Al-based optimization method for the analysis of co-ordinate measurements within integrated CAD/CAM/CAE systems

摘要

To achieve higher geometric accuracy of complex shapes and overall increase in surface quality it is necessary to take full advantage of capabilities available in integrated CAD/CAM/CAE systems as well as in a new generation of CNC production equipment and coordinate measuring machines. Knowledge-based automation of product design, manufacture and coordinate measurements, using artificial intelligence methods, are of great importance for effective determination of global deviations between the virtual product model and the part machined. Due to a large number of points (500 to 100 000 and more), obtained from scanning measurements, it is only recently possible to run the appropriate, specialized software. To find a global optimum in the location and orientation of the cloud of points in relation to the geometric model, a new approach has been applied. A combinatorial-cyclic iteration method is, in fact, an reliable multivariable optimization method based on either deterministic models or selected neural network techniques. The method enables : (a) to analyse all errors in manufacturing processes, even on the most complex surfaces; (b) to effectively apply reverse engineering techniques; (c) to identify all sources of errors in CAD/CAM technology; (d) to eliminate the necessity of a precise set-up of the measured part on CMMs or other measuring systems; (e) to identify a global optimum, among many local optima.