飞机中液滴粒径和速度，液态水含量LWC检测方案(激光干涉仪)

智能文字提取功能测试中

ResearchGateSee discussions， stats， and author profiles for this publication at：https：//www.researchgate.net/publication/266872249 ILASS Americas， 23"d Annual Conference on Liquid Atomization and Spray Systems， Ventura， CA.， May 2011 Aircraft Icing Research： Challenges in Cloud Simulation and CharacterizationArticle CITATIONS READS 0 90 3 authors， including： W.D.Bachalo 160 PUBLICATIONS 2，208 CITATIONS Chad SipperleyStep 2 Consulting， Inc.7 PUBLICATIONS 61 CITATIONS SEE PROFILE SEE PROFILE Aircraft Icing Research： Challenges in Cloud Simulation and Characterization William D. Bachalo， Chad Sipperley， and Gregory Payne Artium Technologies， Inc. 150 West Iowa Ave. Unit 202 Sunnyvale， CA 94086 Abstract Recently， FAA requirements for aircraft icing certification have been updated to include characterization ofsupercooled large drops (SLD). Cloud measurements and icing wind tunnel simulations of existing Federal AviationRegulations， Appendix C， Part 25 and these new conditions (Appendix O) require advances in instrumentationcapable of measuring droplet size and liquid water content (LWC) under a much wider range of conditions. Cloudmicrophysics investigations and measurements have revealed that some clouds containing large drops can havedroplet size distributions that are bi-modal. Such conditions present an additional challenge to the instrumentation.Existing instrumentation based on concepts developed in the 1970’s are known to have difficulties measuring spraydrops in the SLD range. Currently， up to five or six different instruments utilizing different measurement principlesare needed to characterize icing clouds and icing cloud simulations in icing wind tunnels. The phase Dopplerinterferometer flight probe， dual range (PDI FPDR) has been developed to address these new and challenging spraymeasurement conditions. This instrument is capable of simultaneously measuring droplets in the size range of 1 umto 2000 um or larger and LWC over a full range of cloud and cloud simulation conditions. To ensure reliableoperation under the harshest icing conditions， the probe incorporates deicing and temperature controls. Since theinstrument is considered to be relatively new to the meteorological and icing research community， extensive testsand evaluations of the instrument were conducted to demonstrate its measurement capabilities. An extensiveprogram of instrument comparisons between the well-established PMS FSSP and OAP instruments was conducted.Results of these comparisons are reviewed and discussed. Significant measurement differences were found underhigher droplet number density spray conditions and these differences are discussed and reconciled. Comparisons ofLWC measurements are also provided and compared to the well-established hot wire devices and icing blade data. Introduction： Aircraft icing remains a potential threat to flightsafety and continues to be the subject of intenseresearch， testing， and certification to ensure that aircraftcan operate safely under a wide range of atmosphericconditions. Ice buildup on aircraft components occurswhen aircraft encounter supercooled droplets in theatmosphere which then freeze upon or after impact withthe aircraft surfaces producing ice accretion that canseriously affect the drag， lift， and control of the aircraft.Ice accumulates not only on the wings and controlsurfaces but also on exposed engine components andinstrumentation. For example， Air France flight 447from Rio de Janeiro to Paris was lost over the Atlanticnear the Equator due to consequences of what isbelieved to be icing on the pitot-static systems neededfor monitoring air speed.Satellite images and observedweather conditions indicated that the aircraft may haveencountered rime icing and possibly clear or glaze icingconditions. Turbojet， turboprop， and turbo fan engineinduction systems are also affected by ice accretion oninternal engine vanes and compressor blades whenencountering mixed phase conditionss(supercooledliquid drops and ice crystals). Ice particles are known tomelt within the engine and then freeze onto thecompressorbladesirinside of the engineaffectingcompressor performance producing compressor surgesand stalls，flameouts， which can lead to engine damage. Recently， the FAA [1] has proposed a significantexpansion in its icing certification standards to includethe condition known as supercooled large droplets(SLD). SLD refers to the drizzle-sized droplets in theapproximate size range (median volume diameter，MVD) of 100 to 500 um in diameter (Federal AviationRegulations， Part 25， Appendix O).Droplets in thissize range can remain as liquid at temperatures wellbelow freezing (to -40℃) and are especially dangeroussince they can collect and freeze almost anywhere onaircraft surfaces. Coupled with smaller drops， they cancreate a rough surface or structure which spoil aircraftlift and control performance. A key event leading tothese expanded regulations was the tragic 1994 ATR 72icingnrelated accident in Roselawn， Indiana.Investigations of the accident concluded that liquidwater runback of SLD produced a subsequent