Engineering Experiment

8/12/2019 Engineering Experiment

1/58

LL NOUNIVERSITY OF ILLINOIS AT URBANA CHAMPAIGN

PRODU TION NOTEUniversity of Illinois at

Urbana-Champaign LibraryLarge-scale Digitization Project 2 7


2/58


3/58

UNIVERSITY OF ILLINOIS ULLETINISSUED WEEKLY

Vol. XXXI October 31 1933 No 9[Entered as second-class matter December 11 1912 at the post office at Urbana Illinois underthe Act of August 24 1912. Acceptance for mailing at the special rate of postage providedfor in section 1103 Act of October 3 1917 authorized July 31 1918.1

FLAME TEMPERATURES IN AN INTERNALCOMBUSTION ENGINE MEASURED BY

SPECTRAL LINE REVERSAL

BY

ALBERT E HERSHEYAND

ROBERT F. PATON

BULLETIN No. 262ENGINEERING EXPERIMENT STATION

PUBLIBHED BY THE UN VS Ty OF ILLIN018OI URBANA

PaICe: FIFTY-FIVE CENTS


4/58

THE Engineering Experiment Station was established by actof the Board of Trustees of the University of Illinois on De-cember 8 1903 t is the purpose of the Station to conduct

investigations and make studies of importance to the engineeringmanufacturing railway mining and other industrial interests of theState.

The management of the Engineering Experiment Station is vestedin an Executive Staff composed of the Director and his Assistant theHeads of the several Departments in the College of Engineering andthe Professor of Industrial Chemistry. This Staff is responsible forthe establishment of general policies governing the work of the Stationincluding the approval of material for publication. All members ofthe teaching staff of the College are encouraged to engage in scientificresearch either directly or in co6peration with the Research Corpscomposed of full-time research assistants research graduate assistantsand special investigators.

To render the results of its scientific investigations available tothe public the Engineering Experiment Station publishes and dis-tributes a series of bulletins. Occasionally it publishes circulars oftimely interest presenting information of importance compiled fromvarious sources which may not readily be accessible to the clienteleof the Station and reprints of articles appearing in the technical presswritten by members of the staff.The volume and number at the top of the front cover page aremerely arbitrary numbers and refer to the general publications of theUniversity. ither above the title or below the seal is given the num-ber of the Engineering Experiment Station bulletin circular or reprintwhich should be used in referring to these publications.

For copies of publications or for other information addressTHE NGIN RING EXPERIMENT STATION

UNIVERSITY OF ILLINOISURBANA ILLINOIS


5/58

UNIVERSITY OF ILLINOISENGINEERING EXPERIMENT STATION

BULLETIN No 262 OcTOBER 1933

FLAME TEMPERATURES IN AN INTERNALCOMBUSTION ENGINE MEASURED BY

SPECTRAL LINE REVERSAL

Y

ALBERT E HERSHEYRESEARCH ASSOCIATE IN MECHANICAL ENGINEERING

AND

ROBERT F PATONASSOCIATE PROFESSOR OF PHYSICS

ENGINEERING EXPERIMENT STATIONPUBLISHED Y THE UNIVERSITY OF ILLINOIS URB N


6/58

UNIVERSITY12333000 12 33 4897 ,,Pi's S ,


7/58

CONTENTSPAGEI. INTRODUCTION 5

1 Object of Investigation 52 Acknowledgments 6

II COMBUSTION IN AN ENGINE 63 Thermodynamic Variables p V T ci c2 cn 64 Temperature Measurement by Line Reversal 8

III. DESCRIPTION OF APPARATUS AND METHODS OF CALIBRAT ON 115 Engine 116 Engine Control and Metering Apparatus 137 Temperature Measuring Apparatus. 168 Pressure Measuring Apparatus 22

IV EXPERIMENTAL PROCEDURE 239 Method of Conducting Tests. 2310 Lamp Calibration 24

V. TEST RESULTS 2511 General Statement. 2512 Effect of Spark Timing 2713 Effect of Varying Air Fuel Ratio 2814 Concentrations of Combustion Products 33

VI COMPARISON WITH THEORETICAL CALCULATIONS 3415 Method of Calculating Temperatures 3416 Comparison of Calculated and Observed Tempera

tures 37VII CONCLUSIONS 44

17 Summary of Results 4418 Conclusions. 45

VIII APPENDIX. DEFINITIONS AND NOMENCLATURE 46


8/58

LIST OF FIGURESNO P G

1 Diagram of Line Reversal Apparatus 92 Fuel and Air Metering Apparatus 143 Air Meter Calibration Curve 144 Fuel A S T M Distillation Curves 175 Diagram of Optical System 216 Sample Flame Temperature and Lamp Calibration Curves 257 Effect of Spark Timing on Temperature Constant Air Fuel Ratio 278 Variation of Maximum Combustion Temperature with Spark Timing 289 Effect of Variation of Shutter Opening on Temperature Measurements. 29

10 Cylinder and Manifold Pressures 3111 Temperature Curves with Varying Air Fuel Ratios 3212 Concentration of H 20 and CO 2 at Maximum Observed and CalculatedTemperatures 3313 Observed and Calculated Maximum Combustion Temperatures 3614 Crank Angle at Maximum Combustion Temperature 4015 Apparent Losses at Maximum Observed Temperatures 41

LIST OF TABLES1 Sample Test Data 262 Summary of Calculated Results 353 Summary of Test Results 374 Results Calculated from Observed Temperature 39


9/58

FLAME TEMPERATURES IN AN INTERNAL COMBUSTIONENGINE MEASURED BY SPECTRAL LINE REVERSAL

I. INTRODUCTION1 Object of Investigation.-A fuel burning in the cylinder of an

internal combustion engine forms a gaseous mixture whose properties,although not uniform, may be assumed to vary continuously through-out the combustion space. Such a mixture constitutes a homogeneoussystem and its thermodynamic state at any instant can be completelydetermined only when suitable values can be assigned to the vari-ables p V T cl c2 . . n (pressure, volume, temperature, andgas concentrations) either by direct measurement or by calculation.Methods for measuring instantaneous values of p and are welldeveloped, but independent determinations of either the instantaneoustemperatures or concentrations are considerably more difficult. Theconcentrations of some of the gaseous components have been deter-mined, using a sampling valve located in different parts of an enginecylinder*; but the experimental methods employed in making nearlyall previous measurements of the gas temperatures in an enginet areopen to criticism and are entirely inadequate for engine speeds above500 r.p.m. Since knowledge of the temperature is of major importance,it seemed desirable to attempt direct temperature measurement byan optical method capable of following the rapid fluctuations whichoccur in the engine but not affecting the combustion process itself.The method of line reversal, which has been used extensively in thedetermination of the temperature of stationary flames seemed to meetthese requirementst and has, therefore, been used to measure the in-stantaneous flame temperatures during combustion in a spark ignitionengine operating under moderate compression and air-fuel ratios vary-ing from 9:1 to 20:1. This investigation is a continuation of the studyof combustion and explosion phenomena carried on by the EngineeringExperiment Station over a period of years.

*Withrow, Lovell, and Boyd; Ind. and Eng. Chem. Vol. 22 p. 945 1930).tHopkinson: Phil. Mag., Vol. 13, S-VI, p. 84 1907).Calleindar and Dalby Proc. Royal Soc., Vol. 80, p. 57 1908)Coker and Scoble: Proc. Inst. of Civil Eng., Vol. 196, p. 1 1913)IThe results of another optical method used by one of the authors in a study of thesame problem are given in Ind. and Eng. Chem., Vol. 24 p. 867 1932). A Study of Explosions of Gaseous Mixtures, Univ. of Ill. Eng. Exp. Sta. Bul. No .133, 1922. An Investigation of the Mechanism of Explosive Reactions, Univ. of Ill. Eng. Exp.Sta. Bul. No. 157, 1926. A Thermodynamic Analysis of Internal Combustion Engine Cycles, Univ. of Ill. Eng.Exp. Sta. Bul. No. 160, 1927.


10/58

ILLINOIS ENGINEERING EXPERIMENT STATION

2. Acknowledgments.-The authors wish to express their apprecia-tion for the excellent facilities generously made available to themby DEAN M. S. KETCHUM, Director of the Engineering ExperimentStation, and PROF. A. C. WILLARD, Head of the Department of Me-chanical Engineering, where the work was performed. Personal ac-knowledgment is due PROF. A. P. KRATZ for his ready assistance andconstant interest.

II. COMBUSTION IN AN ENGINE3. Thermodynamic Variables p V T c c . c,.-Strictly

speaking, measurement of any thermodynamic property of a systemmay be made, and a definite meaning attached to the result, onlywhen the system investigated is in complete equilibrium. Obviouslygases burning turbulently in an engine are not in complete equilibriumand the only quantity that can be determined with physical certaintyis the total volume occupied by the gases at any instant. If anymeaning is to be attached to the pressure, temperature, and the severalconcentrations it is that of an effective average throughout the totalvolume as recorded by some measuring device. This recorded valuewill be influenced always by the properties of the measuring instru-ment. Measurements of pressure, using any one of several modernhigh speed indicators, give such effective values, since pressureequilibrium is established throughout the region at velocities compa-rable to that of sound. This velocity is far above that of the pistonand the assumption is quite justified that such determinations ofpressure give values which may be interpreted correctly as charac-teristic of the effective pressure equilibrium existing at any giveninstant.Measurements of the concentrations, ca, c2 . cn of the nconstituents in the combustible at different phases of the cycle areopen to more definite questioning than the pressure measurements,for several reasons. In an actual engine the mixture cannot be uniformwhen ignition occurs; and the normal flame velocities, though muchhigher than the piston velocity, are less than that of sound. Hencethe concentrations determined by any sampling valve method, if aver-aged, will give effective values which will have a wider range of un-certainty than that of the corresponding pressure measurements. Alsothe samples themselves will be removed from a rapidly burning mix-ture at high temperatures and of necessity analyzed at some later time

* Flame Movement and Pressure Development in an Engine Cylinder, N.A.C.A. ReportNo 399 1931.


