Local Propagation Research Articles

An effective multi-scale interactive fusion network with hybrid Transformer and CNN for smoke image segmentation

Smoke has visually elusive appearances, especially in low-light conditions, so it is quite difficult to quickly and accurately detect smoke from images. To address these challenges, we design a dual-encoder structure of Transformer and Convolutional Neural Network (CNN) to propose an effective Multi-scale Interactive Fusion Network (MIFNet) for smoke image segmentation. To enhance the presentation of features, we propose a Local Feature Enhancement Propagation (LFEP) module to enhance spatial details. To optimize global and local features for efficient fusion, we integrate LFEP into the original Transformer to replace the traditional multi-head self-attention mechanism. Then, we propose a Multi-level Attention Coupled Module (MACM) to fuse Transformer and CNN features of the dual-encoder. MACM can flexibly focus on information interaction between different levels of two encoding paths. Finally, we design a Prior-guided Multi-scale Fusion Decoder (PMFD), which combines prior knowledge with a multi-scale feature fusion strategy to improve the performance of segmentation. Experimental results demonstrate that MIFNet substantially outperforms the state-of-the-art methods. MIFNet achieves a mean Intersection over Union (mIoU) of 81.6% on the synthetic smoke (SYN70K) dataset, and a remarkable accuracy of 98.3% on the forest smoke dataset.

Vector Field Heterogeneity as a Novel Omnipolar Mapping Metric for Functional Substrate Characterisation in Scar-related Ventricular Tachycardias

BackgroundVector Field Heterogeneity (VFH) is a novel omnipolar metric to quantify local propagation heterogeneities that may identify functionally critical sites for ablation in scar-related VT. ObjectiveAssess the diagnostic value of VFH to identify abnormal propagation patterns during ventricular substrate mapping and compare VFH in VT isthmus sites (IS), low-voltage bystander (LVA) and normal voltage areas (NVA). MethodsSubstrate maps acquired with a 16-pole grid catheter in patients with scar-related VT were segmented into sites corresponding to IS, LVA and NVA (defined as omnipolar voltages >&<1.5mV) based on corresponding VT activation maps. For each 4×4 electrode-array acquisition, omnipolar-derived vector maps of the directions of electrical propagation were computed offline and a VFH value per clique ranging from 0 (perfect planar wave) to 1 (maximal disorganisation) derived. Results16 patients were studied (56.3% ischemic) evaluating 9 endocardial/7 epicardial substrate maps. VFH metric at IS was 0.57±0.26, LVA 0.52±0.34, NVA 0.16±0.24. VFH at IS was higher (more disorganised) than at LVA (p<0.001). VFH in NVA was significantly lower (i.e. more homogenous) compared to IS and LVA (p<0.001, respectively). Within the isthmus, highest heterogeneity was recorded at entry (0.61±0.24), lowest at exit sites (0.44±0.27). ConclusionVFH mapping identified a significant increase in electrical heterogeneity at IS compared to bystanders with highest heterogeneity at the entrance. Absolute differences were small and substantial overlap between sites were recorded precluding its use as a stand-alone mapping approach. VFH may be integrated in existing mapping strategies as complementary information to support identification of ablation targets.

Local Strain-Dependent Etching Patterns of Chemical Vapor Deposited Molybdenum Disulfide.

This study reveals a local strain-dependent etching behavior that enables the formation of distinguished etching patterns in differently strained chemical vapor deposited (CVD) 2D molybdenum disulfide (MoS2) monolayers. It is demonstrated that when the local tensile strain of CVD 2D MoS2 is as uniformly low as ɛ ≈ 0.33% or less, the oxidative etching pattern possesses conventional triangular etching pits (TEPs), while when the local tensile strain is as uniformly high as ɛ ≈ 0.55% or larger, the oxidative etching pattern consist of uniformly oriented hexagonal etching channels (HECs). More interestingly, when the CVD 2D MoS2 monolayer has heterogenous strain distribution from ɛ ≈ 0.55% (center region) to ɛ ≈ 0.33% (perimeter region), the oxidative etching pattern comprise of non-uniformly hexagonal-mixed-parallel etching channels (HPECs). The further characterization and analysis reveal the formation mechanism of such strain-dependent etching patterns is built on the local strain-related fractures propagation under oxidative etching, as well as the anisotropy fractures-based oxidative etching kinetics. This study may enhance the understanding of the relationship between etching and growth features of 2D TMDs, and paves the way to etching-nanostructured (or defect) engineering of 2D TMDs and other 2D materials for potential applications in electrocatalysis and optoelectronics.

