Prediction of mechanical properties of rubberised concrete exposed to elevated temperature using ANN

Considering scarcity of natural sand, waste rubber tyre can be an alternate ingredient for replacement of conventional fine aggregates in the production of concrete. Use of the waste rubber tyre in building materials is beneficial from sustainable and economical points of view. A systematic and comprehensive experimental study was conducted earlier by the authors for the mechanical and durable properties of rubberised concrete subjected to elevated temperature. However, there is non-availability of a mathematical model for rapid prediction of mechanical properties of the rubberised concrete subjected to elevated temperature. To bridge this gap an attempt has been made for development of explicit expressions through artificial neural network (ANN) approach in this paper. The training, validation, and testing data sets for ANN, are compiled from the recent researches of the authors. The input data sets contain six levels of elevated temperature (T) with three exposure durations (t) for all the specimens having six different fiber content (RF) along with three different water-cement ratio (w/c). On the other hand, the output parameters consist of mechanical properties (compressive strength static modulus of elasticity, dynamic modulus of elasticity and mass loss). Sensitivity analysis has also been carried out to investigate the effect of the input parameters on the output parameters. It is found that the average contribution of w/c,RF,T,t to all the output parameter is 6.67%, 10.10%, 80.01% and 3.22% respectively. The parameter T has highest impact on the all output parameters followed by RF whereas, rest of the input parameters (w/c,t) have relatively lower impact.