ISSUE 19 Winter 2002/3

ISSUE 19 Winter 2002/3

1 STAR-CD Dynamics � Winter 2002/03

COMMENT

Accelerating design, reducingcost and improving your productsDidier Halbronn, Vice President, CD adapco Agency Operations

1. Reduce time to market

Comet-WorksTo enable CFD to be introduced earlier into the design process

we have developed Comet-Works, a reduced-functionality code,

integrated into SolidWorks, taking advantage of the latter’s state

of the art parametric modeling and intuitive environment. ‘What-

if’ CFD analyses have never been so easy! Comet-Works and

STAR-Works are the first of a series of codes to be developed for

all the major CAD environments.

es-toolsOur es-tools (expert systems) initiative captures rapid solution

capability for application-specific problems. Each tool provides

high quality automatic meshing, integration with the CFD solver

and special pre-/post-processing facilities. Available tools are:

es-aero (external aero-dynamics), es-uhood (underhood

thermal management with complex geometry), es-ice (moving

mesh for IC engines), es-turbo (rotating machines), es-fsi (fluid-structure

interaction), and es-4u (a more general tool for capturing your own methodology

and process). Next year, we plan to release additional tools such as es-

aftertreatment, es-pass. We will be guided by whatever you tell us will speed your

design process.

Parallel processingFinally, we help you to reduce your computing time. The parallel processing

offered by STAR-HPC applies not only to multiprocessor machines but also to

clusters of low cost boxes managed by our STAR-NET utility. For example, excellent

scalability is achieved on Linux clusters (see our last newsletter, issue 18).

2. Minimize your design costs

Easy meshing processesThanks to the native CAD readers (Catia, Unigraphics, Pro/ENGINEEER, etc.)

combined with pro-surf (the missing link between CAD and CAE/CFD), direct

geometry transfer, with repair and closed surface meshing becomes a quick and

easy automated process. pro-am (incorporating pro-surf) produces a 3D

automatic mesh for any of the CD adapco Group’s solvers, or other CAE packages.

The CD adapco Group is

committed to developing new

products and services to help

you achieve your business

goals. To help you reduce time

to market we have developed

three strategies.

STAR-CD Dynamics � Winter 2002/03 2

COMMENT

STAR-WorksAlternatively, you can use STAR-Works as a SolidWorks front-ended version of

STAR-CD, integrated as Comet-Works to take full advantage of state of the art

parametric modeling and intuitive user interface. The main difference however, is

that at any time you can move to the full STAR-CD user interface to access

unlimited CFD functionality. Once again, please let us know if you need such a

product implemented within a different CAD environment. To help you evaluate

STAR-Works, we offer a free animated demo disk, available on request.

Saving costs by accelerating designOur strategies to reduce your time-to-market (as already described) will also

reduce your design costs. So, Comet-Works, the es-tools and STAR-HPC are great

cost savers. Finally, we have modeFrontier, a uniquely powerful tool for cost-

effective multidisciplinary optimization.

3. Minimize your design costs

Parametric modelingMore what-if ‘fine tuning’ scenarios become practicable using the parametric

modeling capabilities of STAR-Works and Comet-Works.

Physical modelsYou also have access to the wide choice of physical models in STAR-CD, including

its aeroacoustics capabilities, its STAR-CD/KINETICS solver (integrated CFD with

stiff chemistry), fluid structure interaction capabilities and links with other CAE

systems: FEA, and 1D codes such as Flowmaster and GT-POWER.

Engineering ServicesFinally, on ‘the other side of the coin’, we offer the full range of benefits of the

Engineering Services provided by the worldwide offices of the CD adapco Group.

This is a summary of the main initiatives, which the CD adapco Group has

implemented to help you meet your business targets. We value your feedback very

much. So, please tell us what you need from us to help lead you to success.

On a personal note, I look forward to meeting many of you at our User Group

Meeting in London on November 25-26.

APPLICATION STORIES

AUTOMOTIVE

A cascade atomization and drop break-up model 6

CFD modeling and design optimization of a gerotor pump 7

Fuel cell analysis with STAR-CD 9

CORPORATE NEWS

SolidWorks solution partnership 3

Survey of your hardware preferences 3

New look website 3

French User Conference 3

CATIA 5 and STEP readers now available 3

CFD workshops 3

MOTORCYCLING

Cool designs at Harley-Davidson 11

TRANSPORT

Simulating silence with CFD 24

POWER GENERATION

Analysis of low NOx burner refit 19

The impact of CFD on the design of a PBMR 21

Gas flow in gas condensing towers 23

ENVIRONMENTAL

Forest fire management using STAR-CD 13

PRODUCT NEWS

News from the development department 4

The DESTINY poject 5

BUILDINGS

Microclimate study for conservation of Roman murals 15

Cigarette smoke distribution in a casino 17

NEW PUBLICATIONS 25


CORPORATE NEWS

SolidWorks solutionpartnership

Following the successful integration of

STAR-CD and Comet into the 'parts and

assembly' modules SolidWorks, and the

associated release of our "STAR-Works" and

"Comet-Works" products, we are pleased to

announce that we now have 'SolidWorks

Solution Partner' status. Previously, a number

of CAD companies have released limited-

capability CFD codes, and vice versa.

However, we believe that this is the first time

a leading CFD company has collaborated so

closely with a leading CAD company. The

advantages of using two best-in-class

products together

is clear. Every

industry has its

own expertise,

and at the CD

adapco Group we focus on our core CFD

business, since advanced CAD is available

elsewhere. Our customers will benefit from

the tight working relationship we have

confirmed in or partnership with SolidWorks

and its technical staff. Moreover, with our

open environment philosophy, users of

STAR-Works and Comet-Works who want to

go more fully into CAD technology should feel

free to purchase additional licences and

training for SolidWorks. If you need additional

help, such as, for advanced stand-alone

applications of the SolidWorks CAD system,

we can put you in contact with their extensive

local support offices.

STAR-CD French UserConference

The French User Conference 2002 was held on

September 16 2002 at the 'Cité des Sciences

et de l’Industrie La Villette', in Paris.

Presentations given by the CD adapco Group

described future and recent developments

including the STAR-CD's associated Expert

System es-tools, and our more recent product

STAR-Works. The user presentations covered

a variety of applications from clients across

different industry sectors including Renault,

CERN, ENSAM PAris, GE EPE, DCN Indret and

Flowmaster.

Survey of yourhardware perferences

We have a policy always to port STAR-CD

promptly to the latest machines and improved

operating systems. Our aim is to be led in our

porting priorities by your computer

preferences. In order to help us provide you

with the best possible support for STAR-CD,

we recently sent a questionnaire to all of our

licensees, asking about current hardware

usage and future plans. Interestingly we

found that PCs are now more numerous than

Unix, although for number crunching, parallel

UNIX machines continue to provide the

computational backbone for most large

companies. In the PC area, both Linux and

Windows are strong, but your growth plans

are much higher for Linux than any other

operating system. We also asked you about

your preferred version of Linux. We found

Redhat and SuSE to be equally popular, but

with the former stronger in Europe and the

latter stronger in the USA. Concerning parallel

computing, we noted that 80% of you use (or

plan to use) STAR-HPC, and that the average

case size for all users peaks very strongly at

500,000 to 1 million cells, with this applying

equally to both Linux clusters and to

massively parallel UNIX machines. We thank-

you for your feedback and confirm our

commitment to supporting you, whatever

hardware you prefer.

New look website

You may have recently noticed that our

corporate website (www.cd-adapco.com) has

undergone a face lift. This is all part of our

programme to continuously improve its

design and make it more intutive to explore.

Our new drop-down menus enable easier

navigation and the new html news section

makes it easier for you to keep upto date with

STAR-CD. Keep visiting!

Catia 5 and STEPreaders now available

The following readers are now available,

enabling effortless importation of CAD

models into STAR-CD's pro-am automatic

mesher:

● Unigraphics

● Catia 4

● Catia 5

● STEP

● Pro/E

● SolidWorks (integrated in STAR-Works)

CFD workshops

The CD adapco Group regularly hosts

seminars, workshops and specialized training

to show how the best possible industrial

solutions can be achieved with STAR-CD and

its associated tools. Examples of upcoming

workshops include: CFD for the marine

industry, CFD for domestic appliances,

General CFD & STAR-Works, CFD for diesel

engines, CFD for the turbomachinery industry,

CFD for the chemical process industry, CFD in

pumps and compressors, Aeroacoustics and

more…..