Extracting dynamical maps of non-Markovian open quantum systems.

The most general description of quantum evolution up to a time τ is a completely positive tracing preserving map known as a dynamical mapΛ̂(τ). Here, we consider Λ̂(τ) arising from suddenly coupling a system to one or more thermal baths with a strength that is neither weak nor strong. Given no clear separation of characteristic system/bath time scales, Λ̂(τ) is generically expected to be non-Markovian; however, we do assume the ensuing dynamics has a unique steady state, implying the baths possess a finite memory time τm. By combining several techniques within a tensor network framework, we directly and accurately extract Λ̂(τ) for a small number of interacting fermionic modes coupled to infinite non-interacting Fermi baths. First, we use an orthogonal polynomial mapping and thermofield doubling to arrive at a purified chain representation of the baths whose length directly equates to a time over which the dynamics of the infinite baths is faithfully captured. Second, we employ the Choi-Jamiolkowski isomorphism so that Λ̂(τ) can be fully reconstructed from a single pure state calculation of the unitary dynamics of the system, bath and their replica auxiliary modes up to time τ. From Λ̂(τ), we also compute the time local propagator L̂(τ). By examining the convergence with τ of the instantaneous fixed points of these objects, we establish their respective memory times τmΛ and τmL. Beyond these times, the propagator L̂(τ) and dynamical map Λ̂(τ) accurately describe all the subsequent long-time relaxation dynamics up to stationarity. These timescales form a hierarchy τmL≤τmΛ≤τre, where τre is a characteristic relaxation time of the dynamics. Our numerical examples of interacting spinless Fermi chains and the single impurity Anderson model demonstrate regimes where τre ≫ τm, where our approach can offer a significant speedup in determining the stationary state compared to directly simulating the long-time limit. Our results also show that having access to Λ̂(τ) affords a number of insightful analyses of the open system thus far not commonly exploited.

High-resolution observations of shallow-water acoustic propagation with distributed acoustic sensing.

Distributed acoustic sensing (DAS), converting fiber-optic cables into dense acoustic sensors, is a promising technology that offers a cost-effective and scalable solution for long-term, high-resolution studies in ocean acoustics. In this paper, the telecommunication cable of Martha's Vineyard Coastal Observatory (MVCO) is used to explore the feasibility of cable localization and shallow-water sound propagation with a mobile acoustic source. The MVCO DAS array records coherent, high-quality acoustic signals in the frequency band of 105-160 Hz, and a two-step inversion method is used to improve the location accuracy of DAS channels, reducing the location uncertainty to ∼2 m. The DAS array with refined channel positions enables the high-resolution observation of acoustic modal interference. Numerical simulations that reproduce the observed interference pattern suggest a compressional speed of 1750 m/s in the sediment, which is consistent with previous in situ geoacoustic measurements. These findings demonstrate the long-term potential of DAS for high-resolution ocean acoustic studies.

Study on the extension mechanism of fissures in expansive soil based on fracture mechanics test method

Expansive soil, characterized by widespread fissures, is a special type of soil prone to localized fissure propagation, leading to failure and instability, often resulting in landslides. Numerous studies have applied fracture mechanics theory to analyze soil failure along fissures, but no standardized testing method has been established. This paper reviews existing soil fracture toughness testing methods, designs an integrated system combining digital image correlation(DIC) technique and electrical resistance testing, and employs the Cracked Chevron-Notched Brazilian Disc (CCNBD) method to assess the fracture toughness of expansive soil. The deformation and internal damage accumulation processes of the specimens were monitored. The KΙc and KΙΙc of expansive soil were tested at moisture contents of 15 %, 20 %, and 25 % using specimens with three different crack length ratios (a/R). The role of water was explored by injecting water into the fissures. The results showed that KΙc of expansive soil ranged from 10 to 36 kPa·m0.5, with a linear negative correlation to moisture content and a proportional relationship to tensile strength, having a proportionality coefficient of 0.331. The KΙΙc ranged from 20 to 70 kPa·m0.5,and it is greater than KΙc under the same conditions. Based on the load, internal damage, strain, and fissure area during the tests, the failure process of expansive soil along fissures was divided into four stages: initial deformation stage(I), quasi-elastic deformation stage(II), fissure extension stage(III), and failure stage(IV). Water injection into the fissures reduced the soil's fracture toughness, with a more significant reduction as a/R increased and the failure showed progressive behavior. The CCNBD method, combined with DIC technique and electrical resistance testing, effectively measures the fracture toughness of soil, aiding in understanding the failure mechanism along fissures and providing a basis for preventing and controlling landslides and other hazards in expansive soils.

