NASA Technical Memorandum 88791

NASA's Aircraft

(NASA-TM-8879 I) NASA'S

ANALYSIS PROGRA_ (NASA)

Icing Analysis

AIRCRAFT ICING

26 _< CSCL 01C

Program

N86-31548

Unclas

G3/03 g3986

Robert J. Shaw

Lewis Research Center

Cleveland, Ohio

Prepared for theInternational Conference of the Aeronautical Sciences (ICAS)

London, England, September 7-12, 1986

N/ A

https://ntrs.nasa.gov/search.jsp?R=19860022076 2020-04-14T21:54:02+00:00Z

pOOR QUALITY NASA'S AIRCRAFT ICING ANALYSIS PROGRAM

Robert J. ShawNational Aeronautics and Space Administration

Lewis Research CenterCleveland, Ohio 44135

Abstract

An overview of the NASA ongoing efforts todevelop an aircraft icing analysis capability ispresented. Discussions are included of the over-all and long term objectives of the program aswell as current capabilities and limitations ofthe various computer codes being developed. Des-criptions are given of codes being developed toanalyze two and three dimensional trajectories ofwater droplets, airfoil ice accretion, aerodynamicperformance degradation of components and completeaircraft configurations, electrothermal deicer,fluid freezing point depressant antideicer andelectro-impulse deicer. The need for bench markand verification data to support the code develop-ment is also discussed, and selected results ofexperimental programs are presented.

Introduction

The aircraft icing problem has long beenresearched and studied. Reports exist in theliterature which trace icing research activitiesback as early as the late 1920's. Since thenresearch into the aircraft icing problem has beenan almost ongoing effort at varying levels ofintensity by a number of government and privateresearch organizations both in the United Statesas well as in many other countries. A commonthread which binds these various activities isthat for the most part they have been experimentalin nature. The majority of these programs havebeen aimed at studying the performance of variousice protection systems, although some limitedattempts have been made to develop icing analysiscapabilities.

• The current NASA Aircraft Icinq ResearchProgram which was started in 1978( Iy seeks totake advantage of available computational fluiddynamics capabilities and develop a series ofcompatible computer codes which will address thefundamental icing problems. Once these initialcodes are developed, additional codes will thenbe developed to broaden and enhance the capabili-ties. The computer codes will be thoroughlyevaluated by comparison with appropriate verifi-cation data. The development of these computercodes and the acquisition of the required verifi-cation data bases are two major goals of the NASAaircraft icing research program.

Once these various computer codes have been

developed and adequately validated, they willprovide the aircraft designers and certifierswith tools which should help to reduce the timeand, therefore, the cost for the development andcertification of aircraft with the required iceprotection.

Reference 2 presented the overall icinganalysis plan which was developed and the status

of the various computer codes at that time(1984). This report can be viewed as an updateof that document.

Symbols

CD,CL

Cd,CI,Cm

aircraft drag and lift coefficients

airfoil drag, lift, and momentcoefficients

C

LWC

MVD

TI ,T2 ,T3

Too

t

airfoil chord

icing cloud liquid water content, glm 3

icing cloud median volume dropletdiameter, um

electrothermal deicer temperatures, °C

free stream temperature, °C

time, sec

UIU e

Voo

X yC C

boundary layer velocity ratio

free stream velocity, m/sec

nondimensional coordinates

angle-of-attack

Overall Objectives

The long range objective of the NASA aircraft

icing analysis program is to develop a capabilityto predict the details of an aircraft icingencounter for both fixed and rotary wing vehicles.

Figure i suggests four uses which could be madeof such an analytical capability. Namely, theanalysis could be used to predict aircraft per-formance and handling qualities changes due toicing on unprotected and de-iced components andto determine ice protection system performance.In addition, the analysis could be used to designice protection systems for various applicationsand perhaps even be used to design componentswhich are insensitive to icing.

A majority of the current code development/verification efforts are directed toward the two-dimensional airfoil problem. Once the capabilityto analyze the two-dimensional airfoil has beendeveloped and the accuracy verified, it will serveas the basis for developing a fundamental analy-tical capability for treating wings, propellers,helicopter rotors, and complete aircraft con-figurations.

Obviously a large number of computer codesare required in order to develop an overallaircraft icing analysis capability. Figure 2

showsthecodesrequiredandsome(but bynomeansall) of the interfacesrequiredto developaunifiedmethodology.

Theshadedboxesin FigureP indicatethosecomputercodescurrentlyunderdevelopmentandverification. Asthefigure indicates,therearemanycomputercodeswhicharenotcurrentlybeingdeveloped.Whileavailableresourcescertainlylimit the numberof codeswhichcanbedevelopedconcurrently,another,evenmoresignificant,limitation is a lackof knowledgeof appropriatefundamentalphysicsonwhichto basecomputermodels.Forexample,the lackof understandingof basicstructural andfracturepropertiesofimpacticemakeit difficult to developa codewhichmodelsthe pneumaticboot,oneof theoldestof all ice protectionsystems.

Thus,beforecomputercodescanbedevelopedto treat variousaspectsof the aircraft icingproblem,a seriesof basicmodelingexperimentsfirst mustbeconductedin orderthat the keyphysicalelementsbeincorporatedin anysubse-quentmathematicalmodelswhichareformulated.Thisworkis currentlyunderwayin severalareassuchasdeterminingthemechanicalpropertiesofimpactice._3).

Figure2 also pointsout that an integralpart of the icing analysismethodologyunderdevelopmentis the aerodynamicanalysisof bothindividual components and even complete aircraftconfigurations. Since these codes have histori-cally been developed for analysis of aerodynamicperformance in nonicing environments, their useto study icing problems will require that sig-nificant modifications and improvements be made.Obviously, the more robust and accurate thesecodes are, the more robust and accurate the icinganalysis methodology will be.

The development of an aircraft icing analysismethodology requires that a large number of sup-porting benchmark (modeling) and verificationexperiments be run. The experiments require awide range of ground and flight test facilities.Laboratory tests are required to do some of themost fundamental experiments while icing windtunnel tests are needed to acquire data such asice accretion, aero performance, and ice protec-tion system performance for selected components.Dry wind tunnel tests are needed to look at thedetailed flowfield characteristics about airfoilswith artificial leading edge ice accretions.Flight tests are required, both in simulated andnatural icing environments, to determine overallaircraft performance and handling qualitieschanges.

The following sections of the paper willdiscuss in detail the particular icing codesbeing developed as indicated in Figure 2.