For more information, visit:

www.cd-adapco.com/support/CDAevents.htm


PRODUCT NEWS

5

ontinuing the series started in the previous

issue, we present another new feature of

our future software products which

demonstrates their extension from pure CFD to

a wider CCM (computational continuum

mechanics) scope: simulation of phase change.

This feature is now being tested as a "beta

product" and will be released in the near future.

The need to simulate processes involving

phase change is obvious: the melting of solid

materials or casting followed by solidification

are typical examples of engineering

applications. The range of problems to be

solved is vast; typically, flow of molten material

and species transport, temperature distribution

in both solid and liquid, and deformations and

stresses in solid parts are of interest. It would

be of great benefit to industry if such

phenomena can be computed.

The approach adopted for our software

development is based on the pioneering work

by professor Ismet Demirdzic (now Senior

Corporate Consultant of the CD adapco Group),

who started more than a decade ago extending

the Finite-Volume method used in CFD to

computation of deformations and stresses in

solid bodies. The technology is now so mature

that it can be incorporated into a general-

purpose commercial code using arbitrary

polyhedral control volumes (see the report in

the preceding issue) and put to practical use by

the engineering community.

The methodology used to compute phase-

change problems is based on the so-called

fixed-grid enthalpy approach. In subdomains

which remain solid all the time, only the energy

equation and the equations for displacements

are solved; the solid material is at present

assumed to be thermo-elasto-visco-plastic. In

subdomains where both liquid and solid

fractions may be present, both equations for

velocities of the liquid phase and for

displacements of the solid phase are solved and

which parts are in which state is defined by the

liquid volume fraction. Here the interface-

capturing scheme used for computing free-

surface flows is employed to trace the interface

in the case of existence of the third, gaseous

phase, while the interface between solid and

liquid is determined by as a non-linear function

of temperature. The momentum equations

contain an additional source term, representing

the so called mushy region resistance tensor.

As the liquid fraction solidifies, gaps

appear between solidified material and the

mold wall due to shrinkage. These gas gaps

affect the heat transfer substantially and are

taken into account as additional resistance.

Also, due to deformations of the solidified

material, contact problems need to be taken

into account. More details concerning the

method can be found in a recent publication.1

The method now implemented

incorporates all these features, and work is in

progress to model the formation of macroscopic

porosity and to compute residual stresses in the

solidified material. Other constitutive relations

for solid-body deformation and interaction with

electro-magnetic fields are areas for future

development. This will make our future CCM-

software applicable to simulations of such

complex multi-physics problems as welding

and casting, among others. The figures on the

right demonstrate the functionality of the new

methodology for the case of solidification of

molten tin in a rigid sand mold. Applications to

real industrial problems are now under way and

the results will be reported in the near future on

our web site.

1 A. Teskeredzic, I. Demirdzic and S. Muzaferija: Numerical

method for heat transfer, fluid flow, and stress analysis in

phase-change problems, Numerical Heat Transfer, Part B, Vol.

42, pp. 1-23 (2002)

News from the development departmentMilovan Peric, Director of product development operations, CD adapco Group

C

Fig. 4: Distribution of effective stresses in the solidified cast, showing

(magnified) gaps and stress concentrations near corners.

Fig. 1-3: Distribution of the liquid fraction and velocity vectors during

solidification of tin in a rigid sand mold. Note that the magnitude of

velocity reduces substantially as the solidification progresses.

ONGRATULATIONS, with STAR-CD you

have chosen the best software for

Aeroacoustics source predictions, with proven

efficiency of transient solution, accuracy of

discretization and advanced turbulence

modeling, fit for purpose. It has been validated

against experimental data, integrated within a

process tool-kit for acoustics analysis, and is

now beginning to be used widely among

STAR-CD users.

At the start of this year, recognition of the

need for industrial tools to predict flow noise

led to the formation of a CD adapco industry

sponsored strategic partnership. The project,

named DESTINY-AAC (Detached Eddy

Simulation for the Transportation INdustrY –

AeroACoustics), targets ambitious but

realizable goals. These are the provision of

robust LES-type turbulence and advanced

near-wall modeling, and their validation

against a wide variety of aeroacoustics

examples, which are currently a high priority in

the transport sector. This includes noise from

HVAC systems and air delivery, external flow

including A-pillar and wing-mirrors, fans,

cavities, pantographs and high-speed

elevators.

Complementary to the provision of

advanced turbulence modeling, noise

propagation to the near and far field is realized

through coupling to new aeroacoustics

features in LMS International’s software

SYSNOISE, which imports STAR-CD transient

data and compiles equivalent sources.

DESTINY-AAC has completed the first of

two stages. The implementation and functional

testing of Detached Eddy Simulation (DES) [1]

and hybrid-wall functions are finished. The

project now enters its testing phase,

performed by Air International, Audi, BEHR,

Bombardier, DaimlerChrysler, Denso and

Sulzer Innotec on core industrial applications.

Preliminary results have confirmed

expectations that the advanced models are

robust, efficient and strongly suited to

aeroacoustics. The DES implementation in

STAR-CD is valid for the k-ε, k-ω and Spalart-

Allmaras turbulence models and their variants.

These RANS models in unsteady mode are

active in the near-wall region and transition

smoothly to a Smagorinsky-like sub-grid scale

model away from surfaces. The hybrid wall-

function, which is valid for the full low and

high-Reynolds number y+ range, relaxes the

tedious necessity for the user to produce

appropriate near-wall mesh spacing.

Applied to high-speed cavity acoustics,

these new DES modeling techniques

demonstrate real

benefits compared with

traditional unsteady RANS or

LES turbulence modeling. RANS modeling, due

to over-prediction of eddy-viscosity

underestimates the broadband noise content

or incorrectly predicts narrowband modes. LES,

in contrast, greatly improves the broadband

prediction, but requires very fine near wall

resolution or else the low-frequency noise

content is over predicted.

For further details, please contact: [email protected]

[1] Spalart, P.R., Jou, W.H., Strelets, M. and Allmaras, S.R.,

"Comments on the Feasibility of LES for Wings, and on a

Hybrid RANS/LES Approach,", First AFOSR International

Conference on DNA/LES, Ruston, Louisiana, USA. 1997

[2] Henshaw M.J de C., "M219 Cavity case in Verification

and Validation Data for Computational Unsteady

Aerodynamics", RTO-TR-26, AC/323(AVT)TP/19, October 2000


PRODUCT NEWS

Aeroacoustics DESTINY-AAC industrial projectFred Mendonça, Agents’ support and consultancy manager, CD adapco Group

C

A2 side-view mirror, courtesy of Audi AG

HV

AC

blo

we

r y+

distrib

utio

n, co

urte

sy o

f De

nso

Th

erm

al S

yste

ms

HVAC box and flap, courtesy of BEHR

PSD at a point along the cavity ceiling - DES

compared with unsteady RANS and LES for high

speed cavity acoustics. Data [2] from QinetiQ


AUTOMOTIVE APPLICATIONS

he breakup of liquid fuel jets in diesel

combustion engines plays a decisive role

in the evolution of the spray and its

subsequent processes; it has a direct

influence on an efficient and clean

engine operation. Recent

investigations conducted by various

researchers, utilizing different

experimental techniques, show that

transient, high-pressure-driven fuel

jets are broken into liquid fragments

of various shapes and sizes at the time

they exit the injector nozzle or shortly

thereafter. Subsequently, these liquid

fragments are subject to aerodynamic

forces, which lead to further

breakups until the droplets reach a

stable state. The fundamental

mechanisms responsible for the

aerodynamic breakup are either the

Rayleigh-Taylor or Kelvin-

Helmholtz instability on the

liquid/gas interface.

