Cloud Processing Research Articles

Context-based local-global fusion network for 3D point cloud classification and segmentation

3D point clouds have gained much research attention because of their ability to represent the spatial information of real-world environments in a detailed manner. Despite recent progress in point cloud processing with deep neural networks, most of them either implement sophisticated local feature aggregation methods or imitate 2D convolution operations in the range of K nearest neighbors with limited local context information. These methods may struggle to distinguish between similar geometric shapes within the local region of K nearest neighbors, such as doors and walls. To address this issue, we propose a novel local–global fusion network that captures the diverse local geometric shapes with global structure information. The proposed local–global fusion network comprises two main modules. Firstly, we have developed an effective approach for local context learning using incremental dilated KNN (IDKNN) as the neighbor selecting mechanism to enlarge the receptive field and incorporate more reliable points for local geometric shape learning. Secondly, a three-direction region-wise spatial attention (TRSA) algorithm has been developed to explore the global contextual dependencies. For global context learning, we first split the entire 3D space into regions with equal numbers of points, and, then, intra-region context features are extracted to learn the inter-region relations from three orthogonal directions, taking global structural knowledge into account. By fusing the local context information and global contextual dependencies, we establish a Local-Global Fusion Network, end-to-end framework, called LGFNet. Extensive experimental results on several benchmark datasets clearly demonstrate our approach can achieve state-of-the-art (SOTA) performance on point cloud classification, part segmentation, and indoor semantic segmentation. In addition, TRSA and IKDNN can be easily used in a plug-and-play fashion with various existing SOTA networks to substantially improve their performance. Our code is available at https://github.com/jasonwjw/IDKNN

5G and edge: A reinforcement learning approach for Virtual Network Embedding with cost optimization and improved acceptance rate

5G technologies are fueling a revolution across numerous industries, including manufacturing, healthcare, and entertainment, by enabling the development and deployment of novel applications at the network’s edge. To meet the demanding service level agreements of these industries, a dynamic and adaptable infrastructure strategy that combines cloud and edge computing models is needed. This hybrid approach offers the benefits of both centralized cloud processing and decentralized edge computing, optimized for responsiveness and efficiency. A key element for success is an orchestration mechanism that dynamically allocates resources to ensure the infrastructure can adapt to fluctuating demands in real time, optimizing resource utilization and meeting SLA requirements. Among these mechanisms, virtual network embedding (VNE) and dynamic resource management (DRM) have emerged as tools for defining where and how edge technology should be used. However, current VNE approaches struggle to adapt to real-time fluctuations in demand across geographically distributed edge resources. This work introduces a novel resource allocation algorithm, the VNE-CRS, which uses an Artificial Intelligence technique called Reinforcement Learning to orchestrate resources across multiple domains. This approach benefits from the strength of Reinforcement Learning: its ability to consider the entire problem from beginning to end while incorporating various aspects of 5G Quality of Service Indicators for optimal decision-making. Experiments were conducted to evaluate the performance of VNE-CRS against state-of-the-art algorithms for multi-domain edge environments. Results have shown that employing Reinforcement Learning techniques for VNE resource allocation yields performance gains of 12.32 percentage points in comparison with the GRC algorithm and 28.80 percentage points in comparison with the base edge environment, presenting an acceptability rate closer to the Public Cloud environment with all benefits of edge environment. In conclusion, VNE-CRS offers an efficient solution for resource allocation in 5G environments, achieving superior performance and transforming the VNE architecture into a comprehensive orchestration system that optimizes infrastructure utilization for strategic long-term benefits.

