Structural Damage Detection Using FRF Data, 2D-PCA, Artificial Neural Networks and Imperialist Competitive Algorithm Simultaneously

In this study, a promising pattern recognition based approach is introduced for structural damage identification using the measured dynamic data. The frequency response function (FRF) is preferably employed as the input of the proposed algorithm since it contains the most information of structural dynamic characteristics. The 2D principal component analysis (2D-PCA) is used to reduce the large size of FRFs data. The output data generated by the 2D-PCA are used to extract the damage indexes for each of the damage scenarios. A dataset of all probable damage indexes is provided; of which 30% are selected to form the train dataset and to be compared with the unknown damage index for an unidentified state of the structure. The sum of absolute errors (SAE) are calculated between the unknown damage index and the selected indexes from the dataset; of which the minimum refers to the most similar damage condition to the unknown one. The artificial neural networks (ANNs) are used to form a smooth function of the SAEs and the imperialist competitive algorithm (ICA) is utilized to minimize this function in order to find the location and severity of the damages of the unknown state of the structure. To validate the proposed method, the damage identification of a truss bridge structure and a two-story frame structure is conducted by considering all the single damage cases as well as multi damage scenarios. In addition, the robustness of the proposed method to measurement noise up to 20% is thoroughly investigated.

Structural Damage Detection Using Signal Pattern-Recognition

Deterioration significantly affects the structure performance and safety. A signal-based pattern-recognition procedure is applied for structural damage detection with a limited number of input/output signals. The method is based on extracting and selecting the sensitive features of the structure response to form a unique pattern for any particular damage scenario, and recognizing the unknown damage pattern against the known database to identify the damage location and level (severity). In this study, two types of transformation algorithms are implemented separately for feature extraction: (1) Continuous Wavelet Transform (CWT); and (2) Wavelet Packet Transform (WPT). Three pattern-matching algorithms are also implemented separately for pattern recognition: (1) correlation, (2) least square distance, and (3) Cosh spectral distance. To demonstrate the validity and accuracy of the procedure, experimental studies are conducted on a simple three-story steel structure. The results show that the features of the signal for different damage scenarios can be uniquely identified by these transformations, and correlation algorithms can best perform pattern recognition to identify the unknown damage pattern. The proposed method can also be used to possibly detect the type of damage. It is suitable for structural health monitoring, especially for online monitoring applications.

Locating Damage In Waveguides From Point FRF Measurements

A method is presented for locating discontinuities in a uniform waveguide from two or more point frequency response functions (FRFs). The phase of the FRF exhibits modulation when plotted against wavenumber because of interference of waves reflected from the discontinuity, and this is related to the distance of the excitation point from the discontinuity. Such discontinuities might be known boundaries or unknown damage sites. An inverse Fourier transform is used to transform from the wavenumber domain to the spatial domain in order to extract the locations of the discontinuities. The use of the transform relies upon knowledge of the dispersion relation for the waveguide. Experimental results are presented for several uniform isotropic beams which were damaged to differing extents by sawing transverse slots in them. The results show the method to be successful in locating the slots.