A Case-Based Reasoning System Combining Expert Knowledge for Automated Design of Multi-Pass Hot Forging for Hub Bearings

This article highlights a case-based reasoning method combined with expert knowledge in the design of multi-pass hot forging process for hub bearings, as well as the outcomes of our research on the representation and revise of forging design knowledge. Forging design knowledge acquired from manuals and the experience of technologists was refined and then framed in the form of an “IF-Then” variety of production rules, including the selection of the shape complexity factor and the machining allowance, etc. A forging similarity measure based on part sizes of the hub bearing was proposed to retrieve cases. After the dimensional input of the new hub bearing, the synthesis weight was determined using the objective weight calculated from historical cases and the subjective weight recorded in the expert knowledge. Based on the global similarity calculated by the synthesis weight, the two most similar historical cases would be then searched for use as templates in the design of the new forging process. Size ranges of the new forging and dies were established in detail based on design values of two similar cases combined with expert knowledge, including the machining allowance and the fillet, etc. The above system was coded in Visual C#, and its user interface was developed using WinForm, AutoCAD, and UG software. The newly designed dies were imported into Deform 3D for the finite element simulation, and the results demonstrated that the hub bearing forging was adequately filled, proving that CBR would assist technologists in achieving the goal of rapid design.