Fuzzy neural network analysis on the compacted graphite iron with improved tensile and heat transport properties.

In order to find out the most effective method for developing compacted graphite iron with a combination of high tensile strength, ductility and thermal conductivity, the superposed structural effects were investigated by experimental results and the relative significances were ranked on the basis of fuzzy neural network model. The concerned structural parameters consisted of graphite content, vermicularity and microhardness of the matrix. It was found that the relationships between properties and structural parameters become complex due to the mutual disturbances of various characteristics. Irregular and compossible corrections were both observed. The sensitivity level suggested that low microhardness of the matrix and low vermicularity are the optimal directions for improving simultaneously the tensile strength, thermal conductivity and elongation of compacted graphite iron. [ABSTRACT FROM AUTHOR]