ENGINEERING RESEARCH INSTITUTE UNIVERSITY OF MICHIGAN: ANN- ARBOR Final Report EVALUATION OF AN. ICING-INDICATOR PROBE I..S.TUBB Project 2217 THE DETRIT C ONTR OLS CORP ORAT ION REDWOOD CITY, CALIFOR~NIA April, L9'55

s 4 cA`J. E IC% 3 -%

Page LIST OF ILLUSTRATIONS iii ABSTRACT: iv INTRODUCTION 1 FINDEINSENN'S DETECTI0N:1OD. 2 INDICATOR PROPOSED BY TRIBUS AND MOY;LE 3 TEOREICAL PERFO RMANCE INDICAO 8 LABORATOIRY PERF ANCE TESTS 11 GE NERAL 0BSRVATONS 16 CONCLUSIONS- 19 BIBGLIORA 20

LIST OF ILLUSTRATIONS Fig. 1. Sketch of probe Fig. 2. Functions -appearing in e-quation 5 Fig. 3. Icing indicator probe Fig. 4. Electrical circuit Fig. 5. Output of indicator probe in icing conditions ii i

ABSTRACT The; performance of an icing indicator proposed by Tribus and Moyle was evaluated theoretically and experim'entally. The indicator consists of two conducting elements in a nonconducting aerofoil, with the forward element exposed to the impingement of airborne particles and the rear element sheltered. The analysis shows that a probe can be designed to be insensitive to airstream velocity and changes in the ambient temperature. Ice clouds and water clouds above the freezing point will not proiduce a- response which can be misinterpreted as icing conditions. The instrument will give an indication of the icing rate with a: derate ie cap but the performane will deteriorat eIn genal th performance tests confirmed the theoretical analysis. The minimum icing rate which could be detected and the sensitivity of the- device to design details were determined. iv

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN EVALUATION OF AN ICING INDICATOR PROBE INTRODUCTION Continuous indication of the existence and intensity of icing co ditions about aircraft is desirable from an operational and safety point of view. The pilot may take appropriate action upon such indication, or the indicator may actuatCe the ice-protection equipment of the aircraft. The work reported here was undertaken to evaluate an icing indicator conceived by Myron Tribus and Morton Moyle while working on Air Force research contract AF18(600)-51. The indicator is essentially an. improvement on one developed by W. Findeisen in Germany. IL

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN FINDEISEN'S DETECTION METHOD Findeisen suggest:ed an icing detector based on the energy released at a surface by the formation of ice. In general terms, Findeisen's device has two surfaces exposed to the air stream in which the plane is flying. One of these sur:fac.es is placed so that airbore particles impinge on it the other is sheltered from such particles. In nanicing conditions both surfaces assume the temperature of the, airstre:am. In icing conditions the temperature of the expo-sed surfae is raised due to ice fomation, while the sheltered surface is unaffcted. The temperature rise of the exposed surface above that of the sheltered surfaee is thus an indication of icing. The temperature difference Can be measured by any onvenlent method. This scheme has the inportant advantage of distinguishing icing conditions from nonicing water and from ice- clouds. Water above the freezingpoint and solid ice particles either have no; effect or, cool the exposed surface. Certain conditions other than ieing can raise the tempefature of the exposed surface of -Findeisens device above that of the sheltered one. One of these conditions occurs wen: the aircraft flies through regions of variable air temperatuure; if on'e of t the t surfaes rth anbient temperature more rapidly than the other there will be a transient period durig gwhich the' two surfaces are at different temperatures and the instrument may give a spurious indication of icing when there is none. Only if the transient thermal respnse of the two elements is the same will spurious icing signals be avoided for all changes -of ambient temperatur. Frictional heating of a surface in a high-velocity airstrea can also give a apurious icing signal. Th-e fr-ictional heating is different at differaent parts of a body,. and in Findeisen:'s icing indicator the- exposed:element was

