T H E UN I V E R S I T Y O F M I C H: G A N COLLEGE OF ENGINEERING Department of Mechanical Engineering The Heat Transfer and Thermodynamics Laboratory Technical Report No0 3 A STUDY OF POOL BOILING IN AN ACCEL4ERATING SYSTEM H. Merte, Jr. J.o A, Clark.UMRT' Pr6oj-ec.t 6I6 under contract with~ DEPARTMENT OF THE ARMY DETROIT ORDNANCE DISTRICT CONTRACT NO. DA 20-018-ORD-15316 DETROIT, MICHIGAN administered by: THE UNIVERSITY OF MICHIGAN RESEARCH INSTITUTE ANN ARBOR November 1959

TABLE OF CONTENTS Page LIST OF TABLES v LIST OF FIGURES vii NOMENCLATURE xi ABSTRACT xiii I, INTRODUCTION 1 A. Purpose 1 B. The Boiling Phenomena 1 C. Acceleration of a Boiling System 4 Do Existing Relationships in the Literature Showing an Effect of Acceleration 7 E. Previous Experimental Work in Boiling Heat Transfer Embodying an Alteration of the Gravitational Field 10 II. EXPERIMENTAL APPARATUS AND INSTRUMENTATION 12 A. Configuration 12 B. Construction 12 C. Instrumentation 32 III. ESTIMATION OF ERRORS 34 A. Temperature 34 B. Rotation of the Main Shaft 49 C, Acceleration at the Heating Surface 49 Do Pressure on the Heating Surface 53 E o Heat Flux 57 Fo Calculation of Heat Flux Rate from Temperature Gradient in Heater Block 59 G. Measurement of Barometric Pressure 62 H. Measurement of Specific Resistivity of Water in the Test Vessel 62 IV. TEST PROCEDURES 63 A. Fluid 63 B. Heat Flux Range Covered 64 C. Heater Surface Treatment 64 Do Range of Accelerations Covered 65 E. Location of Water Temnperature Thermocouples 66 iii

TABLE OF CONTENTS (Continued) Page F. Variation of Cooling Water 66 Go Attainment of Steady State Conditions 68 H. Criteria for Acceptable Data 69 V. TEST RESULTS 70 A, Natural Convection 70 B. Boiling 74 C. Overall Results 101 VI. ANALYSIS 117 A. General 117 B. Bubble Relationships in Boiling 118 C. Some Observations on Boiling at Large Values of Heat Flux 123 D. The Influence of Acceleration on Boiling Area 129 E. The Influence of Acceleration on the Number' of Nucleating Sites 138 F. Concluding Remarks 142 APPENDICES A. DERIVATION OF EQUATION FOR ERROR IN MEASURING WATER TEMPERATURES 143 B. DERIVATION OF EQUATION FOR HEAT LOSS BY CONDUCTION THROUGH HEATER SKIRT 147 C. APPROXIMATION OF VELOCITY OF FLUID AT HEATING SURFACE DUE TO CORIOLIS' FORCE 153 D. DATA 155 BIBLIOGRAPHY 167 iv

LIST OF TABLES Table Page I List of Mercury Thermocouple Slip-Ring Channels and Radius from Centerline of Rotation................ 36 II List of Mercury Channel Pairs Associated with Each Thermocouple Circuit.0 e..*s *e oo* e.* o,eoo 37 III Values of Uncertainty W(Tw -T1) Calculated from Typical Test Data.. a....... a *....... * 0 7. 44 IV Values of Parameters in Test Vessel Which Vary with Rotational Speed.520...........*......*.......n.. 52 V Loss by Conduction Through Skirt for Convection Tests...,. 61 VI Sample Sequence of Accelerations During Test Runs to Determine Effect of Past History of Acceleration.......... 66 VII Comparison of Heat Flux Calculated from Heater Temperature Gradient with that Calculated from Power Measurement, 111 V.

LTST OF FIGURES Fi-gure Page 1 ~Orientation of Acceleration with Re.spe~t to'Heating Surface o. Q. I ~ af O a' a a 4 a a a o ~ a O 4 4 ~ c o ~* a a a a 6 o ~ Q ~ Q ~ a 0 13 2 Test Vessel a 4 a 0 4 a 14 3 Centrifuge Assembly0.- * lb. a 4 a s Or a a.. a a a..O a O a a a a a a a O a 0 15 4 Detail of Copper Heater-Skirt Intersection v. oo o.,0 a a. 0 0o 17 5 BHeater Thermocouple Locationso o0 o 0o -a..o.o00, o. 0o 18 Underside View of'Heater Block o o... o * o.... o o 0 o ~ 20 7 Cover Installedd on Heater:Block,.....0..... o,.0o0, 0.*. o oo a 20 ~8 Guard Heater Assembly. 4 b 0 0 0. o 21 9 Assembled Heater Unit.. 00000000a000.0 0 o 00 0 o0 21 10 Assembly.of'Inner Container Walls and Heater., o...,.a. o 22 11 Inner Container Test Assembly B o..o a O ao'oo o.ooooo a 22 12 Test Vessel Installed in Centrifuge',,~aon*oo 24 13 Construction of Heater Thermocouple Probeooo..ooo..o.. 4..- 24 14 Equivalent Thermocouple Circuits..E0 0b0000 0 00 27 15 Sketch.of Mercury Thermocouple Slip-Ring Assembly~.o..0..o0 28 16 View of Mercury Thermocouple Slip-Ring Assembly,.0a0,0o,0o0. 31 17 Overall View of Test Apparatus,. o o...... 00.oo. 0 la... 31 18 Typical Plot Indicating -Effect of. Rotation Upon Thermocouple Readings'.. o 4 a'0 0 0 0 a a o O 0 a a P a 0t a a a a'. 38 19 Corrections Applied to Thermocouples to Compensate for Rotation0 0000 0 4o 2Rotation.a 4 a a O 0 a b a O a a e a a a. a a Al a a 42 a 20. ILocation of Heater Thermocouple Holes,,,v,o b o,o,. *o4*., 000o.. 42 vii

LIST OF FIGURES (CONT D) Figure Page 21 Typical Conditions for Thermocouple T with System Subjected to Acceleration 21.15 Times6normal Gravity........ 46 22 Test Vessel Dimensions for Calculation of Acceleration at the Heating Surface.,....,..........,........, o. 51 23 Relation of Heater and Liquid Surfaces Due to Rotationo..,.. 55 24 Estimated Heat Loss by Conduction Through the Heater Skirt for Various Total Heat Fluxes.....,................. 60 25 Various Locations of Thermocouple T6 for Different Test Runs*.a..............0........O....O.a.............#. 67 26 Plot of ATc Versus Acceleration for Natural Convection Indicating the Effect of the Flow Guide...,..........,.. 71 27 Run No. C-5. Temperature Data for Convection as Taken..... 72 28 Correlation of Natural Convection with Acceleration Normal to Heating Surface.................................. 73 29 Run No. B-15. Plot of Tw-Tsat and Tw-T5 Versus Water Temperature T.....................*.*.....*. *...*..s0 75 30 Run No. B-15. Plot of Tw-Tsat and AT in Heater Block Versus Acceleration with Pool Boiling....,............... 82 31 Run No. B-15. Difference in Water Temperatures for Various Accelerations and Subcooling.......,...-,,,..... 83 32 Run No. B-15. Temperature Profile Between Heater and Water Surface for Various Accelerations.........,..,.... 84 33 Run No. B-9. Tw-Tsat and Tw-T5 vs. Water Temperature T. at Various Accelerations.................... 85 34 Run No. B-9,. Temperature Profile Between Heater and Water Surface for Various Accelerations..................... 89 35 Run No, B-9. Difference in Water Temperature for Various Accelerations and Subcooling........................ 91 36 Run No. B-9. Plot of Tw-Tsat and AT in Heater Block vs. Acceleration with Pool Boiling.....,..,................. 92 viii

LIST OF FIGURES (CONT'D) Figure Page 37 Run No, B-14D Plot of TW T a and Water Temperatures vs0. Acceleration with Pool ling o 94 38 -Run No. B-14. Plot of fT in Heater Block vso Accelerationo, 95 39 Run No, B-140 Temperature Profile Between Heater and Water Surface for Various Accelerations00.o 0 96 40 Run No. B-22o Plot of T'-T and AT in Heater Block vs_ W. gat Acceleration with Pool Bo ling.00a...,oo0oOa,,a,,a0 o,00,0*. 98 41 Plot of' Water Temperature vs0 Acceleration for Two Runs Identical Except for Location of Thermocouple T6o o.oooooo, 99 42 Run No, B-22, Temperature Profile Between Heater and Water Surface for Various Accelerations b o o.0 o o o Qo 0aoo 100 43 Run No, B-21. Plot of Tw-Tsat and Water Temperature vs. Acceleration with Pool Boilinga o.o.ooo a a Oo0o0oo.oo 102 44 Run No, B-21o Plot of AT in. Heater Block vso Accelerationo0 103.4.5 Run No. B-21 Temperature Profile Between Heater and Water Surface for Various Acceleration o's a a., o o ao. o ooooo 104 46 Plot of q/A vso Tw-Tsat for Boiling in Standard Gravitation. al Field 1 o0 o0 0 06 47 Influence of Acceleration on Tw.TTsat with Pool Boiling to Saturated Water 0' 1 07 48 Plot of Tw-T vs0 Acceleration.with Pool Boiling tto Saturated Waler o o o e 0, o o a o a 0 o 0 o o o0 o 0 0109 49 Sub cooling at Thermocouple T5 as a Function of Heat; Flux and Acceleration. o 0 0 0 113 50 Example of Coriolis Acceleration on a Particle Constrained to Move Radially on a~Rotating Bar.... a loao o0 ooa00*aa00aa0,00l 113 51 Correlation of Natural Convection with Flow Guide ooo. oooo0 116 52 Representation of Hirano and Nishikawa (49) for Boiling Heat Tran.sfer 0 0 0 0 o o o o o o0o 0 o a o o o 0 0. o 0 a o.. 0 0 o o 0o0 Q 000a o Q 11.9 1X:~~~~~~~~~~~~a

LIST OF FIGURES (CONT'D) Figure Page 53 Interdependence and Complexity of Boiling Elements..,,,,..,. 119 54 Thickness of Boundary Layer on a Horizontal Heating Surface as Measured by Hirano and Nishikawa(54) Using Refraction Method......Is.........,.. 4..a....*.......... 122 55 Effect of Heat Flux on Fluid Temperature Near Heating Surface vith Forced Convection Boiling, Due to TreschovO 55)i * s r 0 W 0 v * r......,..a. V v * 0. i t * r * e v+-. a I 122 56 Representative Plots of q/A vs, A9 Near Peak Heat Flux...... 125 57 Illustration of Area of Influence of Bubble with Boiling Heat Transfer,...........,.... *.**...........*,, 131 58 Convective Fluid Flow Pattern in the Experimental System with and without Flow Guide........................... 131 59 Effect of Acceleration on Peak Heat Flux with Pool Boiling.. 135 60o Calculated Values of y as a Function of Total Heat Flux and Acceleration,.................................. 136 61 Cross Plot of Figure 60..................1...............1 62 View of Heater Surface After Test Run with Slightly Contaminated Water at Flux q/A = 99,500 Btu/hr-ft........ 139 63 Number of Nucleating Sites as a Function of Heat Flux and Acceleration......................... 141 64 Schematic of Thermocouple Tube Shown in Figure 21....... 144 65 Extended Surface Representing the Heater Skirt,.............. 148 66 Model Used to Calculate Maximum Possible Water Velocity Due to Coriolis Acceleration in Test Vessel...................... 154 x

