Local Loss Research Articles

Attitudes Towards Dementia Among a Diverse Group of Refugees Resettled in the United States.

Forced migration results in exposure to trauma, interrupted access to healthcare, and loss of social support and may increase dementia risk. Literature on refugees' knowledge of dementia and its risk factors is scant. This study investigates refugee perspectives on dementia and their access to cognitive healthcare in the United States (US). We conducted 6 focus groups and 30 individual in-depth interviews (total of 69 participants) with Arab, African, and Afghan refugees resettled in San Diego, California. Data was coded using inductive thematic analysis. Organized by the socioecological model of health, the following themes emerged: (1) mental trauma due to migration was linked to dementia (individual); (2) fear of dementia and burdening caregivers due to limited support systems (interpersonal); (3) reliance on virtual communities for dementia information and the stress of local community loss increasing dementia risk (community); (4) healthcare providers, both in the US and in refugee camps, didn't address cognitive health concerns (institutions); and (5) discriminatory immigration and healthcare policies as barriers to healthy aging (policy). Despite being a heterogeneous group, refugees share specific experiences, knowledge gaps, and barriers to healthy aging. Tailored interventions and policies are needed to address this population's cognitive health needs. This includes addressing their mental health and social support concerns as well as training clinicians to screen for/discuss dementia with aging refugee patients.

Are agricultural commodity production systems at risk from local biodiversity loss?

Compelling evidence for feedbacks between commodity crop production systems and local ecosystems has led to predictions that biodiversity loss could threaten food security. However, for this to happen agricultural production systems must both impact and depend on the same components of biodiversity. Here, we review the evidence for and against the simultaneous impacts and dependencies of eight important commodity crops on biodiversity. We evaluate the risk that pollination, pest control or biodiversity-mediated soil health maintenance services are at risk from local biodiversity loss. We find that for key species groups such as ants, bees and birds, the production of commodities including coffee, cocoa and soya bean is indeed likely to be at risk from local biodiversity loss. However, we also identify several combinations of commodity, ecosystem service and component of biodiversity that are unlikely to lead to reinforcing feedbacks and lose-lose outcomes for biodiversity and agriculture. Furthermore, there are significant gaps in the evidence both for and against a mutualism between biodiversity and agricultural commodity production, highlighting the need for more evaluation of the importance of specific biodiversity groups to agricultural systems globally.

Autoimmune astrocytopathy double negative for AQP4-IgG and GFAP-IgG: Retrospective research of clinical practice, biomarkers, and pathology.

The objective of this study is to investigate the presence of astrocyte antibodies in patients, excluding aquaporin-4 or glial fibrillary acidic protein (GFAP) antibodies, while evaluating associated biomarkers and pathologies. Patient serum and cerebrospinal fluid (CSF) were tested for antibodies using tissue- and cell-based assays. Neurofilament light chain (NFL) and GFAP in the CSF were detected using single-molecule array (SIMOA). 116 patients accepted SIMOA. Fifteen functional neurological disorders patients without antibodies were designated as controls. Thirty-five patients were positive for astrocyte antibodies (Anti-GFAP: 7; Anti-AQP4: 7; unknown antibodies: 21, designed as the double-negative group, DNAP). The most frequent phenotype of DNAP was encephalitis (42.9%), followed by myelitis (23.8%), movement disorders (19.0%), and amyotrophic lateral sclerosis-like (ALS-like) disease (14.2%). The levels of CSF GFAP and NFL in DNAP were higher than in the control (GFAP: 1967.29 [776.60-13214.47] vs 475.38 [16.80-943.60] pg/mL, p < 0.001; NFL: 549.11 [162.08-2462.61] vs 214.18 [81.60-349.60] pg/mL, p = 0.002). GFAP levels decreased in DNAP (n = 5) after immunotherapy (2446.75 [1583.45-6277.33] vs 1380.46 [272.16-2005.80] pg/mL, p = 0.043), while there was no difference in NFL levels (2273.78 [162.08-2462.61] vs 890.42 [645.06-3168.06] pg/mL, p = 0.893). Two brain biopsy patterns were observed: one exhibited prominent tissue proliferation and hypertrophic astrocytes, with local loss of astrocytes, while the other showed severe astrocyte depletion with loss of neurofilaments around the vessels. Eighteen patients received immunotherapy, and improved except one with ALS-like symptoms. We identified anti-vimentin in this patient. There are unidentified astrocyte antibodies. The manifestations of double-negativity are heterogeneous; nevertheless, the pathology and biomarkers remain consistent with astrocytopathy. Immunotherapy is effective.

