Continuous Failure Diagnosis for Assembly Systems using Rough Set Approach

Increasingly, companies require faster ramp-up in order to cope with shorter production cycles and greater demand for product variety. Since quality and dimensional problems are one of the major reasons for delay during ramp-up, rapid diagnosis of dimensional failures is of critical concern. Given the lack of historical data and incomplete process knowledge in ramp-up, a Rough Set based diagnosis methodology is proposed which focuses on developing: (1) Self-learning ability to detect new faults as a system undergoes adjustments; and (2) A continuous diagnostic for faults rather than a crisp definition of faults. A case study illustrates the proposed approach.