Document
.ff'Z'T7' ' 2 L,
N90-20928
THE INFLUENCE OF ICE ACCRETION PHYSICS ON THE FORECASTING OF AIRCRAFF ICING CONDITIONS * + R. John Hansman, Jr.
Department of Aeronautics and Astronautics Massachusetts Institute of Technology ABSTRACT Because the physics of mixed and glaze icing are similar, the term mixed icing will refer to both mixed and glaze unless The physics which control aircraft ice accretion axe otherwise noted.
reviewed in the context of identifying and forecasting hazardous icing conditions. The severity of aircraft icing is The physical processes which control ice accretion found to be extremely sensitive to temperature, liquid water content and droplet size distribution particularly near the are distincdy different for dry and wet ice growth. In dry transition between rime and mixed icing. The difficulty in growth, where the droplets freeze on impact, the ice measurement and the variability of these factors with accretion is controlled by the local rate of impingement of altitude, position and time coupl_ with variable aircraft liquid water on the surface. The local impinging mass flux sensitivity make forecasting and identifying icing conditions is an inertially determined quantity which involves the difficult. Automated Pilot Reports (PIREPS) are suggested individual droplet trajectories as they pass through the as one mechanism for improving the data base necessary to flowfield surrounding the body. For wet growth, the ice forecast icing conditions.
accretion is controlled by the rate at which latent heat of fusion can be removed from the surface. The heat transfer 1. INTRODUCTION behavior of the ice surface, therefore, becomes the controlling mechanism for wet ice growth. For mixed icing The accurate and precise forecasting of conditions both the impingement and heat transfer meteorological conditions favorable to aircraft ice accretion is mechanisms play important roles in the ice accretion process.
difficult for several reasons. First, the type and severity of ice accretion is often strongly or nonlinearity dependent on environmental parameters such as temperature, liquid water RIME ICE content, cloud droplet size distribution, a:rbuleace level and water phase. Secondly, several of these parame:ers are difficult to measure or estimate in the forecasting environment. Finally the severity of the ice accretion and its influence on aircraft performance will depend on both the v= type and flight condition of the aircraft.
This paper will examine those meteorological and aircraft factors which influence the physics of ice accretion in order to gain some insight into the limitations of current forecasting techniques and the requirements for improved icing forecasts.
MIXED ICE 2. OVERVIEW OF THE ICE ACCRETION PROCE_ The ice accretion process is controlled by two physically distinct subprocesses. The rust is the inertial transport of liquid water (either in the form of cloud droplets, rain drops or mixed phase hydrometeors) from the ambient environment to the aircraft surface. Once droplets have impacted the aircraft surface, their freezing becomes controlled by thermodynamic processes. If the heat transfer from the surface is sufficient to remove all the latent heat of freezing of the impinging water than the droplets will freeze on impact resulting in a dry ice surface. The ice shape typically protrudes forward into the airstream and is commonly referred to as rime ice (see Figure 1).
Fig. 1 Typical dry (rime) and wet (mixed/glaze) ice shapes.
When the heat transfer from the surface is inadequate to remove all of the latent heat from the impinging droplets 2.1 Wata" Impingement the ice surface becomes wet. This type of accretion is commonly characterized as glaze ice. In some cases, both The physics which control the impingement of liquid wet and dry ice growth can occur at different places on the water onto the aircraft surface are fairly straightforward.
same body. This situation is referred to as mixed ice Because the surface is not permeable to gas, flow growth. Often in glaze or mixed conditions, the resulting ice m'eamlines do not intersect the body as can be seen in Figure shape displays two pronounced growth peaks on either side 2. Water droplets have a higher ratio of inertial to of the stagnation line (see Figure 1). The most severe hydrodynamic forces than gas molecules and will tend to aircraft performance degradation is typically associated with cross the streamlines resulting in impingement as shown in Figure 2.
such horned ice formations (Renaudo, et al., 1984).
* Preprint, Third International Conference on the Aviation Weather System, 1989.
+Some original figures were not available at time of publication.
