电力应用场景局部放电声纹数据集

本数据集为电力应用场景局部放电声纹数据，选取GIS设备模型、变压器模型、水下电缆管路3类典型电力设备，针对性模拟尖端放电、沿面放电等多种典型局部放电故障，采用自研光学麦克风探头及解调模块进行数据采集，通过时域分析与频域分析相结合的方法，完成了3类电力应用场景局部放电声纹数据集的构建。本数据集不仅为多种故障声纹特征标准化数据库的构建提供了核心数据支撑，更可为局部放电智能识别算法的训练、验证与优化提供高质量样本，助力提升电力设备局部放电检测的自动化、精准化水平，具有一定的工程应用价值与学术研究意义。本数据集包含3个子文件夹，分别为“GIS设备”、“变压器”、“水下电缆”，对应了3种电力行业典型应用场景。其中，“GIS设备”文件夹内含尖端放电、沿面放电、升降压过程中放电时采集的声信号波形数据；“变压器”文件夹内含尖端放电、悬浮放电、油隙放电时采集的声信号波形数据；“水下电缆”文件夹内含1kHz、100kHz固定声信号频率以及局放发生时随机频率下采集的声信号波形数据。数据集大小为20 MB。

This dataset is a partial discharge acoustic fingerprint dataset for power application scenarios. Three types of typical power equipment models, namely GIS equipment models, transformer models and submarine cable pipelines, were selected for the dataset. Multiple typical partial discharge faults such as tip discharge and surface discharge were simulated in a targeted manner. Data collection was conducted using self-developed optical microphone probes and demodulation modules. Combining time-domain and frequency-domain analysis methods, the construction of the partial discharge acoustic fingerprint dataset for the three power application scenarios was completed.

This dataset not only provides core data support for the establishment of standardized databases for various fault acoustic fingerprint features, but also offers high-quality samples for the training, validation and optimization of partial discharge intelligent recognition algorithms. It helps improve the automation and accuracy of partial discharge detection for power equipment, and holds certain engineering application value and academic research significance.

The dataset includes 3 subfolders, namely "GIS Equipment", "Transformer" and "Submarine Cable", corresponding to three typical application scenarios in the power industry. Specifically, the "GIS Equipment" folder contains acoustic signal waveform data collected during tip discharge, surface discharge and discharge during voltage step-up and step-down processes; the "Transformer" folder contains acoustic signal waveform data collected during tip discharge, floating discharge and oil gap discharge; the "Submarine Cable" folder contains acoustic signal waveform data collected at fixed acoustic signal frequencies of 1kHz and 100kHz, as well as at random frequencies when partial discharge occurs. The total size of the dataset is 20 MB.