ENGINEERING R.ESEARCH INSTITUTE: UNIVERSITY OF MICHIGAN heated. more than the shelt.ered' one, so that spurious readings were produced by high-spe"ed flight. Findeisen made an attempt to eliminate: the spurious signal's by painting part of his instrument with: a lacquerj but control of th- thickness aof the lacquer coating is difficult, and excessive experimentation would be required tO determine if the spurious signals will be eliminated for all flight velocities, INDICATOR PR:OPOED BY TRIBUS AND MOYLE In' 1952, Tribus anid Moyl proposed'an ice-warning probe which also is based o the energy relase- on a surface where ice is forming, but which is designed to'elimi.nate- spurious icing signals from frictional heating or variable. air-stream temprat'ure. The robe" has two: cnducting elem'ents inserted in a nonc nducting aerOfoil as shown in Fig. 1. The front eelement is exposed to the impingenent of airborne- particles but the rear element is sheltereed beaus: e the aerofoil be:comes narrOer toward the tail. As with Findeisen"s instrument, the temperature of the fr.n elemet bcobms highe than thet rea in icing conditions:., and this tePerture difference is detected by some convenient m-eans. The equations governing convective- heat transfer can be used to explain ho the probe avoids spurious signals f rom frictial heating aid from changing ambient temeratures. The rate of:convetive heat transfer betwen a body a.nd an airstream b-ecmes zero, not.'when the- bdy is at thte samei temperature- as the air stre'am, but when the body is at a higher temperature called the adiabatic wall teerature-. The differTene etw- en the air temerature and the adia- - batic wall tereturea is not significant for low ve2Loities, but may be 3

ENGINEERING: RESEARCH INSTITUTE' UNIVERSITY OF MICHIGAN appreci:able at the velocities attained by -modern aircraft. Data for the local rate of he'at exchan-ge expr-ssed as -energy transferred per unit time per unit area are generally correlated in terms of a heat transfer coeffici-ent, i.e., heat transfer rate = a - ht-a - ts) (1)'where; h ordinary hbeat transfer c'efficient ta adiabatic wall temtper re ts= tpemp`r atur.~e':of sur fa'ce dA = differential eleent Of area. The rat-e on the right must be' integrated over t~he whole surface to give the total rate of'heat transfer The' rate- of change in the themperature of the. body wil in turn be roportional to the rate of heat transfer: dt.'C =.J h(ta ts)dA (2) where: C = thermal caacity of body (dimensions: energy/o) time. The-e adiabatic wall temperature can be express;ed as ta t0 + r(tt - tO) (3) where: to = temperature of the air stream tt =: the total temerature and r = the` ~recovery factor. Introducing Equation 3 into Euation 2 gives (to sts) + (tt A to) dC s

ENGINEERING RESEARCH INSTITUTE *~ UNIVERSITY OF MICHIGAN where the symbols with bars are average values defined by (5) c/ h dA Ai_ A * (6) Equation 4 can be integratedd to give (to- ts) + r (tt - ts) = K'ex(- ). (7) K is a constant equal to the value- of the left side of the equation for 0 equal to zero. Solving xliicitly for the, surface( temperature gives ts = t + r(tt to:) - K ex --. (8) For two, bodies i n the same airstr ( such as the frnt nd rear el ts of the Tribus-Moyle- iing indicator), to'an tt will be th- same. It is then evident from Equation 8 that r must be the same -Tor':the two bodi-es if they are thave the s.me- surface temperatures at eulibrium. If in addition the two bodies are to have the s- tempeature dur-ing the transient period when the' e p:n.tia-l term is Sgnifint the:-fficient ~f 0 in the ex-onential t'must be th'e.sae for b.th bodies. The proper design of -an iin indicat.or in. or der to'aid surilous sigls f rom fricti~o:l heat ing r abint teratre changes t uir tht t th the va.lue's o_ r be the s:e for the frt and rer s, and.al —' that (:/C) be the same for the front and rar elem. ts....~~~~~~

