The use of machine learning and metaheuristic algorithm for wear performance optimization of AISI 1040 steel and investigation of corrosion resistance

Abstract

AISI 1040 steel offers a wide range of industrial applications due to its mechanical characteristics and applicability. The present work investigates the wear performance of AISI 1040 steel under dry sliding conditions and its optimization using combined machine learning (ML) and metaheuristic algorithm. Sliding wear test were carried out on a pin-on-disc tribometer by varying the load (10–100 N), sliding speed (0.5–1.5 m/s) and sliding distance (400–1000 m). The test parameters were varied at three levels. Experiments were carried out following combinations in Taguchi's L27 orthogonal array (OA). Artificial neural network (ANN) was used to model the process parameters with wear rate. The trained network was optimized using genetic algorithm (GA) to predict optimal wear rate. This methodology has been termed as ANN-GA method. The results were compared with conventional Taguchi based and regression based GA optimization. A significant reduction in the wear rate could be realized due to optimization using ANN-GA method. This work also examined the corrosion behaviour of AISI 1040 steel exposed to 3.5% NaCl, 3.5% NaOH, 0.5 M H2SO4 and 3.5% NaCl + 0.5 M H2SO4 to simulate various corrosive environments encountered in industrial applications. A nobler corrosion potential was obtained in 3.5% NaOH. Investigations of corroded samples showed pitting corrosion in 3.5% NaCl, 0.5 M H2SO4 as well as combined chloride and sulphate attack. On the other hand, negligible corrosion was observed in 3.5% NaOH.