Investigation of defect characteristics and heat transfer in step heating thermography of metal plates repaired with composite patches

Nowadays, composite patches are widely used in different industries to repair damaged metal structures. Inspection of such repaired structures is always considered as a challenging task. Different thermography methods such as step heating are commonly used to inspect repaired structures. Some parameters such as defect features or heating procedure play major roles in defect detection. In this work, in order to investigate such effects, step heating thermography of an aluminum plate repaired with a composite patch is modeled and tested. The main goal of this study is to evaluate the effects of defect type (delamination and disbond), size and depth on the detection ability of the test. Moreover, regarding the heat transfer process obtained from the simulation, the appropriate heating procedure for inspecting the repaired metal structures is determined. To validate the simulation outputs, experimental results corresponding to the temperature variations are compared with those predicted from the simulation.     

Health monitoring of a composite wingbox structure

This work was devoted to the development of a health monitoring system assigned to aerospace applications. Those applications concerned the detection of damaging impacts and debonding between stiffeners and composite skins, since they are the major causes of in-service damage of aircraft structures. The chosen health monitoring system was first based on the excitation and reception of Lamb waves along the structure by using thin piezoelectric transducers (active mode) and secondly on a continuous monitoring taking the same transducers used as acoustic emission sensors (passive mode). The composite specimen used was consistent with aircraft wingbox in terms of structure and loading. Several impacts with increasing energy increments were applied on the composite specimen. In passive mode, the study showed the ability of using the acoustic signature of an impact to detect possible damage. Moreover, the damage emergence in the case of damaging impact was confirmed in active mode. Further measurements during fatigue testing were performed. The aim was to demonstrate the ability of the system to monitor disbond growth between the stiffener and the composite skin. The sensitivity of the health monitoring system to the disbond growth was further demonstrated.     

The sensitivity of surface guided modes to the bond quality between a concrete block and a composite plate

Ultrasonic modes guided along the surface of a concrete block covered by a composite plate are investigated. First of all, the propagation phenomenon is studied by predicting the dispersion curves for various modes supposed to exist in this structure. Then, focus is made on a specific mode expected to be sensitive to the quality of the bond between the concrete and the composite, according to its through-thickness displacement field. Its dispersion curve is re-plotted for various states of this bond, which is modelled as a thin layer with varying properties. Limit cases corresponding to total disbond and perfect bond are considered. Experiments are then made on two composite-concrete assemblies for these two extreme cases using a contact PZT transmitter and an air-coupled receiver system. The phase velocities are in good agreement with the predicted dispersion curves, thus showing the potential of the ultrasonic guided mode to control the quality of bonds in such structures.     

基于Gammatone滤波器组时频谱和卷积神经网络的海底底质分类

为了有效利用海底底质信号完成海底底质的分类识别,该文提出一种将深度学习方法和底质信号相结合实现底质分类识别的方法.首先利用Gammatone滤波器组计算底质侧扫图像信号的时频谱,然后通过卷积神经网络对得到的时频谱进行分类识别完成底质分类.利用加利福尼亚州Scott Creek近海采集的侧扫声呐图像数据进行数据分析,结果...     

联合特征选择与改进集成学习模型的海底底质分类

针对侧扫声呐获取类型复杂的海底底质数据分类问题,提出联合特征选择与改进Stacking模型的数据自驱动分类方法。该方法首先在海底散射数据多域态特征的基础上采用ReliefF算法提取有效的低维度特征,然后将人工鱼群算法与Stacking模型结合形成改进集成学习分类器,完成海底底质分类。海上数据处理结果表明该方法可对多种海底底质类型进行分类,分类准确率、Kappa系数和F1-score分别达到85.55%,0.857和0.887,证明了该方法的有效性。      

Hydrogen bond dynamics probed with ultrafast infrared heterodyne-detected multidimensional vibrational stimulated echoes

Hydrogen bond dynamics are explicated with exceptional detail using multidimensional infrared vibrational echo correlation spectroscopy with full phase information. Probing the hydroxyl stretch of methanol-OD oligomers in CCl4, the dynamics of the evolving hydrogen bonded network are measured with ultrashort (<50 fs) pulses. The data along with detailed model calculations demonstrate that vibrational relaxation leads to selective hydrogen bond breaking on the red side of the spectrum (strongest hydrogen bonds) and the production of singly hydrogen bonded photoproducts.     