How non-equilibrium aerosol chemistry impacts particle acidity: the GMXe AERosol CHEMistry (GMXe–AERCHEM, v1.0) sub-submodel of MESSy

Abstract. Aqueous-phase chemical processes in clouds, fog, and deliquescent aerosols are known to alter atmospheric composition and acidity significantly. Traditionally, global and regional models predict aerosol composition by relying on thermodynamic equilibrium models and neglect non-equilibrium processes. Here, we present the AERosol CHEMistry (GMXe–AERCHEM, v1.0) sub-submodel developed for the Modular Earth Submodel System (MESSy) as an add-on to the thermodynamic equilibrium model (i.e. ISORROPIA-II) used by MESSy's Global Modal-aerosol eXtension (GMXe) submodel. AERCHEM allows the representation of non-equilibrium aqueous-phase chemistry of varying complexity in deliquescent fine aerosols. We perform a global simulation for the year 2010 by using the available detailed kinetic model for the chemistry of inorganic and small oxygenated organics. We evaluate AERCHEM's performance by comparing the simulated concentrations of sulfate, nitrate, ammonium, and chloride to in situ measurements of three monitoring networks. Overall, AERCHEM reproduces observed concentrations reasonably well. We find that, especially in the USA, the consideration of non-equilibrium chemistry in deliquescent aerosols reduces the model bias for sulfate, nitrate, and ammonium when compared to simulated concentrations by ISORROPIA-II. Over most continental regions, fine-aerosol acidity simulated by AERCHEM is similar to the predictions by ISORROPIA-II, but simulated aerosol acidity tends to be slightly lower in most regions. The consideration of non-equilibrium chemistry in deliquescent aerosols leads to a significantly higher aerosol acidity in the marine boundary layer, which is in line with observations and recent literature. AERCHEM allows an investigation of the global-scale impact of aerosol non-equilibrium chemistry on atmospheric composition. This will aid in the exploration of key multiphase processes and improve the model predictions for oxidation capacity and aerosols in the troposphere.

Study on Dynamic Evolution and Erosion Characteristics of Cavitation Clouds in Submerged Cavitating Water Jets

The dynamic evolution behavior of submerged water jet cavitation clouds was studied by combining experiments and simulation. The formation, development, shedding, and collapsing process of a void cloud was analyzed by high-speed camera technology, and the influence of jet pressure was studied. Cavitation water jet erosion experiments were carried out on AL6061 specimens with standard cylindrical nozzles, and the correlation between cavitation cloud evolution and material erosion was studied by surface analysis. The results showed that the evolution of a cavitation cloud has obvious periodicity, that one period is about 0.8 ms, and its action region can be divided according to the attenuation rate of the jet velocity of the nozzle axis. The attenuation rate of the jet velocity at the nozzle axis in the central jet action zone is less than or equal to 82.5%, in the mixed action zone greater than 82.5% and less than 96%, and in the cavitation action zone greater than or equal to 96%. The erosion damage characteristics in different regions of the mixed action zone are significantly different.

The Application of Laser-Scanning 3D Model Reconstruction Technology for Visualizing a Decommissioning Model of the Heavy Water Research Reactor

Currently, over 100 nuclear power units globally have been in operation for more than 40 years. Hindered by the limitations of computer technology at the time, these nuclear facilities lack detailed electronic drawings. Activities such as equipment replacement and process circuit system modifications during operation result in discrepancies between paper drawings and actual conditions. Given the complexity and irreversibility of nuclear facility decommissioning activities, virtual simulation technology is often employed before the decommissioning process begins to assist in designing and validating decommissioning plans. Consequently, the creation of high-precision 3D models is crucial for subsequent decommissioning designs. Through innovatively utilizing laser-scanning 3D model reconstruction technology in the reconstruction of the model of China’s first heavy water research reactor undergoing decommissioning, this paper provides an overview of the process of laser-scanning 3D model reconstruction and its application in reconstructing the heavy water research reactor model. Using a 3D laser scanner, four decommissioning areas of the heavy water research reactor, including the reactor building, secondary water pump room, ventilation center, and low-level radioactive wastewater storage tank area, were subjected to 3D laser scanning. The acquired point cloud data from 572 scanning stations were processed using point cloud processing software for denoising, stitching, and triangulation. The triangulated model was then imported into modeling software for 3D reconstruction, ultimately establishing a digitalized model of the heavy water research reactor suitable for subsequent decommissioning simulation and design.

