Hardware Discrete Wavelet Transform Based SoC For Low-voltage Series AC Arc Fault Detection And Identification Method

Arc fault is gradually regarded as one of the main causes for residential building electrical fire incidents.Due to its inherent properties such as complexity,randomness,and uncertainty,traditional circuit breakers are unable to respond in time and malfunction.Thus,signal characteristics of arc fault is studied and analyzed.As a result,a new arc fault feature detection algorithm is proposed.This algorithm is combined with discrete wavelet transform(the Mallat algorithm)to implement a hardware digital detection circuit.This circuit co-works with high speed ADCs to analyze features of arc fault signals.And it overcomes the computational bottleneck of software implementation of traditional MCUs.Based on this detection unit,a power dedicated SoC for low-voltage series ac arc fault detection is proposed.This SoC is designed and fabricated using standard 55nm CMOS process with a die size of 1.55*1.9mm~2 and power consumption of 66m W.And an arc fault detection device(AFDD)is built using this SoC.A method is proposed to locate the initial position where arc fault occurs with the AFDD system.Finally,a series of tests are conducted to verify the system under experimental conditions.The test results indicate that the system can detect low-voltage series AC arc faults in real-time,with detection time less than 80ms and a recognition rate of 97%.