Recent Submissions

  • Performance Characterization of High and Low Power Prism based Tunable Blue Laser Diodes Systems

    Mukhtar, Sani; Holguin Lerma, Jorge Alberto; Ashry, Islam; Ng, Tien Khee; Ooi, Boon S.; Khan, M. Z. M. (IEEE, 2020-11-13) [Conference Paper]
    Comparison of high- and low-power tunable external-cavity blue laser-diode system demonstrates a tunability of 10.6 and 4nm, respectively, with a corresponding SMSR as high as 35 and 32dB and linewidth as low as 97 and 59pm, while showcasing high stability at extreme operating conditions.
  • Deep Learning Approach to Predict Rheological Behavior of scCO Foam Fracturing Fluid Under Different Operating Conditions

    Ahmed, Shehzad; Alameri, Waleed; Ahmed, Waqas Waseem; Khan, Sameer (Society of Petroleum Engineers, 2020-11-09) [Conference Paper]
    CO2 foam as a fracturing fluid for unconventional reservoir has been of huge interest due to its potential in solving various challenges related to conventional water-based fracturing. The rheological property of CO2 foam is a key factor controlling the efficiency of fracturing process and it is strongly influenced by different parameters such as foam quality, temperature, pressure and shear rate. The quantification of these parameters under reservoir conditions leads to the design of optimum injection strategy. However, the traditional modeling approaches are unable to provide fast and accurate prediction while considering combined effect of all these parameters. Here, we proposed a data driven approach based on supervised deep learning to estimate rheological property of CO2 foam as a function of foam quality, temperature, pressure, and shear rate. We exploit deep neural networks (DNNs) that are trained to learn the complex nonlinear aspects of the data. For the data generation, we performed a series of experiments for CO2 foams by varying different process variables. CO2 foams at different qualities were generated using conventional surfactant in a flow loop system and foam viscosity measurements were performed at HPHT under wide range of shear rate. The architecture of DNN was optimized to accurately estimate the foam apparent viscosity for given foam quality, temperature, pressure, and shear rate. The predictive capability of designed network is found to be significantly high, analyzed by regression coefficient approaching unity, low mean squared error, and low average absolute relative deviation (≪ 2.5%). The designed neural network allows robust and accurate prediction of foam apparent viscosity at different foam qualities under various reservoir condition, which demonstrates its practicality for CO2 foam projects for fracturing unconventional reservoirs.
  • In Silico Design of Deep Space Optical Links

    Lee, Carlyn-Ann; Xie, Hua; Lee, Charles H.; Lyakhov, Dmitry; Michels, Dominik L. (American Institute of Aeronautics and Astronautics, 2020-11-02) [Conference Paper]
    As deep space links migrate toward higher frequency bands like Ka and optical, thorough trade-space exploration becomes increasingly valuable for designing reliable and efficient communications systems. In this contribution, we leveraged high-performance, concurrent simulations when the run-time complexity of simulation software overwhelms capabilities of ordinary desktop machines. The first part of this manuscript describes how to run error correcting code simulations concurrently on a high-performance supercomputer. The second part of this study describes a framework to produce azimuth and elevation terrain masks from imagery of the Lunar South Pole.
  • Towards a Taxonomy for Automatic and Autonomous Cooperative Spacecraft Maneuvering in a Space Traffic Management Framework

