Modelling framework to predict shallow-to-deep geometries of milled pockets by incorporating the effect of waterjet flow on nonplanar target

Milling shallow-to-deep features in titanium alloy presents machinability challenges when a conventional process is employed. In this context, maskless milling by abrasive waterjet (AWJ) can be a better alternative due to its unique capabilities. Unlike conventional milling, assuming a linear summation of kerf cross-section profiles (CPs) while following the raster jet path, in AWJ milling will not yield an accurate prediction of surface geometry CP due to the nonlinearities involved in material erosion by multiple phenomena (e.g., reflected jets). The unavailability of measurement techniques for gaining insights into this complex AWJ interaction with the target at the milling zone for studying the flow behavior and the associated erosion hindered the development of accurate models for AWJ milling. This necessitates the need for the AWJ milling models that incorporate the abovementioned nonlinearities to yield reliable predictions. Towards this, the present work proposes a novel modelling framework to predict the geometrical characteristics (CP, maximum erosion depth (hmax), top width (WP), and CP area (AP)) of AWJ milled feature in Ti-6Al-4V alloy, considering pocket as a case study under varying jet’s overlap percentage (JO%) and jet traverse rate (Vf). To get insights into the changes in the AWJ flow behavior and its interactions upon impinging the nonplanar target (NPT) post-first jet sweep and subsequent jet sweeps, computational fluid dynamics (CFD) simulations are performed. Results show an increasing asymmetric stagnation zone at the NPT relative to the planar target (PT), with the increased JO% and Vf leading to strong tangential flow, causing intense jet reflection and unwanted erosion. These understandings are utilized to mathematically model the unexplained erosion caused by the reflected jets, enabling accurate prediction of CP, hmax, WP, and AP. The proposed model accurately captured the nonlinearities that arose in hmax, WP, and AP with a maximum error of 10%, 11.8%, and 14.8%, respectively. In addition, the validated shapes of the pocket CPs result in an average mean absolute error of 42 µm, which agrees well with experimentally obtained ones. The validation results corroborate the hypothesis that the reflected jets caused upon impinging on the NPT compared to PT significantly contribute to unexplained erosion during milling other than the erosion due to the primary jets.