11/58


12/58


assuming the gases in the cylinder to behave as a gas thermometer.But such extrapolation is open to criticism, since the concentrationsof these gases are changing and uncertain.

To obtain the instantaneous values of the effective .temperaturethe thermometric substance must be distributed throughout the entireregion where the temperature is desired, and must have a heat capa-city negligible as compared with that of the working substance. Theseconditions can be satisfied by introducing in very small quantities,compared with the total amount of gas present, a thermometric sub-stance which is easily vaporized. The temperature of this substancecan then be found by comparing its brightness at any given wave-length with that of a continuous radiator whose brightness tempera-ture at that wave-length is known. The amount of vapor presentmust be sufficient to radiate and absorb light according to Kirchhoff sLaw, and this condition can be attained with extremely small quanti-ties of vapor if comparison is made at a wave-length correspond-ing to some resonance radiation of the vapor. Under these circum-stances, investigation has shown that the thermometric substancehas a negligible effect on the flame temperature. So also in the engineno effect on combustion was observed and the temperature thus deter-mined may be interpreted as the effective temperature at the instantof comparison.

When such effective temperatures are known, corresponding con-centrations can be computed, assuming chemical equilibrium. Theseconcentrations will likewise be effective values and, while their un-certainty will be at least as great as that of the temperature measure-ments, they should indicate in a general way the progress of thechemical reactions.

4 Temperature Measurement by Line Reversal. The method oftemperature measurement just outlined is known as the line-reversalmethod, and has been used extensively in measuring the temperaturesof a wide variety of flames.t The principles involved can be readilyunderstood by considering the application of the method in the deter-mination of stationary flame temperatures. In Fig. 1, representsa bunsen flame into which a metallic vapor such as sodium is intro-duced. The vapor must be that of an element whose resonanceradiation is in a wave-length region which can be easily observed.Radiation from the tungsten ribbon filament lamp at L passes

*Jones. Lewis, Friauf and Perrott: Journ. Amer. Chem. Soc., Vol. 53, p 869 1931).tLoc. cit., p 8.Henning and Tingwalt: Z. Physik, Vol. 48. p 805 1928).Loomis and Perrott: Ind. and Eng. Chem., Vol. 20 p 1004 1928).Griffiths and Awberry: Proc. Roy. Soc., Vol. 123, p 401 1929)


13/58

FLAME TEMPER TURES IN AN INTERNAL COMBUSTION ENGINE

E::jA A c l/vFIG. 1 DIAGRAM OF LINE REVERSAL APPARATUS

through the flame to the slit of the spectroscope at S. The flame ispractically transparent to radiation in the visible region except at X,the wave-length of the resonance radiation of the metallic vapor inthe flame. At this wave-length due to the presence of the vaporstrong absorption as well as emission occur. Let Ex and EX be themonochromatic emissive power at the wave-length X for the lampfilament and the flame respectively and let a be the absorptivity ofthe flame at the same wave-length. The energy which reaches theslit of the spectroscope at this wave-length will be E 1 - al)Ex,assuming that reflection at the flame surface is negligible. If

EX > x E, bright line will be observed crossing the continuous spectrum ofthe filament radiation at X. If

EX x

dark or reversed line will appear at this wave-length. For theparticular case whenE = a x 1)

neither the bright nor the reversed line will be seen and the flamewill be emitting s much energy of wave-length X as it is absorbing.

r


14/58


If the flame satisfies Kirchhoff s Law for radiation of wave-length X, then S x 2)a/where E, is the monochromatic emissive power of a perfect radiatorat the wave-length X the true temperature of this radiator being thesame as that of the vapor in the flame. Therefore,

x = Ex 3)From Wien s Equation

Ex . exp. - 4)where Sx is the brightness temperature of the lamp filament at thewave-length X. Also

Ex exp. 5)where T is the true temperature of the vapor in the flame. Hence

T = Sx 6)If it is assumed that the vapor is in thermal equilibrium with thegases in the flame, and that its radiation is the result of purelythermal excitation, then the brightness temperature Sx of the lampfilament must be the same as the true flame temperature. The agree-ment between the values found for the temperature of stationaryflames by the line-reversal method and by other independent meth-odst is satisfactory proof of the validity of these assumptions.

The brightness temperature of the lamp filament can be measuredwith an optical pyrometer, so that the true temperature of the flamecan be readily determined. However, when the effective wave-lengthof the pyrometer screen is not the same as that of the resonance radi-ation of the vapor in the flame, it is necessary to make a slight correc-tion for the variation in emissivity of the lamp filament betweenthese two wave-lengths. If x is the monochromatic emissivity of thefilament at wave-length Xthen

Kohn: Ann. der Physik, Vol. 44, S. 14, p 749, 1914) has shown that this is true forstationary flames colored with alkali-metal vapors.tJones, Lewis, Friauf and Perrott: Jour. Amer. Chem. Soc., Vol. 53, p. 869 1931).


15/58

FL ME TEMPERATURES IN N INTERN L COMBUSTION ENGINE

CiEx = x .exp. _ , 7when T is the true filament temperature. Combining this equation

1 1 Xwith 4) gives - - = - In. 8)T SXEvaluating this equation for the two wave-lengths and eliminating T

1 l X2 1gives finally - = - In. - In.e 9)Sx2 C C2 Sx

where X, is the effective wave-length of the pyrometer screen and X2is the wave-length of the radiation from the vapor. The pyrometermeasurement gives the value of Sx the constant C is the same forall temperatures the emissivity for a tungsten filament is known overa wide range of wave-lengths and temperatures and hence the value ofSx2, which is equal to the true flame temperature can be calculatedfrom Equation 9).

In measuring the flame temperature in an internal combustionengine by the line-reversal method it is only necessary to pass thelamp radiation through the flame in the engine by means of suitablewindows on opposite sides of the cylinder and to introduce sufficientsodium into the combustible to make it possible to observe both thebright and the reversed line. A stroboscopic shutter driven from theengine crankshaft limits the observation to a short interval in suc-cessive cycles and thus gives an approximation to the instantaneousvalues of the effective flame temperature.

III. DESCRIPTION OF PP R TUS AND METHODS OF C LIBR TION5. Engine.-The engine of a farm lighting plant was used for the

experimental work. In order to obtain satisfactory control of operat-ing conditions the original air-cooled cylinder and head were re-placed with special water-cooled units. The design of these new partswas such as to insure adequate and uniform cooling to all parts ofthe combustion space and prevent the formation of hot spots. Thecylindrical combustion chamber was 3 inches in diameter this beingthe bore of the engine cylinder. The jacket space stopped 1 inches


16/58

ILLINOIS ENGINEERING EXPERIMENT ST TION

from the top of the cylinder, so that a solid ring was left in whichfour equally spaced -inch holes were drilled. In two of these holes,on opposite sides of the combustion chamber, the cylinder windowswere placed. These windows were of quartz and were held in shellswhich were screwed into the cylinder so that the inner surfaces of thewindows were flush with the surface of the combustion chamber. Thevalves were located in the flat cylinder head, thus providing a com-bustion space which could be completely machined and which con-tained no cavities. The height of the chamber could be adjusted from1.000 inch to 1.667 inches by means of rings of varying thicknessplaced between the cylinder and head, thus changing the compressionratio from 6:1 to 4:1.A one-inch Stromberg plain tube carburetor was used, but withseveral modifications. Since the engine was operated within a narrowspeed range and with wide open throttle, only the main jets wereused and the idling jet was completely closed. The needle valve forthe main jets was replaced with one having a more gradual taperand a large graduated head to give finer adjustment. With the air-metering system employed, the pressure at the carburetor inlet wasapproximately 2 in. of water below atmospheric, so that it was necessary to seal the float chamber and connect it and the air bleed to theair inlet pipe. The capacity of a one-inch carburetor being consider-ably in excess of the demands of an engine having a displacement of41 cubic inches, the large venturi tube had to be replaced by onewith a smaller bore. Since the mixture could be enriched with themain needle valve for starting, the choke valve was removed.

The spark plug was located in the cylinder head just above one ofthe cylinder windows, and, due to space limitations, a 14-millimeterplug was used. The breaker cam was fastened directly to the cam-shaft, while the case carrying the breaker arm could be securelyclamped to prevent any variation of spark timing when the enginewas running. The current source was a potential divider across the110-volt d.c. power line which supplied a 12-volt ignition coil.

The engine was loaded by means of the direct-connected 32-voltgenerator with which the power unit was equipped. By connectingthe generator to the 110-volt d.c. power line with suitable resistance,it also served as a motor for starting and for motoring tests. Thegenerator field which was originally self-excited, was also connectedto the d.c. supply for starting and for independent voltage controlwhen loading the engine.


17/58

FL ME TEMPERATURES IN AN INTERNAL COMBUSTION ENGINE 13

6 Engine Control and Metering Apparatus. Of the numerousfactors affecting combustion in an engine, the following were selectedas being the most important and, therefore, requiring the closestcontrol:

a) Air-Fuel Ratio b) Compression Pressurec) Inlet Temperature d) Jacket Temperaturee) Spark Timingf) Speedg) Fuel Characteristics

For this particular series of tests the air-fuel ratio was the variablewhile the remaining factors were held as nearly constant as possible.A few tests with constant air-fuel ratio and varying spark timingwere also made.

a) Air-Fuel RatioThe air consumption of the engine depends only on engine speed

and throttle position, and hence the rate was constant, since the loadwas adjusted to give a constant speed, and the throttle was alwayswide open. The fuel consumption, therefore, determined the air-fuelratio and was controlled by the setting of the carburetor needle valve.The air and fuel metering apparatus are shown in Fig. 2. The airwas measured with an Emco No. 1 Dry Test Meter. Between it andthe carburetor inlet was an expansion chamber with a volume 52times the engine displacement. This served to reduce the pressurefluctuations at the meter to less than 0.1 in. of water. The drop inpressure through the meter was measured by means of a manometerconnected to a piezometer ring at the outlet, and was found to beabout 2 in. of water at operating rates. A thermometer indicated thetemperature of the air leaving the meter. The temperature beingknown, and the pressure in the bellows assumed to be less than atmos-pheric by half the total drop through the meter, the density of theair could be determined.

