Predicting Standard Penetration Test N-value from Cone Penetration Test Data Using Gene Expression Programming

Standard Penetration Test (SPT) and the Cone Penetration Test (CPT) are employed in-situ to evaluate soil parameters. In geotechnical engineering practice, engineers often conduct in-situ tests either SPT or CPT to delineate soil profile and evaluate soil parameters for bearing capacity analysis. Most of the geotechnical parameters are correlated with SPT instead and widely employed. Since numerous soil parameters are correlated with SPT N-values, it is very beneficial to establish a correlation between CPT data and SPT N-values. To predict the SPT-N value from CPT data across various soil types such as silty sand, sandy silt, silty clay, and lean clay, this study has developed an empirical model using gene expression programming (GEP). Also a comprehensive GEP model encompassing all soil types has been proposed. The input parameter used in the GEP models are CPT tip resistance (q_c), CPT-Sleeve friction (q_f), and effective overburden pressure (σ_vʹ). The effectiveness of the models is evaluated through the implementation of statistical tests, employing a comprehensive index OBJ, and performing parametric analysis. Moreover, to test the reliability of the proposed GEP models, CPT-SPT data pairs that were not utilized in the model generation were employed. The results of the proposed models testing indicated that the models either under-predicts the targeted value by 3–9% or over-predicts by 3–12%. The OBJ values indicate that silty clay has the highest value of 4.985, making it the weakest model, while the all-soil model achieved the lowest value of 1.656, thus being considered the most effective model. The results indicated that the suggested models are precise and exhibit a strong potential for generalization and prediction.