Surface roughness prediction in end milling process using intelligent systems

• Published in: International journal of machine learning and cybernetics Vol. 5; no. 1; pp. 135 - 150
• Main Authors: Sharkawy, Abdel Badie, El-Sharief, Mahmoud A., Soliman, M-Emad S.
• Format: Journal Article
• Language: English
• Published: Berlin/Heidelberg Springer Berlin Heidelberg 01.02.2014; Springer Nature B.V
• Subjects: Accuracy; Adaptive systems; Aluminum base alloys; Approximation; Artificial Intelligence; Artificial neural networks; Back propagation; Complex Systems; Computational Intelligence; Control; End milling; Engineering; Feed rate; Fuzzy logic; Fuzzy systems; Genetic algorithms; Inference; Intelligent networks; Mathematical models; Mechatronics; Neural networks; Original Article; Pattern Recognition; Process parameters; Propagation; Radial basis function; Robotics; Surface finish; Surface roughness; Systems Biology; Training; Tuning; Workpieces; Prediction accuracy; Surface roughness; Adaptive neuro-fuzzy inference systems (ANFIS); End milling process; Radial basis function neural network (RBFN); Genetically evolved fuzzy inference systems (G-FIS)
• ISSN: 1868-8071; 1868-808X
• DOI: 10.1007/s13042-013-0155-7

The mechanism behind the formation of surface finish is very complicated and process dependent, therefore it is very difficult to calculate the value of surface roughness through analytical formula. In this paper, a study is presented to model surface roughness in end milling process. Three types of intelligent networks have been considered to numerically identify the end milling process. They are (i) radial basis function neural networks (RBFN) (ii) adaptive neuro-fuzzy inference systems (ANFIS), and (iii) genetically evolved fuzzy inference systems (G-FIS). The machining parameters, namely, the spindle speed, feed rate and depth of cut have been used as inputs to model the workpiece surface roughness. The contribution of this work is to investigate different methodologies which could be used to get the best prediction accuracy. The procedure is illustrated using experimental data of end-milling 6,061 aluminum alloy. The three networks have been trained using experimental training data. After training, they have been examined using another set of data, i.e. validation data. Results are compared with previously published results. It is concluded that ANFIS networks may suffer the local minima problem and genetic tuning of fuzzy networks cannot insure perfect optimality unless suitable parameter setting (population size, number of generations … etc.) and tuning range for the fuzzy inference systems (FIS) parameters are used which can be hardly satisfied. It is shown that the RBFN model has the best performance (prediction accuracy) in this particular case.