JADE盲源分离算法在变压器振动信号监测中的应用

铁芯和绕组是变压器振动产生的主要振源,监测铁芯和绕组的振动信号可以对变压器运行状态进行预测和诊断.选择了S13-12500/35型油浸式无励磁调压电力变压器,将2只光纤Bragg光栅振动传感器布设于变压器内部,直接采集变压器负载运行下的振动信号.通过JADE算法将2组负载信号分离,发现:一组信号能量主要集中在100 Hz处,与铁心振动信号频谱特征较为接近;另一组信号在各频率上均有分布,与绕组振动信号频谱特征较为接近.

Application of blind source separation algorithm based on JADE in monitoring of transformer vibration signal

Core and winding are main vibration sources which generate transformer vibration,monitoring vibration signals of core and winding,can predict and diagnose operational status of ransformer.Select general changes made S13-12500/35 type oil-immersed non-excitation voltage regulating power transformer,the two fiber Bragg grating vibration sensors are laid inside transformer,vibration signal is acquired directly under transformer load.By JADE algorithm,two groups load signal are separated,after separation,a groups is found that signal energy is concentrated at 100 Hz which is closer to core vibrational signal spectral characteristics,another group of signals are distributed at each frequency,and closer to spectral characteristics of winding vibration signal.