Modeling of laser welding of stainless steel using artificial neural networks. (2022)

• Record Type: Journal Article
• Authors: Banerjee, N.; Biswas, A.R.; Kumar, M.; Sen, A.; Maity, S.R.
• Is Part Of: Materials today. Volume 66:Part 4(2022)
• Page Start: 1784
• Page End: 1788
• Publication Date: 2022
• Subjects: Artificial neural network -- Normalization -- Laser welding -- Stainless-steel; Materials science -- Congresses -- Periodicals; 620.1
• DOI: 10.1016/j.matpr.2022.05.278
• Languages: English
• ISSNs: 2214-7853

Abstract

Abstract: Laser Welding (LW) is a process which is used to effectively weld metal, alloy or thermoplastics. Laser Welding uses a high concentrated beam of laser to melt the material and to perform welding. Its demand is increasing rapidly in automation as it is fast, more precise, have a deep penetration capacity and have less heat affected zone (HAZ), reduced tendency to spiking thus suitable for Ultra-Narrow Gas Laser Welding (UNGLW). Due to the high cost and time-consuming nature of laser welding experiments, repetition of one experiment in a wide range of data is not feasible. Artificial neural networks have been introduced as an effective tool for predicting values of outcomes and input parameters of various machining processes. In this work, the effect of laser power, welding speed and wire feed rate on the weld quality, i.e., tensile strength and overlap factor is investigated by ANN. Modelling of laser welding of stainless steel is performed using feed forward back propagation trained neural network. The projected outcomes are found to be in good agreement with the previous experimental findings.