In this research the authors propose an integrated intelligent diagnostics scheme for the on-line quality control of the Powder Injection Molding (PIM) process. In specific, this paper concentrates on the computational methodologies related to the process diagnostics. A causality network model and its underlying mathematics to represent the domain and case specific diagnostic knowledge of PIM process are discussed. Alternatives of hypothetical explanations to the product defects are generated by network conversion techniques and abductive reasoning. Diagnostic solutions are generated by applying verification criteria for the hypotheses. The effectiveness of the proposed methodologies are verified by testing real-life problems.
All Science Journal Classification (ASJC) codes
- Mechanical Engineering
- Industrial and Manufacturing Engineering