Prediction of the effect of vacuum sintering conditions on porosity and hardness of porous NiTi shape memory alloy using ANFIS

A neuro-fuzzy model was utilized to predict the hardness and porosity of NiTi shape memory alloy produced by vacuum sintering of powder mixture. Compaction pressure, sintering time and sintering temperature were chosen as input nodes. This procedure allowed successful prediction of porosity and hardness of the NiTi SMA samples. Absolute relative errors were at most 6.3% for hardness and 4.8% for porosity. Mean relative values were 3.4% for hardness and 3.3% for porosity. Results showed that the increasing of the values of input parameters affected outputs, linearly. The most significant parameters influencing the porosity content and the hardness of the under-vacuum combustion-synthesized NiTi specimens were sintering temperature and compaction pressure.