Modeling of process parameters on DSS 2205 through RSM, ANN, fuzzy under cryo-MQL process

Abstract

Duplex stainless steel (DSS) 2205 is considered difficult to cut machine materials. Therefore, an advanced cooling technique, such as a hybrid Liquid carbon dioxide (LCO2)-Minimum quantity of lubrication (MQL) approach, is proposed for drilling duplex stainless steel 2205. In this research study, parametric optimization was performed by using the Response surface method (RSM) with the box-Behnken design of the matrix. For parametric optimization, select three input parameters: drill diameter, spindle speed, and feed rate, while hole deviation and cylindricity error are output responses. The Analysis of variances (ANOVA) test showed that spindle speed is the major factor that influences both responses, e.g., hole deviation and cylindricity error, with contributions of 38.90% and 59.42%, respectively. The modeling and predictive abilities of developed Fuzzy logic system (FLS) and Artificial neural network (ANN) to the comparison of experimental results. It was also found that the predicted values from regression analysis, artificial neural networks, fuzzy logic interfaces, and the proposed method were in good agreement with actual experiment results. Also, the Artificial neural network model provides a minimum % of error of 1.36% and 2.44 %, respectively responses compared to other models.