The transmission of Lamb waves across adhesively bonded lap joints

The transmission of Lamb waves across adhesively bonded lap joints is investigated using finite element analysis. The studies consider three modes for excitation and reception, s0, a0, and a1, applied to lap joints consisting of parallel aluminum sheets bonded with an epoxy adhesive. Transmission coefficient results for a two-dimensional range of bond thicknesses and bond overlap lengths are presented for all three modes. The transmission coefficients are calculated from the spectra of the received and transmitted signals using an approach which is insensitive to the presence of multimode signals and reverberated signals, and which approximates to a power transmission coefficient. Detailed analysis is then performed for one of the modes in order to investigate the nature of the mode conversion in the overlap region of the joint. It is found that the relative amplitudes of the different modes which propagate in the overlap region can be estimated reliably and simply from the properties of the incident wave mode. As well as demonstrating the physics of the mode conversion behavior, the study provides a basis for the selection of modes for nondestructive evaluation (NDE) of the bond region and for measuring the bond dimensions.     

Comparison of three thermographic post processing methods for the assessment of a repaired aluminum plate with composite patch

Composite patches are widely used to repair damaged metal structures, especially in aerospace industry. Perfect patch and bonding are necessary to achieve an effective repair. Various thermographic methods such as step heating thermography are commonly applied to inspect repaired structures. Since accurate determination of defect features are admirable, some techniques are used to process the thermal films. In this study, three common post processing techniques of thermography (namely, principle component analysis (PCA), pulse phase thermography (PPT) and thermal signal reconstruction (TSR)) have been utilized to inspect an aluminum plate repaired with carbon/epoxy patches. Several delaminations with various sizes and locations along with some disbond defects were artificially embedded in five samples of composite patches to experimentally investigate the performance of the three techniques for post-processing of the step heating thermography data. Furthermore, the outputs of the mentioned processing techniques were quantitatively compared to find the most effective one. Based on the comparison results, it was demonstrated that, TSR outputs leads to the more accurate defect sizing.     

联合稀疏表示在水下目标分类中的应用

水下目标分类识别的性能受所选特征的限制,多特征往往可以获得更加稳定的结果,针对这一问题,提出了一种基于联合稀疏表示模型的水下目标分类识别方法。首先对水下目标回波信号提取3种具有信息互补性与关联性的特征:中心矩特征、小波包能量谱特征、梅尔频率倒谱系数特征,然后应用加速近端梯度法对联合稀疏表示模型进行优化,求解得到最优联合稀疏系数,最后根据最小误差准则确定目标类别。在消声水池开展模拟实验,对6类目标进行分类识别,结果表明:与传统算法相比,提出的算法具有更高识别准确率,并且其执行效率较传统算法有很大提升。      

Propagation of ultrasonic guided waves in lap-shear adhesive joints: case of incident a0 Lamb wave

This paper deals with the propagation of ultrasonic guided waves in adhesively bonded lap-shear joints. The topic is relevant to bond inspection by ultrasonic testing. Specifically, the propagation of the lowest-order, antisymmetric a0 mode through the joint is examined. An important aspect is the mode conversion at the boundaries between the single-plate adherents and the multilayer overlap. The a0 strength of transmission is studied for three different bond states in aluminum joints, namely a fully cured adhesive bond, a poorly cured adhesive bond, and a slip bond. Theoretical predictions indicate that the dispersive behavior of the guided waves in the multilayer overlap is highly dependent on bond state. Experimental tests are conducted in lap-shear joints by a hybrid, broadband laser/air-coupled ultrasonic setup in a through-transmission configuration. The Gabor wavelet transform is employed to extract energy transmission coefficients in the 100 kHz 1.4 MHz range for the three different bond states examined. The cross-sectional mode shapes of the guided waves are shown to have a substantial role in the energy transfer through the joint.     

