Multi-scale Characteristics Research Articles

Investigation of foam concrete's mechanical properties and multi-scale damage evolution characteristics under uniaxial loading

The macroscopic mechanical behavior of foam concrete damages its mesoscopic pore structure. To study the macroscopic mechanical properties of foam concrete, a series of uniaxial compression tests were carried out. AE-DIC and SEM characterization were used to reveal the multi-dimensional damage evolution characteristics of foam concrete under load. Furthermore, the porosity response law of the foam concrete damage characteristics was analyzed using the finite element method. The results show a significant linear relationship between the residual and peak stress of foam concrete under uniaxial load (σr = kσp), and the coefficient k decreases with an increase in porosity. During uniaxial compression, tensile and shear damages were dominant on both sides of the yield point of foam concrete. The macroscopic load retention characteristics in the failure stage typically result from microscopic fracture structure reorganization and stress path optimization. The damage-dissipation energy during this period accounts for approximately 100% of absorbed energy. An increase in the porosity led to frequent adjustments in the stress path of the foam concrete, and the tensile damage tended to be significant and diffuse.

Investigating the causes of permeability anisotropy in heterogeneous conglomeratic sandstone using multiscale digital rock

Heterogeneous conglomeratic sandstone exhibits anisotropic physical properties, rendering a comprehensive analysis of its physical processes challenging with experimental measurements. Digital rock technology provides a visual and intuitive analysis of the microphysical processes in rocks, thereby aiding in scientific inquiry. Nevertheless, the multiscale characteristics of conglomeratic sandstone cannot be fully captured by a single-scale digital rock, thus limiting its ability to characterize the pore structure. Our work introduces a proposed workflow that uses multiscale digital rock fusion to investigate permeability anisotropy in heterogeneous rock. We use a cycle-consistent generative adversarial network to fuse computed tomography scan data of different resolutions, creating a large-scale and high-precision digital rock that comprehensively represents the conglomeratic sandstone pore structure. Subsequently, the digital rock is partitioned into multiple blocks, and the permeability of each block is simulated using a pore network. Finally, the total permeability of the sample is calculated by conducting an upscaling numerical simulation using the Darcy-Stokes equation. This process facilitates the analysis of the pore structure in conglomeratic sandstone and provides a step-by-step solution for permeability. From a multiscale perspective, this approach reveals that the anisotropy of permeability in conglomeratic sandstone stems from the layered distribution of grain sizes and differences in grain arrangement across different directions.

Multi-scale pore structure characteristics of coal under alternating hydraulic intrusion pressure

Multi-scale pore structure characteristics of coal under alternating hydraulic intrusion pressure

Pseudo-Eutectic of Isodimorphism to Design Biaxially-Oriented Bio-Based PA56/512 with High Strength, Toughness and Barrier Performances.

The biaxially-oriented PA56/512 has excellent mechanical strength, extensibility and water-oxygen barrier properties and has broad application prospects in green packaging, lithium battery diaphragm and medical equipment materials. The correlation between the aggregation structure evolution and macroscopic comprehensive properties of copolymer PA56/512 under biaxial stretching has been demonstrated in this work. The structure of the random copolymerization sequence was characterized by 13C Nuclear magnetic resonance (NMR). The typical isodimorphism behavior of the co-crystallization system of PA56/512 and its BOPA-56/512 films was revealed by differential scanning calorimetry (DSC) and X-ray diffraction (XRD) tests. And the aggregation structure, including the hydrogen bond arrangement, crystal structure and crystal morphology of PA56/512 before and after biaxial stretching, was investigated by XRD, Fourier-transform infrared spectroscopy (FTIR) and polarized optical microscopy (POM) tests. Furthermore, the effect of the biaxially-oriented stretching process on the mechanical properties of PA56/512 has been demonstrated. In addition, a deep insight into the influence of the structure on the crystallization process and physical-mechanical performance has been presented. The lowest melting point at a 512 content of 60 mol% is regarded as a "eutectic" point of the isodimorphism system. Due to the high disorder of the structural units in the polymer chain, the transition degree of the folded chain (gauche conformation) is relatively lowest when it is straightened to form an extended chain (trans conformation) during biaxially-oriented stretching, and part of the folded chain can be retained. This explains why biaxially stretched PA56/512 has high strength, outstanding toughness and excellent barrier properties at the pseudo-eutectic point. In this study, using the unique multi-scale aggregation structure characteristics of a heterohomodymite polyamide at the pseudo-eutectic point, combined with the new material design scheme and the idea of biaxial-stretching processing, a new idea for customized design of high-performance multifunctional polyamide synthetic materials is provided.