A novel method for estimating avian roost sizes using passive acoustic recordings using deep neural network

Abstract Communal bird roosts serve as information centres and a means of thermoregulation for many species. While some communally roosting species are major pests and cause dis‐amenities, others are of conservation concern. Estimating the population of roosting birds can provide a useful proxy of population size and possibly a more reliable estimate than other sampling techniques. However, estimating these populations is challenging as some roosts are large and often occluded in foliage. Previous acoustic methods such as paired sampling, microphone arrays and use of call rate have been used to estimate bird abundances; however, these are less suited for estimating large roost populations where hundreds of individuals are calling in unison. To address this challenge, we explored using machine learning techniques to estimate a roost population of the Javan myna, Acridotheres javanicus, an invasive species in Singapore. While one may expect to use sound intensity to estimate roost sizes, it is affected by various factors such as distance to the recorder, local propagation conditions (e.g. buildings and trees), weather conditions, and noise from other sources. Here, we used a deep neural network to extract higher order statistics from the sound recordings and use those to help estimate roost sizes. Additionally, we validated our method using automated visual analysis with a dual‐camera setup and manual bird counts. Our estimated bird counts over time using our acoustic model matched the automated visual estimates and manual bird counts at a selected Javan myna roost, thus validating our approach. Our acoustic model estimated close to 400 individual mynas roosting in a single tree. Analyses of additional recordings of Javan myna roosts conducted on two separate occasions and at a different roost location using our acoustic model showed that our roost estimates over time also matched our automated visual estimates well. Practical implication: Our novel approach of estimating communal roost sizes can be achieved robustly using a simple portable acoustic recording system. Our method has multiple applications such as testing the efficacy of avian roost population control measures (e.g. roost tree pruning) and monitoring the populations of threatened bird species that roost communally.

Cascading spatial drought network: A complex networks approach to track propagation of meteorological droughts to agricultural droughts

Meteorological droughts often propagate to agricultural (and other) droughts, both spatially and temporally. The present study proposes a novel complex networks-based cascading spatial drought network to examine the spatial propagation of meteorological droughts in a region to agricultural droughts in other regions. This is done through: (1) establishing stable homogeneous drought communities; (2) investigating inter-community drought propagation; (3) locating drought sources; and (4) evaluating drought connections within major crop belts. The approach is implemented to study droughts in the Indian-subcontinent during the period 1948–2022. Monthly precipitation and root-zone soil moisture data from GLDAS (Global Land Data Assimilation System) are used to compute the standardized precipitation index (SPI) for meteorological droughts and standardized soil moisture index (SSI) for agricultural droughts. Primarily, the drought network is demarcated into several subsets of network communities within which clusters of localized propagation take place. Multi-community subgraphs combining different communities are also formed to understand the long-distance inter-community drought linkages. Using network centrality measures, such as degree, closeness, and clustering coefficient, network properties of scale-freeness, small-worldness, and presence of rich-clubs are checked. Although the overall network does not exhibit any of these properties, certain subgraphs have significant small-worldness, rich-clubs, and partial scale-freeness. Some of the crucial nodes that support these network properties lie in the monsoon pathways (in the Western Ghats), and others have a strong association with El Niño Southern Oscillation (ENSO) teleconnections, thus validating the ability of the drought network to capture seasonal and climatic features. Additionally, subgraphs of nodes with high productivity of different food crops are created to study drought propagation within crop belts. Barring potential shortcomings related to data dependencies, the cascading spatial drought network helps identify an impending agricultural drought that could strengthen our ability to forecast droughts.

