Related Experiment Video
Updated: Sep 30, 2025
11:14
Two-way Valorization of Blast Furnace Slag: Synthesis of Precipitated Calcium Carbonate and Zeolitic Heavy Metal Adsorbent
Published on: February 21, 2017
12.5K
Method for quantifying the reaction degree of slag in alkali-activated cements using deep learning-based electron
Priscilla Teck1,2, Ruben Snellings1, Jan Elsen2
1VITO, Mol, Belgium.
Journal of Microscopy
|March 10, 2022
Summary
This study introduces a novel neural network image analysis method to measure ground granulated blast furnace slag (GGBFS) reaction in alkali-activated cements. The AI model accurately quantifies GGBFS reaction, offering a robust alternative to traditional techniques.
Area of Science:
- Materials Science
- Civil Engineering
- Artificial Intelligence
Background:
- Alkali-activated cements utilize ground granulated blast furnace slag (GGBFS) as a binder.
- Accurate measurement of GGBFS reaction degree is crucial for understanding cement performance.
- Conventional methods for material characterization can be time-consuming and less robust.
Purpose of the Study:
- To develop and validate a novel methodology for quantifying the reaction degree of GGBFS in alkali-activated cements.
- To leverage deep learning-based image analysis for enhanced material characterization.
- To assess the generalizability and robustness of the developed AI model.
Main Methods:
- Implementation of a deep learning U-net architecture for segmentation of back-scattered electron (SEM-BSE) images.
- Application of the methodology to NaOH-activated slag cements and validation against X-ray Diffraction (XRD) results.
- Testing the broader applicability on NaOH-Na2SO4-activated systems.
Main Results:
- The neural network-based image analysis demonstrated strong correlation with independent XRD measurements.
- The trained model achieved fast and accurate image segmentation.
- The model showed good generalizability, being readily applicable to different activation systems.
Conclusions:
- The developed AI-driven image analysis methodology provides an accurate, replicable, and transferable tool for GGBFS reaction degree measurement.
- This approach offers improved performance and robustness compared to conventional threshold-based image segmentation.
- The methodology shows significant promise for advanced material analysis and characterization in cementitious materials.