Measurement report: Bio-physicochemistry of tropical clouds at Maïdo (Réunion, Indian Ocean): overview of results from the BIO-MAÏDO campaign

Abstract. The BIO-MAÏDO (Bio-physicochemistry of tropical clouds at Maïdo: processes and impacts on secondary organic aerosols formation) campaign was conducted from 13 March to 4 April 2019 on the tropical island of Réunion. The main objective of the project was to improve understanding of cloud impacts on the formation of secondary organic aerosols (SOA) from biogenic volatile organic compound (BVOC) precursors in a tropical environment. Instruments were deployed at five sites: a receptor site, Maïdo Observatory (MO) at 2165 m a.s.l. and four sites along the slope of the Maïdo mountain. Observations include measurements of volatile organic compounds (VOCs) and characterization of the physical, chemical and biological (bacterial diversity and culture-based approaches) properties of aerosols and cloud water. Turbulent parameters of the boundary layer, radiative fluxes and emissions fluxes of BVOCs from the surrounding vegetation were measured to help interpret observed chemical concentrations in the different phases. Dynamical analyses showed two preferred trajectory routes for air masses arriving at MO during the daytime. Both trajectories correspond to return branches of the trade winds associated with upslope thermal breezes, where air masses likely encountered cloud processing. The highest mixing ratios of oxygenated VOCs (OVOCs) were measured above the site located in the endemic forest and the highest contribution of OVOCs to total VOCs at MO. Chemical compositions of particles during daytime showed higher concentrations of oxalic acid, a tracer of cloud processing and photochemical aging, and a more oxidized organic aerosol at MO than at other sites. Approximately 20 % of the dissolved organic compounds were analyzed. Additional analyses by ultra-high-resolution mass spectrometry will explore the complexity of the missing cloud organic matter.

Semantic segmentation of large-scale point clouds by integrating attention mechanisms and transformer models

In the current field of point cloud processing, semantic segmentation of large-scale point clouds remains a challenging problem. Traditional methods often underperform when faced with the complexity and density variations present in large-scale point cloud data. This study introduces an innovative model designed for semantic segmentation of large-scale point clouds. The model integrates a CNN-Transformer-based Context Aggregation Module with a Slot Attention mechanism to enhance the understanding of entity relationships and improve semantic segmentation performance. Furthermore, by flexibly adjusting the weight parameters within the loss function, we have optimized the training process, increasing its flexibility. In our experiments, we compared the proposed model with 55 classical and contemporary point cloud semantic segmentation models, evaluating semantic segmentation performance, computational resource consumption, inference time, and the number of model parameters. The results show a significant improvement in performance on the S3DIS and Semantic3D datasets, specifically achieving an mIoU of 71.53% and an OA of 93.92%. Ablation studies were conducted to further ascertain the contribution of each module to the overall performance of the model. This study presents an innovative approach to overcoming the challenges of large-scale point cloud semantic segmentation, offering significant contributions to the advancement of point cloud processing.

Finite Difference Method for Solving Parabolic-Type Integro-Differential Equations in Multidimensional Domain with Nonhomogeneous First-Order Boundary Conditions

