Improved CVMD method applied to partial discharge signal

Partial discharge signals are often interfered by white noise and narrowband noise, which makes it difficult to locate the partial discharge source. To tackle this intricate issue, this paper proposes an improved variational Mode Decomposition (VMD) noise reduction method. Initially, the method optimizes the VMD by employing the Crested Porcupine Optimizer (CPO). Subsequently, modal screening is carried out in accordance with a specific criterion. Finally, noise reduction is implemented through an improved 3σ method. Via noise - reduction simulation experiments, the superiority of this method is validated, and comparisons are made with several other methods, wherein this method is capable of boosting the Signal - to - Noise Ratio (SNR) by up to 50, elevating the correlation coefficient to over 0.99, and reducing the Mean Squared Error (MSE) by 99%. Meanwhile, the practicality of this method is confirmed through localization experiments.