Stationary Domains Research Articles

GCReID: Generalized continual person re-identification via meta learning and knowledge accumulation

Person re-identification (ReID) has made good progress in stationary domains. The ReID model must be retrained to adapt to new scenarios (domains) as they emerge unexpectedly, which leads to catastrophic forgetting. Continual learning trains the model in the order of domain emergence to alleviate catastrophic forgetting. However, generalization ability of the model is still limited due to the distribution difference between training and testing domains. To address the above problem, we propose the generalized continual person re-Identification (GCReID) model to continuously train an anti-forgetting and generalizable model. We endeavor to increase the diversity of samples by prior to simulate unseen domains. Meta-train and meta-test are adopted to enhance generalization of the model. Universal knowledge extracted from all seen domains and the simulated domains is stored in a set of feature embeddings. The knowledge is continually updated and applied to guide meta-train and meta-test via a graph attention network. Extensive experiments on 12 benchmark datasets and comparisons with 6 representative models demonstrate the effectiveness of the proposed model GCReID in enhancing generalization performance on unseen domains and alleviating catastrophic forgetting of seen domains. The code will be available at https://github.com/DFLAG-NEU/GCReID if our work is accepted.

Effect of Axial Clearance Variation on Dual-Rotor Flowmeter Performance.

This study examines the impact of axial clearance variations on the performance characteristics of a dual-rotor flowmeter (DRT-FM) through numerical simulations, with the validity of the numerical results verified by calibration experiments. The results indicate that within the range of 200 L/h to 1600 L/h, the K factors of different groups increase as clearance increases. The K factor of the 0.80 mm group is the largest, showing an average increase of around 6% compared to that of the 0.50 mm group. Additionally, Linearity E also decreased, with a minimum of 1.07% in the 0.65 mm group, significantly lower than the 3.33% in the 0.50 mm group. Furthermore, the pressure loss increased slightly, with the 0.65 mm group having the largest pressure loss; however, at a flow rate of 1600 L/h, the pressure loss only increases by 0.186 kPa compared to that of the 0.50 mm group. Flow field analysis reveals that changes in axial clearance predominantly affect pressure distribution. Larger clearances reduce low-pressure regions on upstream and downstream transition surfaces, thereby reducing energy loss due to pressure changes. Entropy analysis further demonstrates that higher axial clearance decreases energy loss in the upstream and downstream stationary domains, optimizing the DRT-FM's energy characteristics.

Effects of geometrical parameters on the performance of hydrogen regenerative pumps in proton exchange membrane fuel cell systems

Regenerative pumps are considered a promising option for hydrogen recirculation in proton exchange membrane fuel cell (PEMFC) systems. The geometry of the impeller exerts a direct influence on the pressurization process within the regenerative pumps. However, the influence has not been explored systematically in any available publications. This study investigates the effects of geometric parameters on pump performance and provides design guidelines for hydrogen regenerative pumps. A three-dimensional (3D) numerical model is developed and validated. The effects of critical geometric parameters, including the number of blades, impeller radius, and impeller radius ratio, on the performance of the regenerative pumps are then analyzed. Based on the analysis, this work recommends 35 blades and an impeller radius ratio of 0.65. To establish a correlation between operational performance and micro-flow mechanisms, the fluid velocity distribution at the interface of the rotating and stationary domains is quantitatively analyzed. Results indicate that a reduction in the flow rate results in increased axial velocity and decreased tangential velocity. The increased axial velocity enhances mass transfer, thus increasing the pressure and diminishing the efficiency, whereas the decreased tangential velocity intensifies the impact of the fluid on the blades, leading to increased shock losses and reduced efficiency.

Enhancing Evolving Domain Generalization through Dynamic Latent Representations

Domain generalization is a critical challenge for machine learning systems. Prior domain generalization methods focus on extracting domain-invariant features across several stationary domains to enable generalization to new domains. However, in non-stationary tasks where new domains evolve in an underlying continuous structure, such as time, merely extracting the invariant features is insufficient for generalization to the evolving new domains. Nevertheless, it is non-trivial to learn both evolving and invariant features within a single model due to their conflicts. To bridge this gap, we build causal models to characterize the distribution shifts concerning the two patterns, and propose to learn both dynamic and invariant features via a new framework called Mutual Information-Based Sequential Autoencoders (MISTS). MISTS adopts information theoretic constraints onto sequential autoencoders to disentangle the dynamic and invariant features, and leverage an adaptive classifier to make predictions based on both evolving and invariant information. Our experimental results on both synthetic and real-world datasets demonstrate that MISTS succeeds in capturing both evolving and invariant information, and present promising results in evolving domain generalization tasks.