Introduction. We investigate a multidimensional (in terms of spatial variables) parabolic-type integro-differential equation with nonhomogeneous first-order boundary conditions. The locally one-dimensional finite difference scheme developed herein can be applied to solve various applied problems leading to multidimensional parabolic-type integro-differential equations. Examples include mathematical modelling of cloud processes, addressing the issue of active intervention in convective clouds to prevent hail and artificially enhance precipitation, as well as describing the droplet mass distribution function due to microphysical processes such as condensation, coagulation, fragmentation, and freezing of droplets in convective clouds.Materials and Methods. In this study, an approximate solution to the initial-boundary value problem is constructed using the locally one-dimensional scheme of A.A. Samarsky with a specified order of approximation О(h2 +τ). The primary research method employed is the method of energy inequalities.Results. A priori estimates have been obtained in the discrete interpretation, from which uniqueness, stability, and convergence of the solution of the locally one-dimensional difference scheme to the solution of the original differential problem follow, with a convergence rate equal to the order of approximation of the difference scheme.Discussion and Conclusions. The research findings can be utilized for further development of boundary value problem theory for parabolic equations with variable coefficients. Additionally, they may find applications in the fields of difference scheme theory, computational mathematics, and numerical modelling.

A New Comprehensive Cloud Macrophysics Scheme With a Prognostic Dual‐Triangular PDF

AbstractTo improve cloud simulation in a general circulation model, we develop a new cloud macrophysics scheme that treats detrained cumulus generated by convective detrainment process separately from pure stratus; prognoses two symmetric triangular probability density functions of total specific humidity for each detrained cumulus and pure stratus; and diagnoses both cloud fraction and cloud condensate in all liquid, ice, and mixed‐phases in a consistent and unified way without any adjustment to remove empty or very dense cloud. Supersaturation is allowed within ice cloud. The new scheme (“NEW”) is compared with the previous model (“OLD”) using single‐column simulations for subtropical marine stratocumulus (DYCOMS2) and continental deep convection (ARM97) cases. In DYCOMS2, both NEW and OLD produce vertical profiles of grid‐mean cloud condensate similar to large‐eddy simulation. In ARM97, compared with OLD, NEW simulates less sporadic vertical profiles of in‐cloud condensates, due to consistent diagnosis of cloud fraction and cloud condensate; more continuously‐varying detrained cumulus with time, due to prognostic treatment of detrained cumulus; and ice cloud fraction and ice condensate similar to those of OLD, in spite of completely different treatment of ice cloud processes. The global performance of NEW is similar to OLD with improved relative humidity. Compared to OLD, NEW simulates more and improved cloud condensate, but less and degraded cloud fraction, particularly, in the lower troposphere. Detrained cumulus is moister and colder and has a larger moisture variance than pure stratus. Overall, NEW simulates stronger condensation‐deposition rates than OLD, due in part to the separate treatment of detrained cumulus and pure stratus.

Looking Back: An Account of How Ice Nucleation by Bacteria Was Discovered (1963 to about Mid-1980s). Part I: The Basics

Abstract An overview is given of the path of research that led from asking how hailstones originate to the discovery that ice nucleation can be initiated by bacteria and other microorganisms at temperatures as high as −2°C. The major steps along that path were finding exceptionally effective ice nucleators in soils with a high content of decayed vegetative matter, then in decaying tree leaves, and then in plankton-laden ocean water. Eventually, it was shown that Pseudomonas syringae bacteria were responsible for most of the observed activity. That identification coincided with the demonstration that the same bacteria cause frost damage on plants. Ice nucleation by bacteria meant an unexpected turn in the understanding of ice nucleation and of ice formation in the atmosphere. Subsequent research confirmed the unique effectiveness of ice nucleating particles (INP) of biological origin, referred to as bio-INPs, so that bio-INPs are now considered to be important elements of lower-tropospheric cloud processes. Nonetheless, some of the questions which originally motivated the research are still unresolved, so that revisiting the early work may be helpful to current endeavors. Part I of this manuscript summarizes how the discovery progressed. Part II (Schnell and Vali) shows the relationship between bio-INPs in soils and in precipitation with climate and other findings. The online supplemental material contains a bibliography of recent work about bio-INPs.

