Enhancing Concrete Damage Detection through Ultrasonic Rebound Measurements and Deep Learning Techniques

10.32738/JEPPM-2024-0030

Suzheng Zhao

Concrete damage detection ； deep learning ； ultrasonic rebound ； improved algorithm

Journal of Engineering, Project, and Production Management

14卷3期（2024 / 09 / 01）

1 - 11

英文

Due to the construction industry's rapid growth, concrete is now the standard building material used in new construction. In recent days, the development of the construction industry has focused heavily on how to maintain and identify the concrete structure of some older buildings. However, conventional concrete damage identification lacks precision and accuracy. Therefore, this work suggests an ultrasonic rebound enhancement approach based on deep learning. In order to detect concrete damage, the new algorithmic model measures the concrete data using the ultrasonic rebound technique and then analyses it using deep learning. By utilizing the improved Genetic Particle Swarm Optimization algorithm (GA-PSO) combined with the Back Propagation Neural Network, an improved GA-PSO-BP long-term concrete ultrasonic rebound comprehensive strength measurement model was established. The experimental results show that the improved GA-PSO-BP algorithm has a lower root mean square error than the conventional algorithm, which is lower than the Back Propagation neural network 0.008 and lower than the Genetic Algorithm-Back Propagation algorithm 0.001, and a higher accuracy rate than the Back Propagation algorithm model 0.07 and higher than the Genetic Algorithm-Back Propagation algorithm model 0.02. As a result, the improved GA-PSO-BP algorithm performs more precisely and accurately than the conventional approach. This study has practical applications for practitioners in the construction industry. The high accuracy and precision of the new algorithm render it an effective tool for identifying structural damage in old concrete buildings, which provides more reliable data support for maintenance efforts. This not only generates novel methods for enhancing concrete damage detection but also contributes to the sustainable development of the construction industry.