Handbook of Hygiene Control in the Food Industry

of 1.5 m/s,almostindependent of the diameter, for pipes withinnerdiameters
above0.02 m. The thinnerthe viscoussublayer, the betterthe cleaning,as a thin
viscoussublayerallowsa faster(shorter lengthof diffusion,@y) transfer of
detergent and heatfromthe outsideof the viscoussublayerto the soil thana
thickerviscous sublayer.Thisis illustratedby Fourier'sheattransfer equation:

qsàˇk

@T

@y

yà 0

Ö 11 : 3 Ü

whereqsis the heat transfer(W/m^2 ),kis the thermal conductivity(W/(mK)), is
the temperature gradient(K) andyis the distance fromthe soil (m). The effectof
velocityand Reynolds numberfor cleaning in straightpipes is discussedby
Timperley (1981). The conclusionof his study is that specifyinga velocityof
1.5 m/s in pipes withan innerdiameterof 0.038and 0.076 m is moreappropriate
than specifying a Reynolds number when evaluating removal of micro-
organisms.Thisis supported by the findingsof Bergmanand Tragardh (1990)
for the removalof clay in a straight ductunderturbulent flowconditions.
However, recentfindingsof Lelie¡vreet al. (2002,2003)showthat for more
complex equipment the wall shearstress, and therebythe velocity,cannotgive a
coherentexplanationof the results of cleaning tests.Instead, localmasstransfer
to the surfaceis shownto be important.In equipment withcomplex flow
patterns,recirculationand separation create fluctuationsin the flowand in the
boundary layer,creatingdifferentlevels of masstransfer to the surface and wall
shearstresson the surface. In the workof Timperley (1981)and Bergmanand
Tragardh (1990) the viscoussublayerwas hardlyaffected by the rangeof
velocities and Reynolds numbers (the average velocitywas above1.5 m/s)
investigated, which could explain the difference from the conclusions of
Lelie¡vreet al. (2002,2003).
The aboveconsiderationspresentedby Timperleyand Lawson(1980)and
Timperley (1981)are validfor cleaning of straightpipes. In straightpipesof
innerdiameters between 0.02and 0.076m, withfullydevelopedturbulent flow,
a mean velocityof 1.5 m/s producesalmostconstant wall shearstresses (slightly
higherat the smaller diameters).Changingthe velocityfrom1.5 m/s to, e.g.,
0.5 m/s or 2.5 m/s has a large impacton the wallshearstressin the pipe.Thisis
similarto the conditionsin all othertypesof equipment (bends, valves, heat
exchangers,etc.)otherthanstraight pipes;localvelocitiesdifferentfromthe
averagevelocityare encounteredat different locationsin the equipment.
Muchequipment has someareas withvelocitieshigher and someareaswith
velocitylowerthanthe mean (see Fig. 11.2for localvelocitiesinsidea 90Î pipe
bend). Furthermore, the velocityat the wallis alwayszero, so it is impossibleto
statethat a localvelocity at the wallshouldbe of a certainmagnitude.Hence,
specifying a meanvelocity as the onlyindicator for the cleaning effectof fluid
flowin a CIP operationis too weakfor optimisationpurposes. Instead, a
combination of wallshearstress and fluidexchange/mass transfer fromthe bulk
to the viscous sublayershouldbe evaluatedto estimate if certainareasof the

Improvingthe hygienicdesignof closedequipment 195