    Hobbs, Kerianne; Collins, Alexander R.; Feron, Eric (American Institute of Aeronautics and Astronautics, 2020-11-02) [Conference Paper]
    As the number of objects in Earth orbit continues to grow, automatic maneuvering of spacecraft may play an important role in a comprehensive space traffic management framework. Automatic collision avoidance, rendezvous and proximity operations, and station keeping are all possible spacecraft missions where two or more spacecraft may interact in an automatic maneuver scenario. Interaction may be limited to awareness of state information such as the other object’s position and velocity relative to an ownship position and velocity, or it may be as extensive as two spacecraft that communicate and coordinate maneuvers to optimize resource use while still meeting mission objectives. Before standards and policies can be determined, a common vocabulary describing spacecraft interactions is needed. This paper proposes a spacecraft maneuver taxonomy that provides a common set of definitions for categories of spacecraft interactions, maneuver coordination, intent, and maneuver efficiency, as well as related concepts such as centralized, distributed, and hierarchical control. It is envisioned that this taxonomy will provide a basis for specifications, planning, coordination, and on-orbit synchronization of spacecraft automatic maneuvering.
  • Leveraging PaRSEC Runtime Support to Tackle Challenging 3D Data-Sparse Matrix Problems

    Cao, Qinglei; Pei, Yu; Akbudak, Kadir; Bosilca, George; Ltaief, Hatem; Keyes, David E.; Dongarra, Jack (IEEE, 2020-11-01) [Conference Paper]
    The task-based programming model associated with dynamic runtime systems has gained popularity for challenging problems because of workload imbalance, heterogeneous resources, or extreme concurrency. During the last decade, lowrank matrix approximations, where the main idea consists of exploiting data sparsity typically by compressing off-diagonal tiles up to an application-specific accuracy threshold, have been adopted to address the curse of dimensionality at extreme scale. In this paper, we create a bridge between the runtime and the linear algebra by communicating knowledge of the data sparsity to the runtime. We design and implement this synergistic approach with high user productivity in mind, in the context of the PaRSEC runtime system and the HiCMA numerical library. This requires to extend PaRSEC with new features to integrate rank information into the dataflow so that proper decisions can be taken at runtime. We focus on the tile low-rank (TLR) Cholesky factorization for solving 3D data-sparse covariance matrix problems arising in environmental applications. In particular, we employ the 3D exponential model of Matern matrix kernel, which exhibits challenging nonuniform ´high ranks in off-diagonal tiles. We first provide a dynamic data structure management driven by a performance model to reduce extra floating-point operations. Next, we optimize the memory footprint of the application by relying on a dynamic memory allocator, and supported by a rank-aware data distribution to cope with the workload imbalance. Finally, we expose further parallelism using kernel recursive formulations to shorten the critical path. Our resulting high-performance implementation outperforms existing data-sparse TLR Cholesky factorization by up to 7-fold on a large-scale distributed-memory system, while minimizing the memory footprint up to a 44-fold factor. This multidisciplinary work highlights the need to empower runtime systems beyond their original duty of task scheduling for servicing next-generation low-rank matrix algebra libraries.
  • High Performance Asynchronous Reverse Time Migration for Oil and Gas Exploration

    Qu, Long; Abdelkhalak, Rached; Ltaief, Hatem; Said, Issam; Keyes, David E. (IEEE, 2020-11-01) [Conference Paper]
    Reverse Time Migration (RTM) is a state-of-the-art algorithm used in seismic depth imaging in complex geological environments for the oil and gas exploration industry. It calculates high-resolution images by solving the three-dimensional acoustic wave equation using seismic datasets recorded at various receiver locations. Using a finite-difference time-domain (FDTD) scheme, the time integration follows an adjoint-state formulation with two successive phases, i.e., forward modeling and backward integration. Each subsurface image is then generated during the imaging condition that combines a forward propagated source wavefield with a backward propagated receiver wavefield. RTM’s computational phases are predominantly composed of stencil computational kernels for the FDTD, applying the absorbing boundary conditions, and I/O operations needed for the imaging condition. In fact, RTM can be considered as an out-of-core algorithm, which requires offloading to disk snapshots of the domain solution at specific time intervals during the forward modeling phase. During the backward time integration, these snapshots are read back and synchronized at the corresponding times. As far as optimizing the stencil computation, spatial blocking represents the widely-adopted vendor-agnostic technique for increasing data reuse in the high-level of the memory subsystem. In this paper, we integrate in RTM the asynchronous Multicore Wavefront Diamond (MWD) tiling approach that permits to further increase data reuse by leveraging spatial with Temporal Blocking (TB) during the stencil computations. This integration engenders new challenges with a snowball effect on the legacy synchronous RTM workflow as it requires to deeply rethink of how the absorbing boundary conditions, the I/O operations, and the imaging condition operate. These disruptive changes in cascade are necessary to maintain the performance superiority of asynchronous stencil execution throughout the time integration, while ensuring the quality of the subsurface image does not deteriorate. We assess the overall performance of the new MWD-based RTM and compare against traditional SB-based RTM on various shared-memory systems using the SEG Salt3D model. The MWD-based RTM is able to achieve up to 60% performance speedup compared to SB-based RTM. To our knowledge, this paper highlights for the first time the applicability of asynchronous RTM executions, which results in a higher simulation throughput and may eventually create new research opportunities in improving the hydrocarbon extraction for the petroleum industry.
  • Investigation of a New Voltage Balancing Circuit for Parallel-connected Offshore PMSG-based Wind Turbines