The Enhanced Taylor

Analogy Breakup (ETAB)

model simulates this liquid

jet disintegration process

as a cascade of drop

breakups. The breakup

criterion is determined by

Taylor's linear drop

deformation dynamics and

the associated drop breakup

condition. Breakup occurs

when the normalized drop

distortion exceeds a critical value. The

breakup into product droplets is

modeled after the experimentally observed bag

or stripping breakup mechanisms and the

radial velocities of the product droplets

are derived from an energy

conservation consideration.

At the nozzle exit, the liquid jet is

simulated as a sequence of large,

high velocity drops which are very

unstable. In order to avoid an

immediate breakup, they are

assigned a deformation velocity

such that their lifetime is extended

to match experimentally observed

jet breakup lengths. This

computational artifice leads to the

simulation of a fragmented liquid

core, as reported by various research

groups. An additional benefit of this

initial breakup delay is the radial

velocity of the product droplets at

first breakup, which results in an

automatic adjustment of the

spray cone angle to changes in

the gas density. On the other

hand, the model requires an

initial drop size distribution

in order to compensate for

the neglect of the surface

stripping near the nozzle

exit. This phenomenon

determines the fuel-air

mixing near the nozzle exit,

and has a strong influence on

the ignition location. The

performance of the ETAB model

has been compared with the WAVE

model, as implemented in STAR-CD,

and with measurements obtained under

controlled conditions from a constant volume

bomb. The simulations showed good overall

agreement with experimental data, especially

the drop sizes were well predicted. In addition,

the amount of model tuning for a particular

injection condition is considerably reduced due

to the automatic adjustment of the spray cone

angle to the changes in the gas density.

Ossi Kaario, Technical Research Centre of Finland, Martti LarmiHelsinki University of Technology, Finland, Franz Tanner, Michigan Technological University, USA

A cascade atomization and drop breakup model in STAR-CD

T

Spray penetrations for the ETAB and WAVE

computations compared with experimental data

Drop sizes expressed as Sauter mean diameter

(SMD) for the ETAB and WAVE computations

compared with experimental data



erotor pumps are widely used in

the automotive industry for fuel lift,

engine oil and transmission systems.

Volumetric efficiency and cavitation damage,

are causes for concern in gerotor pumps with

high output flow. To optimize pump

performance and reduce cavitation damage, it

is essential to understand the fluid dynamics

inside the pump.

In a gerotor pump (Fig. 1), the fluid is

sucked into the inlet port and shifted to the

outlet port. Due to the rotor clearances (Fig.

2), flow leakage occurs between the high-

pressure and low-pressure sides of the pump.

To limit pressure, excess fluid is re-circulated

to the inlet port through a pressure relief

valve. The flow through the rotor clearances

creates high fluid velocity and localized low-

pressure areas, which produce air and vapor

bubbles hence causing cavitation damage

and noise.

CFD analysis can be used as a cost-effective

design tool for the optimization of pump flow

performance and reduction of fluid borne

noise [1]. In order to optimize the design of

gerotor pumps, a realistic CFD model is

required which takes into account gear

meshing, leakage flow across clearances and

cavitation bubble formation, recompression

and collapse [2,3].

A full 3-D transient model with moving

and deforming boundaries has been

developed specifically for gerotor pumps (Fig.

3). This model can predict cavitation bubble

formation, recompression and collapse, by

realistically modeling the dynamics of gear

rotation, meshing and sliding over the inlet

and outlet ports, and flow leakage through

the rotor set clearances. The grid generation

and time dependent manipulation has been

carried out using PROSTAR, the pre-processor

of STAR-CD. The mesh motion/rotation and

deformation of the pumping chambers has

been defined using a "script". Arbitrary

sliding interfaces have been used to connect

the rotating pumping volumes with the

stationary inlet and outlet ports.

Extensive sensitivity studies on grid

density and distribution, and time stepping

have been carried out to meet the accuracy

criteria. To be able to incorporate design

improvements from the CFD analysis into the

pump design process, calculation time had to

be kept to within a few hours.

A CFD analysis of the flow inside gerotor

pumps has been conducted for the design

optimization of fuel lift pumps. Similar

analyses could be conducted for gerotor

pumps with smaller or larger rotors, and a

wide range of rotor clearances, pump speeds,

and fluid viscosities. A preliminary validation

of the CFD calculations has been carried out

in [2]. The CFD results of the average delivery

flow rate and its fluctuation (flow ripple)

were compared with the experimental

measurements at the different speeds and

pressures. In general, the CFD results of

the delivery flow rate were in good agreement

with the experimental data

and well within the

e x p e r i m e n t a l

scatter. The flow

ripple is over-

predicted due to

the assumption

that the flow was

incompressible and

because leakage was

over-predicted. CFD results for velocity vectors

and pressure distributions are shown in Figs 4

and 5 at different rotation angles in the pumping

cycle.

CFD modeling and design optimization of a gerotor pumpF. Iudicello, Pump Products Division, DANA Engine and Fluid Management Group

G

fig 1: CAD model of a diesel fuel lift gerotor pump

fig 2: Gerotor pump rotor set clearances



A CFD design optimization has been carried

out using a design of experiment (DOE)

method to maximize the pump volumetric

efficiency, and minimize cavitation damage

and noise [3]. The effect of porting geometry

and rotor clearances has been investigated.

The DOE analysis of the CFD results has

shown that the most important parameters for

the flow performance are the rotor clearances

and the presence of the inlet grooves. The

most important variables for the flow

fluctuations RMS are the tip-to-tip

clearance, the inlet minor sealing

angle and the outlet major sealing

angle. Experimental verification of

the CFD optimal design has shown a

reduction of cavitation damage to

an acceptable level. Further

development work will be required

to include compressible flow and

cavitation in the model.

References

[1] Iudicello, F. and Baseley S. CFD modeling of the

flow control valve in a hydraulic pump. PTMC 1999,

ed. Burrows, C.R. and Edge, K. A., University of Bath,

1999, pp.297-312.

[2] Iudicello, F. and Mitchell D. CFD modeling of the

flow in a gerotor pump. PTMC 2002, ed. Burrows, C.R.

and Edge, K. A., University of Bath, 2002, pp.53-66.

[3] Iudicello, F. CFD modeling and design optimization

of a gerotor pump. Eighth European Congress on Fluid

Machinery for the Oil, Gas & Petrochemical Industry,

The Hague, 31 Oct- 1 – 1 Nov 2002.

fig 3: CFD model of the

diesel fuel lift gerotor pump

fig 4: Velocity vectors at different rotation angles in the pumping cycle

fig 5: Pressure distributions at different rotation angles in the pumping cycle



A Proton Exchange Membrane Fuel Cell

(PEMFC), also known as a polymer electrolyte

membrane fuel cell, operates at lower

temperatures than other types of fuel cells

and is a strong candidate for use in future

automobiles and stationary power

generators. These uses will require many

single cells to be connected with bipolar

plates; but understanding how geometry and

operating conditions affect the performance

of single cells is the first step towards high

performance designs. Engineers at the CD

adapco Group and the University of South

Carolina have produced a module for STAR-

CD that can now be used to understand single

cell performance. This article is the first in a

series to highlight how CFD tools and this

module can be used to explain the behavior

of a PEMFC.

Fig 1 shows that components of a PEMFC

include two distinct flow channels separated

by the Membrane and Electrode Assembly

(MEA); one channel is for the anode gas

mixture and the other is for the cathode gas

mixture. In between the flow channels and

the MEA, there are diffusion layers on both

sides of the MEA. These layers are porous to

allow for distribution of the gases to

unexposed areas of the flow channel and this

distribution allows for complete utilization of

the electrode area.

The gases flow along the channels and

diffuse toward the MEA where the

electrochemical reactions occur. On the

anode side of the MEA, hydrogen is oxidized

to protons and electrons as shown in Fig. 2.

The electrons flow through the load

device and to the cathode where they react

with oxygen and protons to form water. The

critical aspect of a PEMFC is that the protons

require water in the present-day membranes

to facilitate transport to the cathode. Water

moves from the anode to cathode with the

protons through a mechanism known as

electroosmotic drag. If the membrane is thin

and the concentration of water on the

cathode is higher than on the anode, water

can move through the membrane from

cathode to anode by diffusion. Recent

advances in polymers and composite

membranes allow for very thin membranes

with maximum structural integrity to be used

in PEMFCs.

