Forecasting Pile Bearing Capacity Using an Innovative RES-Based Approach

In situ tests for determining the bearing capacity of pile (Q_u) in geotechnical and mining projects are integral to the design and implementation of pile-based structures. The selection and execution of in situ tests depend on the expertise of individuals in the field of piling, considering specific geological conditions and operational costs. The pile driving analyzer (PDA) test is widely regarded as the most reliable method for assessing pile behavior. However, predicting Q_u in every case study proves challenging due to the associated high costs, time requirements, delays in mobilization of equipment, variability due to the choice of PDA, and the need to provide necessary testing conditions. To overcome these challenges, an innovative approach based on the Rock Engineering System (RES) has been introduced, allowing for the precise prediction of Q_u without incurring high costs. This research showcases a dataset derived from 50 PDA tests executed on precast concrete piles in Pekanbaru, Indonesia. Employing five key input parameters—drop height, pile set, hammer weight, length pile, and pile cross-sectional area—the study utilized Q_u as the output parameter. The efficacy of the RES method was subsequently assessed through a comparison with polynomial, linear, logarithmic, exponential, and power regression models. The findings unequivocally reveal the superiority of the RES-based method over alternative regression techniques in predicting Q_u, boasting a superior coefficient of determination (R²) and a diminished mean squared error.