관형 철탑 볼트 풀림 진단을 위한 음향방출 신호의 특징 분석

Abstract

We conducted an acoustic emission test for the loosened bolt diagnosis of tubular steel towers. Signal processing techniques and machine learning were applied to acoustic emission signals to confirm the classification possibility of the bolt fastening strength. Consequently, a clear difference between the fastened condition of the bolt and the loosened condition was observed; however, signals with different bolt fastening strengths were not classified. In this process, it was confirmed that the bolt fastening strength was classified up to 74.3N·m using a band-pass filter. In conclusion, we confirmed the performance possibility of the loosened bolt diagnosis through acoustic emission signals.