    Elserougi, Ahmed A.; Bertozzi, Otavio; Massoud, Ahmed M.; Ahmed, Shehab (IEEE, 2020-10-30) [Conference Paper]
    Parallel connection of wind turbines (WTs) is beneficial in high-power applications. For successful operation of parallel WT-based energy conversion systems, a well-regulated voltage is needed at the collection point. Due to wind speed variation, the generated voltage from each WT may differ from one to another. Conventional solutions use regulating converters with full power rating. In this paper, a new concept is presented which depends on using fully-rated uncontrolled rectifier bridges for AC-DC conversion, and partially-rated fully-controlled bridge rectifiers which are used as voltage tuners to guarantee the flow of desired maximum power point DC currents through the parallel connected branches. The proposed system is simple, cost effective, reliable and efficient. The main drawback of the proposed system is the critical need for filters and VAR compensators on the AC side to ensure acceptable performance of the WT generator. Also, smoothing reactors are needed on the DC side for filtering of the transmitted DC current. A simulation model has been built to validate the proposed concept, and the simulation results show the effectiveness of the approach.
  • Symmetrical orientation of spiral-interconnects for high mechanical stability of stretchable electronics

    Qaiser, Nadeem; Damdam, Asrar Nabil; Khan, Sherjeel Munsif; Hussain, Muhammad Mustafa (IEEE, 2020-10-30) [Conference Paper]
    Recently, interconnect based stretchable electronic devices have attained growing interest due to its application for various state-of-the-art technologies. Here, we report an engineered design of spiral interconnects for a series of stretchable networks referred to as the symmetrical series; wherein spirals connect to the island in the symmetry manner. A systematic analysis of Si-based spiral interconnects by numerical modeling, and experiments show that our design provides higher stretchability of 165% in comparison to the conventionally used nonsymmetrical design. The reason for high mechanical reliability is attributed to the favorable unwrapping profile of spiral interconnect due to the nature of forces acting on it during the stretching process. In contrast, for the nonsymmetrical series, the nature of tensile forces produces the rotation, and resultant tilting of spiral arm results in low stretchability of 150%. As a result, nonsymmetrical interconnect fails at earlier stages of stretching. Our study demonstrates the significance of the orientation of spiral interconnects linked to the island to attain the high performance of stretchable electronic devices.
  • Novel Approach to Study the Impact of Asphaltene Properties on Low Salinity Flooding