Review on New-Generation Batteries Technologies: Trends and Future Directions

Battery technologies have recently undergone significant advancements in design and manufacturing to meet the performance requirements of a wide range of applications, including electromobility and stationary domains. For e-mobility, batteries are essential components in various types of electric vehicles (EVs), including battery electric vehicles (BEVs), plug-in hybrid electric vehicles (PHEVs), and fuel cell electric vehicles (FCEVs). These EVs rely on diverse charging systems, including conventional charging, fast-charging, and vehicle-to-everything (V2X) systems. In stationary applications, batteries are increasingly being employed for the electrical management of micro/smart grids as transient buffer energy storage. Batteries are commonly used in conjunction with power electronic interfaces to adapt to the specific requirements of various applications. Furthermore, power electronic interfaces to batteries themselves have evolved technologically, resulting in more efficient, thermally efficient, compact, and robust power converter architectures. This article offers a comprehensive review of new-generation battery technologies. The topic is approached from the perspective of applications, emerging trends, and future directions. The article explores new battery technologies utilizing innovative electrode and electrolyte materials, their application domains, and technological limitations. In conclusion, a discussion and analysis are provided, synthesizing the technological evolution of batteries while highlighting new trends, directions, and prospects.

Decoherence and Turbulence Sources in a Long Laser.

We investigate the turn-on process in a laser cavity where the round-trip time is several orders of magnitude greater than the active medium timescales. In this long delay limit, we show that the universal evolution of the photon statistics from thermal to Poissonian distribution involves the emergence of power dropouts. While the largest number of these dropouts vanish after a few round-trips, some of them persist and seed coherent structures similar to dark solitons or Nozaki-Bekki holes described by the complex Ginzburg-Landau equation. These coherent structures connect stationary laser emission domains having different optical frequencies. Moreover, they emit intensity bursts which travel at a different speed, and, depending on the cavity dispersion sign, they may collide with other coherent structures, thus leading to an overall turbulent dynamics. The dynamics is well-modeled by delay differential equations from which we compute the laser coherence time evolution at each round-trip and quantify the decoherence induced by the collisions between coherent structures.

Localized kriging parameters optimization using local uncertainty

ABSTRACT Estimates of natural phenomena with spatial correlation, i.e. stationary domains, are more precise and accurate when performed using geostatistical techniques (e.g. kriging). The kriging estimates require the definition of the spatial continuity model and a search strategy. Many techniques, such as unfolding and dynamic anisotropy, try to give some improvement in the estimates, considering the variations of the distributions in the geological bodies, however, the definition of the search strategy in the other parameters is unique. This study presents an alternative to this, called Localized Kriging Parameters optimization (LKPO). LKPO considers the best local kriging parameters settings (block by block) through the local uncertainly (simulations). To illustrate this methodology, a synthetic dataset is presented, and the results are compared with the methodologies currently available in the geostatistical literature. Validation checks show a significant improvement in precision and accuracy on the estimates when using local kriging parameters.

A High-Order Compact Gas-Kinetic Scheme in a Rotating Coordinates Frame and on Sliding Mesh

This paper extends the high-order compact gas-kinetic scheme (CGKS) to flow simulations on a rotating coordinate frame. The kinetic equation with the inclusion of centrifugal and Coriolis accelerations is used in the construction of the scheme. With the updates of both cell-averaged flow variables and their gradients in the rotating and stationary domains, a third-order compact reconstruction is developed with sliding interface between them. To capture shock waves and complicated wave interactions, the HWENO-type non-linear reconstruction and gradient compression factors are incorporated in the scheme. For achieving high-order time accuracy, based on the time-accurate flux function the multi-stage multi-derivative time stepping method is implemented in the scheme for the fourth-order time accuracy with two stages. The CGKS is validated in cases from subsonic acoustic wave propagation to high Mach number shock interaction. The compact scheme achieves high-order accuracy and remarkable robustness.