Photo-directed organization of silver nanoparticles in mesostructured silica and titania films

Carbon nanotubes (CNTs) have the potential to enhance the strength, toughness, and multifunctional ability of composite materials. However, suitable dispersion and interfacial bonding remain as key challenges. Composites that are formed by reactions with water, like Portland cement concrete and mortar, pose a special challenge for dispersing the inherently hydrophobic nanotubes. The hydration of Portland cement also offers a specific chemical framework for interfacial bonding. In this study, nanoscale silica functional groups are covalently bonded to CNTs to improve their dispersion in water while providing interfacial bond sites for the proposed matrix material. The bond signatures of treated nanotubes are characterized using Fourier transform infrared spectroscopy. In situ dispersion is characterized using cryogenic transmission electron microscopy and point of zero charge (PZC) measurements. At the nanoscale, interparticle spacing was greatly increased. A slight increase in the PZC after treatment indicates the importance of steric effects in the dispersion mechanism. Overall, results indicate successful functionalization and dramatically improved dispersion stability in water.

     

Effect of bonding parameters on microstructure development during TGTLP bonding of Al7075 alloy

The effect of temperature, pressure and bonding time on microstructure of temperature gradient transient liquid phase (TGTLP) diffusion bonded Al7075 alloy using liquid gallium interlayer was investigated. The selected bonding method relies on using the minimum amount of liquid gallium on faying surfaces, using a very fast heating rate to reach the joining temperature and imposing a temperature gradient across the bond region. The microstructure of the diffusion bonded joints was evaluated by light optical microscopy, scanning electron microscopy and energy dispersive spectroscopy (EDS). Results show that by increasing the temperature, pressure and time of joining, a more uniform microstructure can be obtained at the joint area. The best joint microstructure was achieved at a temperature of 460?°C, pressure of 10?MPa and time of 10?min. EDS spot analysis indicates that brittle silicon-rich precipitates form at the joint line during TGTLP bonding.     

Reference-free crack detection using transfer impedances

A new concept of reference-free damage detection methodology is developed using transfer impedances to detect crack damage in a plate-like structure without using previously collected baseline data. Conventional impedance-based damage detection techniques have been shown to be vulnerable to other types of changes such as temperature variation that may not be relevant to defects of interest. One of the potential disadvantages of the conventional techniques is frequent false-alarms due to these undesirable variations that may occur in field applications. In order to reduce these false-alarms, this paper proposes a new methodology that utilizes transfer impedances obtained between two pairs of collocated PZT patches instead of the electromechanical impedance obtained at one PZT patch. The proposed technique utilizes Lamb mode conversion effects caused by the presence of crack damage in plate structures. Furthermore, an instantaneous damage classification is carried out by comparing mode conversion energy among several combinations of measured signals without any user-specified threshold or relying on the baseline data. The feasibility of the proposed reference-free methodology using transfer impedances is investigated via a series of experiments conducted on an aluminum plate.     

基于空间特征的光谱分类算法研究

随着成像光谱仪器的广泛应用,利用数据立方体进行物质分类与识别成为一项重要的研究内容,分类算法对最终的目标识别准确度与精度具有很大的决定作用。目前常见的分类算法主要利用了光谱维信息,从光谱匹配的角度进行物质分类。由于仪器探测的物质反射光谱不仅反映了物质种类,还与物质表面的几何结构,表面粗糙度等有关,因此仅仅利用物质的反射光谱进行物质分类识别具有一定的误差。该文在利用可见光反射光谱进行分类的基础上,结合图像空间特征,对分类过程进行控制,达到提高分类准确度的目的。利用该分类算法进行真假叶片识别,结果表明其具有较好的空间连续性,很大程度上克服了"麻点"效应,验证了算法的有效性。     

Kernel representation-based nearest neighbor classifier

An improvement to the nearest neighbor classifier (INNC) has shown its excellent classification performance on some classification tasks. However, it is not very clearly known why INNC is able to obtain good performance and what the underlying classification mechanism is. Moreover, INNC cannot classify low-dimensional data well and some high-dimensional data in which sample vectors belonging to different class distribution but have the same vector direction. In order to solve these problems, this paper proposes a novel classification method, named kernel representation-based nearest neighbor classifier (KRNNC), which can not only remedy the drawback of INNC on low-dimensional data, but also obtain competitive classification results on high-dimensional data. We reveal the underlying classification mechanism of KRNNC in details, which can also be regarded as a theoretical supplement of INNC. We first implicitly map all samples into a kernel feature space by using a nonlinear mapping associated with a kernel function. Then, we represent a test sample as a linear combination of all training samples and use the representation ability to perform classification. From the way of classifying test samples, KRNNC can be regarded as the nonlinear extension of INNC. Extensive experimental studies on benchmark datasets and face image databases show the effectiveness of KRNNC.     

