An integrated approach for structural damage identification using wavelet neuro-fuzzy model

Structural damage can be identified by processing structural vibration response signals and excitation data, and thus the suitability of signal processing methods is essential to structural damage identification. To explore an intelligent signal processing method for structural damage identification, the paper integrated wavelet real-time filtering algorithm, Adaptive Neruo-Fuzzy Inference System (ANFIS) and interval modeling technique to process structural response signals and excitation data. With Wavelet Transform (WT) algorithm filtering random noise, ANFIS was found to model the structural behavior properly and interval modeling technique to quantify damage index accurately. The rapid identifications of several unknown damages and small damages indicate the efficiency of this integrated method. The comparison of these results and some other signal processing methods shows that, the proposed method can be used to identify both the time and the location when the structural damage occurs unexpectedly.     

Ultimate strength analysis of structural components using the continuum damage mechanics approach

In this study, an analytical model is developed to include the effect of material fracture on the overall behavior of a structure. The continuum damage mechanics approach is used to describe the continuous deterioration of the material stiffness that leads to total failure. The damage evolution is assumed to be anisotropic and the effect of the triaxiality of stress on the fracture criterion is included in the analysis. The above modifications were applied to an existing finite element program which is capable of including both geometric and material nonlinearities in the analysis. The program was used to analyze a welded tubular connection (T-joint). This joint was chosen because fracture was shown experimentally to control its ultimate behavior. Solid finite elements were used to model the chord and the branch in the intersection region, wedge elements were used for the weld profile, and the rest of the geometry was modeled using shell elements. The joint was subjected to an increasing amount of displacement at the far end of the branch. The loading was continued until total failure. During this process, the evolution of the damage and its propagation were monitored. Divergence of the solution was obtained at a failure load of 42.4 kips compared to 44.0 kips obtained experimentally. In addition, the failure mode of the joint predicted by the analysis was identical to that observed experimentally.     

A piecewise constant level set method for damage identification of continuum structures based on natural frequencies

Structural and Multidisciplinary Optimization - The damage identification of continuum structures from the observation of its natural frequencies is investigated. The piecewise constant level set...     

A randomised approach for NARX model identification based on a multivariate Bernoulli distribution

The identification of polynomial NARX models is typically performed by incremental model building techniques. These methods assess the importance of each regressor based on the evaluation of partial individual models, which may ultimately lead to erroneous model selections. A more robust assessment of the significance of a specific model term can be obtained by considering ensembles of models, as done by the RaMSS algorithm. In that context, the identification task is formulated in a probabilistic fashion and a Bernoulli distribution is employed to represent the probability that a regressor belongs to the target model. Then, samples of the model distribution are collected to gather reliable information to update it, until convergence to a specific model. The basic RaMSS algorithm employs multiple independent univariate Bernoulli distributions associated to the different candidate model terms, thus overlooking the correlations between different terms, which are typically important in the selection process. Here, a multivariate Bernoulli distribution is employed, in which the sampling of a given term is conditioned by the sampling of the others. The added complexity inherent in considering the regressor correlation properties is more than compensated by the achievable improvements in terms of accuracy of the model selection process.     

A study on X-FEM in continuum structural optimization using a level set model

In this paper, we implement the extended finite element method (X-FEM) combined with the level set method to solve structural shape and topology optimization problems. Numerical comparisons with the conventional finite element method in a fixed grid show that the X-FEM leads to more accurate results without increasing the mesh density and the degrees of freedom. Furthermore, the mesh in X-FEM is independent of the physical boundary of the design, so there is no need for remeshing during the optimization process. Numerical examples of mean compliance minimization in 2D are studied in regard to efficiency, convergence and accuracy. The results suggest that combining the X-FEM for structural analysis with the level set based boundary representation is a promising approach for continuum structural optimization.     

