Temperature Prediction During Self-pierce Riveting of Sheets by FEA-ANN Hybrid Model

Abstract

Nowadays, the automobile industries are more concerned about reducing automobile weight for improving fuel efficiency and reduced vehicle emission. Hence, there is a need for relatively lighter weight materials like aluminium alloy replacing steel parts for automobiles. However, joining of these lightweight materials is very difficult using conventional spot welding technique, which is a slow process also. In the present study, an alternative high-speed advanced mechanical fastening technique, namely self-pierce riveting is used for joining of these lightweight sheets of different materials. In this process, a semi-tubular rivet is pressed by a punch into two or more sheets which are supported on a die. Due to the special design of die shape, the rivet flares inside bottom sheet to form a mechanical interlock. Experimental measurement of temperature inside the sheets during the process is a difficult task, as the process takes place for a fraction of a second within a span of 10 mm. In the present study, the temperature rise during SPR process is predicted by finite element simulation using Abaqus® for any combination of sheet materials, process conditions and tool dimensions to understand the process. In addition, an ANN model is developed to predict the temperature rise during joint formation in sheets for any combination of process parameters. The ANN model accurately predicted the temperature evolution during the process. Hence, with the hybrid ANN-FEM strategy one can design and optimize the process, tool and material conditions, efficiently, which otherwise will be time and resource-intensive.