Microstructural Analysis of Deposited Alkali Metal in “Anode-Free” Solid-State Batteries

Lithium metal as an anode material can greatly boost energy and power density in solid-state batteries.[1] However, challenges like local contact loss and dendrite growth limit the applicable current density. It was shown previously that the lithium microstructure can greatly influence the electrochemical performance of lithium, e.g. the stripping capacity until pore formation occurs.[2]Solid-state-batteries (SSBs) manufactured without the use of an alkali metal reservoir are under intense investigation, as they may enable unmatched energy densities without the need to handle reactive and inherently unsafe alkali metal foils. In these reservoir-free cells (RFCs) the alkali metal is usually stored in the discharged cathode active material upon assembly and thus only deposited at the anode current collector within the first charging step. However, despite its expected influence on the electrochemical performance, the microstructure of freshly deposited lithium has yet to be analyzed. Up to now, the lack of a combination of suitable preparation and analyses tools restrict any study of the microstructure of thin alkali metal films.Within this study, we designed a workflow to reliably analyze the microstructure of thermally processed lithium and sodium films. Using a combination of cryogenic focused ion beam (FIB) preparation and electron back-scatter diffraction (EBSD), we were able to exclude any recrystallization during room temperature storage or FIB preparation. Therefore, this newly developed workflow was further used to study the microstructure of otherwise buried alkali metal films, which were deposited in so called “anode-free” or reservoir free cell designs.Our analyses show that the grain size of plated alkali metal films is independent on the microstructure of used current collector and generally consists of very large grains, albeit not as large as as-bought foil. Even more interesting, every single deposited film consists solely of columnar grains with grain boundaries oriented perpendicular to the solid electrolyte interface, marking a clear difference to using an alkali metal foil as the anode. Our results are confirmed at two different RFCs, namely Cu|LLZO|Li and Steel|LPSCl|Li. Generally, we believe our research on metal anodes will open up the possibility of tackling more advanced scientific questions previously impossible to answer.(1) Krauskopf, T., Richter, F. H., Zeier, W. G. & Janek, J. Physicochemical Concepts of the Lithium Metal Anode in Solid-State Batteries. Chem. Rev. 120, 7745–7794 (2020).(2) Singh, D. K., Fuchs, T. et al. Origin of the lithium metal anode instability in solid-state batteries during discharge. Matter 6, 1463–1483 (2023).

Study on transient cavitation performance of centrifugal pump based on the influence of rough impeller

This study employs numerical simulation to investigate the transient flow and cavitation performance of centrifugal pumps with rough impellers, validating the numerical method with experimental data. Initially, the effect of blade roughness on the external characteristics of centrifugal pumps is examined. Subsequently, the study specifically addresses the impact of roughness on internal flow characteristics during cavitation, including vapor volume distribution, three-dimensional vortex structures, and vorticity distribution in the impeller channel. Furthermore, the influence of blade roughness on local energy loss is analyzed using entropy production theory. Finally, several monitoring points are arranged in the impeller channel to assess pressure pulsation effects. The results show that blade roughness generally reduces the head and efficiency of centrifugal pumps. During the non-cavitation and cavitation incipient stages, roughness marginally increases the head, with a maximum increase in only 0.1%. Impeller roughness causes vacuole collapse and vortex structure enlargement, disrupting the stable flow path within the channel. Blade roughness also escalates energy loss within impeller components, particularly under full cavitation conditions, where the impeller's entropy production accounts for up to 50%. Pressure pulsation results reveal that while blade roughness can slightly suppress cavitation, it also disturbs the flow field pressure. These insights provide guidance and data support for mitigating roughness and cavitation, the two primary instability factors in centrifugal pump operations.