Discussion of Icing Computer Codes

Water Droplet Trajectory AnalysisA knowledge of the water droplet trajectories

about any component provides the first indicationof that component's sensitivity to icing. NASAis sponsoring the development of a family of drop-let trajectory codes which can handle a wide

variety of geometries, from simple two-dimensionalairfoils to complete aircraft configurations.Table 1 summarizes the various trajectory codesbeing developed and the geometries which can behandled. As the table suggests, the codes usevarious flowfield analysis methods. Two differenttechniques are used to pass flowfield informationto the trajectory routine. The direct solutionapproach solves for the flowfield at each pointalong each trajectory and the grid generationapproach calculates the velocity at the point ofinterest by interpolating on the two or threedimensional grid generated by the flowfieldsolver.

There are advantages and disadvantages toboth approaches. The direct solution approachcan be more expensive computationally, especiallyas the number of trajectories computed increases.However, this approach frees the user from havingto worry about generating a grid which adequatelyresolves the flowfield acceleration near the bodysurface. In addition, the direct solution approa-ches tend to have more difficulty computing theflowfield around a body such as an airfoil whichhas a leading edge ice accretion. Such calcula-tions are required as part of an ice accretionanalysis as will be discussed later. The gridgeneration approach can handle airfoils with lead-ing edge ice accretions, but the development ofan acceptable grid requires care and skill on thepart of the analyst.

All the trajectory codes indicated in Table 1use a predictor-corrector approach to integratethe droplet equations of motion. Different drop-let drag coefficient expressions are used by thevarious codes, but these differences are judgedto not have a significant effect on the calcula-tion of droplet trajectories.

Historically, water droplet trajectory codeshave been evaluated by comparisons with availableairfoil collection effi#_ncy data obtainedexperimentally by NACA._) In general, thecodes appear to be in reasonably good agreementwith the experimental results which may themselvesbe subject to some appreciable errors. However,this database is limited in scope and in particu-lar, is confined to low speed studies of airfoilsections of interest to the aviation community inthe 1940 to 1955 time period. Also, no experi-mental data had been available for inlet con-figurations of interest.

An experimental program has been initiatedto acquire a comprehensive collection efficiencydatabase for a wide variety of modern airfoil andinlet geomerties which are of interest to theaviation community. The experimental approach, avariation on the dye tracer technique developedby NACA, is discussed in Reference 10.

The major difference from the NACA techniqueis that a laser reflectance technique is used todetermine the local collection efficiency distri-bution curves from the blotter paper strips whichhave been placed at various locations on the testmodel. The NACA approach was to dissolve out thedye from small samples of the blotter paper stripsand then perform a colorimetric analysis of eachresultant water-dye mixture. Figure 3 shows a

2

close-up photo of a quarter scale Boeing 737-300inlet model in the NASA Icing Research Tunnel(IRT) with a series of blotter strips attached atvarious circumferential locations. The laserset-up used to determine the collection efficiencydistribution is shown in Figure 4. A Helium-Neonlaser is used to provide the coherent light source.The amount of light reflected from the blotter canbe related to the local amount of dye on the blot-ter. The dye was dissolved in the spray waterprior to being feed through the spray nozzles.This dye concentration is directly related tolocal collection efficiency. This automated datareduction procedure allows a much larger quantityof data to be reduced relative to the labor inten-sive method which had to be employed by NACAresearchers.

Figure 5 shows measured collection efficiency_ata for the NACA 657-015 airfoil for two anglesof attack (0 ° and 8°7. Also shown are predictedcollection efficiency curves. The results are inclose agreement for the 0 ° angle-of-attack casewhile some differences are noted for the 8 ° case.

Initially, drop impingement tests will beconducted in the NASA Icing Research Tunnel (IRT)using large number of airfoil and inlet models.Once these tests are completed, follow-on testswill be conducted in a high speed icing wind tun-nel to quantify the effects of flowfield com-pressibility on droplet trajectories. Previousanalytical stud_conducted jointly by NASA andthe British RAE_11) have indicated that com-

pressibility becomes important only for the small-est dropsizes. However, this analytical resultmust be verified experimentally.

Some current applications of the trajectoryanalysis codes will be mentioned. Figure 6 showsselected results of three-dimensional trajectoryanalysis(8) for the nose region of the NASAicing research aircraft, a deHavilland Twin Otter.The purpose of this study was to look at theeffects of the flowfield on droplet trajectoriesand therefore, on the measurements of icing cloudliquid water content as measured by various in-struments located on the aircraft. The resultsare shown in the form of concentration factor asa function of droplet diameter. Concentrationfactor indicates the number of droplets per unitvolume at the measurement station ratioed to thenumber of droplets of that size per unit volumein the free stream cloud. As the figure indi-cates, for the selected location, the instrumentsees more water droplets per unit volume thanactually exist in the cloud especially, for arange of droplet sizes between about 20 and 200 _mand thus would indicate a liquid water contentsomewhat higher than the freestream value. Thisanalysis is being used to help in interpretingicing cloud data currently being acquired and toassist in future placement of instruments.

An indepth study of airfoil collection

efficiency characteristics has been investigatedusing the two-dimensional trajectory code ofReference 4. Some 30 different airfoil geometrieswere analyzed for a wide variety of velocity andangle-of-attack conditions for a range of waterdroplet sizes. The results were used to form acomputational database to investigate the effectsof airfoil design parameters such as leading edge

3

radius, camber, and maximum thickness on collec-tion efficiency.(12) In addition, correlationexpressions have been developed to allow quickestimates to be made of airfoil collection effic-

iency characteristics for geometries other thanthose analyzed.

The three-dimensional trajectory analysiscode discussed in References 6 and 7 is beingapplied to inlet configurations which have flow-fields which are highly thre_=dimensional innature. The initial results _IJl are highlyencouraging and indicate complex inlet geometriescan be analyzed for icing sensitivity.

Airfoil Ice Accretion

The NASA airfoil ice accretion analysis code

which is called LEWICE -±_:4_n extension of thework of Lo_ki et al.k J and Ackley andTempleton._ z_) The LEWICE code predicts theice growth rate distribution (and thus ice shape)around the leading edge of the airfoil by locallysolving the quasi-steady enerav balance equation

• (_) The enerfirst proposed by Messinger. gybalance accounts for the governing heat and masstransfer processes thought to occur during theicing process. These terms include convection tothe free stream, latent heat release due to freez-ing, sublimation (or evaporation), and aerodynamicheating. A more detailed discussion of the LEWICEcode is given in Reference 17.

The four major elements of the LEWICE codeare (1) a potential flow panel code to calculate

the airfoil flowfield, (?_ a two-dimensional waterdroplet trajectory code,_ ) (3) an energy bal-ance routine, and (4) integral boundary layerroutine to predict heat transfer distributions.The LEWICE code is constructed so that the flow-field and droplet trajectory calculations arerepeated at user specified time intervals as theice accretion grows on the airfoil leading edge.This updating process attempts to account for theeffect of the leading edge ice accretion on theairfoil flowfield and thus, on the changes inairfoil collection efficiency characteristics.

