Conference Proceedings

Access technical proceedings from past user conferences to learn about the many real-world applications of FLOW-3D products and the recent and upcoming developments for the FLOW-3D product suite.

New Conference Proceedings

The conference proceedings from the 2018 FLOW-3D European Users Conference are now available!

Speakers 2018 FLOW-3D European Users Conference
The distinguished speakers at the 2018 FLOW-3D European Users Conference

Accelerating Simulation Workflows through User Interface Design

John Ditter, Flow Science

As simulation becomes deeply embedded in the product design cycle, the need to improve simulation workflows increases. A number of productivity-enhancing features have been implemented in FLOW-3D’s user interface to improve the workflow. These features, which include runtime simulation monitoring, interactive geometry creation, remote solving, batch post-processing, and active simulation control with global conditions, will be discussed. New features coming soon include a new, high performance scenario results viewer, more efficient handling of raster topography, the HPC version of FLOW-3D CAST, and more plotting features in runtime simulation monitoring.

Additive Manufacturing and Foaming Applications

Raed Marwan, Flow Science Japan

FLOW-3D’s WELD module has been in research and development for six years now with the cooperation of many users in Japan and other parts of the world. We will present new features in the latest version of the module along with sample applications to show the module’s capabilities. We will also present a new foaming module. Polyurethane foam has many excellent features such as thermal insulation and shock and sound absorption, and is used for refrigerator parts, car parts and many other products. Simulating the precise behavior of polyurethane foaming is very effective for reducing the development period and production costs. During several years of cooperation with Hitachi through collaborative experiments and foaming flow simulations, a new module was developed by Flow Science Japan to extend FLOW-3D’s capabilities to simulate foaming flow simulations.

Influence of Floating Debris Racks on Flow Rate at Standard Spillways

Grégory Guyot, Christian Lassus, and Cécile Idelon, EDF-CIH

Flood events may entrain large amounts of floating debris; especially large wood debris may clog or dramatically decrease the spillway capacity. For safety reasons, it is crucial to find a solution that maintains the required discharge and prevents the floating debris from entering the lake upstream. Furthermore, a floating debris rack may be proposed to avoid the clogging of the spillway. Therefore, two numerical models were achieved to precisely determine the best location of a rack. For that purpose, FLOW-3D was coupled with EDF’s Python automation. It enabled around 1300 simulations to be created and run automatically in order to analyze the ranges of all influent parameters. A first simulation run was performed in 2D to efficiently determine the most suitable location. Then a full 3D model was used to investigate the influence of the rack design and the type of debris on the discharge drop. Both of the numerical results were compared with physical experimentation. First, the numerical models help us determine a location where the discharge capacity is preserved despite the debris. Second, they give us an accurate understanding of the different clogging processes depending on the wood debris type.

Complexity & Diversity in Sands & Gravels: A Case Study of Decision Options for Sediment Transport Modelling

Kate Bradbrook, JBA Consulting

FLOW-3D’s sediment transport model is one of the most advanced available to commercial users. For example, it is possible to specify several bed layers, each with a different grain size distribution. This presentation will start by showing a scour simulation at a case-study bridge with such a detailed representation and discusses the information needed as well as the computational overheads and shows interpretation of the results. However, in many real-world examples, the required input data for such a detailed representation will simply not be available; either significant assumptions would have to be made or a simpler approach taken. In either case, sensitivity testing is standard practice. We have therefore repeated our case-study bridge simulation with different levels of detail and assumptions. We will show, for example, the difference in modelling requirements and potential conclusions resulting from decisions such as uniform sediment versus mixed grain sizes, single layer versus several layers, and clear water input versus upstream sediment supply. As with many modelling decisions, the appropriate level of detail will be case-specific depending on specific characteristics and the reason for modelling. It is hoped that the comparisons shown in this presentation will provide useful information to help this decision-making process. The presentation will end with a real-world example where, with the minimum of data, we attempt to replicate a catastrophic scour event.

