An improved CSS for damage detection of truss structures using changes in natural frequencies and mode shapes

Non-destructive structural damage identification can be carried out using the difference between a structure’s characteristics before and after a catastrophic event. An approach is to formulate the problem as an inverse optimization problem, in which the amounts of damage to each element are considered as the optimization variables. The objective is to set these variables such that the characteristics of the model correspond to the experimentally measured characteristics of the actual damaged structure. Since the structures are usually symmetric, this is an optimization problem with several global optimal solutions each representing a probable state of damage, where unlike many other optimization problems, it is not enough to merely find one of these optimal solutions; it is important to capture all such possible states and to compare them. In this paper, structural damage detection of planar and spatial trusses using the changes in structures’ natural frequencies and mode shapes is addressed. An improved Charged System Search algorithm is developed and utilized to tackle the problem of finding as many global optimal solutions as possible in a single run. A 10-bar planar truss and a 72-bar spatial truss are considered as numerical examples. Experimental results show that it is important to incorporate mode shapes in order to determine the actual damage scenario among other possibilities.     

基于单元模态应变能和剩余模态力的结构损伤识别方法在某塔式桁架中的应用

为快速、准确定位工程结构损伤位置,有效提高工程结构安全性能和使用寿命,以某塔式桁架结构为研究对象,运用单元模态应变能法和剩余模态力法对其进行损伤识别.利用MSC Marc对该桁架完整结构和几种不同损伤程度下的损伤结构进行模态分析,通过MATLAB编程从模态分析结果中提取这些结构的模态参数,计算损伤结构单元模态应变能的变化率和损伤结构各节点自由度对应的剩余模态力,并进行结构损伤识别.结果表明单元模态应变能变化率和剩余模态力是有效和准确的结构损伤标志量.     

基于神经网络模型的桥梁结构损伤识别

给出一种神经网络方法在钢桥结构损伤检测中的应用。着重讨论了网络设计和学习算法问题。网络结构模拟桁架桥,训练样本从多个损伤区域产生。仿真表明,本算法只需少量的结构参数就可得到较满意的辨识结果,且适用于现场测量数据不精确或不能完全已知的情况,具有较好的工程实用性。     

CPN神经网络及其在结构识别中的应用

本文建立了基于ＣＰＮ（ＣｏｕｎｔｅｒｐｒｏｐａｇａｔｉｏｎＮｅｕｒａｌＮｅｔｗｏｒｋ）神经网络的计算力学反问题分析方法,并将其应用于薄板振动系统识别。研究结果表明,神经网络计算是工程结构分析中一种很有发展潜力的新方法。     

A swift technique for damage detection of determinate truss structures (2)

This paper introduces a novel and robust probable statistical approach for the applied damage detection of determinate truss structures. This technique involves two steps; the first is called most probable damaged element identification step and the second is called probable damage severity prediction step. In the first step, a new index based on modal residual forces plays a major role to independently identify damage-suspected elements for each considered mode. Then among them, the elements, the most probable to damage, are extracted. In the second step, the probable damage severity for each most probable damaged element is individually predicted using a novel statistical approach. Finally, to justify the validity and robustness of the technique, three commonly used bridge trusses including a 29-bar Pratt truss, a 29-bar Warren truss, and finally, a 37-bar K truss under different damage scenarios are thoroughly studied while their modal parameters are corrupted by noise. The obtained results indicate that the method is innovatively capable of swiftly predicting, for determinate truss structures, not only damaged elements but also their damage severities by carrying out solely few structural analyses.

     

Functional link neural network approach to solve structural system identification problems

System identification problems are generally inverse vibration problems. Sometimes it is difficult to handle the inverse problems by traditional methods and classical artificial neural network. As such, the objective of this paper is to identify structural parameters by developing a novel functional link neural network (FLNN) model. FLNN model is more efficient than multi-layer neural network (MNN) as computation is less because hidden layer is not required. Here, single-layer neural network with multi-input and multi-output with feed-forward neural network model and principle of error back propagation has been used to identify structural parameters. The hidden layer is excluded by enlarging the input patterns with the help of Legendre and Hermite polynomials. Comparison of results among MNN, Legendre neural network, Hermite neural network and desired is considered and it is found that FLNN models are more effective than MNN.

