realisticfoulingbuild-upcan resultin substantial quantitiesof product being
used unless the product is recycled. However, recycling can result in
unrepresentativefoulingcharacteristics as the fouling-sensitivematerials will
tend to deposit on the heatedsurfaces duringthe earlystagesof the run and thus
reducethe concentrationof thesespeciesin the bulk. The fouling curvemay
thentendto indicate a maximumfouling rate at the beginning of the run
followedby a reductionin rate eventuallyreachingan asymptote.
In a pilot-scale plate heatexchanger systemfor heatingproducts to UHT
temperatures (Fryeret al.1996)the capacity of the systemwas 150 L/h, the total
heat transfer area was 1.6 m^2 and the hold-upvolumein the heat exchangerwas
3.1 L. Thismeantthat for a totalproduct volumeof 50 litreseachelement of
fluidis subjectedto the full UHTtemperatureprofilethree times per hourand
henceany heat-sensitive fouling componentsare likely to be deposited in the
relatively earlystages of the run.Thehighly instrumented rig enablesthe
temperatures and pressures overa numberof differentsections of the machine to
be monitoredand hencethe fouling of product underwell-definedtemperature
conditions.
35.4.5 Heat fluxsensor
A number of sensors basedon heat flux havebeenreportedwithapplication at
bothlaboratoryscale(Joneset al.1994)and on commercialplant(Truonget al.
2002).The development and potentialapplication of a patented heat flux sensor
(Baginskiet al.2002)has beenreported (Hastinget al.2005).The sensor can be
usedto generate as wellas monitor bothfouling and the subsequentcleaning
process.Initialconcepts basedon the use of a commercialheatfluxsensor
withinthe completesystem(Joneset al.1994) werepredictedto havelimited
sensitivityto depositthickness (Hastinget al.2005).A modified designusinga
heatingblock directlycoupled to the externalsurface of the flowchannel gave
significantly greatersensitivity.Initial results using a sensor manufactured
according to this designwas shownto be capable of monitoring fouling and
cleaning with whey protein and skimmed milk fluids under laboratory
conditions. The geometry and configuration of the sensor used standard
components to incorporate it into the rig and the construction was considered
hygienic, robust and compatible withthe foodindustry.
Furtherworkwiththis sensor designshowed the potentialvalueas a small-
scaleexperimentaltool. The experimental arrangementused a small volumeof
fluid,20 litres,and recirculated this past the sensorand backto a buffertank.
Owing to the smallheat transferarea of the sensor, 3.6 10 ˇ^4 m^2 , the amount of
heattransferredwas smalland the bulktemperatureof the fluidcould be
maintained at temperaturesbelowwhichfoulingrates wouldbe minimised.The
heat transfer area to volume ratio,0.018m^2 /m^3 , was considerablylowerthan the
plateheatexchangerunit described above,32 m^2 /m^3. The system designalso
meantthat the fluid velocityand temperature drivingforce,whichare both
considered significant in terms of fouling and cleaning, can be varied
Monitoringof fouling, cleaning and disinfection in closedprocessplant 583