NOMENCLATJRE Other nomenclature is defined locally as necessary a -Acceleration normal to heating surface a/g Dimensionless acceleration A Area C Constants Cp Specific heat D Diameter DNB Departure from Nucleate Boiling,.defined in Chapter VI-C f Frequency of bubble formation g Local gravitational acceleration gc Mass-force conversion constant = 320174 lbm/lbf ft/sec2 h Convective heat transfer.coefficient Gr Grashof number Nu Nusselt number Pr Prandtl number N/A Active nucleating sites per unit area q/A Heat flux rate T Temperature Tn Temperature at thermocouple n T, (m-n.) Temperature difference = Tn- Th W Uncertainty c Thermal diffusivityr 3 ~Volumetric coefficient of thermal expansion xi

Y Defined by.Equation 66 Boundary layer thickness Surface tension X Latent heat of vaporization Defined by Equation 63 nA Tw - Tsat Contact angle Absolute viscosity v Kinematic viscosity Subscripts b Bubble c Convection f Film 1 Liquid Sat. Saturation t Total v Vapor w Wall xii

ABSTRACT A study is made of the influence of system acceleration (1 to 21 gCs) on pool boiling heat transfer in saturated distilled water at approximately atmosph4eric pressure, A flat electrically heated chromium plated copper disc serves as the heat transfer area, with a thin stainless steel skirt attached'to the periphery of the disc to provide a continuous surface. The water depth is nmaintained constant at 2-1/2 inches, and a cooling coil on the underside of the cover condenses the vapor formed. Temperatures in the heating disc and water are measured with an uncertainty of + 0,1~Fo Acceleration is attained by use of the centrifuge principle. The boiling system is pivoted from the cross arm on a vertical shaft such that the acceleration is always normal to the heating surface. The magnitude of the acceleration is varied. from that due to one standard gravity up to 21 times gravitational accelerat ion. Heat flux rate is varied from approximately 5,000 to 100,000 Btu/hr-ftt Several tests were conducted with non-boiling convective heat'transfera The data agreed quite well with the correlation~ Nu = 0.14 (GrPr) l/3 With boiling at heat flux values up to 50,000 Btu/hr-ft2, it was found that a small degree of subcoolsing sign.ificantly influenced the results w;ith the system under acceleration. As a means for obtaining a theoretical understanding of the process of'boiling under the influence of high acceleration including the simultaneous effect of natural convection, a concept of the "Area of Influence" of the'bubbles is defined and values calculated:for the various accelerations and heat flux rates, The change in Tw - Tsat with heat flux and acceleration is used'to calculate the influence of heat flux on the number of active nucleating si+es. X~~~~~~~~~~~~~~~ ue J

I. INTRODUCTION A, Purpose The purpose of this study is to investigate the effect of acceleration on a system in which boiling is taking.place from a flat heated surface, Acceleration of the system provides the equivalent to a change in the force field acting.on the system, An understanding,of the role of the force field in the boiling heat transfer process may add to the overall understanding of the boiling. phenomena.and increase its effective application, Bo The Boiling Phenomena The characteristics of the three regimes of boiling are abundantly available in the literature 3) The maximum heat transfer rate attainable with. nucleate boiling is designated as the peak heat flux or "burnout point"', and has been characterized as the condition where the bubble population on the heating surface is so great that (4-6) the bubbles interfer with one another( A further increase in temperature difference between the liquid and the heating surface then results in a decrease of heat transfer rate because of this interference, until a minimum point is reached, after which stable film boiling is present. The higher heat transfer rates associated with nucleate boiling over that of convection have been ascribed to the action of the bubbles as agitators in the "laminar sub-layer" rather than as media of energy transport'7') The wide variation of temperature -1.

difference observed on an electrically heated surface with patchwise (9) boiling would tend to confirm this. The mechanism of nucleate boiling has been considered in the following stages: 1. The presence or formation of a nucleus from which growth of a bubble can take place. It has been shown that because of surface tension effects(2' ) a minimum value of superheat in the liquid is necessary in order for a bubble nucleus of finite dimensions to form and grow. For a given superheat, the minimum size of the nucleus which is unstable and hence will grow has been termed the critical size.(l1) The smaller the surface tension, the lower is the superheat required for the formation of such a nucleus of given size (i.e. with a lower superheat the nuclei are more apt to form for a given superheat). The roughness of the heating surface also affects the ease with which nuclei may be formed.(9) At lower pressures surface boiling is initiated at a lower surface (27) temperature with the presence of a dissolved gas, which provides a source of nuclei. 2. The early and mid growth period of the bubble. Consideration of the hydrodynamic,....heat transfer and surface tension effects on the growing nucleus(113) has indicated that owing to the surface tension the initial growth rate of the bubble from the critical size is small. Once the bubble has reached a size such...that the surface tension effect is no longer important, the growth rate becomes very large. This is roughly that size which is discernable

to the eye, In this stage the rate of growth is governed by the rate of heat transfer across a thin liquid film surrounding the bubble, and hence is dependent upon the degree of superheat of the.liquid, 3. The late growth and departure period of the bubble. In this period the growth of the bubble has "consumed" the liquid superheat in the immediate vicinity of the bubble and the bubble ceases' growing, Depending upon whether the bulk liquid is subcooled or saturated, the bubble may then either collapse immediately, depart: from the heating surface and then collapse, or depart and rise, due to dynamic and buoyant forces, It has been demonstrated in nucleate boiling that bubbles form at preferential active points on the heating surface (2'9) These active points are postulated to be due to minute cavities existing in the surface which trap vapor from an earlier bubble which has departed(, or simply the remnants of an earlier bubble(1), which then serve as nuclei for further growth. A theoretical study of nucleation(21) in the presence.of normal gravity has lead to the conclusion that nucleation always occurs at boundaries of gas or vapor trapped in surface (38) cavities, A microscopic photographic study of boiling confirmed.this, Experiments in which the effect of the entrapped vapor had been minimized or eliminated enabled extremely high values of superheat to be (22-24) d43) attained. However, it has also been theoretically demonstrated that even if no cavities exist on the heating surface, nucleation will take place preferentially at the surface if any finite degree of

-4wettability exists between the fluid and the surface. Increasing the surface temperature results in more active nuclei and an increase in agitation, and hence an increase in heat flux until the peak heat flux is reached. The literature is not clear as to whether the increase in agitation results solely from an increase in the number of active nuclei or if each existing nucleus also contributes more agitation because of the possible greater liquid superheat in the vicinity of the heating surface. It is likely both factors contribute to the effect. C. Acceleration of a Boiling System Consider the boiling system represented in Figure 1 with a spherical vapor bubble just attached to the heating surface. If the system is accelerated as shown in a direction normal to the heating surface in a standard gravitational field it can be shown that the net buoyant force acting on the bubble will be: / g a FB Vb(Pg v) g +p Vb (1) At the instant the bubble is detached from the surface it can further be shown that the acceleration of the bubble with respect to the heating surface will be: aB - Pv (g + a) (2)

-5If the effect of surface tension is the only force holding the bubble to the surface it is obvious from Equation (1) that a smaller bubble volume will suffice to overcome a given value of the adhering force with the system under acceleration. This effect is inherent in (14) the following expression developed by Fritz for the maximum volume of a bubble at departure from a surface: 3/2 Vb(max)= (0.0119 P5)3 F 2 gc ) V (max) L g(P —P) Equation (3), based on the work of Bashforth and Adams(5) and Wark() considers the equilibrium of the surface of-i.cu~i,,vature.- separating 2 phases in a normal force field of standard gravitational acceleration. Assuming a constant contact angle, an increase in the gravitational acceleration will decrease the maximum volume of the bubble at departure from a heated surface. If the effect of an imposed normal acceleration of the boiling system were to cause the bubbles to be detached prematurely in the sense of Equation (3), then a number of postulations of a general nature might be made at this point. With an electrically heated surface the time averaged heat flux must remain constant, and any decrease in agitation due to the smaller bubble sizes must be compensated for in other ways. The frequency of formation of the bubbles at each active point may increase, If this is insufficient, the number of active sites may increase, which most likely would require an increase in the surface temperature(9) If, on the other hand, the major agitation caused by the bubbles takes place during the early growth period, smaller bubble sizes will either

-6have little effect if the frequency of bubble formation remains unchanged or will act to decrease the time averaged surface temperature if the frequency increases. A complicating feature of the consequence of smaller bubble sizes will be the increasing contribution of natural convection, both because of the increased area available and because of the increased force field resulting from the imposed acceleration. The peak heat flux might also be expected to increase under the action of the acceleration. The agency causing the departure of the bubble from the heater surface will probably be one of the major factors influencing this effect. The effect of buoyancy due to normal gravity has been discounted as an explanation for the departure of a vapor bubble from a heated surface since it has been observed that vapor bubbles may be ejected even from the (18) lower side of a horizontal surface. A plausible cause of departure is (18,60) given which attributes the motion to the inertia of the surrounding liquid. During the rapid growth period of the bubble the surrounding liquid acquires momentum. As the growth rate decreases because of decrease in superheat, the inertia of the liquid causes a reduced pressure field on the upper surface of the bubble tending to draw it from the surface. Photographs upon which this description is in part based showed that during the greater portion of growth the bubble maintains a hemispherical shape, but just prior to departure it becomes elongated perpendicular to the heating surface. The effect of surface tension in nucleate boiling, discussed below affirms this view. It has been observed that the introduction of small quantities of additives which decreased the surface tension decreased the peak heat

-7flux and, for a given temperature difference at less than peak heat flux, increased the heat flux(19') The latter effect seems logical in view of the effect of surface tension on the tendency for nucleation, Such a condition requires less liquid superheat to cause nucleation and hence results in lower bubble growth rates, If the liquid inertia is important in drawing the bubble from the surface then lower growth rates mean slower departures, which in turn mean a greater tendency for the heating surface to become vapor bound. Hence, the peak heat flux would have a lower value, as was observed. A discussion similar to the above is given by Larson,() A recent photographic study of boiling in the absence of (32) gravity, however, seems to invalidate the importance of liquid inertia as a mechanism in detaching the bubble from the heating surface, The vapor remained adjacent to the heating surface, and there was no evidence of bubbles being pushed away from the surface to any appreciable extent during their formation, D. Existing Relationships in the Literature Showing an Effect of Acceleration 1. Several correlations for boiling heat transfer which incorporate a term for the gravitational acceleration exist in the literature, The relationship due to Rohsenow(29), 1/3 = OLq/A gca 17 X C t X g(P-Pr (4) contains an acceleration term as a result of using Fritz' relation for the bubble diameter in a bubble Reynold's Ntumber. Rewriting Equation (4) we

-8have q/A =A,- pa r ly. (rG)5 (5) In extending this correlation to a boiling system undergoing an acceleration, its validity will depend upon the continuing fidelity of the assumptions made. Among these assumptions are: (a) The effect of the contact angle p, which is included in the constant Csf. (b) The relationship between the frequency of bubble formation and the diameter as it leaves an active site, f - Db = constant. (c) The overall heat transfer is proportional to the heat transfer to the bubbles while attached to the surface. 2. Gravitational acceleration appears in the expression for boiling heat transfer derived by Chang(30), given in Equation (6), as a result of the extension of the wave analysis of natural convection to boiling, C1 and n are specified as constants to be determined by experiment. 0b [ 146 1 PC1( _ X P -)ij ](6)/ CT P... 2 This equation is not explicit in q/A, but if P c A v p n. -PrC( X P' v > (1 -P) (7) by at least an order of magnitude, it can be written with small error as q/A = K1 gl/-2n (a )4/:-2n (8)

where K1 is a function of properties given in one form by 2/3-2n K1 L.056 ClP(,)1/2 Cp3/2 ( -)nj /32n In attempting to utilize Equation (8) to describe the boiling characteristics in an accelerating system both C1 and n, and hence K1 may be a function of the acceleration. 3. A number of relationships for the peak heat flux with pool boiling have been presented in the literature, (a) Addoms, whose work is cited in McAdams (1-pp 384) correlated experimental peak heat flux data by plotting (qc/A)P.,. ~. versus - Pv hpv(gg)z/3 pv If the deviation from a straight line near the critical state is ignored, the best line of the data yields an expression of the form (q/A)p = c x PV (ga )/3( (10),Pv where C = 2 and n = 1/2 (31) (b) By the application of dimensional analysis Borishanskii derived similarity criteria for the condition of peak heat flux which, in conjunction with experimental data, result in the following(q/A)p =K2 (p/ [a g(P P]1/4 where K2 = 0o13.+ 4N~4 (12) where K2 = 0,13+I. 0 4N (12)