Improving the Accuracy of Nearest-Neighbor Classification Using Principled Construction and Stochastic Sampling of Training-Set Centroids

A conceptually simple way to classify images is to directly compare test-set data and training-set data. The accuracy of this approach is limited by the method of comparison used, and by the extent to which the training-set data cover configuration space. Here we show that this coverage can be substantially increased using coarse-graining (replacing groups of images by their centroids) and stochastic sampling (using distinct sets of centroids in combination). We use the MNIST and Fashion-MNIST data sets to show that a principled coarse-graining algorithm can convert training images into fewer image centroids without loss of accuracy of classification of test-set images by nearest-neighbor classification. Distinct batches of centroids can be used in combination as a means of stochastically sampling configuration space, and can classify test-set data more accurately than can the unaltered training set. On the MNIST and Fashion-MNIST data sets this approach converts nearest-neighbor classification from a mid-ranking- to an upper-ranking member of the set of classical machine-learning techniques.     

Experimental Study and Comparison of Imbalance Ensemble Classifiers with Dynamic Selection Strategy

Imbalance ensemble classification is one of the most essential and practical strategies for improving decision performance in data analysis. There is a growing body of literature about ensemble techniques for imbalance learning in recent years, the various extensions of imbalanced classification methods were established from different points of view. The present study is initiated in an attempt to review the state-of-the-art ensemble classification algorithms for dealing with imbalanced datasets, offering a comprehensive analysis for incorporating the dynamic selection of base classifiers in classification. By conducting 14 existing ensemble algorithms incorporating a dynamic selection on 56 datasets, the experimental results reveal that the classical algorithm with a dynamic selection strategy deliver a practical way to improve the classification performance for both a binary class and multi-class imbalanced datasets. In addition, by combining patch learning with a dynamic selection ensemble classification, a patch-ensemble classification method is designed, which utilizes the misclassified samples to train patch classifiers for increasing the diversity of base classifiers. The experiments’ results indicate that the designed method has a certain potential for the performance of multi-class imbalanced classification.     

基于小波降噪与支持向量机的恒星光谱识别研究

提出了一种对恒星光谱识别的新方法。 根据恒星光谱数据的特性,我们以支持向量机为核心技术构建光谱识别器。 由于恒星光谱数据通常含有较高的噪声,如果直接进行分类,识别率往往较低。 因此作者首先采用小波分析的方法对原始光谱数据进行降噪预处理,提取光谱的特征,然后馈送到支持向量机完成对光谱数据的最终识别。 利用实际光谱数据(Jacoby, 1984)对所提出的技术进行检测,实验结果表明使用这种小波分析结合支持向量机的技术的识别效果要优于使用支持向量机结合主分量分析降维技术的识别方法。 另外,作者还比较了支持向量机与传统甄别分析的分类性能,对实际及合成光谱进行实验的结果显示了支持向量机的识别正确率不但优于常见的5种甄别分析方法的识别率,而且有较强的泛化能力。     

水稻分蘖期无人机高光谱影像混合像元特征分析与分解

开展水稻无人机高光谱解混,获取水稻植株的高光谱反射率信息,对于提高水稻理化参量的反演模型精度具有重要意义.目前大多基于高光谱遥感影像自身数据进行解混,运用算法模型进行高光谱数据解混,将高光谱图像和可见光图像进行优势互补,提出一种基于无人机高清影像与高光谱遥感影像融合的稻田无人机高光谱解混方法,解决单一数据局限性问题,增...     

基于信息增益率的WNB水下机器人故障分类

提高故障诊断能力对于确保水下机器人系统的稳定运行具有重要意义,故障分类是目前水下机器人故障诊断所面临的一个重要问题。针对水下机器人推进器系统数据特征,提出一种基于信息增益率的加权朴素贝叶斯故障分类算法。首先,计算故障训练样本的先验概率,将各属性的信息增益率作为权值;其次,构建基于增益率加权的朴素贝叶斯分类模型;然后,对检测的故障数据利用分类模型获取具有最大后验概率的故障模式,实现故障分类。与朴素贝叶斯算法和决策树算法相比,仿真实验结果表明基于信息增益率加权的朴素贝叶斯算法的分类成功率更高,能够有效地实现水下机器人的故障分类。