The current capabilities of the LEWICE codeare summarized in Figure 7 where ice accretionshapes measured on a NACAO012 airfoil in the IRTare compared with LEWICE predictions. Two setsof comparisons are shown for 5 min rime and glazeice accretions. For each calculation, the flow-field and droplet trajectory analyses were updatedevery i min. The agreement is judged to be goodfor the rime ice case, but the LEWICE code overpredicts the glaze ice accretion, although thegeneral shape appears to be good.

The integral boundary layer code has a simpletreatment of the effects of surface roughness onboundary layer transition and heat transfer levels.However, this model is formulated in terms ofequivalent sand grain roughness and the relation-ship to actual ice surface roughness is still aresearch topic, so currently, "educated guesses"must be input to the model. Other models forwall roughness currently are being considered forincorporation in the LEWICE code.

Recent high-speed photo-micrographic moviesof the icing process( 18] have suggested thatsome of the fundamental assumptions made in the

ice accretionmodelmaybe in error. Effortswill shortlybeginto formulateeither changcstothe existingmodelor analternatemodelin orderto seeif ice shapepredictionscanbe improved.

Otherresearchbegunrecentlyin ordertosupportthe ice accretionmodelingeffortsincludethemeasurementof convectiveheattransfercoefficientsfroma smoothandroughsurfaceairfoil bothin lowandhighturbul,nceenvironmentsandapplicationof aNavier-Stokessolverto predictingheattransfercoefficientdistributionsaroundbodieswith ice accretions.

AerodynamicPerformanceDegradationAnalysisAccurate#re_ct_nsof aerodynamTcperform-

ancedegradationof anaircraft dueto icing isone of the desired end products of the icinganalysis methodology. Currently, the developmentand verification of computer codes to predictairfoil performance degradation in icing is theprimary research effort, but some more empiri-cally based approaches are being developed forpredicting degradations of propellers, helicopterrotors, and complete aircraft. The analysis forthe airfoil will provide the basis for developingmore exact analyses for components such as wings,tails, propellers, rotors, and eventually forcomplete aircraft configurations.

Currently, two advanced analysis codes arebeing evaluated for use in predicting iced airfoilperformance - a Navier-Stokes solver and an inter-active boundary layer code. Code predictions arebeing compared to data being gathered in the IRTand in the Ohio State University 0.9 by 1.5 mlow-speed wind tunnel. These codes were chosensince they have both demonstrated the ability totreat flows which have regions of separation-reattachment, a phenomenon which often occurswith ice shapes. Previously, more conventional

airfoil analysis codes were considered, but thecodes all were incapable of handling the icedairfoil problem without major revisions beingincorporated. A summary of these investigationsis given in Reference 2.

The Navier-Stokes code being evaluated isthe ARC2_ code developed the at NASA Ames ResearchCenter.(19) The code has been used to analyzethe flowfields about a large number of airfoilgeometries including airfoils with extended spoil-ers which create a large separated zone. The codecan be run in either of two modes - the fullNavier-Stokes mode or the so-called thin layerNavier-Stokes mode which removes the viscous terms

in the axial direction. To date, most calcula-tions have been done with the thin layer Navier-Stokes mode, although, limited comparisons of thepredictions made with both modes failed to revealany significant differences. The ARC2D code usesthe Baldwin-Lomax turbulence model, and the so-called GRAPE grid generation,_Q_e developed bySorenson, also of NASA Ames. _Lu)

Initial ARC2D calculations were made for a

NACA632-A415 airfoil with a simulated leadingedge glaze ice shape.(21) The ARC2D resultsare compared in figure 8 to measurements made inthe IRT with a general aviation wing section witha NACA632-A415 airfoil section with a woodenleading ice shape affixed. The predicted results(assuming no laminar boundary layer growth) are

judged to be in good agreement with the IRT datawith the exception of the over prediction ofC1. Similar over predictions of C1 withARC2D have been found by othe_Researchers forclean airfoil configurations. _zz) The agreementin drag polars is thought to be exceptionallygood. The velocity vector plot shown in the fig-ure indicates that the ARC2D code predicted siz-able zones of separated flow on both upper andlower surfaces, but reattachment was aslo pre-dicted for both surfaces.

While these initial results suggested thatthe ARC2D code could predict the performance ofairfoils with leading edge ice shapes, the dataacquired in the IRT was judged to be not detailedenough to serve as a validation/verification database. It was determined that such a databaseshould include, in addition to the airfoil per-formance data (Cl, Cd, Cm), much more detailedsurface pressure distributions, velocity and tur-bulence profiles, especially, in the separation-reattachment zones, and f_ow visualization data.In order to acquire this database, an extensiveexperimental program was initiated in the OhioState University low speed wind tunnel. Twosummary papers of that effort are given inReferences 23 to 24.

The airfoil geometry chosen for the programwas a 0.53 m (21 in.) chord NACAO012 airfoil. Aremovable leading edge was fabricated which hadan ice shape which approximated an ice accretionmeasured in the IRT. The ice shape and its rela-tion to the IRT ice shape are shown in Figure 9.The ice shape geometry can be described by a con-stant radius section, two rounded edges of speci-fied radius, and two straight line sections.Thus, an exact specification could be used toinput the geometry into any flowfield analysiscode. The wind tunnel model was instrumented

with a surface static pressure every 1 percentchord in the leading edge region. This largenumber of pressures gave excellent definition ofthe separation-reattachment zones on the surfaces.

Figure 10 presents the ARC2D predictionscompared to the OSU data. Again, the agreementis judged to be quite good for the lower angles-of-attack. However, it was observed that theARC2D code did not converge to a steady statesolution for angles of attack greater than about6 ° . Currently, flow visualization experimentsare being conducted to determine if this is purelya computational result, or if the experimentalflowfield is not steady at these higher angles-of-attack.

To date, velocity and axial turbulenceintensity profiles have been measured every2 percent chord for the upper surface separation-reattachment zone. Figure 11 shows two selectedprofile comparisons - one just downstream of themeasured point of separation and the other justupstream of the experimental reattachment point.The ARC2D predictions are in reasonable agreementwith the data; however, the predicted reattachmentoccurred more upstream than was observed in theexperiment. Investigations are underway toimprove the code predictions. Areas of investi-gation include the sensitivity to ice shape defi-nition and grid generation as well as to theturbulence model employed.