3.0 INFLUENCE OF METEOROLOGICAL
Much work has been done ondroplet irnpingement
PARA_
trajectories todetermine local collection efficiencies B
(Bergrun, 1947, Bragg, eta1.,1981, Brun, etal,1953, Gelder, etal,1956, Hansman, 19841. ,The local collection The influence of various meteorological parameters efficiency isdefined as the fraction ofthe mass fluxlocally which axe considered important aircraft icing are discussed impinging onto the surface to the freestream mass flux. The below in terms of their effect on the physics and the severity of the ice accretion.
collection efficiency is typically highest near fl_e stagnation point of the body and decreases downstream. The point at which [3 becomes zero is defined as file inlpingement limit. 3.1 Ternlxn'aal_ The collection efficiency is a strong functioq of d,'oplet size and body geometry. Large droplets have high inertia tending Air temperature is one of the most important of the to cross streamlines resulting in high collection efficiencies icing parameters. Meteorologists normally work with the ambient or Outside Air Temperature (OAT) however pilots and impingement limits. Small droplets tend to follow the streamlines resulting in lower collection values. Small and aircraft designers also use Total Air Temperature (TAT) bodies are more efficient droplet collectors because there is to include aircraft velocity effects. The TAT is the temperature at the stagnation point of the aircraft and less room for the droplets to turn prior to impact.
corresponds ,o the OAT plus an additional temperature rise DROPLET -_ due to the deceleration of the incoming flow. This "ram rise" can be significant at high velocities. In Figure 4 it can be seen that the difference between stagnation and ambient
TRAJECTOR,ES ii
temperature (TAT-OAT) may be as high as 30* at 500 kts.
Because the surface temperature can be lower than the TAT aft of the stagnation point of the aircraft the normal STREAMLINES _-_-___tt procedure in jet aircraft is to operate anti-icing equipment at TAT values between +i0 ° C and -10 ° C in the presence of BODY visible moisture.
Fig. 2 Example of the relation of water droplet trajectories ,_, 4O to the streamline field from Bergrun (1947).
2.2 Thermodynamic Heat Bal,'mce 3O As described above, the them_odynamic heat balance on the accrefing ice surface is a controUing factor in determining the rate of ice accretion in mixed icing conditions and is the critical factor in determining the [...
transition between these conditions and rime ice growth. e-, -_- I0 Figure 3 shows the principal modes of energy transfer associated with an icing surface, as depicted by Messinger -..£ (1953)• Heat is added to the surface primarily from the latent heat of fusion released as the droplets freeze, but also "_ 0 .... b from aerodynamic heating and, to an even smaller extent, 0 1O0 200 300 400 500 600 from the kinetic energy of the droplets impacting the surface.
Velocity _knots) Heat is removed from the surface primarily by convection, and to a lesser degree by sublimation (when the surface is Fig. 4 Stagnation temperature rise versus velocity.
dry) or evaporation (when the surface is wet). In addition, An example of the effect of temperature on ice heat is absorbed from the surface as the supercooled droplets accretion and performance degradation can be seen in Figure impinge and warm to 0 ° C. The parameters which primarily 5. which shows ice shapes and section drag coefficients for influence the heat balance are the temperature difference a NACA 0012 airfoil at 4 ° angle of attack obtained by Olsen, between the surface and the free stream, the convective heat Shaw (1984) in the NASA Lewis Icing Research Tunnel. At transfer and the impinging liquid water mass• cold temperatures, the ice accretion was insensitive to temperature and rime accretions were observed with drag increases of 2 to 3 times the clean values. However as the temperature increased above -10 ° C the drag increased sharply with temperature to a peak value of over 8 times the clean drag. What happened was that the ice growth transitioned from a dry rime growth to a wet mixed growth.
The horns characteristic of mixed growth can be observed in the high drag ice accretions. As the Total Air Temperanu'e Latent Heat _:_:_ nears 0 ° C the ice accretions and the resulting performance degradations decrease due to insufficient heat transfer to freeze all of the incoming water.
From this example it is clear that a relatively small _'- Kinetic _'_.
temperature change can cause transition from a relatively benign rime icing condition to a dangerous mixed condition.
