7.7.18. Torque Measurement Regression
Predict PMSM motor torque from current measurements.
7.7.18.1. Overview
This example demonstrates torque estimation for Permanent Magnet Synchronous Motors (PMSM) using phase current measurements. Accurate torque estimation enables sensorless torque control, reducing cost and complexity.
Application: Motor drives, electric vehicles, industrial automation
Task Type: Time Series Regression
Data Type: Multivariate (motor currents)
7.7.18.2. Configuration
common:
target_module: 'timeseries'
task_type: 'generic_timeseries_regression'
target_device: 'F28P55'
dataset:
dataset_name: 'torque_measurement_regression'
training:
model_name: 'REGR_4k_NPU'
training_epochs: 50
batch_size: 32
testing: {}
compilation: {}
7.7.18.3. Running the Example
cd tinyml-modelzoo
./run_tinyml_modelzoo.sh examples/torque_measurement_regression/config.yaml
cd tinyml-modelzoo
run_tinyml_modelzoo.bat examples\torque_measurement_regression\config.yaml
7.7.18.4. Dataset Details
Input Variables:
Phase A current (Ia)
Phase B current (Ib)
Phase C current (Ic)
Rotor position (optional)
Output:
Torque (continuous value in Nm)
7.7.18.5. Recommended Models
Model |
Parameters |
Use Case |
|---|---|---|
|
~1,000 |
Basic estimation |
|
~4,000 |
Balanced accuracy |
|
~8,000 |
High accuracy |
7.7.18.6. See Also
Induction Motor Speed Prediction - Speed prediction