type classification

Prepare the 2-class dataset:

cd examples/electrical_fault
python electrical_fault.py

Prepare the 6-class dataset:

cd examples/electrical_fault
python electrical_fault_6class.py

Set dataset.input_data_path in the config to point to the generated zip file.

7.7.10.3. Configuration

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

dataset:
  dataset_name: 'electrical_fault'
  # Use 'electrical_fault_6class' for 6-class fault type classification

training:
  model_name: 'CLS_4k_NPU'
  training_epochs: 50
  batch_size: 32

testing: {}
compilation: {}

7.7.10.4. Running the Example

LinuxWindows

cd tinyml-modelzoo
# 2-class (fault detection) on F28P55
./run_tinyml_modelzoo.sh examples/electrical_fault/config.yaml

# 6-class (fault type classification) on MSPM0
./run_tinyml_modelzoo.sh examples/electrical_fault/config_MSPM0.yaml

7.7.10.5. Dataset Details

Input Variables:

• Phase voltages (Va, Vb, Vc)

• Phase currents (Ia, Ib, Ic)

Fault Classes:

• No fault (normal)

• Line-to-ground fault (LG)

• Line-to-line fault (LL)

• Line-to-line-to-ground fault (LLG)

• Three-phase fault (LLL)

Simulink Model:

Simulink model used for generating electrical fault training data

7.7.10.6. See Also

• Grid Stability - Power grid stability prediction

• Example on GitHub