Structural health monitoring of a cable-stayed bridge with Bayesian neural networks

In recent years, there has been an increasing interest in permanent observation of the dynamic behaviour of bridges for long-term monitoring purpose. This is due not only to the ageing of a lot of structures, but also for dealing with the increasing complexity of new bridges. The long-term monitoring of bridges produces a huge quantity of data that need to be effectively processed. For this purpose, there has been a growing interest on the application of soft computing methods. In particular, this work deals with the applicability of Bayesian neural networks for the identification of damage of a cable-stayed bridge. The selected structure is a real bridge proposed as benchmark problem by the Asian-Pacific Network of Centers for Research in Smart Structure Technology (ANCRiSST). They shared data coming from the long-term monitoring of the bridge with the structural health monitoring community in order to assess the current progress on damage detection and identification methods with a full-scale example. The data set includes vibration data before and after the bridge was damaged, so they are useful for testing new approaches for damage detection. In the first part of the paper, the Bayesian neural network model is discussed; then in the second part, a Bayesian neural network procedure for damage detection has been tested. The proposed method is able to detect anomalies on the behaviour of the structure, which can be related to the presence of damage. In order to obtain a confirmation of the obtained results, in the last part of the paper, they are compared with those obtained by using a traditional approach for vibration-based structural identification.     

A Structural Damage Neural Network Monitoring System

Abstract: Traditional methods for structural monitoring and damage assessment have been implemented largely through visual inspection and on-site tests. A system for automating this process should be able to record the various signatures of the structure to be monitored and issue a warning signal if there is a damage-related change in those signatures. In this paper, a general system for structural damage monitoring is proposed based on observations of other researchers and the results obtained from a case study of a physical and analytical model of a five-story steel frame. The proposed diagnostic system utilizes neural networks for identifying the damage associated with changes in structural signatures. The system is independent of the type of signatures used for monitoring. Two sets of neural networks were developed. The first set was trained with the results of a series of shaking-table experiments, while the second set was trained with the output produced from a finite-element model of the same test structure. The results show that the proposed system provides a suitable framework for automatic structural monitoring.     

神经网络在结构工程健康监测中的应用探讨

建立结构损伤诊断子系统是建立大型工程结构智能健康监测专家系统的核心问题。人工神经网络技术可以实现结构损伤的自动识别与定位,具有广阔的应用前景。本文介绍基于人工神经网络的两级损伤识别策略,并对采用人工神经网络进行结构损伤诊断的网络输入参数与网络结构选择等关键问题进行了探讨。     

Structural Health Monitoring via Measured Ritz Vectors Utilizing Artificial Neural Networks

Abstract:   A pattern recognition approach for structural health monitoring (SHM) is presented that uses damage-induced changes in Ritz vectors as the features to characterize the damage patterns defined by the corresponding locations and severity of damage. Unlike most other pattern recognition methods, an artificial neural network (ANN) technique is employed as a tool for systematically identifying the damage pattern corresponding to an observed feature. An important aspect of using an ANN is its design but this is usually skipped in the literature on ANN-based SHM. The design of an ANN has significant effects on both the training and performance of the ANN. As the multi-layer perceptron ANN model is adopted in this work, ANN design refers to the selection of the number of hidden layers and the number of neurons in each hidden layer. A design method based on a Bayesian probabilistic approach for model selection is proposed. The combination of the pattern recognition method and the Bayesian ANN design method forms a practical SHM methodology. A truss model is employed to demonstrate the proposed methodology.     

Decentralized Parametric Damage Detection Based on Neural Networks

In this paper, based on the concept of decentralized information structures and artificial neural networks, a decentralized parametric identification method for damage detection of structures with multi-degrees-of-freedom (MDOF) is conducted. First, a decentralized approach is presented for damage detection of substructures of an MDOF structure system by using neural networks. The displacement and velocity measurements from a substructure of a healthy structure system and the restoring force corresponding to this substructure are used to train the decentralized detection neural networks for the purpose of identifying the corresponding substructure. By using the trained decentralized detection neural networks, the difference of the interstory restoring force between the damaged substructures and the undamaged substructures can be calculated. An evaluation index, that is, relative root mean square (RRMS) error, is presented to evaluate the condition of each substructure for the purpose of health monitoring. Although neural networks have been widely used for nonparametric identification, in this paper, the decentralized parametric evaluation neural networks for substructures are trained for parametric identification. Based on the trained decentralized parametric evaluation neural networks and the RRMS error of substructures, the structural parameter of stiffness of each subsystem can be forecast with high accuracy. The effectiveness of the decentralized parametric identification is evaluated through numerical simulations. It is shown that the decentralized parametric evaluation method has the potential of being a practical tool for a damage detection methodology applied to structure-unknown smart civil structures.     

