Ndt & E International最新文献

The thickness and interface roughness of coatings both affect the interface bonded quality. Existed ultrasonic testing methods based on traditional phase screen approximation or spring model assumption are difficult to simultaneously identify the interface roughness and stiffness of coating. This paper, a new method for integrated identifying coating thickness, interface roughness, and interface stiffness using developed ultrasonic reflection phase derivative spectrum (URPDS) is proposed. A phase-screen-approximated spring-model (PSASM) for ultrasound vertically propagating into rough and weak bonded interface is constructed. On basis of PSASM, a URPDS of coating/substrate structure is developed for identifying the interface stiffness and other parameters of coated parts. Cross-correlation analysis is used to eliminate the phase deviation of URPDS introduced by reference signal and initial phase of tested signal. Sensitivity analysis is used to determine the high-sensitivity regions of URPDS to interface roughness and interface stiffness. Genetic algorithm optimization is used to achieve integrated identification of coating thickness, interface roughness, and interface stiffness. The rationality of PSASM is verified through numerical simulation using a series of coating/substrate models with rough and weak bonded interface, and the relationship between the high-sensitivity regions and the high-precision measurement ranges of interface roughness Rq and interface stiffness K_n is clarified. Ultrasonic experiments are implemented on Nickel-coating samples and coated parts using plane wave probe. The coating thickness, interface roughness, and interface stiffness could be identified accurately, which shows that the proposed URPDS method can identify the interface stiffness of rough contacted dissimilar media or coated parts with rough interface.

We proposed a deep-learning attention-based methodology to predict acoustic sources obtained from pendulum impact experiments using the Cluster-Self Adaptive Network (CSAN) and showed that the experimental data required for training can be reduced by 50% without losing significant localization accuracy. Acoustic signals due to pendulum impacts on a homogeneous steel plate were recorded by an asymmetric microphone array. Important wavelet features were extracted by transforming the acoustic signals using continuous wavelet functions and reduced the data dimensionality by principal component analysis. Two data sampling strategies (random and Latin hypercube) were investigated to study the effect of the density of training domains on the model performance. The attention-based modulation strategy was employed on microphone positions for data augmentation and prediction of acoustic sources. A comprehensive analysis of the CSAN-based localization results including error estimation was performed. The outcome was contrasted against delay-and-sum beamforming localization results.

Online monitoring of weld crack leakage in pressure pipelines of nuclear power ship based on acoustic emission (AE) technology is of great significance for maintaining the safe and stable operation of the system. However, most of the current leakage studies are conducted through artificially designed pipeline hole types, which deviate from the actual crack morphology and are weakly online, with low identification accuracy and slow monitoring speed. Therefore, a convolutional network of FGI and multi-scale channel information cross fusion based on AE technology is proposed in this paper. First, the FBank feature of the AE signal of pipeline weld leakage are extracted. On this basis, the Gini Index (GI) preference feature is used to filter the useless information in the FBank feature. Then, a multi-scale channel information cross fusion module is designed to improve the feature learning ability of the network through the interaction and fusion of different channel information. Finally, the superiority of the proposed FGI feature extraction method and the effectiveness of the proposed multi-scale channel information cross fusion CondenseNet (MCCF-CondenseNet) convolutional neural network are verified by the pipeline leakage AE monitoring experiments under three crack morphologies. The results show that the identification accuracy of the proposed method is as high as 96.42 %, and the identification speed is significantly faster than other state-of-the-art approaches under the premise of ensuring the identification accuracy. This work provides a new method for the online leakage monitoring of nuclear power pressure pipelines, and has important supporting significance for the online leakage monitoring of other large and complex equipment.

Currently, in terms of resolution and excitation efficiency for pipeline inspection, the high-frequency Rayleigh-like wave excited by an EMAT with a traditional Rayleigh wave EMAT structure is not optimal when using the same magnet volume. This paper introduces an EMAT performance evaluation method focused on 'bandwidth' in the high-frequency-thickness region of circumferential guided waves. A wavenumber spectrum analysis method utilizing combined equivalent surface stresses is proposed to quantify this optimize design. Comparative studies, including theoretical analysis and experimental validation, demonstrate that incorporating bandwidth significantly improves the design of Rayleigh-like waves at high frequencies. The proposed EMAT achieves a performance improvement of 2.4 times for inside pipe excitation and 2.6 times for outside pipe excitation over the conventional structure. The occurrence of multiple wave packets outside the optimal excitation frequency range is acknowledged. Therefore, this method offers a new approach for optimizing EMATs for Rayleigh-like waves.