4

A similar evaluation of the capabilities ofthe interactive boundary layer code was begunmore recently. This code was developed by Cebeciand associates under grant to the National ScienceFoundation(25) to predict clean airfoil behavior

including the higher angles-of-attack and theonset of stall. The code has two major modules -a potential flow panel routine to calculate theinviscid region and an inverse finite differenceboundary layer routine which is capable of treat-ing regions of separation-reattachment. The tworoutines are coupled together to solve furtherairfoil flowfield by iterating back and forthuntil a converged solution is acquired.

Initial studies have been completed for lowerangle-of-attack results for the "iced" NACAO012airfoil already discussed with comparisons made

"of airfoil performance (i.e., Cl, Cd). The pre-dictions were in very close agreement with theexperimental data. Some oscillations in the finalsolutions obtained suggested some additionalimprovements are required in the numerics. Theseimprovements are currently being made and themore detailed velocity profile comparisons arealso being made. Also, comparisons are beingmade for the higher angle-of-attack conditions.

Currently, it is judged that both the Navier-Stokes and interactive boundary layer methodsshow great promise for being able to predict icedairfoil performance. It is felt that the twomethods are more complementary than competitiveas both can be used for icing analysis. TheNavier-Stokes approach, while computationallymore expensive (approximately 10 to 20 min persolution on the Cray XMP) is thought to be capableof treating larger, more extreme ice shapes.Navier-Stokes results, possibly, can be used tohelp in the more approximate modeling in theinteractive boundary layer code. The interactiveboundary layer code is much faster (approximately4 to 30 sec per solution on the Cray XMP) andwould be very attractive for those problems whichrequire many aero performance calculations to bemade (e.g. the helicopter rotor in forwardflight).

Besides those areas of improvements currentlybeing investigated for both codes which have beendiscussed, each will require a model to be devel-oped to handle the extreme levels of surfaceroughness which exist for any ice shape. At pre-sent, neither code can handle anything other thansmooth surfaces. Several approaches for treatingsurface roughness are being evaluated. Regardlessof the model employed, some means must be foundfor quantifying the nature of the ice shaperoughness. Currently, the equivalent sand grainapproach is employed to treat rough surfaces, butit is not clear that that approach Would work forice accretion roughness.

As already indicated, more approximate meth-ods have been developed for predicting the aeroperformance degradation of propellers, helicopterrotors (hover and forward flight), and completeaircraft. References 26 to 30 review theseefforts. These various efforts all use the sameapproach - a set of correlation equations is usedto predict airfoil performance degradation due toicing and, then an appropriate performance code

is used to determine the total component or air-craft performance degradation.

The correlation equations are relationsdeveloped from an airfoil icing data base whichrelate the change in lift, drag, and pitchingmoment to known aerodynamic and icing variables.The first suc_#Rrrelations were developed byGray of NACA._a±j More rece_#ly, Bragg devel-oped a rime ice correlation. _"j Both of thesecorrelations, used data gathered in the NASA _RT.Flemming,(29) acquired a large data base in theCanadian NRC high speed icing wind tunnel for aseries of reduced scale rotor airfoil sectionswhich he used to develop a series of correlationequations.

Unfortunately, it is not possible to a

priori select one set of correlation equationswith assurance that the predictions will yieldaccurate results. Studies made with the abovementioned sets of correlations have determinedthat the errors can be as large as 100 to200 percent for selected conditions. Some ofthis error is undoubtedly due to the rather limi-ted airfoil icing data base which exists. Also,questions can be asked about how the icing resultsfrom one facility can relate to those from anotherfacility which were calibrated using differenttechniques. Thus, extreme care must be exercizedwhen using these correlations. Miller(28) givesan excellent discussion of typical effects of

using the different correlation equations topredict propeller performance.

In spite of the above concerns, some reason-able results have been acquired. Figure 12 showspropeller performance predicted in rime icingconditions when compared to natural icing datagathered by NACA researchers. The propeller effi-ciency variation as a function of advance ratiomeasured in flight icing conditions was closelyduplicated by the analytical methodology for thisparticular set of tests. As Reference 26 indi-cates, the agreement was not as good for some ofthe other rime icing encounters.

Korkan, et al., used a similar analytical

approach_analyze the helicopter rotor_2_irstin hover_ _°j and then in forward flight._The particular rotor configuration analyzed wasthe front rotor of the Boeing Chinook CH47.Unfortunately, no experimental data existed forcomparison with the predictions. All that couldbe said was the results were reasonable when com-pared with previously reported torque rises mea-sured in icing for other heliocpter configurations.

Flemming used the previously discussed air-foil performance in-icing correlations along withappropriate rotor performance codes to predictrotor degradation for the S-76, UHIH, and UH-60Ahelicopters. The predicted results compared toavailable experimental data are shown in Fig-ure 13. Generally the agreement was better forthe hover comparisons than for the forward flight.As Figure 13 indicates, the forward flight pre-dictions were generally lower than the experimen-

tal measurements. At present_ it is not knownwhy this is true, although, as Flemming pointsout,(29) it is not clear how accurate are theexperimental measurements. These results point

out the needfor dedicatedicing flight programsfor acquiringhighquality verification data.

Figure14showsselectedresults fromthepredictionof overall aircraft performancedegrad-ation in icing usingtheNASAcode(30)comparedtodatagathered_2_art of the NASAicing researchflight program_J usingtheTwinOtter4) Theairfoil correlationdevelopedbyBragg wasusedto determineairfoil performancedegradation.Eventhoughthis correlationwasdevelopedi.ohandlerimeicing conditions,thefigure indicatesit appearedto providereasonableinputsfor themixedicing conditionsof the icing flight shown.Thecode,ascurrentlyconfigured,canpredictaeroperformancedegradationsonlyfor liftingcomponents(e.g. wings,tails, struts). Thecon-tributionsof othercomponentssuchaswheels,flap'hinges,antennas,etc., mustbeestimatedandinput. Forthepredictionsshownin Figure14,this contributionwasestimatedbyforcing thecodeto agreewith thedragincreasemeasuredforthe completelyicedaircraft for the lowestliftcoefficientdatapoint (filled symbolin thefig-ure). Thisamountof dragincreasewasnotvariedbutkeptconstantfor all th_othercode--p-re_Cztions. Notethereasonableagreementbetweentheoryandexperimentnotonlyfor thefully iced,but alsofor thewingandwingplustail deiceddata. Additionalcomparisonsof the codepredic-tions usingthe variousairfoil correlationsavailablewith additionalnaturalicing flightdataareplanned.Also, simplemethodsfor esti-matingthe dragincreasedueto nonlifting com-ponentsarebeingconsidered.Ice ProtectionSystems...... -NASAhassponsoredthedevelopmentof aseriesof computercodeswhichmodeltheelectro-thermaldeicer, a systemcurrentlybeingusedformanyhelicopterrotor andfixedwingaircraftapplications.Table2 givesthe importantcharac-teristics of the six different codesdevelopedtodateandhowtheydiffer fromoneanother. Asthetable suggests,thecodesvarygreatlyincomplexityandtherefore,in the sizeof computerrequired. Codes1 and2 whichmodeltheelectro-thermaldeicerona one-dimensional,timevaryingbasiscanberunona personalcomputer.Code6whichanalyzesthecompletetwo-dimensionalgeo-metry,includingthevariablethicknessice layerandthemanylayersof the airfoil geometryrequiresa supercomputer(i.e. aCray1Sor XMP)to achieve reasonable run times. A unique featureof all codes is that they treat the melting ofthe ice through a phase change routiner_AC_s 2through 6 use the weawhile code 1 uses an p

