Energy dissipation‐based material deterioration assessment using random decrement technique and convolutional neural network: A case study of Saigon bridge in Ho Chi Minh City, Vietnam

Abstract

Summary The bridge structures must work under random and complex excitation conditions. The vibration response of these structures includes two main components as a determining component and a stochastic component. Thus, using vibration data, the structural health monitoring (SHM) process for these structures requires eliminating random parts impact. The random decrement (RD) signature, a known technique to serve this requirement, is applied to analyze the bridge's vibrations under the ambient load (random excitation) in this study. The nonlinear viscoelastic model is used to evaluate the energy dissipation of material. Then, a new damage index, called the loss factor function (LF), determined from the power spectral density (PSD) of vibration modes, is used to assess material deterioration. In fact, only some vibration modes of structures that occur with large amplitude are considered to be determined. Therefore, the article evaluates some of the first vibration modes' energy dissipation to monitor the change of material's mechanical properties. The vibration of the Saigon bridge under actual traffic loadings over 9 years is used as an illustrative example. One single beam with the same modes of the bridge span is modeled and used to extract numerical data to train convolutional neural network (CNN). CNN is widely known at the current time with outstanding capabilities in the typical image classification. The distribution of these LF values is overall evaluated by using a featured image built from their contour plot images. With the supervised learning algorithm, the proposed network is trained to assess the deterioration level of materials. The output is the label corresponding to the deterioration level, and the inputs are featured images about energy dissipation.

Deep Learning for Structural Health Monitoring: Data, Algorithms, Applications, Challenges, and Trends

“…They compared four different DL methods, MLP, LSTM, 1DCNN, and 2DCNN, for damage identification, localization, and evaluation, with the 2DCNN providing the best results. Toan et al [227] proposed a CNN to evaluate energy dissipation to monitor the material degradation of the Saigon Bridge in Vietnam.…”

Damage assessment in beam-like structures by correlation of spectrum using machine learning

“…A predictive model using Machine learning algorithms in the problem of damage identification has been proposed in many studies. Machine learning as neural network pattern recognition (NNPR) serves the damage detection process in beams [11][12][13][14][15] or conditional assessment in bridges [3,16,17]. Besides, machine learning methods are also applied in other structures, such as plates, pipes, and frames [18][19][20].…”

“…Several performance metrics are calculated from these four results to show the performance of the decision tree:  Accuracy: The proportion of correctly classified samples in the data set. It is calculated as follows:TP+TN accuracy= (TP+TN+FP+FN)(17)…”