Corrosion behavior of 2A97 Al−Cu−Li alloys in different thermomechanical conditions by quasi-in-situ analysis

As a promising material in the aircraft industry, 2A97 Al−Cu−Li alloy exhibits high corrosion susceptibility that may limit its application. In the present work, to illustrate the influences of precipitate and grain-stored energy on localized corrosion evolution in 2A97 Al−Cu−Li alloy, cold working and artificial aging were carried out to produce 2A97 Al−Cu−Li alloys under different thermomechanical conditions. Quasi-in-situ analysis, traditional immersion test and electrochemical measurement were then conducted to examine the corrosion behavior of 2A97 alloys. It is revealed that precipitate significantly affects Cu enrichment at corrosion fronts, which determines corrosion susceptibility of alloys, whereas grain-stored energy distribution is closely associated with localized corrosion propagation. It is also indicated that quasi-in-situ analysis exhibits a consistent corrosion evolution with traditional immersion tests, which is regarded as a proper method to explore localized corrosion mechanisms by providing local microstructural information with enhanced time and spatial resolutions.

The qP- and qSV-wave separation in 2D vertically transversely isotropic media and their applications in elastic reverse time migration

Compressional and shear wave separation is an important step in anisotropic elastic reverse time migration (ERTM). With separated wavefields, we can generate images between different wave types and reveal more subsurface physical properties, as well as remove unwanted crosstalk and improve image quality. Traditional Helmholtz decomposition based on divergence and curl operations for isotropic media cannot be directly extended to vertically transversely isotropic (VTI) media. Wave-mode separation methods, similar to the nonstationary spatial filter and Poynting vector, cannot preserve the phase and amplitude information of the original coupled wavefield, thus limiting their application in ERTM. Currently, the anisotropic wavefield decomposition methods include the wavenumber-domain approach, the low-rank approximation, and other separation approaches based on different approximations, e.g., weak or elliptical anisotropy. These methods are either computationally intensive or involve large separation errors, especially in models with strong heterogeneities and anisotropy. The VTI wavefield separation operator is essentially defined in a mixed space-wavenumber domain. We introduce scalar operators to transfer operations from this mixed space-wavenumber domain to the space domain. The local wave propagation direction in the scalar operators is estimated using the Poynting vector, which is highly computationally efficient in the space domain. When we apply the wave separation method to ERTM, we not only obtain the vectorized qP and qSV waves but also retrieve the scalar qP wavefield. The scalar and vector wavefields preserve the amplitude and phase information in the original coupled wavefield. For anisotropic ERTM, we suggest using the scalar imaging condition to generate the PP image and the magnitude- and sign-based vector imaging condition to produce the PS image, both having higher image accuracy than the dot-product imaging condition. Numerical examples are used to validate our anisotropic wavefield separation method and the related ERTM workflow.

Adjustable stereo path design method based on pin-sculpture acoustic topological insulator with Z-dislocation defect immunity

The field of topological protected wave engineering, inspired by quantum mechanics, has generated significant interest. Acoustic analogs of electronic topological insulators provide new opportunities for manipulating sound propagation with unconventional acoustic edge modes that are immune to backscattering. Numerous reports have been published on the design of two-dimensional acoustic topological insulators (ATIs). However, the sound path of a two-dimensional design is simple, and its ability to control sound waves is limited. On the other hand, the design of 3D ATIs is relatively complex, making it difficult to manufacture and limiting its versatility. Based on the design idea of the 2D ATIs, inspired by the art named 3D pin-sculpture, an adjustable structure of a finite size consisting of spindle-shaped units with a variable cross section is designed to realize flexible path transformation. Furthermore, unlike two-dimensional structural defects, such as cavities and disorder, the analysis of vertical dislocation defects in finite-sized structures allows for the design of local sound propagation along the z-direction, providing a concept for constructing a stereo path. The designed structure also serves two functions: acoustic switch and delay. This idea offers an alternative approach to designing complex sound transmission paths.

