MEMS Data-Driven Intelligent Identification of Rotation-Angle Response and Shear Band Position in Gravelly Soil
Di Wu1, Yongzhe Feng1, Gurong Yao1
1School of Architecture and Transportation Engineering, Guilin University of Electronic Technology, Guilin 541004, China.
Sensors (Basel, Switzerland)
|August 13, 2026
Summary
This study introduces a novel MEMS sensor framework to predict internal rotation angles and locate shear bands in gravelly soil slopes. This data-driven approach enhances early warning systems for progressive slope failure.
Area of Science:
- Geotechnical Engineering
- Geomechanics
- Civil Engineering
Background:
- Conventional slope monitoring methods struggle to detect internal shear band evolution and localized deformation.
- Early identification of shear bands is crucial for recognizing precursors to progressive failure in gravelly soil slopes.
- Existing techniques are limited in capturing the spatial dynamics of internal slope instability.
Purpose of the Study:
- To develop and validate a Micro-Electro-Mechanical Systems (MEMS) data-driven framework for predicting spatial rotation-angle responses.
- To enable the localization of potential shear bands within gravelly soil slopes.
- To improve the perception of internal shear deformation and detect zones of potential instability for enhanced early warning.
Main Methods:
- Conducted scaled laboratory model tests with embedded MEMS sensors to measure cumulative rotation-angle responses during shear band formation.
- Developed a Discrete Element Method (DEM) model incorporating particle morphology to analyze rotation-angle field differentiation and validate experimental results.
- Established a shear band localization framework using PDL-GAN-based data augmentation and PCA + Gaussian regional rotation-angle field prediction.
Main Results:
- DEM simulations showed good agreement with laboratory test results for cumulative rotation-angle responses (MAPE < 9.79%).
- Observed distinct rotation-angle patterns within shear bands: negative accumulation at upper points and positive at lower points.
- The PCA + Gaussian model accurately predicted regional rotation angles (MAE=0.0803, RMSE=0.1024) and preserved dominant deformation modes.
Conclusions:
- The proposed MEMS data-driven framework effectively predicts internal rotation-angle responses in gravelly soil slopes.
- Potential shear band locations can be indirectly identified through analysis of predicted rotation-angle fields.
- This approach provides crucial data-driven support for precursor recognition and intelligent early warning systems for slope failure.
