The subgrade provides a foundation for supporting the pavement structure. The sub grade whether in cut or fill should be well compacted to utilize its full strength and to economize thereby on the overall thickness of pavement required. For design, the sub grade strength is assessed in terms of the CBR of the sub grade soil in both fill and cut sections. For determining the CBR value, the static penetration test procedure should be strictly adhered to. This is described in IS: 2720 (part 16) “Methods of test for soils laboratory determination of CBR”. The test must always be performed on molded samples of soils in the laboratory. CBR test is laborious and time consuming; but sometimes the results are not accurate due to the poor laboratory conditions. Further if the available soil is of poor quality, suitable additives are mixed with soil and the resulting strength of the soil will be assessed by CBR value, which is cumbersome. Hence a method is proposed for correlating CBR values with the liquid limit, plastic limit, plasticity index, OMC, Maximum dry density, UCC values of various soils taken from in and around three different districts in Tamil Nadu namely Thanjavur, Tiruchirapalli and Pudukkottai districts using Artificial Neural Network. The results of these analyses are compared with experimental results. Multiple linear regression based model for CBR prediction performs better than neural network based model in the present study.
CBR, UCC, OMC, LL, PL, PI, Artificial Neural Network, Multi Linear Regression
