CAD TOOLS, REVERSE ENGINEERING, 3D MEASURING AND ARTIFICIAL NEURAL NETWORKS IN AXIAL PISTON PUMPS STUDY.

The correct definition of technical parameters in the construction of volumetric pumps, especially those with axial pistons, must take into account both their mode of operation, characteristic load curves, and their operating duration. Since hydraulic pumps in general, and axial piston pumps in particular, are equipment capable of utilizing the energy transmitted by the fluid at high powers and moments of operation, great attention is paid to them from the design phase to maintenance. This paper aims to address essential principles in the CAD modeling of an axial piston pump, a study on the inspection of the piston block using a specific reverse engineering measurement technique, which involves 3D scanning of these components and their measurement using specialized software (GOM Inspect), as well as the comparison, through overlay, of CAD models created using CATIA software with scanned models, highlighting the advantages of reverse engineering techniques compared to traditional CAD design. By using coordinate measuring machines, the free-form, internal, and external surfaces of the piston block were measured and inspected to generate coordinates of measured three-dimensional points on the surfaces, with the aim of defining the maximum distances between the real profile and the adjacent circle of the axis channels (deviations from circularity), through parameter specifications such as eccentricity, concentricity, and radial runout. Additionally, a neural network model is proposed for predicting the theoretical fluid flow rate of axial piston pumps based on their technical parameters, such as pump speed, number of pistons, effective piston diameter, or stroke, which aims to optimize the synchronization process between the drive shaft rotation and the piston block rotation and ensure efficient power transfer to the overall hydraulic system. [ABSTRACT FROM AUTHOR]