The key to optimum performance is a

membrane that is wet enough for maximum

conductivity but dry enough around the

electrodes so that the transport of gas is not

limited.

Fig 3 shows the experimentally measured

current for a laboratory scale cell operated at

a voltage of 0.6 V and a temperature of 70 oC

at 202 kPa on both the anode and cathode

sides. Currents for 10 hrs are shown for four

levels of inlet humidity. At lower humidifier

temperatures (65/55oC and 75/65oC), the

membrane is not sufficiently wet and the

current oscillated in a very unstable manner.

The amplitude of the oscillations decreased

when the humidifier temperature was

increased and the maximum current occurred

for humidifier temperatures of 85oC at the

anode and 75oC cathode (i.e., 85/75oC).

When the temperatures were raised to

95/85oC, the current was very stable but the

measured current was lower. We attribute

the lower current at the higher temperatures

to water accumulation on the cathode side of

the MEA. CFD can predict this accumulation.

Fig 4 shows the geometry and detailed

mesh created for the with the gas channel

flow-field plate used in the laboratory single

cell. This model also has a thin membrane

that is sandwiched between anode and

Using STAR-CD to explain PEM

fuel cell behaviorS. Shimpalee and J. W. Van Zee Center for Fuel Cell Research, University of South Carolina, USA

Fig 1: Schematic of single PEMFC assembly displaying different essential components of the system



cathode

d i f f u s i o n

layers. The

reaction area

of this single

PEMFC is 10 cm2.

There are twenty serpentine

passes in the flow path and we

used 313,400 computational cells.

Figs 5a and 5b show predictions of

local current density on the membrane

surface at conditions of 85/75oC and

95/85oC humidifier temperatures,

respectively. For 85/75oC, the current density

decreases along the flow path toward the

outlet due to the lower anode water activity

and the depletion of reacting gases. For this

condition, the current density varies from 0.80

A/cm2 to 0.34 A/cm2. There is a non-

uniformity in electrochemical reaction on the

catalyze layers of MEA and this will affect the

temperature distribution inside PEM fuel cell.

Fig 5b shows the local current density

distribution

on the membrane

surface for 95/85oC

inlet humidity condition.

Here, the local current density also decreases

from the inlet toward the outlet but the values

are lower than for the case shown in Fig 5a.

The maximum current density in Fig 5b is

about 0.68 A/cm2 and the minimum is 0.28

A/cm2. This is because higher humidity

generates membrane flooding on the cathode

side and also in the first channel of the anode.

The flooding creates higher resistance for

access of the oxygen and hydrogen to the

catalyst on the

m e m b r a n e

surface. Other

profiles are

available by

post processing

the results from the

STAR PEMFC module and

these profiles will be discussed in future parts

of this series. Documentation for the

equations can be found on our web site.

www.che.sc.edu/centers/PEMFC/index.html

Fig 2: Schematic of water transport and

electrochemical reactions in a PEMFC

Fig 4: Geometrical m

odel of complete PEM

FC shown without the

graphite current collector

Fig 3: Humidity effects on PEMFC performance at 70oC cell

temperature, P(A/C) : 202/202 kPa, flow rate (A/C): 76/319

cm3/min, and cell voltage: 0.6V

Fig 5a: Current density distribution for condition of 85/75oC

humidifier temperatures

Fig 5b: Current density distribution for condition of 95/85oC

humidifier temperatures

Cool designs at

Harley-Davidson Motor company

he Harley-Davidson Motor Company

manufactures heavyweight motorcycles

and offers a complete line of motorcycle

parts, accessories, apparel and general

merchandise. These products exemplify a

distinct look, sound and feel which is closely

tied to the company’s heritage. Harley-

Davidson is firmly committed to preserving

and enhancing its unique heritage. To

achieve this goal, remain competitive in the

market place, and meet future customer and

regulatory requirements, Harley-Davidson

powertrain engineers incorporate computer

aided engineering (CAE) tools into the

product development process.

CAE tools are the key to

shorter development

time, reduced development cost, and

improved product quality. Computational

fluid dynamics (CFD) is one of the CAE tools

Harley-Davidson powertrain engineers use to

meet these challenges. This article provides

an overview of CFD usage by Harley-Davidson

engineers.

The air-cooled V-Twin engine is the

traditional soul of a Harley-Davidson

motorcycle. The use of air-cooling is an

integral part of the vehicle’s character – "the

look". With ever increasing power density,

powertrain thermal management is a

challenging task. It is also a very important

part of the product development process.

Traditional empirical methods require testing

entire vehicles either in a wind tunnel or on a

test track. Whole vehicle testing requires

considerable expenditures to prepare and

run the experiments. The quality and quantity

of information generated by these

tests is typically limited by

instrumentation constraints.

In addition, test track

ambient conditions rarely

seem to cooperate with

product development

schedules. To reduce

costs and improve the

final product, Harley-

Davidson engineers use

CFD models to address

powertrain thermal management.

Recent publications indicate that the

automotive industry is regularly using CFD

for powertrain (under hood) thermal

management. An air-cooled motorcycle

yields unique challenges not seen by the

automotive industry. Consequently, the first

step for Harley-Davidson engineers was to

identify the modeling strategies required to

properly solve these challenges. Figure 1

illustrates the results of an early

investigation into the meshing and solver

parameters required. The particular example

shown is an extruded fin array similar to

those used as heat sinks in the electronic

industry. Figure 2 shows an excellent

correlation between measured and predicted

metal temperatures. Comparing the

predicted results to test data provided

feedback regarding the suitability of different

modeling strategies. Lessons learned from

such simple exercises were carried forward

into the analysis of more complicated

real world problems. As a result of this

methodical approach, Harley-

Davidson engineers can now analyze

engine designs before costly

prototypes are produced.

Thermal management activities are

not limited to air-cooling. The V-Rod

motorcycle represents the fusion of

traditional Harley-Davidson styling with

MOTORCYCLE APPLICATIONS

T

Paul Troxler, Harley-Davidson Motor Company


liquid-cooled, contemporary

performance to create a new

family of power infused

custom motorcycles. The

Revolution engine powering

the V-Rod is Harley-Davidson’s

first mass production water-cooled

powertrain. During the development

program, CFD was used to analyze water

flow rates and heat transfer coefficients in

the Revolution engine’s water jackets. Figure 3

illustrates an example of the results generated

during this project.

The lubricating oil is another liquid flow

system Harley-Davidson engineers use CFD to

evaluate. The pipe flow portion of the lubrication

supply system is modeled using a one-

dimensional (1-D) "network" analysis program.

The 1-D program is an efficient tool for

evaluating system operating pressures and flow

distributions. Not all lubrication system

components meet the assumptions made by the

1D program. The flow characteristics of these

components are evaluated using three-

dimensional (3-D) CFD. The results are

parameterized for use by the 1-D code. Although

it is possible to couple the two codes together,

Harley-Davidson engineers have not found a

need to

use that

capability.

The present

approach allows

rapid evaluation of

proposed designs concepts

and changes.

The multi-phase nature of oil

splashing in the crankcase is treated

using three-dimensional CFD. Spray and

droplet breakup sub-models are used to track

oil particle generation. The Lagrangian/Eulerian

framework predicts the motion of a dispersed

phase (oil particles) within a continuous phase

(air). Simultaneous solution of the energy

equation allows for heat transfer between the oil

and the metal parts to be evaluated. Moving grid

capability (cell addition and deletion with vertex

motion) allows the motion of the pistons to drive

the flow within the crankcase. Incorporating CFD

gives Harley-Davidson engineer’s the capability

to separate and independently study different

aspects of the problem. This capability was not

possible with the traditional empirical

techniques.

Harley-Davidson uses CFD in the develop-

ment of the power cylinder components.

Evaluation of intake and exhaust port flow

coefficients is accomplished without the need to

manufacture costly flow boxes and run time

consuming tests. An automated mesh generator

and existing solid models make the

computation of port flow curves a nearly

automated process. Final engine performance

predictions are made by incorporating the port

flow curves into a cycle simulation code. Using

the same techniques, the flow characteristics

of induction and exhaust system components

are evaluated. After static flow

comparisons of several

designs, the best are

evaluated for engine

performance by

coupling the three-

dimensional CFD model

with a cycle simulation code.

