A Wireless Acoustic Emission Sensor System with ACMD-IGWO-XGBoost Algorithm for Living Tree Moisture Content...

Area of Science:

• Forestry science
• Ecology
• Sensor networks

Background:

• Trunk water content significantly impacts tree metabolism, forest ecosystems, and the global moisture cycle.
• Accurate, real-time moisture content (MC) measurement is crucial for tree cultivation and forest management.

Purpose of the Study:

• To design and implement a non-destructive system for diagnosing water content in living tree trunks.
• To develop a robust MC prediction model using advanced signal processing and optimization techniques.

Main Methods:

• A wireless acoustic emission sensor network (WASN) was developed for high-speed sampling of trunk epidermis acoustic emission (AE) signals.
• Adaptive chirp mode decomposition (ACMD) was used for optimal characteristic wavelet sequence decomposition.
• An improved grey wolf optimizer (IGWO) optimized XGBoost model was established for MC prediction, enhanced by multi-strategy joint optimization.

Main Results:

• The developed system demonstrated high diagnostic accuracy, reaching an average of 96.75% across field tests.
• The system proved effective in monitoring moisture content in various tree species, including Robinia Pseudoacacia, Photinia serrulata, Pinus massoniana, and Toona sinensis.

Conclusions:

• The wireless acoustic emission sensor network provides an effective non-destructive method for diagnosing tree trunk moisture content.
• The system exhibits excellent applicability across diverse working conditions and tree species, supporting sustainable forest management.