Document Type : Research Paper

Authors

1 Department of Mechanical Engineering, Arak University of Technology, Arak, Iran

2 Department of Mechanical Engineering, Tarbiat Modares University, Tehran, Iran

3 Department of Mechanical Engineering, Faculty of Engineering, Tarbiat Modares University, P.O.Box 14115/143, Tehran, I.R. Iran

4 Department of Industrial and Manufacturing System Engineering, Beihang University, Beijing, China

Abstract

In this paper, by considering the processing parameters including Blank Holder Force, Blank Holder Gap, and Cavity Pressure as the most important input factors in the hydroforming process, an experimental design is performed and an adaptive neural-fuzzy inference system (ANFIS) are applied to model and predict the behavior of aluminum thinning rate (upper layer and lower layer), the height of wrinkles and achieved depths that are extracted in hydroforming process. Also, the optimal constraints of the network structure are obtained by the gray wolf optimization algorithm. Accordingly, the results of experimental tests have been utilized for training and testing of the ANFIS. The accurateness of the attained network has been examined using graphs and also based on the statistical criteria of root mean square error, mean absolute error, and correlation coefficient. The results show that the attained model has been very effective in approximating the aluminum thinning rate (upper layer and lower layer), the height of wrinkles, and achieved depth in the hydroforming process. Finally, it can be seen that the root means square error of aluminum thinning rate (upper layer and lower layer), the height of wrinkles, and achieved depth respectively) of the test section, are 1.67, 2.25, 0.05, and 2.67. It is also observed that the correlation coefficient for the test data is very close to 1, which demonstrates the high precision of the ANFIS in predicting the outputs of the hydroforming procedure.

Graphical Abstract

Experimental investigation and modeling of Fiber Metal Laminates hydroforming process by the GWO optimized neuro-fuzzy network

Keywords

Main Subjects

CAPTCHA Image