Complete three-dimensional

CFD calculations of the entire

power cylinder are used to evaluate

engine performance. These models

combine advanced capabilities such as moving

mesh with fuel spray and combustion sub-

models. This type of analysis provides

information regarding mixture preparation,

combustion efficiency, heat transfer, and

emissions formation. Figure 4 presents a power

cylinder model sample result.

Incorporating CFD

into the product

development process

allows Harley-Davidson

Motor Company engineers to improve product

quality while simultaneously reducing

development time and cost. The resulting

products incorporate advanced technologies

while maintaining the corporate heritage.

fig 4:

Predicted

Combustion

Gas Temperature

fig 1: Predicted metal temperatures

for an extruded fin array


MOTORCYCLE APPLICATIONS

fig 3: Predicted water cooling

system flow velocity

fig 2: Comparison of predicted and measured extruded fin

array metal temperatures

Forest fire management using STAR-CDDr. Stephen T. McIlwain and Kevin Jones, CD adapco Group

ENVIRONMENTAL APPLICATIONS


orest fires cost the United States millions of

dollars per year in property and land loss

and were responsible for the deaths of

numerous firefighters and civilians last year. In

addition, this year, there have been highly

publicized aviation disasters that have occurred

in attempting to extinguish the infernos.

What if the likely ignition locations of these

fires could be predicted? What if the spread of

these fires could be predicted and therefore

controlled before becoming uncontrollable?

Unusually for a CFD challenge this is not about

reducing design costs, nor about meeting

production deadlines: it is, however, about

protecting the environment and saving property

and lives where the deadlines are even

more critical.

With this in mind, it is no surprise that the

use of advanced analysis tools such as STAR-CD

are being used to help provide an insight into

such a damaging and life threatening

phenomenon.

Leading the research and development of

forest fire management are the team of

engineers and scientists at the Forest Service’s

Fire Sciences Laboratory in Missoula, Montana.

They have recently acquired STAR-CD to further

advance their already comprehensive numerical

modeling techniques and gain a better

understanding of how to contain and minimize

forest fire spread. Their facility is equipped with

an instrumented wind tunnel burn-rig, capable

of burning a range of solid fuels in a controlled

environment and generating valuable data. The

study of fire growth in the wind tunnel can then

be compared to predictions performed using

STAR-CD in an effort to correlate combustion

and turbulence models required to predict this

chaotic flow domain.

On the macro-environment scale, analyses

carried out by the Forest Service using STAR-CD

have involved the study of the flow of large air

masses over topo-

graphical domains. Using

satellite-based digital elevation

map (DEM) data, the landscape

surface can be read automatically into

STAR-CD to provide an immediate shell

representation of the ground surface. The

CD adapco Group worked together with the Fire

Sciences lab to develop this automated process

to build the computational model based on the

DEM data. Once the boundary conditions have

been added, which represent actual measured

wind velocities in both 3-D spatial and temporal

reference frames, the analysis can be run and

results post-processed all within STAR-CD.

The ability of STAR-CD’s modular

framework to interact with third

party codes and in-house

software has been a

major boost to the

Scientists and Engineers

at the Fire Research Center.

Using STAR-CD as the backbone

of their computational effort, the

investment in their existing in-house codes is

protected by linking these directly to STAR-CD

to produce a dedicated framework suited to

highly specific problems. This approach is

exemplified by the creation of dedicated user-

panels to fully automate the analysis set-up,

execution and post-processing.

Future work will include the analysis of

local buoyancy effects due to fire

progression and also the impact of

solar radiation on mountain

slopes, which can act as

a catalyst in the spread

of vegetation fire.

F

PROSTAR surface representation of 7 mile square DEM

Pressure contours for 9 m/s wind speed

Velocity magnitude contours

User defined panel

ISTeC (Interdepartmental Centre of

Science and Technology for Historical and

Architectural Conservation) has investigated

thermofluid dynamics phenomena inside

Domus Aurea in Rome, to understand

the causes of decay of the ancient and

precious mural paintings (fig. 1). Domus Aurea

was Emperor Nero’s residence (first century

A.D) and is now completely below ground

level. Based on previous studies on several

historical hypogeum monuments (e.g.

Hebrew Catacombs at Torlonia Villa in Rome,

Catacombs in Tarquinia), CISTeC defined

the optimal conditions for the conservation of

paintings: air temperature: 17°C (± 2°C),

relative humidity: 92% (± 5%), air velocity <

0.4 m/s.

With STAR–CD we studied the

microclimatic changes (e.g. wind velocity, air

temperature and its relative humidity) inside

the rooms of Domus Aurea and how they

differed from the above optimal values.

These simulations have been supported

by experimental monitoring of the

environmental parameters: air temperature,

contact temperature of the walls, relative

humidity of the air, wind velocity.

Simulations were especially concerned

with the so called "Room with the Golden

Vault" (fig.1), where openings in the masonry

produced turbulent air flow.

Geometrical model of the room

The geometrical model of "Room with the

Golden Vault" is 1:1 scale, the volume of air

storage capacity is equivalent to 813.10 m3.

Environmental data andboundaries

We chose the microclimatic data for the

months of January and August 2000, because

they are the coldest and hottest periods of the

year in Rome. In figure 2a the original case is

shown with environmental average conditions

(measured parameters). Table 1 reports the

data for the case of August 2000.

Other simulations of the original case were

carried out changing only the wind velocity:

the maximum and the minimum values of

wind were chosen from the experimental

measurements.

Finally, simulations were repeated using a

modified condition where the openings in the

masonry were closed (fig. 2b). We will show

how this proposed modification brings

potential benefits.

Results – August 2000

The three values of the inlet for this month

are: average=0.09 m/s, minimum=0.03 m/s,

maximum=0.73 m/s.

There are no significant differences between

the summer and winter cases for what

concerns fluidodynamic and thermo-

hygrometric distributions. It goes without

saying that the absolute environmental

values change.


BUILDING APPLICATIONS

Microclimate study helpsconserve Roman muralsSabatino Albero, Carlo Giavarini, Maria Laura Santarelli, CISTeC - Interdepartmental Centre of Science andTechnology of Historical and Architectural Conservation, University of Rome "La Sapienza", Italy

C

Eastern door 0.09 292 98 0.0135Vault openings 0.01 291.5 99.5 0.01286Southern opening 0.01 291.5 99 0.01322Noth opening 0.1 292 99.5 0.01371

Walls (*) 291 ÷ 292

Room zones Velocity (m/s) Temperature (K)Relative Humidity (%)

Specific Humidity (kg/kg)

Tab. 1 – Average conditions for a summer month in Domus Aurea (August 2000)

Fig. 1: Mural paintings inside Domus Aurea (Room with Golden Vault)

Fig. 2 : Geometrical model of the

"Room with Golden Vault" with the

(a) inlet for original case and (b)

modified case with closed openings

a

b

Fig. 3: average velocity, original case

(August 2000): (a) velocity

map, (b) inlet section of

thermal distribution

The minimum

values of air wind

velocity is 0.26 m/s and the maximum value is

0.96 m/s. In Table 2 the maximum data of air

velocity inside the room are shown:

The optimal value for the indoor air

velocity (0.4 m/s) was exceed only in the

maximum velocity case. In the other cases,

the critical zones are: vault, pillars and

western wall (fig. 3a). In summer the

temperature difference between walls and air

is lower than in winter. In summer the

difference is 0.5°C with air temperature in the

range 18.5 - 19 °C (fig. 3a), while in winter the

difference is 2.1°C with air temperature in

the range 7.6 - 9.7°C.

An isothermal profile of the inflow air is

shown in fig. 4. The hygrometrical trend

follows the winter case, but the specific

humidity value is double. The higher wet

zones are the north wall and the vault (fig.

5a). In the modified case the microclimatic

conditions are better and close to the optimal

values (fig. 5b).

Conclusion

In the original cases, the simulations led to

the following results: The inflow air damages

the masonry season in both seasons, but for

different reasons:

● In winter the optimal value is never reached

● In summer optimal temperatures are

achieved but the humidity is too high

In the modified masonry cases, simulations

show that:

● Microclimatic parameters are distributed

more homogeneously

● The room is more insulated from the

external environment

It is therefore easier to reach optimal values

for air temperature and relative humidity and

so guarantee the conditions for the

conservation of the mural paintings.