Currently, the code predictions are beingcompared with electrothermal deicer data fromicing wind tunnel and helicopter natural icingflight tests. These comparisons are aimed atproviding a better understanding of how accu-rately the various codes can predict electro-thermal deicer performance for various levels ofmodeling complexity.

Figure 15 shows the test model used in arecent IRT test. The model was a section of aBell Helicopter Textron UHIH rotor (0.53 m (21 in.chord)) with a spanwise electrothermal deicerconfiguration installed. The deicer, provided by

the BF Goodrich Company, is typical of a flightdesign. The model was instrumented with a largenumber of thermocouples to measure the transienttemperature response at various locations in themodel. The test model was run over a wide variety

of dry, wet (above freezing), and icing condi-tions, and a large data base was acquired.Reference 36 gives a description of the testprogram.

Figure 16, taken from Reference 38, showsone selected result of the comparison of code 2predictions with experimental measurements. Asthe figure shows, thermocouples were located at(I) the abrasion shield surface, (2) the heater,and (3) the internal spar of the model. For thiscomparison, the model predictions were judged tobe in close agreement with the experimental datafor all three cycles of electrothermal deiceroperation. The model predicted a melting of theice for the second and third heater cycles, and

the experimental results showed the same occur-rence. While the agreement was not as good forall cases as the one shown here, the one-dimensional code predictions were found to begenerally in good agreement with the data.Currently, efforts are underway to make similarcomparisons with the other codes to see if theincreased complexity in the numerical modelingcan give better agreement with the experimentalresults.

Ice shedding is a key aspect of electro-thermal deicer performance. The one-dimensionalmodel of Reference 34 attempts to model this

phenomenon by assuming that shedding occurs whenthe thickness of melted ice reaches a user speci-fied amount. A better treatment of shedding maybe required.

Another ice protection system, the fluidfreezing point depressant system, has beenmodeled. A simple engineering procedure for pre-dicting the minimum anti-icing flowrates requiredhas been developed. (39) The code predictionscompared with flowrates measured in the IRT fora full-scale general aviation wing section modelequipped with this system are shown in Fig-ure 17. In general, the method is judged to givereasonable agreement with th_ experimental data.As Reference 39 points out, some of the error canbe attributed to the experimental values measuredfor minimum anti-icing flowrates. This deter-mination based on visual observations was foundto sometimes vary significantly for repeat testconditions and was also sensitive to the inter-

pretations of different observers as to when theminimum flow rate was achieved. Given theseinherent uncertainties in the experimentalapproach, the predictive method is felt to beacceptably accurate.

The third ice protection system beingmodeled is the electromagnetic impulse deicer(EIDI). EIDI is a concept which is receiving

much attention, both in the United States andEurope. It shows great promise for being a lowpower consumption, highly efficient deicing sys-tem. NASA has sponsored an intensive researcheffort, the results of which are summarized inReference 40, to acquire an experimental database to aid in the design of the system forvarious aircraft applications.

._,,_.........POOR QL?L_L!'Pi

Work is also underway to develop the required capability have been developed, many validation/structural and electrodynamic codes to model thekey features of the EIDI system. The currentstatus of the structural dynamic modeling capa-bilities is discussed in Reference 41. The simpleconfiguration which is being modeled is shown inFigure 18. The half cylinder is meant to model awing leading edge and has a single 2.54 cm radiuscoil located 18 cm (0.07 in.) beneath the surface.Finite element model predictions for surfaceacceleration at two locations and the circumferen-

tial and longitudinal strains at the coil locationare shown in the figure. Agreement in all casesis judged to be at least reasonable and surpris-ingly good in some instances (e.g. peak circum-ferential strain).

A key input into the structural dynamicanalysis is the force produced by the coil. Botht'her temporal and spatial characteristics of thisforce must be known. Until recently, no analyti-cal solutions had been obtained for this forcefunction, and thus experimental data had to beused. The results shown in Figure 18 wereacquired using a simple force model. However,recently, a transmission line model obtained fromthe governing field theory equations has beenachieved which appears to yield good predictionsof the force for the case of a.coil with its axisperpendicular to a flat plate.(42) Figure 19shows a comparison of the force versus time pre-dicted by this transmission line model with theforce versus time calculated from the measuredinduction fields for a 3.08 cm (2 in.) diameter0.478 cm (0.188 in.) thick coil rigidly mountednext to a fixed 0.0813 cm (0.032 in.) thick 2024T3 Aluminum disc with a 12.7 cm (5 in.) diameter.It is hoped that the use of this force model willimprove theoretical predictions such as thoseshown in Figure 18.

While the modeling of the EIDI system showsgreat promise, the effort is still in its infancy.A better understanding of the ice adhesion prop-erties and removal mechanisms must be gainedbefore a complete modeling of the system can beattempted. Also, the EIDI system modeling formost applications will probably require three-dimensional structural analysis codes.

As already indicated, NASA has sponsoredefforts to measure some of the key physicalproperties of ice(3) which are required asinputs into any surface deflection type iceprotection system model. The results are beingused along with a finite element analysis approachto begin a quasi steady modeling of a genericpneumatic boot ice protection system. Companionexperiments are being planned to measure the icefracture process of the generic boot configurationfor comparison with modeling predictions.

Concluding Remarks

This paper has presented a current statusreview of the progress being made toward devel-oping and validating the computer codes requiredto assemble an aircraft icing analysis method-ology. While it is felt that much progress hasbeen made since the initial review presented inReference 2, much work remains to be done. Whilemany of the codes which are needed to form a core

verification experiments are still required. Thisis especially difficult (and expensive) when itcomes to determining the performance in icing ofcomplex configurations such as helicopter rotorsand complete aircraft configurations. Analysisof the steady state and dynamic behavior of thosecomplex configurations in icing may require sig-nificant levels of empiricism to be incorporatedin the modeling.

