Prediction of Crack Path in Reinforced Concrete using Acoustic Emission Analysis

Abstract

This paper presents a probability-based approach for predicting crack propagation in lightly reinforced concrete beams using acoustic emission techniques. A novel methodology combining Gaussian Mixture Model clustering and spatial binning is employed to simulate the crack path and plane. Experiments were conducted on notched lightly reinforced concrete beams under flexural loading, with AE data continuously recorded. GMM clustering was used to categorize AE events based on their average frequency and rise angle, enabling the selection of mode I crack events based on probability threshold. Then the filtered AE data was processed using a spatial and temporal binning strategy to simulate the evolving crack path. The predicted crack evolution in terms of path and length was validated against the results obtained from digital image correlation, demonstrating good agreement. This research highlights the effectiveness of acoustic emission technology, combining probabilistic clustering with spatial and temporal binning of AE data for real-time crack path prediction in reinforced concrete structures, offering valuable insights for structural health monitoring and damage assessment.