Neural Predictor for Surface Roughness of Turned Parts

Journal of control, automation and electrical systems

article

Mizuyama, Demerval¹  da Silva, Carlos Elias¹  Goedtel, Alessandro¹  Graciola, Clayton L.¹  Palácios, Rodrigo H. C.¹ 

[1] Departments of Electrical, Mechanical and Computer Engineering, Federal Technological University of Paraná

关键词: AC induction motor;    Turning process;    Power quality;    Artificial neural network;    Surface roughness;   

DOI  :  10.1007/s40313-018-0376-9

学科分类：自动化工程

来源: Springer

【 摘 要 】

Lathes and other turning machines are widely used in the metalworking and manufacturing sector which are powered by induction motors and subject to disturbances in power quality. The dimensions and surface roughness of turned parts produced by these machines are a function of the machining parameters, material properties of the parts and tools, and tool geometry. The surface roughness of machined parts varies with changes in the electromagnetic torque on the motor shaft, and it is considered one of the main indices of finished product quality. This paper seeks to present artificial neural networks as a surface roughness predictor for turned parts, based on the values of the effective current feeding a three-phase induction motor in a machining process. Simulation and experimental results are presented to validate the performance of the proposed method under different unbalanced voltage and machining conditions.

【 授权许可】

CC BY