Comparison of processing techniques for optimizing the diagnosis of solid insulation based on acoustic emissions from partial discharges

A partial discharge (PD) is a phenomenon related to the degradation of the insulation systems in HV equipment like power transformers. PD is present in the entire life of the HV equipment but its presence is not an indicative of the existence of a critical defect in the insulation. However, if a set of PD are located in a small region or the repetition rate is high in a small temporal window, then there is a high probability of an imminent failure of the equipment. The detection of the acoustic emission (AE) generated by PD is a non-intrusive technique which can be used for the detection and location of PD in order to perform online condition assessment of insulation systems in power transformers. Although the multiple advantages of the acoustic detection, it has several difficulties to overcome like strong attenuation, signal distortion or accurate arrival time determination. In the last years, new digital signal processing (DSP) techniques has been studied to improve the signal to noise ratio (SNR) of the signals and the detection of the beginning of the detected signals. The objective of this work is the analysis of these DSP techniques in order to improve the diagnosis of power transformers by means of the detection of the AE.