F6
Particle Methods 2

Back to overview

10:40
conference time (CEST, Berlin)
E-drives Spray Cooling Optimization by Moving Particle Simulation
26/10/2021 10:40 conference time (CEST, Berlin)
Room: F
I. Deac, J. Wang (NIDEC PSA emotors SAS, FRA); M. Merelli (Enginsoft, ITA)
I. Deac, J. Wang (NIDEC PSA emotors SAS, FRA); M. Merelli (Enginsoft, ITA)
Over the next decade, e-drive units for Hybrid and Electric traction will continue to evolve towards the industry-specific requirements: increase in power density, efficiency improvement, cost reduction, reliability. On the other hand, the EVs and hybrid platforms are expected to become mainstream even for less expensive cars with high production volumes placing even more pressure on the cost of materials and process for the e-drives. Thermal management of the e-drive is a crucial factor affecting both the performance and life. Internal spray cooling concepts are already used by several players in the automotive and transportation sector in order to cool down the e-drives components with impressive results. Even more, the growing integration of the e-drive and transmission units allows for the use of the transmission fluid to be used as the coolant in the e-motor. The development of this type of cooling is still very dynamic with multiple players exploring the limits of this technology (see recent patents). Nidec-PSA e-Motors joint venture took on the challenge of developing and producing high-performance and competitive e-motors, bridging e-Motor design and manufacture know-how and automotive expertise together in a very ambitious enterprise. The company has adopted simulation techniques since the early stages of the development process to drive the thermal design of their line of e-motors and to predict and optimize the motor temperature. This presentation will focus on the simulation of oil distribution system in the motor, and its use on predicting the heat removal and temperature distribution in the motor. Fast comparison of different concepts and the choice of the best design solution is done by running mesh-less CFD simulations using the Moving Particle Simulation method, which is ideal for simulating oil splashing and jet impingement phenomena in the e-drive. In this presentation engineers from Nidec PSA emotors and EnginSoft will explain the simulation process that allows inspecting virtual prototypes of the e-drive. In the first part, the coolant flow rate and distribution in the cooling channels are predicted for different load cases and flow rate conditions. In the second part, the coolant particles trajectory and heat transfer coefficient on the parts to be cooled are calculated considering the influence of air flow inside the motors. Lastly, the coupling between mesh-less CFD and thermal calculation allow the prediction of critical electrical components temperatures. The results of this procedure and the validation of the numerical results are presented and explained.
emotor, thermal management, spray cooling, moving particle simulation, cfd, experimental validation
11:00
conference time (CEST, Berlin)
Thermal Simulation of an Oil-Cooled E-Motor
26/10/2021 11:00 conference time (CEST, Berlin)
Room: F
D. Percival (EnginSoft UK Ltd, GBR); M. Brada (Ricardo Shanghai Company, CHN)
D. Percival (EnginSoft UK Ltd, GBR); M. Brada (Ricardo Shanghai Company, CHN)
With the huge drive towards electrification, the ability to predict temperature variations during the design process in oil-cooled electric motor windings is of ever-growing importance. The long run times of traditional CFD make it unfeasible to perform this kind of analysis within the design turnaround of a new electric motor. However, a new method of performing these simulations using Particleworks, a meshless particle based CFD solution, has been tested and validated against experimental data. This method can simulate violent multi-phase flow far more efficiently than traditional finite volume based (FVM) CFD methods. The fluid simulation was used to acquire a heat transfer coefficient (HTC) map with an advanced technique of simulating air and oil flow fields separately in order to achieve the most accurate results in the shortest possible time. The HTC map was subsequently applied to a steady state finite element thermal model in order to solve for the full thermal profile throughout the windings. The correlation with the experimental data was well within the targeted tolerance. With the new Particleworks 7.1.0 update, exciting new features provide further functionality and efficiency increases. Post processing tools can now be executed during the solver to allow for finer resolution in the time domain and reduced man-hours per design iteration. Improved speeds of GPU solving and the new FVM solver allows for more advanced simulations than ever before to be ran and faster design turnaround achieved. Finally, a viewing only Graphical User Interface has been implemented to allow management to explore simulations in more detail than from a static video gaining valuable insight into the design without occupying a full license. Join this presentation, given jointly by David Percival (EnginSoft UK) and Martin Brada (Ricardo), to learn more about Particleworks, its applications and how it can benefit you in streamlining your design process.
Particleworks, CFD, Thermal Simulation, E-motor, Heat Transfer,
11:20
conference time (CEST, Berlin)
