Genetic Optimization Of Material Properties Based On ANFIS Model And It's Application

Grey cast iron is a type of alloys containing more than one elements, and their properties are determineed by it's compositions, melting and pouring conditions in the preparing of melt. At the same time, it requres charging in the lowest cost. This is a common problem which must be coped with in the applications of grey cast iron, and the solving of it is of great practical values.In this paper, by the coupled use of adaptive neuro-fuzzy inference modeling and artificial neural network learning ability, a set of "IF-THEN" rules, which could help us better understanding the basic principles of alloying, have been generated directly from the experimental data. The developed adaptive neuro-fuzzy inference system (ANFIS) could map the relationships between compositions, melting parameters and mechanical properties accurately. In spite of this, the developed ANFIS modeling method provided a tool for knowledge discovering from material data.The above ANFIS model is embed in the routine of genetic optimization process for finding best combination of compositions and process parameters. The proposed ANFIS model based optimization framework provides a newmethodology, which could be used in other data based modeling and optimization practices.By the use of the ANFIS model based optimization method, the design of grey cast iron under prescribed specification and the least cost charging for cupolas melting have been performed. The experimental results show that, the optimization framework could provide useful guides in the grey cast melting practices. The algorithms and software could extend to the similar modeling and optimization practices in other fields, such as oil, metallurgical industry in coping with the compositions and mixing design problems.