Nevertheless, it is felt that once developedand validated, this icing analysis capability willbe of great benefit to the industry in reducingthe fixed and rotary wing development and certi-fication costs for future aircraft configurationswhich are required to safely fly into forecasticing conditions.

References

1. Reinmann, J.J.; Shaw, R.J.; and Olsen, W.A.,Jr.: Aircraft Icing Research at NASA. NASATM-82919, 1982.

2. Shaw, R.J.: Progress Toward the Developmentof an Aircraft Icing Analysis Capability.NASA TM-83562, 1984.

3. Scavuzzo, R.J.; Chu, M.L.; and Olsen, W.A.:Structural Properties of Impact Ices. AIAAPaper 86-0549, Jan. 1986.

4. Bragg, M.B.: Rime Ice Accretion and ItsEffect on Airfoil Performance. NASACR-165599, 1982.

5. Chang, H.P., et al.: Influence ofMultidroplet Size Distribution on IcingCollection Efficiency. AIAA Paper 83-0110,Jan. 1983. f

6. Kim, J.J.: Particle Trajectory Computation ona 3-Dimensional Engine Inlet. NASA CR-175023,1986.

7. Norment, H.G.: Calculation of Water DropTrajectories To and About Arbitrary Three-Dimensional Bodies in Potential Airflow.NASA CR-3291, 1980.

8. Norment, H.G.: Calculations of Water DropTrajectories To and About Arbitrary Three-Dimensional Lifting and Nonlifting Bodies inPotential Airflow. NASA CR-3935, 1985.

9. Gelder, T.F.; Smyers, W.H., Jr.; andVonGlahn, U: Experimental DropletImpingement on Several Two-DimensionalAirfoils with Thickness Ratios of 6 to16 Percent. NACA TN-3839, 1956.

i0. Papadakus, M., et al.: An ExperimentalMethod for Measuring Droplet ImpingementEfficiency on Two-and-Three-DimensionalBodies. AIAA Paper 86-0646, Jan. 1986.

11. Gent, R.W.: Calculation of Water DropletTrajectories About an Aerofoil in Steady,Two-Dimensional, Compressible Flow. RAETR-84060, June 1984.

12. Bragg, M.B.; and Gregorek, G.M.: AnAnalytical Evaluation of the Icing Propertiesof Several Low and Medium Speed Airfoils.AIAA Paper 83-0109, Jan. 1983.

13. Kim, J.J., et al.: Sand Separate EfficiencyCalculation for the JUX T.H. Rake AircraftInlet. Presented at the 42nd American

Helicopter Society Forum, June 2-4,1986.

14.Lozowski,E.P.;Stallabrass,J.R.; andHearty,P.F.: TheIcing of an UnheatedNon-Rotating Cylinder in Liquid Water Droplet:Ice Crystal Clouds. National ResearchCouncil of Canada, LTR-LT-96, Feb. 1979.

15. Ackley, S.F.; and Templeton, M.K.: ComputerModeling of Atmospheric Ice Accretion, CRRELReport 79-4, Mar. 1979.

16. Messinger, B.L.: Equilibrium Temperature ofan Unheated Icing Surface as a Function ofAirspeed. J. Aeronaut. Sci., vol. 20, no. i,Jan. 1953, pp. 29-42.

17. MacArthur, C.D.: Numerical Simulation ofAirfoil Ice Accretion. AIAA Paper 83-0112,Jan. 1983.

18. Olsen, W.A., Jr.; Walker, E.D.; and Sotos,R.G.: Microscopic High Speed Movies Showingtile Droplet Freezing Process of Icing. AIAAPaper 84-0019, Jan. 1984.

19. Pulliam, T.H.: Euler and Thin Layer Navier-Stokes Codes: ARC2D, ARC3D. ComputationalFluid Dynamics, University of Tennessee,UTSI Publication No. E02-4005-023-84, 1984,Section 15.1.

20. Sorenson, R.L.: A Computer Program toGenerate Two-Dimensional Grids About Airfoilsand Other Shapes by the Use of Poisson'sEquation. NASA TM-81198, 1980.

21. Potapczuk, M.G.; and Gerhart, P.M.: Progressin the Development of a Navier-Stokes Solverfor Evaluation of Iced Airfoil Performance.

AIAA Paper 85-0410, Jan. 1985.22. Mehta, U.; Chang, K.C.; and Cebeci, T.: A

Comparison of Interactive Boundary Layer andThin-Layer Navier-Stokes Procedures.Numerical and Physical Aspects of AerodynamicFlows III, T. Cebeci, ed., Springer-Verlag,1986.

23. Bragg, M.B.; and Coirier, W.J.: DetailedMeasurements of the Flowfield in the Vicinityof an Airfoil with Glaze Ice. AIAA Paper85-0409, Jan. 1985.

24. Bragg, M.B.; and Coirier, W.J.: Hot FilmMeasurements of the Separation Bubble on anAirfoil with Glaze Ice. AIAA Paper 86-0484,January 1986.

25. Cebeci, T., et al.: Airfoils with Separationand,the Resluting Wakes. Third Symposium onNumerical and Physical Aspects of AerodynamicFlows, California State University, LongBeach, CA, 1984, pp. 2-13 to 2-25.

26. Korkan, K.D.; Dadone, L.; and Shaw, R.J.:Performance Degradation of Propeller/RotorSystems Due to Rime Ice Accretion. AIAAPaper 82-0286, Jan. 1982.

27. Korkan, K.D.; Dadone, L.; and Shaw, R.J.:Performance Degradation of Helicopter RotorSystems in Forward Flight Due to Rime Ice

Accretion. J. Aircr, vol. 22, no. 8,Aug. 1985, pp. 713-718.

28. Miller, T.L.: Analytical Determination ofPropeller Performance Degradation Due to IceAccretion. NASA CR-175092, 1986.

29. Flemming, R.J.; and Lednicer, D.A.: HighSpeed Ice Accretion on Rotorcraft Airfoils.NASA CR-3910, 1985.

30. Gregorek, G.M., et al.: NASA Twin OtterFlight Test Program - Comparison of FlightResults with Analytic Theory. SAE Paper No.850924, Apr. 1985.

31. Gray, V.H.: Prediction of AErodynamicPenalties Caused by Ice Formation on VariousAirfoils. NASA TN D-2166, 1964.

32. Ranaudo, R.J., et al.: PerformanceDegradation of a Typical Twin Engine CommuterType Aircraft in Measured Natural IcingConditions. AIAA Paper 84-0179, Jan. 1984.