-10Y y3/2 and N =/2 (13) 4 2ig(P Pv)]1/2 N is a correction factor to K2. For water boiling at atmospheric pressure the term 4N- has a value of approximately.013, or 10% of 0.13. (c) By considering the peak heat flux as a hydrodynamic (5) stability problem, Zuber arrived at Equation (14) without the necessity of an empirical constant (q/A)p = X (pv) /2[ g(p p- )]1/I 1/2 (14) (d) Using two different models for the peak heat flux Chang and Snyder(44) obtained correlations identical to Equation (11) with K2 = 0.145. (e) The correlation for peak heat flux for both pool boiling and forced convection derived by Griffith(45) specifies that the peak heat flux is proportional to gravitational acceleration with an exponent of 1/3. It is noted that except for (a) and (e) above, the peak heat flux is given as proportional to gravitational acceleration with an exponent of 1/4. E. Previous Experimental Work in Boiling Heat Transfer Embodying an Alteration of the Gravitational Field (32) In a paper referred to earlier, Siegel and Usiskin made a photographic study of boiling from several heater configurations in the absence of a gravitational field. No attempts were made to measure heat fluxes or temperatures for the series of tests reported. With water flowing in a vortex in an electrically heated tube a peak heat flux on the order of 55 x 106 Btu/hr.-ft2 was attained.

-11This was attributed to the effect of the centrifugal acceleration on the bubbles forming at the heating surface, estimated to be 18,000 times normal gravity at the exit from the heating tube. However, the contribution of forced convection could not be isolated, A number of investigations have been made on the effectiveness of an increased force field in promoting heat transfer by natural convection and condensation, 37)

II. EXPERIMENTAL APPARATUS AND INSTRUMENTATION The experimental work was performed using the facilities of the Heat Transfer and Thermodynamics Laboratory in the Mechanical Engineering Department. A. Configuration In order to most effectively isolate the influence of the increase in buoyant force on the bubbles with nucleate boiling in an accelerating system, the orientation of a heater surface as shown in Figure 1 was selected for the experimental work, that is, with the acceleration applied normal to the surface. For practical reasons the centrifuge principle was used to obtain the acceleration. To maintain an approximately uniform depth of liquid over the surface and the acceleration normal to the surface at all speeds it was necessary that the vessel be pivoted, as in a fly-ball governor, Figure.2 is a drawing of the pivoted test vessel and Figure 3 shows the overall centrifuge assembly. If the center of gravity of the test vessel is located at the surface of the liquid, a plane tangent to the liquid sun.face at the center of the vessel will be parallel to the heater surface. The liquid surface will have the shape of a parabola of revolution. If the center of gravity of the test vessel is located at the heater surface the acceleration will be normal to it. Obviously both conditions cannot be maintained for a finite depth of liquid, and a compromise is necessary. This aspect of the apparatus is discussed in more detail in Chapter III. B. Construction 1. Heater Assembly The heater itself consists of a cylindrical piece of copper, containing 1% lead for machineability, 3 inches in diameter and 1 inch -12

-13HEATING SURFACE Figure 1. Orientation of Acceleration with Respect to Heating Surface.

-14PIVOT ARM TC -6, WATER TC-5, WATER CONDENSER COIL OUTER CONTAINER DRIP PLATE FLOW GUIDE,, I — t -1t HOMELRESSTNC ~.~ % ) ~. I,, H;4 f / tIHEATER CONTAINER " S S NSUPPOOR. /TO-, HEATER _ HEATER BLOCK GHROIEL RESISTANCE Figure 2. Test Vessel. RIBBON,,~.-;,1.S.S. SPACER RING Lf9P- 7- - j,~.-..~. ~ ~- HEATER GUARD L~D~cr.,.,:c~ I:' — GUARD HEATER (3' ~' ~'~"?" ~ ELEMENT AT-G, DIFFERENTIAL THERMOCOUPLE Figure 2. Test Vessel.

T~AC~H~O~METER.COOLING WATER INLET TACHOMETER GENERATOR (TO ELECTRONIC COUNTER) BELT DRIVE KEROSENE THERMOCOUPLE COMPENSATION BATH _ MERCURY THERMOCOUPLE SLIP-RING ASS'Y 10 CHANNEL'S l mVARIABLESPEED HYDRAULIC TRANSMISSION COUNTERWEIGHT MAIN SHAFT- COOLING \ ELECTRIC WATER OUTLET DRIVE MOTOR TEST VESSEL -' 11 P POWER SLIP-RING ASS Y TO DRAIN Figure 3. Centrifuge Assembly.

-16long. One end of the cylinder serves as the heat transfer surface. In order to provide a continuous surface and to keep the heat loss by conduction to minimum, this end was undercut with a bevel and a mating piece machined from stainless steel.002 inch undersize,as shown in Figure 4. By immersing the copper in liquid nitrogen the stainless steel skirt slipped over the bevel, and upon return to normal temperature a water tight smooth surface was obtained. The heating surface was then chromium plated. Four 1/32 inch diameter holes were drilled radially into the cylinder for the insertion of thermocouple probes. Two of the holes extended to the center from opposite sides as shown in Figure 2, the centerline of one being approximately 1/16 inches below the heater surface and the other being 7/16 inches below the heater surface. The two remaining holes were located at 90~ from the above holes, extended half-way to the centerline of the cylinder, and were also located 1/16 inches and 7/16 inches below the heating surface. Figure 5 shows the location of these holes with the respective thermocouple designations. In the lower end of the pylinder 32 parallel slots 0.00o8 inches wide by 5/16 inches deep were machined to accomodate 6 feet of Chromel A heater ribbon 1/4 inch wide by 0.002 inches thick. The ribbon was insulated from the copper with 0.003 inch thick strips of mica. Gold foil O.001 inch thick shunted the ribbon where it emerged from one slot and entered the next in order to minimize "hot spots," Figure 6 is a view of the underside of the heater showing the slots and one of the thermocouple holes near the stainless steel skirt, To protect the ribbon

-17HEATER SURFACE 026S.S. SKIRT 0 u 1/16" Figure 4. Detail of Copper Heater-Skirt Intersection.

,-18/'~~~~~~~ //x~ " ~~~~~-SE 3" DIA Fo I- T I (NEAR SIDE) TC-3(NEAR SIDE)!F z- " —---------—'. -~.... — "'- - 6=__... _........."TC-4(FAR SIDE)J TC-2(FAR SIDE) Figure 5. Heater Thermocouple Locations.

-19and to provide a uniform temperature on the underside of the heater a nickel and chromium plated copper cover 1/4 inch in thickness was assembled to the heater block as shown in Figure 7. In order to minimize heat losses by radiation and convection from the underside, the heater is installed in a 1/4 inch thick chromium plated copper base, seen in Figure 8, which serves as a guard heater. The thin stainless steel spacer ring (refer to Figure 2) supports the underside of the heater cover, Differential thermocouples are installed between the heater cover and the heater guard (Figure 2), and a heating element located under it is controlled to keep the temperature difference at a minimum. The assembled heater unit, Figure 9, is quite compact, measuring 6 1/2 inches in diameter and 2 inches in depth. 2. Inner Container and Cover The inner container side walls consist of a double wall sheet stainless steel welded assembly separated by an air space. The inner surface was chromium plated to prevent contamination of the water due to corrosion at the welds. Figure 10 is a view of the inner container walls attached to the heater assembly. Fittings for installation of the water temperature thermocouples are installed in the cover, and a coil of 3/8 inch O.D. copper tubing is silver-soldered to the underside as a condenser. This assembly was also chromium plated. Teflon gaskets are used wherever contact with the test water is necessary. The inner chamber is vented to the atmosphere thru a small tube which is surrounded by the cooling water as it enters the condenser. To prevent a change in pressure within the vessel due to an aspiration effect upon rotation, the

0 CI C)0 0 H a) 0 -,-I~~~~~~~~~~~~~~~~~~~~~~~~~~~ 0 o3 H, —! —! rd.........o..................................................CU.,.....,............... 0............................................................................................................................................................................................................................................................................................ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ i~I~........................................................................................................................................................................................................................................ I.................................................................................................................................................................................................i~i~'iiii~iii iiii~ii~i~iiii:........................................................ ~ ~ ~ ~ - i- ii ii::i ~-ii~~iii-:-:iii~iii~iii~-::i~~iiii-i -ii~-iii-iii-i~ii.......................................................................................................................:: -: -i ii~ i i- ii ii~ i i~ -i i ~:: -i i-~ i -i i~ i: i: i- ii -: i............................................................................................................"I,........................I" I,

-21-........... f0...0W.000t0..0.;0Xl0.x........................ X0.t0XXX0.0.0.00000...000040X;000000jt00000000-00;~~~~~~~~~~~~~~~~~... e.......................................j 1....-..,-,...................................'.'.'.'.'.'.'.. 0~ ~ ~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~...;;;.;._.;;........................................g....-.g.-.g.ii-.....g.........-i'.......'.............................. R...... R......... m~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~..............................................S::.S.S.:.:::::::::.::::::.::::::.::::::.::::::.::::::.::::::::.::::::.::::::.::::::.::::::.::::::.::::::.::::::.::::::::.::::::.::::::.::::::.::::S:.:::::S.SS:...:.:.:.::.:.:.::.:..:.:.:.:S..:.:..:.:.S..S.:.:..:.:.:..:.:.:.::.:.:.::.S.:.::.:.:.:. ~ ~~ ~.....-,. I. I I.I I...........................................sg.iE g-gE.E -E - -. —E -. SS..........................g g S R ~................................................ tg:S S u~ ~ ~ ~ ~~~~~~~~~~~~.. I..I I I.. II II............. II II.I.II.I....................................................................::::::::::::::::::::.:::::::::::::::::S::S::::S:::.::S::::S.........................................................:....S:j::::::::::Sj:::S:::::::::::':::::.:S:.:::S.....:::::S:.::::.:..::..::.::...::...:...............:..:'~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~~~~~~~~~~. j:S::.:S::::::S:::::::::::::::S:::::::::j::::''.::::S:X.RR:........::::::::::.::::::.::::'..:....:..::.:':::::.::::SSSR:...........................................................'.......................... g..-s...gg..g...-..s...................................................j..S...SS..............SS..R'......j..''.....''........j.....................................................................R..R..II I. II II I II.......I.............................rs g.sy.gg r j g g ~ ~ g g s g. RR.g~.......~~ggB.~'jg S a gR -.... g''...''....................~S.RR.,...SR.....-gs,,, s.:............. gg g g g ~.'..'..g.'... gg.'...............Sj j..,'.....E..,............... —-.s':"g. S..'..'."'",..'.':.. j..-.-.::.:::.::................................ Fi ur. 9..........ebe He te.U.. nit........

Figure 10. Assembly of Inner Container Walls and Heater Figure 11. Inner Container Test Assembly.