Flow in fracture Intersections: Deflection flow assumption reexamined

Deflection flow, which represents the process of flow redistribution at fracture intersections, is essential for accurately predicting fluid infiltration and contaminant transport in fracture networks. However, previous studies focused on the outcome of flow redistribution but neglected the process, leading to poor knowledge of deflection flow at intersections. In this study, numerical investigations of the deflection flow in cross fractures were carried out based on Navier-Stokes equations and the results were compared with previous experiments. Geometrical characteristics such as angle(φ), aperture ratio(k), and hydraulic characteristics like inlet flow ratio(n) were considered to explore the mechanism of deflection flow. The deflection flow intensity(Xφ) was introduced to characterize the magnitude and direction of the deflection flow. The results showed that the deflection flow is caused by unequal local head loss. The effect of φ on Xφ is limited by n and k. Increasing n or k enhances the deflection flow but the directions of the induced deflection flow are reversed. The obtained relationship was further supported by theoretical results derived from the cubic law. Subsequently, sensitivity analysis of impact factors to the deflection flow was carried out by the Morris method, the sensitivity in descending order was k, n, φ. Finally, based on the analysis of Xφ, the conditions to limit the occurrence of deflection flow was obtained as a linear positive correlation between the critical inlet flow ratio(nc) and aperture ratio(k). The results from this study may provide insights and physical constraints on construction of flow models based on discrete fracture networks.

TST_MFL: Two-stage training based metric fusion learning for few-shot image classification

Addressing the limitations of most few-shot learning (FSL) methods, particularly their insufficient single-feature discriminability and generalization during pre-training encoding, this paper introduces a novel approach: Two-Stage Training based Metric Fusion Learning (TST_MFL) for few-shot image classification. TST_MFL innovatively integrates a two-stage fusion metric learning strategy, enhancing both pre-training and meta-training phases to overcome the prevalent focus on single-level feature extraction in current few-shot image classification methods. In pre-training stage, we employ prediction distillation loss and local difference loss to facilitate mutual learning between global and local features, thus obtaining more discriminative features. Moreover, meta-training further enhances feature generalization and integrates global-local metrics to improve few-shot classification accuracy. Our approach demonstrates competitive performance across three fine-grained datasets and one coarse-grained dataset. Notably, experimental results in 1-shot and 5-shot settings of fine-grained datasets underscore the superiority of TST_MFL. The code for this work is available at: https://github.com/ZitZhengWang/TST_MFL.

Connectivity and climate influence diversity–stability relationships across spatial scales in European butterfly metacommunities

AbstractAimAnthropogenic‐driven biodiversity loss can impact ecosystem stability. However, most studies have only evaluated the diversity–stability relationship at the local scale and we do not fully understand which factors stabilize animal populations and communities across scales. Here, we investigate the role of species dispersal ability, climate, spatial distance and different facets of biodiversity on the stability of butterfly populations and communities across multiple spatial scales.LocationPrimarily Western Europe.Time Period2005–2016.Major Taxa StudiedButterflies (Rhopalocera) of Europe.MethodsWe assembled a continent‐wide database of European butterflies' abundance and used Structural Equation Modelling to evaluate the direct and indirect effects of multiple stabilizing mechanisms. In parallel, we tested the effect of dispersal ability on the stability at multiple spatial scales, using a butterfly mobility index as an indicator of dispersal capacity.ResultsRegional stability strongly reflected local stability, which in turn was driven by both taxonomic and functional α‐diversity. Spatial asynchrony was also important for regional stability and it was driven by both functional β‐diversity and metapopulation asynchrony, which in turn increased with spatial distance among communities. We observed a positive effect of temperature on functional α‐diversity and on local stability, whereas precipitation negatively influenced local diversity. Finally, spatial asynchrony contributed more to the regional stability of less mobile species compared to highly mobile ones, indicating that both extrinsic and intrinsic determinants of connectivity impact regional stability indirectly.Main ConclusionsOur results demonstrate the importance of local and regional processes for regional stability. However, the relative contribution of spatial asynchrony and metapopulation asynchrony increases with connectivity loss, especially for less mobile species, indicating that landscape management should be tailored depending on the dispersal capacity of organisms. Both local biodiversity loss and regional biotic homogenization destabilize metacommunities, with potential implications for the reliable provision of ecosystem functions.

Cultural relics fragment assembly based on fracture surface contour features

Addressing the issue of registration deviations caused by local fragment loss and shape diversity during the process of cultural relics fragment assembly, this paper proposes a method for assembling cultural relics fragments based on the contour features of fracture surfaces. Initially, the curvature of discrete point clouds is calculated using the quartic surface method and the Weingarten matrix, and curvature discontinuity points are identified using a proportional method as a candidate feature point set. Subsequently, the angle difference between a point and its neighboring points’ directed segments is defined as a neighborhood feature parameter, which further optimizes the feature point set and selects the optimal matching set. Then the 4-points congruent sets (4PCS) algorithm is utilized for the initial registration of the fragments, and the unit quaternion method is employed to calculate the rigid body transformation matrix during the iteration process. Finally, corresponding point pairs are determined by the intersection of a point’s normal with a set of points, improving the iterative closest point (ICP) algorithm to achieve precise assembly of the fragments. The experimental results show that, compared to other assembly algorithms, the proposed method is more accurate in feature point extraction, robust, and effectively enhances the accuracy and efficiency of fragment assembly.

