DEVELOPMENT OF SUPPORT SYSTEM FOR THE EVA LUATION OF UNIFORMITY IN GRAIN-STRUCTURE BY FUZZY-REASONING

The uniformity in the microstructures of materials is one of the most important factors in the properties and behavior of materials. However , the uniformity has not been characterized well. Therefore, the unifo rmity evaluation is revealed at present from inspection such as sensor y test performed by expert. The evaluation of the uniformity by subjec tive sense inspection of an expert generally includes fuzziness or amb iguity. This paper attempts to construct the objective and quantitativ e evaluation systems based on the technique of fuzzy reasoning for the uniformity on grain structure. The systems are modelled by the fuzzy inference rules with two input variables (U-sp, U-size) and a single o utput variable (sensory evaluation), and are trained by the results of the expert sensory test. The results obtained by the fuzzy reasoning are in good agreement with the results of sensory test by experts. Fur thermore, the avaiability of the present fuzzy reasoning system is con firmed by the comparison of the performance for training data and test data with the statistical method using linear regression analysis.