The meter was calibrated under actual operating conditions ofpulsating flow by allowing the engine to pump a known weight of airthrough it from the air-weighing tank. Atmospheric pressure wasmaintained at the meter inlet by adjusting the throttling valves ofthe air-control manifold and the engine was operated at constantspeed. From Fig. 3, in which the results of the calibration are showngraphically, it is evident that the correction is independent of the

See Engineering Experiment Station Bulletin No. 2 7 for a description of the air weigh-ing equipment.


18/58

ILLINOIS NGIN RING XP RIM NT STATION

FIG 2. FUEL AND AIR METERING APPARATUS

~5/.S I -- ---

1 90

^~~~~~~ ZZIZIIZIZ^^II

- - -- - - - - - - - - - - -~ % * / A ^ T ,/ '

J l I I I 1 1 1 1 1 1 I

Rate of A/ F ow t? /b p r /?.FIG 3 AIR METER CALIBRATION CURVE

KKK=3

5s


19/58

FL ME TEMPER TURES N N INTERN L COMBUSTION ENGINE

rate of flow. The meter was found to read about 3 per cent high and,since such meters can generally be relied upon for an accuracy of atleast 1 per cent, this greater error was probably caused by the pul-sating flow. Calibrations were made at the beginning and end of thetests with an interval of more than a year between, and the change inthe meter correction factor was from 0.971 to 0.964 or 0.7 per cent.

The fuel was measured volumetrically by means of the smalltanks shown in Fig. 2. Using a fuel of known density, the volumeof these metering tanks was found, and the weight of a measuredamount of any fuel could then be found, once its density was known.In calibrating the tanks a chemical balance was used to weigh theircontents and a Westphal balance was used for all density determina-tions.

b) Compression PressureThe compression pressure depends upon the compression ratio and

the density of the incoming charge. The compression ratio was un-changed throughout the tests, being determined by the volume of thespace between the cylinder and piston head. The charge density wascontrolled by the inlet temperature, since the volumetric efficiency wasconstant, the engine being operated at constant speed with wide openthrottle, and the change in atmospheric pressure had a negligible effect.

c) Inlet TemperatureThe inlet temperature was controlled by a jacket around the inlet

manifold between the carburetor and cylinder through which hotwater from the cylinder jacket could be circulated in varying amounts.No attempt was made to determine the exact temperature of the com-bustible passing through the manifold, but a thermometer at the endof the manifold jacket served as an indicator for duplicating inlettemperature conditions.

d) Jacket TemperatureThe cooling water entered the cylinder jacket at the bottom and

passed through the solid ring at the top of the cylinder by means ofeight equally spaced -inch holes into the cylinder head. Part. ofthe hot water leaving the outlet in the cylinder head returned to theinlet in the cylinder where it was recirculated by means of an injectorthrough which the cold water entered the jacket. The advantages ofthis arrangement are more rapid circulation, small temperature differ-ence between the water inlet and outlet, and small heat capacity inthe cooling system. A thermometer in the outlet in the cylinder head


20/58

ILLINOIS NGIN RING XP RIM NT ST TION

indicated the jacket temperature which was controlled by means of avalve in the supply line.

e) Spark TimingThe spark timing was controlled in the usual manner by shifting

the breaker housing carrying the breaker points. For the earlier teststhe spark indicator incorporated in the timer for the balanced dia-phragm pressure indicator was used but later a spark indicator ofsimilar design was mounted on the crankshaft of the engine to pro-vide for more accurate spark setting. For all tests with constantspark timing 36 deg. advance was used this being the timing whichgave the maximum power with the theoretical air-fuel ratio.

f) SpeedThe throttle position being fixed, the speed was controlled entirely

by changing the load on the engine. This was done by means of arheostat in the armature circuit of the generator. An Elgin tachometergave speed indication while operating conditions were being adjustedbut during a test the speed was determined from the total revolutionsand the elapsed time occurring while the fuel metering tank was beingemptied. A breaker operated by a cam on the engine camshaft in-terrupting the circuit of a magnetic counter constituted the revolutioncounter while a manually-operated stop watch measured the time.

g) Fuel CharacteristicsThe desirable fuel characteristics were: fairly high volatility to

insure reasonably homogeneous combustible mixtures at both low andhigh air-fuel ratios and a minimum amount of gum to prevent ex-cess carbon deposits on the cylinder windows. A commercial grade ofaviation gasoline met these requirements satisfactorily. The fuel wasstored in a barrel which was kept filled with nitrogen to minimize anychanges in its unsaturated components. The curves in Fig. 4 are theresults of A.S.T.M. distillations made at intervals during the testsand they indicate satisfactory fuel stability. The ultimate analysisof the fuel showed the carbon and hydrogen percentages to be inreasonably close agreement with the corresponding percentages in oc-tane C8H18 which was the formula assumed for all combustion cal-culations.

7 Temperature Measuring Apparatus. In general the necessaryapparatus for making line-reversal temperature measurements on aflame consists of a source of continuous radiation a pyrometer for de-termining its temperature a spectroscope and an aspirator for intro-


21/58

FL ME TEMPER TURES IN N INTERN L COMBUSTION ENGINE 17

K

FIG. 4. FUEL A.S.T.M. DISTILL TION CURVES

ducing the metallic vapor into the flame. In the first few engine testsa tungsten ribbon-filament lamp was used as a source; but the highintensity required, together with the vibration of the engine causedthe lamp to fail after only a few hours use. Another lamp with a Vshaped ribbon filament* supported by a Julius suspension operatedsatisfactorily during the remainder of the tests. In order to measurethe maximum temperatures inside the cylinder, a brightness tem-perature of the lamp filament at 0.589 t of about 4800 deg. F. abs.was necessary. Tungsten has an emissivity at this wave-length andtemperature of 0.433 so that the corresponding true temperature of a

*This lamp was one of several made by Mr. Colbey of the Physics Department Shops.

I


22/58

ILLINOIS ENGINEERING XP RIM NT STATION

ribbon filament would be 5285 deg. F. abs.* By using a filament witha V-shaped fold the true temperature for the same brightness temper-ature is much lower since the emissivity of such a cavity is a charac-teristic of the reflectivity of the filament and the angle of the V.Mendenhallt has shown that the emissivity of a wedge-shapedcavity is

x = 1 - rx 10)where x is the reflectivity of the surface for the wave-length X, and

180n = - 11)

a being the angle of the wedge. For tungstenro0.89 5

and with an angle of 60 deg.E0.589 = 0 875

This gives a true temperature of 4871 deg. F. abs. for a brightnesstemperature of 4800 deg. F. abs., or a reduction of 400 deg. F. as com-pared with the straight ribbon filament. Also, since the reflectivity r,does not change appreciably in this wave-length region, the bright-ness temperature at the wave-length of the sodium line, X 0.589 uis the same as that at the wave-length of the pyrometer screen, orfor = 0.665 p and the correction outlined in discussing the methodneed not be applied when a V-filament source is used. Filaments withthe angle less than 60 deg. were tried, and, while their emissivitieswere even higher, they failed before reaching sufficiently high tempera-tures due to an arc forming across the opening of the V.

The brightness of the lamp filament was controlled by varyingthe current through it by means of a rheostat in the lamp circuit.Since it was inconvenient to measure the brightness temperature of thefilament when the spectroscope was being used, the lamp current wasdetermined whenever a balance between the flame and lamp radiationhad been obtained. The corresponding brightness temperatures werethen found by calibrating the lamp at the completion of each test.The lamp was calibrated by measuring the brightness temperature for

See Forsythe and Worthing: Astrophys. Journ., Vol. 61, p 146 1925) for radiation char-acteristics of tungsten filaments.tAstrophys. Journ., Vol. 33, p 91 1911).


23/58

FL ME TEMPERATURES IN AN INTERN L COMBUSTION ENGINE 9

a number of different currents and plotting these results to determinea calibration curve. A Leeds and Northrup optical pyrometer wasused in measuring the brightness temperature and the current wasmeasured by means of a potentiometer across a standard 75-ampereammeter shunt in the lamp circuit.

The optical pyrometer was checked against a standard ribbon-filament lamp and was found to be in satisfactory agreement. Thisstandard lamp had been calibrated by Dr. W. E. Forsythe at theNela Park Laboratory of the General Electric Company and wasused as a primary standard for all temperature measurements. Inmeasuring the temperature of the V-filament lamp however the py-rometer objective was replaced by the lens which focused light onthe spectroscope slit and hence it was necessary to compare the py-rometer with the standard lamp using this lens. When the same lenswas used with both the pyrometer and the spectroscope the V-fila-ment lamp could be calibrated without changing the optical systemit being only necessary to interchange the two instruments. This sim-plification saved considerable time since the lamp was calibrated afterevery test. The pyrometer with this lens was found to read about10 deg. low at 3180 deg. F. abs and since this represents an error ofonly 0.3 per cent no corrections were made.

A small Kriiss spectrometer with a glass prism was used andbeing of very rigid construction it. was well adapted for this par-ticular work. The engine subjected the entire optical system to con-siderable vibration but by clamping the spectrometer to a tablesecured to the engine foundation satisfactory optical alignment wasmaintained. It was of vital importance that this alignment remainunchanged during a test because only the V of the filament was freefrom temperature gradients and this area of the image which wasrelatively small had to fall upon the slit of the spectroscope. Thebottom of the V which appeared as a dark line across the imageacted as a convenient reference line when setting up either the spec-troscope or the pyrometer.

