An approach for using iterative learning for controlling the jet penetration depth in abrasive waterjet milling

This paper presents a new methodology for controlling the jet penetration in abrasive waterjet milling. The generation of milled parts by means of abrasive waterjet is traditionally done by trial and error, relying mainly on the operator's expertise. An Iterative Learning Control (ILC) based approach is proposed to improve the prediction and correction of depth of the generated jet footprints in a systematic way every time the same part is milled. This approach utilizes a P-type (proportional) ILC algorithm to control the milling process, and uses a geometrical model based on non-linear partial differential equations for predictions of footprints to provide a fast convergence to the process input and reducing errors. The experimental validations of this approach were carried out using a Titanium alloy (Ti6Al4V) workpiece where a slope-varying profile was generated using 220 mesh grit-size abrasives with 0.04 kg/min mass flow rate and 1380 bar pressure. The results show that the achieved depth accuracy is indeed improved by more than 50% after four iterations when using this approach; providing basis for generating precision freeform surfaces using abrasive waterjet machines in a controlled manner.