Dual-camera off-axis holographic particle tracking velocimetry: Development and application to air-blast swirl spray measurement

Quantifying the droplet field near the atomizer exit poses a challenge due to its three-dimensional (3D), multiscale and turbulent characteristics. An optical measurement method named dual-camera high-resolution holographic particle tracking velocimetry (DCHR-HPTV) is proposed. This method enables the extraction of near-field spatial Sauter mean diameter (SMD) distribution and three-dimensional two-component (3D-2C) velocity field, from off-axis hologram pairs with a frame interval of 6 μs, recorded using two synchronized industrial cameras and pulsed lasers. The system achieves a resolution of 9344 × 7000 pixels with an individual pixel size (dpix) of 3.2 μm. The droplet size measurement range is from 10 μm to above 500 μm with an absolute error within 3 μm. The velocity measurement range is adjustable and is set to be 0 to 38.4 m/s in this study, with errors of 0.23 m/s and 0.15 m/s for horizontal and axial components, respectively. The method is applied to investigate a spray region of 29 mm × 20 mm with a depth of 35 mm (equivalent to 20.3 cm3) at the exit of an air-blast atomizer. The spatial distribution of circumferentially integrated SMD along radial and axial directions is obtained. Statistical analysis is performed on both spatial and quantity distributions of droplet 3D-2C velocities. Concurrently acquired droplet size and velocity in the central toroidal recirculation zone (CTRZ) indicate that both the increasing air pressure and enlarged size range lead to a weakened droplet backflow motion. This method, distinguished by its large field of view and high resolution for particle velocimetry, is suited for dynamic analysis of near-nozzle droplet fields and can be employed in various testing scenarios involving 3D motion of micron-sized particles within multiscale flow fields.

Learning Aligned Vertex Convolutional Networks for Graph Classification.

Graph convolutional networks (GCNs) are powerful tools for graph structure data analysis. One main drawback arising in most existing GCN models is that of the oversmoothing problem, i.e., the vertex features abstracted from the existing graph convolution operation have previously tended to be indistinguishable if the GCN model has many convolutional layers (e.g., more than two layers). To address this problem, in this article, we propose a family of aligned vertex convolutional network (AVCN) models that focus on learning multiscale features from local-level vertices for graph classification. This is done by adopting a transitive vertex alignment algorithm to transform arbitrary-sized graphs into fixed-size grid structures. Furthermore, we define a new aligned vertex convolution operation that can effectively learn multiscale vertex characteristics by gradually aggregating local-level neighboring aligned vertices residing on the original grid structures into a new packed aligned vertex. With the new vertex convolution operation to hand, we propose two architectures for the AVCN models to extract different hierarchical multiscale vertex feature representations for graph classification. We show that the proposed models can avoid iteratively propagating redundant information between specific neighboring vertices, restricting the notorious oversmoothing problem arising in most spatial-based GCN models. Experimental evaluations on benchmark datasets demonstrate the effectiveness.

Scaling and Masking: A New Paradigm of Data Sampling for Image and Video Quality Assessment

Quality assessment of images and videos emphasizes both local details and global semantics, whereas general data sampling methods (e.g., resizing, cropping or grid-based fragment) fail to catch them simultaneously. To address the deficiency, current approaches have to adopt multi-branch models and take as input the multi-resolution data, which burdens the model complexity. In this work, instead of stacking up models, a more elegant data sampling method (named as SAMA, scaling and masking) is explored, which compacts both the local and global content in a regular input size. The basic idea is to scale the data into a pyramid first, and reduce the pyramid into a regular data dimension with a masking strategy. Benefiting from the spatial and temporal redundancy in images and videos, the processed data maintains the multi-scale characteristics with a regular input size, thus can be processed by a single-branch model. We verify the sampling method in image and video quality assessment. Experiments show that our sampling method can improve the performance of current single-branch models significantly, and achieves competitive performance to the multi-branch models without extra model complexity. The source code will be available at https://github.com/Sissuire/SAMA.