基于计算机视觉的建筑施工期临时结构损伤识别方法

建筑施工临时结构是施工现场的事故主要风险源。以往的基于振动的临时结构监测方法依赖于在预先分析确定的监测关键部位放置的加速度传感器。但由于临时结构存在构件搭设不规范、施工现场不确定性等因素,通过有限元分析等手段得到的监测关键部位可能与实际情况相差较大,存在不确定性。为此提出一种基于欧拉运动放大算法的临时结构损伤识别方法,充分利用计算机视觉技术的全域覆盖及监测高效的优点。采用数码摄像机采集临时结构的数字图像序列,经过基于相位的欧拉运动放大算法处理,获取运动放大后的数字图像序列;运用Canny边缘识别算法获取边缘图像序列并消除运动放大造成的噪声,通过基于形心的运动跟踪算法获取临时结构的位移时程数据,并利用快速傅里叶变换进行频谱分析;与预先建立的损伤动力指纹库进行对比判断临时结构的损伤状态。以存在10种损伤状态的门式脚手架为测试对象,证明该方法的可行性与适用性。与加速度传感器测量进行对比,该方法平均误差为0.95%,满足临时结构损伤状态识别的精度要求。     

New results on PWARX model identification based on clustering approach

This paper deals with the problem of piecewise auto regressive systems with exogenous input(PWARX) model identification based on clustering solution. This problem involves both the estimation of the parameters of the affine sub-models and the hyper planes defining the partitions of the state-input regression. The existing identification methods present three main drawbacks which limit its effectiveness. First, most of them may converge to local minima in the case of poor initializations because they are based on the optimization using nonlinear criteria. Second, they use simple and ineffective techniques to remove outliers. Third, most of them assume that the number of sub-models is known a priori. To overcome these drawbacks, we suggest the use of the density-based spatial clustering of applications with noise(DBSCAN) algorithm. The results presented in this paper illustrate the performance of our methods in comparison with the existing approach. An application of the developed approach to an olive oil esterification reactor is also proposed in order to validate the simulation results.     

Stress-constrained continuum topology optimization: a new approach based on elasto-plasticity

A new approach for generating stress-constrained topological designs in continua is presented. The main novelty is in the use of elasto-plastic modeling and in optimizing the design such that it will exhibit a linear-elastic response. This is achieved by imposing a single global constraint on the total sum of equivalent plastic strains, providing accurate control over all local stress violations. The single constraint essentially replaces a large number of local stress constraints or an approximate aggregation of them – two common approaches in the literature. A classical rate-independent plasticity model is utilized, for which analytical adjoint sensitivity analysis is derived and verified. Several examples demonstrate the capability of the computational procedure to generate designs that challenge results from the literature, in terms of the obtained stiffness-strength-weight trade-offs. A full elasto-plastic analysis of the optimized designs shows that prior to the initial yielding, these designs can sustain significantly higher loads than minimum compliance topological layouts, with only a minor compromise on stiffness.     

A neural network approach for defect identification and classification on leather fabric

In this paper, an automated vision system is presented to detect and classify surface defects on leather fabric. Visual defects in a gray-level image are located through thresholding and morphological processing, and their geometric information is immediately reported. Three input feature sets are proposed and tested to find the best set to characterize five types of defects: lines, holes, stains, wears, and knots. Two multilayered perceptron models with one and two hidden layers are tested for the classification of defects. If multiple line defects are identified on a given image as a result of classification, a line combination test is conducted to check if they are parts of larger line defects. Experimental results on 140 defect samples show that two-layered perceptrons are better than three-layered perceptrons for this problem. The classification results of this neural network approach are compared with those of a decision tree approach. The comparison shows that the neural network classifier provides better classification accuracy despite longer training times.     

An efficient approach for damage identification based on improved machine learning using PSO-SVM

Engineering with Computers - Structural health monitoring (SHM) and Non-destructive Damage Identification (NDI) using responses of structures under dynamic excitation have an imperative role in the...     

A feature selection approach based on a similarity measure for software defect prediction

Software defect prediction is aimed to find potential defects based on historical data and software features. Software features can reflect the characteristics of software modules. However, some of these features may be more relevant to the class (defective or non-defective), but others may be redundant or irrelevant. To fully measure the correlation between different features and the class, we present a feature selection approach based on a similarity measure (SM) for software defect prediction. First, the feature weights are updated according to the similarity of samples in different classes. Second, a feature ranking list is generated by sorting the feature weights in descending order, and all feature subsets are selected from the feature ranking list in sequence. Finally, all feature subsets are evaluated on a k-nearest neighbor (KNN) model and measured by an area under curve (AUC) metric for classification performance. The experiments are conducted on 11 National Aeronautics and Space Administration (NASA) datasets, and the results show that our approach performs better than or is comparable to the compared feature selection approaches in terms of classification performance.     