Simulation-based Development of a Casting Process to Produce Clad Aluminum Strips

Stefan Heugenhauser1, Erhard Kaschnitz1, and Peter Schumacher1,2

1 Österreichisches Gießerei-Institut
2 University of Leoben

Casting liquid melt on a preheated substrate is a promising way to produce clad aluminum strips. A small-scale pilot plant to cast pure aluminum (Al99.8) on aluminum alloy plates (Al7075) was developed to investigate the formation of a metallurgical bond at the interface. A three-dimensional numerical simulation model of the casting device was set up using FLOW-3D to find suitable thermal conditions for the melt flow in the casting device as well as for the composite contact region. The model was iteratively calibrated using measured temperatures obtained from various casting experiments. However, the temperatures at the interface between the substrate and the clad alloy are not accessible by direct measurement. Therefore, a more accurate two-dimensional simulation of the middle section of the casting device was modeled. Furthermore, a sub-model of the two-dimensional simulation was derived to compute the thermal conditions at the interface as well as in its vicinity in extremely high temporal and spatial resolution. The simulation shows the occurring re-melting and solidification processes during the compound formation. The obtained results correlate very well with electrochemically etched cross sections of the cast bilayer aluminum strips.

Improvement of Shrinkage Macro-Porosity Prediction Capability

Daniele Grassivaro, Form Srl

Many factors influence the formation of shrinkage macro-porosity, such as the geometry of the cast, position and geometry of gates, temperature of the mold, temperature and solid fraction of the metal after filling, intensification pressure, and alloy properties. In this study, we will focus on the influence of the position and geometry of gates. Simulations should help us to design a feeding system in order to minimize the shrinkage porosity, but we have seen that this is not true. Simulations usually show that gates get solidified suddenly after filling and then it seems they are not useful to compensate for the shrinkage, while in reality they continue to feed during the intensification phase. We will see how to tune our simulations to get results closer to reality. Thanks to: Denso Manufacturing Italy, site Barberino (FI).

Support of the Design Process for Iron Castings

Malte Leonhard, Flow Science Deutschland GmbH

The design process of casting parts consists of several phases, each having different requirements in regard to the simulation of the process. The process starts with an approximate dimensioning of the gating and feeding system in order to estimate the amount of return scrap and by this provide a quick and simple assessment of the cost of the casting. Next is the design phase, in which different gating and feeding concepts are evaluated, and finally the design of the gating and feeding system is fixed. The result is a prototype with correct dimensions and minimal porosities. The final phase is the optimization of the tool design for mass production with the goal of minimizing material usage and scrap. This presentation illustrates how FLOW-3D CAST can aid the different phases of the design.

Spillway Assessment – When is a Splash a Splash?

Mary Jeddere-Fisher, Mott MacDonald

As standards change and improvements are made in hydrological methods, existing spillways often need re-assessment. Typically, water depths and velocity measurements are required along the length of the spillway. The high level of detail in results made accessible by FLOW-3D is very desirable and can be used to provide high resolution long profiles of water depths and velocities, making it possible to refine and minimise planned spillway upgrade works. However, such measurements from unsteady simulation of turbulent flows are not straightforward. Careful consideration must be made in terms of how data are extracted, and what sampling frequency and period are sufficient for a robust assessment. This presentation will discuss some of the sampling methods used to extract such data and make a comparison of FLOW-3D results against a recent physical model study which looked at splash heights and training wall requirements on stepped spillways.

Flow in a Peristaltic Pump: Two-Fluids Approach

Julien Bœuf, Roche Diagnostics GmbH

Peristaltic pumps are commonly used in the pharmaceutical industry. However, in case of high sensibility to shear forces of the dispensed biopharmaceutical solution, limitations in process parameters must be carefully considered. For example, denaturing of protein during processing (aggregation, unfolding, fibril formation, loss of activity, etc.) can occur at a too high or a too long shearing. CFD models basically allow a better theoretical understanding of the flow and of the shear stress applied to the proteins. In a peristaltic pump, the fluid is contained within a flexible tube. The tube is occluded by a solid part, forcing the fluid to be pumped to move through the tube. Of course, this is basically an FSI problem with high deformation of the solid (tube), and the structural mechanical part of the problem is not fully covered by the current FLOW-3D software version. However, because the only physical aspect of interest is the fluid flow, the tube deformation can be roughly modeled using a visco-elastic fluid in a two-fluids approach. The delivered precision of the tube deformation is sufficient for the purpose of this study. Within this modeling approach, the shear rate field of different processes can be compared. In this presentation, commercial linear and radial peristaltic pumps are considered. To evaluate the ability of the devices to dispense small volumes of pharmaceutical solution, the flow and the resulting shear rates are analyzed. The numerical models are described, the strength and limits of this numerical approach are discussed, and the main results reviewed.

