Novel modelling approach based on fuzzy set theory to graded problem diagnosis and its application

Abstract A modelling and decision approach to multi-level trouble diagnosis is presented on the basis of fuzzy set theory and a pattern recognition analysis method. The diagnostic results from the analysis method of graded trouble sources of trouble are very useful to reasonably select measures to remove the problems diagnosed. A heuristic searching method is used in the real-time scheme for deciding the diagnosis, which is of great significance in saving operation time for on-line diagnosis. Meanwhile, to test the proposed approach, a model for the diagnosis to contamination problems for a specific industrial antibiotic fermentation process and the situation of its industrial tests are presented. The results of the tests are satisfactory.