Applying Edge AI to DC Arc Fault Detection (Part 2): Software Development to Deployment (Download)
Part 1 of this article series explored how edge AI-enabled microcontrollers can more effectively detect dangerous DC series arc faults in solar, electric-vehicle (EV), and battery storage systems compared to traditional protection methods. Part 2 explores the practical software workflow, from data collection through model training to deployment on real-time control MCUs.
In a typical arc fault detection system, a current transformer senses current between solar panels and an inverter, while an analog front-end circuit conditions the signal through gain and filtering stages. In a system with a real-time MCU, such as the Texas Instruments (TI) C2000 MCU, its integrated analog-to-digital converter (ADC) digitizes the signal for processing.