iceformation aft of the deicing boots near the leading edge.Simulations and characterization of mixed phase and SLD icing conditions remain a challenge forexisting iicingwindtunnels andinstrumentation. Available I tunnels were designed for genraloperation with large spray droplets. Challenges thatmust be confronted include production of a large dropsize range with a relatively uniform droplet size andconcentration distribution within the test section. Flowthrough the contraction and gravitational force can beexpected to produce segregation of droplet trajectoriesbased on their inertial characteristics and air flowspeed. Furthermore， large droplets require much greatertime to reach supercooled conditions than smallerdrops. Thus， the relative residence times of the dropletsmay need to be controlled which is very difficult inicing wind tunnels. Existing instruments based on lightscattering can be frustrated by the mixture of sphericalliquiddroplets and icecrystals. Forexample，commonly used PMS FSSP instruments [2，3] based onnear forward scattered light intensity measurements areknown to produce significant error under mixed phaseconditions [3，4]. More recently， the phase Doppler interferometrytechnique developed into a configuration for aircraft-based and wind tunnel studies has been applied tomeasurements of droplet size distributions under mixedphase and SLD conditions. With a dual range capableof covering droplet sizes from 1 to 2000 um or largerhas proven to be more resistant to errors in reportingliquid particle measurementsandl in rejecting icecrystals. Instruments for measuring liquid water content(LWC) consisting of hot wire and heated collectiondevices also experience difficulties when attempting tomeasure a wide range of drop sizes included under SLDconditions. Uncertainties in droplett collectionefficiency and shattering are problems identified assources of measurement error. A goal of this work is to address measurementrequirements for icing research under the new FAAcertification requirements including characterizations ofmedian droplet diameter (MVD) and liquid watercontent (LWC) for icing simulations including SLD.Based on the impending FAA regulations， instrumentrequirements， capabilities and limitationswill bbediscussed.l.As a starting point. measurementcomparisons of droplet mean size (Median VolumeDiameter， MVD)， LWC， and number density will beprovided and compared to other established methods insize ranges where these methods should be reliable.Differences are analyzed and discussed and someobservations made as to the reasons for significantdifferences in the results. Experimental Investigation National and international icing facilities includingwind tunnels， open free jet systems， and other facilitiesare used along with a wide range of instrumentation forcharacterizing icing cloud simulations requiredinaircraft component certifications. Cloud measurementsincluding icing clouds are also being made by a numberof institutions using aircraft-based instruments. Thesewind tunnel simulation data are acquired with differenttypes of instruments leaving open the possibility ofdifferent certification results. Measurements are oftencarried outby groupswith varying degrees ofsophistication and understanding of the instrumentationand metrology capabilities. Often， measurements. insimilar cloud conditions produce significantly differentresults. Questions remain as to what measurements， ifany， are reliableandd whatiaare theemeasurementuncertainties associated with these cloudcharacterizations. Measurement challenges anduncertainties increase when measuring in mixed phasecloud conditions consisting of droplet and ice crystalspectra typified by a relatively broad range of dropletsizes and number densities. Added to the complexity isthe recent requirement of characterizing supercooledlarge drops； (SLD). Existing instruments use tocharacterize clouds over the past decades encounterdifficulties when attempting to measure under theseconditions. Furthermore， different types of instrumentsusing different physical principles are required formeasuring in mixed phase conditions and SLD. Theseconditions further add to the questions regardingmeasurement uncertainty and reliability. Instrumentation： Forward Scattering Spectrometer Probe (FSSP) The FSSP instrument was developed by ParticleMeasuring Systems (PMS) in the early 1970s and hasserved the meteorology community. Originally， theinstrument was designed for measuring atmosphericparticulate consisting ofboth solid and liquid particlesover a size range of approximately 0.5 to 50 um. It is asingle particle counting method [2，4，5，6，7] based on themeasurement of light scatter intensity by particlespassing through a focused laser beam and detected inthe near forward direction. The measurement principleis based on the fact that particles scatter light inproportion to their diameter squared (Isca = kI，(r)d).Innovative'e means were incorporated into ttlhis instrument to limit the detection to only particlespassing through the central nearly uniform intensitypeak of the focused laser beam. The masking systemalso limits the