Image based structural performance prediction for the reinforced concrete interior beam-column joint

• Main Author: Manivannan Barath
• Other Authors: Li Bing
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2023
• Subjects: Engineering::Civil engineering
• Online Access: https://hdl.handle.net/10356/172604
• Institution: Nanyang Technological University

This study employs computer vision technology to relate superficial crack patterns in RC interior Beam-Column Joints (BCJ) to damage levels, aiming to augment a broader database for reinforced concrete beams. Over the past four decades, extensive research has focused on assessing the seismic behavior of RC interior BCJs with varying aspect ratios. Loading-induced cracks significantly affect structural strength, ductility, and stiffness, making accurate assessment crucial. The study combines image processing and machine learning to recognize structural damage and estimate performance degradation in RC BCJs. The study's broader goal is to enhance current site inspections by considering the feasibility of incorporating computer vision technology. Using MATLAB R2023b, images are segmented and converted into a consistent binary format, then proceeds to capture crack information from image detection, including crack width and compressive strength, which are used as input for machine learning. A triad predictor-responses are established, and MATLAB models are trained to assess accuracy under various predictors and required responses. Characteristic response values are linked to hysteretic lateral force-drift ratio curves, along with experimental details, design parameters, and extracted features. The comprehensive database consists of 323 crack images from 155 specimens, based on prior experimental research, and features drift ratios, failure modes, geometric properties, and experiment details. Regression analysis results indicate high predictive accuracy, suggesting the potential application of image-based damage recognition in visual structural inspections of RC BCJs.