    Hassan, Saleh F.; Yutkin, Maxim; Kamireddy, Sirisha; Radke, Clayton J.; Patzek, Tadeusz (Society of Petroleum Engineers, 2020-10-21) [Conference Paper]
    Low salinity water flooding (LSW) has gained significant attention, because of its advantages compared with other enhanced oil recovery (EOR) methods. LSW's positive contribution to recovery factor has been demonstrated in the literature at lab and field scales. However, LSW flooding does not always increment oil recovery. It is a specific combination of properties of an asphaltenic crude oil, chemically equilibrated brine, and rock surface that may explain the success or failure of LSW. In this work, we introduce a novel experimental approach to study asphaltene-like chemical interactions with surfaces rock minerals to evaluate the effectiveness of applying LSW. When studying the impact of asphaltene properties on incremental recovery, one aims to detach some of the immobile oil, which is semi-irreversibly stuck on rock surface. This is a difficult task, because of varying crude oil composition, as well as asphaltene interfacial and chemical properties. To overcome these issues, we split the problem into several parts. We study how mono- and poly-functional chemical compounds mimic asphaltene interactions with mineral surfaces, like silica and calcium carbonate, which are proxies for sandstones and limestones, respectively. For example, amines, quaternary ammonia or carboxylates represent asphaltene functional groups that are mainly responsible for crude oil base and acid numbers, respectively. Adsorption of polymers and oligomers containing such groups mimics the irreversible asphaltene deposition onto rock surface through formation of chemically active polymerlike structures at the oil-brine interface. The silica surface is negatively charged in brines with pH above 2. Silica attracts positively charged ammonia salts, such as cetrimonium chloride (CTAC). However, negatively charged mono-functional carboxylates, i.e. anionic surfactants, like sodium hexanoate (NaHex), hardly adsorb onto silica, even in the presence of a bridging ion, like calcium. In contrast to silica, calcium carbonate surface has both positive and negative charges on its surface. We found that CTAC adsorbs onto calcium carbonate in any brine tested. NaHex shows minimal adsorption onto calcium carbonate only in the presence of calcium ions suggesting a contribution of an ion-bridging mechanism. Adsorption of all studied mono-functional surfactants is fully reversible and, consequently not representative of asphaltenes. Multifunctional compounds, i.e., polymers, demonstrate irreversible, asphaltene-like, adsorption. We studied adsorption of carbohydrates decorated with individual amines and quaternary ammonia functional groups. The carbohydrates with amine functional groups adsorb irreversibly on calcium carbonate and silica in all tested brines with pH up to 10. Therefore, a lower base number (BN) in crude oils indicates a higher potential for LSW. Our findings demonstrate the proof of concept that contribution of different functional groups to asphaltene adsorption/deposition can be studied using functionalized water-soluble polymers. This framework is useful for assessment of adsorption strength vs. number of active groups as well as screening of efficient detachment process of asphaltenic crude oils from rock surface
  • Interfacial Viscoelasticity in Crude Oil-Water Systems to Understand Incremental Oil Recovery

    Saad, Ahmed M.; Aime, Stefano; Mahavadi, Sharath C.; Song, Yi-Qiao; Patzek, Tadeusz; Weitz, David (Society of Petroleum Engineers, 2020-10-21) [Conference Paper]
    Improved oil recovery from asphaltenic oil reservoirs may provide the world with a significant source of lower-cost energy over many decades. However, the mechanisms through which the surface-active components in crude oil, such as asphaltenes and organic acids, affect incremental oil production are still unclear. In this study, we investigate crude oil/water interfacial films using shear and dilational rheology for mechanical properties and Fourier-Transform Infrared Spectroscopy (FTIR) to better understand its molecular species present at the interface that contribute to the development of viscoelastic behaviors. Dilational rheology has proven to be more sensitive to early time development of elasticity. In contrast, shear rheology provided more insights regarding the formation of elastic films at the macroscopic scale and late time interfacial changes. The presence of salts such as sodium chloride in the aqueous phase played a critical role in altering the dynamics of both the rheological properties development and the interfacial tension.
  • A Dual-mesh Framework for Multiphysics Simulation of Photoconductive Terahertz Devices