Unveiling the quantitative relationship between microstructural features and quasi-static tensile properties in dual-phase titanium alloys based on data-driven neural networks

The quasi-static mechanical properties of α+β dual-phase titanium alloys are susceptible to their microstructural features, presenting a complex, high-dimensional nonlinear relationship, which hinders the rapid development of high-performance materials. In this work, 4065 micro-representative models were virtually constructed with varying volume fractions of α and β phases and characteristic dimensions via high-throughput finite element simulation, incorporating a cohesive zone model to simulate the interfaces between the two phases. Especially, the established representative models were experimentally verified by two groups of real material microstructures, and the results showed that the relative errors were not more than 9.5 % in microstructural characteristics and quasi-static mechanical properties. Afterward, a neural network model was developed to correlate the quasi-static tensile properties with the microstructural features of the dual-phase TC6 titanium alloys, achieving an 88.2 % accuracy in predicting overall mechanical performance. Utilizing the Shaply Additive Explanation method, it was found that the primary α phase's volume fraction and the secondary α phase's width were the most significant microstructural features affecting quasi-static strength. Specifically, the volume fraction of the primary α phase and the width of the secondary α phase negatively affected strength, while the width of the secondary α phase positively influenced plasticity. Notably, the primary α phase's volume fraction had a quadratic curve pattern of influence on plasticity. The intrinsic mechanisms behind these laws were further revealed based on local stress-strain responses and crack propagation analysis. Ultimately, the optimal microstructural features with strength-plasticity balance were identified through the lower threshold method: a secondary α phase width of about 1 μm and a primary α phase volume fraction ranging from 0.1 to 0.2, effectively facilitating microstructure design.

Localized Corrosion Behavior and Corrosion Microstructure of LZ91 and LAZ931 Mg Alloys

LZ91 (Mg-9 wt%Li-1 wt%Zn) and LAZ931 (Mg-9 wt%Li-3 wt%Al-1 wt%Zn) alloys are emerging ultra-lightweight Mg alloys, which can be used in applications that require weight reduction. Although LZ91 alloy has found some success in the 3C industry, the relatively inferior mechanical strength limits its widespread application. To improve the mechanical properties, Al was added to LZ91 alloy to form LAZ931 alloy. However, both LZ91 and LAZ931 alloys are primarily composed of a dual-phase structure with α (HCP) and β (BCC) phases, which form a microgalvanic couple in the corrosion environment. Furthermore, adding Al introduces additional secondary phases, such as AlLi and MgLi2Al, which complicate the corrosion behavior and mechanism. Therefore, this study investigates the corrosion behavior and microstructure of LZ91 and LAZ931 Mg alloys in 3.5 wt% NaCl solution and the relationships to the phase distribution in the alloy substrates.By in situ optical microscope (OM) observations and open circuit potential (OCP) monitoring, localized corrosion initiates on the LAZ931 alloy earlier than on the LZ91 alloy. The early initiation and propagation of localized corrosion on the LAZ931 alloy result in worse corrosion resistances and faster corrosion rates measured by electrochemical impedance spectroscopy (EIS) and potentiodynamic polarization curves. By scanning Kelvin probe force microscopy (SKPFM), the Li-rich β phase is more active than the Mg-rich α phase. However, interestingly, localized corrosion was found to preferentially propagate inside the α phase, which seems to contradict the SKPFM measurements. The preferential propagation of localized corrosion in the α phase could be related to the difference in the surface corrosion film thicknesses on the α and β phases before localized corrosion propagation. The localized corrosion behavior and corrosion microstructure of the alloys will be discussed based on electrochemical measurements and site-specific microstructure characterizations. Figure 1

Coplane-constrained sparse depth sampling and local depth propagation for depth estimation

Depth estimation with sparse reference has emerged recently, and predicts depth map from a monocular image and a set of depth reference samples. Previous works randomly select reference samples by sensors, leading to severe depth bias as this sampling is independent to image semantic and neglects the unbalance of depth distribution in regions. This paper proposes a Coplane-Constrained sparse Depth (CCD) sampling to explore representative reference samples, and design a Local Depth Propagation (LDP) network for complete the sparse point cloud map. This can capture diverse depth information and diffuse the valid points to neighbors with geometry prior. Specifically, we first construct the surface normal map and detect coplane pixels by superpixel segmenting for sampling references, whose depth can be represented by that of superpixel centroid. Then, we introduce local depth propagation to obtain coarse-level depth map with geometric information, which dynamically diffuses the depth from the reference to neighbors based on local planar assumption. Further, we generate the fine-level depth map by devising a pixel-wise focal loss, which imposes the semantic and geometry calibration on pixels with low confidence in coarse-level prediction. Extensive experiments on public datasets demonstrate that our model outperforms SOTA depth estimation and completion methods.

