Experimental investigation and validation on the effect of nickel addition on properties of the pressureless sintered boron carbide composites using machine learning models

This research work aims to investigate the effect of a low amount of nickel (< 0.6 wt%) on the properties of the boron carbide. The samples are synthesized through pressureless sintering at 2150 C under an argon atmosphere (1 bar). Firstly, the thermodynamic simulation is carried out using the HSC software to determine the free energy change of reactions and their viability. The prepared samples are analyzed by employing various characterization techniques such as XRD, FESEM, and EDS. It is concluded that the presence of nickel (i.e., a low melting point than the boron carbide) results in the formation of the liquid phase during the sintering process. Thus, the sinterability of the samples is significantly enhanced. Moreover, the optimum sample is obtained by adding the 4 wt% of nickel in boron carbide. The maximum density, hardness, elastic modulus, and fracture toughness for the sample containing 4 wt% nickel are obtained as 91.00%, 21.83 GPa, 374.45 GPa, and 3.11 MPa m(0.5), respectively. Furthermore, machine learning (ML) techniques (i.e., M5P tree model, random tree, REPTree (Reduced-Error Pruning), and artificial neural network (ANN)) are also utilized to check the adequacy of the results. It is concluded that the maximum value of CC, i.e., 0.9929, is obtained for the REPTree technique with a minimum value of MAE (i.e., 0.0869) and RMSE (i.e., 0.0968). It is observed that the ML technique showed the validation of experimental results.