Neural Network Applied to Detect Burn in Grinding

P.R. Aguiar, E.C. Bianchi, F.R.L. Dotto, R.A. Flauzino, and D.H. Spatti (Brazil)


Manufacturing, Neural Networks, Acoustic Emission, Grinding, Signal Processing, Automation


In this paper, a new methodology is presented for detecting burn in surface grinding process. A multi perceptron neural network was employed to generalize the process and, in turn, obtain the burning threshold. It has been investigated that burning occurrence in grinding can be detected by DPO, FKS and other studied parameters. However, those parameters were not good enough for the grinding conditions used in this work. Acoustic emission and electric power of the grinding wheel drive motor are the input variables and the burning occurrence the output variable of the neural network. In the experimental setup, one type of steel (ABNT-1045 annealed) and one type of grinding wheel referred to as TARGA model ART 3TG80.3 NVHB were employed.

Important Links:

Go Back