Deep Hierarchical Video Compression

Recently, probabilistic predictive coding that directly models the conditional distribution of latent features across successive frames for temporal redundancy removal has yielded promising results. Existing methods using a single-scale Variational AutoEncoder (VAE) must devise complex networks for conditional probability estimation in latent space, neglecting multiscale characteristics of video frames. Instead, this work proposes hierarchical probabilistic predictive coding, for which hierarchal VAEs are carefully designed to characterize multiscale latent features as a family of flexible priors and posteriors to predict the probabilities of future frames. Under such a hierarchical structure, lightweight networks are sufficient for prediction. The proposed method outperforms representative learned video compression models on common testing videos and demonstrates computational friendliness with much less memory footprint and faster encoding/decoding. Extensive experiments on adaptation to temporal patterns also indicate the better generalization of our hierarchical predictive mechanism. Furthermore, our solution is the first to enable progressive decoding that is favored in networked video applications with packet loss.

A comprehensive numerical investigation of multi-scale particle shape effects on small-strain stiffness of sands

The effects of multi-scale particle shape characteristics on the small-strain stiffness of granular soils remain controversial. This study revisits this topic using well-calibrated three-dimensional discrete-element simulations incorporating a particle roughness–embedded contact model and particles of realistic shape. Based on the numerical simulation results, the multi-scale particle shape effects on the small-strain stiffness of sands and the magnitude of Hardin's equation parameters are systematically investigated, and the underlying micro-mechanisms are also thoroughly explored. Results indicate that the small-strain stiffness increases with the increase of particles’ overall irregularity due to the increased mechanical coordination number, but decreases with the increase of particle surface roughness because of the decreased contact normal stiffness. In addition, the constant term parameter and void ratio term parameter of Hardin's equation increases and decreases linearly, respectively, with the particles’ overall regularity, but reduces and grows with the particle surface roughness, respectively. Furthermore, the stress exponent is almost unchanged with the particles’ overall regularity, but increases with the particle surface roughness, which determines the relative proportions of contacts under asperity-dominated, transitional and Hertzian stages. The study helps to advance the cross-scale understanding of multi-scale particle shape information in relation to small-strain stiffness of sands.

Signal processing collaborated with deep learning: An interpretable FIRNet for industrial intelligent diagnosis

Due to the neglect of prior characteristics and the lack of explicit constraints on fault knowledge, conventional intelligent diagnosis methods suffer from great hardships in exacting fault-sensitive information and making explainable decisions, resulting in poor interpretability and inferior robustness. Motivated by the excellent multiscale analysis performance of signal processing and the powerful feature mining ability of deep learning, this study proposes an interpretable FIRNet for equipment intelligent diagnosis under strong noise environments. It consists of a well-crafted FIRLayer and a deep learning backbone. Inspired by the modulation principle that fault-sensitive components are normally modulated into multiscale mode characteristics, multiple ex-ante interpretable filters with two learnable parameters, including center frequency and bandwidth, are analytically designed to process the sequence signal sets, represented as signal-processing-based FIRLayer where the extracted multiscale feature maps are taken as an interpretable status information expression. Subsequently, multiscale convolutional kernels are established to extract the high-level feature maps and further make the final diagnostic decisions, represented as the deep learning backbone. The simulated and experimental results show that the proposed FIRNets have higher identification precision compared to the other nine deep learning models. Specially, three aspects, including model interpretability, noise robustness capacity and edge intelligent diagnosis, are further analyzed to illustrate the interpretable advantages of FIRNets. Hereinto, feature visualizations of the FIRLayer and mode decision contributions of the deep learning backbone are in-depth analyzed to verify the interpretable feature representation and decision-making principle of FIRNets. The results indicate that FIRNets have superior ex-post interpretability compared to other methods. Focusing on industrial practices, an efficient edge diagnosis system based on a pruned FIRNet is established, and an online diagnosis accuracy of more than 99% has been achieved. It can be foreseen that the proposed FIRNets show great potential and competitiveness to promote the edge computing application of equipment intelligent diagnosis.

