Balancing the surface erosion and hardening in discrete waterjet peening by the data-driven approaches of response surface method and desirability function

Discrete waterjet (DWJ) technology has potential for surface strengthening due to its precise parameter control and strong impact performance. However, the possible excessive material erosion during peening processes may compromise the service performance of the materials. This study focuses on balancing surface hardening and material erosion caused by DWJ peening in 304 stainless steels. Using data-driven techniques of response surface and desirability function methodologies, a multi-factor multi-objective parameter optimization was performed. Regression models of target responses were developed for the material responses to DWJ parameters (exposure time, pressure, frequency, and pulse length), including residual stress, hardness, erosion volume, and depth. The analysis of variance (ANOVA) and model diagnostics verified the goodness-of-fit, predictability, and reliability of established models. Perturbation plots indicated that residual stress was negatively correlated with pressure and frequency, but positively correlated with pulse length, whereas hardness showed positive correlations with exposure time, pressure, and frequency. Erosion depth and volume were positively correlated with all DWJ parameters. Additionally, significant correlations were identified between erosion depth and hardness, as well as residual stress and hardness. The F-value ranking indicated pressure has the highest influence degree on residual stress, hardness, and erosion depth, whereas pulse length predominantly affected erosion volume. For second-order interaction terms, the interaction between exposure time and pulse length significantly influenced residual stress, whereas the combined effect of pressure and frequency had the greatest effect on hardness and erosion depth. Furthermore, the interaction between exposure time and pulse length was identified as the most influential factor on erosion volume. Using the desirability function method, optimal parameters were determined within the experimental range, achieving maximum hardening effects with minimal material erosion and relative errors of 3%-8%. These findings offer valuable insights for optimizing surfaces treated with DWJ peening technology.