Research on the influence of system drainage holes on the head loss of water transmission systems

Abstract To ensure the structural safety of the thin lining structure of the tailwater tunnel, this study used three-dimensional flow field numerical simulation technology to conduct an in-depth analysis of the drainage holes set along the tunnel body, to evaluate the impact of drainage holes on the overall and local head loss of the tunnel under different system layout types. It aims to provide scientific theoretical support and practical reference for engineering design and achieve more efficient and stable tunnel hydraulic operation conditions.

Exploiting Optical Flow Guidance for Transformer-Based Video Inpainting.

Transformers have been widely used for video processing owing to the multi-head self attention (MHSA) mechanism. However, the MHSA mechanism encounters an intrinsic difficulty for video inpainting, since the features associated with the corrupted regions are degraded and incur inaccurate self attention. This problem, termed query degradation, may be mitigated by first completing optical flows and then using the flows to guide the self attention, which was verified in our previous work - flow-guided transformer (FGT). We further exploit the flow guidance and propose FGT++ to pursue more effective and efficient video inpainting. First, we design a lightweight flow completion network by using local aggregation and edge loss. Second, to address the query degradation, we propose a flow guidance feature integration module, which uses the motion discrepancy to enhance the features, together with a flow-guided feature propagation module that warps the features according to the flows. Third, we decouple the transformer along the temporal and spatial dimensions, where flows are used to select the tokens through a temporally deformable MHSA mechanism, and global tokens are combined with the inner-window local tokens through a dual-perspective MHSA mechanism. FGT++ is experimentally evaluated to be outperforming the existing video inpainting networks qualitatively and quantitatively.

Influence of geometric parameters on aerodynamic characteristics of serpentine convergent-divergent nozzle

The serpentine convergent-divergent nozzle represents an optimal configuration for next-generation fighter aircraft characterized by low detectability and high thrust-to-weight ratio. In contrast to the serpentine convergent nozzle, such configuration offers increased design flexibility with additional parameters, leading to heightened interactions among these parameters. As such, it is crucial to reveal the influence of design parameters on the aerodynamic performance of the serpentine convergent-divergent nozzle and the multifactor interaction, as well as its mechanism. Therefore, the influence, interaction and sensitivity of parameters on the aerodynamic performance of the nozzle were numerically investigated using the orthogonal test method. Additionally, the influence mechanism of the convergence angle, throat aspect ratio, and axial length to inlet diameter on the flow characteristics of the nozzle was investigated in detail. The results show that the convergence angle is identified as the main factor affecting the aerodynamic parameters of the nozzle. As the convergence angle increases, the thrust coefficient, total pressure recovery coefficient and discharge coefficient gradually decrease. The interaction between throat aspect ratio and other parameters is obvious. Different design parameters affect the local loss and the friction loss by affecting the curvature and wetted perimeter area, resulting in different aerodynamic characteristics of serpentine convergent-divergent nozzle.

FedLAP-DP: Federated Learning by Sharing Differentially Private Loss Approximations

Conventional gradient-sharing approaches for federated learning (FL), such as FedAvg, rely on aggregation of local models and often face performance degradation under differential privacy (DP) mechanisms or data heterogeneity, which can be attributed to the inconsistency between the local and global objectives. To address this issue, we propose FedLAP-DP, a novel privacy-preserving approach for FL. Our formulation involves clients synthesizing a small set of samples that approximate local loss landscapes by simulating the gradients of real images within a local region. Acting as loss surrogates, these synthetic samples are aggregated on the server side to uncover the global loss landscape and enable global optimization. Building upon these insights, we offer a new perspective to enforce record-level differential privacy in FL. A formal privacy analysis demonstrates that FedLAP-DP incurs the same privacy costs as typical gradient-sharing schemes while achieving an improved trade-off between privacy and utility. Extensive experiments validate the superiority of our approach across various datasets with highly skewed distributions in both DP and non-DP settings. Beyond the promising performance, our approach presents a faster convergence speed compared to typical gradient-sharing methods and opens up the possibility of trading communication costs for better performance by sending a larger set of synthetic images. The source is available at https://github.com/a514514772/FedLAP-DP.