Simulation of the vulnerability of the steppe landscape and climate zone of Ukraine to climate changes based on space image data

Now, the whole world, including Ukraine, is facing one of the biggest environmental problems, namely, climate change. The steppe landscape-climate zone is one of the territories considered the most sensitive to Ukraine's current and future climate change threats. Studying the vulnerability to climate change of the steppe zone of Ukraine based on data from space surveys requires analysing a large amount of objective data, namely the products of remote sensing data processing. The article presents the results of combining remote sensing, geographic information systems, and multi-criteria decision analysis to identify vulnerable areas to the impact of climate change in the steppe landscape-climatic zone. This information will be used to recommend adaptation systems to modern conditions and reduce the impact of adverse climate changes. The proposed decision-making structure was developed in three stages: 1) collection and processing of available data from space surveys; 2) development of a model of vulnerability to climate change of the steppe landscape-climatic zone of Ukraine based on the method of analysis of hierarchies; 3) construction of the resulting map, which includes degrees of vulnerability to climate changes of the studied territory. Modern cloud processing methods for space survey data provide access to a large number of geo-informational products, including the characteristics of the earth's surface and the spatial distribution of climatic indicators accumulated over a long period. These products allow the processing of these data for large areas to be implemented quickly. This technique allows, based on expert assessments, to assess the combined impact of the most significant characteristics of the earth's surface and regional climate, prioritise their impact on the studied territory's vulnerability to climate changes, and implement its quantitative multi-criteria assessment.

Variability in morphology of soot particles during non-cloud and in-cloud processes

The morphology of soot particle influences their optical properties and their ability to act as cloud condensation nuclei (CCN). However, the morphology of soot particles has not been thoroughly examined from non-cloud to in-cloud processes. This study analyzed the morphology and mixing structures of cloud interstitial (INT) and cloud residual (RES) soot particles collected at three mountain sites. Soot particles were classified into four categories based on their mixing structures: INT-Bare, INT-Thinly coated, INT-Thickly coated, and RES soot particles. The fractal dimensions (Df) of INT-Thinly coated (1.93) and INT-Thickly coated (1.99) are larger than those of INT-Bare (1.86) soot particles, indicating significant soot compaction due to non-cloud atmospheric aging processes. Interestingly, we found that the Df of RES soot particles (1.95) is close to that of non-cloud internally mixed soot particles in this study (1.97) and previous literature (1.96). Our results indicate that cloud processing may not further compact the soot particles. This discovery has significant implications for the accurate portrayal of soot morphology within cloud droplets in models, ultimately impacting the modeling of radiative and climate effects. Additionally, our study highlights the potential significance of coating density as a factor influencing soot compaction.

Compact rover surveying and laser scanning for BIM development.

This paper presents a custom made small rover based surveying, mapping and building information modeling solution. Majority of the commercially available mobile surveying systems are larger in size which restricts their maneuverability in the targeted indoor vicinities. Furthermore their functional cost is unaffordable for low budget projects belonging to developing markets. Keeping in view these challenges, an economical indigenous rover based scanning and mapping system has developed using orthogonal integration of two low cost RPLidar A1 laser scanners. All the instrumentation of the rover has been interfaced with Robot Operating System (ROS) for online processing and recording of all sensorial data. The ROS based pose and map estimations of the rover have performed using Simultaneous Localization and Mapping (SLAM) technique. The perceived class 1 laser scans data belonging to distinct vicinities with variable reflective properties have been successfully tested and validated for required structural modeling. Systematically the recorded scans have been used in offline mode to generate the 3D point cloud map of the surveyed environment. Later the structural planes extraction from the point cloud data has been done using Random Sampling and Consensus (RANSAC) technique. Finally the 2D floor plan and 3D building model have been developed using point cloud processing in appropriate software. Multiple interiors of existing buildings and under construction indoor sites have been scanned, mapped and modelled as presented in this paper. In addition, the validation of the as-built models have been performed by comparing with the actual architecture design of the surveyed buildings. In comparison to available surveying solutions present in the local market, the developed system has been found faster, accurate and user friendly to produce more enhanced structural results with minute details.