-23vent tube is connected by rubber tubing to a point near the centerline of rotation. Figure 11 shows the entire test vessel assembly with thermocouple and heater leads prior to installation in the welded aluminum outer container shown in Figure 12. 3o Flow Guide The heating surface consists of a heated section in the center of an unheated section. In order to eliminate the possible influence of the gross convection currents on the bubbles forming near the edge of the heated section, a flow guide was constructed, consisting of a cylinder of thin sheet stainless steel open at both ends. The inner diameter of the flow guide is about 1/32 inch larger in diameter than the heated section, and is installed so that it rests over the heater on the skirt. The net effect is that of having a test vessel whose entire bottom surface is heated. Several test runs were made with this flow guide removed. 4. Thermocouple Construction A representative thermocouple installed in the 1/32 inch diameter holes in the heater block is shown in Figure 13. These were constructed by filling s short piece of 1/32 inch O.D. copper tubing with a ceramic cement, passing a 30 gage constantan wire thru the tube and welding the tube and wire near the tip by pinching the tube, after applying a voltage with a charged condenser. The tip was then trimmed so that the junction occurred at the extreme end of the tube. A 30 gage copper wire was welded to the other end of the tube on the outer side in a similar fashion.

Figure 12. Test Vessel Installed in Centrifuge. 30 GAGE COPPER WIRE ~ WELDED JUNCTION WELDED TO TUBE I / 32"O0.D.. COPPER TUBE SPACE FILLED WITH TECHNICAL "G" COPPER CEMENT CONSTANTAN WIRE Figure 13. Construction of Heater Thermocouple Probe.

-25The water temperature probe designated T5 in Figure 2 consists of a 30 gage copper constantan thermocouple cemented inside a 1/16 inch OD. stainless steel tube closed at one end and extending 1 inch from a 1/8 inch O.D. s.s. tube for rigidity. The thermocouple tip is located 1/4 inches from the heater surface in the center. The small extension tube is used to keep errors due to conduction in the tube to a minimum, Water temperature thermocouple T6 is similarly constructed except that only 1/8 inch O.D. s.s. tubing is used for the outer casing. Several were made with different shapes to permit water temperature measurements at different locations, 5. Thermocouple circuit In order to remove the thermocouple EMF's from the rotating member, a mercury slip ring was selected as being capable of the greatest precision, A similar assembly was used by Fultz and Nakagawa(39) with copper-constantan thermocouples. It was stated the spurious EMF's gave no trouble, even with voltages measured on the order of several microvolts. The temperature errors in the present work are discussed in Chapter IIIo It was found that 30 gage copper wires moving through the mercury deteriorated completely due to amalgamation after about 1/2 hour of operation, Iron wire was the most commonly available material which would not be attacked by mercury, and 24 gage wire was introduced as an intermediary between the copper and me.rcury both on the moving and stationary sections. By the thermoelectric "law of intermediate metals," no error should result if the corresponding junctions on the rotating and stationary members are maintained at the same temperature.

-26The equivalent circuit for the measurement of temperature with respect to the ice point is shown in Figure 14a. Ten concentric mercury channels were provided in a plexiglas piece, shown schematically with two channels in Figure 15. The stationary iron wires enter the mercury through the bottom, and the moving wires dip into the mercury through small holes in a plexiglas dust cover which rotates with the shaft. The corresponding iron-mercury junctions are the stationary and moving wires in the same channel of mercury. The mercury is maintained at a uniform temperature by having the plexiglas assembly rest on a heavy block of aluminum, which also serves as the upper bearing block, and by the stirring action of the wire as itrmrves through the mercury. A difference in temperature between the moving and stationary iron wires might be anticipated because of the stagnation effect, but calculations showed it to be negligible at the rotational speeds employed. The corresponding copper-iron junctions are maintained at a uniform temperature by insertion in a circular kerosene compensation bath which rests on the dust cover and rotates with it. Again, the stirring action of the thermocouples is relied upon to provide the uniform temperature. The water thermocouples T5 and T6, and the heater thermocouple T1 located in the center of the block near the heating surface (Figure 5) are used with respect to the ice point, and employ a common constantan wire in passing through the mercury slip ring. The remaining thermocouples in the heater block (Figure 5) are measured as differences as follows, by

-27Junctions in "Isothermal" Bath Copper L Hg I e Copper To Measuring Potentiometer Junction Copper I Const Fe Hg Fe nst. Stationary Rotating Section Section Reference Junction a. Equivalent Circuit for Measurement with Respect to Ice Point. Copper, Fe I Hg I Fe, Cons t. To > T I'~' Temperature topper Difference to Potentiometer Copper + Fe Hg Fe / Const. be Measured Stationary Section Rotating Section b. Equivalent Circuit for Measurement of Differences. Figure 14. Equivalent Thermocouple Circuits.

-28STATIONARY WIRE ROTATING WIRE MAIN SHAFT KEROSENE COMPENSATION BATH DUST COVER MERCURY (ONLY 2 OF 10 CHANNELS SHOWN ) ALUMINUM BASE Figure 15. Sketch of Mercury Thermocouple Slip-Ring Assembly.

-29a suitable switching arrangement~ T2 T1 = a (2l1) TI1 T 3 = (1-3) 1: - T4 = A (24) Owing to the contact between the copper tubes serving as one of -the thermocouple leads and the heater block it was necessary -to use the circuit represented in Figure l4b to measure the voltage differences, In addition to the above thermocouple circuits, the differential thermocouples between the heater guard and the underside of the heater block are taken through the mercury channels. Also, to check for the effect of differences in velocity.of the wires moving through the mercury, the innermost channel (radius = 1-1/8 inches and the outermost channel (radius = 3 15/16 inches) are short-circuited on the rotating assembly, and periodic measurements taken at the stationary wires at various RPMo All connections between various parts of *the circuit are made in tight copper boxes to avoid air currents and ot;her thermal gradients0 Figure 16 is a view of the thermocouple slip-ring assembly showing the mercury channels, dust cover, kerosene compensation bath and support for the stationary junctions in the kerosene bath0 The wires from the test assembly to the slip rings are led around the bearing through slots in the main shaft0 6. Cooling Water System Cooling water for the condenser is transmitted to the rotsating system through a rotating seal at the upper end of the main

-30shaft. A hole drilled at the center of the shaft extends below the upper bearing, and connection with the pivoted test vessel is made with rubber tubing, as seen in Figure -12. The water then enters the lower part of the shaft, again via rubber tubing, bypasses the lower bearing through the hole drilled in the shaft, and enters a drain at the lower end of the shaft. For certain tests with low flux rates it was found necessary to drastically reduce and control accurately the rate of cooling water flow. For these tests small sharp-edge orifices were installed at the discharge from the lower end of the shaft and calibrated with a needle valve and pressure gage located near the stationary water inlet to the shaft. 7. Electric Power System Electric power for the main and guard heaters is transmitted to the rotating shaft via a slip ring assembly located under the lower bearing, as seen in Figure 3. The wires bypass the lower bearing through slots in the shaft and coiling of the wires prevents their imposing any restraint on the pivoted test section. 80 Power Drive The main shaft is rotated by means of a belt connected to a variable-speed hydraulic transmission, driven in turn by an electric motor 9. Counterweight Both the test assembly and the counterweight are pivoted on a cross-arm attached to the main shaft. The counterweight consists of a

.......... Sk.................................................................................................................................................................. Figure 16. View of Mercury Thermocouple SlipRing Assembly........................................................................................ - 1111,11,11.111.11,.......................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................

-32threaded rod on which the desired weights are attached, with a pivot at one end. The counterweight was made equal in weight to the filled test assembly, and the center of gravity was made equal by a trial and error process, by moving the weight up and down the threaded rod until no vibration could be detected. Hence the system was dynamically balanced for all rotational speeds up to 220 PRM, the maximum for which it was designed. C. Instrumentation 1. Temperature All thermocouple EMF were measured with a Leeds & Northrup Type K-3 potentiometer with the exception of the guard heater differ6,ntial thermocouple. It was considered adequate to maintain the temperature difference between the heater and guard within 5~F, and a G-M Laboratories portable galvanometer readily detected this difference. 2. Power The main heater power was furnished by a 7 KW Variac and the guard heater power by a 1.2KW Variac. A standard high quality calibrated voltmeter and ammeter measured the power to the main heater. It was necessary to use an auxiliary voltmeter to measure the voltage at the heater terminals on the test vessel because of the voltage drop across the carbon brushes in the slip-ring assembly. Obviously this was done only during nonrotation, During a particular run the electric current to the heater was maintained at a given value, but fluctuations in line voltage required periodic adjustment of the variac. No attempt was made to measure guard heater power,

-333. RPM of Main Shaft A tachometer - generator, seen in Figure 16, generating 60 pulses per revolution was geared to the main shaft with a 1:2 ratio. Its output was fed to a Model 522B Hewlett-Packard electronic counter. Figure 17 is an overall view of the test apparatus and instrumentation.

III. ESTIMATION OF ERRORS A. Temperature 1. Potentiometer The manufacturer of the potentiometer (L & N, K-3) states that the maximum error at the range used is + (.15% + 0.5tv). Used in conjunction with the potentiometer were a certified standard cell and a mirror-type galvanometer with a sensitivity of 0.00031A/mm. In order to determine the reproducibility$, of readings with the potentiometer, a series of 20 successive independent measurements were made with a thermocouple at the steam point in a hypsometero The standard deviation of these readings was found to be 0o35v, less than the maximum error given by the manufacturer. With copper-constantan thermocouples the standard deviation amounts to approximately 0.010F. In tests run at the higher values of heat flux, a tendency for the surface temperature to vary periodically with time was noted. Readings of the potentiometer were taken at the mid-point of the oscillation, and the extent of the deviation on the galvanometer scale recorded. To convert this deviation to a temperature, the deflection of the galvanometer was calibrated and found to correspond to 1.74), volts per mm of deflection. 2. Thermocouples All thermocouples were constructed from the same spools of wire and calibrated at the steam point, and at other temperatures by comparison with Bureau.of Standards calibrated mercury-in-glass thermometers. The thermocouples agreed within + lev. No difference in calibration could be detected between the water temperature thermocouples -34

and the heater block thermocouples made with the 1/32 inch O.D. copper tubing serving as one of the conductors. The composition of the copper thermocouple wire was not available, but the manufacturer of the tubing listed its composition as oxygen-free high conductivity copper - copper 99,96%, phosphorous -.0003% max. A thermal EMF arising from any inhomogeniety between the wire and tube would be smalltince the junction exists in an essentially isothermal zone. 3. Effect of circuit on accuracy of measurements, The determination of errors in temperature measurements introduced by the complex circuit consisted of a two-step process. A check of the system under static conditions (i.eo - nonrotating) was made, whereby any error due to the presence of the iron wire, the kerosene compensating bath, the mercury and the switching arrangement may be detected. A check also was made under rotating conditions, which would then indicate the sole effect of the rotation, if no. errors are detected in the first check, The first check is absolute, the second relative, For the first check, one of the thermocouples in the test vessel was disconnected at a junction box.on the rotating.assembly and another previously calibrated one substituted in its place. This thermocouple was again calibrated at the steam point, but the connection to the potentiometer was now made by way of the entire test circuit as it is to be used. The measurement came within the + lmv deviation obtained with the original calibrations.

