Random forest and counter propagation neural network model used to study the consequence of nano silica on the mechanical performance of high strength concrete at elevated temperature

In the building industry, the usage of high strength concrete (HSC) has increased dramatically in the last several years. High strength and remarkable durability are only a few of the numerous attractive advantages of HSC. Present research presents the mechanical and durability characteristics of the HSC developed with cement and nano silica (NS). The NS added to the cement at 1.34%, 2.67%, 4%, 5.34%, 6.66%, 8% and 9.34% respectively. The specimens were developed with addition of the 5 Kg/m3, 10 Kg/m3, 15 Kg/m3, 20 Kg/m3, 25 Kg/m3, 30 Kg/m3 and 35 Kg/m3 of NS. The specimens are also subjected to the elevated temperature ranging 27 0C, 100 0C, 300 0C, 500 0C and 700 0C. The counter propagation neural network (CPNN) and random forest (RF) have been used to predict the mechanical properties of the concrete specimens. The maximum compressive strength was obtained for the specimen NS25 of 51.59 MPa. The compressive strength increases as the temperature increased from 27 0C to 100 0C beyond that the compressive strength decreased. The durability properties of the concrete were tested using water absorption and rapid chloride penetration test (RCPT). The specimen NS25 shows the maximum strength and less water absorption and RCPT values. The prediction of the mechanical properties has been also analysed with CPNN and RF. The CPNN performed well during training, while the RF provided better prediction results compared to the CPNN during testing.