33. DeWitt, K.J. and Baliga, G.: NumericalSimulation of One-Dimensional Heat Transferin Composite Bodies with Phase Change. NASACR-165607, 1982.

34. Marano, J.J.: Numerical Simulation of anElectrothermal Deicer Pad. NASA CR-168097,1983.

35. Chou, D.F.K.: Numerical Simulation ofTwo-Dimensional Heat Transfer in CompositeBodies with Application to De-lcing ofAircraft Components. NASA CR-168283, 1983.

36. Masiulaniec, K.K., et al.: Full Two Dimen-sional Transient Solutions of ElectrothermalAircraft Blade Deicing. AIAA Paper 85-0413,Jan. 1985.

37. Leffel, K.L.: A Numerical and ExperimentalInvestigation of Electrothermal AircraftDeicing. NASA CR-175024, 1986.

38. Leffel, K.L., et al.: A Numerical andExperimental Investigation of ElectrothermalAircraft Deicing. Presented at the AmericanHelicopter Society 42nd Forum, June 2-4, 1986.

39. Albright, A.E.: Experimental and AnalyticalInvestigation of a Freezing Point DepressantFluid Ice Protection System. NASA CR-174758,1984.

40. Zumwalt, G.W., et al.: Analysis and Testsfor Design of an Electro-lmpulse De-lcingSystem. NASA CR-174919, 1985.

41. Bernhart, W.D.; Gien, P.H.; and Wilson,B.K.: Structural Dynamics InvestigationsRelated to EIDI Applications. AIAA Paper86-0550, Jan. 1986.

42. Henderson, R.A.: Theoretical Analysis of theElectrical Aspects of the Basic Electro-Impulse Problem in Aircraft De-icingApplications. Ph.D. Dissertation, WichitaState University, 1986.

TABLE 1. - DROPLET TRAJECTORY ANALYSIS CODE CHARACTERISTICS

Geometries Flowfield analysis code used Comments iReference

I. Single Element Airfoils

2. Single Element Airfoils

3. Multiple Element Airfoils

4. 2D, Axisymmetric Inlets

5. 3D Inlets

6. Complex Configurations(incl. Complete Aircraft)

Theodorsen Transformation

Navier-Stokes

Panel Method

Panel Method

Transonic, Full Potential

Panel Method

Incompressible Flowfield,Direct SolutionCompressible Flowfield,Grid Generation

Incompressible Flowfield,Direct Solution

Incompressible Flowfield,Direct Solution,2D TrajectoriesCompressible FlowfieldGrid GenerationIncompressible Flowfield,Direct Solution

4

5

5

5

6

7,8

TABLE 2. - ELECTROTHERMAL DEICER CODES

Geometry Ice thickness Number of heaters Phase change Reference

I. one Dim#nsional2. One Dimensional3. Two Dimensional

(Rectilinear)4. Two Dimensional

(Rectilinear)5. Two Dimensional

(Rectilinear)6. Two Dimensional

(Exact)

ConstantConstantConstant

Constant

Variable

Variable

OneOneOne

Variable

One

Multiple

ApproximateWeak EnthalpyWeak Enthalpy

Weak Enthalpy

Weak Enthalpy

Weak Enthalpy

333435

37

36

AIRCRAFTPERFORMANCE,HANDLINGQUALITIESCHANGES

ICE PROTECTIONSYSTEMSPERFORMANCE

ICEPROTECTIONSYSTEMDESIGN

COMPONENTDESIGN

CS-86-1714

Figure 1. - Uses of aircraft icing analysis.

FLIGHT, /

E"V%",T_O_A'_I BODY

I o_o_.__' I

HOT GAS ]SYSTEM

ICE SHEDDING I

IME_.A.,OAL] _.EOMAT,OVIBRATION BOOT

SYSTEM SYSTEM

ICEPHOBICS

I ON ICE

,, J

NOTE:

BY NASA LeRC

CD-85-15526

Figure 2. - Aircraft icing analysis methodology.

OF POOR QU?,J_!_'CI

Figure 3. - Boeing 737-300inlet n Icing Research Tunnel with blotter strips attached.

OR|GIN_ _::_:'_ '_i_i

OF. POOR QUaLiTY

, Figure 4. - He-Ne laser system to reduce drop impingement blotter samples.

Io0a =8o

.8

.6

.4>-C_)Z

.2C_)

L_J

z 00

LJ___ 1.0̧0

0

.6'

0

!

THEORY/_ EXPERIMENT

2 4 6 8 i0SURFACECOORDINATE,~cm

a : 0°

-4 -2 0 2 4 6SURFACECOORDINATE,~cm

Figure 5. - Preliminary collection efficiency results for NACA6.52-015airfoil. Voo =82 m/sec; MVD = 15pm.

r,,,"O

ZC)

t---

m,,I--Z

(..)ZO(..)

1.2

1.1

1.0

"- INSTRUMENT

-- LOCATION ,/_

I I0 I00 i000DROPLETDIAMETER,~pro

Figure 6. - Concentration factor calculations for NASA icingresearch aircraft, Voo= 67 mlsec.

CALCULATED COMPARISON EXPERIMENTAL

LWC=1.02gm/m3 MVD=12pm Voo=52m/sec Too=-26°C

CALCULATED COMPARISON EXPERIMENTAL

LWC=1.20gm/m3 MVD=20pm Voo=89m/sec Too=-11°C

' Figure7. - LEWICEpredictions of airfoil ice accretions comparedto TRTdata. NACA0012, 0..53mchord, 5 minute ice accretion, 0° angle of attack.

/

1.50

,25 --

-i. 00-I0. 0

• ii0

.055

J

-_ _. -_'"_-- SEPARATION-REATTACHMENTZONES

_T DATA

-D- ARC2D PREDICTION

I I2.5 15Cl

_Pl

-0- IRT DATA I--El.- ARC2D I

PREDIC-(__

-- TION ,,

o I I-.5 .5 1.5

C_

Figure8.- Navier-Stokesanalysis

of NACA632-A415airfoil withglaze ice (ref. 21).

\\

• i0XIC

I.20

MEASUREDIN IRI

SIMULATEDC$-86-1715

Figure 9. - Comparison of measuredand simulated ice shapesNACA0012airfoil, 0..53m (21 in. ) chord.

1.0

5

0

.i0

.05

0

.i

E 0

-.i

.20

THEORY(ARC2D)-- O EXPERIMENT

NO CONVERGEDSOLUTIONSFOR a >6° --_

\O0

00

0

0

I0 5

a, deg

00

0

I10

Figure 10. - Comparison of measuredand predicted airfoil performancefor NACA0012airfoil with artificialice shape.

