This study is concerned with the development of a practical and accurate technique for off-line determination of the variables characterizing the macro-geometry of the kerf generated in the course of abrasive waterjet (AWJ) machining of ductile materials. The study involved generation and processing of a database relating operational conditions with kerf dimensions. The total number of generated samples exceeded 1500.

A physical model relating the process results with operational conditions was constructed and used for the selection of a statistical technique for analysis of the acquired experimental information. The semiempirical model developed here is based on a simple theoretical model which assumes that the particle distribution within the AWJ is statistically uniform. The good correlation found between the experimental results and the semiempirical model demonstrate the validity of this assumption. Regression equations, determining the depth of jet penetration, kerf width and taper, were constructed. The correlation coefficients between predicted and measured values of the kerf characteristics exceed 0.94. Only in one case out of 20 was the correlation coefficient 0.9. The observed result demonstrates the geometry of the kerf is controlled by the diameter of the jet and kinetic energy of particles.

A practical, reliable procedure for construction of a prediction technique for AWJ machining of a material in question was suggested and tested. The proposed technique will be used for control of AWJ machining as well as for the development of an expert system.