The nonlinear dependance of icing severity on Total Air Temperature (TAT) combined with aircraft velocity effects f-'vaporation _ make it difficult to accurately identify regions of moderate or severe ice potential from the ambient air temperature alone.
Fig. 3 Modes of energy transfer from an icing surface.
O AIRSPE[D. _0gk_r_r: LWC.. ]_lmJ: TIIVk_ 8rntn.
X 08 _. O4 1T.M_ERATUR[ -L_ O C -_ C -[ 8 0 C -[5 ° C -LZ ° C -tl ° C -50 C -_ C .|o C 0 a C O AIR$_[D, ]3$kWhr: LWC. L_gtm3; TI/V_. 6,2mm.
0.2 ==_ 0'0 TOTAL 2 z, 6 8 l0 TTMFTRATUR( -_,Q C .fro c -12 ° c -# C -# C _ C EXPOSURE TIME. t (MINUTES) RUN CONDITIONS : • I0 F (_O v. _tmlnr LWC- L, 3glm 3 T. 8ram V- 3311knVflr LWC • | Q_glm 3 T-_.2mJn Fig. 7 Plot of liquid water content versus exposure time showing typical fluctuations observed in natural icing conditions. Also shown are ultrasonically measured periods of wet, dry and transitional ice growth.
stagnation point ice accretion was monitored by ultrasonic techniques which couid determine weather the accretion was -30 -_ -10 0 TOTAL TEMPERATURE, °C dry, wet or transitional. It can be seen that the variation in (b) S Klion arag c_efficimt.
liquid water content caused the accretion to vary from dry growth characteristic of rime accretion to wet growth Fig. 5 Effect of total temperature on iceshape characteristic of mixed accretion within the same cloud.
and drag. MVD = 20pro; 0.053 m cloud NACA 0012 airfoil at 4° angle of attack. Taken 3.3 Droplet Size from Olsen, Shaw and Newton (1984).
3.2 Liquid Water Content The size of the ambient water droplets both in terms of the Median Volumetric Diameter (MVD) and the actual shape of the Droplet Size Distribution can be important to the Liquid Water Content (LWC) influences the severity of the icing in two primary ways. First, increasing LWC ice accretion process. As described in Section 2.1, large droplets are more efficiently caught by body. These effects irn.plies more potential water and larger accretions within a have been quantified for .tyt_icalcloud and rain size given tune. Thus, high LWC implies a greater urgency and distributions by Hansman (1984). Figure 8 Shows the therefore severity ofxhe icing encounter. The second effect impinging mass distribution function and the ambient size of high LWC is to cause the ice accretion to transition from distribudon function for a Khrigian-Mazan distribution with time to mixed icing due to the higher impinging water load 20 micron mean effective droplet diameter. It can be seen on the icing surface. This can be seen in the schematic plot that the bulk of the impinging mass results from the small of icing severity versus LWC in Figure 6. Rime ice growth number of large droplets in the tail of the distribution. The will occur even at relatively warm supercooled temperatures MVD and the shape of the distribution therefore determine for low liquid water contents. In this regime there is a linear the effective collection efficiency of the aircraft. It is not increase in the icing severity with LWC. At some value of LWC however, the growth will transition from rime to uncommon to have trace or negligible icing even at high liquid water contents for clouds of predominantly small mixed and the severity of the icing will increase. At colder temperatures the threshold for the transition will occur at a droplets. It is also possible to have moderate to severe icing at relatively low liquid water contents for distributions higher liquid water content.
consisting primarily of large droplets.
I0 Warm Temp.
Z 09 1 O8 --_ Mixed Cold Temp.
u_ 0.6 o5 .__ 0.4 N 0.3 w_0.2 • u • I • U " ! • I O Z 0.1 Liquid Water Content Fig. 6 Schematic plot of icing severity versus liquid water o 2o 4o 60 8o too content.
