A novel model is presented for global health monitoring of large structures such as high-rise building structures through adroit integration of 2 signal processing techniques, synchrosqueezed wavelet transform and fast Fourier transform, an unsupervised machine learning technique, the restricted Boltzmann machine, and a recently developed supervised classification algorithm called neural dynamics classification (NDC) algorithm. The model extracts hidden features in the frequency domain of the denoised measured response signals recorded by sensors on different elevations or floors of a structure. The extracted features are used as an input of the NDC to detect and classify the global health of the structure into categories such as healthy, light damage, moderate damage, severe damage, and near collapse. The proposed model is validated using the data obtained from a 3D 1:20 scaled 38-story reinforced concrete building structure. The results are compared with 3 other supervised classification algorithms: k-nearest neighbor (KNN), probabilistic neural networks (PNN), and enhanced PNN (EPNN). NDC, EPNN, PNN, and KNN yield maximum average accuracies of 96%, 94%, 92%, and 82%, respectively. KEYWORDS health monitoring, high-rise building, machine learning, neural dynamics model of Adeli and Park, neural networks, tall building 1 | INTRODUCTION | System identification methodsStructural health monitoring (SHM) can be divided into two classes, vision-based SHM [1] and vibration-based SHM. [2,3] The latter often relies on system identification to evaluate the overall health of the structure and detect the damage location and severity. [4] System identification methods can be divided into two categories: (a) parametric system identification [5][6][7] and (b) nonparametric system identification (NSI). [8,9] Parametric system identification methods determine structural static or dynamic properties such as flexibility [10] or stiffness, [11][12][13] damping, [14] natural frequencies, [15,16] and mode shapes [17] usually assuming a linear behavior. [18] In contrast, NSI methods do not provide the aforementioned structural properties and can model nonlinear dynamic systems more effectively. They provide an estimate of the structural time series response based on a given time series input excitation. An NSI method should be able to detect features in time series responses based on the variation of the input signal. [19,20] These features are usually not meaningful from a structural engineering point of view but can be used for structural system identification and vibration-based SHM. [4] Among recent approaches used for NSI are autoregression moving average exogeneous neural networks, [21] dynamic fuzzy wavelet neural network, [22,23] and synchrosqueezed wavelet transform (SWT). [24,25] 1.2 | Vibration-based SHM In general, vibration-based SHM can be divided into five general steps: (a) sensor selection and installation, [26] (b) data collection, (c) data processing, (d) feature extraction, and (e) damage detection. [27][28][29] Th...

No abstract

No abstract

The keys for the development of an effective classification algorithm are: 1) discovering feature spaces with large margins between clusters and close proximity of the classmates and 2) discovering the smallest number of the features to perform accurate classification. In this paper, a new supervised classification algorithm, called neural dynamic classification (NDC), is presented with the goal of: 1) discovering the most effective feature spaces and 2) finding the optimum number of features required for accurate classification using the patented robust neural dynamic optimization model of Adeli and Park. The new classification algorithm is compared with the probabilistic neural network (PNN), enhanced PNN (EPNN), and support vector machine using two sets of classification problems. The first set consists of five standard benchmark problems. The second set is a large benchmark problem called Mixed National Institute of Standards and Technology database of handwritten digits. In general, NDC yields the most accurate classification results followed by EPNN. A beauty of the new algorithm is the smoothness of convergence curves which is an indication of robustness and good performance of the algorithm. The main aim is to maximize the prediction accuracy.

No abstract