Modeling of Surface Roughness Belonging to Oriental Beech Wood with an Artificial Neural Networks

The aim of this study was to model the surface roughness of Oriental beech (Fagus orientalis Lipsky) wood with the aid of artificial neural network (ANN) approach. Firstly, the working parameters of a wide belt sanding machine were adjusted to be the sanding belt grit size of 60-100, feeding speed from 4 m/min to 10 m/min, and sanding cutting depth from 0.1 mm until 0.3 mm, respectively. Secondly, after the surface roughness of the samples was experimentally recorded, the data were divided into two basic cathagories: namely, (I) the training sets and (II) test data sets. Thirdly, they were modeled by the approach of artificial neural networks so that the fundamental surface roughness parameters (Ra, Rq and Rz) can be anticipated thoroughly. The comparisons between the experimental results and theoretical findings (RRa = 0.99869, RRq = 0.9982 and RRz = 0.99882) show well-agreement with each other’s. In this respect, this study declares that the surface roughness of the solid beech was perfectly predicted within far higher accuracy and relatively lower error by using the artificial neural networks approach