Cogging process design of M50 bearing steel for billet quality

Abstract

This study proposes a method to determine cogging process parameters aiming at homogeneous effective strain distribution and grain size uniformity and refinement during the manufacturing M50-bearing steel billets. The design parameters were selected as feed, number of rotation, depth schedule, and forging ratio. The Billet Quality Index (BQI) is introduced to comprehensively evaluate the effective strain distribution and grain size number (ASTM E112) of the final billet and used as the objective function for the process design. The results of effective strain distribution and grain size number are obtained through finite element simulation, and these values are used to investigate the BQI of the final billet. For the finite element analysis to calculate the temperature field and grain size, the heat transfer coefficients determined by reverse engineering and the grain growth model coefficients were used compared with the experimental measurements. The Taguchi method investigates the influence of parameters on BQI, excluding the forging ratio. Optimal process parameters, including forging ratio, are determined using a Deep Neural Network (DNN) regression model, targeting BQI minimization. The BQI prediction accuracy of the DNN regression model is 98.83%, emphasizing the effectiveness of DNN despite the limited dataset size. The optimized cogging process reduces the total process time by 2.394 h and improves the BQI by 6.91% compared to the minimum BQI in the dataset. The optimized cogging process promotes homogeneous effective strain distribution, grain refinement, and uniformity, enhancing billet manufacturing efficiency.