DC Fault Arc Compressive Sampling Based on Mean-Enhanced Generalized Variable-Parameter Fibonacci Sequence
Chenhu Yuan, Zijia Liang, Bowen Zheng, Zhenhui Zhang, Yun Peng, 范训明, Haoran Kong · IEEE Access · 2026
To mitigate data redundancy and communication bandwidth pressure in low-voltage DC (LVDC) system monitoring, this paper proposes a DC fault arc compressive sensing (CS) method based on the Mean-Enhanced Generalized Variable-Parameter Fibonacci (ME-GVF) sequence. The GVF sequence is constructed by cascading a 3D-Logistic chaotic map with the Fibonacci sequence. Mean enhancement is achieved via logical OR operations with an M-sequence, maintaining the low correlation of chaotic sequences while establishing a "low-frequency direct channel" to suppress high-frequency noise. Furthermore, the phase space characteristics of the chaotic sequence and the Restricted Eigenvalue Condition (REC) of the observation matrix are analyzed. Simulation results compare the proposed sequence against M, Gold, and various typical chaotic sequences. The results demonstrate that under multi-load conditions and high compression ratios, the ME-GVF sequence exhibits superior performance in reconstruction error and signal-to-noise ratio (SNR). Finally, adaptability analysis confirms that the ME-GVF sequence, combined with the OMP algorithm, facilitates precise fault feature extraction with bounded computational complexity. This approach holds significant potential for developing low-power, lightweight DC monitoring systems.