A Bayesian Probabilistic Approach for Acoustic Emission‐Based Rail Condition Assessment

The investigation described in this article aims at developing a Bayesian‐based approach for probabilistic assessment of rail health condition using acoustic emission monitoring data. It comprises the following three phases: (i) formulation of a frequency‐domain structural health index (SHI), via a linear transformation method, tailored to damage‐sensitive frequency bandwidth; (ii) establishment of data‐driven reference models, using Bayesian regression about the real and imaginary parts of the SHI derived with monitoring data from the intact rail; and (iii) quantitative evaluation of discrimination between the new observations representative of current rail health condition and the baseline model predictions in terms of Bayes factor. If the deviation of the new observations from the predictions is within an acceptable tolerance, no damage is flagged, and the new data are further used to update and refine the reference models. If the observations deviate substantially from the model predictions in a probabilistic sense, damage is signaled, damage severity is quantified, and damage location determined. The proposed approach is examined by using field monitoring data acquired from an instrumented railway turnout, and the coincidence between the assessment results and the actual health conditions demonstrates its effectiveness in damage detection, localization, and quantification.     

Large-scale structural health monitoring using composite recurrent neural networks and grid environments

The demand for resilient and smart structures has been rapidly increasing in recent decades. With the occurrence of the big data revolution, research on data-driven structural health monitoring (SHM) has gained traction in the civil engineering community. Unsupervised learning, in particular, can be directly employed solely using field-acquired data. However, the majority of unsupervised learning SHM research focuses on detecting damage in simple structures or components and possibly low-resolution damage localization. In this study, an unsupervised learning, novelty detection framework for detecting and localizing damage in large-scale structures is proposed. The framework relies on a 5D, time-dependent grid environment and a novel spatiotemporal composite autoencoder network. This network is a hybrid of autoencoder convolutional neural networks and long short-term memory networks. A 10-story, 10-bay, numerical structure is used to evaluate the proposed framework damage diagnosis capabilities. The framework was successful in diagnosing the structure health state with average accuracies of 93% and 85% for damage detection and localization, respectively.     

Computational framework for risk-based planning of inspections,maintenance and condition monitoring using discrete Bayesian networks

This paper presents a computational framework for risk-based planning of inspections and repairs for deteriorating components. Two distinct types of decision rules are used to model decisions: simple decision rules that depend on constants or observed variables (e.g. inspection outcome), and advanced decision rules that depend on variables found using Bayesian updating (e.g. probability of failure). Two decision models are developed, both relying on dynamic Bayesian networks (dBNs) for deterioration modelling. For simple decision rules, dBNs are used directly for exact assessment of total expected life-cycle costs. For advanced decision rules, simulations are performed to estimate the expected costs, and dBNs are used within the simulations for decision-making. Information from inspections and condition monitoring are included if available. An example in the paper demonstrates the framework and the implemented strategies and decision rules, including various types of condition-based maintenance. The strategies using advanced decision rules lead to reduced costs compared to the simple decision rules when condition monitoring is applied, and the value of condition monitoring is estimated by comparing the lowest costs obtained with and without condition monitoring.     

Neural Network Approaches to Aid Simple Truss Design Problems

Abstract: Diverse problems in engineering may be solved accurately with computers. In structural engineering, many solution techniques exist. Over the past few years, neural networks have evolved as a new computing paradigm, and many engineering applications have been studied. This paper describes configuring and training of a neural network for a truss design application and explores the possible roles for neural networks in structural design problems. The specific problem considered is a simple truss design where, given a geometry and a loading, economical cross-sectional areas of all the members are to be selected. For this problem, a two-layer neural network is trained using the back-propagation algorithm with patterns representing optimal designs for diverse loading conditions. The performance of the trained neural network is evaluated with a sample problem.     

