Statistical Analysis of Soil Parameters Affecting the Bearing Capacity and Settlement Behaviour of Gravel Soils

Understanding the geotechnical behavior of soils is fundamental for the safe design and performance of engineering structures. This study investigates the bearing capacity and settlement behavior of gravel soils using extensive field and laboratory data collected from 27 boreholes in Düzce, northwest Türkiye. Key soil parameters, including excavation depth, groundwater levels, unit weight, water content, particle size distribution, internal friction angles, and cohesion coefficients, were analyzed. Correlation analysis revealed a significant relationship between bearing capacity and the No. 200 sieve value, while relationships with other parameters were less pronounced. Predictive models for bearing capacity and settlement were developed using multiple linear regression, quadratic, and cubic models. The cubic model demonstrated the highest accuracy, predicting bearing capacity with an R2 of 0.79 and settlement with an R2 of 0.904. These results underscore the potential of advanced statistical models to reliably estimate geotechnical properties based on measurable soil parameters. The findings emphasize the importance of combining field and laboratory analyses with robust statistical approaches to accurately predict soil behavior. This methodology provides a foundation for improving geotechnical design practices and may be extended to other soil types and regions, thereby enhancing the reliability and efficiency of soil parameter estimation for engineering applications.