Application of Multiscale Increment Entropy and InceptionTime Model for Structural Health Monitoring
Chia-Ju Lin, Ahmed Abdalfatah Saddek, Tzu-Kang Lin, Y. Lin, Clive Chin-Kang Shen
Aging civil structures are increasingly vulnerable to environmental degradation and natural hazards, highlighting the need for reliable and automated structural health monitoring (SHM) systems. This study proposes a novel SHM framework that integrates Multiscale Increment Entropy (MIE) with the InceptionTime deep convolutional neural network (CNN) to detect and localize structural damage with high accuracy. The MIE technique is employed to analyze the velocity responses of structures under ambient excitation, providing robust and scale-independent entropy features that capture nonlinear and nonstationary characteristics of structural behavior. These entropy-based features are input into the InceptionTime model for automated damage classification and localization. To validate the performance of the system, both numerical simulations and laboratory-scale experiments were conducted using a seven-story steel frame benchmark structure. The proposed method achieved 99.78% accuracy in numerical simulations and 94.21% accuracy in experimental verification, demonstrating strong consistency, robustness, and generalization capability. The integration of MIE with InceptionTime effectively enhances the interpretability and reliability of entropy-based damage assessment, offering a scalable, data-driven approach for real-time SHM applications in complex engineering systems.