[22p-C105-1~17] 1.6 Ultrasonics

[22p-C105-17] Damage evaluation of fixed beams at both ends based on unsupervised machine learning using responses of impedance-loaded surface acoustic wave sensor

Naoki Hashimoto¹, Shinji Baba¹, 〇Jun Kondoh¹ (1.Shizuoka Univ.)

Keywords:surface acoustic wave, unsupervised machine learning

Damage evaluation of fixed beams at both ends was carried out using an impedance-loaded surface acoustic wave sensor for bridge health monitoring. The sensor responses were analyzed using a continuous Wavelet transform and the attenuation coefficient was obtained for each frequency. Then, the coefficients were applied to an unsupervised machine learning method. For unsupervised machine learnings, k-means and Gaussian Mixture Model (GMM) were used. The evaluation results indicate that the GMM is the effective method.