“CISTec of Rome”, says Anthony

Massobrio of CD adapco (Turin Office), “is a

leading example on how to merge art and

technology. What archaeologists discover,

CFD studies help to preserve.”


BUILDING APPLICATIONSC

ase

Average Vav

inlet (m/s)

Original case(with openings)

Modified case(closed openings)

Minimum vmin Maximum vmax

vi,max = 0.39 vi,max = 0.24 vi,max = 0.84

vi,max = 0.35 vi,max = 0.23 vi,max = 0.75

Fig. 4: Isothermal

distribution of the inflow air

at18.9C (maximum velocity,

original case – August

2000)

Tab.2 – Maximum values for the air wind velocity inside the room (August 2000)

a

b

Fig. 5:

Hygrometrical

distribution

(average velocity,

August 2000): (a)

original case and

(b) modified case

with closed

openings

a

b

Analysis of cigarette smokedistribution in a casinoJan-Hendrik Grobler, CFD Centre, CSIR, South Africa

TAR-CD has recently been used by the

CSIR (Council for Scientific and Industrial

Research) to analyze the airflow on the main

floor of a casino complex to determine the

distribution of cigarette smoke particles.

Strict laws aimed at curbing smoking in

public venues have recently been

promulgated in South Africa. Many

businesses had to make significant structural

changes to their buildings in order to comply

with the new laws, which generally require

that designated smoking areas be completely

separated from non-smoking areas by means

of full height partitions.

The STAR-CD model of the main floor of

the casino complex enabled ventilation

engineers to investigate the possibility of

confining smoke to certain areas with

effective ventilation and partial partitions

instead of full height partitions which would

harm the aesthetical appeal of the venue.

The engineers were also interested in the

effectiveness of the ventilation system in

removing smoke from smoking areas in the

event that they had to be separated with full

height partitions. The STAR-CD model was

therefore adapted to accommodate such a

scenario as well.

The grid presented a modeling challenge

due to the many geometric details that had to

be captured. The ability of PROSTAR

to generate three-dimensional

cells by extruding two-

dimensional shell cells came

in very handy. A floor plan

detailing the position of the

machines was used to

generate a similar picture in

PROSTAR consisting of shell

cells, via the PATCH

command. These cells were then

extruded in the vertical direction to form

solid cells to represent four different types of

machines as well as fluid cells representing

the airspace (see figure 1).

The roof detail was also complex, but had

a repetitive pattern and the relevant

commands in PROSTAR made it

possible to generate the complete

set of shell cells from one

segment. Due to the

difference between the

roof and floor detail, the

two sections were joined by

means of an arbitrary

coupled interface, situated

approximately 4 meters from the floor level. If

the two sections could not be joined in this

way, it would have been virtually impossible

to generate the grid with hexahedral cells

alone, and grid development would have

taken far longer.

The remaining part of the grid consists of

the passage surrounding the

main floor and includes

several domes. This part

of the model was

generated using blocks

and is shown in figure 2.

Figure 2 also shows the

location of some of the

boundaries, which were

automatically applied by

running a script of

PROSTAR commands.

The final grid consisted of

approximately 800 000 fluid cells and was

used to perform several runs, each with

different boundary conditions, representing

different ventilation designs. Figure 3 shows

one of these solutions and indicates the

expected path that smoke particles will take

when released by smokers at selected

machines.

The model predicted that it would be

difficult to contain the cigarette smoke to

certain areas without significant changes to

the ventilation system. A combination of a

modified ventilation system and partitions

around the centre section of the main floor

was found that made it possible to contain

and effectively ventilate the cigarette smoke

from that area.


Figure 1: Detail of the main floor of the casino

Figure 2: Dome and roof detail

indicating the position of inlet

boundaries

S

Figure 3: Particle tracks showing the movement of cigarette smoke

BUILDING APPLICATIONS

his study concerns an existing 260 million

BTU/hr tangentially fired process boiler

which was re-fitted with a low NOx natural gas

firing system. The re-fit involved changing the

burners, the windbox and the Air/FGR

ductwork, as shown, with the boiler, in figure 1.

After a year of operation, some very

unusual damage was witnessed in some of the

burners’ Fuel-Lean ports. This was

rather puzzling to the

manufacturer as this

burner design had

been tried and

tested for almost 20

years without

suffering such

problems. The

only difference

between this

particular unit

and the other

t a n g e n t i a l l y

fired boilers

where this

technology was

installed is in

boiler capacity;

all the

other boilers had capacities ranging

from 500 to 1000 million BTU/hr.

Due to the impossibility of carrying

out any kind of experimental

measurements inside the

furnace area of an operating

boiler, Cerrey S.A. de C.V., an

industrial boiler manufacturer,

requested that a CFD analysis of

this boiler’s operation be

carried out in order to determine the cause

of these unusual problems.

In order to determine if the combustion

system had any operational difficulties

when fired according to design

specifications, BMA first carried out a

detailed furnace simulation (3.6 million

cells). Details of the burner and windbox

are shown in figure 2. The burner is divided

into six zones: a lower Fuel-Lean port with

pre-mixing chamber, a lower pure re-

circulated flue gases (FGR) port, a central

Fuel-Rich port, an upper FGR port, an upper

Fuel-Lean port and an Over-Fire-Air (OFA)

port. The pre-heated combustion air is diluted

with FGR prior to entering the windbox with an

average FGR mass fraction of 13%. The first

simulation was carried out at 100% load using

boundary conditions corresponding to a

homogeneous Air/FGR mixture entering the

windbox.

The results of the simulation showed that

this firing system operates within

specifications when subject to design

conditions. Furthermore, no examples could

be found of situations likely to cause the

damage to the Fuel-Lean ports. Consequently,

it was concluded that the problem

with this unit probably lay

upstream of the firing system. To

explore this further, a second

simulation (6.2 million cells)

including all the Air and FGR

ductwork was carried out at

60% load. The partial load

operating conditions

were chosen because,

as this is a process

boiler, it spends half of its time operating in

turndown mode, which is when it is deemed

that the damage to the Fuel-Lean ports is most

likely to occur.

Figure 3 shows the FGR mass fraction

distribution in the ductwork and windbox. The

pipes shown in red contain 100% FGR and lead

either directly to the FGR ports or into the

pre-heated air stream. The figure clearly

shows that that the FGR distribution within the


POWER GENERATION APPLICATIONS

Analysis of a low NOx burner re-fit of a tangentially fired boilerDr. F. McKenty*, L. Gravel*, M. Mifuji**

Figure 1: Boiler, burners, windbox & ductwork

Figure 3: Ductwork Air/FGR distribution

Figure 4: FGR mass

fraction at the face

of burners 1– 4

Figure 2: Tangentially fired burner

T



pre-heated air stream is uneven. Figure 4

shows the FGR mass fraction distribution at

the burner face. Burners 1 and 2 have FGR

concentration distributions that vary widely

from port to port. Burners 3 and 4 have near-

homogeneous FGR distributions; these are the

burners that are located on the far side of the

boiler and so there is more space for the FGR

and air streams to mix. Nevertheless, the

average FGR mass fraction at burners 3 and 4

is only 7%, which places the Fuel/Air/FGR

mixture in the pre-mixing chamber within

flammability limits. The FGR distribution

imbalance was thus demonstrated to be the

source of the damage to the Fuel-Lean ports

caused by the mixture igniting within the port.

The specified design concentration for FGR of

13% would have made this situation

impossible.

Figure 5 shows the resulting fireball. It is

asymmetrical (Figure 6) as the flame from the

Fuel-Rich port of burner 2 is severely lifted due

to the very high FGR concentrations. This

behavior was confirmed by visual inspection of

the fireball through the boiler view-ports.

Figure 7 shows flames developing inside the

upper Fuel-Lean port of burner 3; this is the

ultimate cause of the damage.

