TY - JOUR AU - Ray, Rahul AU - Samui, Pijush AU - Roy, Lal Bahadur PY - 2023/02/04 Y2 - 2024/03/29 TI - Reliability analysis of a shallow foundation on clayey soil based on settlement criteria JF - Journal of Current Science and Technology JA - JCST VL - 13 IS - 1 SE - Research Article DO - UR - https://ph04.tci-thaijo.org/index.php/JCST/article/view/208 SP - 91-106 AB - <p>Soil is a heterogeneous medium and the involvement of its many effective attributes in geotechnical behaviour for soil-foundation system makes the prediction of settlement of shallow foundation on soil a complex engineering problem. As the understanding about the soils are improving, the variability in soil attributes is taken into consideration. As result, the present research approach has also shifted from deterministic to probabilistic approach. The present paper describes the application of two probabilistic based soft computing techniques i.e. Adaptive Network based Fuzzy Inference System (ANFIS) and Functional Network (FN) to study the shallow foundation reliability based on settlement criteria. These models are simple and reliable and can be used for routine design practice. In addition, FN and ANFIS were tested to find their adoptability for shallow foundation settlement prediction considering different soil attributes. Models performance was tested based on different fitness parameters i. e. RMSE, VAF, RSR, β, etc. Functional network (FN) model outperformed in terms of all fitness parameters (RMSE=0.0017, VAF=98.512, RSR=0.1416, NS=0.979, RPD=7.062) as compared to ANFIS (RMSE=0.0026, VAF=95.687, RSR=0.2148, NS=0.953, RPD=4.655). The results show that FN approach can be used as a reliable soft computing technique for non-linear problems like settlement of shallow foundations on soils.</p> ER -