The sodium in the form of a NaOH solution was introduced intothe engine manifold at the elbow just above the carburetor. A glassaspirator formed a fine spray of the solution which could be pickedup and carried into the combustion chamber by the air stream in themanifold. The air for aspirating the solution was drawn from thesurge tank in the carburetor air supply line and was thereforemetered along with the main air supply for the engine The aspi-rator outlet was connected to the engine manifold with rubber tubing


24/58

ILLINOIS NGIN RING EXPERIMENT STATION

so that a pinch cock could be used to control the amount of solutionentering the engine cylinder.A line drawing of the optical system is shown in Fig. 5. Thelamp was placed at the principal focus of the first lens and the slitof the spectroscope at that of the second lens. Thus a parallel beamof light passed through the engine cylinder. This represented a slight

modification of the optical system used in measuring the temperatureof a stationary flame by line-reversal. In the latter case, the lampand flame are placed at conjugate foci of the first lens while the flameand slit are at conjugate foci of the second lens. The foci being thosewhich give equal object and image size, a full-sized image of thelamp filament is produced in the middle of the flame by the first lenswhile the second lens refocuses this image on the slit. The light passesthrough the flame in a narrow conical beam and the temperaturefound is characteristic of a very small region of the flame. In theengine however the desired temperature was the effective temperaturewhich would be characteristic of the entire region at any instant andaccordingly a parallel beam of light of maximum cross section whichwould traverse a larger portion of the region was used instead of asmall conical beam.

The temperature of the gases inside an engine cylinder is fluctu-ating rapidly and in order to obtain approximately instantaneousvalues of the temperature at different points in the operating cyclea stroboscopic shutter was used to limit the interval during whichradiation from the lamp and flame reached the slit of the spectroscope.The shutter operated at half the crank-shaft speed and had anangular opening of 1 deg. However it was located just outside thecylinder window where it moved across the parallel beam of lightwhich passed through the cylinder and thus it allowed radiation toreach the spectroscope during considerably more than 2 deg. ofcrank-shaft rotation. Moving the shutter by means of its phase-changing gear while the engine was stationary it was found that lightreached the eye-piece of the spectroscope for an interval correspondingto about 18 deg. of crank rotation. In several tests made to deter-mine the effect of the length of shutter opening on the observedtemperatures the optical system was changed so that the shutter wasin the focal plane of the first lens. With this arrangement a shutteropening of one degree allowed light to reach the eye-piece duringthree degrees of crank rotation.

An aluminum disc 6 inches in diameter with an opening partlycovered by a thin brass plate formed the shutter. This plate was


25/58

FL ME TEMPER TURES IN N INTERN L COMBUSTION ENGINE

zFCLCQ

C

CCL

CC


26/58


held in place by screws passing through circular slots so that itcould be shifted to change the angular opening. The drive for theshutter was taken from the camshaft through a train of bevel gearswhich included a phase-changing gear, by means of which the timingof the shutter relative to the crankshaft could be changed. Rotatingthe housing of the phase-changing gear through a given angle movedthe shutter through twice this angle relative to the crankshaft. Thehousing was rotated by a worm gear, and a counter, driven from theshaft of the worm, indicated the phase relation between the shutterand crankshaft. As it was difficult to eliminate lash entirely fromsuch a gear train, which included five pairs of bevel gears, reason-able care was exercised in setting the gears, and the actual phasedifference due to this lash was measured. This was done by placinga spark plug in the window opening on the opposite side of the cylin-der and determining the apparent spark timing by changing the phaseof the shutter until the spark was first observed, the line of sightbeing determined by two screens with narrow openings. The enginewas operated at normal speed and the actual spark timing wasfound from the spark indicator on the crankshaft. The differencebetween the actual and the apparent timing was taken to be the phaseerror of the shutter due to lash. Except for a narrow range of about -5 degrees from top center, the phase error was constant in amountand direction, so that it was only necessary to subtract this amountfrom the counter reading to obtain the actual timing of the shutter.

8 Pressure MeasuringApparatus. A Bureau of Standards bal-anced-diaphragm indicator was used in measuring the gas pressure inthe engine cylinder, the pressure element being screwed into one of theholes in the cylinder wall at 90 degrees from the windows. At firstconsiderable difficulty was experienced with the indicator due todiaphragm failure; the regular plated diaphragms quickly corrodedeven though they were frequently removed and cleaned. This wasovercome by using stainless steel diaphrams 0 006 in thick. Thecontact timer for the indicator was driven from the camshaft throughthe phase-changing gear which drove the stroboscope shutter. Thus,after the timer and shutter were set in phase, they could simultane-ously be shifted to any point, in the cycle by adjusting the phase-changing gear without entailing any change in their relative timing.The phase difference with respect to the crankshaft was found bymeasuring the spark timing simultaneously with the indicator incor-

The steel, an 8 per cent nickel, 8 per cent chrome alloy, was furnished by the ColdMetal Process Company, Youngstown, Ohio.


27/58

FL ME TEMPER TURES N N INTERN L COMBUSTION ENGINE 23

porated in the timer and the one mounted on the crankshaft, thedifference between two corresponding readings being the phase errorcaused by lash in the driving gear. Corrections could then be madeas previously described for the phase error of the rotating shutter.

Instantaneous pressures in the inlet manifold were also measuredwith the same type of indicator. A piezometer ring around the inletmanifold close to the cylinder head served as a pressure tap to whichthe indicator was connected with a short -inch pipe nipple. Pressuremeasurements were taken during the induction period, starting sometime before the inlet valve opened and continuing until after it hadclosed.

IV EXPERIMENTAL PRO E URE9 Method of Conducting Tests. Before starting the engine for

a test, the rotating shutter was removed and the spectroscope care-fully aligned with the rest of the optical system so that the samepart of the image of the filament V always fell on the slit. The spec-troscope was then clamped in place and, after the shutter had beenreplaced, the engine was started. No sodium was introduced into thecylinder until the engine had been running for a sufficient time toheat the cylinder and manifold jacket approximately to operatingtemperatures. Any unvaporized portions of the NaOH solution thatreached the combustion chamber collected on the walls and windowsand quickly pitted the latter. t was necessary, therefore, to avoidintroducing any excess of the solution, only enough being aspiratedinto the manifold to give satisfactory reversal of the spectral line,and this only after the engine was thoroughly warmed up. The car-buretor needle valve was next adjusted to give the desired air-fuelratio, the fuel and air consumption being measured only when theengine operating conditions had become constant. A series of flametemperature measurements were then made, starting with the pistonat its top center position and continuing at intervals of ten degreesof crank rotation to a point in the operating cycle well past that atwhich the maximum temperature was observed. No temperatures weremeasured before the piston reached top center, because, with the ex-ception of those tests in which the mixture was very rich, the burn-ing had not progressed sufficiently before this point in the cycle toproduce conditions within the cylinder giving definite reversal ofthe spectral line. n determining the balance point between the-bright and reversed lines it was necessary to select an average con-dition representative of a large number of separate cycles. For the


28/58

ILLINOIS ENGINEERING XP RIM NT ST TION

temperature w s found to vary noticeably from one cycle to anotherand so the lamp intensity was adjusted until the bright and reversedlines e ch appeared with about equal frequency. When the balancepoint had been found the current through the lamp was measured bymeans of the potentiometer and recorded. As soon as the traversearound the cycle was completed the air and fuel consumptions weremeasured again and a second series of temperature measurementstaken. These observations were also made at ten-degree crank inter-vals starting midway between the first two readings of the previousset thus giving temperatures at five-degree intervals as the fin l re-sult. When this series was completed the air and fuel consumptionwere measured once more and the engine stopped. While a test wasin progress the jacket and manifold temperatures and the speed werefrequently checked and adjustments were made when necessary butno change was made in the carburetor setting after the initial ad-justment.

10 Lamp Calibration. When flame temperatures were beingmeasured the intensity of the radiation from the sodium vapor andthat of the radiation from the lamp filament having the same wave-length were balanced in the region within the engine cylinder. There-fore the brightness temperature of the lamp radiation inside the cylin-der was determined for the required range of filament currents by re-moving the rotating shutter and the left cylinder window see Fig. 5and calibrating the lamp as described in Section 7. Any absorptionor reflection due to the shutter and this window or to the opticalparts of the spectroscope reduced the intensity of the radiation fromthe lamp and the sodium vapor by exactly the same amount andhence they were removed from the optical system during calibration.In order to determine the reduction in intensity of the lamp radia-tion caused by the deposit on the inner surface of the right cylinderwindow the lamp was calibrated at the end of each test. With le nmixtures the decrease in intensity was negligible but with rich mix-tures the brightness temperature was sometimes lowered by as muchas 100 deg. F. In any case most of the deposit occurred while theengine w s warming up or while operating conditions were being ad-justed so that a calibration at the end of a test was found to be repre-sentative of conditions during the period in which temperaturemeasurements were being made.


29/58

FLAME TEMPER TURES IN AN INTERN L COMBUSTION ENGINE

4400

4300

>1

KK

4000

389

Q 7 7

6

A 7K

I.r6 77/

A/ el RP/ o /3.93\czz340 0 20 40 60 86 A0 /20Crolr? Ag; e 2egr ees Af/er op Ce/?/er

9 /0 //

FIG. 6 SAMPLE FLAME TEMPERATURE AND LAMP CALIBRATION CURVES

V. TEST RESULTS11 General Statement. The line-reversal method while well

established as a means of measuring the temperature of stationaryflames has not been used extensively in measuring rapidly changingtemperatures such as those which occur in a bomb or an enginecylinder. Accordingly it was essential in applying the method tothe measurement of flame temperatures in an engine to repeat theexperiment many times to determine what accuracy might be expected.As already pointed out every effort was made to keep all conditions

^ \ / enffer~a i/ ~ff

es A o35t

.