Extended Implicit PIS-ALE Method to Efficient Simulation of Turbulent Flow Domains with Moving Boundaries

In this work, an implicit finite-volume-element (FVE) method is extended to efficiently simulate the vortical structure of unsteady turbulent flows in domains with moving meshes. The arbitrary Lagrangian-Eulerian (ALE) approach is used to consider the motion of a hybrid mesh distributed in the solution domain. Conventional turbulence models are applied to simply confirm the sample achieved efficiency and accuracy in solving complex turbulent flow domains with moving boundaries. In this regard, the advective terms in the Navier-Stokes equations, including those in the transport equations for the applied turbulence models, are treated in a rather innovative manner. In other words, an advanced physical influence scheme (PIS) is suitably introduced in the context of extended ALE formulations. The accuracy and efficiency of the extended method are carefully evaluated by simulating various turbulent flows, including the fluid flow in stationary domains, separated turbulent flow over a bluff body problem, and the dynamic stall of fluid flow over a flapping airfoil. Comparing the current solutions with experimental data, it is shown that the current PIS-ALE method provides better accuracy and efficiency than those of past numerical methods, which used similar turbulence models in their algorithms.

Evaluation of low fidelity and CFD methods for the aerodynamic performance of a small propeller

The increasing use of unmanned aerial vehicles and micro air vehicles creates a strong demand for the accurate aerodynamic performance of small-diameters fixed-pitch propellers. The techniques range from low-fidelity to high-fidelity methods. Among the low-fidelity techniques, there is the blade element momentum theory (BEMT). Computational fluid dynamics (CFD) is categorized as a high-fidelity method. This study aims to employ these methods for the analysis of an APC propeller 14 × 7e. Geometric modeling was performed in a 3D scanner and simulations via BEMT and CFD. Analyses using the BEMT model were conducted with two formulations. The first one considered a linear lift coefficient up to the stall limits, without specifying the post-stall lift coefficients. The second one incorporates three-dimensional flow equilibrium effects, and best represents the post-stall conditions. In the CFD models, two approaches for the compatibility of the rotating and stationary domains were evaluated, the stationary frozen rotor (FR) and the transient arbitrary mesh interface (AMI). The kω SST turbulence model was employed. To investigate the influence of the transition laminar-turbulent boundary layer, the Gamma Theta transitional model was also considered. In the comparison of the results with the experimental tests, it was observed that the appropriate choice of analysis method had a strong relationship with the Reynolds number. For low advanced ratio velocities, the best results were obtained with the BEMT models. BEMT linear model overpredicted the power and efficiency for high advance ratio, while BEMT with tridimensional flow equilibrium presented consistent results with a low computational cost. On the other hand, the CFD-AMI model shows better results for higher advanced ratio velocities but is computationally more expensive. The CFD-FR model maintains a constant pattern of error that, in the analysis of efficiency, did not compromise the results.

Performance of clustering for the decision of stationarity; A case study with a nickel laterite deposit

The decision of stationarity is a fundamental prerequisite to geostatistical estimation and uncertainty characterization of natural resources. Great effort is given to delineate relevant spatial-statistical populations for modeling. Recently developed spatial clustering methodologies claim improvements over traditional methodologies for stationary decision making. However, these novel methods are generally unproven with respect to the geostatistical algorithms that generate the models upon which important and costly decisions are based. In this study, a new method to generate the prior parameter uncertainty for the decision of stationarity is established, and the effects of considering different (and uncertain) decisions of stationarity are explored through a K-Fold cross validation study applied to a nickel laterite dataset with complex multivariate relationships.Results show that, in terms of univariate regression metrics, geostatistical models generated with clustering-based modeling domains are competitive or better than those generated with the traditional merged-lithological domains. However, models generated from the clustering-based stationary domains are superior in terms of multivariate feature reproduction when compared to the merged-lithology domains. We also demonstrate a geostatistical modeling workflow that incorporates uncertainty associated with stationary domaining, with only a minimal extension to established techniques for geostatistical modeling with parameter uncertainty.

Damage zone characterization combining scan-line and scan-area analysis on a km-scale Digital Outcrop Model: The Qala Fault (Gozo)

Fault damage zones can act as a preferential corridor for fluid flow in the subsurface, and for this reason the characterization of their structure, including the attributes of the associated fracture network, is fundamental. In this work, we characterize the damage zone of the Qala fault, a normal fault developed in platform carbonates of the Gozo Island (Maltese Islands). We propose a new workflow that combines scanline and scan-area analysis applied on a high resolution DOM. Linear scanlines allow to characterize fracture spatial distribution, detect stationary area and identify damage zone width. Areal sampling permits to extract the fracture parameters matching the stationary 1D domains. This new approach allows us to: (1) univocally separate the damage zone from the background fractures, (2) identify fracture corridors, (3) collect fracture parameters (length, trend, density, intensity, spacing and topology), (4) identify the REV of the fracture density, intensity and topology and (5) characterize the fracture network connectivity.