Dimers and discrete breathers in Bose-Einstein condensates in a quasi-periodic potential

A quasi-one-dimensional Bose-Einstein condensate loaded into a quasi-periodic potential created by two sublattices of comparable amplitudes and incommensurate periods is considered. Although the conventional tight-binding approximation is not applicable in this setting, the description can still be reduced to a discrete model that accounts for the modes below the mobility edge. In the respective discrete lattice, where no linear hopping exists, solutions and their dynamics are governed solely by nonlinear interactions. Families of nonlinear modes, including those with no linear limit, are described with a special focus on dimers, which correspond to breather solutions of the Gross-Pitaevskii equation with a quasi-periodic potential. The breathers are found to be stable for negative scattering lengths. Localization and stable propagation of breathers are also observed for positive scattering lengths at relatively weak and moderate nonlinearities. Published by the American Physical Society 2024

Crack initiation and propagation from a closing flaw with frictional contact in a natural quasi-brittle orthotropic solid

This work considers the frictional closing of a pre-existing flaw in a natural quasi-brittle orthotropic medium (wood) under compression. The effect of the flaw face friction at multiple orientations of the flaw, on the peak load-bearing capacity are investigated. The failure due to crack growth initiation from the periphery of the pre-existing flaw and its propagation was investigated experimentally and modelled numerically. The experimental investigation was carried out using New Zealand Pine wood with a pre-existing flaw under a quasi-static compression test. In the numerical investigation, an eXtended Finite Element Method (XFEM)-based cohesive zone model (CZM) is employed in the modelling of the local crack initiation and propagation. The onset of the crack initiation is modelled using a stress-based criterion and the progressive degradation due to crack propagation is traction-separation based damage evolution. The XFEM-based CZM captures the crack initiation and propagation consistent with the experimental results. The numerical results also show the influence of flaw face friction on the peak load bearing capacity as well as the local stress response.

A new nonlocal macro-micro-scale consistent damage model for layered rock mass

Layered rock mass exhibits transverse isotropic characteristics in deformation and strength, and the strength anisotropy is difficult to achieve in traditional local action modes. Therefore, in the framework of nonlocal macro–micro-scale consistent damage (NMMD) theory, this paper constructed a micro damage criterion suitable for layered rock, where both the critical elongation and the brittleness index of material bond are defined as an elliptic function of the bond angle. Next, based on the geometric damage reflecting continuity loss of material point, which is obtained by the weighted average of bond damages around the point, an energy degradation model of a transversely isotropic stiffness matrix is established to achieve localized damage propagation of macroscopic cracks in the layered rock mass. The reliability and advantage of the NMMD model for transversely isotropic material are verified by comparison with the experimental and phase field results through a three-point bending shale beam with different bedding orientations. Finally, the splitting mechanism and bearing capacity of the Brazilian disk with different bedding directions and different inclinations of notch are explored through the established NMMD finite element model. The model is easily implemented, enabling automatic tracing of cracks and accurate prediction of load–displacement curves.

An end-to-end method for predicting compound-protein interactions based on simplified homogeneous graph convolutional network and pre-trained language model