An image watermarking scheme: combining the transform domain and deep learning modality with an improved scrambling process

ABSTRACT This paper proposes a novel image watermarking scheme with the following two processes. In the embedding process, the cover image is subjected to NSCT to obtain the properties of multi-scale and directional characteristics and fed to DWT which decomposes the contourlet image into various sub-bands. The wavelet decomposition image is subjected to a region selection step, where the regions are selected based on the DL strategy for selecting the region from the contourlet of the cover image where the features, including LOOP feature, mean feature, Canny features, and ResNet features are extracted and subjected to the DNN classifier. Conversely, the watermark image performs NSCT and DWT as the cover image and is fed into further steps to eliminate the correlation of pixels. For that, it proposes the FTTS approach to scramble the contourlet of watermark images, in which tent map and dyadic transformation are employed to eliminate the correlation among pixels. Then the resultants of region selection and FTTS are embedded to obtain the watermarked image. On the other hand, the extraction process takes place, which extracts the embedded watermarked image and applies inverse scrambling and inverse an NSCT approach to recover the watermark image against any malicious attack.

CCDN-DETR: A Detection Transformer Based on Constrained Contrast Denoising for Multi-Class Synthetic Aperture Radar Object Detection.

The effectiveness of the SAR object detection technique based on Convolutional Neural Networks (CNNs) has been widely proven, and it is increasingly used in the recognition of ship targets. Recently, efforts have been made to integrate transformer structures into SAR detectors to achieve improved target localization. However, existing methods rarely design the transformer itself as a detector, failing to fully leverage the long-range modeling advantages of self-attention. Furthermore, there has been limited research into multi-class SAR target detection. To address these limitations, this study proposes a SAR detector named CCDN-DETR, which builds upon the framework of the detection transformer (DETR). To adapt to the multiscale characteristics of SAR data, cross-scale encoders were introduced to facilitate comprehensive information modeling and fusion across different scales. Simultaneously, we optimized the query selection scheme for the input decoder layers, employing IOU loss to assist in initializing object queries more effectively. Additionally, we introduced constrained contrastive denoising training at the decoder layers to enhance the model's convergence speed and improve the detection of different categories of SAR targets. In the benchmark evaluation on a joint dataset composed of SSDD, HRSID, and SAR-AIRcraft datasets, CCDN-DETR achieves a mean Average Precision (mAP) of 91.9%. Furthermore, it demonstrates significant competitiveness with 83.7% mAP on the multi-class MSAR dataset compared to CNN-based models.

Multi-Level Feature-Refinement Anchor-Free Framework with Consistent Label-Assignment Mechanism for Ship Detection in SAR Imagery

Deep learning-based ship-detection methods have recently achieved impressive results in the synthetic aperture radar (SAR) community. However, numerous challenging issues affecting ship detection, such as multi-scale characteristics of the ship, clutter interference, and densely arranged ships in complex inshore, have not been well solved so far. Therefore, this article puts forward a novel SAR ship-detection method called multi-level feature-refinement anchor-free framework with a consistent label-assignment mechanism, which is capable of boosting ship-detection performance in complex scenes. First, considering that SAR ship detection is susceptible to complex background interference, we develop a stepwise feature-refinement backbone network to refine the position and contour of the ship object. Next, we devise an adjacent feature-refined pyramid network following the backbone network. The adjacent feature-refined pyramid network consists of the sub-pixel sampling-based adjacent feature-fusion sub-module and adjacent feature-localization enhancement sub-module, which can improve the detection capability of multi-scale objects by mitigating multi-scale high-level semantic loss and enhancing low-level localization features. Finally, to solve the problems of unbalanced positive and negative samples and densely arranged ship detection, we propose a consistent label-assignment mechanism based on consistent feature scale constraints to assign more appropriate and consistent labels to samples. Extensive qualitative and quantitative experiments on three public datasets, i.e., SAR Ship-Detection Dataset (SSDD), High-Resolution SAR Image Dataset (HRSID), and SAR-Ship-Dataset illustrate that the proposed method is superior to many state-of-the-art SAR ship-detection methods.

Interpretable machine learning model of effective mass in perovskite oxides with cross-scale features

The interpretability of machine learning reveals associations between input features and predicted physical properties in models, which are essential for discovering new materials. However, previous works were mainly devoted to algorithm improvement, while the essential multi-scale characteristics are not well addressed. This paper introduces distortion modes of oxygen octahedrons as cross-scale structural features to bridge chemical compositions and material properties. Combining model-agnostic interpretation methods, we are able to achieve interpretability even using simple machine learning schemes and develop a predictive model of effective mass for a widely used material type, namely perovskite oxides. With this framework, we reach the interpretability of the model, understanding the trend of the effective mass without any prior background information. Moreover, we obtained the knowledge only available to experts, i.e., the interpretation of effective mass from the s–p orbitals hybridization of B-site cations and O2− in ABO3 perovskite oxides.

