2021 年 141 巻 1 号 p. 70-77
Overheating due to the screw slackening of a circuit breaker, an electromagnetic switch, a terminal, or a plug-receptacle in a distribution panel or an electric apparatus is one of the main causes of fire accidents originating from electrical equipment in industrial and domestic environments. In this study, we propose a novel screw slackening sensing method for a commercially available 100V blanch circuit breaker. Attention is paid to the conducted emission spectrum obtained during discharge, which occasionally occurs in a small gap between the slackened connected conductors. An investigation of the screw slackening is conducted based on a convolutional neural network. Thus, the proposed approach serves as a potentially effective method for the prevention of fire accidents due to screw slackening.