Identification of interactions between chemical compounds and proteins is crucial for various applications, including drug discovery, target identification, network pharmacology, and elucidation of protein functions. Deep neural network-based approaches are becoming increasingly popular in efficiently identifying compound-protein interactions with high-throughput capabilities, narrowing down the scope of candidates for traditional labor-intensive, time-consuming and expensive experimental techniques. In this study, we proposed an end-to-end approach termed SPVec-SGCN-CPI, which utilized simplified graph convolutional network (SGCN) model with low-dimensional and continuous features generated from our previously developed model SPVec and graph topology information to predict compound-protein interactions. The SGCN technique, dividing the local neighborhood aggregation and nonlinearity layer-wise propagation steps, effectively aggregates K-order neighbor information while avoiding neighbor explosion and expediting training. The performance of the SPVec-SGCN-CPI method was assessed across three datasets and compared against four machine learning- and deep learning-based methods, as well as six state-of-the-art methods. Experimental results revealed that SPVec-SGCN-CPI outperformed all these competing methods, particularly excelling in unbalanced data scenarios. By propagating node features and topological information to the feature space, SPVec-SGCN-CPI effectively incorporates interactions between compounds and proteins, enabling the fusion of heterogeneity. Furthermore, our method scored all unlabeled data in ChEMBL, confirming the top five ranked compound-protein interactions through molecular docking and existing evidence. These findings suggest that our model can reliably uncover compound-protein interactions within unlabeled compound-protein pairs, carrying substantial implications for drug re-profiling and discovery. In summary, SPVec-SGCN demonstrates its efficacy in accurately predicting compound-protein interactions, showcasing potential to enhance target identification and streamline drug discovery processes.Scientific contributionsThe methodology presented in this work not only enables the comparatively accurate prediction of compound-protein interactions but also, for the first time, take sample imbalance which is very common in real world and computation efficiency into consideration simultaneously, accelerating the target identification and drug discovery process.

Richtmyer-Meshkov instability when a shock wave encounters with a premixed flame from the burned gas

We present a linear stability analysis of the Richtmyer-Meshkov instability that develops when a shock wave reaches a sinusoidally perturbed premixed flame from behind. In the hydrodynamic regime, when acoustic contributions dominate the flame growth rate, the problem is analytically addressed by the direct integration of the sound wave equations at both sides of the flame, which are bounded by the reflected and transmitted shock waves and the flame front that acts as a contact surface in the hydrodynamic limit. The resolution involves: a hyperbolic change of variables to modify the triangular spatio-temporal domain, a transformation in the Laplace variable, the resolution of the functional equations in the frequency domain, and the final inverse Laplace transform. The latter involves a novel resolution method that is proven beneficial for long-time dynamics. Asymptotic analysis is also carried out to describe the early time and late time hydrodynamic response. The nonuniform flow field resulting from distorted oscillating shocks is characterized by acoustic, rotational, and entropic disturbances, each of which exerts a substantial influence on flame dynamics. These disturbances contribute to the intricate interplay of factors shaping the behavior of the flame in response to the nonuniform flow. In particular, the sensitivity of local flame propagation to temperature disturbances is investigated. This work contributes to a deeper understanding of Richtmyer-Meshkov instability dynamics and offers insights into instability reduction through the modulation of temperature disturbances generated by the transmitted shock.

Verification and characteristics of continuous rotating detonation waves in a novel rounded-rectangle hollow combustor

A noncircular cross-sectioned combustor holds the potential to facilitate the flexible design of continuous rotating detonation (CRD) engines, meeting the demands of airframe/engine integration. In this study, we propose and validate a novel rounded-rectangle hollow combustor comprising arc and straight segments. We conduct experimental and numerical investigations into the detailed propagation and combustion characteristics of the CRD wave in this combustor. The results demonstrate that a single hydrogen–air CRD wave can be achieved and sustained over a wide equivalence ratio (ER) range. As the ER increases from 0.71 to 1.25, the CRD wave's propagation velocity escalates from 1837.1 to 2190.8 m/s, with a substantial decrease in the relative standard deviation from 9.76% to 0.85%. The curvature variation along the combustor circumference influences the local propagation velocity, combustion intensity, shock wave strength, and shock wave system. Specifically, the constraints imposed by the arc segment significantly enhance combustion intensity and amplify shock wave strength. This is evident through a higher heat release rate fraction (fHRR), increased pressure peaks, and elevated propagation velocity. At the straight segment, the shock wave system comprises the CRD wave, leading, and trailing diffracted shock waves. However, the constraint from the arc segment leads to the formation of an additional reflected shock wave, altering the shock wave system to include the CRD wave, the reflected shock wave, and the leading diffracted shock wave. Thus, this study advances our understanding of CRD characteristics in noncircular cross-sectioned combustors, providing insight to the design theory for CRD engine combustors.