Electric AC Arc Faults Simulation Using Measured Data Based on PCA

AC arc faults’ detection aims to detect the early arc faults to prevent potential accidents. To develop a small and low-cost testing device for arc fault detection devices (AFDDs), an ac arc faults’ simulation method is vital. Generally, physical arc models are used to generate arc fault signals. However, signals can hardly address the high-frequency component, such as the transient spikes. From a data-driven perspective, we propose an electric ac arc faults’ simulation method using measured data based on principal component analysis (PCA) to simulate more accurate and more realistic arc current signals. Without extra physical models, this method can adaptively simulate actual ac arc fault current signals in different application scenarios using a few measured data. Based on a set of standard orthogonal basis derived from PCA, the simulated ac arc current signals can maintain the time and frequency features more effectively than classical models. Compared with other simulated arc current signals, real-world experimental results indicate that the proposed method can simulate more accurate and more actual ac arc current signals.