Structural damage detection with wavelet support vector machine: introduction and applications

Abstract

Based on wavelet packet decomposition and conditions of the support vector kernel function, a non-linear wavelet basis is introduced to construct the kernel function of support vector machine (SVM), and a tighten wavelet support vector machine (WSVM), which has strong generalization ability is obtained. Wavelet packet decomposition is applied to the structural response signals under ambient vibration, feature vectors are obtained by feature extraction according to wavelet energy spectrum. The feature vectors are used for training and classification as the inputs of the tighten SVM. Hence, a new structural damage detection method called complete WSVM is established. This method is used to a single-layer spherical lattice dome for damage detection. The structural damage location and degree can be detected and classified, and the result is highly accurate. This approach has some advantages, such as engineering oriented, low cost and convenient. ensure the accuracy [6]. Flexibility matrix method is hard to reflect structural local damages [7]. Modal strain energy method can obtain different results using different test modals [8].Structural health monitoring (SHM) is a multi-disciplinary and integrated technology; hence, it is difficult to resolve many practical problems for large and complex structures based on merely vibration testing. Improving the finite element updating technology for health monitoring and state evaluation, combining modern signal analysis technology and soft computing theory, mining structural characteristic data deeply, realizing the real-time, online dynamic monitoring and control, is the developmental direction for SHM.Recently, many advanced technology and intelligent methods, such as digital filter technology, wavelet transform (WT) analysis [9][10][11][12], artificial neural networks (ANNs) [13,14] and genetic algorithm (GA), are studied to detect the structural global and local damage information, and to optimize the SHM system.Wavelet packet transform (WPT) is a new method of signal processing that has recently been applied to various science and engineering fields with great success. The specific local properties of wavelet packet can be particularly useful to describe signals with sharp spikes or discontinuities. It is effective for the structural damage signals process by wavelet packet because its multi-resolution analysis capability and time-frequency localization. As an effective time-frequency tool, wavelets could detect whether damage has occurred but it is not effective to detect different damage states. Intelligent tools are needed to solve the damage classification problem, which includes damage location and degree.ANNs have been applied in structural damage detection as generalization or classification problems based on learning pattern from examples or empirical data model [15]. The ANN method for structural damage detection has some advantages such as strong self-adaptive and fault-tolerant capabilities, and it has been approved to be effective for some elements and simple...

Structural Damage Information Fusion Based on Soft Computing

“…The intelligent methods such as support vector machine (SVM) can effectively improve the accuracy of structural damage identification [15]. The original structure can be divided into multiple substructures, and then the detail damage can be identified using support vector machine (SVM) for each substructure [16]. If the number of the components in the substructure is still large, the featurelevel fusion for damage information can be carried out by the making full use of the multiple sensors placed on the substructure.…”

“…Based on above equations, the parameters a j i , b j i , and w j could be adjusted in the following seven steps [16].…”

“…The damage element number is 29, 105, 126, 32, and 91, as shown in Figure 3. Wavelet packet decomposition is applied to the structural response signals under ambient vibration, and feature vectors are obtained by feature extraction method with energy-damage state in [16]. Four group of feature vectors are used for training and classification as the inputs of the tighten support vector machine.…”

A substructure approach to local damage detection of shear structure

“…Saito [9] used the probabilistic method to estimate structural damage, which includes identifying the damage existence, location, and severity, by utilizing the vibration data observed before and after a building was damaged. He and Yan [10] introduced a non-linear wavelet basis. Wavelet packet decomposition is applied to the structural response signals under ambient vibration, and feature extraction is used to obtain the feature vectors according to the wavelet energy spectrum.…”

Nearest neighbor and learning vector quantization classification for damage detection using time series analysis

“…Such methods can broadly be divided into supervised techniques when data from damaged structures are available for training of pattern recognition algorithms, and unsupervised when they are not. Widely employed supervised techniques include artificial neural networks [9][10][11], genetic algorithms [12], support vector machines [13,14] and discriminant analysis [15]. A number of unsupervised methods have also been investigated and these include outlier detection [16][17][18], control chart analysis [3,19] and clustering analysis [5,20,21].…”