摘要
变压器作为供电系统的核心设备,其稳定性和健康状态直接关系到电力的供应,因此对其故障诊断方法的研究显得极其重要。本文阐述的变压器诊断方法基于大数据、物联网及AI等技术,其核心是故障数学模型的建立。本文给出了基于声振的变压器诊断系统架构。通过对故障数学模型的深入分析,发现由生成样本加少量真实故障样本训练的故障诊断模型具有更高的准确性。
As the core equipment of power supply system, the stability and physical condition of transformer are directly related to power supply, so the study on its fault diagnosis method is very important. The transformer diagnosis method described in this paper is based on big data, Internet of things, AI and other technologies, while the core is the establishment of fault mathematical model. In this paper, the structure of transformer diagnosis system based on acoustic-vibration is presented. Through the in-depth analysis of the fault mathematical model, it is found that the fault diagnosis model trained by generated samples and a small amount of real fault samples has higher accuracy.
作者
赵学举
ZHAO Xueju(Kehua Data Co.,Ltd.,Xiamen 361006,China)
出处
《电声技术》
2022年第4期113-115,共3页
Audio Engineering
关键词
数学模型
故障诊断
对抗学习
样本训练
mathematical model
fault diagnosis
adversarial learning
sample training
