Three-dimensional Visualization Technology for Ultrasonic Detection of Partial Discharges in Power Transformers

Abstract

Partial discharges are crucial sources for insulation deterioration in power transformers, and precise detections of partial discharge is of great significance for transformer condition monitoring. Recently, visualized ultrasonic detection technology has been increasingly used as a live detection method for locating partial discharges inside transformer. It has the advantages of strong penetrating ability and fast response, but still has the problems of poor detection sensitivity and low positioning accuracy when the atmosphere noises are high. Further, the sonography indicating the region of partial discharges measured by visualized ultrasonic technology is only the two-dimensional projection of the region of interest, and the precise spatial positioning of the partial discharge cannot be evaluated. This paper proposes an improved visualized ultrasonic technology for panoramic visualization of partial discharge voiceprint on power transformers. The refined power transformer model is established through 3-D laser scanning and virtual reconstruction technology. Ultrasonic sensors with center frequency of 100 kHz were used to collect ultrasonic signals presented on the surface of power transformer. A self-developed data extraction algorithm was used to obtain ultrasonic signal digital matrix from the original collected data. In the following, 2D to 3D geometric transformation is applied to perform digital ultrasonic data on the meshed grid of transformer model. Finally, the ultrasound image on the entire hologram can be displayed holographically, which can be used to indicate partial discharge defects that occur inside the transformer in the 3-D space. Experimental verification on a converter transformer with a capacity of 382MVA confirmed that using this new visualization method, the location of the ultrasonic hot spots has a higher accuracy.