The detection of partial discharge (PD) in the field is critically important for the maintenance of electrical power systems. Traditional methods often employ the affine projection algorithm (APA) for denoising PD signals, aiming to retain their fundamental characteristics. However, the conventional APA approach exhibits limitations in effectively eliminating pulse noise. To address this issue, this paper introduces a hyperbolic cosine function and proposes an innovative denoising technique, termed the affine projection exponential hyperbolic cosine algorithm (APEHCA). This enhanced APA-based method is specifically designed to mitigate noise in PD signals. Both simulation and empirical tests substantiate that the newly proposed algorithm excels in system identification and significantly improves the quality of PD signal denoising.

现场局部放电 (PD) 检测对于电力系统的维护至关重要。传统方法通常采用仿射投影算法（APA）对局放信号进行去噪，旨在保留其基本特征。然而，传统的 APA 方法在有效消除脉冲噪声方面存在局限性。为了解决这个问题，本文引入了双曲余弦函数，并提出了一种创新的去噪技术，称为仿射投影指数双曲余弦算法（APEHCA）。这种基于 APA 的增强方法专门用于减轻 PD 信号中的噪声。仿真和实证测试都证实了新提出的算法在系统辨识方面表现出色，并显着提高了局放信号去噪的质量。