Modeling the Surface Integrity of Ball Burnished Biocompatible Magnesium Alloy by Soft Computing Techniques

Soft computing techniques induce artificial intelligence to the machines, which are able to differentiate patterns or information. Design and development of soft computing models for analyzing the surface integrity of ball burnished biocompatible magnesium alloy are quite essential for elucidating ball burnishing process as potential surface modification method for biomaterials. Due to nonlinear and complex interactions between ball burnishing parameters, exact modeling of surface integrity is very challenging. Soft computing techniques can be used in distinguishing information from experimental data sets and can assist in predictive modeling of surface integrity of ball burnished surface. In this study, soft computing techniques such as FIS, ANN and ANFIS are deployed to predict the surface integrity of ball burnished Mg Ze41A alloy. The ANFIS model surpassed the other two models in estimating the surface roughness with a lowest mean error of 2.20%. Similarly, the ANN model surpassed the other two models in estimating the microhardness with a lowest mean error of 0.37%. The mean error of all the three soft computing models is less than 5%, indicating a confidence level greater than 95%. Therefore, these soft computing models can be regarded as robust, reliable and accurate models for estimating the surface integrity during ball burnishing process.