Correlation coefficients of vibration signals and machine learning algorithm for structural damage assessment in beams under moving load

Abstract

This paper presents a novel method of assessing structural damage in beams exposed to moving loads via acceleration signals through experimental studies. In this study, beams are supported on both ends, and their dynamic response to moving loads is assessed. The raw signal has been improved using a random decrement technique. Take measurements from different locations and calculate correlation coefficients between them, then use these as features to evaluate the structure. In order to create a reliable and potential framework for predicting damage efficiently, these features are used as input variables to the machine learning model. The proposed methodology exhibits promising results in accurately discerning and predicting damage in beam structure. It demonstrates a high level of precision to subtle changes in structural integrity when trained by machine learning on the statistical feature extracted from acceleration signals. As a result of this research, methods for detecting structural damage can be made more reliable and efficient by employing machine learning techniques. Additionally, structures operating in dynamic environments can benefit significantly from the proposed methodology.

Keywords

• beam structures
• correlation coefficient
• machine learning
• artificial neural network
• structural health monitoring