Structural health monitoring for condition assessment using efficient supervised learning techniques

Pattern recognition is known as one of the applicable technologies for structural health monitoring (SHM) based on statistical characteristics extracted from raw vibration data. Structural condition assessment is an important step in SHM since either linear or nonlinear changes in the relevant properties may adversely alter the behavior of any structure. It looks therefore necessary to adopt efficient and robust approaches for the classification of different structural conditions using features extracted from the said raw vibration data. To achieve this goal, it is essential to correctly distinguish a normal or undamaged state of the structure, from an abnormal or damaged one. The primary aim of this work is to present and compare efficient classification methods using feature selection techniques to classify the structural conditions, even characterized by different levels of damage severity. All of the utilized classifiers need a training set pertinent to the undamaged and (foreseen) damaged conditions of the structure, as well as known class labels to be adopted in a supervised learning strategy. Autoregressive (AR) modeling and principal component analysis (PCA) are used in an effort to extract the features for the classification process. The performance and accuracy of the considered classification methods are assessed through a numerical benchmark concrete beam and an experimental benchmark laboratory frame.