ENGINEERING RESEARCH INSTITUTE' ~ UNIVERSITY OF MICHIGAN In the Tribus-Moyle indicator, the rear element is placed in the region- of turbulent flow where the recovery factor is equal to the third root of the Prandtl modulus.2 The recovery factor may then be taken as 0.9 for the rear element. The front element is placed in the cylindrical leading edge -of the indicator probe. In the laminar flow which occurs near the stagnation point of a cylinder, both the heat transfer coefficient h and the recovery factor r are functions of angular deviation from the stagnation point. In particular, the recovery factor: is higher than 0.9 at the stagnation point and lower' than this value farther back, and it is possi'ble to choose the size of the front element so that it will have an average recovery factor r equal to that of the rear' element. The process is as follows: The local heat transfer co-efficient for a.cylinder can be evaluated from an expression such as that of Schmidt and Wenner:3 hD = 1.14 Pr.5 [1 3(/3 (9) 1.,1.54 where: D = diameter of cylinder k = thermal conductivity of gas Pr. = Prandtl modulus = 0.73 for air v = velocity of gas stream V = kinematic viscosity =:angle from stagnation point (in radians) Local values of the recovery factor given by Eckert and Weise- are reproduced in Table I. - - - ~~~6 —

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN T-ABLE I (degr ees) 0 1.00 15 -98 30.90 45.79 65.70 When- Equatio-n 9 is ittrod -c into Equation 5, the expressio be comes'L.5r [ 3-] (.. ) The integrands thatap in Equation 10 are plottd in Fig. 2. It is nec-essary to choose an upper liit for the inte-gration so that r will be equal to 0.9. The corret value for the upper limit is about 55 i.e.:, element if it extends t; a psitio 55 back from the stagaatio porint ozf the cylinder. The transient response Of the front and rear elements can then bie made equal by adjusti:ng the hetat ccity of the two lemlets so that (i/c)fnt = (/C)rear (11) The ratio of the thermal cap'acities. of the two elements can be evaluated with the- aid of Equation 9 which is applicable t t the frnt elent and an:equation given by Jacob and Do0 which is applicab le to the rear element: - -. -.'

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN =L 0Q280() I + oA.40 j ) (12) where: L' = distane fromn stagnation point to beginning of elem~ent L = distance from stagnation point to end of -elment. The ratio of the heat cap-aities then bec-onelS 2.75 0Crear D 3 1+ (13) Cfront LI \D,/\j / where: 1 = length of front element along axis of cylinder. An examination of Equation 13 shows that the tim constnts of the two elemnts cannot be equalized, irrespetive of the Reynolds modulus - ever, the Reyn.lds modulus apear only to the 0.3 power in the expression so that t heratio of thermal p.aciti:es is not very sensitive to the velo:city In, fct, when the velocity changes by a factor of 22, the correct ratio for the thermal capacities will change only 23 pr cent. By designing for the center of the velocity range f the aircTaft, the ti:e Zonstant s can be pretty well matched for;all velocities. It does.appear however, that it would be desirable to- use different thermal capacities for aircraft whose flying speeds differ considerably ~fam each other. THEORETICAL PERFOBMANCE OF INDICATOR It will now be worthwhilet toh prfr ih the rfoane ih theoretical csiderations inicate for an icing detector as suggste by Tribus and Moyle. In clear air the two elents will have the same equilib8

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN rium temperature no matter what the ambient temperature of flight velocity. If the aircraft encounters an air mass with a different temperature than that in which it has been flying, the two elements will continue to- have equal t-emperatures if the aircraft is flying at the "design speed" (i.e., the speed used in Equation 13). If the speed is less than the "design speed", the front element will respond to the new ambient temperature more rapidly than the rear element; if the speed is greater than the "design speed", the front element will approach the new ambient less rapidly than the rear element. A spurious signal similar: to that induced by icing conditions will occur when entering oblder air at speeds higher than the design speed, and when entering warme air at speeds lower than the design speed. In any case the- temperature difference will be small unless the flight speed differs greatly from the design speed. For example if an aircraft suddenly enters an air mass with a different temperature while fly ing at a speed 10 per cent above the design speed, the maximum temperature difference between the elements will be about 1 per cent of the difference in temperature between the two air masses. When the detector flies through water clouds above'.the freezing point, water will collect on the front element where: some of it will- evaporate, cooling the element. Water may r may not run back to the rear element depending on how fast drops collect on the front surface of the indicatQor If water does run to the rear element, it will also be cooled and the two elements will have the same temperature in any case-, the front element will not be:.warmer than the rear. In cold ice clouds the ice particles will bounce off the front elnement arnd produce no effeet.different from that which wwould.occur in clear airs If the air cloud is only a little belowfreezing, frictional 9