Extension of FLOW-3D with a New Phase Change Model to Simulate Cryogenic Tank Flows

Martin Konopka, ArianeGroup GmbH

Cryogenic liquid propelled rockets are equipped with tanks containing subcooled liquid propellants such as hydrogen and methane. The pressure evolution in such tanks is governed by the heat exchange of the vapor with the tank walls and liquid surface, and the phase change at the liquid-vapor interface. Since phase change is a major driver for the pressure evolution, the accurate prediction of the phase change rate is of crucial importance for the design of cryogenic liquid tanks and their pressurization systems. Therefore, in this study FLOW-3D is extended with a customized phase change model based on the temperature gradients at the phase boundaries. To validate the customized phase change model, liquid evaporation at a heated wall, liquid evaporation at a superheated liquid body at a wall, the bubble growth problem in superheated liquid and meniscus evaporation in a heat pipe is considered. The numerical solutions of all computations show a good agreement with reference solutions. Furthermore, the phase change model is applied to the micro-gravity tank flow problem SOURCE-2 where HFE-7000 evaporates at a superheated wall. With the new phase change model, FLOW-3D can predict the pressure evolution with a maximum deviation of 0 % to 20 % compared with measurements.

Solver Developments: FLOW-3D v12.0 and FLOW-3D CAST v5.1

Michael Barkhudarov, Flow Science

FLOW-3D users will see many exciting and far-reaching developments in 2018. Hydraulic applications will benefit from a significant upgrade of the sediment transport model and a new outflow pressure boundary condition, both resulting in more accurate and stable solutions. A new multi-component alloy model will mark a major enhancement of the solidification simulation capabilities in FLOW-3D CAST, opening the doors to predicting microstructure and mechanical properties. A full radiation model, along with the shell generation feature, will enable investment casting modeling. All users will benefit from the addition of general unit systems. The accuracy of the two-fluid solution has been advanced by extending it to a two-temperature approach, while the new Immersed Boundary Method improves the accuracy of the near-wall flow solution. There will be a debut of the HPC version of FLOW-3D CAST, which will be available on the POD cloud service, along with the HPC version of FLOW-3D.

IMPROVEit, a Modern Optimization Engine for FLOW-3D

Raul Niccolò Pirovano, XC Engineering Srl

In the panorama of optimization software and DOE schedulers there is a new tool, designed specifically for FLOW-3D users: IMPROVEit software. IMPROVEit is a very modern and easy-to-use software that performs automated optimization research and regression analysis in a very user-friendly environment. There are several advantages compared to other solutions for CAE users. First, it is a fully automated solver where users do not need to know anything about optimization algorithms. Second, it is ‘budget’ based in the sense that users will enter their available time and the software will achieve the best result within that time. Third, it is optimized for long simulation runtimes and for FLOW-3D users. Regarding FLOW-3D users, the software embeds a “FLOW-3D solver” node, an automated “FLOW-3D postprocessing” node to extract the variable of interest, and additional nodes to morph geometries whatever they are: STLs or third-party CAD tools.

Faster FLOW-3D Time-to-Solution with HPC Cloud or on Premise

Rod McAllister, Penguin Computing

This presentation describes Penguin Computing On Demand (POD), Penguin Computing’s on-demand HPC cloud. Penguin Computing has been working with Flow Science for several years to assure the best performance and user experience for running FLOW-3D on POD. With Penguin’s remote 3D desktop solution Scyld Cloud Workstation attached to the cluster fabric, customers can use the FLOW-3D GUI, run simulations on a true HPC cluster, and process the results without downloading large output files and with no specialized HPC skills. Penguin Computing supplies this solution as a cloud or on-premise resource.

A Case Study of Flood Modelling from Tailings Storage Facility Failure

Rudolf Faber, Pöyry Energy GmbH

Tailings material is deposited in stacks which may reach heights of 100 meters. Depending on the material properties, in the case of failure such stacks pose risks of liquefied material runout, flooding, pollution, etc. Several aspects from a recent client project on tailings material flood modelling will be introduced. After a brief outline of the study area with the areas at risk around the storage facility, the failure and flooding process based on the literature are outlined first, including the non-Newtonian fluid character of the tailings material. Second, the model setup in terms of topography, fluid parameters, meshing and boundary conditions are discussed as well as sensitivity aspects. Third, the definition of failure scenarios under investigation is addressed with a focus on the tailings stack’s breach outflow as a main factor for the intensity of the simulated flooding. And finally, selected results are presented and specific experiences from the entire modelling project are discussed.

