Classifying Damages in Engineering Material Using Bayesian Networks

Mohamed Addin, Addin OsmanNajran University, Kingdom of Saudi Arabia and University PutraMalaysia, Malaysia.

Since its foundation in 1980s, Bayesian networks have been widely and successfully implemented in many research and industrial areas. Nevertheless, they have not been thoroughly investigated and implemented for damage detection in engineering materials. This book provides a through introduction to Bayesian networks as a competitive probabilistic graphical model in general and as a classification tool (the Naïve bayes classifier) in particular for damage detection in engineering material. Since the feature extraction is essential for the classifiers, the book introduces the f -folds feature extraction algorithm. The derivation of the algorithm is based on empirical study on a data set, which represents voltage amplitudes of Lamb-waves produced and collected by sensors and actuators mounted on the surface of quasi-isotropic graphite/epoxy laminates contain different artificial damages.