Investigation of vacuum arc modes in 252 kV single-break vacuum interrupters based on deep learning image recognition
Siyuan Liu, Xianglai Bo, J. Chen, Zhiyuan Liu, Yingsan Geng
Abstract
Replacing traditional SF 6 circuit breakers with environmentally friendly 252 kV single-break vacuum circuit breakers effectively reduce SF 6 greenhouse gas emissions in transmission systems. Understanding the evolution of vacuum arc modes under different conditions is crucial for analyzing 252 kV single-break vacuum interruption. This paper built a test platform using a detachable vacuum chamber and a motor-driven operating mechanism to investigate the vacuum arc modes of a 252 kV single-break vacuum interrupter with 2/3-turn coil-type axial magnetic field contacts under varying arc currents. The arc modes were classified, and a deep learning network was developed for arc mode image recognition. The results revealed the evolution characteristics of vacuum arc modes at a hundred-millimeter range contact gap and demonstrated the influence of arc current from 10 kA to 45 kA on the duration of arc modes. • 252 kV single-break vacuum arc modes are investigated with 2/3-turn coil-type axial magnetic field contacts at a hundred-millimeter range contact gap. • Experimental tests investigated the 252 kV single-break vacuum arc. Then, a deep learning network was developed for arc mode recognition and classification. • Observation experiments revealed the evolution and transient process of the vacuum arc modes. • The test results demonstrated the influence of arc current from 10 kA to 45 kA on the duration of arc modes.