1 Create Motor fan state project in NXP-Eiq
With 3-axis sensor on FRDM-MCXN236, to train and deploy a machine learning model for a predictive fan state application
Fan States:
-
FAN-Off — The fan is turned off.
-
FAN-On — The fan is turned on and running in a normal condition.
-
FAN-Clogged — The fan is turned on and the airflow is obstructed.
-
FAN-Friction — The fan is turned on and excessive friction or a blade fault is detected.
2 Training
Add data in csv file
-
Each sample window comprises original signals from the three channels (ax, ay, az) of an accelerometer. There are 128 sampling points in each channel, totaling 3 * 128 = 384 float data points. Each sample time window is written as one line.
-
Data format: x0 y0 z0 x1 y1 z1…, separated by spaces.
Then start training
3 Emulating
The model during training can be emulated to validate it.
4 Generate code
If the value is OK. Then generate the code.
This code comes with read-only library file and one header file in .h. This can reduce time to code and it limit the model selection as well.
5 Generate code and download
After the process , there is problem. The On-board MCU-Link is missing, and cannot be downloaded to the FRDM -MCXN236 board.
Try many time, still can not probe and recover the port. Really hard time for this project.
Before figure out what to do. Try more times for new wayout.
