A Novel Hybrid Transform approach with integration of Fast Fourier, Discrete Wavelet and Discrete Shearlet Transforms for prediction of surface roughness on machined surfaces

Abstract

Abstract Milled surfaces contain features of geometrically similar appearances under various magnifications in different orientations. Fourier transform is useful in obtaining accurate information from periodic data than wavelets, wavelets are effective in handling multi-resolution data than fourier transform, shearlets show much higher directional sensitivity than both fourier transform and wavelets but are computationally complex. Hence a novel hybrid transform approach with integration of Fast Fourier Transform (FFT), Discrete Wavelet Transform (DWT) and Discrete Shearlet Transform (DST) to characterize surface roughness on machined surfaces is presented. For the experimentation, Taguchi’s L9 orthogonal array was used on Computer Numerical Control (CNC) mill with High Speed Steel (HSS) end mill to machine aluminum 3025 alloy work samples. The Artificial Neural Network is trained with cutting parameters, vision parameters as input and the experimental surface roughness (Ra, Rq, Rz) as output. The validation results of proposed models show better performance with reasonable accuracy.