30/58


TABLESAMPLE TEST DATAData from Test No. 5

Air-Fuel Ratio 13.93 Observers: RFP and AEHSpark Advance 36 deg. Date: January 25, 1933

Flame Temperature Measurements

Phase GearCounterReading

920930940950960970980990945955965975985995

CrankAngledeg.

716151413121

463626 6

356

LampCurrent

8.7489.2389.7151 3531 92211.44011.63411 27710.0681 7 611 28511.62811.45211.083

Temp.deg.F.abs.

3825396340834215433743834298401641654300438243404252

Calibration of Lamp W-79)

PyrometerLamp Pyrometer Current Temp.Current Screen ma. deg.F.abs.

9.291 Ex. High 406 38379.956 Ex. High 417 399310.583 Ex. High 427 412011.177 Ex. High 438 427411.378 Ex. High 443 433911.823 Ex. High 450 4430

Potentialdrop across shunt in light circuit in milli-volts.

uniform during any given test. The results of one such test, togetherwith the lamp calibration, made immediately after its completion, aregiven in Table and plotted in Fig. 6 It may be noted that thedeviation of the individual points from a smooth curve drawn torepresent the observed changes is surprisingly small; but it was foundthat the agreement between comparable sets of observations takenat different times was not as close as might be expected from the,consistency of the results obtained during a single run. In searchingfor the cause of comparatively large variations between runs, varia-tions in atmospheric humidity were checked, but such changes asoccurred were found to have little or no effect on the measured re-sults. The installation of a more accurate spark indicator increasedthe precision with which analogous runs would check and for this

T


31/58

FL ME TEMPER TURES IN AN INTERN L COM USTION ENGINE

T st 1/

C 2Z 40 60 80 /00CrcrA Ang/e i Degrees 1/ fer 7To2 Ce erFIG. 7 EFFECT OF SPARK TIMING O TEMPERATURE-CONSTANT AIR-FUEL R TIO

reason a series of runs were made with three settings of the carburetorgiving different air-fuel mixtures and with different spark settings foreach mixture. These together with a large number of runs made withthe 36-deg. spark setting and different air-fuel ratios constitute themain experimental results included in this bulletin.

12 Effect of Spark Timing. The results of the experiments onspark timing are summarized in the curves of Figs. 7 and 8 Thecurves of Fig. 7 show how the temperatures changed with the sparktiming the air-fuel ratio being kept as nearly constant as possible.The phase in the cycle at which the maximum temperature occurs is


32/58


33/58

FLAME TEMPERATURES IN AN INTERNAL COMBUSTION ENGINE 29

qu u fv ouu CC ~rt A 7g e Oe grees Af e op enoerFIG 9 EFFECT OF VARIATION OF SHUTTER OPENING ON TEMPERATURE

MEASUREMENTS

it was concluded that the layer of cooler gases in contact with thecylinder walls was too thin to affect the results appreciably. Anytemperature measured was of course an average taken along a diame-ter of the cylinder and was averaged also over a sequence of manycycles. The method very definitely showed the variations in tempera-ture from cycle to cycle and estimates indicate that it could be usedto detect temperature fluctuations of approximately 10 deg. F. at thetemperatures measured provided the temperature remained constantwhile a balance was being observed. To obtain sufficient visibility afinite shutter opening was used most of the observations being madewith light entering the spectroscope during about 18 deg. of crank-angle. However runs made with the light entering the spectroscopeduring but 3 deg. of crank angle gave results for maximum tempera-tures not measurably different from those made with the larger angle.In Fig. 9 two curves are drawn from runs with essentially identical air-


34/58


fuel ratios, one made with a 3-degree interval and the other with the18-degree interval. With proper corrections made for differences inthe optical setup, the maxima of the two curves are seen to agreewithin the limit of experimental error, though the one made withthe wider interval is noticeably broader. This indicates that tem-perature at the maximum remains constant over an appreciablepart of the cycle. The flicker accompanying observations with the 3-degree timing was more noticeable, but the variations in the tempera-tures from cycle to cycle in all the measurements were considerablygreater than the precision of the setting.

Complete measurements of gas pressures for six widely differingair-fuel ratios had been made with the engine used in this study byone of the authors in connection with a previous study of radiationmeasurement. Pressure measurements were therefore, not made foreach run in this investigation, only two being made as checks. Oneof the curves taken from the previous study is included here in Fig.10. This shows also the data from which the volume at the start ofcompression is determined. The data for the temperatures during partof the cycle for the same air-fuel ratio, but taken by the line-reversalmethod, are also shown here. In another run, the data for which arenot included, it was found that the line-reversal method could beused to measure the temperature in the engine well after the exhaustvalve opened and even beyond the bottom center position. Thetemperature seems to reach maximum values before the pressure, butduring the later portions of the cycle the pressure and temperaturecurves parallel each other quite closely. Most of the runs were madewithout pressure measurements, and with only enough readings oftemperature to completely determine the maximum. The set of curvesin Fig. 11 were selected as good examples of the experimental dataand as illustrating also the manner in which the temperature varies asthe air-fuel ratio is changed. The curves with broken lines weredrawn from data taken with fuel mixtures richer than the theoretical,while the full line curves were drawn from data obtained with mix-tures leaner than this. The trend toward and away from a maximumtemperature is very conspicuous. Observations were made over aswide a range of air-fuel mixtures as could be used, still retainingreasonably smooth engine operation. The highest temperature wasobtained with an air-fuel ratio of 13 9 to 1 which corresponds to amixture containing 91 per cent of the theoretical amount of air.

Loc. cit., p 5


35/58


36/58


SA -4- Fe/ Fal X ILess 1 /- T12eorel la/

iGreeer /Fhaearef/iN:r~ea 1; Th earellec

Crank Ane e D erees fter 7-o enTerFIG. 11 TEMPER TURE CURVES WITH V RYING AIR FUEL R TIOS


37/58


/0

7 -:S6

y

: Obser~ed - rSTeeo t/y r te ^

.0

0___

FIG. 2

/0 /2 /4 /6 6Alr lf e/RP/o

CONCENTRATION OF sO AND C 2 AT MAXIMUMCALCULATED TEMPERATURES

OBSERVED N

14. Concentrations of Combustion Products.-When the instan-taneous temperatures of the gaseous mixture inside the engine cylinderare known the corresponding concentrations of the components canbe computed if correct equilibrium data are available. Thus the concentrations of HO0 and CO 2 were computed for the maximum observed temperature at six representative air-fuel ratios from the equi-librium data for the three reactions


38/58


H2 ,O . H20OCO + 2 O2 - CO 2

and H2 CO 2 H20 COassuming that combustion occurs at constant volume. The resultsof these calculations are shown graphically in Fig. 12, where thepercentages of the concentration for complete combustion are plottedagainst air-fuel ratio. In a similar manner concentrations were com-puted for the maximum temperatures determined by theoretical con-siderations and these results are also shown in Fig. 12. The datafrom which the curves in this figure were plotted are given in Tables2 and 4.

It may be noted that the concentrations found for the observedand calculated temperatures are in reasonable agreement up to anair-fuel ratio of 12.5 to 1. As the air-fuel ratio increases beyond thisvalue, both the H 20 and CO concentrations for the observed tempera-tures increase more rapidly than the corresponding concentrations forthe calculated temperatures. This is due to an increasing differencebetween the observed and calculated temperatures, the former beinglower and hence allowing more complete combustion.

VI COMPARISON WITH THEORETICAL CALCULATIONS15. Method o Calculating Temperatures By means of the

method developed by Goodenough and Felbeck,* the maximum flametemperature during combustion in an engine may be calculated. Thishas been done by Goodenough and Bakert for a wide range of com-pression pressures and air-fuel ratios. By interpolating between theirresults, the maximum temperatures could be found for a compressionratio of 3.86 to 1, which was the actual ratio from the closing of theinlet valve in all the present tests. This was done for a number of air-fuel ratios, and the results are represented graphically by the upperfull line in Fig. 13. The broken line in the same figure represents cal-culated temperatures found by slightly modifying the method men-tioned.

In considering the Otto cycle with insufficient air, Goodenough andBaker assumed that the residual gas mixture in the clearance volume,when the fresh charge enters, contains no 0.2. This is not strictly

* An Investigation of the Maximum Temperatures and Pressures Attainable in the Com-bustion of Gaseous and Liquid Fuels, Univ. of Ill. Eng. Exp. Sta. Bul. No. 139. 1924).tLoc. cit., p 5.:Page 28 loc. cit., p. 5.


39/58


HA-4:

H l CDA: ~U~ QACDr~Q

CHCD

0C tC DV000t txcOOOC

-* OGOC,-CDVI-V00000 0

0r

0V C 1CO 10VU.CCS rC^)C~ -CC0-t

wQT cowroc cc

[ (CDNNY000^

D00 x .C

CC X 0UCC00OCC

co ~ t

a1 0C2 00i^t~oCCC TCC(CD CC

- 0 CMO-00

Dio - o CDiwo0 CC0 0 0

iCO C-l CD-V O CC>CCC0Ci

O CDVO-I

VCOD00GO00

i00 > i

VF- 0 S000

D 0 -D 0t UI 0X i C -OM CO ifeVtFic t C CC


40/58


C~a ce e7 edemperai6're.s

S

4000

55 4SO

^.~3500'K

iO i

//NN

Ar-e/ A1 t/A-,L/--uel Raz7;/b

/

/8 20 eeFIG. 13 OBSERVED AND CALCULATED MAXIMUM COMBUSTION TEMPERATURES

true because at the temperature of this residual gas 2800 to 3400 deg.F. abs.) some dissociation must always occur. The gas mixture whichwould be in equilibrium under the existing conditions was found byevaluating the two equilibrium equations

log KP co), + Y log T = log x - log 1 - x) +NiT2Y log - log n.P2

y 1 - x)K 1w..)y)x 1 - y)

12)

13

for the temperature T of the residual gases, and solving for x and y.With this exception the method of calculation was the same as thatused by Goodenough and Baker. However, instead of the initial con-ditions assumed by them, such as the temperature and pressure of thecharge in the inlet manifold when the inlet valve opens, actual ex-perimental values were used. The temperatures were found by meansof the thermometer in the manifold, while the pressures were obtained

0

N

~ ~ _

Ob5et-vede/l2 leza't6/res


41/58

FLAME TEMPERATURES IN AN INTERNAL COMBUSTION ENGINE 37TABLE 3

SUMMARY OF TEST RESULTSCompression Ratio 3.86:1 Spark Advance-36 deg.