Flow stirred by open turbine: A modal analysis of multiscale flow characteristics

AbstractThe turbulence observed in the stirred tank is a result of the superposition of flows at different scales, each exerting various influences on the operational processes. To understand the flow characteristics associated with stirring, this paper undertakes particle image velocimetry (PIV) experimental research on an unbaffled tank stirred by an open turbine featuring four distinct blade deflection angles. The flow field in the unbaffled tank manifests as a double‐circulation flow pattern; however, it is worth noting that the blade deflection angle exerts a significant impact on the stirring process. Specifically, when the blade deflection angle reached 90°, the flow discharge aligned closely with the radial direction, accompanied by pronounced fluctuations in the high‐velocity region. The Lagrangian coherent structures (LCS) formed were subsequently extracted and subjected to finite‐time Lyapunov exponent (FTLE) analysis. Notably, a more regular wake pattern emerged in proximity to the blade tip, while a larger scale flow structure persisted along the flow direction. To systematically analyze the flow field stirred by the open turbine, modal analysis was employed. By reconstructing the flow field based on the principle of 50% energy content, the decomposition of flows at different scales was achieved. The turbulent kinetic energy (TKE) generated by the various scale flows was subsequently calculated, revealing that the high TKE values primarily originate from the large‐scale flow structure. Conversely, the TKE generated by the small‐scale flow exhibits a uniform distribution across different regions. Notably, in regions characterized by higher velocities, the peak TKE associated with the small‐scale flow also manifests; however, it is significantly smaller in magnitude compared to that generated by the original flow or the large‐scale flow.

Dilatancy and Multi-Scale failure characteristics of a foliated rock under triaxial confinement unloading conditions

Controlled by the foliation structure, foliated rocks exhibit mechanical anisotropy and cause squeezing deformation and subsequent supporting structure failure during tunnel excavation. Here, the deformation and failure characteristics of a phyllite under unloading stress paths were experimentally investigated. The strength and elasticity parameters of phyllite exhibited moderate ‘shoulder’ shaped anisotropy, and specimens failed in three different modes: cross-plane shear-tensile fracture, in-plane shear sliding, and interlayer buckling, depending on the foliation angle and initial confining pressure. Meanwhile, the evolution of dilatancy angle during volume dilatancy also showed significant anisotropy, either increasing or decreasing with the increasing plastic shear strain during post-peak loading. We further discussed the mechanism of the evolution of dilatancy angle based on the multi-scale observation of the failure characteristics. Generally, interlayer buckling improved the dilatancy angle and roughness loss during shear failure resulted in dilatancy angle degradation. These findings provided evidence for a deeper understanding of the failure mechanisms of tunnel rock mass in foliated strata.

Cracking process and microstructural characteristics of granite under heating–cooling alternations

Studying the damage correlation mechanism of rocks subjected to heating–cooling alternations is of important significance for deeply understanding the permeability increase mechanism of reservoirs during heat extraction from hot dry rocks and assessing the reservoir stability. By carrying out the uniaxial compression test, Brazilian splitting test, nuclear magnetic resonance test, and scanning electron microscopy test on granite after heating–cooling alternations, the evolution laws of physical and mechanical properties as well as the microstructures of granite were studied. In addition, the multi-scale response characteristics and the damage correlation mechanism of granite after heating–cooling alternations were analyzed. Research results show that, when the heat treatment temperature rises to 600 °C, the physical and mechanical parameters, including the wave velocity, strength, and elastic modulus, reduce abruptly. The heat treatment temperature exerts more significant influences on the physical and mechanical properties of the granite than the heating–cooling alternation cycles. Both the temperature and alternation cycles are beneficial to the microcrack development and propagation, while the temperature more significantly affects the development degree of microcracks. Under the external load, the microscopic defects in the granite after heating–cooling alternations guide the propagation of macrocracks and cause more serious damage to the microstructures in the granite, exhibiting more complex failure modes. The results provide a theoretical basis for fracturing technology and stability evaluation of high-temperature reservoirs.