depth of field to form the measurementvolume. Nonetheless， forward scatter light detectioninevitably produces a relatively large sample volumewith dimensions of approximately 200 um in diameterand 2000 um in length. This limits the number densitiesof the clouds in which this instrument can be usedreliably. Although it is not described in the literature，the size of the sample volume must inevitably be afunction of the droplet size being measured. Largerdrops will be detected over a larger sample volume thanthe smallest drops. As the cloud droplet number densityincreases， probability of coincidence (more than oneparticle in the sample volume at a time) increases whichleads to measurement uncertainty. In addition， as thedroplet number density increases droplets within thelaser beam path but outside the probe volume willscatter light to the detector and will appear in thebackground of droplets that pass the validation logic.Presumably， the particle upper size measurements arelimited by the need to measure only those particlespassing the nearly uniform peak intensity region of thelaser beam. The beam would need to be proportionatelylarger to measure large drops whichwould thenexacerbate coincidence problems anddmeasurementerrors. The FSSP design incorporates the use of a tube tostraighten the flow and limit the length of the exposedbeam path for the reasons described. Under someconditions droplet shattering on the rim of this tube canbe significant producing errors in the measurements.This problem may be expected when measuring SLDsince drops in this condition are relatively large and canproduce a cascade of smaller drops when they impactthe rim of the tube. The probe was designed foratmospheric testing from aircraft and due to powerlimitations， it has inadequate heating capacity to keep itice free under the higher LWC loading conditions inicing wind tunnels. Accumulation of water on thereceiving optics has also been reported as a problem.Similar to other instruments utilizing light scatteringintensity，thelemethodnneedsrelativelyffrequentcalibration due to detector drift， changes in laser beamintensity， and optical contamination. PMS Optical Array Probe Optical array probes (OAP) were developed byParticle Measuring Systems， Inc. (PMS) in the early1970s and have been used extensively for cloud dropletand1 ice particlee spectra measuruermeemnetnst.s. The OAPinstrument utilizes a linear detector array to capture theimages as they sweep past at a predetermined speed.Shadows of the particles are measured by the detectorarray to obtain their size and some information onshape. Although the instrument incorporates strategiesto identify and limit the depth of field for the particlespassing through the exposed laser beam， these methodshave been shown to be somewhat unreliable [8，9]. OAPinstruments have been subject to numerouSinvestigations [8，9] with attention focused on varioussources of measurement error related to the underlyingmeasurement principles used. A key finding of thiswork Was thee dependence of the particlesizemeasurement on the particle distance from the objectplane which was found not to be monotonic and isdependent on the particle size. Similar depth of fieldeffects on the image size are generally present inshadow imaging techniques [10]. A detailed investigation by Korolev et al. [8]evaluated the effects of depth of field on the dropletsize measurements. Streams of monodisperse waterdroplets were traversed along the beam axis to obtainquantitative information on the measurement error as afunction of distance from the focal plane of theinstrument. The authors observed that as the droplettrajectories moved out of the expected depth of field ofthe instrument， the rate at which the droplets aremeasured did not diminish sharply as expected. Datarate gradually fell off as the droplets were traversedoutside of the acceptable depth of field of theinstrument.The： authors foundthattithe detectiondistance of the droplets as they were moved away fromthe focal plane depended strongly on the drop size. Theauthors observed that these depth of field effectsresulted in size measurement error for the smallerparticles of as much as 85% (measured as larger thanactual size). They also noted that for droplets largerthann125 um in diameter， the depth of field wasconstant and limited by the separation distance betweenthe probe windows. This affects the measurements ofSLD where the droplets will be much larger than 125micrometers. Although the depth of field will not affectthe measurement of larger droplets， collisions andshattering of droplets from the probe aarms will undoubtedly result in recording smaller droplets， addingto the measurement uncertainty. PDI Flight Probe， Dual Range (FPDR) Phase Doppler interferometry (PDI， also known asPDPA，PDA) is now a well-known and well-establishedspray flow measurement method [11]. This instrumenthas been demonstrated to have numerous advantagesoverlight scattering intensity-basedinstruments.Measurements are based on the wavelength of light andhence， are not significantly affected by interveningdroplets in the laser beam or in the light