Prediction of Ultimate Bearing Capacity of Eccentrically Inclined Loaded Strip Footing Resting Over Dense and Medium Dense Sand Using Generalized Regression Neural Network

Abstract

AbstractSometimes in remote areas, geotechnical testing becomes more complex which makes conventional methods more complicated and cumbersome. So to conserve more time and money, computer science has developed a “Neural network technique” which is a mimic of the biological neural network like ANN, GRNN, PNN, etc., to attain results in less time with more accuracy. In this study, the ultimate bearing capacity of eccentrically inclined loaded strip footing is predicted resting over dense and medium dense sand with the help of RF (Reduction factor) value. A model test results were utilized for modeling GRNN (Generalized regression neural network) model using DTREG software to predict this RF value using Embedment ratio, inclination ratio, eccentricity ratio as input parameters, and RF as output parameter using Gaussian type activation function. The results of experimentally calculated RF on the same study are compared with GRNN results and found more convenient and reasonable in terms of error minimization and accuracy. Also, ANN MATLAB results were also analyzed with the GRNN results in which no such variations were spotted.KeywordsUltimate bearing capacityReduction factorStrip footingGRNN