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN heating may raise the temperature of the elements above freezing so that the ice particles melt when they hit the front element and thus cool it. Again no situation arises in which the front element becomes warmer. When the instrument encounters icing conditions, the ice forming on the front-element liberates heat. The temperature of the front element, therefore, rises until a steady-state condition is produced when the heat liberated by the heat of fusion of the ice is carried away from the element by convection and by sublimation. The steady-state condition has been analyzed by Messinger6, and the temperatures of the front and rear elements can be calculated: by the methods which he presents. However, as a rough approximation, the heat supplied is proportional to the rate of ice formation, and the heat carried away is proportional to h(ts - ta). Since the rear element assumes the adiabatic wall temperature, the temperature difference between the elements will be an appr-oximate measure of the icing rate, provided h remains constant. However, h depends on the velocity of the airplane and no universal calibration of the instrument can be achieved. In any one icing encounter the readings of the instrument should be an indication of the relative intensity of icing from instant to instant. A qualitative picture may be obtained of the performance of the indicator as an ice accretion builds up on the front element. Heat exchange between the front element and the exposed surface will have to take place through a layer of ice and the temperature of the element will therefore respond less quickly to changing conditions at the surface. Also the ice cap will add its heat capacity to that of the' front element and will alter the effective angular extent of the front element, so that the balance of recovery factors and time constants between the two elements will gradually 10

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN be distorted. However, the basic detection principle will continue to operate with the front surface of the ice layer rising in temperature by an amount that is about proportional to the icing rate. In short, the instrument continues to operate, but in a way equivalent to a poorly designed instrument. LABORATORY PERFORMANCE TESTS In order to check the foregoing theoretical analysis, to examine questions which theory cannot answer, and to obtain practical experience in the operation of the icing indicator, an experimental program was undertaken. The program can b-e considered in two- parts. The first part of the experimental program was conducted at low velocities (about 100 ft/sec) with the indicator exposed to various conditions of rain and icing clouds. The second part was conducted at higher velocities (up to about 800 ft/sec) and was designed primarily to check the calculations for the: angle of the front element and for the ratio of the thermal capacities of the two elements. Several indicator probes were constructed of aluminum and plastic as shown in Fig. 3. Thermistors were inserted in the front and rear elements and connected in a bridge circuit as shown in Fig. 4. This circuit was suggested by Findeisen and is a rather obvious way to make the instrument sensitive only to- differences in temperature between the two elements. A vacuum-tube voltmeter was usually used as the detector. The first part of the.experimental program.was conducted in a small wind tunnel situated in a refrigerated room. The tunnel had a fixed Some were supplied byWestern Electric Company and Others by Victory Engineering Corporation. 11

ENGINEERING -RESEARCH INSTITUTE * UNIVERSITY OF MICHIGAN air velocity of about 100 ft/sec and was driven by an exhaust fan down.stream from the test section. Water could be sprayed into the tunnel entrance to simulate cloud conditions. A DeVilbis.air-atomizing nozzle was used to produce a spray similar to that found in natural clouds. The equipment was not suitable, however, for measurement of the water content of the air stream. One series of tests in this tunnel showed that the front element became cooler than the rear when the instrument was exposed to water clouds at temperatures above freezing. This result confirms the expectation that water clouds above the freezing point will not produce an indication that c an be confused. with icing. Another series of tests was run with the.air temperature below freezing. The results are shown in Fig. 5 where the.output of the detector is plotted. against the ambient temperature. With the ambient temperature only slightly below freezing and a high water content~, the rate of ice formation is controlled by the maximum rate at which heat can be transferred from the surface. The icing rate, therefore, can be expected to increase linearly as the ambient temperature falls below the ice point. The theoretical maximum output will occur if the rear element remains at the ambient temperature and the front rises to 320F. The maximum output can be calculated from the constants of the bridge circuit and the temperature coefficient of the thermistors. It can be seen that in most cases the indicator realizes the theoretical value. The one point above the curve is probably An.erroneous reading. A few points lie significantly below the theoretical curve. It is not certain what caused these deviations, but water may have impinged so fast on the front surface that some of it ran back as far as the rear element before it completely froze. This would then heat the rear element and reduce the temperature difference between the elements. 12