DIVIDINGSTREAMLINE(EXPERIMENTAL)_

\\

/

y/c

_

j._ PREDICTED-_ II ,L I _I

I, I

[ ....9' ' >'....I yJc \"\ 1#

u/u e ,----...,._....._ /

__. U_Ue

CS-86-1717

Figure 11. - Comparison of measured and predicted velocity profiles using ARC 2Dcode.

1.0

.9>-

(J2:"' .8

mL.I.-I.a--

'" 7

,._1i.i

o .6

13_

.5

.4

-- /- CLEAR /-CLEARi (EXPERIMENTAL) / (THEORETICAL)I

-- / /

(EXPERIMENTAL) I ICE(THEORETICAL)

I I I I I I I1.0 i.i 1.2 1.3 1.4 1.5 1.6

ADVANCE RATIO1.7

Figure 12. - Comparison of measuredand predictedpropeller performance in rime icing conditions(ref. 26).

_D

o

oI.LIl-.-

r_1.1-I

3O

2O

i0

0

20

i0

0 S-76 '81[] S-76 '82A UH-IH /_

$7 UH-60A /

oo-_za

I_ [] / x__LINEOF PERFECT

[] AGREEMENT,

>1 _/ D I I I(al Hover.

- / \

\\ LINEOF PERFECT

_--_1_ _1 I Ii0 20 30 40

OBSERVEDTOROUERISE, percent

(b) Forward flight.

Figure 13. - Comparisonsof measured and predictedhelicopterperformance in natural icing (ref. 29).

•12

.10

•08@

O6

.04

•020

ORIGIN:t:L P_;_:_::_i_.C

OF POOR QUAISTY

EXPT THEORY

-- -mOB FULLYICED

---El--- WING DEICEDW + T DEICED

0 CLEAN _-...___ _.- -

ooo4#o ooo <_@o

I I I I I.2 .4 .6 .8 1.0

Figure 14. - Comparison of predicted and measured aircraft per-formance in natural icing conditions (ref. 30).

ELECTROTHERMAL DE-ICER EXPERIMENT IN IRT

6D.-15-16183

Figure 15. - Electrothermal de-icer experiment in I RT.

O

I_1_1

ICE (~0. 16 cm)

_////,_ T3_ SH,Em,,_I...,.............,._--.-HH_J HEATER

INSULATION--4-1 _T2___ I

TI (NOTTO SCALE)

'I

EXPERIMENT

THEORY

TI

F-ICE MELTING

0 i00 200 300

B

T2

# /

r

k_

T3

I I I I I0 100 200 300 0 100 200

TIME, sec

Figure 16. - Comparison of transient temperature profiles measured in IRT with predictions of one dimensional, transientheat conduction code(ref. 38).

I300

E(_}

r-.m

E.....,.

I.L

...II..i,_

r-,,

.-..IL.I..

(..)

rml_i_l

rv,CL

>,.,-.I

0

l.--

Z,:IZ

•28

•24

.20

.16

.12

•08

.04

m/

V : 96, 145KNOTS /a:2.1 °, 5.90 /T =25 OF,5 OF //

0 ///

-- 0 // 0///

_+25%/0 / /"

i/ J ///

o/6_b///"_25 _0-- / / i/"

o/ / .;40 Jr,O_//'-'./f L LINEOF PERFECT

- __E_'ME_i .....

_._"'o1_ I1 I I I I

•04 .08 .12 .16 .20 .24

AVERAGED ANTI-ICESPECIFICFLUIDFLOW, SFF,glmincm2

Figure17.- Comparisonofpredictedand measured

minimum anti-icing flow rates for NACA2412(rood)airfoil with fluid freezing point ice protec-tion system (ref. 3g).

800z"

_._Z___ 400

Z"'{::3

"_C3_- 0

r_

-400

800

Z

rv' Z ,I-.-- _< 400._.jr"*'

Z m

¢.v--n (_.__--- 0

0--1

2.54cm RADIUSCOIL --x\

\\

_ ,,

z--O.18cm GAP

-4000

_- /,_x' o s.oMEASURED

, / "_XX___ PREDICTED z"

o__ 2.5,,::E

i.J.J

I I °" z -5.0CIRCUMFERENTIALSTRAINAT COIL

I I I I INORMALACCELERATIONAT COIL

z_ 1.5

t---

eY

_ 0_,,J

-- -1.5

r_

I I I I I °:_ -3.

200 400 600 800 1000

LONGITUDINALSTRAINAT COIL

TIME, psec

m

.._f/'_ _ _\\\

I I I _/'I ''' I0 100 200 300 400 500

NORMALACCELERATION11.4 cm FROMCOIL

Figure 18. - Comparison of measured and predicted accelerations and strains for idealized deicer model(ref. 41).

,&0

'-rl--°

I I_ I I I _o_J__ I i I i

o_'o _ _ __o ! I _:0_ 1 _

- 91 __

_ I !

1. Report No.

NASA ]M-88791

2. Government Accession No.

4. Title and Subtitle

NASA's Aircraft Icing Analysis Program

7. Author(s)

Robert J. Shaw

9. Performing Organization Name and Address

National Aeronautics andLewi_ Research CenterCleveland, Ohio 44]35

Space Administration

12. Sponsoring Agency Name and Address

National Aeronautics and

Washington, D.C. 20546

Space Administration

3. Recipient's Catalog No.

5. Report Date

6. Performing Organization Code

505-68-li

8. Performing Organization Report No.

E-3121

10. Work Unit No.

11. Contract or Grant No.

13. Type of Report and Period Covered

Technical Memorandum

14. Sponsoring Agency Code

15. Supplementary Notes

Prepared for the International Conference of theLondon, England, September 7-12, 1986.

Aeronautical Sciences (ICAS),

16. Abstract

An overview of the NASA ongoing efforts to develop an aircraft icing analysis

capability is presented, Discussions are included of the overall and long termobjectives of the program as well as current capabilities and limitations of the

various computer codes being developed. Descriptions are given of codes being

developed to analyze two and three dimensional trajectories of water droplets,

airfoil Ice accretion, aerodynamic performance degradation of components and

complete aircraft configurations, electrothermal deicer, fluid freezing point

depressant antldelcer and electro-lmpulse deicer. The need for bench mark andverification data to support the code development is also discussed and selected

results of experimental programs are presented.

17. Key Words (Suggested by Author(s))

Aircraft icing; Analytical methods;

Computer programs

18. Distribution Statement

Unclassified - unlimitedSTAR Category 03

19. Security Classif. (of this report) 20. Security Classif. (of this page)

Unclassified Unclassified21. No. of pages 22. Price*

*For sale by the National Technical Information Service, Springfield, Virginia 22161