Experimental investigations and parametric optimization of laser beam welding of NiTinol sheets by metaheuristic techniques and desirability function analysis. (April 2020)

• Record Type: Journal Article
• Authors: Datta, Susmita; Raza, Mohammad Shahid; Das, Amit Kumar; Saha, Partha; Pratihar, Dilip Kumar
• Is Part Of: Optics & laser technology. Volume 124(2020)
• Publication Date: 2020-04
• Subjects: Laser beam welding -- NiTinol sheet -- Grey Wolf Optimizer -- Cricket Algorithm -- Bonobo Optimizer -- Genetic algorithm -- Desirability function analysis; Optics -- Periodicals; Lasers -- Periodicals; Electronic journals; 621.366
• DOI: 10.1016/j.optlastec.2019.105982
• Languages: English
• ISSNs: 0030-3992

Abstract

Highlights: Butt-welding of 1 mm thick NiTinol sheet by fiber laser was systematically studied. Effect of process variables on bead geometry, mechanical properties were analyzed. A constrained optimization problem was formulated. Bonobo optimizer showed the fastest convergence among the metaheurestic algorithms. Abstract: Nitinol is widely used as a functional advanced material in various fields. Poor machinability and lack of available joining techniques are the major drawbacks in the application of NiTinol. The effects of process parameters on the bead geometry, microstructure, new phase formation and mechanical properties in laser welding of one mm thick NiTinol sheet in butt-joint configuration were established through this study. Laser power, scan speed and focal position were considered as input parameters, whereas bead area and microhardness value of the bead were taken as output parameters. Statistical regression analysis was performed in order to establish the input-output relationships. Optimization technique was applied in order to get the minimum bead area satisfying the condition of minimum deviation of microhardness of the bead area from that of the parent material. This was formulated as a constrained optimization problem and solved using three recently developed metaheuristic techniques, namely, Grey Wolf Optimizer (GWO), Cricket Algorithm (CA), Bonobo Optimizer (BO), apart from Genetic algorithm (GA), and Desirability function analysis. A good agreement was found between the results predicted by optimization tools and the experimental results.