The piezoelectric impedance-based technique is always regarded as one of the most promising structural health monitoring and nondestructive evaluation methods. In recent years, impedance measurement chip AD5933 with the characteristics of high integration and cost-effectiveness makes it possible to address the huge and high-cost problems of the commercialized impedance measurement instrument during the process of structural health monitoring and defect identification. However, it still faces lots of challenges for the chips to be utilized in practical applications due to several limitations, such as short distance for data transmission, single measurement channel and artificial attendant requirement. In this paper, a wireless multichannel miniature impedance measurement system composed by the front-end measurement device and the remote measurement and control platform on the server, is firstly developed and presented with the functions of wireless data transmission, multi-channel acquisition and remote data post-processing. Subsequently, the design concept and composition of the system were introduced in detail. Then, a series of piezoelectric transducers related tests were conducted to validate its impedance measurement performance, especially when comparing with the ones measured by commercialized instruments. In addition, to verify its effectiveness and feasibility of the developed system for the structural damage detection, a bolt loosening detection experiment on the flange connection of a pipeline specimen was investigated for its damage localization and severity quantification. Finally, all the results demonstrated that the developed system provides a great possibility to be used as a convenient and portable impedance measurement tool for the civil structural health monitoring and damage identification in practical applications.

Continuously welded rails (CWR) are prone to the development of high thermal-induced load along the axial direction. Excessive levels of load lead to risk of rail buckling and potential for derailment. Knowledge of the in situ rail axial load in CWRs is therefore important to ensure safe rail management. Field-deployable, nondestructive evaluation techniques for measuring the rail load, or a widely adopted alternative called rail neutral temperature (RNT), are desired. This study uses a data-driven approach to investigate if rail dynamic response data, collected in a non-destructive fashion, can be used to predict RNT. The study is based on a data set comprising rail equivalent strain, temperature and vibration resonance frequencies that was collected from a revenue-service rail over a period of nearly two years. All excited vibration resonance peaks are identified from other peaks caused by noise using spectral amplitude variance. Among these resonance peaks, potentially useful resonances are identified with respect to stacked spectra collected across a testing day using an assumed temperature-frequency relation. A subset of the identified useful resonances is then identified based on their consistent appearance across both testing locations and all testing days, strong correlation to effective strain, and strong correlation to each other. Three particular vibration resonances (or vibration modes -- these terms will be used interchangeably throughout this paper unless specified otherwise. The term mode does not necessarily indicate mode shapes or mode families.) emerge from this process as best candidates. A classic feature selection technique, Lasso linear regression, is then employed to identify critical power combinations of the three resonant mode frequencies. Two power combinations exhibit unique correlation to the measured equivalent axial strain at both test locations across all testing days, and thus show particular ability to predict RNT. The RNT is predicted at one test location using different models based on the power combination data from the other location, and vice versa, where the predictions satisfy standard RNT measurement accuracy expectations.

Pulse-compression is a correlation-based measurement technique successfully used in many nondestructive evaluation applications to increase the signal-to-noise ratio in the presence of huge noise, strong signal attenuation or when high excitation levels must be avoided. In thermography, the pulse-compression approach was firstly introduced in 2005 by Mulavesaala and co-workers [1], and then further developed by Mandelis and co-authors that applied to thermography the concept of the thermal-wave radar developed for photothermal measurements [2-3]. Since then, many measurement schemes and applications have been reported in the literature by several groups by using various heating sources, coded excitation signals, and processing algorithms. The variety of such techniques is known as pulse-compression thermography or thermal-wave radar imaging.

Even despite the continuous improvement of these techniques during these years, the advantages of using a correlation-based approach in thermography are still not fully exploited and recognized by the community. This is because up to now the reconstructed thermograms' time sequences after pulse-compression were affected by the so-called sidelobes, i.e. the temperature time trends of the pixels exhibit oscillations, especially in the cooling stage, so that they do not reproduce the output of a standard thermography measurement. This is a severe drawback since it hampers an easy interpretation of the data and their comparison with other thermography techniques.

