Artificial intelligence supported optimization of piled raft foundations based on three-dimensional finite element analyses

The growing demand for high-rise building designs has necessitated optimizing piled raft foundations using innovative methods for efficient and safe design. Multiple parametric three-dimensional finite element (3D FE) analyses were conducted using high-capacity workstations to optimize various parameters of piled raft foundations in this study. Data from a structure located in Frankfurt (Germany) were used to verify the simulations. Following the verification process, ninety distinct combinations of 3D FE analyses were performed by systematically varying the piles' diameter, number, and elastic modulus within the parametric framework of the piled raft foundation. The impact of changes in the diameter, number, and elastic modulus of the piles on the settlement was analyzed. An optimization process was conducted using a novel artificial intelligence-aided Goal Attainment technique, which generates robust regression functions based on data obtained from 3D FE analysis of the piled raft foundation system. A balance was achieved between the parameters of the raft foundation system and reliable settlements in the multi-objective optimization to ensure the model's economical and safe design. The pile diameter was determined to be 1.1 m, the number of piles to be 30, and the modulus of elasticity to be 33.18 GPa as a result of the optimization. Although the settlement value obtained with these parameters equals the base case, a more cost-effective solution system was developed. Consequently, a technique supported by artificial intelligence has been introduced that yields significant results for general piled raft foundation systems and identifies optimum design parameters to guide future research. The method developed in this study has the potential to support engineering practice in real-world projects by enabling a more efficient and reliable determination of pile dimensions and configurations in foundation systems.