scatteringintensity path to the receiver. After factory calibration，additionalil calibrationis unnecessary.Sincelaserwavelengthsrremain constant and theoopticalconfiguration does not change， field calibrations areunnecessary. Signals produced by the instrument have aunique sinusoidal characteristic which allows easy andreliable detection of the signals even in low signal-to-noise ratio environments using digital means and thefull complex Fourier transform. A feature sometimesoverlooked is the fact that the sample volume may bereduced or increased over a relatively large range toaccommodate the prevailing number density. Themethod has undergone more than two decades ofdevelopment andevolutiona wellas extensiveevaluations on a wide range of measurement tasks. It isnoW amature technologywithhtheemeasurementcapability fully validated and sources of error identifiedand reconciled. Surprisingly， in the meteorological andaircraft icing cloud research community， the method isnot well-known nor understood and is not yet widelyused. An early version of the phase Doppler icing probewas developed under NASA funding and tested in theNASA Icing Research Tunnel (IRT)， Figure 1， [12，13]as well as on the NASA Twin Otter aircraft. Similarversions of the probes have been used by BoeingAirplane Company to obtain measurements in icingclouds from their 737 and 777 aircraft. The PhaseDoppler icing probes are also used in the CIRA icingfacility in Italy. These tests have revealed advantages ofthe PDI method as compared to the other light scatterdetection instruments. In icing wind tunnels and insome clouds， mixed phase conditions can prevail (liquidwater droplets and ice crystals). The PDI expectsspherical particles and has several signal validationtests to reject particles that are not spherical (crystals orsolid particles with morphologies significantly different than spherical). This capability is important in isolatingthe measurements to only supercooled liquid droplets.In wind tunnel testing， ice crystals shed from windtunnel walls， spray bars， screens and other surfaces thatmay be present after extended periods of operation. Thephase Doppler method has been demonstrated capableofrejecting more than >99% of these particles. After a hiatus in the development and applicationof the phase Doppler method for meteorological andicing cloud characterization， advanced phase Dopplerinstruments have been developed for cloud studiesunder U.S. Navy Office of Naval Research， NASAGlenn Research Center， US Air Force， McKinleyClimatic Laboratory， and U.S. Army funding. Theseefforts resulted in an advanced instrument capable ofmaking measurements over a size range of 0.5 to 2000um in the most challenging icing conditions. LWC Methods Currently，，hotwire iprobesand icingg bblades[9，14，15] are commonly used to estimate liquid watercontent (LWC) in clouds. The first device known as theJohnson-Williams (JW) was developed in 1955. In the1980s， PMS produced constant temperature hotwiredevices referred to as the "King"probe.The devices areanalogous to well-known hotwire anemometers whichare used for aerodynamic mass flow (pu) measurementsand use both constant current and constant temperaturemethods. The device responds to the cooling of the wiredue to evaporation of cloud droplets that have impactedand collected on the wire. It is argued that constanttemperature devices have a response that isnmorepredictable based on first principles and tthus， istheoretically independent of calibration [14]. The wirediameters for these devices are in the range ofapproximately(0.5 to2mm. Investigations hhaverevealed that at typical aircraft speeds of 100 m/s andpressures of 70 KPa， collection efficiencies rapidly fallbelow 90% for droplets smaller than about 9 to 12 umin diameter for the JW and King probes [9]. Otherinvestigations of thee ccollection efficiency havedetermined that the response of the hotwire devices alsodrops off with increasing drop size. For example， it washypothesized that for droplets above 30 um， the JWprobes experienced partial aerodynamic removal ofcaptured water mass before full evaporation. Thus，researchers have found that there is not only a rapid lossof collection efficiency with the decreasing dropletdiameter below about 10 um but there is also a loss of captured water for droplets larger than 30 um. Ingeneral， LWC measurementswith the cylindricalhotwire sensors showed a roll off in the LWC responsewith increasing spray MVD. This is believed to be dueto entrainment or shedding of captured water before thewater was fully evaporated from the wire. Extensive research by SStrapp， et aal.l.