Series DC Arc Characteristic and Diagnosis Strategy for Distributed PV Power Generation

In order to realize the nation’s goal of achieving carbon peaking before year 2030 and the carbon neutrality by year 2060, the installed capacity of distributed PV power generation system has increased year by year in China, which increases the probability of DC arc fault. The features of DC arc may be very weak under some conditions. When the current of DC arc is large and the voltage level of photovoltaic array is high, DC arc will be very stable, and it is hard to be detected. Besides, if there are many high-power devices around DC arc detector, the output signal of current sensor will be seriously interfered, and it will be more difficult to distinguish between normal state and arc fault state. DC arc can be divided into unstable combustion stage and stable combustion stage. When in stable combustion stage, DC arc is also hard to be detected. Fortunately, the features of DC arc are obvious when in unstable combustion stage and the value of current spectrum energy will significantly increase. When the spectrum energy exceeds the set threshold in five consecutive time windows, it is preliminarily determined that there is a suspected arc. If the spectrum energy still exceeds the set threshold within five consecutive time windows, it is finally determined that there is an arc fault, otherwise, short PV array for further diagnosis. If there is an arc fault, the fluctuation characteristics of DC arc will be enhanced and the loop current will go fluctuating intensely again, then spectrum energy calculated will exceed the set threshold and an arc fault alarm can be sent out. Taking into consideration the different features of DC arc’s unstable combustion stage and stable combustion stage, a Series DC arc diagnosis method based on current spectrum energy is proposed, which amplifies the fluctuation characteristics of arc current and keeps arc current fluctuating strongly for a long time by shorting the PV array. The spectrum energy is calculated through FFT analysis of each PV string’s current, compared to the calculated spectrum energy with the set threshold, and then decided whether DC arc fault exists. The results to an experimental platform verify the effectiveness of the arc detection method, which has advantages of low calculation cost and high accuracy.