DateofTest

10- 6-3210-10-3210-31-3211 7 3211- 7-3211-14-3211-14-3211-14-3211-21-3211-21-3212 5-3212-22-321- 9-331- 9-331-16-33 6 33 23 331-25-331-28-332 3 332-13-332-13-332-14-332-14-332-16-334-24-334-24-33

AirPressureat Meterin. ofHg. abs.

29.3429.2429.2529.1729.0529.1229.0929.1029.6629.6729 1829.6729.3529.2829.2029.1129.1828.9529.4729.2929.1529.1329.4529.4629.0329.0528.93

AirTemper-atureat Meterdeg. F

87.589.077.790.093.586.285.087.084.284.590.883.087.088.585.788.587.282.584.081.283.883.782.082.784.885.385.7

MoistureContentlb. H20vapor per100 lb.of air

0.3860.5200.5020.7360.9970.3420.5350.6500.1330.1810.3340.3780.3060.3170.6900.6210.4210 5640.2570.1990.3800.4270.2970.2720.3090 4490.383

MaximumAi x rFuelRatio aturedeg.F.abs.

InletTemper-aturedeg. F

11511 812012 212 212011912 012 011612 212 212012512512512512512 412312 512 611911 712 112 0121

Calculated for dry air.

from curves of manifold pressure such as that shown in Fig. 10. Fromthese same curves the volume at the beginning of compression andhence the actual compression ratio could also be found. A summaryof these calculated results is given in Table 2. The net effect of thesemodifications was to leave the temperatures for lean mixtures un-changed and to raise those for rich mixtures by 50 to 75 deg. F. asindicated by the broken line portion of the upper curve in Fig. 13.

16. Comparison o Calculated and Observed Temperatures. Theobserved maximum temperatures given in Table 3 were also plottedagainst air-fuel ratios and are represented by the lower curve in Fig.13. The general shape of this curve is the same as that of the curvefor calculated values but the latter is from 600 to 1000 degreeshigher. The smallest difference occurs with an air-fuel ratio of 14to 1 when both calculated and observed temperatures are highest.

CrankPositionat Max.Temp.deg. pasttop center

11.013.515.014.079.0

1511.010.01816.016.018.011.017.013.525.021.021.019.011.013.014.021.012.0


42/58


This difference is undoubtedly due to a number of factors, the moreobvious ones being the following:

a) Effect of heat loss and work done up to position of maxi-mum temperatureb) Effect of NaOH solution

c) Effect of humidityd) Effect of finite shutter openinge) Effect of fogging of the cylinder window f) Effect of precision of the line-reversal methodg) Effect of chemical equilibrium conditions

a) Effect of heat loss and work done up to position ofmaximum temperatureIn calculating the temperatures, the assumption was made that

combustion takes place adiabatically and at constant volume. Undersuch conditions the energy equationAQ = AU + AW 14)

becomesHm ni 1 - x Hco - n2 1 - y HH = U3 U2 15)

using the notation of Bulletin No. 160.* In the actual engine, neitherof these conditions is satisfied since combustion continues over a finiteinterval during which heat loss must occur to the cylinder walls, andwork must be done on the moving piston. Therefore, the energy equa-tion should be

Hm - ni 1 - x Heo - n2 1 - y HH, h =U3 - U + W 16)

where h is the heat loss, and AW the work done. For any given test,AW may be found from the area under the pressure-volume curve be-tween top center position and that at which maximum temperatureoccurs. Now if it is assumed that the difference between calculatedand observed temperatures is due entirely to this work of expansion

*See Equation 21), page 21, loc. cit. Hm is the lower heat of combustion at constantvolume for the combustible mixture at the end of compression. This mixture contains the gaso-line of the fresh charge together with the H2 and CO of the residual gases left in the cylinderat the end of exhaust. The values given by Goodenough and Felbeck were used in calculatingall heats of combustion. For gasoline they assume the higher heat of combustion at constantpressure to be 20 000 B.t.u. per lb. while the gasoline used in these tests had a heat of com-bustion of 20 400 B t u per lb. The use of this latter value would have resulted in higher cal-culated flame temperatures; for any increase in Hm in Equation 15) above would necessitatean increase in U3 or an increase in T3 the temperature at the end of combustion.


43/58

FLAME TEMPERATURES IN AN INTERNAL COMBUSTION ENGINE 39TABLE

RESULTS CALCULATED FROM OBSERVED TEMPERATURE

Air Energy Calculations onFuel Observed mi mols of Combustion AW Apparent a Temp. er Yer per Heat Lossatio deg. F per per cent per centby abs. cent cent ff AW H IIegh 104 B t u 104 B t u 104 B t u9.55 3754 25.43 63.50 222.2 0.795 24.69 0 4 11 111.14 4155 39.89 79.81 220.0 0 896 27.33 0 4 12.412.44 4378 56.16 89.16 218.4 1.261 30.81 0 6 14.114.10 4435 80.69 96.47 216.6 1.736 44.82 0 8 20.715.27 4384 92.49 98.75 215.5 2.371 50.90 1.1 23 617.52 4115 98.29 99.70 215.4 2.626 56.55 1.2 26.3

and heat loss, the latter quantity may also be determined. When anobserved temperature has been substituted in the equilibrium equa-tions 12) and 13), the corresponding values of x and y may becalculated, and h found from Equation 16) after substituting thesevalues of x y and AW Table 4 contains the results of such calcula-tions for six different air-fuel ratios. The point in the cycle at whichmaximum temperature occurs is somewhat indefinite, as may be seenfrom Fig. 14, in which is shown the crank-angle at maximum tempera-ture plotted against air-fuel ratio for a large number of tests. Thepositions of maximum temperature, used in calculating AW werefound from the curve in this figure. VW and h expressed as percent-ages of the heat of combustion H were plotted against air-fuelratios, as shown in Fig. 15. It is evident from this graph that thework of expansion AW is far too small to account for much of thedifference between calculated and observed maximum temperatures.Moreover, the heat loss would have to attain unreasonable values ifit alone were to account for the temperature difference For, as maybe seen in Fig. 15, this would require losses of from 11 per cent atrich mixtures up to 26 per cent at lean mixtures; while experimentalevidence would apparently indicate that the heat loss during combus-tion up to the point of maximum temperature should not exceed 5per cent.*

b) Effect of NaOH SolutionThe amount of NaOH solution introduced into the engine cylinderwas so small that no very appreciable lowering of the temperature

could be expected unless it had some catalytic effect on combustion.Jones, Lewis, Friauf, and Perrott have determined the effect of NaCl

* Some Properties of the Working Fluid of Gas Engines, W. T David. Engineering,Vol. 113, p. 281. 1922)


44/58


8

2

/

*&

A/r Fue/ Pea ioFIG. 14. CRANK NGLE T MAXIMUM COMBUSTION TEMPERATURE

solution on the temperature of a stationary flame in connection withtheir investigation of flame temperatures by line reversal. Frommeasurements of the brightness temperature of a platinum ribbonheated in the flame, they estimated that with a flame temperature of3875 deg. F. abs., the introduction of enough solution to give definitereversal reduces the flame temperature about 27 deg. F.

c) Effect of humidityGoodenough and Felbeck have found that water vapor added to

the combustible mixture lowers the calculated flame temperatures.tHowever, this effect becomes appreciable only when the amount ofwater added is greater than 10 per cent of the fuel burned, the de-crease in temperature for this particular mixture being about 50 deg.F. From the humidity data included in Table 3 it may be seen that

Loc. cit., p. 8.tLoc. cit., p. 34 .


45/58


46/58


was open. Since the opening was about 18 deg. and readings weretaken at 5 deg. intervals the effect would be that of a running averagein which the temperature at any given crank position would be included in at least four observations. In spite of this it might be pos-sible to obtain observed maximum temperatures considerably lowerthan the actual temperatures due to this averaging process if the timeduring which the latter persisted were extremely short. Decreasing theshutter opening to 3 deg. of crank travel showed no temperatureincrease so that it may safely be concluded that the finite shutteropening would not account for the difference between observed andcalculated maximum temperatures.

e) Effect of fogging of the cylinder w n owAny deposit which accumulated on the window between the time

when observations were being made and that when the lamp was cali-brated would tend to lower the brightness of the lamp radiation in-side the engine cylinder. Hence the curve drawn for an actual cali-bration would be lower than one drawn for a calibration made simul-taneously with the flame temperature observations. However if thewindow deposit were increasing very rapidly during the time theseobservations were being made the apparent temperatures as read fromthe actual calibration curve should be consistently higher for thesecond traverse than for the first. That this is not the case, may beseen from the curves in Fig. 11, where the circles are used to indi-cate the first set of readings and the filled circles the second set.Therefore the fogging of the cylinder window does not furnish asatisfactory explanation of the difference between observed and calcu-lated temperatures.

f) Effect of precision of the line reversal methodSeveral investigators have measured stationary flame tempera-

tures both by the line-reversal method and by various other inde-pendent methods and have found the results to be in satisfactoryagreement. One such method consists of heating a platinum stripelectrically in a vacuum and determining the relation between thebrightness temperature and the heating current. The electrically-heated strip is then placed in a flame and a similar relation betweentemperature and current is found. When these two temperature-current curves are plotted they will intersect and from the tempera-ture at the point of intersection the flame temperature may be de-termined. For at this point the strip is dissipating the same amount ofenergy whether it is in a vacuum or in the flame. When it is in the