神经网络在土木工程领域的应用

分析了神经网络在结构工程、建筑管理等土木工程领域的应用研究现状及其发展方向 ,以一个建筑管理领域的实际问题为例提出了将神经网络用于解决土木工程领域问题的一般方法和步骤     

Structural parameter identification and damage detection for a steel structure using a two-stage finite element model updating method

A two-stage eigensensitivity-based finite element (FE) model updating procedure is developed for structural parameter identification and damage detection for the IASC-ASCE structural health monitoring benchmark steel structure on the basis of ambient vibration measurements. In the first stage, both the weighted least squares and Bayesian estimation methods are adopted for the identification of the connection stiffness of beam-column joints and Young’s modulus of the structure; then the damage detection is conducted via the FE model updating procedure for detecting damaged braces with different damage patterns of the structure. Comparisons between the FE model updated results and the experimental data show that the eigensensitivity-based FE model updating procedure is an effective tool for structural parameter identification and damage detection for steel frame structures.     

Nonparametric Identification of a Building Structure from Experimental Data Using Wavelet Neural Network

Abstract:   This study presents a wavelet neural network-based approach to dynamically identifying and modeling a building structure. By combining wavelet decomposition and artificial neural networks (ANN), wavelet neural networks (WNN) are used for solving chaotic signal processing. The basic operations and training method of wavelet neural networks are briefly introduced, since these networks can approximate universal functions. The feasibility of structural behavior modeling and the possibility of structural health monitoring using wavelet neural networks are investigated. The practical application of a wavelet neural network to the structural dynamic modeling of a building frame in shaking tests is considered in an example. Structural acceleration responses under various levels of the strength of the Kobe earthquake were used to train and then test the WNNs. The results reveal that the WNNs not only identify the structural dynamic model, but also can be applied to monitor the health condition of a building structure under strong external excitation.     

采用扩展的贝叶斯人工神经网络选择结构损伤监测模型

对采用规则的动态数据进行结构损伤监测时,模式识别是一个有效的方法,人工神经网络作为匹配模式特征的系统方式广泛应用于模式识别研究中。人工神经网络设计是影响模型识别性能和效率的最基本因素。由Lam等人提出的贝叶斯人工神经网络设计法则为单隐层前馈人工神经网络确定大量隐性神经单元提供了严格的数学手段。本文的第一个目标是对贝叶斯人工神经网络设计法则进行拓展,包括选择隐层中神经单元的传递函数。所提出的法则具有高效的特点,适用于实时人工神经网络设计。目前,许多人工神经网络设计技术需要在训练前已知人工神经网络模型的类型,因此,最基本的问题是自动选择优化的人工神经网络模型类型的技术。由于模型参数和Ritz向量一般用于描述模式的特征,本文的第二个目标是采用模式识别对结构损伤监测中这两个模式特征进行比较。为了清楚判断这两个特征,研究中采用了IASC-ASCE准则。研究结果显示:采用模型参数进行训练的人工神经网络性能略优于采用Ritz向量进行训练的人工神经网络性能。     

Pipe network leak detection: comparison between statistical and machine learning techniques

ABSTRACT

This paper investigates an inverse analysis technique to find leaks in water networks and compares different solution strategies. Although a number of strategies have been proposed by different authors to identify leaks on a vast selection of pipe networks, limited research has been done to compare strategies and point out their weakness. Three strategies, a Bayesian probabilistic analysis, a support vector machine and, an artificial neural network were combined with the inverse analysis technique on different numerical and experimental networks to point out each strategy’s weakness. Two numerical networks are investigated and one experimental network. It is shown that the Bayesian probabilistic analysis struggles to find unique solutions when a few observations are available, while the support vector machine and the artificial neural network struggle when only flow measurements are available. Additionally it is shown that the artificial neural network struggles to estimate unique solutions for leak size and location.     

Monitoring Structural Health Using a Probabilistic Measure

A Bayesian probabilistic methodology for structural health monitoring is presented. The method uses a sequence of identified modal parameter data sets to continually compute the probability of damage. In this approach, a high likelihood of a reduction in model stiffness at a location is taken as a proxy for damage at the corresponding structural location. The concept extends the idea of using as indicators of damage the changes in model parameters identified using a linear finite-element model and modal parameter data sets from the structure in undamaged and possibly damaged states. This extension is needed because of uncertainties in the updated model parameters that in practice obscure health assessment. These uncertainties arise due to effects such as variation in the identified modal parameters in the absence of damage, as well as unavoidable model error. The method is illustrated by simulating on-line monitoring, wherein specified modal parameters are identified on a regular basis and the probability of damage for each substructure is continually updated. Examples are given for abrupt onset of damage and progressive deterioration.     

