Assessing the Coefficients of Porosity-to-Binder Index Formulations for Stabilized Clay Through Automated Calibration...

Area of Science:

• Geotechnical Engineering
• Materials Science

Background:

• The porosity-to-cement relationship is crucial for predicting stabilized soil properties since 2007.
• Traditional parameter calibration relies on manual regression, which can be time-consuming and less accurate.

Purpose of the Study:

• To develop an automated methodology for calibrating the porosity-binder index parameters (A, B, x).
• To enhance the predictive accuracy of mechanical strength, durability, expansion, and stiffness in stabilized soils.

Main Methods:

• An iterative optimization framework using the sum of absolute errors (SAE) and a Monte Carlo search algorithm was employed.
• The methodology was applied to cement-stabilized clay with ground glass (GG), recycled gypsum (GY), and limestone residues (CLW).
• Two calibration strategies were compared, with the second incorporating all mixtures and both unconfined compressive strength (q_u) and initial shear stiffness (Go).

Main Results:

• The automated calibration, particularly Strategy 2, yielded robust and physically consistent parameters.
• High coefficients of determination were achieved: 0.9318 for q_u and 0.9412 for Go.
• The algorithm-driven framework demonstrated improved predictive capabilities.

Conclusions:

• The proposed automated calibration methodology offers a systematic and improved approach for determining porosity-binder relationship parameters.
• This enhances the reliability of predicting the performance of stabilized soils in various engineering applications.