47/58

FL ME TEMPER TURES IN AN INTERNAL COMBUSTION ENGINE 43former the loss is by radiation alone and, since its temperature isunchanged, this must also be the case in the flame. If there is no direct loss to the flame it and the strip must be in thermal equilibriumand the flame temperature will then be equal to the true temperatureof the strip. Loomis and Perrott,* using this method, found the tem-perature of a natural gas-air flame to be 3672 deg. F. abs. or about36 deg. F. above the corresponding line-reversal temperature. Griffithsand Awberry,t employing the same method, found temperatures from11 to 5 deg. F. above the line-reversal temperatures, depending uponthe size of the platinum .wire which was used. These results certainlyindicate that temperatures found by the line-reversal method differfrom the actual flame temperatures by amounts which are well withinthe limits of experimental error.

g) Effect of chemical equilibrium conditionsIn discussing the influence of heat loss and work of expansion

under a), it was pointed out that the calculated temperatures are forconstant volume adiabatic combustion. Departure from these assumedconditions affects not only the energy equation as already explained,but also the equilibrium equation 8). Since the actual combustionoccurs while both pressure and volume are changing, the correct formof this latter equation cannot be readily determined. As a first ap-proximation, it is satisfactory to assume constant volume combustion,for the volume change is much smaller than the pressure change, buta more exact analysis should also include the effect of the changingvolume during combustion. In this respect the analysis of the spark-ignition cycle would appear to be quite similar to that of the mixedcycle for compression-ignition engines.

Relative volume changes for combustion with various air-fuelratios may be estimated from the curve in Fig. 14, which shows theangular position of the crankshaft at the time that maximum tem-perature occurs for different air-fuel ratios. From this curve it isevident that a greater volume change takes place with leaner mixturesthan with rich ones, the variation being practically linear for ratiosgreater than 13 to 1. Since the greatest difference between calculatedand observed temperatures was found for the leanest mixture, thiswould seem to indicate, qualitatively at least, that the departure fromconstant volume combustion is an important factor in explaining thedifference between calculated and measured temperatures. For, as-suming that each of the factors discussed under a), b), c), and f)

*Loc. cit., p 8.tLoc. cit., p 8.


48/58


49/58

FL ME TEMPER TURES N N INTERN L COMBUSTION ENGINE 45gas, variable specific heat and dissociation but neglecting heat loss.The effect of spark timing on the flame temperature has also been de-termined for a limited range of operating conditions.

18. Conclusions. The results of the investigation indicate that thefollowing conclusions are justified:

1) The flame temperature found from a line-reversal measure-ment is characteristic of thermal equilibrium established throughoutthe gases in the engine cylinder early in the process of burning andmaintained during the subsequent expansion.

2) The maximum flame temperatures for a compression ratio of3.86 to 1 have been found to vary from 3750 deg. F. abs., with therichest and leanest air-fuel mixtures to 4450 deg. F. abs. with an air-fuel ratio of 13.9 to 1.

3) The maximum observed flame temperatures and the cor-responding calculated temperatures were found to be in closest agree-ment for the normal operating range of air-fuel ratios between 12 to 1and 14 to 1. Throughout this range the calculated values are ap-proximately 600 degrees higher than the observed values. With eitherricher or leaner mixtures the difference increases and reaches a maxi-mum of 1000 degrees at the lean combustion limit.4) Since some of the discrepancy between the calculated and theobserved maximum temperatures is undoubtedly due to experimentalerrors the most likely of these have been considered and their upperlimits estimated. If the observed values were increased by theseamounts they would still be from 200 to 600 deg. F. below the cal-culated temperatures. It would therefore seem probable that thisremaining difference is due to an inadequate analysis of the actualcombustion process.

5) The concentrations of the gases in an engine cylinder maybe calculated when the temperature and pressure are known hencethe measurement of temperature independent of the other thermody-namic variables makes it possible to study the progress of the chemi-cal reactions and equilibrium at high temperatures.


50/58

VIII. APPENDIXDEFINITIONS ND NOMENCLATURE

The following definitions are for the terms and symbols usedthroughout this bulletin.

1) T True Temperature, is the temperature of a body as indi-cated by a suitable thermometer in contact and in equilibrium with it.

2) T Radiation Temperature of an imperfect radiator is thetrue temperature of a perfect radiator which is emitting the sametotal radiant energy, hence

Tr e

For a perfect radiator e = 1 and Tr = T 3) S, Brightness Temperature of an imperfect radiator is the

true temperature of a perfect radiator which is emitting the sameradiant energy in the spectral range X to X dX In specifying abrightness temperature the wave-length X must also be specified.

4) E Total Emissive Power, is the total radiant energy emittedin unit time by unit area of radiating surface into a solid angle 2 ir.

5) Ex Monochromatic Emissive Power, is the emissive powerin the spectral range X to X + d

E = Ex dX

6) J Total Intensity is the total radiant energy emitted inunit time by unit area of radiating surface into unit solid angle normalto the surface, then

E rJ7) Jx Monochromatic Intensity is the intensity in the spectral

range X to X dXE x

8) E,Total Emissivity, is the ratio of the total emissive power ofan imperfect radiator to that of a perfect radiator at the same truetemperature.

9) cx Monochromatic Emissivity, is the emissivity in the spec-tral range X to X dX The total and monochromatic emissivitiesmay also be defined as the ratios of the corresponding intensities.


51/58

FLAME TEMPERATURES N N INTERNAL COMBUSTION ENGINE 4710) a, Total Absorptivity, is the ratio of the total absorbedradiant energy to the total incident radiant energy. From Kirch-hoff s Law

Eawhen E is the total emissive power of an imperfect radiation, a thetotal absorptivity for the same, and E the total emissive power for aperfect radiator at the same temperature as the imperfect radiator.

11) ax Monochromatic Absorptivity, is the absorptivity in thespectral range X to X dX.12) r, Total Reflectivity, is the ratio of the total reflected radiantenergy to the total incident radiant energy.13) rx Monochromatic Reflectivity, is the reflectivity in thespectral range Xto X + dThe following terms occur in the discussion of the combustionreactions.14) n, Number of mols or pound-molecular weights of a singleelement or compound. The subscripts indicate the substance andstate. Thus 1) indicates mols of C02 with complete combustion,

2) mols of H20 with complete combustion, (e ) mols of 02 atequilibrium.15) N, Total number of mols in a mixture of elements and com-pounds. The subscripts indicate the state. Thus i) indicates thebeginning of compression and

i N N.where N 1 is the number of mols of entering charge of fuel and air,and Nc is the number of mols of residual exhaust gas in the clearancespace.

16) x Percentage of CO burned at equilibrium.17) y Percentage of H2 burned at equilibrium.18) C, Molecular Concentration of a substance in a mixture. In

nlgeneral C1 = --NFor a gaseous mixture

Cz - P


52/58


where pi is the partial pressure of component 1) and P is the totalpressure.

19) Kp Equilibrium Constant. If the chemical reaction equa-tion is written

riA V A + vsA3 = 0where A1 A2 etc. are the molecular formulae of the gases and vV2 etc. are the number of molecules of each taking part in thereaction v being positive when the substance is formed and negativewhen it is consumed then

K, = C C2 Cfrom the Law of Mass Action. The subscripts indicate the reactioninvolved. Thus CO) indicates the reaction

-CO YO CO 2 0H2) indicates -H 2 - Y2 + H20 =

and w.g.) indicates-H2 - CO 2 + H20 + CO = 0

the water-gas reaction.20) H Heat of Combustion is the lower heat of combustion at

constant volume. The subscripts indicate the fuel. Thus CO)indicates carbon monoxide H 2) indicates hydrogen and m) indicatesa fuel mixture.

21) U, Intrinsic Energy of a mixture. The subscripts indicatethe state. Thus 1) indicates the beginning of compression 2) indi-cates the end of compression and 3) indicates the end of constantvolume combustion.