Hydrodynamic Forces on a High-head Slider Gate

Boris Huber, Vienna University of Technology

Slider gates are commonly used to operate bottom outlets in high-head power plants. When the gates are opened, flow passes underneath, generating hydrodynamic forces which can be lift or drag forces. Usually the weight of the gate is sufficient so that it closes automatically, but if lift forces are too high, the gate might not close. In the case under consideration, hydrodynamic forces on a rather unconventional gate were simulated with different CFD codes.

Definition of a Robust Aluminum HPDC Process by Mean of Virtual Simulation, Design of Experiment and Taguchi Methodology

Claudio Mus, Endurance Overseas and Raul Pirovano, XC Engineering

In die casting processes, a high quality of the produced parts is required more and more. High quality products in high pressure die casting can be achieved by optimizing, often through numerical process simulations, the geometry of runners, gates, overflows and of the part itself, and by precisely calibrating the different process parameters like melt temperature, holding time, injection pressure and velocity, cooling rate and so on. Despite all of this, during the real production in a foundry there is always a certain variability of the process parameters which could lead to variations, sometimes significant, of the desired performances, and consequently to an increase of the scrap. For this reason, it is important to study the influence of the main process parameter on the product’s functional characteristics due to manufacturing variations and sources of noise which are variables that are impossible or expensive to control. The Taguchi methodology allows the effect of multiple variables to be analysed without performing an expensive full factorial design, which would require a large number of simulations or experiments, by choosing a limited and well-thought out selection of tests. In the present work, the Taguchi methodology has been applied to an HPDC process, and simulated with FLOW-3D CAST, analysing the influence of metal temperature and plunger speed to assess their effect on gas and shrinkage porosity. Consequently, the main guidelines for a more robust process have been identified.

Feasibility Study on the Simulation of Ultrasonic Treatment of Liquid and Solidifying Aluminium A356

Eric Riedel, Otto-von-Guericke-University Magdeburg

Ultrasonic treatment and its associated effects, namely cavitation and acoustic streaming, are promising for targeted treatment of liquid light metal alloys for achieving grain and structure refinement, and therefore, better mechanical properties. Furthermore, this treatment is considered a comparatively clean technology since no additions to the liquid are necessary. The simulation-based prediction of these process phenomena is an important requirement for better understanding, implementing and scaling of this technology for foundry processes. However, challenges remain including the high-frequency movement (20 kHz), small amplitude (35 mm) of the ultrasonic radiator and the resultant small time-step definition. Using FLOW-3D, we studied the ability to calculate the onset and expansion of cavitation and acoustic streaming for A356 (AlSi7Mg0,3) in the liquid state as well as during solidification. Our investigation shows that the results obtained with FLOW-3D are in good agreement with the theoretical and practical outcomes of other studies. The simulation of the solidification accompanying treatment indicates that the streaming caused by the ultrasonic treatment on the one hand supports the distribution of particles within the melt, and on the other hand counteracts a classical exogenous solidification process, thus promoting a more homogeneous casting structure.

Orlik Dam – Protection of the Dam Against Large Floods

Tomas Studnicka and Jan Sehnal, AQUATIS

The Vltava Cascade is a sequence of 9 retention reservoirs on the Vltava River, and the Orlik reservoir forms a fundamental part of it. Holding 720 million m3 of water, it is the largest retention reservoir in the Czech Republic. In August 2002, catastrophic floods hit a considerable part of the Czech Republic. The region most affected was the Vltava River basin. During the flood event, the water level in the Orlik reservoir considerably exceeded the maximum allowable level, which resulted in the flooding of the Orlik hydropower plant that put it out of operation. The purpose of the project “Orlik Dam – Protection of the Dam Against Large Floods” was to design technical measures that would withstand extreme floods. The new standards for dam safety in the Czech Republic require that the most important water structures of this type must safely stand up to a Ten Thousand Years – Control Flood. In the project, numerical modelling followed by physical modelling have been used. This presentation describes the performed numerical simulations and compares their results with the data obtained through physical modelling.