A KNOWLEDGE-BASED ADVISER FOR THE AUTOMATIC SELECTION AND SEQUENCING OF ORIENTING DEVICES FOR VIBRATORY FEEDING

The selection and sequencing of orienting devices used in vibratory fe eders is integral to the design of the feeders. However, because of th e lack of organized, published guidelines, the selection and sequencin g of orienting devices depends almost entirely on the experience of th e skilled artisans who tool up the vibratory feeders. This paper discu sses a knowledge-based framework by which a CAD model of the part to b e oriented and fed is analysed for its significant orienting feature(s ), and a suitable match of orienting devices is determined. The outcom e of a consultation with the expert advisor typically yields move than one feasible sequence of orienting devices, leaving the user free to decide on the most appropriate sequence to adopt. The paper also discu sses the hierarchical classification of the features of the part, as w ell as those of the orienting devices. The knowledge-based advisor was implemented on a 486-DX33 personal computer within the Microsoft(R) W indows 3.1 environment, which integrates both the CAD and the expert s ystem. A benchmarking exercise carried out for two rotational and two prismatic parts yielded solutions that compared favorably with those u sed in industry, and those proposed by the UMASS system.