Evaluation of Bending Strength of the Vibratory Welded Joint Using Regression Technique
- P. Govinda Rao (GMR Institute of Technology) | P. Srinivasa Rao (Centurion University) | A. Gopala Krishna (Jawaharlal Nehru Technological University)
- Document ID
- International Society of Offshore and Polar Engineers
- International Journal of Offshore and Polar Engineering
- Publication Date
- September 2015
- Document Type
- Journal Paper
- 227 - 230
- 2015. The International Society of Offshore and Polar Engineers
- bending strength, Generalized Regression Neural Network (GRNN), Vibratory welding, voltage, grain structure, time of vibration
- 0 in the last 30 days
- 23 since 2007
- Show more detail
Previous researchers developed the vibrating table for producing mechanical vibrations into the weld pool during the welding process. The designed vibrating table produces the required frequency with suitable amplitude and acceleration in terms of voltages. This helps in producing a uniform and fine grain structure in the welded joints, which results in an improvement of the bending strength of the welded joints. This paper presents the implementation of the Generalized Regression Neural Network (GRNN) to establish a relation between vibration parameters such as the input voltage to the vibromotor, the time of vibration, and the bending strength of the vibratory welded joints. In order to validate the feasibility of the developed prediction tool, a comparison is made with the experimental results.
In manufacturing industries, welding is widely used for joining metals. The welding joints prepared by the arc welding process generally offer good strength and hardness properties. Metal arc welding is the most flexible fusion welding and one of the most widely used welding processes. Mechanical vibrations into the weld specimen during the welding process improve the welded joint properties significantly. The enhancement of the welded joint properties can be altered by the variation of the vibration parameters.
Vibrations applied during welding generally reduce the residual deformation and stress and improve the mechanical properties of the weldments (Lu et al., 2006; Xu et al., 2006; Lakshminarayanan and Balasubramanian, 2010). An enhancement of the mechanical properties and the quality of the fusion metal through the use of vibration during welding was considered recently and was found to improve the bending property of the welding line, tensile strength, and morphology (Hussein et al., 2011; Munsi et al., 2001; Tewari and Shanker, 1993; Weglowska and Pietras, 2012).
The Generalized Regression Neural Network (GRNN) is a type of supervised network and has been widely accepted for its excellent ability to train rapidly on sparse data sets. The GRNN usually performs better and faster in the approximation of continuous functions. Tseng (2006) implemented the GRNN to create approximate models to establish a relation between the spot welding parameters, welded joint strength, and power required to prepare the welded joint. Kathersan et al. (2012) addressed the modelling of the welding parameters in the arc welding process by using a set of experimental data, utilizing regression analysis, and employing optimization via the particle swarm optimization algorithm.Though there is literature that describes the phenomenon of improving the welded joint strength properties, the relation between the vibration parameters and welded joint properties has not been established. Hence, the present work is aimed at building a relation between the vibration parameters and welded joint properties from the experimental data through the use of the Generalized Regression Neural Network.
|File Size||879 KB||Number of Pages||4|