Investigation of the effect of hydromechanical deep drawing process parameters on formability of AA5754 sheets metals by using neuro-fuzzy forecasting approach
No Thumbnail Available
Date
2015
Journal Title
Journal ISSN
Volume Title
Publisher
IOS PRESS
Access Rights
info:eu-repo/semantics/closedAccess
Abstract
Adaptive neural-network based fuzzy logic inference system (ANFIS) is a useful method instead of costly Finite Element Analysis (FEA) in order to reduce investigation cost of forming processes. In this research, the effect of hydromechanical deep drawing (HDD) process parameters on AA5754-O sheet was investigated by FE simulations with analysis of variance (ANOVA) and Adaptive Neuro-Fuzzy Modeling approach. In order to determine the prediction error of the ANFIS model according to FEA, firstly a series of FEA of the HDD process were conducted according to Taguchi's Design of Experiment Method (DOE). The results of the FEA were confirmed by comparing the thickness distributions of the formed cups by experimentally and numerically. Moreover an adaptive neural-network based fuzzy logic inference system (ANFIS) was created according to results of simulation to predict the maximum thinning of AA5754-O sheet without needing FE simulations. The calculation performances of the ANFIS model were determined by comparing the estimated results with the results of the FE simulations. By using the results of the FE simulations which were conducted according to a matrix plan, the effects of the parameters to the thinning of the blank were determined by the analysis of variance (ANOVA) method. ABAQUS and MATLAB/ANFIS/Simulink softwares were used to realize and simulate proposed techniques. Mean error of prediction result of ANFIS is found as 0.89% according to FEA.
Description
Keywords
Hydromechanical deep drawing, limiting drawing ratio, prediction, finite element, Taguchi DOE, adaptive neuro-fuzzy
Journal or Series
JOURNAL OF INTELLIGENT & FUZZY SYSTEMS
WoS Q Value
Q3
Scopus Q Value
Q2
Volume
28
Issue
2
