A comparative study of experimental and adaptive neuro fuzzy inference system based prediction model of machined AM60 magnesium alloy and its parameter effects. (2021)

• Record Type: Journal Article
• Authors: Sivam, S.P. Sundar Singh; Babu Loganathan, Ganesh; Harshavardhana, N.; Kumaran, D.; Prasanna, P.
• Is Part Of: Materials today. Volume 45:Part 2(2021)
• Page Start: 1055
• Page End: 1062
• Publication Date: 2021
• Subjects: Comparative study -- AM60 -- Subtractive manufacturing -- ANFIS; Materials science -- Congresses -- Periodicals; 620.1
• DOI: 10.1016/j.matpr.2020.03.158
• Languages: English
• ISSNs: 2214-7853

Abstract

Abstract: Today in industrialized nations, the machining cost adds up to over 15% of the estimation of every fabricated item." Reason why machined component have been proven as one most significant parts of manufacturing science and technology. Magnesium - Mg and its amalgams are widely used due to their remarkable properties such as low density, excellent machinability and recyclability. It is evident that the magnesium alloys are produced into cast. In the aircraft industry, Magnesium alloys have been recognized as an important material. It is important to choose the most optimum machining parameters to obtain very high quality components. The machining variables are completely subjected to the required quality and the nature of the material which is undergoing machining operation. This paper studies the optimization of outputs, namely surface Roughness - Ra and Material Removal Rate – MRR for input variables, for instance cutting speed – IP1, feed rate – IP2 and depth of cut – IP3 for Magnesium alloy AM60 by employing the computational methodologies ANFIS (Gaussian membership function). Investigational approval trails were directed to approve the ANFIS demonstrate. The anticipated Ra (μm) and MRR (mm 3 /min) was related with measured data, and the maximum prediction error for surface roughness was 0.001115 & 0.0000399% and minimum prediction error for MRR (mm3/min) was 27.8% & 0.7%, while the average prediction error was 0.000452% &7.7% for Ra and MRR.