Prediction of surface finish and optimization of machining parameters in turning

Abstract

Surface roughness plays a crucial role in the functional capacity of machined parts. In this work, experiments were carried out on a conventional lathe for different cutting parameters namely feed, spindle speed, depth of cut and tool nose radius according to Taguchi Design of Experiments. Radial acceleration readings were taken with an accelerometer. Optimum cutting parameters and their level of significance were found using Taguchi analysis (ANOVA). Regression analysis was carried out to identify whether the experimental roughness values have fitness characteristic with the process parameters. Recurrence Plots (RP) were obtained using the sensor signals which determine surface roughness qualitatively and Recurrence Quantification Analysis (RQA) technique was used to quantify the RP obtained. Surface finish was predicted using a feed forward back propagation neural network with RQA parameters, cutting parameters and acceleration data as inputs to the network. The validity and reliability of the methods were verified experimentally. © (2012) Trans Tech Publications.