Implementation of Data-Driven Approaches for Condition Assessment of Structures and Analyzing Complex Data

Abstract

Structural components are exposed to external or/and internal degradation agents and damage during service life, which may endanger their structural integrity. Physical and numerical modeling of some structures is not always feasible and efficient due to structural complexity, computational efficiency, and the challenge associated with model verification and validation. Nondestructive testing and structural health monitoring approaches such as acoustic emission has become favorable due to recent advances in digital sensing and sensors. Acoustic emission is a passive structural health monitoring method, which is very sensitive to small-scale (e.g., micro-scale) surface vibrations of structures caused by internal or external damage. Acoustic emission data include signals or parametric features. Management, analyzing, understanding, and interpreting data are sometimes challenging due to the data complexity attributed to structural detailing, monitoring duration, and the temporal evolution of acoustic property of material. This chapter introduces some data-driven approaches such as machine learning and deep learning algorithms in combination with signal processing methods and their application to analyze and interpret acoustic emission time-series data for different structures and materials.