The Use of Mesh Free Methods in System CFD to Model Multi-Component Transient Duct and Pipe Flows
26/10/2021 11:20 conference time (CEST, Berlin)
Room: F
D. Hunt (Siemens Digital Industries Software, GBR); R. Drury, Z. Pan, (Mentor a Siemens Company, GBR)
D. Hunt (Siemens Digital Industries Software, GBR); R. Drury, Z. Pan, (Mentor a Siemens Company, GBR)
Many flows of industrial interest contain multiple materials moving and interacting within a pressurized system. Where these materials are moving within a pipe or ducted system, then the materials move as batches with little axial diffusion. Examples including the priming of a sprinkler system as high-pressure water moves through an air-filled network. The practice of sending batches of different materials through the same pipeline is utilized in the oil and gas industry or as a way of moving slurries from quarry to the shipping point. System CFD can be used to model large piping systems and capture both the hydraulic and the dynamic behavior. However, finite volume methods do not preserve the concentration fronts between material slugs due to numerical dispersion. Historically, the Method of Characteristics (MOC) approach has been used in 1D CFD to capture pressure dynamics with low dispersion [2-4]. However, these have typically been limited to fixed grid methods with corresponding limitations in time-step and the range of physical phenomena that can be captured. Meshless or particle-based methods have been available for several years in 2D and 3D CFD [1] but are becoming more common for capturing dynamics within multi-material simulations. This paper demonstrates a new approach to modelling pipe and duct flows within a System CFD framework that takes ideas from both 3D CFD and 1D MOC methods to model multi-material flows that preserve fronts and minimize numerical dispersion. The method tracks “particles” in the flow and solves the 1D-flow equation in the fluid reference fame. A particle re-positioning algorithm is used to handle particles leaving the domain and ensures a minimum loss in fidelity by trapping key features between particles. The paper will demonstrate that a particle-based approach, implemented within the Simcenter Flomaster system CFD tool, can be used to model a variety of physical scenarios that are not possible using traditional system CFD approaches including: Batched slurry flows; Oscillating column in a cavitating vertical pipe; and Pipette flows. The paper will further demonstrate that the approach can be used with a wide range of fluid types including liquids, gases and Non-Newtonian materials. References [1] Liu, GR and Gu, YT, “An Introduction to Meshfree Methods and Their Programming”, Springer, 2005. [2] Hunt, DL, et al, “How To - Model Fluid Flow Systems: Computational Fluid Dynamics versus Fluid System Simulation”, NAFEMS, 2017 [3] Toro, EF, “Riemann Solvers and Numerical Methods for Fluid Dynamics”, Springer 2009 [4] Chaudhry, M. H. “Applied Hydraulic Transient”, 3rd Edition, Springer, 2014
system cfd, meshless, 1d cfd, simulation, fluid, flow
11:40
conference time (CEST, Berlin)
Gridfree Simulations of Fluid Structure Interaction (FSI)
26/10/2021 11:40 conference time (CEST, Berlin)
Room: F
J. Kuhnert (Fraunhofer ITWM, DEU)
J. Kuhnert (Fraunhofer ITWM, DEU)
Since 1999, Fraunhofer ITWM develops the so called Generalized Finite Difference Method (GFDM), a purely gridless simulation idea for flows and continuum mechanics. The method bases on moving pointclouds. The points represent numerical nodes, where the differential equations of motion are discretely solved. The points carry no mass, but are purely numerical nodes. As it moves with the flow, the pointcloud’s quality has to be maintained throughout the simulation (point filling, point removal), taking into account local refinement strategies. The moving pointcloud makes it a preferred tool to model free surface flows, FSI, moving geometries. Any partial differential equation can be solved. The central requirement for this is a reliable approximation of spatial derivatives of the discrete functions. The local approximation stencils are established by a local least squares procedure, which allow for desired orders of approximation. This approach represents a generalization of classical Finite Differences. The method is employed in different industrial applications, among them water crossing, rain water management, filling and sloshing, airbag inflation, freezing. In this presentation, we will focus on free surface flow modelling in conjunction with fluid structure interaction (FSI). Usually, FSI is put into praxis by an explicit coupling of a CFD and a structural code. In our particular example, the CFD-method is GFDM, that is coupled to some structural codes, being on the market, or to our own multibody solver (MBS) implementation. Usually, explicit coupling suffers from inherent numerical instabilities, and we show approaches how to neutralize them. As example for the coupling to a structural code, we will show the roll-over simulation of a vehicle in a sand bed, and the interaction of a vehicle with water during water-crossing. The crossing of heavy trucks over a floating bridge will be a demonstration of the coupling of GFDM to our own MBS.
meshfree simulation, fluid structure interaction (FSI), water crossing, floating bridges
×

[TITLE]

[LISTING

[ABSTRACT]

[DATE]

[ROOM]

[KEYWORDS]