The STAR-CD simulations allowed BMA to

pinpoint the cause of the damage as being the

result of incomplete mixing of the pre-heated

air and FGR streams. This mixing problem only

presented itself on this small process boiler

because the Air/FGR ductwork is considerably

scaled down when compared to the larger

installations where this technology had

previously been installed. Consequently, the

FGR stream has much less room to thoroughly

mix with the pre-heated air stream before the

ductwork splits the flow to each side of the

boiler. This behavior would have been hard to

predict without CFD as most aerodynamic

mixing phenomenon do not scale linearly with

geometrical size and flow rates. In this case

maintaining geometrical similarity with larger

units was insufficient to ensure identical flow

patterns.

In summary, the STAR-CD simulations

showed that a problem that was first thought

to be the result a serious flaw in the burner

design was instead linked to a simple

imbalance in the FGR distribution. This was a

very important conclusion as the problem was

easily remedied once it was identified, but

failure to balance the FGR distribution would

have doomed any attempts to eliminate the

problem by modifying the burner design.

*BMA - Brais, Malouin and Associates Inc 5450 Côte-des-Neiges, suite

600, Montréal, (Québec) Canada, H3T 1Y6 www.bma.ca

**Cerrey S.A. de C.V. Av. Republica Mexicana 300, San Nicolas de Los

Garza, N.L. Mexico, C.P. 66450Figure 5: Luminous flame contour Figure 6: Luminous flame contour; Fuel-Rich port level

Fig7: Luminous flame contour – upper Fuel-Lean port – Burner #3



The impact of CFD on thedesign of the PBMRSarel Coetzee, CFD Department PBMR (PTY) Ltd., Centurion, South Africa

he Pebble Bed Modular Reactor (PBMR) is

a next generation nuclear power plant with

high thermal efficiency and inherent safety

characteristics. The extensive use of CFD in the

design of the PBMR allows the engineers to

tackle challenges during the design phases that

would have otherwise only been encountered at

high cost during the commissioning or

operation of the plant.

The PBMR utilizes a direct cycle high

temperature gas cooled Reactor Unit (RU) and

Power Conversion Unit (PCU). The plant has a

reactor of a pebble bed type and a three-shaft

helium Brayton Cycle (Fig 1). The helium gas is

heated by the reactor, passes through a high-

pressure Turbine, low-pressure Turbine and

Power Turbine, driving the generator. It passes

through a Recuperator, Pre-Cooler, low-

pressure Compressor, Inter-Cooler and high-

pressure Compressor, back through the

Recuperator to the Reactor. Helium is chosen as

the working fluid due to the particular benefits

that it brings to closed cycle high temperature

reactors. Its advantages are that it is a

chemically inert gas and thus not affected by

radiation, high specific heat and its high sonic

speed (three times higher than air), allows

higher circumferential velocities on turbo

machinery blades. The disadvantages are that

some PCU components need to be either

specifically developed for helium or adapted

from existing components.

CFD in the design process

CFD provides detailed information to System

Engineers from PBMR as well as external

suppliers and serves as input to FEM analyses

when integrated CFD-FEM results are required.

CFD provides component characteristics for

complex geometries to Flownet, a one-

dimensional thermo-hydraulic network solver.

This interaction is graphically shown in Figure 2.

Reactor Unit, RU

The RU of PBMR consists of a central column of

graphite spheres surrounded by an annular fuel

pebble bed, enclosed by graphite blocks on the

inside of the core barrel. Between the core

barrel and the reactor pressure vessel is a gap

filled with helium. Between the reactor

pressure vessel and the concrete is an array of

water pipes, protecting the concrete against

high temperatures (Fig 3). CFD is employed to

investigate local as well as global

thermal and fluidic effects.

These CFD results have led to

several design changes to satisfy

the PBMR specifications, ranging

from the design of the water pipes,

support structure of the vessel to

the design of the helium inlet and

outlet slots. It is clear

that the design of the

RU has been greatly

influenced by

the detailed

CFD results.

T

Fig 1: Layout of the PBMR

Fig 2: CFD interaction with internal and external groups

Fig 3: Model of the PBMR reactor unit

Fig 4: Meshed volumes of

the PCU



Power Conversion Unit, PCU

Cycle pressure losses and leakage flows have a

major effect on cycle efficiency. The cycle

pressure losses are primarily a function of the

individual component designs and layout.

Leakage and cooling flows are also a function of

the component design and component cooling

strategy. These pressure losses, leak flows and

cooling flows must be determined across

interfaces between components from different

suppliers and through the components

themselves. Some of these interfaces have a

high temperature and/or pressure gradient.

This calls for integrated CFD and FEM analyses.

Therefore, a complete CFD model of the PCU

was constructed, containing all the different

components and interfaces (Fig 4 and Fig 5).

All fluids and solids were solved

simultaneously to obtain temperature and

pressure fields that were mapped onto a FEM

mesh. The CFD results were also used to

calculate pressure drops across the different

components. The calculated loss coefficients

are used by Flownet to improve the accuracy of

the cycle calculations. Detailed information

could also be supplied to the component

designers regarding the thermal

environment in which their components

will operate.

Spent fuel storage tanks

The Spent Fuel Storage Tanks are used to

store spent fuel from the power plant

generated during its production lifetime

(40 years). Thereafter, the tanks must

store the spent fuel for another 40 years

before being decommissioned. Detailed

and accurate temperature distributions

throughout the complete Spent Fuel

Storage Area are needed, ensuring that

the temperature limits for the fuel, tank,

supports and concrete are not exceeded. CFD

was used to simulate this complete Spent Fuel

Storage Area. This model included the fuel,

helium in the tank, the tank itself, the air

surrounding the tank and the concrete walls of

the area (Fig 6).From the results, temperature

distributions in all of the materials could be

obtained. The temperature distribution for the

tanks is shown in fig 7 and the temperature

distribution of the fuel is shown in Figure 8. The

high heat source from the "youngest" fuel can

be clearly seen. Note also the effect this has on

the tank temperatures. CFD supplied answers

to the tank designers, the HVAC designers, the

building designers and the nuclear physicists.

Conclusion

CFD has been a major contributor to improving

PBMR design. Optimizing many design aspects

before commissioning and operating the plant

has saved time and cost.

Fig 5: Close-up of the PCU mesh

Fig 6: Geometry of the spent fuel tanks

Fig 7: Temperature contours of the

spent fuel tank walls

Fig 8: Temperature contours

of the fuel pebbles

n process plants, for example those operated

in cement works, Gas Conditioning Towers

(GCTs) are often placed upstream of

electrostatic precipitators as an important

component of after-treatment. Fig 1 shows

a GCT used to cool flue gases from

temperatures of 300 - 400˚C down to

approximately 140˚C, which is a suitable

temperature for the electrostatic

precipitator. To obtain optimum GCT

performance, wet dust build-up

on the tower wall must be avoided

which can be achieved by

introducing gas flow without wall

separation. A special swirler unit was

recently developed and optimized

with the help of a STAR-CD simulation

of the swirling type of flow. The

swirler, installed just above the

diffuser, is a totally new type, which

replaces traditional gas distribution

screens. Better control enables the

gas distribution to achieve enhanced

gas conditioning with lower pressure

drop.

Model validation

Extensive validation investigations were carried

out with various differencing schemes and

turbulence models, and the results were also

tested for mesh independency. Experimental

data included high accuracy Laser Doppler

Anemometry (LDA) measurements and full-

scale data. The LDA and full-scale

measurements showed good agreement with

the numerical results (Fig 2). It was concluded

that second order upwind difference and the

high Reynolds number form of the standard

k-epsilon turbulence model give accurate

solutions for this type of swirling flow, if the

swirl number is less than approximately 0.8.

Improvements achieved

The performance of the swirler was investigated

using a standard FLS Airtech GCT design

including a low abrupt bend inlet section with

guide vanes. A pronounced effect on the

velocity profile was found including a central

back-flow region in the diffuser section

ensuring down-streaming flow at the

tower wall with a gas distribution

tending towards uniformity further

down the tower (Fig 3). If large droplets

are in play, the design can be extended by

installation of a flow straightener in the outlet of

the diffuser. This effectively reduces the swirl

component in the tower (Fig 4) and prevents

wet dust build-up on the tower wall. The key

advantages of the new GCT design are: lower

pressure drop compared to constructions with

gas distribution screens; hot down-streaming of

gases at walls; no dust build-up on the tower

wall and gas distribution internals; and

vibration or rapping are not needed.