To overcome this issue and unleash the full potential of the approach, this paper shows how it is possible to implement a pulse-compression thermography procedure capable of suppressing any sidelobe by using a pseudo-noise excitation and a proper processing algorithm.

At the end of the procedure, time-sequences of thermograms are reconstructed that correspond to the sample response to a well-defined virtual excitation, namely a rectangular pulse, making the pulse-compression procedure “transparent”. This allows the analysis of pixel time trends by using thermal theory-driven processing such as thermal signal reconstruction, pulsed-phase thermography, etc. Moreover, by tuning the characteristic of the pseudo-noise excitation, it is possible to pass from simulating a very short excitation pulse, retrieving results analogous to pulsed-thermography, to simulating long-pulse excitation to match the sample spectral characteristics maximizing the signal-to-noise ratio. This makes the procedure very flexible and extremely attractive in many applications such as high-attenuating materials, characterization of fast thermal phenomena, and inspection of fragile samples inspection, e.g. paintings or other artworks, etc.

Ultrasonic scattering matrices contain rich defect information and have great potential for characterising small crack-like defects. However, experimentally measured scattering matrices often exhibit some level of distortions compared to those of the idealised defects, posing challenges for accurate defect characterisation. In this paper, defect characterisation was performed by adopting a nearest neighbour approach based on a scattering matrix database of reference defects, and the test data were contaminated by coherent measurement noise of varying amplitudes. The performance of different similarity metrics on characterisation accuracy was studied, including the Euclidean similarity, cosine similarity, Pearson correlation coefficient, and the structural similarity index. Based on a comprehensive analysis of the strengths and weaknesses of different similarity metrics, we propose a defect characterisation framework by constructing similarity graphs and leveraging advanced graph neural networks. Within the proposed approach, multiple metrics were adopted to quantify the similarity between the scattering matrices of different defects, and an improved dynamic graph attention network was developed based on a customised neighbour sampling strategy to learn the optimal metric from the graph-structured data. Experimental results show that compared to the conventional approach which adopted a globally optimal similarity metric, the proposed method can reduce the root mean squared error for the length and angle predictions by 60.5% and 67.1%, respectively.

Electromagnetic tomography (EMT) is an emerging imaging technique that presents the property distribution of conductive or magnetic materials based on signals from the coil array on the EMT sensor. However, conventional EMT researches generally involve large-size EMT sensors, which are ineffective in detecting and discerning the size of small objects due to limited spatial resolution, sensitivity and relatively high noise level. As such, this paper presents a novel miniature EMT sensor, which incorporates 8 small coils around the circumference for imaging and 2 larger horizontal coils that generate a vertical field for target size distinction. Moreover, a novel equivalent theory is proposed to approximate the effect of the horizontal coils by the cumulative effect of the 8 small coils. An EMT testing system is established with the proposed sensor array and the multi-channel instrument developed in our lab. Experiments based on multiple sample distributions and different reconstruction algorithms validate the ability of the sensor to detect and distinguish the size of small objects of different materials. Furthermore, the equivalent theory was validated through the experiments.

This study presents an automated technique combining ultrasonic pulse echo method with machine learning algorithms to detect and classify the depth of interface defects in adhesively bonded joints. After data preprocessing for machine learning and extracting 32 ultrasonic features, the binary and ternary datasets were established for “defect”-“no defect” and its depth classifications. The importance and classification accuracy of various feature subsets—initial, single interface, minimised, tree-based, recursive, sequential, and LDA—were explored. A support vector machine (SVM) model was trained on these datasets. For “defect” vs. “no defect” classification, the initial feature subset achieved over 90 % accuracy on train/test data and 83 % on unseen data. For the ternary dataset, depth classification accuracy on unseen data in recursive feature subset was 97 % for “depth 1,” 62 % for “depth 2,” and 91 % for “depth 3.” The obtained results demonstrate prediction accuracy and suitability of ML models for classifying defects and predicting their depths in adhesive bonds.