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN A number of the points plotted in Fig. 5 were obtained with an ice cap on the instrument. The ice did not affect the magnitude of the detector readings but noticeably influenced the response time of the instrument. With a clean probe the detector voltage rose to its full value in less than a second; with about 0.1 inch of ice on the probe about 4 or 5 seconds were required to obtain the full value. It appeared that the instrument would work tolerably well with up to 0.1 inch of ice cap. The second part of the experimental program was conducted in one of the University of Michigan's supersonic wind tunnels. A sonic throat was produced downstream of the test section by partially closing a valve so that subsonic velocities were produced in the test section. The velocity in the test section could be controlled by the valve, but since the total temperature of the airstream could not be controlled,, a particular ambient temperature was associated with each velocity and could not be controlled independently from the velocity. The coupling of ambient temperature and velocity was not a serious shortcoming since the probe was designed to have no response to either velocity or temperature changes, and this series of tests was intended to see if the design has been successful. The general procedure in these tests was to mount the indicator probe in the test section,. start the tunnel, and then change the velocity and temperature of the airstream. Throughout the test the output of the indicator was observed or recorded. In a final series of runs, the probe was connected to equipment constructed by The Detroit Controls Corporation which permitted the output of theprobe to be expressed directly in terms of temperature. The results of these runs are sunmmarized in Table II. 13

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN TABLE II TESTS IN HIGH-SPEED TUNNEL Run Tunnel Conditions Output at Equilibrium Comments 13 Initially 44~F, 130 knots Probe X Changed to 130F, 379 knots +3.0~F Changed to 44~F: 130 knots -1.5 14 Initially 440F, 130 knots —--- Probe X Changed to 17~F, 362 knots +2.0 Changed to 420F, 159 knots ~0.3 15 Initially 440~F 130 knots Probe Y Changed to 120~F 385 knots 0.0 Front element went 2.0~F cold during transient. Changed to 43~F, 145 knots +1.0 Front went l.01:F warm during transient. 16 Initially 430F, 144 knots Probe Y Changed to 10F, 506 knots +0.3 Front went 2.0~F cold during transient. Changed to 44~F, 112 knots +0.3 Front went 2.00F warm during transient. -,, * -;:' - ~ 1 ~ - The output is the change in the temperature of the front element relative to the.rear which occurred with the change in.tunnel conditions. For example, in Run 13, when the temperature was dropped.the quantity (tfrofnt-trear) increased 30F. When the temperature was raised again the quantity decreased 1.50~F Probe X, used in the first two runs was one of standard design. Probe Y was built with a smaller angle on the front element (450) and with a smaller mass in the front element (0.7) than the standard. The test of Probe Y served to indicate how important small deviations from the design would be. It is concluded from these tests that spurious responses from- tte instrument have a magnitude corrsponding to about + 1 F. The spurious resp6nses set a limit on the sensitivity of the instrument, and it, therefore, appears that the probe will be effective in indicating icing conditions l.t

ENGINEERING RESEARCH INSTITUTE * UNIVERSITY OF MICHIGAN which produce about a 20F temperature rise on the front element. The corresponding icing rate is a function of air speed, but as an ekample) at 300 mph an icing rate of 0.015 in/min would produce a 2~F temperature rise in.the front element. This then is the minimum icing rate that the instrument could reliably detect at 300 mps. A second conclusion is that the angle of the front element is not very critical. The difference in response between the two probes is about the saame as the differente in response between the runs with the same probe, and the probe with the 45 front section is perhaps the better. On the other hand, it.appears essential to match the time constants of the two elements fairly closely. In the unmatched probe, Y, the transient output amounted to about 20F when the wall temperature changed about 4~F. Since temperature changes considerably greater than 4~F will be encountered in flight, mismatched elements would produce spurious signals that would be confused.with signals of icing. It further appears that Equation 13 is reliable for computing the balance.of the time constants of the two elements. Another observation made in the course of the tests deserves mention. There was always a-.large transient swing in the output of the instrument after the air in the tunnel was turned.on:. This was because the two elements were not initially at the same temperature. The convective heat transfer from the elements in still air is very small and especially so at small temperature differences. Consequently, if one element is heated by the touch of a hand-or by radiation it does not immediately return to the air temperature. The instrument should not, therefore, be balanced. or adjusted in.still air. 15