Beamforming optimization of meta-learning algorithm based on gradient descent in RIS-assisted ISAC system

In this paper, we study an integrated sensing and communication (ISAC) system assisted by reconfigurable intelligent surface (RIS). Our goal is to minimize the transmit power of the BS by jointly optimizing the active beamforming and the passive beamforming under the constraints of the signal-to-interference-plus-noise ratio (SINR) of users, the Cramer–Rao bound (CRB) of sensing and the unit-modulus of RIS reflection elements. The optimization problem is difficult to solve because the objective function is non-convex and the variables are coupled. In addition, a large number of base station (BS) antennas lead to high-dimensional transmit beamforming vectors. We first prove that the optimization variables in the optimization problem are low-dimensional coefficient matrices and passive beamforming, rather than active and passive beamforming, which reduces the complexity compared with the original problem. In addition, when the sensing channel space belongs to the communication channel space, the dedicated sensing signal is not needed. Inspired by the meta-learning technology that can solve non-convex problems, we propose a meta-learning algorithm based on gradient descent (MLAGD) to design beamforming, which tends to learn dynamic optimization strategies and iteratively update optimization variables. Specifically, we first introduce Lagrangian multipliers to transform the constrained optimization problem into the Lagrangian dual domain, and the manifold-based optimization is employed to handle the unit-modulus constraint. The meta-learning neural networks are constructed to update optimization variables through local loss function, and then network parameters are updated through global loss function. The numerical results show the effectiveness of the algorithm.

Exergy analysis of a novel ironmaking process combining coal gasification with the smelting reduction of iron ore

The global iron and steel industry is an energy-intensive industry, and energy-saving in this industry is of great significant. A novel ironmaking process combining coal gasification with the smelting reduction of iron ore (CCGSR) is proposed, and an exergy analysis is applied to evaluate its thermodynamic performance. The results show that the exergy efficiency of the entire system is 88.15 % whereas the energy efficiency is 98.64 %. The exhaust gas of pre-reduction furnace is the main exergy output and has great energy-saving potential. The local exergy loss rates in the pyrolysis, pre-reduction, and smelting reduction furnaces are 55.81 %, 7.24 %, and 36.95 %, respectively. The pyrolysis and smelting reduction furnaces exhibit the largest exergy loss, owing to the mixing of high-temperature gas of the smelting reduction furnace and circulating gas, coke powder gasification, and heat transfer. In addition, after the exhaust gas is utilized to form a cogeneration system of iron and electricity, the exergy efficiency increases from 40.18 % to 50.06 % comparable with 600 MW ultra-supercritical coal-fired generating units. The present work also provides some suggestions for reducing exergy destruction.

Arbitrary stylized light-field generation for three-dimensional light-field displays based on radiance fields

Three-dimensional (3D) light-field display technology can reconstruct the spatial characteristics of 3D scenes and provide users with an immersive visual experience without the need for any additional external devices. Here, an arbitrary stylized light-field generation method for 3D light-field displays is presented, and the tripartite constraints are addressed by conducting style transformation in a refined feature space. A multi-dimensional feature refinement module is designed to learn which aspects and regions should be prioritized within the high-level feature grid of the scene, which allows content and style patterns to be better matched and blended. To preserve more structural details and reduce artifacts, a novel global quality and local structure joint loss function is used for optimization. A mask-guided light-field coding method based on ray-casting is employed to accelerate the synthesis of stylized light-field images. Experimental results demonstrate that the proposed method can generate higher-quality stylized 3D light-field content with any given style in a zero-shot manner. Additionally, the method provides two user control extensions to further enrich the ways for the personalized editing of 3D content displayed on 3D light-field displays.

Global self-sustaining and local inheritance for source-free unsupervised domain adaptation

