A timber health monitoring system using piezoelectric sensor for civil engineering applications using artificial intelligence

Summary

No reviewed summary yet. Read the abstract below or open the original publication.

Abstract

Timber is widely used in civil engineering, and its structural condition greatly influences overall safety and performance. Over time, wood is affected by factors such as termite attack, microbial decay, cracking, and mechanical or environmental stress, making damage detection difficult. To address this challenge, a Timber Health Monitoring Technology is proposed, using an AI-driven system that analyzes data from piezoelectric sensors attached to wooden elements. These sensors enable continuous, real-time monitoring by capturing signals that represent the dynamic behavior of the timber. Piezoelectric active sensing is first used to collect vibration-based waveform data, which reflect changes in structural integrity. Multiscale attention entropy (MAE) is then applied to extract quantitative features that serve as indicators of wood health. These features are fed into a Capsule Network (CapsNet) enhanced with data augmentation to classify different damage states with high accuracy. Basic validation is performed through timber damage modeling and vibration testing. The results demonstrate that the proposed method significantly improves detection performance and provides reliable, intelligent monitoring for timber structures. Principal component analysis was also explored for dimensionality reduction, but only minor decreases in classification accuracy were observed, indicating strong robustness of the system.

Abstract as supplied by the source. Full text may be restricted by the publisher.