Exploratory Investigations for Intelligent Damage Prognosis using Hidden Markov Models

Structural damage occurs due to structural overloading or due to environmental conditions or combined effects. Extensive research on structural health monitoring, damage diagnosis and damage pattern recognition has been developed over the last two decades. Recently, damage prognosis has become a new research focus that tries to use damage diagnosis knowledge to inform the responsible authorities on the expected remaining service life of the structure such that efficient maintenance operations can be scheduled. We demonstrate here the use of hidden Markov models to model damage accumulation and propagation in structures and to predict the residual service life of the structure given damage diagnosis observations. The proposed model is described and is tested using damage diagnosis simulated data of a pre-stressed concrete bridge