Data Validity Determination of HVDC Audible Noise Using Temporal and Spatial Correlation Features

The corona discharge of high-voltage direct current (HVDC) transmission lines can lead to severe audible noise and has become a considerable constraint in constructing transmission lines. However, during outdoor testing of audible noise, a large amount of invalid data is often measured due to environmental noise interference, posing significant challenges for data analysis. This paper proposes an effective method for identifying the validity of audible noise. After utilizing a microphone array to capture sound signals and extract various features from both the temporal and spatial dimensions, we employ a support vector machine (SVM) classification algorithm to identify whether environmental noise is present in the audible noise. Experimental results demonstrate that compared to solely using temporal or spatial features, the proposed method using joint features achieves higher accuracy in sample data testing, reaching 96.94%. The method for identifying invalid data provides a new perspective for screening and processing audible noise data.