-36For the second check it was necessary that a stable temperature be available in the test vessel in order to compare rotating and nonrotating measurements. A number of trials indicated that the temperature of the water in the test vessel was not uniform enough over the period of time required in spite of the large thermal inertia.of the test vessel. The acceleration upon rotating caused convection currents which changed readings by as much as 8Lv (i.e. -0.3~F), hence overshadowing the error to be detected. The temperature of the copper heater cylinder-however, was found to be sufficiently stable if no power were applied for at least the previous 48 hours. As was stated earlier, thermocouple T1, located nearest the heating surface in the center, was the only one of the four thermocouples in the heater cylinder to be measured with respect to the ice point. The remainder were used as differential thermocouples. Table I lists the designations of the mercury slip-ring channels and Table II lists the channel pairs associated with each thermocouple circuit. TABLE I LIST OF MERCURY THERMOCOUPLE SLIP-RING CHANNELS AND RADIUS FROM t OF ROTATION Channel Radius from'L Channel Radius from d No. of Rotation-In. No. of Rotation-In. 1 1 1/8 6 2 11/16 2 1 7/16 7 3 3 1 3/4 8 3 5/16 4 2 1/16 9 3- 5/8 5 2 3/8 10 3 15/16

-37TABLE II LIST OF MERCURY CHANNEL PAIRS ASSOCIATED WITH EACH THERMOCOUPLE CTRCUIT Thermocouple Thermocouple Channel Pair Designation T1 2 & 1 AT(2-1) 3 & 4 AT(1-3) 3 & 5 AT(2-4) 4 & 6 T5 2 & 7 T6 2& 8 Readings of the differential thermocouples in the heater cylinder changed by no more than liv from nonrotation to the maximum of 220 RPM. The influence of the rotation upon thermocouple T1, through mercury channels 2 and 1, was determined by taking a rapid succession of readings at 0 RPM, quickly bringing the assembly to a given speed and taking another quick succession Of readings, taking more readings at 0 RPM, and so on for the other spends. This was repeated for thermocouples T5 and T6 by substituting thermocouple T1 in mercury channels 2 and 7 and 2 and 8 respectively. Figure 18 is a representative plot of the data taken with thermocouple T1 in mercury channels 2 and 7. The error which will be introduced in thermocouple T5 is thus indicated. Both the change measured in going up to speed and the cloange in coming down to 0 RPM are measured since they usually were not equal, and are plotted as corrections to be made to the readings.

RPM 1.018 0 4 8 12 16 20 24 28 32 1.016Figure 18. Typical Plot Indicating Effect of Rotation Upon Therocouple Readings. >. 1.014 -9 C)E~C-'! THERMOCOUPLE TI 1. 012 SUBSTITUTED FOR Ts IN MERCURY CHAN NELS 2 & 7 0 4 8 12 16 20 24 28 32 TIME, MINUTES Figure 18. Typical Plot Indicating Effect of Rotation Upon Thermocouple Readings.

-39Figure 19 is a plot of the corrections to be applied to the thermocouples measured with respect to the ice point as a function of rotation. Many more data points were taken than are shown, but are not included for the sake of clarity. No net corrections were applied to T1 since the data deviated quite uniformly plus and minus. The cause of the EMF is not certain, but is believed due to the action of the earth's magnetic field on the uncompensated length of wire between two channels of a circuit, resulting in the presence of a unipole generator. The largest corrections must be applied to thermocouple T6, whose uncompensated length of wire has the greatest length and average velocity of (46) any of the circuits. Using tabulated values of local horizontal magnetic intensity and dip, the EMF generated at 220 RPM were calculated: Circuit T1 = 0.355v (vs + 2[iv experimental) Circuit T5 = 3.17tv (vs 3o06v experimental) Circuit T6 = 4o06tv (vs 9.8tv experimental) Also, by reversing the direction of rotation the polarity of the EIF was observed to change. Fultz and Nakagawa(39) state that they wired their rotating thermocouples noninductively to eliminate the induced EMF's due to the horizontal component of the earth's magnetic field. In any case the net effect is known and corrections can be made. The mercury system is the largest source of error of any of the components in the temperature measuring system. It should be noted in Figure 19 that none of the data points lie more than 2iv from the curve used in the reduction of data. With copper-constantan thermocouples and taking into consideration the other sources of error, it is believed conservative

uO.'l.q.Ot.O. aO'JGSUGduIo0 o0. sa'GTnoouiOaOGL O' PGTTaddV suoTq.Go -3 * 6 a6JnLGT. INdl' C33dS VNlNOI.lOVI OO 091 001 09 0 8 le S13NNVHO'91 010 89-D'- 0 t- Cl) 0 rl 0 (I) 0 Z U 8- 0n ~1- v O ~ ~ ~~0 _- -I 0 > 0 ( 0

to state that the temperature at any thermocouple on the rotating assembly can be measured within + 0,10F. 4. Determination of heater surface temperatures, The measurement of the difference in temperatures of T1 and T2 (Figure 5), AT(2-1), in the heater cylinder are made primarily to enable an extrapolation of temperature T1 to the heater surface. To do this the location of the bottom of the 1/32 inch diameter drilled holes must be known, With such deep holes small drills often have a tendency to deviate from a straight line, Two different techniques were used to locate these positions. Straight, solid, 1/32 inch diameter rods could be slid easily into the holes, indicating that the holes were straight. With a height gage on a surface plate the measurements were then extrapolated to the bottom of the holes, Figure 20 shows the locations of the holes, These are believed accurate within + o003 inches. As a check, an ultrasonic transducer was applied to the bottom side of the heater. Reflections from interferences showed up as "pips" on a Reflectoscope, and by taking ratios of distances between "pips" on the screen the distance to the near side of a hole could be calculated, Because of other interferences it was possible to obtain clear readings for only one hole. This came within 0,oo006 inches of the value determined by the previous method, which is about the extent of the resolution of the Reflectoscope. To determine the uncertainty in the determination of the heating surface temperature by extrapolation from the interior, the procedure outlined by Kline and McClintockl40 is followed. The uncertainty int1erval

AX, =. 063 HEATER SURFACE Tw A /X2.364 _056, I T4 TZ Figure 20. Location of Heater Thermocouple Holes. Figure 20. Location of Heater Thermocouple Holes.

of a result which is obtained from a number of variables, each in themselves possessing.an uncertainty., is given by' WR W1= [ wV + w( a' 2 + +. + W(15 RL VlI uV2 2) 2 where Vn = independent variables WR = uncertainty interval W of the resulzt R Wn = uncertainty interval of the independent, variables. The uncertainty of each variable is specified by using a mean of the readings and an uncertainty interval (ioe,, M + W). Referring to Figure 20 the heater surface temperature can be obtained from Equation (16)Q Tw [T(2 (16) Using the term (Tw T1) as the result in Equation (15) and dividing both sides by Equation (16) to put the uncertainty in terms of percentages gives W(TWTl) V $W1.2 +(Wx2\ 2 (W(21)>2 A2,/21 Tw T1 l T(2-1) (1) For the dimensions of Figure 20, 6x,1 = ~063 + ~020 inches (18) AX,2 = ~364 + o050 inches with a confidence limit of 95%o The large values of W listed are necessary because it is not known which temperatulre across the diameter of the hole the thermocouples are actually detecting, The third term on the right hand side of Equation (17) is negligible in comparison to the first two. S~ubstituting the above values into Equation (17)~ W(T-Tl) = o (l95 Tw~l.T 35 i(.9) T.T1

-44Values of W(Tw-T1) are calculated from typical test data and listed in Table III. TABLE III VALUES OF UNCERTAINTY W(TwTl) CALCULATED FROM TYPICAL TEST DATA Approx. q A AT(2-l) T.-T1 WTZ1 Btu/Hr-Ft OF oF W2 l) 10,000 1.2 0,2 0.1 25,000 3.2 0.5 0.2 50,000 6.2 1.0 0.3 75,000 10.2 1.7 0.6 100,000 13.4 2.3 o.8 The values seem large, but it must be pointed out that these result from the geometry of the system, that is, because of the unknown exact point of contact of the thermocouple junction with the heater block. For the series of tests reported here the thermocouples were not disturbed after the initial installation, Hence, although the absolute value of the heater surface temperature cannot be stated with any greater certainty than that listed in Table III, for purposes of comparison between different tests, the uncertainty will be given by the third term of Equation (1'7) which will give a value of W(TwT1) much less than + 0.10F. Since the intent is to study the effect of acceleration upon boiling heat transfer, comparisons will be made between various tests, For purposes of comparison, then, the uncertainty

of the heater surface temperature will be governed primarily by the uncertainty of the measurement T1, which was previously given as + O.10F. 5. Determination of fluid temperatures. As described earlier, the thermocouples used for measuring the water temperature are encased in stainless steel tubes which are supported from the cover of the test vessel, The cover is at ambient temperature because of the cooling water flow, The thermocouple will not give a true indication of the water temperature if conduction heat transfer up the tube wall is significant. Since the conditions present cannot be specified precisely, an error will be determined analytically by assuming the most adverse conditions possibleo Figure 21 is a sketch showing the principle dimensions and temperatures for T6, which has the largest cross sectional area and is immersed the least in the water. The solution to the fin-type differential equation applied to this particular problem, derived in Appendix A is t2(M-K) K tw t~ +B (20) m mi(1Z ) m(tanh m0+ P)( m m where, tw = water temperature at the tip of the tube ty = temperature of the tube at the tip M = temperature gradient in the tube at the liquid surface K = saturation temperature gradient in the water hC m = kA = depth of immersion of the tube in the liquid hAl kA

-46t ambient = 70~ F / —COVER,. —_1I/8 DIA 1. two 2110 F (SATURATION)'t =216.5~ F - (SATURATION) THERMOCOUPLE T6 HEATER SURFACE Figure 21. Typical Conditions for Thermocouple T6 with System Subjected to Acceleration 21.15 Times Normal Gravity.

A = cross-sectional area in the body of the tube A1 = cross sectional area at the closed tip of the tube h = coefficient of convective heat transfer between the fluid and the fin' Making the condition even more severe, the fin is taken to be a solid rod instead of a hollow cylinder. Then A = A1, and B = h/k, Also it is assumed: h = 100 Btu/hr-ft2 k = 8 Btu/hr-ft-~F (18-8 s,s,) The largest value of K will occur at the maximum acceleration of the test chamber, At 21,15 times standard gravitational acceleration in the apparatuso K = 26.4 ~F/ft This assumes that the liquid is saturated at all depths, likewise the most severe gradient which can exist, Assuming that the rod in the vapor space is insulated, the temperature gradient in the rod at the water surface can be approximated byo -t; -t wO amb _ 21,70. = 564OF/ftr QL 0 25 For the 1/8 inch diameter solid rod~ A = 8,5 x 10-5 ft2 C = 0,0327 fto = 1 1/4 inches = 0.104 ft.

then m = 69~4 mr = 7.21 B = 12.5 Substituting the appropriate values into Equation (20): Error = tw-te 564-2= x 2 (21) 69.4 160( 12.5 694(12.5) 69.4 69.4 =.01 +.32 = 0.33~F It should be noted that the second term of Equation (20) contributes the greatest part of the error. Experimental measurements of T5 and T6 in a line perpendicular to the heating surface have shown that virtually no temperature gradient exists in the part of the water probed, henceK = 0 and the error is negligible. Thermocouple T5 has a smaller diameter tube and is immersed to a greater depth than T6, hence its error is also negligible. Two additional factors aid in keeping the error small; the thermocouple supports are tubes rather than rods, and the portion of the tubes above the liquid level are exposed to saturated steam rather than being adiabatic. 6. Guard heater temperature difference The combination of the guard heater differential thermocouple and portable galvanometer were calibrated by the use of an auxiliary temporary thermocouple installed in a small hole in the heater guard (Figure 2) to measure its temperature, and thermocouple T1 to measure the temperature of the heater block, With no water in the test vessel, a small current was passed through the primary heater

o49and simultaneous readings taken of thermocouple Ti, the auxiliary thermocouple and the portable galvanometer. It was found that one millimeter deflection of the galvanometer corresponded to a 5OF temperature differential0 B.o Rotation of the Main Shaft The tachometer generator geared to the main shaft generates 1.20 pulses for each revolution of the shafto. With the electronic counter set with a one second gate time the direct reading was twice the actual RPMo The gate time is calibrated against an internal standard crystal~ Since the direct readings are accurate within +./2 unit, the rotational speed of the main shaft is accurate within + 1/4 RPMo During operation, the rotational speed did not vary more than 1/2 RPM over a period of several hours because of the low friction and large inertia of the systemo C. Acceleration at the Heating Surface In order to calculate the centrifugal acceleration at the heating surface the angle which the test vessel assumes must be known0 It would be both difficult and inconvenient to measure this angle with any precision. Since the restraints on the pivoted test vessel were maintained at a minimuum it was deemed sufficient to calculate the angle, which requires a knowledge of the location of the center of gravity0 The vessel is symmetrical about its own axis, and the axial location of the center of gravity is determined by calculating that of the counterweighto The vessel and counterweight differ in mass within

