Gene expression programming model to predict driven pipe piles set-up

Abstract

In this paper, Gene Expression Programming (GEP) is developed to predict the increase in pipe piles axial capacity with time known as set-up. Pile set-up is significant and continues to increase with time after installation. Analysis of the collected data identified pipe pile diameter, time after installation, soil type, effective vertical effective stress at the pile tip and the initial axial capacity as playing a substantial role in predicting pile set-up. The results indicated that GEP could predict the pipe pile set-up with a remarkable accuracy (R 2 = 0.95, MAE = 377.5 and RMSE = 484.25) for the training data-set. The developed GEP model is provided in the pattern of expression tree and transformed into a mathematical equation. A statistical performance comparison is presented between the developed GEP model and previously published empirical formulas. Based on the results, the GEP model is substantially superior the traditional empirical relationships and has been identified to provide better solutions to predict set-up.