OVMD-MF算法用于漏电流光纤传感
2025,47(1):181-189
吴健华
海军工程大学 电气工程学院, 湖北 武汉 430033 ;
中国人民解放军92853部队, 辽宁 葫芦岛 125100
张晓锋
海军工程大学 电气工程学院, 湖北 武汉 430033
陈亮
海军工程大学 电气工程学院, 湖北 武汉 430033
海军工程大学 电气工程学院, 湖北 武汉 430033 ;
中国人民解放军92853部队, 辽宁 葫芦岛 125100
张晓锋
海军工程大学 电气工程学院, 湖北 武汉 430033
陈亮
海军工程大学 电气工程学院, 湖北 武汉 430033
摘要：
针对船舶电力网络漏电流光纤传感测量响应度低、测量能力弱等问题，提出了一种基于最优变分模态分解与形态学滤波组合的降噪方法，以提升微弱电流测量能力。以希尔伯特谱熵为适应度函数，利用捕食者优化算法搜寻分解模式数及二次惩罚因子的最优组合，并完成变分模态分解获取本征模态函数分量。设置相关系数阈值确定有效模态函数分量，完成信号重构。以信噪比为适应度函数，通过捕食者优化算法确定形态学滤波的结构因子及比例系数，对重构后的信号进行时域降噪处理，实现对漏电流信号进一步降噪。通过仿真分析以及实验验证，确定该方法相对于现有的滤波方法，改善了信号的信噪比，降低了最小均方误差，泄漏电流的分辨力可提升至3 mA。
针对船舶电力网络漏电流光纤传感测量响应度低、测量能力弱等问题，提出了一种基于最优变分模态分解与形态学滤波组合的降噪方法，以提升微弱电流测量能力。以希尔伯特谱熵为适应度函数，利用捕食者优化算法搜寻分解模式数及二次惩罚因子的最优组合，并完成变分模态分解获取本征模态函数分量。设置相关系数阈值确定有效模态函数分量，完成信号重构。以信噪比为适应度函数，通过捕食者优化算法确定形态学滤波的结构因子及比例系数，对重构后的信号进行时域降噪处理，实现对漏电流信号进一步降噪。通过仿真分析以及实验验证，确定该方法相对于现有的滤波方法，改善了信号的信噪比，降低了最小均方误差，泄漏电流的分辨力可提升至3 mA。
基金项目：
OVMD-MF algorithm for fiber optic leakage current sensing
WU Jianhua
College of Electrical Engineering, Naval University of Engineering, Wuhan 430033 , China ;
The PLA Unit 92853, Huludao 125100 , China
ZHANG Xiaofeng
College of Electrical Engineering, Naval University of Engineering, Wuhan 430033 , China
CHEN Liang
College of Electrical Engineering, Naval University of Engineering, Wuhan 430033 , China
College of Electrical Engineering, Naval University of Engineering, Wuhan 430033 , China ;
The PLA Unit 92853, Huludao 125100 , China
ZHANG Xiaofeng
College of Electrical Engineering, Naval University of Engineering, Wuhan 430033 , China
CHEN Liang
College of Electrical Engineering, Naval University of Engineering, Wuhan 430033 , China
Abstract：
Aiming at low response and weak measurement ability of the all-fiber optic leakage current sensor in the ship power network, a denoising method based on the combination of VMD(variational modal decomposition) and MF(morphological filtering) was proposed to improve the ability of weak current measurement. The hunter-prey optimization algorithm was used to search for the optimal combination of model number and quadratic penalty term based on the minimum entropy of the Hilbert spectrum, and the intrinsic mode functions component were obtained by completing the variational mode decomposition. The practical modal function component was determined by setting the correlation coefficient threshold to complete the signal reconstruction. The structure factor and scale coefficient of MF were determined by the hunter-prey optimization algorithm with the signal-to-noise ratio as the fitness function. The reconstructed signal was subjected to time-domain denoising processing to further denoise in the leakage current signal. This method significantly improve the signal-to-noise ratio and reduce the minimum mean square error to the existing filtering methods by simulation analysis and experimental verification. The leakage current resolution is achieved to 3mA with the proposed denoising algorithm.
Aiming at low response and weak measurement ability of the all-fiber optic leakage current sensor in the ship power network, a denoising method based on the combination of VMD(variational modal decomposition) and MF(morphological filtering) was proposed to improve the ability of weak current measurement. The hunter-prey optimization algorithm was used to search for the optimal combination of model number and quadratic penalty term based on the minimum entropy of the Hilbert spectrum, and the intrinsic mode functions component were obtained by completing the variational mode decomposition. The practical modal function component was determined by setting the correlation coefficient threshold to complete the signal reconstruction. The structure factor and scale coefficient of MF were determined by the hunter-prey optimization algorithm with the signal-to-noise ratio as the fitness function. The reconstructed signal was subjected to time-domain denoising processing to further denoise in the leakage current signal. This method significantly improve the signal-to-noise ratio and reduce the minimum mean square error to the existing filtering methods by simulation analysis and experimental verification. The leakage current resolution is achieved to 3mA with the proposed denoising algorithm.
收稿日期：
