A SEMANTICALLY CONSTRAINED BAYESIAN NETWORK FOR MANUFACTURING DIAGNOSIS

The diagnostic problem is posed as recognizing patterns in rejection d ata and the subsequent mapping to causes. A new network architecture h as been proposed which should overcome many of the disadvantages of th e existing diagnostic tools. The network is based on the authors' earl ier work (Ransing et al. 1995) on representing the causal relationship in the defect-metacause-rootcause form. Although the algorithm is bas ed on the Bayesian analysis, many of the laws of probability have been altered to suit the complexities involved. For example, the notion of conditional probability has been generalized to enable the belief rev ision even in the presence of partial evidence. The inherent presence of the degree of ignorance or uncertainty in the quantification of a r elationship has also been considered. Rigorous constraints, again base d on the laws of probability, have been developed to check the consist ency among the network values. The network is required to be initializ ed with only a few values or the range for the same and then a set of globally consistent values is generated automatically and efficiently. Using the most suitable set of consistent values, the diagnosis is pe rformed using the generalized Bayesian analysis. The network has been tested for a pressure die casting process, however, it is generic in n ature and can also be applied to other manufacturing processes.