    Chen, Liang; Bagci, Hakan (IEEE, 2020-10-21) [Conference Paper]
    The operation of a photoconductive terahertz (THz) device relies on optoelectronic interactions and THz radiation. Although these two processes are coupled, in simulations, they are often modeled separately due to the large difference between the frequencies of optical and THz electromagnetic waves. To model both processes in a single simulation, we propose a dual-mesh discontinuous Galerkin (DG) scheme. Optoelectronic interactions and THz radiation are accounted for by solving, respectively, a coupled system of Maxwell and drift-diffusion equations discretized on a fine mesh and only the Maxwell equations discretized on a coarse mesh. This approach uses an efficient high-order interpolation scheme to “connect” electric field and current discretized on these two meshes. The proposed scheme is validated against experimental results.
  • Assessment of polymer-induced formation damage using microfluidics

    Sugar, Antonia; Torrealba, Victor; Buttner, Ulrich; Hoteit, Hussein (Society of Petroleum Engineers, 2020-10-21) [Conference Paper]
    Polymers have been successfully deployed in the oil&gas industry in various field implementations, including mobility control in waterflood, flow divergence, and well conformance control. However, lab and field applications of polymer injections often encounter polymer-induced formation damage related to pore-throat clogging from polymer entrapments, leading to permeability reduction. This phenomenon manifests as a loss of injectivity, which can diminish the recovery performance. The first principles of polymer interaction with porous rocks are poorly understood. In this work, we use microfluidics to assess formation damage induced by polymer flood. Microfluidic techniques offer convenient tools to observe polymer flow behavior and transport mechanisms through porous media. The microfluidic chips were designed to mimic the pore-size distribution of oil-bearing conventional reservoir rocks, with pore-throats ranging from 1 to 10 µm. The proposed fabrication techniques enabled us to transfer the design onto a silicon wafer substrate, through photolithography. The constructed microfluidic chip, conceptually known as "Reservoir-on-a-Chip", served as a two-dimensional flow proxy. With this technique, we overcome the inherent complexity of the three-dimensional aspects of porous rocks to study the transport mechanisms occurring at the pore-scale. We performed various experiments to assess the mechanisms of polymer-rock interaction. The polymer flow behavior was compared to that of the water-flood baseline. Our observations showed that prolonged injection of polymer solutions could clog pore-throats of sizes larger than the measured mean polymer-coil size, which is consistent with lab and field observations. This finding highlights a major limitation in some polymer screening workflows in the industry that suggest selecting the candidate polymers based solely on their molecular size and the size distribution of the rock pore-throats. This work emphasizes the need for careful core-flood experiments to assess polymer entrapment mechanisms and their implication on short- and long-term injectivity.
  • Tail Entity Recognition and Linking for Knowledge Graphs

    Zhang, Dalei; Qiang, Yang; Li, Zhixu; Fang, Junhua; He, Ying; Zheng, Xin; Chen, Zhigang (Springer International Publishing, 2020-10-16) [Conference Paper]
    This paper works on a new task - Tail Entity Recognition and Linking (TERL) for Knowledge Graphs (KG), i.e., recognizing ambiguous entity mentions from the tails of some relational triples, and linking these mentions to their corresponding KG entities. Although plenty of work has been done on both entity recognition and entity linking, the TERL problem in this specific scenario is untouched. In this paper, we work towards the TERL problem by fully leveraging KG information with two neural models for solving the two sub-problems, i.e., tail entity recognition and tail entity linking respectively. We finally solve the TERL problem end-to-end by proposing a joint learning mechanism with the two proposed neural models, which could further improve both tail entity recognition and linking results. To the best of our knowledge, this is the first effort working towards TERL for KG. Our empirical study conducted on real-world datasets shows that our models can effectively expand KG and improve the quality of KG.
  • Fully printed VO2switch based flexible and reconfigurable filter

