Artificial neural network corrosion modeling for metals in an equatorial climate

摘要: Abstract The majority of the metals used in distribution and transmission electric energy lines, such as cables, towers accessories are susceptible to corrosion degradation process. For that reason, studying factors influence atmospheric is an important issue. In this paper, artificial neural network model was developed with linear sigmoidal functions, aiming predict low-carbon steel, copper aluminum rates according environmental parameters area Sao Luis – Maranhao, Brazil. along “702 II –Presidente Dutra” 500 kV line, located equatorial region, employed for purpose. A specific methodology determine local corrosivity rate these metals. Five stations (ACS) were installed 702 line extension 200 km. Along meteorological data, pollutants collected analyzed during a period two years. same period, specimens exposed atmosphere periodically evaluation. obtained results indicate can be good estimator.