[9]wasconducted：dcoveringa MVDsizeerange lfromapproximately 13 to 236 um to address issues of liquidwater content measurement reliability， especially whenextending the measurements to SLD conditions. Thisstudy concluded that measurements of large droplets inthe SLD range are difficult both for MVD and LWC. Inthat study， they also concluded that up to three differentdroplet sizing instruments with overlapping size rangesmay be necessary to cover the size range for SLDcharacterizations. They also concluded that estimationsof MVD wereverysensitiveto tthechoice ofinstruments which.iiin turn. casts doubt onmeasurements of large MVD. Hence， biases probablyexist between different icing wind tunnels and aircraftmeasurements. Measurement Procedure Extensive testing has been conducted to evaluateperformance of the most common instruments used foricing cloud characterization. Since the newly developeddual rangephaseDoppler interferometer probeidentified as PDI FPDR is relatively new to the icingand meteorology research community， these tests werein part， an effort to demonstrate the capabilities of theinstrument. Although the existing wind tunnel datacannot be concluded to be completely reliable oraccurate， wind tunnelspray characterizationsShaveevolved over extensive and numeroustests withcarefully calibrated instrumentation[16]. This isespecially true for data obtained in the NASA IcingResearch Tunnel (IRT) and in the icing facility at CIRA(Centro Italiano Ricerche Aerospaziali， Italy). The IRTfacility was calibrated using primarily older PMS FSSPand OAPinstruments.. Liquid water contentwasmeasured using both hotwire devices and icing blades(see Strapp et.al. [9]). In some cases， cloud data wereobtained using a PDPA instrument developed in theearly 1990s. As a historical reference， figure 1 shows a directcomparison of spray MVD measurements conducted inthe NASA IRT and described by Rudoff， et al.， [12，13]. In the IRT， there are 8 spray bars in the inlet sectioneach with10 to 15sspray nozzles makingupapproximately 100 nozzles. Testing was conductedusing two sets of spray nozzles in the facility. Becauseoff nnumber densityylimitations of tthe FSSP.measurements with that instrument are performed withone half of the spray nozzles shut off. The tests showeda systematic shift with the PDPA measuring smallerdrops by as much as 50% in the small end of the MVDsize range considered. For larger MVD values， thepercent difference between the IRT calibration results(FSSP and OAP merged data) decreased. For sprayconditions with the largest MVD values， the differenceincreased again. The PDPA had an upper size limit ofapproximately 150 um and hence， was not detecting allof the larger droplets in the distribution. Questionsremained as to why the PDPA consistently reportedMVD values atthe lower size range where theinstrument should have reported reliable measurements.Data analyses offered in the following sections mayhelp explain these earlier results. In subsequent tests， direct comparisons betweenvarious instruments were conducted. A complete set ofdata from the new PDI instrument were recordedincluding MVD， flow velocity， droplet number density，volume flux， and liquid water content. Since the PDIinstrument incorporates two size measurement ranges，dropsizes from1 to 2000 um， areemeasuredsimultaneously. Comparisons were made to updatedPMS FSSP and OAP instruments and to the hotwiredevices used for measuring liquid water content. Sprayconditions were varied over a range of liquid waterflow rates and atomization air pressures. Consequently，droplet number densities variedwidely for theseconditions. Acquired data were carefully analyzed andevaluated in an effort to reconcile and understand whythere were significant differences in the measurementsunder a certain range of spray conditions. Results and Discussion Measurements reported here were acquired undertypical icing conditions with air temperatures in therange from -5° to -30°C. Flow speeds ranged from 10m/s to 100 m/s. As seen in figure 2， ice accumulated onthe instruments that were not properly heated whichnecessitated routinely shutting down the facility anddeicing the probes. The first version of the PDI flightprobe instrument shown here had sufficient heating tomaintain ice free operation under most icing spray conditions. However， at speeds of above 80 m/s andhigh LWC， the probe had some trouble maintaining theinternal temperature within the operating range of thelaser. The design and heating capability has beenchanged in the most recent design so it can operateunder the most severe icing conditions. Comparison ofPDI Results Initially， the new PDI probe was compared withstandard Artium PDI systems to ensure that the systemwas working properly and producing results that wereconsistent with our highly developed laboratory scaleinstruments. In one key facility， we have accumulatedextensive data over the past 15 years using a PDIsystem that was installed external to the wind tunnelandproduced measurements throughh the tunnelwindows. Thiss iinstrument hassproduced consistentyear-over-year results as shown in figure 3. Variationsin the MVDmeasurements are withina fewmicrometers. The PDI flight probe was installed insidethis wind tunnel and direct comparisons measurementswere made between the flight probe and instrument thatwas external to the tunnel. Figure 4 provides anexample of the results obtained under these tests. Ingeneral， the results agreed to within a difference of+/-3%. Hence， between like instrument technologies， theagreement was within the expected uncertainty range. Comparison of PDI and PMS FSSP and OAPResults An extensive