    Yang, Shuai; Li, Weiwei; Vaseem, Mohammad; Shamim, Atif (IEEE, 2020-10-14) [Conference Paper]
    Frequency reconfigurable filters are high in demand because they can cover multiple frequency bands, thus minimizing system level cost and size requirements. Another emerging trend is the flexibility or conformability of the electronic components, so that they are suitable for mounting on non-planar objects as well as for wearable applications. In this work, we demonstrate a frequency reconfigurable bandpass filter that has been fully printed on a flexible Kapton substrate. The frequency reconfigurability has been achieved through a switch made of Metal Insulator Transition (MIT) material Vanadium-di-oxide (VO2). The VO2 switch has been printed through a custom ink. The switch is in OFF condition (insulating condition) at room temperature and turns ON (becomes conductive) at MIT temperature of 68°C. The microstrip bandpass filter employs dual mode resonators and can switch from 4.0 GHz to 3.7 GHz. The required thermal bias is provided through a printed heater which is attached to the backside of the filter. Due to the flexible Kapton substrate and the printing process, the prototype of the filter is highly flexible and low cost. Measured results are promising and in good agreement with the simulation results.
  • Multi-Dimensional Integration and Packaging of Devices for Internet-of-Things Applications

    Elatab, Nazek; Suwaidan, Reema; Alghamdi, Yara; Alhazzany, Alhanouf; Almansour, Reema; Shaikh, Sohail F.; Khan, Sherjeel M.; Hussain, Muhammad Mustafa (IEEE, 2020-10-13) [Conference Paper]
    IoT applications are increasingly becoming widespread with more stringent system requirements. In this work, we demonstrate a nature-inspired integration and packaging technology that achieves self-powered multi-functional systems with optimized performance and small footprint area. The integration technique is based on bifacial usage of the substrate where devices on both sides are interconnected via through-substrate-vias. Multiple substrates are then integrated and folded into a 3D architecture using side-interlocks following a puzzle-like fashion. On the outer sides of the 3D architecture, sensors, RF devices and energy harvesters are integrated while on the inner faces, a solid-state battery in addition to power- management and data-management circuitry are embedded. To package the system, a polymeric encapsulant is used to protect the inner circuitry and enhance the mechanical resilience of the system. Finally, the system is used to send the collected data wirelessly to a phone using an embedded Bluetooth Low Energy unit.
  • Multi-Dimensional Integration and Packaging of Devices for Internet-of-Things Applications

    Elatab, Nazek; Suwaidan, Reema; Alghamdi, Yara; Alhazzany, Alhanouf; Almansour, Reema; Shaikh, Sohail F.; Khan, Sherjeel M.; Hussain, Muhammad Mustafa (IEEE, 2020-10-13) [Conference Paper]
    IoT applications are increasingly becoming widespread with more stringent system requirements. In this work, we demonstrate a nature-inspired integration and packaging technology that achieves self-powered multi-functional systems with optimized performance and small footprint area. The integration technique is based on bifacial usage of the substrate where devices on both sides are interconnected via through-substrate-vias. Multiple substrates are then integrated and folded into a 3D architecture using side-interlocks following a puzzle-like fashion. On the outer sides of the 3D architecture, sensors, RF devices and energy harvesters are integrated while on the inner faces, a solid-state battery in addition to power- management and data-management circuitry are embedded. To package the system, a polymeric encapsulant is used to protect the inner circuitry and enhance the mechanical resilience of the system. Finally, the system is used to send the collected data wirelessly to a phone using an embedded Bluetooth Low Energy unit.
  • Finding the right cloud configuration for analytics clusters