-501 ounce out of approximately 30 pounds for each unit. During balancing of the unit, it was noted that moving the counterweight 0.03 inches along the threaded rod was sufficient to cause vibration at the highest rotational speed of 220 RPP4. Having a simple geometrical shape, its center of gravity is readily calculated and must be equal to that of the test vessel, within an uncertainty of + 0.03 inches. Figure 22a shows the dimensions of the test vessel necessary for calculating the acceleration at the heating surface. Referring to Figure 22b for notation, by elementary mechanics the angle 9 can be obtained from tanG = ~2. (b + i sin ) (22) g From trial and error solutions, G is listed as a function of w in Table IV for the rotational speeds commonly used in the tests. The assumption is made that the center of gravity of the assembly does not change due to the water surface taking the form of a portion of a parabola.of revolution. This is justified since the mass of the water is approximately 2 1/4 pounds as against 28 pounds for the remainder of the vessel. The centrifugal acceleration component ac at the heater surface is given by ac W= 2 (b + iH sinG ) (23) The total acceleration is the vector sum of Equation (23) and the gravitational acceleration, which is 32.17 ft/sec2 at the local latitude of 450: at a c + g (24)

-51/Z40 CENTER OF GRAVITY HEATING SURFACE a - PHYSICAL REPRESENTATION ~~c —. b - b- SCHEMATIC REPRESENTATION Figure 22. Test Vessel Dimensions for Calculation of Acceleration at the Heating Surface.

-52TABLE IV VALUES OF PARAMETERS IN TEST VESSEL WHICH VARY WITH ROTATIONAL SPEED Column 1. Rotational speed Column 2. Angle between heater surface and horizontal Column 3. Dimensionless total acceleration of heater surface Column 4. Angle between heater surface and acceleration vector Column 5. Total head of liquid at heater surface Column 6. Difference between angle of heater surface and liquid surface 1 2 3 4 5 6 Xn 9 Total Acceler- Y h RPM Degrees ation aT/g Degrees Inches 9 - cp of water Degrees 0 0 1.00 90.0 2.5 0.0 65 58.8 1.95 89.7 4.6 1.9 85 71,5 3.21 89.7 7.5 1.3 110 78.9 5.29 89.8 12.2 o.8 155 84.4 10.47 89.9 24.1 0.4 190 86,'3 15073 90.0 36.2 0.3 220 87.2 21.15 90.0 48.4 0.2 The angle y which the total acceleration makes with the heating surface is y = 9 + arctan g (25) ac Values of at/g and y also are listed in Table IV. Applying the uncertainty relation Equation (15), to the centrifugal acceleration given by Equation (23), we have Wac 4 2 sin 9 2 H cos 2 1/2 a b H + W 7I + I c W9 (26) aM 2 X -b + Q sin 9 H Lb + BH sin 9$

-53Take as the uncertainties of the variables: W - +_ 1/4 RPM = + 0.0262 rad/sec W = + 0.03 inches = + 0.0025 fte (27) We = + 20 = 0.0349 radians Substituting into Equation (26): For.@ = 65 RPM, the uncertainty in the calculation of the acceleration at the heating surface is Wat (59. 3 x 10-6 + 3 36 x 10-6 + 146 x 10-6/ + 144% (28) at For w = 220 RPM, Wat (5.16 10 + 76 x 1 + l09 x 10 ) 032 (29) at D. Pressure on the Heating Surface With nucleate boiling taking place from a heated surface, one of the established parameters is the saturation temperature at the surface, which can be determined if the pressure is known, For a given depth of liquid, the pressure is easily determined if the acceleration normal to the surface is linear, However, with an acceleration produced by rotation, the free surface of the liquid in Figure 22a, for example, will take on the form of a parabola of revolution, thereby complicating the calculation of the pressure at the heating surface. Moreover, the pressure may not be uniform across the heating surface, These aspects will now be considered,

-54Figure 23 shows the parabolic shape of the liquid due to rotation together with that portion in the test vessel indicated by the dotted lines. Using the notation of Figures 22 and 23 the head of liquid above the center of the heater surface is h = Y2 + d cos (0) Substituting the relations 2 2 D2r2 X X1 Y2 2; Y 2g (31) 2g 2g x1 r - d sin G (32) r - b + YH sin G (33) into Equation (30): h = 2 [2b sin G + (2eH - d) sin2 G] + d cos 9 (34) 2 g The assumption is made above that the depth of the liquid at the center of the test vessel does not change between the stationary and rotating conditions. Actually, the volume of liquid within the vessel is constant, and the determination of the true centerline depth would require an integration between the boundaries of the container and the liquid surface. The integral must be constant and presumably the precise location of the liquid surface could be determined as a function of rotational speed. The assumption of constant depth of liquid is justified when the difference between the heater surface and liquid surface angles is examined. Referring again to the notation of Figures 22 and 23, the angle of the liquid surface cp at the vessel centerline with respect to

-55tC!J',""' LIQUID SURFACE XC __ _ r ~~ "I- HEATER SURFACE _ _ r _ _ Figure 23. Relation of Heater and Liquid Surfaces Due to Rotation.

-56the horizontal can be determined from 2 2 tan cp = = - (r-d sin G) (35) g g The angle of the heater surface if given by 2 tan = - (b + sin G) (36) g and the difference, substituting Equation (33), is 2 tan G - tan cp = - sin G (e + d - JH) (37) from which (G - c) can be determined, With the values of G existing, (G - p) is a weak function of d, which justifies its use as a constant in Equation (37). Values of h as determined by Equation (34) and (G - cp) are listed in columns 5 and 6 of Table IV for various RPM's with liquid depth d = 2,5 inches. The values of (9 - p) are largest at the lowest rotational speeds but the total liquid heads h are small, so the net effect of the error due to assuming constant liquid depth will be small. At higher rotational speeds the difference in angles (0 - p) becomes negligible. By an analysis similar to the above, it can be shown that the difference in head between the upper and lowerJ edge of the heated surface due to the curvature of the liquid surface is given by Lh = D[- cos G (b + iH sin 9) - sin 9] (38) where D = diameter of heater surface. At c of 220 RPM, Ah is equal to 0o15 inches of water, which is negligible.

Values of the pressure and saturation temperatures at the thermocouples in the liquid are determined from expressions similar to Equation (34). An increase in pressure at the heater surface owing to acceleration of the vapor above the liquid surface is negligible. E, Heat Flux The meters used for power measurement were calibrated by comparison with precision laboratory standard instruments and found to be accurate within the resolution of the meters on all parts of the scales. Assuming that no losses exist, the heat flux is given by q/A E = C x (39) A r2 To obtain an estimate of the error in heat flux resulting solely from uncertainties in the measurements of voltage, AE, current, I, and heater radius, r, use is again made of the general uncertainty expression Equation (15). Performing the proper manipulations on Equation (39) the uncertainty of heat flux is Wq/A 2 r ( + (Wr2 ]l/2 (40) At a heat flux of q/A % 10,000 Btu/hr-ft2, typical values are E = 31.90 + 0.05 volts I = 4.70 + 0.05 Amps. (41) r = 1.470 + 0.005 inches

-58Substituted into Equation (40) the uncertainty is Wn/A /A = + 1.3% (42) q/A At a heat flux of q/A "~ 100,000 Btu/hr-ft2, typical values are E = 96.90 + 0.05 Volts I = 14.20 + 0.05 Amps. Substituted into Equation (40) the uncertainty is: Wq/A A + o.8_ (44) q/A - The heat losses by radiation and convection from the underside of the heater are negligible because of the presence of the guard heater. A large source of error in the determination of heat flux arises from the heat loss by conduction and convection from the stainless steel heater skirt, which is in contact with the water on one face and the heater cylinder at its edge. In order to determine the order of magnitude of the loss, the skirt is treated as a straight, extended surface with a step-change in cross section. The effect of the fin being.a circumferential one is negligible in this case, since the radius of curvature is large in comparison with the thickness and effective length of the fin. Two different expressions giving the rate of heat transfer by conduction from the root of the fin for the geometry shown in Figure 4 are derived in Appendix B. One result per unit area of the main heating surface is nqoss/A = 2.02 (h) /2 (5)

-59where h = heat transfer coefficient between the fluid and the extended surface - Btu/hr-ft2-F. 00 = temperature difference between the root of the extended surface and the fluid - OF. Two tests were conducted with heat transfer to water in the convective nonboiling region, with the flow guide removed. These provided values of h for the heated surface and values of 90 for different acceleration rates. Assuming that the same h applied to the skirt, it is possible to obtain the heat loss through the skirt with Equation (45). These are tabulated in Table V along with the per cent heat loss calculated from Equation (110, The agreement between the two is noted. The percentage heat flux losses were calculated using Equation (110) for the test runs in which boiling was taking place, and are plotted in Figure 24k. F, Calculation of Heat Flux Rate From Temperature Gradient in Heater Block In order to compare values of heat flux measured directly with those calculated from the temperature gradient measured in the heater block, it is essential that the uncertainty in such a calculation be determined. The heat flux is calculated from q/A k T(2-1) (46) Applying the uncertainty equation again: / [i =r + Xt)\ +2 2(47) q/A k; + +

-60o. S~ 5t~ ~(q/A)T: = 10,870 BTU/HR-FT.4 R 3 S (q/A)T 24,450 0o P(q/A)T- 48,800 " | (q/A), = 73,000 (q/A)T =99,500 ACCELERATION 20 Figure 24. Estirmated eat Loss by Conduction Through the Heater Skirt for Various Total Heat Fluxes

-61TABLE V LOSS BY CONDUCTION THROUGH SKIRT FOR CONVECTION TESTS Test h Heat Loss No. RPM (q/A) 1 Based on F0 % off (q/A)total l tota |(/A)total Equatio Equation (45) (110) C-1 0 4,700 204 23.1 13.8 12.6 110 4,700 277 16.7 11.7 11.2 155 4,700 312 15.2 11.3 11,0 190 4, 700 344 13.8 10.7 10.5 220 4,700 372 12.6 10.2 9.8 C-2 0 9,840 316 31.5 11,4 9.1 110 9,840 378 25.9 10.1 9.3 155 9,840 437 22.4 9.4 8.7

-62The value of k given by the manufacturer for the leaded copper is the same as that for pure copper, and its reliability is uncertain. However, the uncertainty of Ax most likely will overshadow this. Taking k = 217.5 + 10 Btu/hr-ft -~F Ax =.364 + 0.05 Inches (48) AT(2-1) = 3.2 + 0.10F i/A + 15% (49) q/A G. Measurement of Barometric Pressure Prior to use the barometer-was calibrated in the Meteorological Laboratory and corrections obtained. The barometer is accurate within + 00005 inches of mercury. H. Measurement of Specific Resistivity of Water in the Test Vessel In order to determine the purity of the water in the test vessel, a portable conductivity cell and bridge are used to measure its specific resisticity. The instrument was checked by comparing the readings obtained with a sample of water previously measured on a precision instrument in the Physical Chemistry Laboratories of the Chemistry Department of the University. The readings agreed at one megohm within the resolution of the portable instrument, which is approximately + 5% at this range.

