Library of Technical Faculties, Maribor (KTFMB)
  • Neural network model for analysis of fixtures
    Župerl, Uroš ; Čuš, Franc, 1953- ; Balič, Jože
    The paper deals with a neural-mathematical model for the analysis and rationalization of fixtures, suitable for clamping thin-wall products likely to undergo deformation due to clamping and cutting ... forces during machining. The program FIXAN has been developed for the evaluation of the fixturing scheme, for the calculation of the optimum magnitude and positioning of clamping forces, required to assure safe clamping of the workpiece during machining. The model is suitable for the analysis of fixtures intended for fixing prismatic and rotational products. The model takes into consideration the friction occurring between the workpiece and the fixture components. When the calculated clamping forces are greater than the permissible forces and would result in workpiece deformation during sraining, the system proposes to change the position of the basic fixture components or to change the cutting conditions (feeding). The cutting forces have been modelled by the BP neural network (ANN). Due to the use o ANN the duration of the calculation is very short, so the procedure can be carried out in real time. The described procedure reduces the fixture planning time and prevents defects and deformation during the machining process.
    Type of material - conference contribution
    Publish date - 2001
    Language - english
    COBISS.SI-ID - 6688790