     

Damage detection in Timoshenko beam structures by multilayer perceptron and radial basis function networks

This study investigates the efficiency of artificial neural networks (ANNs) in health monitoring of pristine and damaged beam-like structures. Beam modeling is based on Timoshenko theory. Two commonly used network models, multilayer perceptron (MLP) and radial basis neural network (RBNN), are used. Beam material and geometrical properties, beam end conditions and dynamically obtained data are used as input to the neural networks. The combinations of these parameters yield umpteenth input data. Therefore, to examine the effectiveness of ANNs, the frequency of intact beams is first tried to be determined by the network models, given the material and geometrical characteristics of beam elements and support conditions. The methodology to compute the vibrational data utilized in training the networks is provided. Showing the robustness of network models, the second stage of the study is carried out. At this stage, the crack parameters (e.g. the location and severity of crack) are estimated by the ANNs using the beam properties, beam end conditions and vibrational data, which consist of natural frequencies and mode shape rotation values. Despite the multiplexed input data, no data reduction schemes or multistage computations are executed in training and validation of neural network models. As a result of analysis runs, the optimal MLP and RBNN models are determined. Comparison of these models shows that the optimal RBNN algorithm performs better. The effectiveness of optimal ANN models in the presence of noise is also presented. As a conclusion, the trained network can be used as a diagnosis method in structural health monitoring of beam-like structures.     

Damage assessment of structures using hybrid neuro-genetic algorithm

Techniques for detecting elemental level damage using the traditional methods receive the setback because of the difficulties in formulating the problems mathematically, specially in case of inverse problems. Artificial neural networks (ANN) have been proved to be an effective alternative for solving the inverse problems because of the pattern-matching capability. But there is no specific recommendation on suitable design of network for different structures and generally the parameters are selected by trial and error, which restricts the approach context dependent. A hybrid neuro-genetic algorithm is proposed in order to automate the design of neural network for different type of structures. The neural network is trained considering the frequency and strain as input parameter and the location and amount of damage as output parameter. The performance is demonstrated using two test problems: (i) clamped-free beam and (ii) plane frame. The outcomes of the results are quite encouraging and prove the robustness of the proposed damage assessment algorithm.     

Configuring and enhancing measurement systems for damage identification

Engineers often decide to measure structures upon signs of damage to determine its extent and its location. Measurement locations, sensor types and numbers of sensors are selected based on judgment and experience. Rational and systematic methods for evaluating structural performance can help make better decisions. This paper proposes strategies for supporting two measurement tasks related to structural health monitoring - (1) installing an initial measurement system and (2) enhancing measurement systems for subsequent measurements once data interpretation has occurred. The strategies are based on previous research into system identification using multiple models. A global optimization approach is used to design the initial measurement system. Then a greedy strategy is used to select measurement locations with maximum entropy among candidate model predictions. Two bridges are used to illustrate the proposed methodology. First, a railway truss bridge in Zangenberg, Germany, is examined. For illustration purposes, the model space is reduced by assuming only a few types of possible damage in the truss bridge. The approach is then applied to the Schwandbach bridge in Switzerland, where a broad set of damage scenarios is evaluated. For the truss bridge, the approach correctly identifies the damage that represents the behaviour of the structure. For the Schwandbach bridge, the approach is able to significantly reduce the number of candidate models. Values of candidate model parameters are also useful for planning inspection and eventual repair.     

A parallel implementation of the backward error propagation neural network training algorithm: experiments in event identification.

An artificial neural-network-based (ANN) event detection and alarm generation system has been developed to aid clinicians in the identification of critical events commonly occurring in the anesthesia breathing circuit. To detect breathing circuit problems, the system monitored CO2 gas concentration, gas flow, and airway pressure. Various parameters were extracted from each of these input waveforms and fed into an artificial neural network. To develop truly robust ANNs, investigators are required to train their networks on large training data sets, requiring enormous computing power. We implemented a parallel version of the backward error propagation neural network training algorithm in the widely portable parallel programming language C-Linda. A maximum speedup of 4.06 was obtained with six processors. This speedup represents a reduction in total run-time from 6.4 to 1.5 h. By reducing the total run time of the computation through parallelism, we were able to optimize many of the neural network's initial parameters. We conclude that use of the master-worker model of parallel computation is an excellent method for speeding up the backward error propagation neural network training algorithm.     

A structural defect identification approach based on a continuum damage model

This paper introduces a structural identification technique built on finite element (FE) model updating. The FE model is parameterized by a structural parameter that continuously describes the damage in the structure, and besides, an evolution equation of this damage parameter is presented. The model updating is accomplished by determining the subset of this damage parameters that minimizes a global error derived from the dynamic residue vectors, which is obtained by introducing the experimental modal properties into the original model eigenproblem. A mode-shape projection technique is used in order to achieve compatibility between the dimension of the experimental and analytical models. The adjusted model maintains basic properties of the analytical model as the sparsity and the symmetry, which plays an important role in model updating-based damage identification. The verification and assessment of the current structural defect identification is performed on a analytically derived bidimensional truss structure and on a cantilever bidimensional Euler–Bernouilli beam through a virtual test simulator. This simulator is used to realistically simulate the corrupting effects of noise, filtering, digital sampling and truncation of the modal spectrum. The eigensystem realization algorithm along with the common-based normalized system identification were utilized to obtain the required natural frequencies and mode shapes.     

