Neural network prediction of stress corrosion cracking
H.M.G Smets, Walter Bogaerts · Materials Performance · 1992
Based on case histories of service experience with austenitic stainless steels in chloride-bearing water, it is possible to predict stress corrosion risk in an intelligent, automated way by means of neural network technology. Two neural networks to forecast chloride-induced stress corrosion cracking risk of type 304 stainless steel have been developed. One network reflects temperature and chloride concentration dependency. The other considers the influence of oxygen and chloride concentrations in high-temperature water