series of tests with directcomparisonsbetweenl uthePMSFSSPaandd OAP.consideredto be the standardwithin thele licingcommunity， and the dual range PDI instrument wereconducted to evaluatethee newBPDIinstrument.Unfortunately， these tests revealed some very largedifferences between the measurements of as much as80%. In general， the FSSP data under the prevailingspray conditions were systematically much larger thanthe MVD measured by the PDI probe. An example ofthe results is shown in figure 5， wherein the sprayconditions were changed over a range of water flowrates and atomization pressures. In an effort to reconcilethese differences， a number of tests were conductedincluding moving the probes to the same relativepositions and repeating the measurements to eliminateany differences that may have been produced by spraynonuniformity. A reviewof tthemeasurementcharacteristics of the FSSP instrument and reports relating to this general method suggested that highdroplet number densities and coincidence (more thanone droplet passing the sample volume at one time)could potentially be the source of the differences.Therefore， the number density measured by the PDIwas plotted versus the range of water flow rates appliedto the nozzles， figure 6. The number densities rangefrom 500 to 3000/cc. In the literature (e.g. Strapp， et al.[9])， number densities above 300/cc or 400/cc havebeen recognized as the conditions for the beginning ofmeasurement uncertainty for the FSSP probe. However，in tthese reports， it is suggested that the dropletcoincidence at the higher number densities would affectthe counting efficiency but do not mention an effect onthe MVD measurements. In the present investigation，the droplet number densities measured by the PDIinstrument were well beyond the limits of reliableoperation for the FSSP. Note that as the number densityfell with reduced water flow rate to the nozzles， thedifference in the MVD measurements between the PDIand the FSSP decreased， figure 7. The measured valuesare within a few percent at the low end of the numberdensity distribution. Differences in the MVD measurements betweenthe PDI and FSSP instruments are plotted in figure 8 fora range of atomization conditions. Measurements withlower number density were produced using loweratomization pressure and lower water flow rate so thatthe MVD values were larger (~20um) for these lownumber density conditions. Note that the differences athigh number densities are as large as 45%. At lownumber densities， the differences are negative since thePDI is measuring larger MVD values than the FSSP.For the low number density conditions， the MVD's arelarger and the FSSP has an upper size measurementlimit of 47um. Hence， the largest drops in thedistribution are not being measured and therefore， theFSSPi1sSunder-sizing the sprayddistribution.Toemphasize this point， figure 9 shows measurements atlow number densities for the FSSP and combined FSSPand OAP compared to the PDI. The PDI combined data(small and large range) reported smaller MVD in thePDI small size range alone. This unexpected behavior，albeit only a difference of 2 um， is due to a fault in themerging of the resultsswhich：hiscurrently underinvestigation and further development. This problemoccurs when the drop size distribution is such that thesmall and large measurement ranges do not haveadequate overlap. In thisi case， the complete size distribution can be measured by the small size rangepart of the instrument. FSSP measurementt results shouldrnot beunexpected or surprising since they were predicted byHovenac [17] in 1991. He conducted an analytical andexperimental study to evaluate the effects of dropletnumber density on measurements reported by the FSSP.His results determined that for droplet number densitiesgreater than approximately 300/cc， the instrumentwould under-predict the reported number density， over-predict MVD， and under-predict liquid water content.Figure 10 shows the predicted results of Hovenac fornearly symmetric and for highly skewed droplet sizedistributionss.. Also plottedinthi1sSfigure are thedifferences lbetween FSSP aandPDI data obtainedduring these investigations for a wide range of spraynumber densities. Since the size distributions weremoderately skewed， the results fall within the bounds ofthe predicted measurement error. Liquid water content Liquid water content (LWC) data were acquiredwith the PDI and PMS instruments as well as withvarious hot wire devices identified here as J-W #1 andJ-W #2. The hot wire devices are considered to be astandard for LWC measurements although as describedabove， they do have their sources of measurementuncertainty [9，18]. No additional description of thesemethods will be provided here. Over the series of tests，direct comparisons of the optical data were made withthe results from these devices. Figure 11 shows atypical result under a range of atomizer water flow ratesand a single atomizing air pressure (500 kPa). Thecomparison between the PDI and hot wire devices arein good agreement except at the highest water flow rate.These tests were taken by cycling from the highestwater flow rate to the