ARTIFICIAL NEURAL NETWORK MODELLING OF THE SURFACE ROUGHNESS OF FRICTION STIR WELDED AA7020-T6 ALUMINUM ALLOY

Document Type : Research articles

Authors

1 Mechanical Engineering Department, Shoubra Faculty of Engineering, Benha University, Cairo, Egypt

2 Mechanical Engineering Department, Shoubra Faculty of Engineering, Benha University, Cairo, Egypt.

Abstract

In the present investigation, lap joints of the AA7020-T6 aluminum sheets were joined using friction stir welding (FSW). The AA7020-T6 sheets has 3 mm thickness. The FSW was carried out at three different tool rotational speeds of 1200 rpm, 1400 rpm and 1600 rpm; and three different welding speeds of 20 mm/min, 40 mm/min, and 60 mm/min. During FSW, the tool tilt angle and plunging depth were kept constant at 3o and 0.5 mm, respectively. The FSW was performed using a tool with a tapered pin profile and a flat shoulder. The surface quality of the FSW specimens was evaluated by the arithmetic average roughness value (Ra). The results revealed that increasing of the tool rotational speed and/or the welding speed increases the surface roughness of AA7020-T6 lap joints. The developed artificial neural network (ANN) model showed a good agreement between the predicted and experimental results. The ANN model exhibited mean relative error (MRE) of 3.8%.

Keywords

Files

Share

How to cite

Statistics