Arc Fault Circuit Interrupter Solution
Overview
AFCI Solution Based on GD32H7 Series MCU
During the process of converting solar energy into electrical energy in photovoltaic (PV) power systems, faults such as poor contact or short circuits on the surface of photovoltaic panels may occur due to factors like dust accumulation or equipment aging. These faults can lead to arcing phenomena. Arcing is a phenomenon of air ionization caused by a significant voltage difference. The high temperature and high voltage effects produced by arcing not only cause damage to the photovoltaic panels and reduce the efficiency of energy conversion, but may also trigger fire and other safety accidents. Therefore, the implementation of an Arc Fault Circuit Interrupter (AFCI) solution for real-time detection of arcing phenomena in photovoltaic systems is crucial for the safe application of photovoltaics.
Main Features
Gigadevice AFCI solution is based on GD32H7 series 600MHz high-performance general-purpose MCU and adopts AI algorithm for DC arc detection solution, which can support up to 12 ADC channels under 500K sampling rate with real-time AI inference capability for arc detection. The deep convolutional neural network model analyses 4096 sampling points to get the result, using only 1.7k size flash and taking 0.6ms. The model is compatible with data collected by different hardware, compatible with 0~40A current, achieving 100% accuracy.
Gigadevice provides a one-click training tool to generate code that can be compiled directly. The data collection tool is used to gather arc signals and normal signals, and it has the capability to test and evaluate the deployed model. The one-click training tool uses a large amount of data to train and generate an AI model for arc detection, and supports the one-click generation of compliable and runnable arc detection code, ensuring accurate identification of arc characteristics in various scenarios, reducing the risk of missed and false alarms.
The AFCI solution paired with a 16-bit external ADC, supporting 8-channel external ADC acquisition. The solution is integrated with a memory card and a WIFI module based on GD32VW553, which can automatically upload data to the cloud platform, reserving the serial port and CAN interface.
The AFCI solution supports cloud data collection, cloud training, and OTA upgrades of model parameters. Users can continuously optimize and update the model through the cloud platform to adapt to various complex and changing working environments.
Recommended Products
