Vol. 52 No. 4 December 2021
Prediction of California Bearing Ratio (CBR) from Index Properties of Fine-Grained Soil
Arunav Chakraborty and Anasuya Goswami
ABSTRACT: The California Bearing Ratio (CBR) value is an important variable in pavement design since it determines the strength of the subgrade soils. However, it should be noted that the CBR test is arduous and time-consuming. As a result, this work attempts to establish relationships between CBR and several soil index parameters such as liquid limit (LL), plastic limit (PL), optimum moisture content (OMC), and maximum dry density (MDD). Regression analysis and neural networks were used to develop three prediction models for correlating soaked CBR values with LL, PL, OMC, and MDD for soil samples taken from different locations in Guwahati, Assam, India. Because Assam is prone to flooding, and some rural roads are inundated for two to three days under water, a soaked CBR is considered. According to the results, ANN can more accurately predict soaked CBR values.
KEYWORDS: California bearing ratio, Artificial neural network, Multiple regression, Optimum moisture content, Maximum dry density.
