ABSTRACT : |
Surface roughness (Ra) is a significant upshot in the manufacturing process and it materializes a major part in the manufacturing system. It depends on different machining parameters and its prediction and control is a query to the researchers. In this study, a regression model and two artificial neural networks (ANNs) namely: Back propagation and radial basis function, were developed to predict surface roughness in electrical discharge machined surfaces. In the development of predictive models, machining parameters of discharge current (Ip), pulse duration (Ton) and duty cycle () were considered as model variables with a constant voltage 50 volt. For this reason, extensive experiments were carried out in order to collect surface roughness dataset. The developed models are validated with a new set of experimental data, and predictive behavior of models is compared, subsequently relative advantages of each model are analyzed. Analysis of variance (ANOVA) and F-test were used to check the validity of regression model and to determine the significant parameter affecting the surface roughness. The statistical analysis exemplify that the Ip, Ton and ô were the factors in sequence have significant influence the on surface roughness. The microstructures of machined surfaces were also studied by scanning electron microscopy (SEM). The SEM investigations revealed that EDMed produced were increased significantly with Ip.
Keywords: Surface roughness, Regression analysis, Back propagation, radial basis function |
|