Assessing geologic model uncertainty - a case study comparing methods

Abstract Evaluating mineral resources requires the prior delimitation of geologically homogeneous stationary domains. The knowledge about the ore genesis and geological processes involved are translated into three dimensional models, essential for planning the production and decision-making. The mineral industry usually considers grade uncertainty for resource evaluation; however, uncertainty related to the geological boundaries are often neglected. This uncertainty, related to the location of the boundary between distinct geological domains can be one of the major sources of uncertainty in a mineral project, and should be assessed due to its potential impact on the ore tonnage, and consequently, on enterprise profitability. This study aims at presenting three different methodologies capable of generating multiple geomodel realizations and thus, assessing uncertainty. A real dataset with high geological complexity is used to illustrate the methodology. The results are compared to a deterministic model used as a reference scenario.

Effect of cavity depth on thermal performance of a cooled blade tip under rotation

In this paper, the effect of cavity geometry of a squealer blade tip on its thermal performance is investigated. Numerical analyses were performed for three depth-to-width ratios for a cooled squealer tip. Also investigated is the effect of cooling introduction into the flow tip and the effect of rotation via the comparison of stationary and rotational domains where the latter was simulated by rotating the blade. The rotational speed and boundary conditions used in the numerical study are chosen to represent the turbine operating environment. The blade tip used had a cooling configuration consisting of seven holes of the same size located on the tip near the blade pressure side. The study presents a discussion on the flow physics within the tip gap and heat transfer on the blade tip. Film-cooling performance of the considered blade tip geometries is demonstrated using distributions of film-cooling effectiveness. It is observed that both the rotation and the addition of cooling at the tip significantly alter the tip gap flow, which is found to have a strong effect on the blade tip heat transfer. The addition of cooling increased the heat transfer coefficient on the blade tip, while an increase in the rim height resulted in higher film effectiveness since the amount of leakage flow entering the cavity floor was reduced.

Sampling Rate Effects on Resting State fMRI Metrics.

Low image sampling rates used in resting state functional magnetic resonance imaging (rs-fMRI) may cause aliasing of the cardiorespiratory pulsations over the very low frequency (VLF) BOLD signal fluctuations which reflects to functional connectivity (FC). In this study, we examine the effect of sampling rate on currently used rs-fMRI FC metrics. Ultra-fast fMRI magnetic resonance encephalography (MREG) data, sampled with TR 0.1 s, was downsampled to different subsampled repetition times (sTR, range 0.3–3 s) for comparisons. Echo planar k-space sampling (TR 2.15 s) and interleaved slice collection schemes were also compared against the 3D single shot trajectory at 2.2 s sTR. The quantified connectivity metrics included stationary spatial, time, and frequency domains, as well as dynamic analyses. Time domain methods included analyses of seed-based functional connectivity, regional homogeneity (ReHo), coefficient of variation, and spatial domain group level probabilistic independent component analysis (ICA). In frequency domain analyses, we examined fractional and amplitude of low frequency fluctuations. Aliasing effects were spatially and spectrally analyzed by comparing VLF (0.01–0.1 Hz), respiratory (0.12–0.35 Hz) and cardiac power (0.9–1.3 Hz) FFT maps at different sTRs. Quasi-periodic pattern (QPP) of VLF events were analyzed for effects on dynamic FC methods. The results in conventional time and spatial domain analyses remained virtually unchanged by the different sampling rates. In frequency domain, the aliasing occurred mainly in higher sTR (1–2 s) where cardiac power aliases over respiratory power. The VLF power maps suffered minimally from increasing sTRs. Interleaved data reconstruction induced lower ReHo compared to 3D sampling (p < 0.001). Gradient recalled echo-planar imaging (EPI BOLD) data produced both better and worse metrics. In QPP analyses, the repeatability of the VLF pulse detection becomes linearly reduced with increasing sTR. In conclusion, the conventional resting state metrics (e.g., FC, ICA) were not markedly affected by different TRs (0.1–3 s). However, cardiorespiratory signals showed strongest aliasing in central brain regions in sTR 1–2 s. Pulsatile QPP and other dynamic analyses benefit linearly from short TR scanning.

