This paper analyses the effect of the abrasive waterjet cutting parameters' modification on the condition of the workpiece surface layer. The post-machined surface of casting aluminium alloys, AlSi10Mg and AlSi21CuNi, was characterised in terms of surface roughness and irregularities, chamfering, and microhardness in order to reveal the effect that variable jet feed rate, abrasive flow rate, and sample height (thickness of the cut material) have on the quality of surface finish. From the analysis of the results, it emerges that the surface roughness remains largely unaffected by changes in the sample height h or the abrasive flow rate ma, whereas it is highly susceptible to the increase in the jet feed rate vf. It has been shown that, in principle, the machining does not produce the strengthening effect, that is, an increase in microhardness. Owing to the irregularities that are typically found on the workpieces cut with higher jet feed rates vf, additional surface finish operations may prove necessary. In addition, chamfering was found to occur throughout the entire range of speeds vf. The statistical significance of individual variables on the 2D surface roughness parameters, Ra/Rz/RSm, was determined using factorial analysis of variance (ANOVA). The results were verified by means of artificial neural network (ANN) modelling (radial basis function and multi-layered perceptron), which was employed to predict the surface roughness parameters under consideration. The obtained correlation coefficients show that ANNs exhibit satisfying predictive capacity, and are thus a suitable tool for the prediction of surface roughness parameters in abrasive waterjet (AWJ) technology.
Keywords: ANOVA statistical analysis; Al-Si alloys; abrasive waterjet technology; artificial neural networks; irregularities; microhardness; specimen chamfering; surface roughness.