minimum and then back to thesame maximum value. Closer observation of theseresults showed that the beginning LWC measurementswere higher than the ending LWC measurement at thesame water flow rate condition. This behavior repeatedvery closely for each of the three runs shown here sothe behavior was systematic. Further investigation willbe required to determine why this was happening. Themeasured droplet number density also showed thisbehavior whereas the other parameters affecting LWC(flow velocity， D30， and probe area) did not changesignificantly. As expected and predicted by Hovenac，the LWC measured by the PMS combined instruments (FSSP and OAP) showed significantly lower LWCvalues. For these spray drop size conditions， the FSSPdata dominate since most of the droplets were withinthe range of the FSSP. Also as expected and predicted，the LWC measurements converged and agreed verywell at the lower water flow rate (lower numberdensity) conditions. At the lowest water flow rate， theFSSP was expected to produce reliable measurements. Summary Extensive testswere conductedi on the newlydeveloped PDI FPDR icing research instrument. In aprogrammatic effort to evaluate theemeasurementcapabilities of the instrument， direct comparisons ofPDI data were madee with data obtained by theestablished PMS FSSP and OAP instruments.Comparisons were also made ofthe LWCmeasurements obtained with the well-established hotwireeddevices. Ingeneral，1，tthereeWweresignificantdifferences in the MVD values measured by the PDIand FSSP instruments. Analysis of the results indicatedthat the number densities for these tests were wellabove the upper limits suggested for the FSSP. At lowernumber densities， thee instruments produced1MVDmeasurements that were in excellent agreement. Theseresults were alsoconsistentwith measurementsobtained in the NASA IRT wherein the PDI and windtunnel calibration data agreed to within a few percentfor the smaller MVD values (within the measurementrange of the FSSP). LWC measurements between thePDI and the hot wire devices were in good agreementexcept at the highest water flow rate values. Differencesin the higher LWC conditions need to be reconciled infuture tests. Consistent with predictions， the LWCmeasured by the FSSP instrument was significantlylower at higher droplet number densities. One can conclude that these instruments work wellunder the measurement conditions for which there wereoriginally intended. The FSSP was designed for cloudstudies wherein droplet concentrations are relativelylow (<500/cc). Unfortunately， wind tunnel certificationtests have been conducted using the FSSP and OAP inspray conditions for which they were not designed.These tests require reevaluation to determine whetherthese data are within acceptable uncertainty limits. The newly developed dual range PDI FPDRinstrument has been demonstrated to be capable ofmeasuring droplet size distributions over the full size range of interest iincluding SLD. Droplet numberdensitiesprevalenttininwindl tunnel icing cloudsimulations are well within the limits of this instrument.In addition， the PDI instrument can measure the flowvelocity， spray droplet number density， and liquid watercontent under the full range of icing conditions. Acknowledgements The authors are grateful to the US Navy，ONR， USArmy， and NASA Glenn Research Center fundingunder Phase I and Phase II SBIR grants used indeveloping this instrument. Funding and support werealso received from NASA Glenn Research IRT and theU.S. Air Force McKinley Climatic Laboratory. ( References： ) ( 1. FAA’s Aviation Rulemaking Advisory Committee (ARAC) and the National T ransportation Safety Board (NTSB)， h tt p ： / / edocket .a cces s.gp o .gov/2 01 0/2 01 0- 15726.htm. ) ( 2. . h Baumgardner， D.， D y e， J . E.， a nd C o oper， W. A .， “The Effects of Measurement U n certainties On the Analysis of Cloud Particle D ata，” Conference on Cloud1 Physics， American Meteorological Society，313-316，1986. ) ( 3. R iley， James T .， M ixed-Phase Icing Conditions： A Review， FAA Final Report， DOT/FAA/AR-98/76， 1998. ) ( 4. G (ardiner， B.A. and Hallett J.， 1 985： Degradation of in-cloud Forward Scattering Spectrometer measurements in the presence o f ic e cr y stals. J.Atmos. O ceanic Technol.， 2 ， 171-180. ) ( 5. B I aumgardner， D.， J. E . Dye，and J. W. Strapp， 1985：“Evaluation of the Forward S c attering Spectrometer Probe. Part II： C orrections for coincidence an d dead-time losses， ” J. Atmos. O ceanic T e chnol.， 2：626-632. ) ( 6.1 B renguier， J. L.， 1989： “Coincidence and dead-timecorrections f or p article counters，” Part I ： High concentratio n measurements with ： a an FSSP. J. Atmos. Oceanic Technol.，6：585-598. ) ( 7. Brenguier， J. L. ， D. Baumgardner， and B . B aker， 1994： A review and d i scussion o f processingalgorithms for FSSP concentration measurements. J. Atmos.Oceanic Technol.，11：1409-1414. ) ( 8. Klorolev， A .V.， Kuznetsov， S .V.， Makarov， Y . E .，and Novikov， S.V，“Evaluation of Measurements of Particle Size and Sample Area f rom Optical A rray Probes，” Journal of Atmospheric and OceanicTechnology， Vol. 8 8， American MeteorologicalSociety， 1991. ) ( 9. .s Strapp， J. W.