Meta-Transfer Metric Learning for Time Series Classification in 6G-Supported Intelligent Transportation Systems

Deep learning-based time series classification in 6G-supported Intelligent Transportation Systems (ITS) helps transport decision-making. Deep learning classifier training necessitates a large amount of labeled data for feature extraction. Labeling time series data in 6G-supported ITS is tough. Meta-learning can be used to train deep classifiers with limited data. However, in meta-learning, the tasks are frequently modeled by a low-complexity base learner. It is unable to use more complicate and powerful structures. The meta-learning-pretrained classifier can only perform new classification problems with the same number of classes. Most pre-training strategies do not prioritize enhancing the pre-training phase’s convergence rate and lowering the computational cost. Most research work aims to improve classification performance by increasing the complexity of the classification model. However, this raises computing costs. In this paper, we propose a one-dimensional Multi-Scale Dilated Convolution Neural Network time series classifier (MSDCNN). MSDCNN combines multi-scale CNN and dilated CNN. It can extract multi-scale characteristics from time series and reduce the complexity of the classifier. Furthermore, we propose a pre-training strategy, called Meta-transfer metric Learning using Scale function (MLS). MLS allows the classifier to gain experience from different tasks with various numbers of classes. Experiments show that MLS reduces pre-training computation costs during the pre-training phase. The pre-trained classifier, without using any fine tuning techniques, achieves the highest accuracy by comparing with the state-of-the-art methods. Finally, we present a case study of applying MSDCNN and MLS to detect road accidents in 6G-supported transportation systems.

Research on the Multiscale Microscopic Pore Structure of a Coalbed Methane Reservoir

Coal rock pores are the space in which coalbed gas is stored and flows. Accurately characterizing the pore structure of coalbed gas is the foundation of coalbed gas reserve assessment and production forecasting. Traditional experimental methods are unable to characterize the multi-scale pore structure characteristics of coal rock. In this paper, a multi-scale pore structure characterization method is proposed by coupling various experimental methods, including low-pressure nitrogen gas adsorption experiments, X-ray computed tomography (XCT) imaging technology, and scanning electron microscopy (SEM). Using Zhengzhuang coalbed gas as an example, the micro-pore structure of coalbed gas reservoirs is characterized and depicted from a multi-scale perspective. The results indicate that a single experimental approach can only partially reveal the microstructure of coal rock pores. The combined use of multiple methods can accurately reveal the full-scale microstructure of coal rock pores. The pore structure of the experimental coal rock samples exhibits multi-scale characteristics, with a complex variety of pore types, including inorganic pores, organic pores, and fractures. Organic pores are predominant, with a small number of inorganic pores, and their sizes range from 2 nm to 50 μm. Mineral particles and fractures are observed at both the nanoscale and microscale, exhibiting typical multi-scale characteristics, with quartz being the predominant mineral.

Intrinsic Mode-Based Network Approach to Examining Multiscale Characteristics of Sea Surface Temperature Variability

Variability of sea surface temperature (SST), characterized by various spatiotemporal scales, is a proxy of climate change. A network analysis combined with empirical mode decomposition is newly presented for examining scale-dependent spatial patterns of SST variability. Our approach is applied to SST anomaly variability in the East/Japan Sea (EJS), consisting of satellite-derived daily datasets of 0.25° × 0.25° resolution from 1981 to 2023. Through the spatial distribution of instantaneous energy in intrinsic modes and features of intrinsic-mode networks, scale-dependent spatiotemporal features are found. The season-specific spatial pattern of energy density is observed only for weekly to semiannual modes, while a persistent high-energy distribution in the tongue-shaped region from East Korea Bay (EKB) to the Sub-Polar Front (SPF) is observed only for annual-to-decadal modes. The seasonality is apparent in the time evolution of energy only for weekly-to-annual modes, with a peak in summer and an increasing trend since the 2010s. Hubs of intrinsic-mode networks are observed in the whole southern area (some northern part) of EJS during the summer (winter), only for monthly to semiannual modes. Regional communities are observed only for weekly to seasonal modes, while there is an inter-basin community with annual-to-biennial modes, incorporating two pathways of East Sea Intermediate Water (ESIW).