Optimizing parameters of freeform surface reconstruction using CMM

Abstract

In this paper, Design of Experiment (DOE) techniques are used to optimize parameters of surface reconstruction in Reverse Engineering technology for freeform surfaces. The input factors are noise reduction, number of points in (%), triangle counts in (%), and Sampling when the responses are the surface accuracy of the final reconstructed surface, computational time and the required space in computer memory. Accuracy of the model is measured in terms of the minimum standard deviation values. The proposed approach was confirmed through a case study carried out using a real metallic part. The case study involves two types of point cloud which have been obtained from both fixed CMM laser line scanner and portable CMM laser line scanner. The 3D CAD models of the selected part are developed based upon combination of parameters given by Taguchi L32 orthogonal array design for point cloud obtained from fixed CMM and Taguchi L16 orthogonal array for point cloud data obtained from portable CMM. This study also describes the development of predictive models for the given responses utilizing response surface methodology (RSM). The developed predictive models are manipulated using contours and response surfaces. It has been concluded that the final accuracy level of reverse engineered surfaces is depending on number of input points of point cloud data, number of triangles in polygon model and noise reduction.