ENGINEERING RESEARCH INSTITUTE I~ UNIVERSITY OF MICHIGAN GENERAL OBSERVATIONS The following remarks, although not directly pertaining to the foregoing tests, may be of interest2 and assistance in the further development of the icing indicator. As pointed out by Findeisen, the instrument should be located on the aeroplane in such a position that it is not affected by the thermal effects of the engines. It should also be located far enough from the wing and fuselage to be out of the boundry layer and out of the region where the drop content of the air is modified by the presence of the aeroplane. In general, these considerations argue in favor of locating the instrument forward in the plane and having the probe protrude several inches beyond the plane- surface. Findeisen also points out that solar radiation can heat the front element and thus give a spurious indication of icing. The spurious nature of such a signal will always be obvious to a pilot since it will o:ccur only in bright sunlight when it is evident that there is no danger of icing. However, the confidence of the pilot in the instrument might be undermined if the device gives erroneous warnings. Use of the instrument with completely automatic ice-protection equipment would be difficult if spurious responses are encountered. In any case, the signal can be avoided by placing a radiation shield around the probe. The probes used for the tests described.above were constructed by machining the metal and plastic parts, inserting the thermistors and associated. wiring in the elements7 and then securing the parts together with screws. Although this method of construction was satisfactory for production of a small number o'f probes, it is likely that better methods can be found for production of larger numbers. One suggestion is that the metal 16

ENGINEERING RESEARCH INSTITUTE * UNIVERSITY OF MICHIGAN elements, thermistors, and wiring be fabricated first in a unit and then the plastic part be cast around the unit by some suitable process, perhaps injection molding. A probe constructed in this way would not offer any access to the thermistors after it is once finished, but this does not seem to be a serious objection. The probes used in the tests were about 4 inches long. This size was chosen arbitrarily and thee i's no reason to believe that it is critical. As a rule smaller probes react more quickly to changes in the icing conditions but are harder to work with and are more fragile. Another consideration is that the rate at which a body accumulates ice is a function of its size. This is because the very small drops of an icing cloud do not impinge on large bodies but do on small ones. The indicator probe will indicate the icing rate appropriate.to its own size, and it appears desirable that its front diameter should be about the same size as the smaller parts of the aircraft such as antenna masts so that a direct indication is given of the icing rate on the most susceptible parts of the plane. If an aircraft carrying the indicator probe flies for a considerable time in icing conditions, the instrument will become unusable because of the ice accumulation, and it will be necessary to remove the ice to restore the instrument to effective operation. It is difficult to conceive of any.method.for removing the ice accretion.that does not involve considerable equipment, since any ice-removal system must be operated intermittently and, therefore, will require switching equipment. Furthermore, while the ice is being removed the signals from the instrument will not be indicative of the icing conditions, and will probably be t-eprarily disabled. These consid.rations do not preclude the possibility that a simple method of ice removal may be devised, but at present the prospects appear pooor. 17

ENGINEERING RESEARCH INSTITUTE * UNIVERSITY OF MICHIGAN In contrast, the instrument without a deicing system requires very little equipment. If temperature-sensitive resistors are put in the elements and connected in a bridge circuit as was done in the experimental probes, the power for the bridge can be taken directly from the 24-volt D.C. supply of the plane, and the output of the bridge can be used directly to actuate a small meter that.will give a rough indication of the icing severity. In addition or alternatively, the output of the bridge can be used in conjunction with. a small rectifier to operate a light that will go:n when icing conditions are- en-cxntered. The ch-ief virtue of an instrume:nt such as that proposed by Tribus and Moyle, in comarison to other icing indicators, is its etreme simplicity. The most suitable.use for the instrument appears to be on aircraft that do not expect to remain in icing conditions for extended periods. In such an application, it offers a very simple and inexpensive device that will warn the pilot when the plant encounters icing, and then give him- an indication on a.meter whether the maneuvers which he undertakes to es-apethe icing conditio ns a taking the plane into more or less se.vere icing.conditions. When the plane emrges from the icing.cloud, the ice on the probe will either melt or sublime. Finally, it should be notd that the rear element of the probe has possibilities as a good thermometer for measuring the temperature of the air stream, since it is sheltered..from impinging articles and has a recovery factor that is independent of vlocity,. An. additional thermistor could be located in the rear element d its resistance used as a lmeasure o the temper.ature. A c'rrtin for the.sped of the p'laneJ would be.necessary, but the correction ould be simpler than for thfrermometric elements that do not have Colnstant re'overy factors. 18

