RT Journal Article
SR Electronic
A1 Vrabeľ, Marek
A1 Maňková, Ildikó
A1 Paľo, Miroslav
A1 Lattner, Radek
T1 The Effect of Artificial Neural Network Architecture on Surface Roughness Parameter Prediction Capability when Turning Inconel 718
JF Manufacturing Technology Journal
YR 2016
VO 16
IS 4
SP 834
OP 839
DO 10.21062/ujep/x.2016/a/1213-2489/MT/16/4/834
UL https://journalmt.com/artkey/mft-201604-0032.php
AB This paper investigates the influence of Artificial Neural Network (ANN) architectures on its prediction capability when machining nickel based super alloy. The ANN was employed to determine surface roughness parameter Ra through cutting conditions, tool wear and process monitoring indices such a cutting force components. The ANN structure was optimized by methods like a reduction of input vector parameters, dimensions of input data pattern, combined reduction and modification of hidden layers. Calculated and experimentally measured values were compared for each optimized ANN model. The work concludes that optimization of ANN has significant influence on prediction capability and accuracy for the task proposed.