Model updating of multistory shear buildings for simultaneous identification of mass,stiffness and damping matrices using two different soft-computing methods

In this research, two novel methods for simultaneous identification of mass–damping–stiffness of shear buildings are proposed. The first method presents a procedure to estimate the natural frequencies, modal damping ratios, and modal shapes of shear buildings from their forced vibration responses. To estimate the coefficient matrices of a state-space model, an auto-regressive exogenous excitation (ARX) model cooperating with a neural network concept is employed. The modal parameters of the structure are then evaluated from the eigenparameters of the coefficient matrix of the model. Finally, modal parameters are used to identify the physical/structural (i.e., mass, damping, and stiffness) matrices of the structure. In the second method, a direct strategy of physical/structural identification is developed from the dynamic responses of the structure without any eigenvalue analysis or optimization processes that are usually necessary in inverse problems. This method modifies the governing equations of motion based on relative responses of consecutive stories such that the new set of equations can be implemented in a cluster of artificial neural networks. The number of neural networks is equal to the number of degree-of-freedom of the structure. It is shown the noise effects may partially be eliminated by using high-order finite impulse response (FIR) filters in both methods. Finally, the feasibility and accuracy of the presented model updating methods are examined through numerical studies on multistory shear buildings using the simulated records with various noise levels. The excellent agreement of the obtained results with those of the finite element models shows the feasibility of the proposed methods.     

Fractional-Neuro-Optimizer: A Neural-Network-Based Optimization Method

Up to now, there have been many attempts in the use of artificial neural networks (ANNs) for solving optimization problems and some types of ANNs, such as Hopfield network and Boltzmann machine, have been applied for combinatorial optimization problems. However, there are some restrictions in the use of ANNs as optimizers. For example: (1) ANNs cannot optimize continuous variable problems; (2) discrete problems should be mapped into the neural networks’ architecture; and (3) most of the existing neural networks are applicable only for a class of smooth optimization problems and global convexity conditions on the objective functions and constraints are required. In this paper, we introduce a new procedure for stochastic optimization by a recurrent ANN. The concept of fractional calculus is adopted to propose a novel weight updating rule. The introduced method is called fractional-neuro-optimizer (FNO). This method starts with an initial solution and adjusts the network’s weights by a new heuristic and unsupervised rule to reach a good solution. The efficiency of FNO is compared to the genetic algorithm and particle swarm optimization techniques. Finally, the proposed FNO is used for determining the parameters of a proportional–integral–derivative controller for an automatic voltage regulator power system and is applied for designing the water distribution networks.     

基于模糊神经网络的冗余度变几何桁架机器人自适应控制

本文提出了一种基于模糊神经网络（FNN）的机器人位置自适应控制方法．利用模糊 神经网络模型来辨识冗余度变几何桁架机器人的逆动力学模型,用常规反馈控制器完成外部 干扰的补偿和闭环控制．并以四重四面体变几何桁架机器人为例进行仿真计算,表明该控制 方法具有良好的轨迹跟踪精度和抗干扰能力．     

Multi-zone parametric inverse analysis of super high arch dams using deep learning networks based on measured displacements

Parametric inverse analysis/identification provides significant information for structural damage detection and construction in dam engineering. The main challenge in inverse analysis is to enhance the computational accuracy and efficiency for complex structures, especially for super high arch dams with many zone parameters. This study developed a high-precision deep learning-based surrogate model for rapid inverse analysis of concrete arch dams. The relationship between mechanical parameters and multi-point displacement response is interpreted by convolutional neural networks (CNN)-based surrogate model. The proposed model is integrated with the Latin hypercube sampling and a meta-heuristic optimization algorithm for rapid inverse analysis strategy. The objective function is defined as the distance between the displacement predicted by the surrogate model and the measured displacement. The proposed approach is tested on an actual super high concrete arch dam. Results show that the proposed approach can achieve high accuracy and improve the computational efficiency by 95.83 % compared with the direct finite element method.     