， J. Oldenburg，R. I d e， L. Lilie，S. Bacic，Z. Vukovic， M . Oleskiw， D. Miller， E . E m ery， and ) ( 14.King， W.D.， Parkin， D.A.， and Handsworth， R.J.，“A Hotwire ILiquid d VWaterDevice H avingg FullyCalculable R esponse C h aracteristics，” Journal of .Meteor.， 17：1809-1813，1978. ) 15.Mazin， I. P.， A. V. Korolev， A. Heymsfield， G. A.Isaac， S. G. Cober， 2001： Thermodynamics of IcingCylinder for Measurements of Liquid Water Contentin Supercooled Clouds.J. Atmos. Oceanic Technol.，18， 543-558. 16.Ide， R. F. and J. R. Oldenburg， 2001： Icing cloudcalibration of the NASA Glenn Icing ResearchTunnel. 39th Aerospace Sciences Meeting aandExhibit， Reno，NV， AIAA， AIAA 2001-0234. 17.Hovenac. E.A.，“Operation，Calibration. andAccuracy of the Forward Scattering SpectrometerProbe in an Icing Environment，”Atomization andSprays， Vol. 1， pp 269-301，1991. 18.Strapp， J.. W. andl R. S. Schemenauer， 1982：Calibrationss of Johnson-Williams liquid watercontent meters in a high-speed icing tunnel. J. Appl.Meteor，21：98-108. Figure 1. Example of data acquired in the NASA IRT with the original PDPA flight probe showing the systematicsmaller MVD measured by the phase Doppler instrument [12]. The IRT data were obtained using an FSSP. Figure 2. Phase Doppler Interferometer FPDR with extended size range of 1 to 2000 um installed in the McKinleyClimatic Laboratory， Eglin AFB. Also shown are the older PMS FSSP and OAP instruments. Figure 3. PDI Icing tunnel spray measurements over a number of years showing the repeatability of the results. Figure 4. Comparisons of the PDI FPDR and PDI External Instrument Figure 5. Typical results from direct comparisons of the PDI (lower points) and FSSP (upper points) during icingspray tests taken over several different runs at the same atomization conditions. Flow rates shown were distributeduniformly over approximately 100 spray nozzles. Figure 6. Number density plotted as a function of water flow rate for four different repeated measurements atsimilar atomization conditions. Arrows indicate the sequence of measurements as they were acquired for each of thefour tests. Figure 7. MVD plotted versus measured droplet number density over a range of conditions with the minimumfollowing within the acceptable range for the FSSP. Figure 8. A plot of the difference between the FSSP and the PDI MVD measurements for a range of sprayconditions with number densities as high as 4000/cc. Figure 9. Measurements of the MVD produced by the PDI small range (PDI s)， PDI combined small range and largerange (PDI Comb)，FSSP， and combined FSSP and OAP (PMS Comb) for low number density conditions. Figure 10. Error in MVD measurements as predicted by Hovenac， 1991 in the difference between the FSSP and PDIMVD measurements. Figure 11. Series of LWC measurements obtained by the PDI， PMS FSSP and OAP compared to the hot wiredevices (J-W). All content following this page was uploaded by W. D. Bachalo on January The user has requested enhancement of the downloaded file. View publication stats       Recently, FAA requirements for aircraft icing certification have been updated to include characterization of supercooled large drops (SLD). Cloud measurements and icing wind tunnel simulations of existing Federal Aviation Regulations, Appendix C, Part 25 and these new conditions (Appendix O) require advances in instrumentation capable of measuring droplet size and liquid water content (LWC) under a much wider range of conditions. Cloud microphysics investigations and measurements have revealed that some clouds containing large drops can have droplet size distributions that are bi-modal. Such conditions present an additional challenge to the instrumentation. Existing instrumentation based on concepts developed in the 1970’s are known to have difficulties measuring spray drops in the SLD range. Currently, up to five or six different instruments utilizing different measurement principles are needed to characterize icing clouds and icing cloud simulations in icing wind tunnels. The phase Doppler interferometer flight probe, dual range (PDI FPDR) has been developed to address these new and challenging spray measurement conditions. This instrument is capable of simultaneously measuring droplets in the size range of 1 μm to 2000 μm or larger and LWC over a full range of cloud and cloud simulation conditions. To ensure reliable operation under the harshest icing conditions, the probe incorporates deicing and temperature controls. Since the instrument is considered to be relatively new to the meteorological and icing research community, extensive tests and evaluations of the instrument were conducted to demonstrate its measurement capabilities. An extensive program of instrument comparisons between the well-established PMS FSSP and OAP instruments was conducted. Results of these comparisons are reviewed and discussed. Significant measurement differences were found under higher droplet number density spray conditions and these differences are discussed and reconciled. Comparisons of LWC measurements are also provided and compared to the well-established hot wire devices and icing blade data.