Currently, the process of finding such families is based on identifying parts with similar shapes, either using visual inspection or by finding common process routes using production flow analysis. However, such methods can be slow and inaccurate, are costly and can lead to sub-optimal cell designs.

The CADFind Sketch and Search system, developed by Applied Search Technology AST), is said to solve these problems by identifying groups of similar parts automatically and then allowing the user to refine the groups interactively. It can search up to 40,000 parts in three seconds, but also makes use of an engineer`s product and manufacturing knowledge.

"Although some research and commercial systems currently offer part clustering, none of these have been designed specifically for this task," explains Dr Doug Love, AST’s research director. Love was one of the cofounders of AST which was set up in 2004 as a spin-off from Aston University.

The parts identification system is based on AST’s CADFind technology which can locate 3D models or 2D drawings from a sketch provided by a designer. The technology, which is said to be easy to use, even for casual users, can search through thousands of a company’s designs. A video demonstration is available on AST’s Web site.