Files
Abstract
The prevalence of electrical fires due to arc faults in both AC and DC systems has necessitated advancements in detection technologies. This dissertation introduces innovative artificial intelligence (AI)-based methods for efficient arc fault detection, enhancing both safety and reliability in electrical systems. For AC systems, a convolutional neural network (CNN)-based arc fault detection algorithm has been proposed, which has an arc fault detection accuracy of 99.47% with a balanced sampling rate of 10 kHz. The performance of the proposed algorithm has been verified using the Raspberry Pi 3B platform.Although traditional CNN algorithms exhibit high accuracy, they require optimization for real-time arc fault detection on resource-limited hardware. To address this, a lightweight CNN architecture combined with a model compression technique using a knowledge distillation-based teacher-student algorithm has been proposed. This model maintains a high detection accuracy of 99.31% and achieves a minimal runtime of 0.20 ms per sample on the Raspberry Pi 3B platform, demonstrating its suitability for commercial embedded microcontrollers (MCUs) with limited computational capability. Extending the research to DC systems, a cost-effective, AI-driven Arc Fault Circuit Interrupter (AFCI) for DC applications has been proposed. Utilizing an STM32 MCU and a silicon carbide (SiC) MOSFET-based solid-state circuit breaker (SSCB), this system achieves a detection accuracy of 98.14% with an arc fault interruption time of 19 ms. To enhance robustness, instead of making decisions on a single cycle of arc, a few consecutive cycle wait time has been proposed. This research demonstrated an arc fault clearing time of 95 ms considering 5 cycles of arcs as one arc fault event. This AFCI solution stands out for its rapid response and high reliability, promising significant improvements in safety for DC-powered installations.Together, these contributions signify a leap forward in the domain of electrical safety, presenting viable solutions for the timely detection and interruption of arc faults in both AC and DC systems. The outcomes of this research are expected to influence future standards and practices in electrical safety management across residential, commercial, and industrial sectors.