Z. Zhang, X. Wang, H. Pan, and Z. Lin
North Dakota State University, United States
Keywords: data analytics, machine learning, structural health monitoring, corrosion-induced damageCorrosion responds for the huge maintenance cost of the nationwide civil structures, including highway bridges and pipelines. Although there are many structural health monitoring techniques available in the market, a large amount of sensory data set with high uncertainty post great challenges in timely information fusion. In this study, we explored a machine learning approach to extract information from the sensory data for early-age corrosion-induced damage identification and classification. Acoustic signals of steel samples collected from simulated corrosion damage were used for model training and calibration. The results showed that the machine learning method allowed effective information fusion for early-age corrosion. Moreover, different noise levels that are often experienced in the actual structures were discussed to evaluate the effectiveness of the methods.