Chemical detection is one of the most sensitive and reliable techniques for detecting the chemical composition generated during the decomposition process of transformer oil. However, it cannot provide any information about the location and hotspots of faults, so it is often only used to verify the existence of faults Another detection method is high-performance liquid chromatography (HPLC), which is used to detect the presence of liquids in transformer oil. By scanning the infrared thermal image, the state of the transformer can be estimated, providing a heating mode for the load

The commonly used analysis method is dissolved gas analysis (DGA), which is mainly used to detect dissolved gases in transformer oil. The detection of initial and transient faults in transformers is effective. However, it cannot locate partial discharge (PD)

Partial discharge is an electrical phenomenon that originates from gaps in insulation or is largely caused by electrical breakdown in an asymmetric electric field.

. Internal insulation faults can lead to destructive breakdown, deterioration, and irreversible damage to system performance. Therefore, the development of partial discharge diagnosis technology based on electrical phenomena and aging precursors of high-voltage equipment is of great significance. There are two main methods for diagnosing partial discharge sources based on the different types of signals generated by partial discharge: electrical detection and acoustic detection

For electrical methods, pulse current method is a traditional PD measurement method. When there is partial discharge in the transformer, transient voltage changes are detected at both ends. This method has high sensitivity and easy quantification of changes. However, the current detected by the pulse current method is not the actual partial discharge

The capacitance component is another method for detecting partial discharge.

. The disadvantage of this method is that its sensitivity decreases with the increase of sample capacitance. In recent years, with the development of digital measurement technology, improved capacitance component method, endpoint current pulse frequency method, and energy ratio line method have been proposed. By using the improved capacitance component method and the data obtained from the current pulse signal of the winding, the positioning error rate can be reduced from 5 to 4.6%. However, due to the high level of electrical and environmental noise caused by various interferences and disturbances in the power transformer environment, PD measurement devices have subtle limitations in distinguishing between electrical noise and PD. For this purpose, an ultra-high frequency (UHF) sensor has been developed to detect electromagnetic waves generated by PD in the 300 MHz 3 GHz frequency range. UHF sensors can locate low noise level PD sources and resist local interference by connecting to power transformers in non electrical units. However, this method cannot provide a reliable indication of remaining life. Therefore, it is necessary to improve the anti-interference stability of the ultra-high frequency method