Neural Networks in Civil Engineering: 1989–2000

The first journal article on neural network application in civil/structural engineering was published by in this journal in 1989. This article reviews neural network articles published in archival research journals since then. The emphasis of the review is on the two fields of structural engineering and construction engineering and management. Neural networks articles published in other civil engineering areas are also reviewed, including environmental and water resources engineering, traffic engineering, highway engineering, and geotechnical engineering. The great majority of civil engineering applications of neural networks are based on the simple backpropagation algorithm. Applications of other recent, more powerful and efficient neural networks models are also reviewed. Recent works on integration of neural networks with other computing paradigms such as genetic algorithm, fuzzy logic, and wavelet to enhance the performance of neural network models are presented.     

人工神经网络在结构工程中的应用和发展

依据神经网络原理及其自身的特点，对其应用在结构优化设计、结构分析及可靠度分析等方面进行了综述和研究，并在此基础上分析了神经网络在结构工程中的研究方向。     

Active learning structural model updating of a multisensory system based on Kriging method and Bayesian inference

Model updating techniques are often applied to calibrate the numerical models of bridges using structural health monitoring data. The updated models can facilitate damage assessment and prediction of responses under extreme loading conditions. Some researchers have adopted surrogate models, for example, Kriging approach, to reduce the computations, while others have quantified uncertainties with Bayesian inference. It is desirable to further improve the efficiency and robustness of the Kriging-based model updating approach and analytically evaluate its uncertainties. An active learning structural model updating method is proposed based on the Kriging method. The expected feasibility learning function is extended for model updating using a Bayesian objective function. The uncertainties can be quantified through a derived likelihood function. The case study for verification involves a multisensory vehicle-bridge system comprising only two sensors, with one installed on a vehicle parked temporarily on the bridge and another mounted directly on the bridge. The proposed algorithm is utilized for damage detection of two beams numerically and an aluminum model beam experimentally. The proposed method can achieve satisfactory accuracy in identifying damage with much less data, compared with the general Kriging model updating technique. Both the computation and instrumentation can be reduced for structural health monitoring and model updating.     

Neural Networks in Structural Engineering

In the past few years literature on computational civil engineering has concentrated primarily on artificial intelligence (Al) applications involving expert system technology. This article discusses a different Al approach involving neural networks. Unlike their expert system counterparts, neural networks can be trained based on observed information. These systems exhibit a learning and memory capability similar to that of the human brain, a fact due to their simplified modeling of the brain's biological function. This article presents an introduction to neural network technology as it applies to structural engineering applications. Differing network types are discussed. A back-propagation learning algorithm is presented. The article concludes with a demonstration of the potential of the neural network approach. The demonstration involves three structural engineering problems. The first problem involves pattern recognition; the second, a simple concrete beam design; and the third, a rectangular plate analysis. The pattern recognition problem demonstrates a solution which would otherwise be difficult to code in a conventional program. The concrete beam problem indicates that typical design decisions can be made by neural networks. The last problem demonstrates that numerically complex solutions can be estimated almost instantaneously with a neural network.     

A hybrid neural network strategy for identification of structural parameters

This paper presents a multistage identification scheme for structural damage detection using modal data. Previous studies of damage assessment using neural networks mostly involved training a backpropagation neural network (BPN) to learn damage patterns that were obtained either experimentally or by simulation for different damage cases. Damage identification for large structures, especially those involving multiple member damage, could result in large training data sets that require a large BPN and consequently greater computational effort. The proposed scheme involves using a counterpropagation neural network (CPN) in the first stage for sorting the training data into clusters and giving an approximate guess of the damage extent within a very short time. After an approximate estimate of the damage is obtained, a new set of training patterns of reduced size is generated using the CPN prediction. In the second stage, a BPN trained with the Levenberg–Marquardt algorithm is used to learn the new training data and predict a more accurate result. A superior convergence and a substantial decrease in central processing unit (CPU) time are observed for three numerical examples.