A Data-Driven Soil Mix Design for Rammed Earth Based on Experimental Particle Size Distribution

Rammed earth construction, a traditional and sustainable building technique, strongly depends on soil properties—particularly particle size distribution (PSD) and plasticity—for its mechanical performance and durability. Ensuring optimal granulometric composition remains a key challenge for achieving consistent structural quality using locally available materials.
This study presents a practical framework for optimizing soil mixtures for rammed earth construction by combining experimental analysis and computational modeling. Particle size distribution tests were conducted on locally sourced soils, including clay, sand, and gravel. Based on these data, a granulometric optimization algorithm was developed and implemented in Python to determine optimal mixing proportions that conform to recommended grading curves. The tool takes the particle size analysis data as input to provide accurate ratios to ensure that the final mixture fits perfectly within the reference range. It also provides graphical validation showing the optimal curve for the actual mixture.
The results show that soil mixtures in compliance with recommended grading curves exhibit an increase in compressive strength often exceeding 50% compared to non optimized soil.
The proposed framework show that soil mixtures in compliance with recommended grading curves exhibit an increase in compressive strength often exceeding 50% compared to non optimized soil offering a reproducible and efficient approach for designing high-performance earthen materials using local resources.