Position-Feature Attention Network-Based Approach for Semantic Segmentation of Urban Building Point Clouds from Airborne Array Interferometric SAR

Airborne array-interferometric synthetic aperture radar (array-InSAR), one of the implementation methods of tomographic SAR (TomoSAR), has the advantages of all-time, all-weather, high consistency, and exceptional timeliness. As urbanization continues to develop, the utilization of array-InSAR data for building detection holds significant application value. Existing methods, however, face challenges in terms of automation and detection accuracy, which can impact the subsequent accuracy and quality of building modeling. On the other hand, deep learning methods are still in their infancy in SAR point cloud processing. Existing deep learning methods do not adapt well to this problem. Therefore, we propose a Position-Feature Attention Network (PFA-Net), which seamlessly integrates positional encoding with point transformer for SAR point clouds building target segmentation tasks. Experimental results show that the proposed network is better suited to handle the inherent characteristics of SAR point clouds, including high noise levels and multiple scattering artifacts. And it achieves more accurate segmentation results while maintaining computational efficiency and avoiding errors associated with manual labeling. The experiments also investigate the role of multidimensional features in SAR point cloud data. This work also provides valuable insights and references for future research between SAR point clouds and deep learning.

Improved MLP Point Cloud Processing with High-Dimensional Positional Encoding

Multi-Layer Perceptron (MLP) models are the bedrock of contemporary point cloud processing. However, their complex network architectures obscure the source of their strength. We first develop an “abstraction and refinement” (ABS-REF) view for the neural modeling of point clouds. This view elucidates that whereas the early models focused on the ABS stage, the more recent techniques devise sophisticated REF stages to attain performance advantage in point cloud processing. We then borrow the concept of “positional encoding” from transformer literature, and propose a High-dimensional Positional Encoding (HPE) module, which can be readily deployed to MLP based architectures. We leverage our module to develop a suite of HPENet, which are MLP networks that follow ABS-REF paradigm, albeit with a sophisticated HPE based REF stage. The developed technique is extensively evaluated for 3D object classification, object part segmentation, semantic segmentation and object detection. We establish new state-of-the-art results of 87.6 mAcc on ScanObjectNN for object classification, and 85.5 class mIoU on ShapeNetPart for object part segmentation, and 72.7 and 78.7 mIoU on Area-5 and 6-fold experiments with S3DIS for semantic segmentation. The source code for this work is available at https://github.com/zouyanmei/HPENet.

Point-to-Spike Residual Learning for Energy-Efficient 3D Point Cloud Classification

Spiking neural networks (SNNs) have revolutionized neural learning and are making remarkable strides in image analysis and robot control tasks with ultra-low power consumption advantages. Inspired by this success, we investigate the application of spiking neural networks to 3D point cloud processing. We present a point-to-spike residual learning network for point cloud classification, which operates on points with binary spikes rather than floating-point numbers. Specifically, we first design a spatial-aware kernel point spiking neuron to relate spiking generation to point position in 3D space. On this basis, we then design a 3D spiking residual block for effective feature learning based on spike sequences. By stacking the 3D spiking residual blocks, we build the point-to-spike residual classification network, which achieves low computation cost and low accuracy loss on two benchmark datasets, ModelNet40 and ScanObjectNN. Moreover, the classifier strikes a good balance between classification accuracy and biological characteristics, allowing us to explore the deployment of 3D processing to neuromorphic chips for developing energy-efficient 3D robotic perception systems.

From dense point clouds to semantic digital models: End-to-end AI-based automation procedure for Manhattan-world structures

The paper presents a novel method for automatically creating semantic digital models for buildings in Manhattan environments from dense point cloud data. Unlike previous approaches, which rely solely on data-driven methods, our method integrates artificial intelligence with domain engineering knowledge to overcome challenges in indoor point cloud processing and geometry representation in complex layouts. A feature-based Decision Tree classifier extracts main building elements, utilized in a knowledge-based algorithm for 3D space parsing. On this basis, an optimization process generates parameterized floor plans, used to finally create volumetric digital models. The method was validated on datasets from the Technical University of Munich and Stanford University, achieving a mean accuracy of approximately 0.08 m for model placement and 0.06 m for estimating element parameters, which highlights its effectiveness for generating a building's semantic digital model. This approach underscores the potential of AI integration in digital twinning workflows for more automated solutions.

