Integrating the Capsule-like Smart Aggregate-Based EMI Technique with Deep Learning for Stress Assessment in Concrete

Area of Science:

• Structural Health Monitoring
• Deep Learning Applications
• Materials Science

Background:

• Concrete structures require reliable stress monitoring for safety and maintenance.
• Existing methods may lack precision or robustness in real-world conditions.
• Smart aggregate sensors offer a promising approach for embedded structural monitoring.

Purpose of the Study:

• To develop and validate a deep learning-based method for concrete stress monitoring using electromechanical impedance (EMI) data.
• To assess the performance of a 1D Convolutional Neural Network (CNN) model for stress estimation.
• To evaluate the robustness of the proposed method against noise and untrained stress conditions.

Main Methods:

• A capsule-like smart aggregate (CSA) sensor prototype was developed for EMI measurements.
• A 2 degrees of freedom (2 DOFs) EMI model was established for the CSA sensor embedded in concrete.
• A 1D CNN deep regression model was designed to process raw EMI signals.
• Experimental tests were conducted on CSA-embedded concrete cylinders under varying compressive loads.

Main Results:

• The 1D CNN model effectively estimated concrete stress from raw EMI signals.
• The method demonstrated feasibility and robustness in handling noise-contaminated EMI data.
• The model showed promising performance even for untrained stress levels.

Conclusions:

• The proposed 1D CNN deep learning approach provides an effective and robust method for concrete stress monitoring using CSA-based EMI signals.
• This technique holds potential for real-time structural health monitoring of concrete structures.
• Further research can explore advanced deep learning architectures and sensor integration.