A new multi-solution approach suitable for structural identification problems

In this paper, a multi-solution strategy aimed at greatly decreasing the cost of structural identification problems is proposed. It is based on the large time increment method and on the possibility offered by this approach to obtain a family of admissible displacement and stress fields in order to solve a nonlinear time-dependent problem. This technique relies on a time and space decomposition of the solution. It has been extended to solve at a low cost, the inverse problem of identification, for various techniques. In particular, it has been adapted to the minimization of the gradient of the gap between test and simulation. In this first attempt, it is applied to the simple case of bending of an elastic–viscoplastic beam. The issue of the numerical efficiency of this type of strategy is studied.     

一种基于主题模型的软件缺陷预测技术研究

软件缺陷预测通常针对代码表面特征训练预测模型并对新样本进行预测,忽视了代码背后隐藏的不同技术方面和主题,从而导致预测不准确。针对这种问题,提出了一种基于主题模型的软件缺陷预测方法。将软件代码库视为不同技术方面和主题的集合,不同的主题或技术方面有不同的缺陷倾向。采用LDA主题模型对不同主题及其缺陷倾向进行建模,根据建模结果计算主题度量,并将传统度量方式和主题度量结合进行模型训练和预测。实验结果显示,该方法相对传统的软件缺陷预测技术有高的准确性,并且可以在软件演化中保证模型相对稳定,可以适用于各种缺陷预测任务。     

A time-domain approach to z-domain model identification

A time-domain parameter-optimization approach to discrete linear-model identification is presented, z-domain transfer functions are sought with outputs closely matching corresponding outputs of observed systems. Optimal transfer-function coefficients are determined through the minimization of mean-square output error. The general technique developed is applied to problems in model reduction and identification in the presence of measurement noise.     

基于NARMAX模型和NOFRF结构损伤检测的实验研究

非线性输出频率响应函数是由Volterra级数发展而来的频域概念,可方便在频域对非线性系统进行分析,它是频率的一维函数.本文主要介绍了利用NARMAX模型以及NOFRF对结构进行损伤检测的方法,并利用实验研究证实了该损伤检测方法的可行性.另外,由于系统非线性特性可用来做结构损伤检测,且具有对系统状态比较敏感的优点,而基于NOFRF的损伤检测方法是利用非线性方法来分析系统的状态,该方法提取出的特征属于非线性特征,所以该损伤检测方法可以用来做结构损伤检测,且具有对系统状态比较敏感的优点.     

Genetic algorithm based defect identification system

In this paper we develop a genetic algorithm based defect identification system for machined-parts inspection purposes. The system is based on genetic algorithm and knowledge system techniques.     

基于改进SSD模型的工件表面缺陷识别算法

工件表面缺陷是影响机械设备性能的重要因素,快速高效的检测方法是目前研究的重点。为了解决工件表面缺陷检测问题,提出一种基于改进SSD模型的检测算法。该算法用本文提出的DH-MobileNet网络代替SSD结构中的VGG16网络,从而简化检测模型,减少了运算量。同时采用反向残差结构进行位置预测,并用空洞卷积代替下采样操作以避免信息损失。利用扫描电子显微镜得到工件表面图像,建立工件表面缺陷数据集并进行扩充,最后针对碎屑、剥落和梨沟这3类高频缺陷进行训练和测试,并与YOLO、Faster R-CNN和原始SSD模型进行效果比较。检测结果表明该算法能够更准确、快速地检测工件表面缺陷,为实际工业场景中的缺陷检测提供了新的思路。     

Simultaneous identification of moving masses and structural damage

A method for simultaneous identification of moving masses and damages of the supporting structure from measured responses is presented. The interaction forces between the masses and the structure are used as excitation. Masses and damage extents are used as the optimization variables; compared to the approaches based on identification of the interaction forces, it allows ill-conditioning to be avoided and decreases the number of required sensors. The virtual distortion method is used; the damaged structure is modeled by the intact structure subjected to response-coupled virtual distortions and moving forces. These are related to the optimization variables via a linear system, which allows the optimization variables of both kinds to be treated in a unified way. A moving dynamic influence matrix is introduced to reduce the numerical costs. The adjoint variable method is used for fast sensitivity analysis. A numerical experiment of a three-span beam with 10% rms measurement error and three types of model errors is presented.     

A blind approach to the Hammerstein-Wiener model identification

In this paper, we propose a blind approach to the sampled Hammerstein-Wiener model identification. By using the blind approach, it is shown that all internal variables can be recovered solely based on the output measurements. Then, identification of linear and nonlinear parts can be carried out. No a priori structural knowledge about the input nonlinearity is assumed and no white noise assumption is imposed on the input.