Direct identification of structural parameters from dynamic responses with neural networks

A novel neural network-based strategy is proposed and developed for the direct identification of structural parameters (stiffness and damping coefficients) from the time-domain dynamic responses of an object structure without any eigenvalue analysis and extraction and optimization process that is required in many identification algorithms for inverse problems. Two back-propagation neural networks are constructed to facilitate the process of parameter identifications. The first one, called emulator neural network, is to model the behavior of a reference structure that has the same overall dimension and topology as the object structure to be identified. After having been properly trained with the dynamic responses of the reference structure under a given dynamic excitation, the emulator neural network can be used as a nonparametric model of the reference structure to forecast its dynamic response with sufficient accuracy. However, when the parameters of the reference structure are modified to form a so-called associated structure, the dynamic responses forecast by the network will differ from the simulated responses of the associated structure. Their difference can be assessed with a proposed root mean square (RMS) difference vector for both velocity and displacement responses. With the associated structural parameters and their corresponding RMS difference vectors, another network, called parametric evaluation neural network, can be trained. In this study, several 5-story frames are considered as example object structures with simulated displacement and velocity time histories that mimic the measured dynamic responses in practice. The performance of the proposed strategy has been demonstrated quite satisfactorily; the error for the estimation of each stiffness or damping coefficient is less than 10% even in the presence of 7% noise. Numerical simulations show that the accuracy of the identified parameters can be significantly improved by injecting noise in the training patterns for the parametric evaluation neural network. The proposed strategy is extremely efficient in computation and thus has potential of becoming a practical tool for near real time monitoring of civil infrastructures.     

Artificial neural networks for vibration based inverse parametric identifications: A review

Vibration behavior of any solid structure reveals certain dynamic characteristics and property parameters of that structure. Inverse problems dealing with vibration response utilize the response signals to find out input factors and/or certain structural properties. Due to certain drawbacks of traditional solutions to inverse problems, ANNs have gained a major popularity in this field. This paper reviews some earlier researches where ANNs were applied to solve different vibration-based inverse parametric identification problems. The adoption of different ANN algorithms, input-output schemes and required signal processing were denoted in considerable detail. In addition, a number of issues have been reported, including the factors that affect ANNs’ prediction, as well as the advantage and disadvantage of ANN approaches with respect to general inverse methods Based on the critical analysis, suggestions to potential researchers have also been provided for future scopes.     

Background Information of Deep Learning for Structural Engineering

Since the first journal article on structural engineering applications of neural networks (NN) was published, there have been a large number of articles about structural analysis and design problems using machine learning techniques. However, due to a fundamental limitation of traditional methods, attempts to apply artificial NN concept to structural analysis problems have been reduced significantly over the last decade. Recent advances in deep learning techniques can provide a more suitable solution to those problems. In this study, versatile background information, such as alleviating overfitting methods with hyper-parameters, is presented. A well-known ten bar truss example is presented to show condition for neural networks, and role of hyper-parameters in the structures.     

频响函数作为神经网络输入的BP算法实现

研究将实测结构频率响应函数作为反向传递人工神经网络的输入数据,用来进行结构健康检测.一般情况下,把频率响应函数应用到人工神经网络的困难在于需要压缩频率响应函数的庞大数据,因为直接使用全部的频率响应函数数据使得神经网络具有大量的输入节点,从而导致网络训练收敛和计算效率方面的困难.仅仅使用部分频率响应数据,或不合适的频率窗数据点选择会引起重要信息的损失.为解决上述困难,用FORTRAN语言编写了一个简化的BP神经网络程序,把某结构的频率响应函数作为网络的输入参量.每个实测频率响应函数具有8192个数据点,神经网络采用8192-8-4结构,网络训练获得了较快的收敛速度.经过训练的网络成功识别了某结构的四种不同状态,识别误差小于10%.     

Neurocomputing strategies in structural design — decomposition based optimization

The present paper introduces a scheme utilizing neurocomputing strategies for a decomposition approach to large scale optimization problems. In this scheme the modelling capabilities of a backpropagation neural network are employed to detect weak couplings in a system and to effectively decompose it into smaller, more tractable subsystems. When such partitioning of a design space is possible (decomposable systems), independent optimization in each subsystem is performed with a penalty term added to an objective function to eliminate constraint violations in all other subsystems. Dependencies among subsystems are represented in terms of global design variables, and since only partial information is needed, a neural network is used to map relations between global variables and all system constraints. A featuresensitive network (a variant of ahierarchical vector quantization technique, referred to as the HVQ network) is used for this purpose as it offers easy training, approximations of an arbitrary accuracy, and processing of incomplete input vectors. The approach is illustrated with applications to minimum weight sizing of truss structures with multiple design constraints.