Study of Tool Wear & Surface Roughness Using Artificial Neural Network

Abstract

In any machining operation, tools wear and surface roughness results inaccuracy and inefficiency. So it is always desirable to ensure minimum level of tool wear and surface roughness. Both of them depend on some parameters like feed, spindle speed, depth of cut and operation time. Optimization of these parameters ensures existence of tool wear and surface roughness under the tolerance limit. In this project, our aim is to devise a way of predicting tool wear and surface roughness for a given set of parameters. To do that, we collected experimental results of tool wear and surface roughness for thirty sets of parameters which were selected by Central Composite Rotatable Design (CCRD). Both tool flank and nose wear and surface roughness were measured by taking microscopic images of tool edge and job piece surface after each machining operation and then by using Image Processing Tool of MATLAB. The obtained results were then used for developing Artificial Neural Network (ANN) which was then used for the prediction of tool wear and surface roughness for a given set of parameters. The prediction and actual result were then compared and it was seen that both results coincide with each other