REVISTA DYNA NEW TECHNOLOGIES

Key words:

Sistema experto, lógica difusa, red neuronal artificial, análisis metalográfico, Expert system, fuzzy logic, artificial neural network, metallographic analysis

Article type:

ARTICULO DE INVESTIGACION / RESEARCH ARTICLE

Section:

RESEARCH ARTICLES

ABSTRACT:
This paper describes the development of an intelligent integrated system able to estimate the damage on 2.25Cr-1Mo low alloy steel rated according to ASTM A200 grade T22. The system is composed of a fuzzy logic architecture focused to assessing the damage present in the steel and a multi-layer perceptron artificial neural network applied to the recognition of metallographic patterns. The analyzed steel samples were subjected to work of water and steam conduction in high temperature conditions, suffering consequently a physical deterioration that must be constantly monitored from its microstructure. The patterns studied in the microstructure of the steels are: spheroidal pearlite, decarburization and graphite nodules. The results showed that the estimation of the damage and the recognition of patterns in the material were predicted correctly with the system developed in comparison with an expert user with a confidence of 95%.

Keywords: Expert system, fuzzy logic, artificial neural network, metallographic analysis, damages generated by temperature.