Point Cloud Completion Based on Nonlocal Neural Networks with Adaptive Sampling

Raw point clouds are usually sparse and incomplete, inevitably containing outliers or noise from 3D sensors. In this paper, an improved SA-Net based on an encoder-decoder structure is proposed to make it more robust in predicting complete point clouds. The encoder of the original SA-Net network is very sensitive to noise in the feature extraction process. Therefore, we use PointASNL as the encoder, which weights around the initial sampling points through the AS module (Adaptive Sampling Module) and adaptively adjusts the weight of the sampling points to effectively alleviate the bias effect of outliers. In order to fully mine the feature information of point clouds, it captures the neighborhood and long-distance dependencies of sampling points through the LNL module (Local-NonLocal Module), providing more accurate information for point cloud processing. Then, we use the encoder to extract local geometric features of the incomplete point cloud at different resolutions.Then, an attention mechanism is introduced to transfer the extracted features to a decoder. The decoder gradually refines the local features to achieve a more realistic effect. Experiments on the ShapeNet data set show that the improved point cloud completion network achieves the goal and reduces the average chamfer distance by 3.50% compared to SA-Net.

Unified gas-kinetic wave–particle method for polydisperse gas–solid particle multiphase flow

The gas-particle flow with multiple dispersed solid phases is associated with a complicated multiphase flow dynamics. In this paper, a unified algorithm is proposed for the gas-particle multiphase flow. The gas-kinetic scheme (GKS) is used to simulate the gas phase and the multiscale unified gas-kinetic wave–particle (UGKWP) method is developed for the multiple dispersed solid particle phase. For each disperse solid particle phase, the decomposition of deterministic wave and statistic particle in UGKWP is based on the local cell's Knudsen number. The method for solid particle phase can become the Eulerian fluid approach at the small cell's Knudsen number and the Lagrangian particle approach at the large cell's Knudsen number. This becomes an optimized algorithm for simulating dispersed particle phases with a large variation of Knudsen numbers due to different physical properties of the individual particle phase, such as the particle diameter, material density, etc. The GKS-UGKWP method for gas-particle flow unifies the Eulerian–Eulerian and Eulerian–Lagrangian methods. The particle and wave decompositions for the solid particle phase and their coupled evolution in UGKWP come from the consideration to balance the physical accuracy and numerical efficiency. Two cases of a gas–solid fluidization system, i.e. one circulating fluidized bed and one turbulent fluidized bed, are simulated. The typical flow structures of the fluidized particles are captured, and the time-averaged variables of the flow field agree well with the experimental measurements. In addition, the shock particle–bed interaction is studied by the proposed method, which validates the algorithm for the polydisperse gas-particle system in the highly compressible case, where the dynamic evolution process of the particle cloud is investigated.

Investigation of turbulent flow induced by particle sedimentation using RIM-PIV

The spatiotemporal structure of the flow field during particle–cloud sedimentation has not been sufficiently investigated. In this study, experiments on the sedimentation process of particle clouds in water are conducted using hydrogel particles with a refractive index similar to that of water as the settling particles. The flow field during the sedimentation process of particle clouds in water is measured using particle image velocimetry (PIV). Although the measurement conditions in this study are restricted to one condition owing to the limitations of the measurable area by our PIV system and the available hydrogel particles, the measurement target is novel because it has not been measured so far. The spatiotemporal structure of the turbulent flows is investigated by analyzing the turbulent flows induced by the sedimentation particles using the PIV system. Furthermore, the turbulent structure of the vortex formed by particle sedimentation is examined.