Coupled Finite Element and Artificial Neural Network Analysis of Interfering Strip Footings in Saturated Cohesive Soils

This study uses numerical analysis to investigate the behavior of the footing under various variables, such as footing spacing, depth, soil undrained shear strength, and groundwater table levels. The use of artificial neural network (ANN) predictions to estimate settlement behavior for each configuration was a unique component of the research. The results showed the importance of soil cohesion and footing depth ratio on interference effects between closely spaced footings. The observation across different cohesion values is that increased footing depth and elevated groundwater tables reduce the required spacing to mitigate interference. The ultimate bearing capacity (UBC) of interfering footings diminishes as the spacing-to-footing-width ratio (S/B) grows until it is equal to that of an isolated footing at higher spacings. At a S/B ratio of 1, the UBC of two footings equals to that of an isolated footing and stays constant as the S/B ratio grows. Furthermore, deeper footings are connected with higher UBC. The incorporation of ANN predictions into the analysis improves settlement estimation and provides a methodological gain in evaluating interfering strip footings in saturated cohesive soils. The impressive RMSE value of 3.6% observed in the ANN model assessment strengthens the dependability of the results, emphasizing the significance of this technique in engineering practice.