Intelligent computational methods for corrosion damage assessment

Corrosion is one of the damage mechanisms affecting the structural integrity of aging aircraft structures. Various nondestructive inspection (NDI) techniques are being used to obtain images of corroded regions on structures. A computational approach using wavelet transforms and artificial neural networks to analyze and quantify the extent of corrosion damage from the NDI images is described. The wavelet parameters obtained from the images were first used to classify between corroded and uncorroded regions using a clustering algorithm. The corroded regions were further analyzed to obtain the material loss due to corrosion using an artificial neural network model. Experiments were carried out to investigate the developed methods for aircraft panels with engineered corrosion obtained from the Federal Aviation Administration Validation Center in Albuquerque. The results presented indicate that the computational methods developed for corrosion analysis seem to provide reasonable results for estimating material loss due to corrosion damage.