摘要
高速铁路供电箱式变电站长期工作在高温潮湿环境,针对箱变内部变压器等关键设备运行安全问题,提出一种利用声波和振动信号联合特征进行变压器运行状态辨识方法。首先,通过非侵入式传感器获取伴随配电变压器运行的声波和振动信号,再经过滤波消噪处理声波信号提取其时频谱图纹理特征,并与振动信号变模态分解后的排列熵联合形成联合特征向量;通过概率分布特征构造最优分类超平面,由支持向量机分类算法实现联合特征向量的变压器典型状态的辨识。实验证明,声振联合特征分析方法对变压器正常、轻载、重载和三相不平衡等各运行状态下判别准确率均达96%以上。通过对辨识发现的变压器异常工况进行预警,及时实施设备状态检修可大大提高铁路供电可靠性。
The box-type substation is responsible for high-speed railways power supply and works in a high temperature and humid environment.Aiming at its internal transformers and other equipments running safety,a state identifying method for transformers by utilizing combined characteristics of sound waves and vibration signals is proposed.Acoustic and vibration signals accompanying the operation of distribution transformers are acquired by non-intrusive sensors.Acoustic signal texture spectrogram features are extracted after filtering and denoising processing,and combined to form joint feature vector with the modal decomposition permutation entropy of vibration signal.The optimal classification hyperplane is constructed by the probability distribution feature,and the support vector machine classification algorithm is used to realize the typical transformer state identification with joint eigenvectors.Experiments show that the combined sound-vibration characteristic analysis method has an accuracy of more than 96%for transformersnormal,light-load,heavy-load and three-phase unbalanced operating conditions.Based on timely identifying and warning of equipment abnormal work conditions,the reliability of high-speed railway power supply can be improved through timely condition-based maintenance.
作者
郑晓庆
ZHENG Xiaoqing(China Railway Design Corporation,Tianjin 300308,China)
出处
《铁道标准设计》
北大核心
2023年第6期180-186,共7页
Railway Standard Design
基金
中国铁路设计集团有限公司重点课题(0100540)
中国铁路设计集团有限公司科研项目(0101223)。
关键词
高速铁路
配电变压器
纹理特征
排列熵特征
状态辨识
high-speed railway
distribution transformer
texture feature
permutation entropy feature
state identification