EQUIVOLUMETRIC DIAMETER (microns) Fig. 8 Khrigian.Mazan ambient doud droplet size An additional factor which complicates the the ice distribution and the resulting impinging mass flux accretion process is the large variability in LWC often within distribution.
the sam_ cloud. Figure 7 shows an example of LWC The presenceof large droplets within the cloud measured by Hansman and Kirby (1987) with a Johnson- .dism.'bution can result in additional hazard due to increased Williams hot-wire probe mounted in the nose of the NASA mapmgement limits. The limit of droplet impingement on an "Twin Otter" Icing Research Aircraft. During this flight, the aircraft component increases with droplet size. Current 4.1 Velocity design guidance (FAA AC-20-73, 197 I) recommends that a diameter of 40 microns be used to determine impingement The aircraft velocity effects both the collection of limits. The presence of significant numbers of droplets in liquid water and the thermodynamics of the icing process.
excess of 40 microns can therefore result in ice accretions Increasing velocity results in higher impinging liquid water occurring behind the protected regions of the aircraft. These exposure by increasing both the path swept out by the factors are thought to contribute to the anomalously high aircraft trajectory in a give time and the collection performance degradations observed by Cooper, Sand, efficiency.of the aircraft surfaces. Thermodynamically, the Politovich and Neal (1984) in clouds with droplets ranging velocity effects the heat load through the increased impinging from 40 to 300 microns.
water mass and through increasing the stagnation point temperature at high velocities. This is shown in Figure 4 The problems resulting from large droplets are where the stagnation temperature rise is plotted as a function exacerbated in freezing rain where both large droplet sizes of velocity.
and large liquid water contents are combined. Freezing rain 4.2 Shape commonly results in clear glaze ice accretions with significant runback icing. It is considered an extremely The shape of the accreting body has a large effect on hazardous condition.
the local collection efficiency. Generally, smaller bodies are 3.4 Cloud more efficient collectors than larger bodies. Therefore, slender components such as propellers, fan blades and The icing potential for a particular cloud is directly antennas will tend to be the most sensitive to ice accretion.
related to the phase of the hydrometeors. As described As a result of their high collection efficiencies, windshield above, icing normally results from the impact of supercooled wipers or Outside Air Temperature probes are often used by water droplets. Dry ice crystals, generally, do not adhere to crews to detect icing conditions in flight.
the aircraft surfaces after impact and are therefore not Three dimensional shape effects can also be considered an icing hazard. If however, the ice crystals are important for many aircraft components. For example wing wet due to partial melting or the aircraft surface is wet due to sweep commonly results in a spanwise variation of the de-icing or the recent penetration of a high LWC region then accretion in mixed icing conditions. This "lobster tail" ice the impinging crystals will stick.
can result in significant performance degradation. It should be noted that current forecasting procedures are based on Mixed phase icing (not to be confused with mixed straight wing propeller driven aircraft (USAF, 1980).
rime/glaze icing) is relatively rare. Some examples of mixed phase icing have however, been observed in flight tests by 4.3 AircraR Catego_ Gayet, Bain and Soulage (1984) who noted that the presence of snow in supercooled clouds significantly reduced the The effect of icing varies significantly with individual icing rate. Ice crystal sticking on wet aircraft surfaces was aircraft design. While it is beyond the scope of this paper to discuss the icing sensitivity of individual aircraft, the observed during NASA Icing Research Aircraft flight tests of ultrasonic ice detector arrays where the wet aircraft sensitivity of broad aircraft categories will be discussed briefly below.
surface was documented by ultrasonic techniques (Hansman, Kirby, McKnight and Humes; 1988).
4.3.1 Tur bojet/l'urlxffan Air,a-aI_ In general, forecasting efforts are directed towards Jet aircraft are considered to be the least susceptible identifying regions of supercooled cloud. Techniques are to icing. Jet aircraft normally operate with significant available to predict cloud phase in stratiform cloud and quantities of excess thrust which can be used to offset glaciation results in overestimation of the icing severity performance degradation. In addition, the typical flight (USAF, 1980). However, in cumuliform clouds, cloud profile of a jet aircraft is to climb and descend rapidly phase uncertainty represents a potential source of error in the through the lower troposphere where the icing potential is forecasting of icing conditions.
