Modelling of Selected Surface Roughness Parameters using ANN in Waterjet Cutting

This paper investigates the effect of changes in selected technological parameters (cutting speed, pressure, and abrasive flow rate) on abrasive water jet machining of AZ91D magnesium alloy, focusing on the surface quality defined by roughness parameters (Ra and Rz). Additionally, artificial neural networks were employed to model the roughness parameters, aiming to assess their utility as a predictive tool for roughness in abrasive water jet machining. The findings reveal that pressure has the most significant impact on roughness during AWJM. The networks developed demonstrated satisfactory predictive capability, as evidenced by the achieved R correlation values (RRa=0.956 and RRz=0.9768). This confirms that artificial neural networks can serve as a valuable tool to achieve the desired surface roughness. This study's findings are crucial for technologists seeking innovative, data-driven solutions for improving machining performance and safety.