In this paper, we investigate a practical problem called source-free unsupervised domain adaptation, which adapts a source-trained model to the target domain with unlabeled target data. To address this problem, we propose a novel GlObal self-sustAining and Local inheritance (GOAL) method. GOAL contains three components. (1) A backbone follows a mean teacher scheme. The teacher model serves as a smoothing functionality, facilitating a more consistent convergence of the student model. This capability alleviates the student model’s sensitivity to minor input data variations and enhances the overall robustness of the model. Additionally, disparities in predictions between the student and teacher models can be leveraged to identify potential noise in the data. (2) A Global Consistency Self-Sustaining mechanism for learning a stable, discriminative, and diverse prediction space. On the one hand, we employ neighbor samples and mean-teacher schemes to enhance the discriminability and stability of model predictions. On the other hand, non-neighbor samples are leveraged to augment the diversity of model predictions. Furthermore, to mitigate the impact of potential negative neighbors, we derive a weighting factor by incorporating both neighbor entropy and the top-nd similarity of features. (3) A Local Topology Inheritance mechanism to improve the semantic structure of the feature space. We construct a semantic topology graph based on the output predictions of the teacher model and subsequently transmit the teacher topology to the feature space of the student utilizing a local topology inheritance loss. Combining these three components, GOAL can effectively solve the source-free unsupervised domain adaptation. To the best of our knowledge, GOAL is the first attempt to perform topology inheritance for global consistency domain adaptation. Comprehensive experiments illustrate the effectiveness and superiority of GOAL in addressing source-free unsupervised domain adaptation.

Research on hydraulic characteristics’ sensitivity analysis of the hydraulic inclined shaft turning

Abstract The parameters of the hydraulic inclined shaft turning directly affects the flow field characteristics in the bending section and the construction difficulty of the bending section. Three-dimensional computational fluid dynamics (CFD) numerical simulation method was used to systematically study the hydraulic characteristics under different flow velocities, turning radii and turning angles. The results indicate that the flow velocities and the size of the turning radius have little influence on the flow velocity distribution in the bending section. The larger the turning angle, the larger the fast flow velocity area in the bending section. But the extreme values of fast flow velocity remain basically unchanged, and the length required for adjusting the flow velocity distribution after the curved section becomes longer. The faster the flow velocity, the smaller the turning radius, and the larger the turning angle, the greater the centrifugal force generated and the more pronounced the difference in pressure distribution between the inside and outside of the bend section. The head loss in the bending section is consistent with the empirical formula in the specifications. As the radius of curvature gradually increases from 1D to 3D, the head loss (primarily due to local head loss) decreases gradually. When the radius of the bending section continues to increase, the reduction in local head loss becomes less significant, but the increase in frictional head loss leads to a slight overall increase in head loss.

Evaluation of Different Numerical Approaches to Modeling Flood Flows Over Groynes

AbstractThe hydraulic resistance of groynes is an important factor in the determination of design flood water levels on rivers and the assessment of how much these levels are lowered by modifying the groynes. In standard one‐ or two‐dimensional numerical hydrodynamic models for flood risk management, groynes are commonly represented as subgrid features with a local energy loss according to a weir formula. We tested this representation by using a two‐dimensional horizontal mesh at various groyne submergence degrees by comparing the results with those of flume experiments. We also compared the results with simulations using different 2D and 3D approaches on finer grids that incorporate groynes in the bed topography. In one of the two tested 3D models, complete Reynolds‐averaged Navier‐Stokes equations were solved. The second tested 3D model was constructed simpler by assuming hydrostatic pressure distribution in the vertical direction. We employed Delft3D software in construction and execution of all models. One of the 3D models did predict the hydraulic resistance at low submergence better than the standard model, but it slightly underestimated the resistance at higher submergences. Despite differences in flow characteristics, weirs and groynes were found to produce similar flow resistances for the same height and boundary conditions. Simulations of groyne modifications showed that hydraulic resistance decreased nonlinearly with groyne lowering and streamlining.

Influence of impeller energy dissipation characteristics of multi-blade axial flow pump impeller by the cascade density

Abstract This paper investigates the influence of cascade density on energy dissipation in the impeller of a multi-blade axial-flow pump. Three-dimensional transient numerical simulations were conducted using the SST k-ω turbulence model for impeller schemes with different cascade density, corresponding to different blade chord length and pitch. Entropy production theory was applied to locate the regions with high energy loss in the impeller. The relationships between local entropy generation, energy loss, and unsteady flow were analyzed for different cascade density. The results indicate that impeller entropy output of the multi-blade axial-flow pump is highly consistent with the head loss during testing. Turbulence dissipation accounts for more than 50% of the total energy loss, followed by wall friction, while direct dissipation is the smallest contributor. The relative loss of the impeller can be minimized by decreasing the number of blades or increasing the blade chord length. By comparing energy dissipation in blade rotation with different chord length and pitch, it is found that turbulent dissipation ability can be controlled more effectively by changing chord length, while the ability of direct dissipation can be controlled more effectively by changing the pitch.