greatest and to cruise at high altitudes (20,000 ft to 45,000 3.5 F'me Scale Turbulence Level ft.). Occasionally, Air Traffic Control (ATC) requirements will dictate sustained operation at low altitudes particularly in An additional factor which has recently emerged as busy terminal areas. The primary icing hazard to jet _t potentially important to the ice accretion process is the fine is engine failure due to Foreign Object Damage (FOD). This scale (centimeter and below) ambient turbulence level. The results from the ingestion of chunks of ice which are shed turbulence level is known to strongly influence the off of other aircraft components such as engine inlets or in convective heat transfer from the icing surface. As shown in some aircrafts the wings. Therefore, critical regions are Section 2.2, the convective heat transfer is one of the normally anti-iced with hot bleed air from the engines to primary parameters in determining the transition between prevent any ice accumulation rime and mixed icing with the resulting effect on icing 4.3.2 Turlx_ropand RedproealingAirera_ severity. Flight test observations by Hansman and Kirby (1987) observed a wide variability in the parametric There is a tremendous variability in the sensitivity of threshold between rime and mixed icing. The variability was propeller driven aircraft to icing. Large turboprops tend to thought to be due to variations in heat transfer resulting from be fully ice protected and can operate successfully in regions the ambient turbulence level. While the effect of free scale of high icing potential. Small reciprocating engine aircraft turbulence level on icing severity has not been directly are generally not equipped with ice protection and are not demonstrated it may be an additional source of uncertainty in approved for flight in icing conditions. Even light icing the forecasting process.
conditions are a potential hazard to unprotected aircraft and 4.0 INFLUENCE OF AIRCRAFT PARAMETERS forecasting uncertainties have the greatest impact on this aircraft category.
The influence of various aircraft parameters which are considered important to the icing problem are discussed Propeller driven aircraft operate at low altitudes where icing potential may exist over the entire flight. The below in terms of their effect on the physics and severity of the ice accretion.
most critical components are generally the propellers because measurement and the variability of these factors with
loss ofpropeller efficiency translates direcdy into loss of
altitude, position and time coupled with variable aircraft thrust. However, even airframe icing can pose a significant sensitivity make forecasting and identifying hazardous icing hazard because propeller driven aircraft normally operate at conditions difficult. Automated Pilot Reports (PIREPS) are much lower excess thrust margins than jet aircraft. Other suggested as one mechanism for improving the data base hazardous factors include; reduced stability, loss of control necessary to forecast icing conditions.
authority and reduced visibility due to windshield icing.
ACKNOWLEDGMENTS 4.3.3 RotorcraR This work was supported by the National Helicopter icing has become important during the last Aeronautics and Space Administration, the Federal Aviation decade where helicopter flight in Instrument Meteorological Administration under the Joint University Program for Air Conditions (IMC) has become more commonplace.
Transportation, Grants NGL-22-009-640 and NAG-3-666, Helicopter operations occur almost exclusively at low and the National Science Foundation Presidential Young altitudes where there is significant icing potential.
Investigator Award Program.
Helicopters are extremely susceptible to icing conditions.
Rotor icing simultaneously degrades the lift and thrust REFERENCY_,S efficiency of the vehicle. In addition, helicopters typically operate with very slim power margins and can therefore only Bergrun, N.R., 1947: A Method for Numea'ieally Calculating the Area tolerate minimal ice accretion. Other hazardous factors and Distril_don of Watex Impingement on the Leading Edge of an Airfoil in a Cloud. NACA TN-1397.
resulting from helicopter operations in icing conditions include; blockage of engine inlets, reduced control authority, Bragg, M. B., Gregorek, G. M.. and Shaw, R. J., 1981 : Analytical vibration due to asymmetrical ice load on rotors and reduced Approach to Airfoil Icing. AIAA Paper 81-0403.
visibility due to canopy icing.
