The high-voltage disconnector is one of the important electrical equipment in the power system, and its main function is to form an obvious disconnection point in the circuit. The traditional switch state recognition uses manual patrol, which has low work efficiency, large misjudgment probability, and increases the personal risk of staff. In order to improve the reliability of identification and promote the intelligent development of the power grid, this paper improves the conventional static contact structure to a visual static contact, and uses the deep learning algorithm to identify the operating state of the disconnector with the intelligent sensing binocular terminal, and designs a highly reliable smart disconnector. It realizes many functions such as image recognition, contact temperature measurement, abnormal state alarm, etc., and completes the intelligent state identification and remote monitoring of the disconnector through three detection methods: image recognition, auxiliary switch and microswitch. According to the GB/T 1985-2014 and DL/T 486-2010, the type test was carried out. The results show that the design scheme of smart disconnector meets the standard requirements, and the dynamic thermal stability and electromagnetic compatibility are good. Through the actual network running test, the ability of the smart disconnector to accurately identify and operate for a long time under actual working conditions is verified, indicating that it has engineering application value.

Smart disconnector; Intelligent sensing binocular terminal; Image recognition; Type test