Multivariate Categorical Modeling with Hierarchical Truncated Pluri-Gaussian Simulation

Multiple categorical variables such as mineralization zones, alteration zones, and lithology are often available for geostatistical modeling. Each categorical variable has a number of possible categorical outcomes. The current approach for numerical modeling of categorical variables is to either combine the categorical variables or to model them independently. The collapse of multiple categorical variables into a single variable with all combinations is impractical due to the large number of combinations. In some cases, lumping categorical variables is justified in terms of stationary domains; however, this decision is often due to the limitations of existing techniques. The independent modeling of each categorical variable will fail to reproduce the collocated joint categorical relationships. A methodology for the multivariate modeling of categorical variables utilizing the hierarchical truncated pluri-Gaussian approach is developed and illustrated with the Swiss Jura data set. The multivariate approach allows for improved reproduction of multivariate relationships between categorical variables.

Prediction of entry compression waves induced by a high-speed train entering tunnel

Tunnel wave produced by an entering high-speed train has always been one of the most prominent problems in the field of aeroacoustics. These waves are formed at the entrance of the tunnel and propagate ahead of the train at the local speed of sound. The formation process of the first of the successive compression wave at the entrance decides the nature and intensity of the subsequent waves. Thus, to visualize and understand the formation process; a numerical study has been conducted using commercial computational fluid dynamics (CFD) solver FLUENT 17.1. The dynamic mesh update techniques have been implemented to update the stationary domains as the train moves inside it. Pressure inside the tunnel is measured at distinct points to study the changes in the intensity of the compression wave. The flow physics has been comprehended using pressure plots and contours.

Bayesian Parameter Identification for Turing Systems on Stationary and Evolving Domains

In this study, we apply the Bayesian paradigm for parameter identification to a well-studied semi-linear reaction–diffusion system with activator-depleted reaction kinetics, posed on stationary as well as evolving domains. We provide a mathematically rigorous framework to study the inverse problem of finding the parameters of a reaction–diffusion system given a final spatial pattern. On the stationary domain the parameters are finite-dimensional, but on the evolving domain we consider the problem of identifying the evolution of the domain, i.e. a time-dependent function. Whilst others have considered these inverse problems using optimisation techniques, the Bayesian approach provides a rigorous mathematical framework for incorporating the prior knowledge on uncertainty in the observation and in the parameters themselves, resulting in an approximation of the full probability distribution for the parameters, given the data. Furthermore, using previously established results, we can prove well-posedness results for the inverse problem, using the well-posedness of the forward problem. Although the numerical approximation of the full probability is computationally expensive, parallelised algorithms make the problem solvable using high-performance computing.

Domain-growth-induced patterning for reaction-diffusion systems with linear cross-diffusion

In this article we present, for the first time, domain-growth induced pattern formation for reaction-diffusion systems with linear cross-diffusion on evolving domains and surfaces. Our major contribution is that by selecting parameter values from spaces induced by domain and surface evolution, patterns emerge only when domain growth is present. Such patterns do not exist in the absence of domain and surface evolution. In order to compute these domain-induced parameter spaces, linear stability theory is employed to establish the necessary conditions for domain-growth induced cross-diffusion-driven instability for reaction-diffusion systems with linear cross-diffusion. Model reaction-kinetic parameter values are then identified from parameter spaces induced by domain-growth only; these exist outside the classical standard Turing space on stationary domains and surfaces. To exhibit these patterns we employ the finite element method for solving reaction-diffusion systems with cross-diffusion on continuously evolving domains and surfaces.

Analysis and observation of moving domain fronts in a ring of coupled electronic self-oscillators.

In this work, we consider a ring of coupled electronic (Wien-bridge) oscillators from a perspective combining modeling, simulation, and experimental observation. Following up on earlier work characterizing the pairwise interaction of Wien-bridge oscillators by Kuramoto-Sakaguchi phase dynamics, we develop a lattice model for a chain thereof, featuring an exponentially decaying spatial kernel. We find that for certain values of the Sakaguchi parameter α, states of traveling phase-domain fronts involving the coexistence of two clearly separated regions of distinct dynamical behavior, can establish themselves in the ring lattice. Experiments and simulations show that stationary coexistence domains of synchronization only manifest themselves with the introduction of a local impurity; here an incoherent cluster of oscillators can arise reminiscent of the chimera states in a range of systems with homogeneous oscillators and suitable nonlocal interactions between them.