IV. TEST PROCEDURES A. Fluid Water was selected as the fluid to be used for this study of the effect of acceleration on boiling heat transfer. Its properties are well established and large quantities of heat transfer data are available for comparison. Double distilled water was obtained from the University's Chemistry Department and distilled again in the Heat Transfer and Thermodynamics Laboratory shortly before being used, The resistivity of the water -in the test vessel was measured immediately before and after each test run. The specific resistivity of the water before each test was always 1.5 x 106 f -cm or greater. The test vessel was filled with 1000 mi of water, which resulted in a depth of 2.51 inches over the heater surface. For purposes of calculating the saturation temperature at the heating surface this was corrected for the increase in specific volume with temperature. The depth was measured again at the end of each test run. For the high heat flux test runs.it did not change, but for the runs at low heat flux it was necessary to vary the cooling water flow rate to attain saturated or as near saturated conditions as possible in the water. This sometimes resulted in a net loss of vapor through the atmospheric vent. The quantity lost each time was estimated, and the total estimate prorated over the course of a run, using the total measured loss. In no case did the total change in depth exceed0.25 inches for the test runs reported here. -63

-64After filling the test vessel and reassembling the apparatus, power was turned on and the water was boiled vigorously to degas both it and the heating surface. Prior to each test run,this was done for a minimum of 4 hours and in the majority of cases, degassing over a period of 16 hours was used. B. Heat Flux Range Covered Several test runs were made solely for obtaining convection data. These was performed with the water highly subcooled with nominal heat fluxes of 5,000 and 10,000 Btu/hr-ft2. For the test runs with boiling heat transfer the nominal heat fluxes were: 10,000 Btu/hr-ft2 25,000 Btu/hr-ft2 50,000 Btu/hr-ft2 75,000 Btu/hr-ft2 100,000 Btu/hr-ft2 For any particular test the heat flux was maintained constant while the acceleration was varied. C. Heater Surface Treatment Prior to chrominum plating, the heater surface was polished with emery cloth with successively finer grits. Finishing was done with the finest crocus cloth available. Before each lest the surface was again polished with crocus cloth and cleaned with reagent grade acetone and wiped with kleenex. After drying completely it was rinsed

-65twice with water of the same purity as the test water. All internal parts of the test vessel which may be in contact with either the water or vapor were subjected to the same treatment. No roughness measurements were made of the heater surface. Because of the similar treatments above, however, it is believed that the roughness will be approximately the same for all tests. D. Range of Accelerations Covered The acceleration rates for the tests were varied from 1 to 21.15 times normal gravitation acceleration in 5 to 8 steps, In earlier tests it was noted that the wall temperatures at a/g = 1 shifted gradually over along period of time. To isolate the effect of the acceleration, the stability of a reference base was improved by taking a set of measurements at a/g = 1 before and after each higher acceleration. Thus any gradual shift in the gravitational boiling surface temperature could be compensated by considering the change only. The maximum shift in a run which fulfilled all the requirements for acceptability was 0.80F over the average test period of 10 to 12 hours. To determine any history effect of the acceleration on the data, a sequence of accelerations similar to that given in Table VI was followed as part of the test. In most cases point numbers 2 and 6 agreed, indicating a negligible influence of an effect of history.

-66TABLE VI SAMPLE SEQUENCE OF ACCELERATIONS DURING TEST RUN TO DETERMINE EFFECT OF PAST HISTORY OF ACCELERATION Point No. a/g 1 1 2 5' 29 3 1 4 21.15 5 1 6 5.29 7 1 E. Location of Water Temperature Thermocouples Thermocouple T5 in Figure 2 is located on the centerline of the test vessel 0.25 inches from the heating surface for all. tests, Two different thermocouples were used for T6. One was straight, as shown in Figure 2, and the other curved, permitting its location in different positions within the test vessel. Its position was varied from test to test, but for a particular run remained fixed. The purpose was to determine, if possible, the effect of the Coriolis acceleration on the flow pattern in the water, which might influence the results obtained. Figure 25 indicates the various locations used for T6, where the numbers indicate corresponding positions in the two views. F. Variation of Cooling Water Earlier tests conducted at a heat flux of 25,000 Btu/hr-ft2 with a non-restricted flow of cooling water through the condenser coils were completely non-reproducible under acceleration. It was noted that

-67DIRECTION OF MOTION VIEW TOWARD'l - r3 4 - HEATER SURFACE 15 15 16 16 DIRECTION OF DE_ ~~MOTION __-1 1 VIEW FROM ABOVE DURING ROTATION FLOW GUIDE HEATER SURFACE Figure 25. Various Locations of Thermocouple T6 for Different Test Runs.

-68the water had become subcooled in varying degrees, and at the highest acceleration of a/g = 21.15 it appeared likely that no boiling was taking place, even with a wall superheat of 18~F. In order to achieve reproducibility and to assure saturated or as near saturated conditions as possible it was necessary to control the cooling water flow rate for heat fluxes up to 50,000 Btu/hr-ft2. Hence it was also possible to obtain limited data on the influence of subcooling for each of the various accelerations. The subcooling occurs because of the increasingly strong convective process in the vapor space above the liquid. As will be noted in the next section, at the heat flux of 50,000 Btu/hr-ft2 variation of the cooling water had little effqct on the subcooling of the test water. The coolant flow was decreased until a net quantity of vapor passed through the atmospheric vent. At heat fluxes of 75,000 and 100,000 Btu/hr-ft2 the cooling water was not controlled. To prevent highly subcooled condensed water vapor from returning to the main body, a drip plate,.shown in Figure 2, was installed over the flow guide. A number of small holes were drilled in the plate to act as a countercurrenth-eat'exchanger, with the condensed water passing down and the vapor rising up..G. Attainment of Steady State Conditions When conditions such as acceleration and cooling water flow rates were changed, sufficient time was allowed for the attainment of steady state conditions before data were taken. This ranged from 10 minutes to 1/2 hour, depending primarily upon the heat flux. At the low boiling heat flux of 10,000 Btu/hr-ft2 changes were sometimes so slow that pseudo-steady state data were taken,

-69H. Criteria for Acceptable Data For the tests reported, three conditions had to be fulfilled for acceptability of the data: 1, The specific resistivity of the test water at the conclusion of a run must be at least 800,000 JL. -cm. 2. No discolorations or spots should be present on the heating surface. 3* The value of Twall- Tsat with the system under gravitational acceleration must not change more than 0.8~F over the entire period of the test run. In almost all of the tests which were discarded, these three conditions simultaneously failed to be met. The tests performed before control of the cooling water flow was adopted also were discarded because of non-reproductibility.

V. TEST RESULTS A. Natural Convection Two tests, C-1 and C-2, were conducted with the flow guide removed at different values of heat flux in the non-boiling region. A third, Run No. C-5, was conducted with the flow guide installed to simulate the condition of heating the entire bottom surface of a container. Selected data are given in Appendix D-1., Figure 26 is a plot of the temperature difference between the heating surface and the water} Tw - T5, as a function of dimensionless acceleration a/g. The decreasein w-T5 for a given heat flux caused by the presence of the flow guide is noted, No attempt was made to investigate the phenomena further at this time, Figure 27 shows the temperature data for Run No. C-5 as taken, as a function of time, Thermocouple T6 was located in position 2 of Figure 25, and indicated a water temperature approximately 20F higher than T as a result of the effect of the Coriolis acceleration on the flow pattern. The level of water temperature was not controlled, The data were correlated with the standard Nusselt-GrashofPrandtl:modii and. are plotted in Figure 28, together with a sketch indicating the difference in configurations. Also included for reference is the correlation recommended by McAdams(l) for a horizontal heated plate facing upward. -70

-7135 30 RUN NO. C-2, q/A = 9840 BTU/HR-FT2 NO FLOW GUIDE 25 E- 20- \ RUN NO. C -5, lj 20. q/A ~ 10,220 BTU/HR-FT WITH FLOW GUIDE RUN NO. C-I, q/A = 4700 BTU/HR-FTt 10 NO FLOW GUIDE 0 I I I I I I I 0 4 8 12 16 20 ACCELERATION, a/g Figure 26. Plot of L~Tc versus Acceleration for Natural Convection Indicating the Effect of the Flow Guide.

230 210 22 ~ = 155 RPM ~ 190RPM 220RPM20 Tw -.Tsat 18' B7 220 200 16 14 u. o.. 0 RPM \ o,..... II 0 RPM 3 210 0190H 12 H~~~~~~~~~~~~~~~~~~~~~~~- H 10 RUN NO. C-5'q/A 10,220 BTU/HR-FT WITH FLOW GUIDE 200 180 - - - -- Tw — El —-T6 i i/-. Tw- T 190 1701 0 1 2 3 4 5 TIME-HOURS Figure 27. Run No. C-5. Temperature Data for Natural Convection as Taken.

-730 RUN NO. C-I, q/A 4,700 BTU/ HR-FT A RUN NO.C-2, q/A =9840 BTU/ HR-FT 1000 E3 RUN NO.C-5, q/A =10,220 BTU/HR-FT2 900 800 - 0 RPM 700 2 110 RPM 6 155 RPM 4 190 RPM 500 - 5 220 RPM 5 BOILING 4 400 3 2 300 3 z2!> —UN = 0.14 ( G x Pr 1 1001/ RUN NO. C-I,,, 0C-2 RUN NO. C-5 FLOW GUIDE \HEATER/ 4 5 6 7 8 9 1O9 2 3 4 5 6 7 8 9 10 Gr x Pr Figure 28. Correlation of Natural Convection with Acceleration Normal to Heating Surface.

B. Boiling For convenience and the sake of clarity, the data are presented subdivided according to the nominal heat flux at which the tests were conducted. Except in one case, the data is representative of at least two reproducible test runs. Because of the sensitivity of the boiling process to even small degrees of subcooling at the lower values of heat flux, it was necessary to vary the water temperature in order to determine when the liquid was saturated. One of the parameters used in boiling heat transfer is the difference between the heater surface temperature and the saturation temperature of the liquid, Tw - Tsat. Because the saturation temperature varies to a large relative degree in the body of the liquLid undergoing an acceleration, the saturation temperature referred to in this sense will always mean that at the heating surface. The saturation temperature of the liquid at the liquid thermocouples will also be indicated on certain plots, referring to the local saturation temperatures. 1. clA'S 10,000 Btu/hr-ft2. The temperature data for Run No. B-15 are given in Appendix D-2, The data were carried along with the second decimal place until the last step before rounding off. Figures 29(a-d) are plots of Tw - Tsat as a function of the measured water temperature T5 for the various accelerations. The convection parameter Tw - T5 is also plotted in order to indicate when boiling has ceased. With natural convection only, Tw - T5 should be

-7530.K NUMBERS INDICATE -s \SEQUENCE IN WHICH TW- TSAT 29 DATA WERE TAKEN t* LOSING STEAM 28 q/A 10,870 BTU/HR-FT2 27 26 3 0 RPM 25 _\/ a/g= 1 "24~~~~ ~4 iS 23 | R TSAT@ H.S. LL 22 5 TSAT@ H.S. 21 \1 20 20 2 0 2 2 2 2 2 10 218 110 7 65 RPM loo 3 0/g=1.95 3.9-' 16 -I 3 TSAT T 9P 15 -TSAT @ H.S. 200 201 04 205 206 207 208 209 210 211 212 213 WATER TEMPERATURE, T5 OF Figure 29a. Run No. B-15. Plot of Tw-Tsat and Tw-T5 vs. Water Temperature T5 at O and 65 RPM.

