Forecasting bearing capacity of the mixed soil using artificial neural networking

Abstract

The bearing capacity of soil changes owing to the mechanical properties of the soil and it influences the structural stability. In most of the geotechnical engineering projects, there are several soil mechanic experiments, that need interpretation before application. The mechanical properties of soil interaction make the prediction of soil bearing capacity complex. However, the enhancement of construction project safety needs the interpretation of soil experiments and design results for proper application in a geotechnical engineering project. In this study, artificial neural network is proposed for the evaluation of the mixed soil characteristics to forecast the safe bearing capacity of soil due to the mechanical properties of the soil interaction phenomenon. The results for prediction of the safe bearing capacity reveal that the R 2 and RMSE for all mechanical properties effects on safe bearing capacity are 0.98 and 0.02, these values can provide a suitable accuracy for the prediction of the safe bearing capacity of the mixed soil. The higher inaccuracy is obtained when only the influence of single mechanical property on the mixed soil is considered in the prediction of the safe bearing capacity. This study supports the enhancement of geotechnical engineering design quality through the prediction of safe bearing capacity from characterized mechanical properties of the soil.

Sustainable Cement Stabilization of Plastic Clay Using Ground Municipal Solid Waste: Enhancing Soil Properties for Geotechnical Applications

“…For the prediction of the geotechnical engineering characteristics of the soil mixture, R 2 and RMSE provided appropriate measures of the precision of prediction for the safe bearing capacity of the mixed soil results. The cohesion of the mixed soil plays a main role in predicting the safe bearing capacity of the soil mixture [59]. In addition, the linear regression results show how the appropriate soil mixing designs are accompanied by selecting the proper width of a concrete footing, resulting in the development of a safe concrete footing design [60].…”

“…In addition, random data were selected from all produced data for the training to have a reliable prediction assessment. ANNs and linear regression analysis are two methods that have been applied to prediction generation in geotechnical engineering with respect to the mixing soil method [59,60]. The data training and reproduction phases are the two main parts of ANNs for the prediction.…”

“…ANNs [59] and regression analysis [60] have been adopted to predict the mixing soil method results to solve the problem of soil weakness. There is no universal method for explaining ANN architectural design to predict a phenomenon in geotechnical engineering.…”

The use of polyurethane foam-sand mixtures in sandy embankment design- predicting seismic response using FEM, catastrophe theory, B-spline method, and artificial neural networks

“…Many researchers have utilized ANNs to predict liquefaction [ 27 ], compaction properties of fine-grained soils [ 28 ], mixing soil [ 29 , 30 ], displacement at selected points of the clayey cover of the landfill model [ 4 ], soil thermal conductivity [ 31 ], and soil compaction [ 32 ]. In addition to geotechnical engineering, artificial neural networks have been applied in many fields of engineering [ 33 ], and to improve the accuracy of ANN predictions, genetic algorithms combined with ANNs [ 34 ] and deep learning combined with ANNs [ 35 ] have been applied.…”

Crack simulation for the cover of the landfill – A seismic design

“…Due to the nonlinearity of the seismic load, numerical simulation results need to be validated using statistical analysis. The current study used artificial neural networks to predict embankment displacement [11][12]26]. The prediction is made from the validation and optimization results of the numerical simulation [27][28][29].…”