Case-based polishing process planning with Fuzzy Set Theory

Abstract

Polishing is widely used as a final processing operation for many products and components. Although the level of automation increases gradually over the years, manual or semi-automatic polishing is still commonly practised. The choice of polishing process parameters is largely based on experience of polishing technicians and involves a lengthy trial and error iteration before reaching an acceptable level. This paper proposes to acquire successful projects and build up a case repository of polishing parameters of both products and processes. Case-based reasoning (CBR) is then applied to mimic the experience-based polishing process planning. A problem case is first well-structured and then matched against all cases in the repository. The most similar ones are retrieved for further reasoning for their potentials of being revised and adapted to form an optimal solution. This research combines Fuzzy Set Theory with CBR to address two fundamental problems in polishing process planning. One is that values of product features and process parameters such as polishing force, amount of polishing compounds, polishing wheels, rotating speed, and feed rate cannot be exactly measured and controlled. The other is that influencing relationships between process parameters and polishing quality indicators as measured by surface roughness (Ra) and grossness (Gu) cannot be scientifically established mathematically. A case study is conducted within the collaborating company and the results from the proposed system are generally consistent with the actual decisions.