53/58

RECENT PUBLICATIONS OFTHE ENGINEERING EXPERIMENT STATION

Bulletin No. 211 The Torsional Effect of Transverse Bending Loads on ChannelBeams, by Fred B Seely, William J Putnam, and William L. Schwalbe. 1930Thirty five cents.Bulletin No. 212. Stresses Due to the Pressure of One Elastic Solid uponAnother, by Howard R. Thomas and Victor A. Hoersch. 1930 Thirty cents.Bulletin No. 213. Combustion Tests with Illinois Coals, by Alonzo P. Kratzand Wilbur J. Woodruff. 1930 Thirty cents.Bulletin No. 214. The Effect of Furnace Gases on the Quality of Enamels forSheet Steel, by Andrew I Andrews and Emanuel A. Hertzell. 1930 Twenty cents.Bulletin No. 215. The Column Analogy, by Hardy Cross. 1930 Forty cents.Bulletin No. 216. Embrittlement in Boilers, by Frederick G. Straub. 1930Reprinted November, 1933 Eighty five cents.Bulletin No. 217. Washability Tests of Illinois Coals, by Alfred C Callen andDavid R. Mitchell. 1930 Sixty cents.Bulletin No. 218. The Friability of Illinois Coals, by Cloyde M. Smith. 1930Fifteen cents.Bulletin No. 219. Treatment of Water for Ice Manufacture, by Dana Burks, Jr.1930 Sixty cents.Bulletin No. 220. Tests of a Mikado-Type Locomotive Equipped with Nichol-son Thermic Syphons, by Edward C. Schmidt, Everett G. Young, and Herman J.Schrader. 1930 Fifty five cents.Bulletin No. 221 An Investigation of Core Oils, by Carl H. Casberg and Carl E.Schubert. 1931 Fifteen cents.Bulletin No. 222. Flow of Liquids in Pipes of Circular and Annular Cross-Sections, by Alonzo P. Kratz, Horace J. Macintire, and Richard E. Gould. 1931Fifteen cents.Bulletin No. 223. Investigation of Various Factors Affecting the Heating ofRooms with Direct Steam Radiators, by Arthur C. Willard, Alonzo P. Kratz, MauriceK. Fahnestock, and Seichi Konzo. 1931 Fifty five cents.Bulletin No. 224. The Effect of Smelter Atmospheres on the Quality of Enamelsfor Sheet Steel, by Andrew I. Andrews and Emanuel A. Hertzell. 1931 Ten cents.Bulletin No. 225. The Microstructure of Some Porcelain Glazes, by Clyde L.Thompson. 1931 Fifteen cents.Bulletin No. 226. Laboratory Tests of Reinforced Concrete Arches with Decks,by Wilbur M. Wilson. 1931 Fifty cents.Bulletin No. 227. The Effect of Smelter Atmospheres on the Quality of DryProcess Enamels for Cast Iron, by A. I. Andrews and H. W. Alexander. 1931Ten cents.CircularNo. 21 Tests of Welds, by Wilbur M. Wilson. 1931 Twenty cents.Bulletin No. 228. The Corrosion of Power Plant Equipment by Flue Gases,by Henry Fraser Johnstone. 1931 Sixty five cents.Bulletin No. 229. The Effect of Thermal Shock on Clay Bodies, by William R.Morgan. 1931 Twenty cents.Bulletin No. 230. Humidification for Residences, by Alonzo P. Kratz. 1931Twenty cents.Bulletin No. 231 Accidents from Hand and Mechanical Loading in Some IllinoisCoal Mines, by Alfred C. Callen and Cloyde M. Smith. 1931 Twenty five cents.Bulletin No. 232. Run-Off Investigations in Central Illinois, by George W.Pickels. 1931 Seventy cents.Bulletin No. 233. An Investigation of the Properties of Feldspars, by Cullen W.Parmelee and Thomas N. McVay. 1931 Thirty cents.Bulletin No. 234. Movement of Piers during the Construction of Multiple-SpanReinforced Concrete Arch Bridges, by Wilbur M. Wilson. 1931 Twenty cents.Reprint No. 1 Steam Condensation an Inverse Index of Heating Effect, byAlonzo P. Kratz and Maurice K. Fahnestock. 1931 Ten cents.

Copies of the complete list of publications can be obtained without charge by addressing tEngineering Experiment Station, Urbana, Ill


54/58


Bulletin No. 235. An Investigation of the Suitability of Soy Bean Oil for CoreOil by Carl H. Casberg and Carl E. Schubert. 1931. Fifteen cents.Bulletin No. 236. The Electrolytic Reduction of Ketones, by Sherlock Swann,Jr. 1931. Ten cents.Bulletin No. 237. Tests of Plain and Reinforced Concrete Made with HayditeAggregates, by Frank E. Richart and Vernon P. Jensen. 1931. Forty-five cents.Bulletin No. 238. The Catalytic Partial Oxidation of Ethyl Alcohol, by DonaldB. Keyes and Robert D. Snow. 1931. Twenty cents.Bulletin No. 239. Tests of Joints in Wide Plates, by Wilbur M. Wilson, JamesMather, and Charles 0. Harris. 1931. Forty cents. BulletinNo. 240. The Flow of Air through Circular Orifices in Thin Plates, byJoseph A. Polson and Joseph G. Lowther. 1932. Twenty-five cents. BulletinNo. 241 Strength of Light I Beams, by Milo S. Ketchum and Jasper0. Draffin. 1932. Twenty-five cents. BulletinNo. 242. Bearing Value of Pivots for Scales, by Wilbur M. Wilson,Roy L. Moore, and Frank P. Thomas. 1932. Thirty cents. BulletinNo. 243. The Creep of Lead and Lead Alloys Used for Cable Sheathing,by Herbert F. Moore and Norville J. Alleman. 1932. Fifteen cents. BulletinNo. 244. A Study of Stresses in Car Axles under Service Conditions,by Herbert F. Moore, Nereus H. Roy, and Bernard B. Betty. 1932. Forty cents. BulletinNo. 245. Determination of Stress Concentration in Screw Threads bythe Photo-Elastic Method, by Stanley G. Hall. 1932. Ten cents. BulletinNo. 246. Investigation of Warm-Air Furnaces and Heating Systems,Part V, by Arthur C. Willard, Alonzo P. Kratz, and Seichi Konzo. 1932. Eightycents Bulletin No. 247. An Experimental Investigation of the Friction of ScrewThreads, by Clarence W. Ham and David G. Ryan. 1932. Thirty-five cents. BulletinNo. 248. A Study of a Group of Typical Spinels, by Cullen W. Parmelee,Alfred E. Badger, and George A. Ballam. 1932. Thirty cents. BulletinNo. 249. The Effects on Mine Ventilation of Shaft-Bottom Vanes andImprovements in Air Courses, by Cloyde M. Smith. 1932. Twenty-five cents. BulletinNo. 250. A Test of the Durability of Signal-Relay Contacts, by EverettE. King. 1932. Ten cents. BulletinNo. 251 Strength and Stability of Concrete Masonry Walls, by FrankE. Richart, Robert B. B. Moorman, and Paul M. Woodworth. 1932. Twenty cents. BulletinNo. 252. The Catalytic Partial Oxidation of Ethyl Alcohol in the VaporPhase. The Use of a Liquid Salt Bath for Temperature Control, by Donald B.Keyes and William Lawrence Faith. 1932. Ten cents.Bulletin No. 253. Treatment of Water for Ice Manufacture, Part II by DanaBurks, Jr. 1933. Forty-five cents.Bulletin No. 254. The Production of Manufactured Ice at Low Brine Temper-

ature, by Dana Burks, Jr. 1933. Seventy cents. BulletinNo. 255. The Strength of Thin Cylindrical Shells as Columns, byWilbur M. Wilson and Nathan M. Newark. 1933. Fifty cents. BulletinNo. 256. A Study of the Locomotive Front End, Including Tests of aFront-End Model, by Everett G. Young. 1933. One dollar. BulletinNo. 257. The Friction of Railway Brake Shoes, Its Variation withSpeed, Shoe Pressure and Wheel Material, by Edward C. Schmidt and Herman J.Schrader. 1933. One dollar. Bulletin No. 258. The Possible Production of Low Ash and Sulphur Coal inIllinois as Shown by Float-and-Sink Tests, by D. R. Mitchell. 1933. Fifty cents. Bulletin No. 259. Oscillations Due to Ionization in Dielectrics and Methods ofTheir Detection and Measurement, by J. Tykocinski Tykociner, Hugh A. Brown,and Ellery Burton Paine. 1933. Sixty-five cents. BulletinNo. 260. Investigation of Cable Ionization Characteristics with Dis-charge Detection Bridge, by Hugh A. Brown, J. Tykocinski Tykociner, and ElleryBurton Paine. 1933. Fifty cents. BulletinNo. 261 The Cause and Prevention of Calcium Sulphate Scale inSteam Boilers, by Frederick G. Straub. 1933. Eighty-five cents. Bulletin No. 262. Flame Temperatures in an Internal Combustion EngineMeasured by Spectral Line Reversal, by Albert E. Hershey and Robert F. Paton.1933. Fifty-five cents.

A limited number of copies of bulletins starred are available for free distribution.


55/58


56/58


57/58

UNIVERSITY OF ILLINOIS

Colleges and Schools t Urbana OLLEGE OF LIBERAL ARTS AND SCIENCES.-General curriculum with majors in thehumanities and sciences; specialized curricula in chemistry and chemical en-gineering; general courses preparatory to the study of law and journalism;pre-professional training in medicine dentistry and pharmacy. OLLEGE OF COMMERCE AND BUSINESS ADMINISTRATION.- Curricula in generalbusiness trade and civic secretarial service banking and finance insuranceaccountancy transportation commercial teaching foreign commerce indus-trial administration public utilities and commerce and law.COLLEGE OF ENGINEERING.-Curricula in ceramics ceramic engineering chemicalengineering civil engineering electrical engineering engineering physics gasengineering general engineering mechanical engineering metallurgical engi-neering mining engineering and railway engineering.COLLEGE OF AGRICULTURE. Curricula in agriculture floriculture general homeeconomics and nutrition and dietetics. OLLEGE OF EDUCATION. Curricula in education agricultural education homeeconomics education and industrial education. The University High Schoolis the practice school of the College of Education.COLLEGE OF FINE AND APPLIED ARTs.-Curricula in architecture landscape archi-

tecture music and painting.COLLEGE OF LAW.-Professional curriculum in law.SCHOOL OF JOURNALISM.-General and special curricula in journalism.SCHOOL OF PHYSICAL EDUCATION.-Curricula in physical education for men andfor women.LIBRARY SCHOOL.-Curriculum in library science.GRADUATE SCHOOL.-Advanced study and research.

Colleges in ChicagoCOLLEGE OF MEDICINE.-Professional curriculum in medicine. OLLEGEOF DENTISTRY.-Professional curriculum in dentistry.COLLEGE OF PHARMACY.-Professional curriculum in pharmacy.

Experiment Stations and Scientific Bureaus t UrbanaAGRICULTURAL EXPERIMENT STATION STATE GEOLOGICAL SURVEYENGINEERING EXPERIMENT STATION STATE NATURAL HISTORY SURVEYBUREAU OF BUSINESS RESEARCH STATE WATER SURVEYBUREAU OF EDUCATIONAL RESEARCH

For general catalog of the University special circularsand other information addressTHE REGISTRAR UNIVERSITY OF ILLINOIS

URBANA ILLINOIS


58/58

Documents

Engineering Experiment