Settlement estimation of the piles socketed into rock employing hybrid ANFIS systems

In civil engineering applications, piles, categorized as deep foundations, provide solid support for buildings by being pushed into the soil. The careful evaluation of the settling of foundations is crucial during the design phase due to their significant load-bearing capability. Therefore, the management and evaluation of settlement present a significant challenge within the piling design and construction domain. The novelty of this study could be combining the Adaptive Neuro-Fuzzy Inference System (ANFIS) with the Equilibrium Optimizer (EO), the Black Widow Optimization Algorithm (BWOA), and Particle Swarm Optimization (PSO), specific application to rock settlement prediction, use of local data, selection of key input parameters, and practical implications. The findings suggest that all ANFE, ANFB, and ANFP have significant promise in properly forecasting the pile settlement (SP). Uncertainty analysis depicts a better performance of the ANFE compared to ANFB by gaining 0.4916 lower than 0.7485 in the train part and 0.7099 smaller than 0.94 in the test part. It is important to note that deleting the UCS variable from the input category leads R² to decrease and RRSE, MAE, and U₉₅ to increase.