5.0 IMPLICATIONS FOR THE FORECASTING Brun, R. I., and Mergler, H. W., 1953 : Impingement of Water OF ICING CONDITIONS Droplets on a cylinder in an Incompressible Flow Field and Evaluation of Rotating Muldcylinder Method for Measurement Some of the difficulties in forecasting or identifying of Droplet-Size Distribution, Volume-Median Droplet Size, and Liquid Water-Content in Clouds. NACA TN2904.
hazardous icing conditions are clear from the above. Many of the parameters which were shown to be important to the Cooper, W. A., Sand, W. R., Politovitch, M. K., Veal, D. L., 1984 : ice accretion process such as the droplet size distribution and Effects of Icing on Performance of a Research Airplane.
cloud phase are not available to the forecaster. Other Journal of Aircraft, 21, pp. 708-715.
parameters, which may be available, such as temperature or liquid water content:are nonlinearly related to icing severity.
FAA, 1971: Advisory Circular. AC-20-73 Limitations in accuracy and spatial resolution of these parameters result in over or under prediction of the icing Gayet, J. F., Bain. M., and Soulage, R. G., 1984: Rote of Ice Crystals severity. Finally the variable susceptibility of different on Ice Accretion Processes. Proceedings of the Second lnt'l aircraft types implies that a single icing hazard analysis will Workshop on Atmospheric Icing of Structures, Norway.
result in over or under predictioh of the severity for certain Gelder, T. F., Smyers, W. H., Jr., and yon Clahn, U., 1956 : aircraft categories.
Experimental Droplet Impingement on Several Two- Given the limitations on the forecasting process, Dimensional Airfoils with Thickness Ratios of 6 to 16 current techniques do a remarkably good job at identifying Percent. NACA TN3839.
general regions of potential icing conditions. One of the key indicators which are used to validate or initiate an icing Hansman. R. J., Jr., 1984 : The Effect of the Atmospheric Droplet forecast axe pilot reports (PIREPS). By actually penetrating Size Distribution on Aircraft Ice Accretion. Journal of AircragCt, 22, pp. 503-508.
the icing environment, aircraft can directly measure the severity of the icing condition. One of the difficulties with Hansman, R. J., Jr., and Kirby, M. S.,1987: Comparison of Wet and PIREPS is that the reports are not well calibrated and the Dry Growth in Artificial and Flight Icing Conditions. Journal variable susceptibility of aircraft discussed in Section 4.3 of Thermophysics and Heat Transfer, 1, pp. 215-221..
must be considered in their interpretation. Another problem is the timely dissemination and generation of PIREPS.
' I-tansman, R. J.. Jr., Kirby, M. S., McKnight R. C., and Humes, R.
Because of other responsibilities, ATC cannot always L., 1988 : In-Flight Measurement of Airfoil Icing Using an process PIREPS rapidly which may discourage the Array of Ultrasonic Transducers. Journal of Aircraft, 25, pp.
voluntary pilot reports. 531-537.
One potential improvement, which has been Messinger, B. L., 1953 : Equilibrium Temperature of an Unheated implemented to a limited extent, is the automated generation Icing Surface as a Function of Airspeed. Journal of the Aeronautical Sciences, pp. 24-42.
and transmission of PIREPS over digital data links. By the use of onboard ice accretion sensors, automated continuous Olsen, W., Shaw, R., and Newton, J., 1984 : Ice Shapes and the reporting of the presence or lack of icing conditions could be Resulting Drag Increase for a NACA 0012 Airfoil. NASA accomplished. If an adequate fleet of aircraft were so TM-83556.
equipped, a significant improvement in the forecasting and identification of hazardous icing conditions could be Renaudo, R.J., Mikkelson, K. L., MeKnight, R.C., amd Perkins, P. J., realized. Other potential technical developments which may Jr., 1984 : Performan_ Degradation of a Typical Twin improve icing forecasting include: vertical sounders, Engine Commuter Type Aircraft in Measured Natural Icing improved satellite imaging, and improved data synthesis Conditions. NASA TM-83564.
systems such as the PROFS system.
6.0 CONCLUSIONS USAF Air Weather Service, 1980:. Forecaster's Guide to Aircraft Icing.
Aws/rR-80/001 The severity of aircraft icing is found to be extremely sensitive to temperature, liquid water content and droplet size distribution particularly near regions of transition between rime and mixed icing conditions. The difficulty in