36
Meeting the Challenges of Fluid-Structure Interaction Alan Mueller, Chief Technology Officer

Meeting the Challenges of Fluid-Structure Interaction

  • Upload
    others

  • View
    3

  • Download
    0

Embed Size (px)

Citation preview

Meeting the Challenges of Fluid-Structure Interaction

Alan Mueller, Chief Technology Officer

Air Interaction with a Flexible Structure

What is FSI?

Water/Air Interaction with a Structure

What is FSI?

Vortex Induced Vibration and Galloping

What is FSI?

4

Aeroelastic Flutter

What is FSI?

5

Hydroplaning

What is FSI?

6

Gulping

What is FSI?

7

Ask 20 engineers “What is FSI?” and you will likely get 20 different answers There is not simply one approach valid for all FSI problems The analyst must be presented with a range of options and chose the most suitable

What is FSI?

8

Mapping data techniques – Finding neighbors and interpolating

Protocols and formats for exchanging data – Getting data from Code A to Code B

Coupling methods – Algorithms for accuracy, stability

Dynamic fluid mesh evolution – Topology changes in the fluid domain

Validation of FSI results

The Challenges of FSI

9

Presenter
Presentation Notes
If you take away anything from my presentation, I hope it is an understanding of major challenges of FSI

The Challenges of FSI

10

MAPPING

Presenter
Presentation Notes
The first challenge is Mapping

Searching for opposing neighbors Interpolating source stencil data on a target point – Source and targets may be face or vertex

Often requires integration (quadratures) – intensive extensive variables

• pressure force • heat flux heat

– FEA nodal loads: integration of intensive variables against the shape function.

The 3 steps of “Mapping”

11

Presenter
Presentation Notes
Mapping has 3 important steps, the first of which is the most expensive and difficult

Search requires little user intervention The search excludes potential neighbors based on proximity and orientation Critically important for sheet metal parts – resolve ambiguities of poor geometry – thin solid parts may be on the wrong side of the fluid surface

Parallel Mapping is a must! – Takes advantage of distributed memory

Source Neighbor Search Imperative

12

Co-Simulation Vehicle Thermal Management

Presenter
Presentation Notes
When done with the calculation we have the changing temperatures of all the parts. We see the exhaust system cool down through radiation. Some of the parts of the car will get hotter before they cool down. In the next slides I will show the changing temperature for some individual parts.

Mapping Displacement in Low Y+ Prism Meshes

14

C0 continuous mapping very important for low y+ meshes

Otherwise very easy for morpher to invert high aspect ratio cells in prism layer

The Challenges of FSI

15

DATA EXCHANGE

Presenter
Presentation Notes
We move on to the next challenge of how to exchange data between CAE codes

File Based Transfer: Import/Map/Export – Data exchange via files on a hard-disk – CAE codes need not be resident in memory – Often called “Loose Coupling” – User responsible for exchange synch

Socket Based Transfer: Co-Simulation API – CAE code and STAR-CCM+ in memory – Data exchanged via sockets – API controls exchange synchronization

Methods for Exchanging Data

16

Presenter
Presentation Notes
It is important to understand that STAR-CCM+ offers two methods for exchanging data. File-based tansfer and socket based transfer.

Coupling via Abaqus Co-Simulation API of SIMULIA – Manages Coupling Synchronization/Exchange/Mapping – Abaqus v6.12/STAR-CCM+ v7.04 (implicit coupling)

STAR-CCM+ Abaqus (explicit or standard) – Initial geometry – Pressure(relative or absolute pressure) – Shear traction – Surface heat flux

Abaqus STAR-CCM+ – Displacement, velocity – Temperature

Abaqus/STAR-CCM+ Co-Simulation

17

Abaqus/STAR-CCM+ Co-Simulation Interface

18

Hit the Step or Run button to commence the co-simulation

Presenter
Presentation Notes
The graphical interface is very familiar to STAR-CCM+ users.`

The Challenges of FSI

19

Coupling Technique

Two-way coupling for aeroelastic or hydroelastic equilibrium

One way static coupling for thermal stresses – structural temperature computed by CFD CHT

One way dynamic coupling – Loads only go from fluid to structure – Loads only go from structure to fluid

Two way dynamic coupling – explicit (exchange once per time step) – or implicit (multiple exchange per time step)

Degrees of Coupling

20

Dynamic Mesh Evolution

The Challenges of FSI

21

Presenter
Presentation Notes
A huge challenge in FSI is how to evolve the fluid mesh so that it follows the deformed structure

Completely define a good quality fluid mesh – Using as little user intervention as possible

Options within STAR-CCM+ – Rigid body motion, sliding interfaces – Morphing – User Defined Motion

Overset Mesh

Dynamic Mesh Evolution

22

Overset Mesh Motion and Fluid Flow : Stop Valve

The Challenges of FSI

24

Validations

AeroElastic Prediction Workshop: HIRENASD

2.3M cells

53K nodes

Static Structure, Steady airflow at deformed shape Ma=0.8, Re=23.5x106, q/E=0.48x10-6

Aerodynamic Equilibrium at different AOA

Wing Tip Displacement Lift Coefficient

Cp: AOA 2°

FUN3D

STAR-CCM+/Abaqus

Cp: AOA 2°, Station 7

x/c

STAR-CCM+/Abaqus NASA FUN3D

Wind-off Vibration Modes : Abaqus vs Experiment

f=25.55 Hz (26.25)

f=80.25 Hz (78.20)

f=106.20 Hz

f=160.35 Hz (165.25)

AeroElastic Prediction Workshop: Forced Vibration of 2nd Flap Bending Mode

Magnitude

Wing Tip Response of Cp

Phase

Fluid-Elastic Instabilities in a Tube Bundle

Weaver & Abd-Rabbo, “A Flow Visualization Study of a Square Array of Tubes in Water Crossflow”

31

HydroDynamic Damping: Simulation vs Measured Vibration RMS vs Inlet Velocity: Simulation vs Measured Tube Bundle Vorticity

Presenter
Presentation Notes
It is extremely important to accurately compute the fluid damping – it is critical to the prediction of the onset of the fluid-elastic instability. By varying the inlet flow in the simulation one can determine the onset and RMS of the limit cycle oscillations. The animations show the vorticity and deformations of the rods.

Co-Simulation DOT Tank Impact

32 Experimental Test Impact

Presenter
Presentation Notes
This test conducted at Sandia National Labs is, as you can imagine was quite expensive but resulted in only one condition. Simulation could be used to examine a range of impact conditions

DOT Tank Impact Simulations

33

Von Mises Stress (Abaqus Explicit)

STAR-CCM+ Pressure

Presenter
Presentation Notes
The abaqus model uses shells to model the double wall tank with contact between outer wall and impacter, outer wall and back barier and contact between inner and outer wall. STAR-CCM+ models the partially filled tank using VOF

34

DOT Tank Impact Comparisons

Presenter
Presentation Notes
The comparison to experiments show excellent agreement in the peaks, unloading, and duration of the impact event.

Mapping Data Techniques Procedures for Exchanging Data Coupling Methods Dynamic Fluid Mesh Evolution Validations

The Challenges of FSI

35

Thank You For Your Attention & Enjoy

36