Conclusions

The study has demonstrated that a

STAR-CD simulation of turbulent

swirling gas flow with low swirl

numbers is an accurate and

e f f e c t i v e

investigation tool that

can partially replace model and full scale

testing, resulting in shorter delivery time from

problem formulation to result. Prospects are

promising as the new GCT design can be

applied for new units and optimization of

existing installations. From a broader

perspective, it is expected that similar swirler

designs can achieve optimal operation in many

other industrial processes.



Applied CFD research for gasflow in gas conditioning towersNiels Finderup Nielsen, Particulate Process & Development, F.L.Smidth Airtech A/S, Denmark

Fig 3:

Typical FLS

Airtech GCT inlet

and swirler

section and

calculated axial

velocity distribution

in diffuser and tower. Note

that pink to red colors indicate

back flow

Fig 2: Comparison between numerically calculated, model

scale, and full-scale measured axial velocities. Thin

vertical line indicates zero velocity

Fig 4: Swirl number versus downstream position of a typical FLS

Airtech GCT

I


TRANSPORT APPLICATIONS

he basic design goal for a combustion

engine silencer is to achieve a certain

noise attenuation within a maximum

allowable pressure loss and within a defined

envelope. Silentor A/S has specialized in

excellent low frequency attenuation at low-

pressure loss in exhaust systems for the

automotive, marine and industrial market.

This is achieved by optimizing the design of

diffusers in our silencers.

CFD is well suited to calculate the

pressure loss through the silencer, and using

STAR-CD, the pressure loss can be calculated

with an accuracy of 10-20% of measurements.

At Silentor A/S, we use the same geometrical

model for both the pressure loss calculation

and the noise attenuation prediction.

The noise attenuation performance of a

silencer can be measured by sending

loudspeaker-generated noise through the

silencer, and measuring the transmitted

noise. The same thing can be done in CFD

calculations where pressure waves are set as

the input transient boundary condition, and

the signal emitted from the silencer is

calculated to predict the noise attenuation.

The secondary silencer in a two-silencer

system for a low noise diesel engine is shown

in Figure 1. The left-hand side shows the

velocity field used in the pressure loss

calculation and the right hand side shows the

local sound pressure for a 200 Hz tone.

Normally, a broad sound source is used, but

for illustration a pure tone is chosen here.

Figure 2 shows a comparison between

calculated and measured attenuation in the 0-

3000 Hz range. There are detailed

discrepancies between calculations and

measurement, but the overall shape is

accurately captured. The start of attenuation

and the first deep resonance at 600 Hz, which

originate from the inlet pipe, is well captured.

Even though not shown here, we have found

that the effect of using absorption material in

the silencer can also be calculated, thereby

enabling the amount and position of

absorbent to be optimized.

In conclusion, we found that the

calculation of pressure loss through a silencer

is accurately predictive, and that CFD is also

effective in noise attenuation prediction and

can be used for calculating the frequency

where attenuation starts where resonances

occur in the damping spectrum and the

overall level of attenuation. Thus, we have a

tool for both design optimization, as well as

for understanding silencer performance. The

calculations can also be used for more

fundamental questions such as the trade offs

between different diffuser designs, which can

be difficult to do in any other way.

For further information, please contact Mads

Røgild, [email protected]

Silence simulated

with CFDMads Røgild, Silentor A/S, Denmark

Figure 1: Axis

symmetric velocity

field through the

silencer (left) and

sound pressure

propagation of a 200

Hz tone (right)

Figure 2: Comparison between measured and calculated

noise attenuation through a Silentor low frequency

attenuation silencer

Figure 3: Port side Silentor silencer for a fast ferry

T

es-turbo

The CFD Expert System for the analysis of rotating

machinery

Guides you through the intuitive process of using es-turbo.

The entire process from geometry import to post-

processing, is clearly defined and examples of es-turbo

applications are also illustrated.


NEW PUBLICATIONS

Technical publicationsTo obtain any of these publications, please send an email request to [email protected], stating your preferred brochure andcontact details. Alternatively, you can download them as pdf’s from: www.cd-adapco.com/products/brochuredownload

es-uhood

A no compromise knowledge based Expert System for all

your underhood flow and simulation needs

Provides an overview how es-uhood can accurately

simulate flow and thermal processes in the complex

engine compartment. This includes the preparation and

cleaning of CAD surfaces and geometry, generation of

appropriate CFD grids, effective choice of boundary

conditions and physical models, to controlling the CFD

solution and post processing of results.

pro-am

A universal automatic volume mesher for all your analysis

gridding needs

Illustrates how pro-am employs state of the art technology

to provide optimized 3D meshes across the entire

spectrum of industries. Based on a unique cell trimming

technique, it provides the most efficient mesh for a given

volume, whether it is for the underhood of a car or the

external aerodynamics of an aircraft.

STAR-Works video

The new STAR-Works video

demonstrates how users of STAR-

Works are able to experience the

benefits of having a CAD system

as a front end to STAR-CD. The

power of parametric optimization

is shown on a CD-ROM which talks

you through a typical CFD case,

from building the model in the

CAD system, to running CFD, to

visualizing the results, and then

simply varying geometric

parameters to look at new flow

solutions until an optimized

design is reached. Mainstream

CAD with (real) CFD is here at last!


NEW PUBLICATIONS

CFD & CAE for the

turbomachinery industry

Describes how CFD can be applied to virtually every major

component of contemporary turbomachines. Applications

described include: Compressors and fans, blade cooling,

combustors, tip and casing treatments, turbines and

nozzles, cavities, ducts and nacelles.

CFD & CAE for industry

Provides a general overview of STAR-CD’s ability to provide

insight into complex fluid flow, heat transfer, chemical

processes and combustion processes.

CFD & CAE for the marine

industry

Illustrates successful applications of STAR-CD and Comet

in the Marine industry with examples from well-known

clients. Applications include interaction between ship hull,

propeller and rudder, flows around ships with breaking

waves, flow around high-speed planing boats, sloshing

and slamming, flow around surface-piercing propellers,

underwater exhausts, cavitation and more.

Europe Far East

Yokohama

CD-adapco JAPAN

Nisseki Yokohama, Building 16F

1-1-8, Sakuragi-cho, Naka-Ku

Yokohama, Kanagawa 231, JAPAN

Tel.: (+81) 45 683 1998

Fax: (+81) 45 683 1999

[email protected]

www.cd-adapco.co.jp

New York

adapco

60 Broadhollow Road

Melville, NY 11747, USA

Tel.: (+1) 631 549 2300

Fax: (+1) 631 549 2654

[email protected]

www.cd-adapco.com

London

Computational Dynamics Ltd

200 Shepherds Bush Road

London, W6 7NY, UK

Tel.: (+44) 20 7471 6200

Fax: (+44) 20 7471 6201

[email protected]

www.cd-adapco.com

Global offices of the CD adapco Group

Austin, TX

Cincinnati, OH

Detroit, MI

San Diego, CA

Miami, FL

Seattle, WA

[email protected]

Korea

CD-adapco Korea

Seoul office

[email protected]

China

CD-adapco Japan Co. Ltd

Beijing office

[email protected]

Resellers

North and South America

Australia

Veta Pty

[email protected]

Czech Republic

SVS FEM s.r.o

[email protected]

India

CSM Software Pvt Ltd

[email protected]

Iran

Mohandesin Moshaver Dynamic Sayallat

[email protected]

Malaysia

Numac Systems Technologies S/B

[email protected]

Russia

CAD-FEM GmbH

[email protected]

Singapore

CAD-IT CONSULTANTS

[email protected]

South Africa

CSIR

[email protected]

Taiwan

FLOTREND Corp.

[email protected]

Turkey

info(+)TRON A.S

[email protected]

A-Ztech Ltd

[email protected]

France

CD-adapco France

Paris office

[email protected]

CD adapco-Lyon office

[email protected]

Germany

CD adapco-Nuremberg office

CD adapco-Hamburg office

[email protected]

Italy

CD adapco-Turin office

[email protected]

Documents

ISSUE 19 Winter 2002/3