Experimental investigation and optimization of plasma arc cutting of structural steel grade St.52-3 using grey-fuzzy algorithm

Abstract

Virtually all steel products are cut at some stage during fabrication, manufacture and construction mostly by using thermal cutting. Material transformation as well as cut surface characteristics after cutting plays vital role in determining its fatigue behavior. At the same time, productivity improvement is the main concern for metal cutting industries. So, challenges of today’s research are related to enhancing quality without compromising its productivity. In this study, literature has been studied with more focus on parametric optimization of plasma arc cutting. Responses considered are the mean height of profile (R5z), material removal rate (MRR) and size of heat affected zone (HAZ). Parameters chosen are current, arc voltage, pressure and speed. Box-Behnken response surface design is adopted to design and conduct an experiment resulting in 27 runs (four factors and three levels for each). Structural steel st.52-3 sheet (5mm thickness and 50*50mm) are cut with the help of CNC air plasma arc cutting machine (Hypertherm Powermax 105A) at three different level of speed (1000 mm/min,1300 mm/min and 1600 mm/min), arc voltage (125, 135 and 145), air pressure (5 bars, 5.5 bars and 4.5 bar) and current (55A, 65A and 75A). R5z and HAZ are measured with the help of surface roughness tester (Motutoyo SJ 210) and inverted type optical microscope. ANOVA showed that current and speed have a significant effect on MRR followed by arc voltage. R5z found to be significantly affected by arc voltage and speed. Similarly, speed is having a significant effect on HAZ followed by interaction between current and arc voltage. The Grey-fuzzy algorithm used for parametric optimization provided optimal parameters setting as the current at 75 A, arc voltage at 125, pressure at 5 bars and speed at 1600 mm/min. The confirmation test showed 18.6% improvement in GFRG. Finally, comparison with oxyfuel cutting is made in terms of quality and cost.