    Bilal, Muhammad; Canini, Marco; Rodrigues, Rodrigo (ACM, 2020-10-12) [Conference Paper]
    Finding good cloud configurations for deploying a single distributed system is already a challenging task, and it becomes substantially harder when a data analytics cluster is formed by multiple distributed systems since the search space becomes exponentially larger. In particular, recent proposals for single system deployments rely on benchmarking runs that become prohibitively expensive as we shift to joint optimization of multiple systems, as users have to wait until the end of a long optimization run to start the production run of their job. We propose Vanir, an optimization framework designed to operate in an ecosystem of multiple distributed systems forming an analytics cluster. To deal with this large search space, Vanir takes the approach of quickly finding a good enough configuration and then attempts to further optimize the configuration during production runs. This is achieved by combining a series of techniques in a novel way, namely a metrics-based optimizer for the benchmarking runs, and a Mondrian forest-based performance model and transfer learning during production runs. Our results show that Vanir can find deployments that perform comparably to the ones found by state-of-the-art single-system cloud configuration optimizers while spending 2X fewer benchmarking runs. This leads to an overall search cost that is 1.3 - 24X lower compared to the state-of-the-art. Additionally, when transfer learning can be used, Vanir can minimize the benchmarking runs even further, and use online optimization to achieve a performance comparable to the deployments found by today's single-system frameworks.
  • Cellular network Marine Sensor Buoy

    Przybysz, Alexander; Duarte, Carlos M.; Geraldi, Nathan; Kosel, Jürgen; Berumen, Michael L. (IEEE, 2020-10-12) [Conference Paper]
    Studies in the marine environment require devices to gather data in remote locations. These devices commonly use some form of data logging that require the user to retrieve the device after a period of time. This is not only a large effort, but often presents an unnecessary delay and risk in the event that the device malfunctions, goes missing or if the data collected is redundant. We propose a modular, remote and autonomous sensing solution that enables multi-sensor readout and wireless data communication, providing scientists with real time data. This sensor buoy is comprised of a primary module which is a data logger that connects to an online server via mobile network. The secondary portion is the hub that connects to an array of sensors that are customized to the needs of the marine scientists.
  • Particle concentration variation for inflow profiles in high reynolds number turbulent boundary layer

    Rahman, Mustafa M.; Samtaney, Ravi (American Society of Mechanical Engineers, 2020-10-12) [Conference Paper]
    Large-eddy simulations (LES) of incompressible turbulent boundary-layer flows can simulate a fundamental unsteady turbulent flow, including time-variant streamwise and wall-normal velocity as well as the near-wall locations of significant turbulence intensities. A typical illustration of turbulent flows with such high Reynolds numbers can be roughly approximated to atmospheric boundary-layer flows. To bypass the demanding mesh criteria of near-ground field and direct numerical simulations, we adopt a virtual-wall model with a stretched-vortex subgrid-scale model. We simulate the dynamics of solid particles in this wall-modeled LES approach toward incompressible flow. The particles considered are both charged and uncharged, and have a fixed concentration profile with no fluctuations at the inflow. An extended streamwise simulation domain is implemented as an alternative to rerunning the simulation with a turbulent inflow profile from the simulation of the previous downstream profile. By extending the streamwise domain, the fluctuation dynamics of the particles reach a steady state far downstream from the inflow. The streamwise and altitude variation of the particle parameters are compared for various particle-concentration inflow profiles. Furthermore, an estimate of the streamwise variation of parameters is also observed. This study is the first step towards enhancing our understanding of the particle dynamics in turbulent flows.
  • How Does Lipschitz Regularization Influence GAN Training?

    Qin, Yipeng; Mitra, Niloy; Wonka, Peter (Springer International Publishing, 2020-10-10) [Conference Paper]
    Despite the success of Lipschitz regularization in stabilizing GAN training, the exact reason of its effectiveness remains poorly understood. The direct effect of K-Lipschitz regularization is to restrict the L2-norm of the neural network gradient to be smaller than a threshold K (e.g.,) such that. In this work, we uncover an even more important effect of Lipschitz regularization by examining its impact on the loss function: It degenerates GAN loss functions to almost linear ones by restricting their domain and interval of attainable gradient values. Our analysis shows that loss functions are only successful if they are degenerated to almost linear ones. We also show that loss functions perform poorly if they are not degenerated and that a wide range of functions can be used as loss function as long as they are sufficiently degenerated by regularization. Basically, Lipschitz regularization ensures that all loss functions effectively work in the same way. Empirically, we verify our proposition on the MNIST, CIFAR10 and CelebA datasets.

View more