Simulation and Analysis of Acoustic Wave Propagation due to Partial Discharge Activity

Acoustic detection of partial discharges is based on the retrieval and analysis of mechanical signals produced by partial discharges. Acoustic method is widely used in locating partial discharge sources in transformers. For source location constant velocity of acoustic signals is used. Acoustic method also has potential to classify the partial discharges for better assessment of insulation condition. This paper presents a computer simulation of acoustic signal and analysis of its propagation behavior. It is found that acoustic velocity is not constant over distances of practical interest. An algorithm proposed for source location with distance dependent acoustic velocity leads to reduction in location error. Acoustic pulse produced by partial discharges is deformed and attenuated while propagating through transformer to sensor. This change depends on propagation distance and medium. So, frequency spectrum of sensor output AE pulse is not true representation of source pulse frequency spectrum. Partial discharge classification based on the output acoustic signal can lead to wrong classification. Knowing the location of partial discharge source, frequency dependent attenuation characteristics and output acoustic signal´s frequency spectrum, an estimation of input pulse frequency spectrum and its parameters is made for the classification of partial discharges