Jensen(2003)shows this for different types of componentcommonlyfoundin
the foodindustry. Someof the detailsin eachstep are presentedbelow.
Step1: A criticalwallshearstressis required for the cleaningtest method
investigated. The critical wall shearstress(mechanical)depends on the natureof
the soil,the surfaceof the equipment and the threeremainingcomponentsof
Sinner's circle(cleaning time,detergenttemperature and chemical strength).
The straightforward approach to obtaina criticalwallshearstressis to use the
radialflowcellassay(RFC) (Fowler and McKay,1980)in whichremovalof an
arbitrary soil froma surfacecan be investigatedwhenexposed to a wallshear
stressthat graduallydecreasesfrominletto outletin the test section. A critical
wallshearstressof approximately 3 Pa is applicable for the EHEDG cleaning
test using a pickledstainlesssteel316Lsurfacewitha meanroughness (Ra) of
0.5m (Jensenand Friis,2004a).Thecritical wallshearstressis foundby
comparing the cleaningresultsof EHEDGtestswithwall shearstressin the RFC
predictedby CFDsimulations.
Step 2: Prediction of wall shear stress and fluid exchange should be
performed usingappropriatemeshand models(for wallshearstresssee Jensen
and Friis,2004b,and for fluidexchange see Jensen,2003 and Jensen and Friis,
2004c). The near-wall regionshouldbe resolvedand flow simulated using two-
layer models or similar approaches. Second order spatial and temporal
discretisationshouldbe chosen if possible. Ideally, flowpatternsshould be
validatedby use of experimental methodssuchas laserDoppler anemometry
(LDA), lasersheetvisualisation(LSV)or particleimagevelocimetry(PIV).
However, suchdata are seldomavailablefor designers of equipment,hencethe
CFD simulations must be performed based on best practice (Casey and
Wintergerste,2000)and experiencefrompreviousvalidatedsimulations(thisis
an acceptedand widely used methodin otherengineeringapplications).
Researchers haveexperimentally shownthat different degreesof cleanability
in a straightpipe or a ductare linkedto the meanwallshear stress(e.g.
Duddridgeet al., 1982;Bergman and Tragardh, 1990).Others,however, have
shownthat for more complex flowsthe meanwall shearstressis not explanatory
for the differentdegreesof cleanability found(Jensen and Friis,2005);the effect
of detergent availability(masstransport)at the soil shouldalso be considered
(Lelie¡vreet al., 2002,2003;Jensenand Friis,2004c).Hence, bothwallshear
stressand fluid exchange haveto be known and evaluatedfor eachand every
area insidethe equipmentinvestigated(Jensen, 2003). One of the mainconcerns
of this approachis the fact that predictionof wallshearstressusingCFDis
known to be difficult ± obtainingdata for validation of the wallshearstressand
fluidexchange predictions is not trivial.Eventhough,confidence maybe high
that the areas predictedas difficultto clean are in fact more difficultto clean
than other areasin the component, this may be a problem. All wall shearstresses
used(thosefor obtaininga critical wallshearstressand thosefor evaluating
cleanability in a component)are predictedfromCFDsimulations.It is assumed
that the levelof wallshearstresspredictedand fluidexchangeestimatedare
comparable.
Improvingthe hygienicdesignof closedequipment 201