ENGINEERING RESEARCH INSTITUTE. UNIVERSITY OF MICHIGAN CONCLUSIONS Theoretical and experimental studies of the thermometric icingindicator probe suggested by Tribus and Moyle indicate that the instrument can be designed to give a warning of icing and an approximate indication of the severity of the icing. The instrument %ill be unaffected by frictional heating at high speeds and only slightly affected by sudden thermal changes in the environment. The minimum icing rate to which the instrument will respond is somewhat a function of velocity, but the order of magnitude is about0.015 inch per minute. The design criteria, which are given in the body of this report, are not critical and the instrument will operate satisfactority with reasonable' manufacturing tolerances. One of the chief advantages of this instrument is its extreme simplicity. However, the simplicity can be achieved only if it is not necessary to introduce equipment to remove ice from the probe of the instrument. The probe,therefore, appears most suitable for use in aircraft that do not expect to remain in icing conditions for extended periods. 19

BIBLIOGRAPHY 1. Findeisen, W., "Das thermometrische Vereisungswarngergt", Deutsche Luftfahrtforschung, Untersuchungen und Mitteilungen, No. 691. Trans. by B. A. Uhlendorf and Myron Tribus, The Thermometric Ice Warning Indicator, University of Michigan, Engineering Research Project M992-B, August, 1952. 2. Ackerman, G., "Plattenthermometer in Stroemung mit grosser Geschwindigkeit und turbulenter Grenzschicht", Forschung Ing. Wes., 13, 226 (1942). 3. Schmidt, E. and Wenner, K., Heat Transfer Over A Heated Cylinder in Transverse Flow, NACA TM 1050 (19 4. Eckert, E. and Weise, W., Forschung Ing. Wes., 13, 246 (1942). 5. Reported in Jakob, M., Heat Transfer, Vol. I., 555, Wiley, New York (1949). 6. Messinger, B. L., "Equilibrium Temperatures of an Unheated Icing Surface as a Function of Air Speed"t Journal of the Aeronautical Sciences, January, 1953; also Lecture 6, Energy Exchanges During Icing, University of Michigan icing-information course. 20

~ ruoT-m a-otz;.s daduo;'a I;Ti (seaaJap) #'lNIOd NOIltVNgVLS 0UhVMOJ WOi0 31E)9N 09 t O 0'zo ~'0 8'0 16 69'0 -~ *~ ~ ~ ~ ~ ~ ~~m o0 -1LN3~.I313 t38- iN3.,313 INO-,

4.2" 2.740 l' _, 4.115 ~' \o'-' -.;0 2 415 1~)1-26'.328 O. 2 ". RAD - 4 CYLINDER:1 I SCREWS.250. 292 XI.,~i! t::-: -.I -.....:,,::.... rl.25o.292

R _ _i i__- ~THERMISTORS ARE -R3 / \ R4 - VICTORY ENGINEERING CORP# 51AII R, AND R2ARE APPROX. loo,oooQ /-\Detector) 22.5 V R3 AND R4 ARE APPROX. I7 222 Fig. 4. Electrical circuit. 2.4 6 points 2.2_. 2.0 1.8 \...'....... iv, %,- ~Theoretical Maximum ~- 1.6 1 1 0 1.4 5 0 rr 1.2 0. 00.8_ 0.6 I0.....4'.... w 00.2~~~~~~~0 8 10 12 14 16 18 2022 24262830 3234 AlR TEMPERATURE,0F Fig. 5. Output of indicator probe in icing conditions.

UNIVERSITY OF MIOH~IAN II 9015 02827 4879