25 4^1 - J 2 110 RPM 1A —7~ ~ a/g=5.29 20 0 Tw_ Tsat Q Tw- T5 19 0 QUESTIONABLE HYSTERESIS 18 -.: CONVECTION 14 //<2 11 * LOSING STEAM 17 3:14 /; Tsat. T T ab Tsat. H.S. i- " q/A 10,870 BTU/HR-FT2 155 RPM I NUMBERS INDICATE SEQUENCE 3 | IN WHICH DATA WERE TAKEN! —- 3 17 _ __ __ -- 15 - 14- Tsat @ T 5 13 ~13| ~/ Tsat'@ HS 206 207 208 209 210 211 212 213 214 215 WATER TEMPERATURE T5,~F Figure 29b. Run No. B-15. Plot of Tw-Tsatand Tw-TS vs. Water Temperature T5 at 110 and 155 RPM.

-7716 98 15 14 13 7 -- -,4 —----- __ - * LO —SIG- -ST A 15 8 10 6* 149 atWAE TEPR R,1 TSAT @ T o. 13 TSAT @ H.S. s e pr1019R TW- TSAT 8 --- Tw T5 HYSTERESIS 7 14 * LOSING STEAM q/A =10,870 BTU/HR-FT2 ~~~~~~~6 ~~190 RPM 15(~~~~~~~~~a /g = 15.73 5 NUMBERS INDICATE SEQUENCE IN WHICH DATA WERE TAKEN I I I I I I I I I I I 205 206 207 208 209 210 211 212 213 214 215 WATER TEMPERATURE, T. ~F Figure 29c. Run No. B-15. Plot of Tw-Tsat and Tw-T5 vs. Water Temperature T5 at 190 RPM.

-7817 v,16 NUMBERS INDICATE SEQUENCE IN WHICH DATA WERE TAKEN 10 1~~~~~~~~~~5~~ ~15 14 21 6 22 A 319 2 19 5I 14 ". --. 2 7 13 10 ~14 A 2 13 L. 19 o 12 2. -T HS. LOSING SAT AM I10/A 10,870BTU/HR-F ~~~21~a/921.1 WATER TEMPERATSATURE, T -Figure 29d. Run No. B-15. Plot of T-Tst and T-T vs. 7Water Temperature T 0 RPM.QUESTIONABLE * LOSING STEAM 6 q/A ='10,870 BTU/HR-FT' 5 22 220 RPM a/g = 21.15 4 206 207 208 v 211 212 213 214 215 216 217 218 WATER TEMPERATURE, T5 OF Water Temperature T5 at 220 RPM.

-79independent of the water temperature except for changes in liquid properties, The saturation temperatures at the heater surface and at liquid thermocouple T5 are indicated for reference, The numbers assigned to each plotted point indicate the sequence in which the data were taken. Where gaps in the sequence appear, the points were omitted from the graph to prevent overcrowding, but are listed in Appendix D-2, Points represented by "SPy are somewhat questionable because inspection of the raw data shows that sufficient time may not have been permitted for steady-state conditions to be reached after changing the cooling water flow rate. Data points marked with an asterik may also be somewhat questionable as these were taken while steam was issuing from the vent tube, with an attendent possible increase in pressure. Referring to Figure 29a, for a/g = 1 as the subcooling decreases the wall temperature first increases, then decreases to a minimum just prior to the point where the water has reached its saturation temperature. For a given acceleration Tsat is constant and Tw - Tsat is a function of the heater surface temperature Tw onlyo As the subcooling decreases the contribution of the natural convection process to the total heat flux decreases. The surface temperature then increases to provide more active sites for boiling to compensate for this decrease, A further decrease in subcooling most likely causes a thicker or more highly superheated liquid boundary layer to be formed resulting in more rapid bubble growth rates and larger bubbles which in turn result in increased agitation of the boundary layer and hence lower surface temperatures.

~80 For a/g = 1 95 the subcooling was very small, but a distinct minimum of Tw - Tsat similar to that with a/g = 1 is notedq In Figure 29b, for a/g = 5%29, the wall temperature again goes through a maximum as the subcooling'is deareasqde only more pronounced Tw - Tsatthen levels off as the saturation temperature is reached. Upon subsequent increasing of the subcooling a -.hysteresis effe-t is observed, The largest subcooling for this particular acceleration was taken at point 16, whose value of Tw - T5 is 20o8~F. This is identical to the value of Tw T5 for Run No. C-5 in Figure 26, indicat~ing that no boiling is taking place. For a/g = 10.a47, the heater surface temperature does not go through.a maximum but levels off as the subcooling is decreased4 The value of Tw - T5 from Run No. C-5 in Figure 26 is included as an extension of the curve of Tw - T5 to indicate the proximity of point 1 to.complete non-boiling, Figure 29c for a/g =.1573 again shows a plateau of Tw - Tsat with.a decrease in subcooling4 A hysteresis effect also is noted, and the transition between natural convection and boiling is quite well defined. With a/g = 21,15 in Figure 29d, the transition between natural convection and boiling is nebulous, with only slight "'humps" in the curves present, Points 10 and ll1 are not true _stat "t-'ate values as the temperatures were changing very slowly with time, The cooling water flow was completely shut off, and the high acceleration suppressed boiling until a superheat of approximately 4.55~F at thermocouple T6 was reached4 The number of active boiling sites was nok;

-81doubt very small, else the large superheat at thermocouple T5 (points 13, 14, & 15) with a net loss of steam would not be possible. The constancy of Tw - Tsat over small ranges of water temperature T5 serves as an excellent indication that well established pool boiling is taking place. The upper curve of Figure 30 is a plot of these values of Tw - Tsat as a function of acceleration. The value at a/g = 21.15 is discretionary since no well defined change in Tw - Tsat is present. The lower curves are values of the temperature differences in the heater block, with the numbers indicating the sequence of readings. For a given acceleration these values are consistant. Thermocouple T6 in the water was located in position 3 of Figure 25. Figure 31 shows the difference in temperature indicated by the two water thermocouples T6 - T5 for the various values of T5 and acceleration, It is noted that the temperatures essentially were the same, This is also illustrated in Figure 32, where the temperatures measured between the heating surface and the liquid surface are plotted in profile for the conditions of pool boiling for thedifferent accelerations. The numbers correspond t-.:the data points of Figure 29 (b-d). The local saturation temperature lines also are indicated, *2.q/a 225,000 Btu/hr-ft2. The temperature data for Run No, B-9 are given in Appendix D-3. Plots of Tw - Tsat and Tw - T5 as a function of water temperature T5 for the various accelerations are given in Figure 33 (a,b). Several items should be noted from the curves: a, The range of subcooling possible is much less than at the lower flux, which is to be expected with the cooling system used. With

-8219 RUN NO. B -15 18 q/A -10,870 BTU/HR-FTz NUMBERS INDICATE 17 |\ 8SEQUENCE IN WHICH E8 | %_DATA WERE TAKEN o 16,:[ 15 10 -14 4 13 12 po)11 8 4 +1 2 0 10 +1p90 o2 G6 O 0 +1 50 0 110 o 9 2 06 10 0 2 N 8 o6 10 4 g 0 00 1-Il 0 L I I I I I,I I I 0 2 4 6 8 10 12 14 16 18 20 22 TOTAL ACCELLERATION, o/g Figure 30. Run No. B-15. Plot of Tw - Tsat and AT in Heater Block versus Acceleration with Pool Boiling.

q/A =10,870 BTU/HR-FT| THERMOCOUPLE LOCATIONS TSAT@ T6 TSAT @ T c=220 RPM +1 ___. -I Fg =150 RPM Ot0 fO =6 Ra a uo 0 _CN. cJ - 65 RPM O A O ( =0 RPM J 207 208 209 210 211 212 213 214 215 216 217 218

-84234 q /A =10870 BTU/HR-FFT NOTE: NUMBERS CORRESPOND TO DATA 232 POINTS OF FIGURE 29. 17 230 28 = 220 RPM 3 a 190RPM a/g 21.15 228 o/g= 15.73 218 15 216 7 ---— 3 214 - -- I o LL 212 21 ATE TSAT 210 a. w 230;I- ~~~~~6 228 3- i w=110 RPM 155RPM 4 0/g 5.2/ = 10.47 226 214' 4 5 -T~~~~~~~~9 212 -- 3 210 TA HEATER SURFACE HEATER SURFACEA WATER SURFACE WATER SURFACE Figure 52. Run No. B-15. Temperature Profile Between Heater and Water Surface for Various Accelerations.

-8526 25 W.110 RPM 135 RPM a/g =5.29 a/g = 7.84 24 22-8 - | \p Tst H.S. \ T set () H. 7 1-2 _- TW - T5 3/g 95 20 Z -0- Tw Tsot 65 RPM 23 214 TataiuAcl5 */ RPM 20 T T8 6 /gTw. T3.2ot — Tst S H/NUMBERS INDICATE SEQUENCE IW a/gI= IWI 19 I I. I. I I IN WHICH DATA WERE TAKEN 209 210 211 212 213 214 210 211 212 213 214 215 WATER TEMPERATURE T5,OF Figure 33a. Run No. B-9. Tw-Tsat and Tw-T5 vs. Water Temperature T5 at Various Accelerations.

-86-.2 5.25 F 190 RPM 220 RPM 24 a/g=15.73 o2 /g 21.15 23 2 Tsat( H.S.;sat~ T522 6| T0at @ H.S. 28~~~ 81~ t; Tsat ) T5\ 21 2 LL 20 7 N99 o9 2 11_L 22 24,450 BU H \ - FT 2 21- 19 6 6 7 L I WHICH DATA WERE TAKEN STEAM 211 212 213 214 215 212 213 214 215 216 217 Figure 33b. Run No. B-95 vs. Water TemperatureTat Variou- TTt 2w' Tsat 3 * LOSING STEAM 1 6 NUMBERS INDICATE QUENCE ature T5 at Various Accelerations.

the heat flux input maintained constant the temperature of the water is determined by the rate of heat transfer between the surface of the water and the cooling coils above. The mechanism is a combination of liquid-water vapor interchange and natural convection of the air in the vapor space. The contribution of the natural convection process is essentially independent of the heat flux input. At low fluxes it is the primary mechanism while at high fluxes it is minor. As the acceleration increases its influence increases, as evidenced by the increasing range of subcooling with acceleration in Figure 33. b. In the range of subcooling covered the heater surface temperature decreases with the decrease in subcooling,;indicating that not only are no new active bubble sites being formed, but the existing ones are becoming.more effective in decreasing the thermal resistance of the boundary layer at the heating surface. This can be explained qualitatively by considering the bubbles as turbulence promotors. The turbulence induced at a single active site will be a function of both the frequency of bubble formation and the bubble volume at departure. Taking the heat transfer per bubble as proportional to the turbulence, we have: a f Vb af a-o 3 (50) Ellion(8 ) has shown that the average growth velocity of a bubble is

-88independent of the degree of subcoolingo Neglecting any delay time between the departure of one bubble and the formation of the next, T constant (51) Since f = 1/vT Db ~ f = constant (52) which Jacob(2) had previously found at low values of heat fluxo If this relation holds true under accelerations greater than gravity, substitution into Equation (50) gives: b Db2 (53) Ellion(18) and Gunther(41) observed that the maximum bubble size increased with a decrease in subcooling. According to Equation (53) then the heat transfer per bubble increases with decreased subcooling. Since the total heat transfer rate is maintained constant, with no change in the number of nucleating sites the surface temperature will decrease, as is observed in Figure 33b, Co At the higher accelerations it is no longer possible to attain saturated conditions near the heater surface, A temperature profile between the heater and water surfaces for several accelerations, shown in Figure 34, aids in accounting for this. Thermocouple T6 was located - inch from the heating surface for this test run at position 1 of Figure 25, but on the basis of the previous and subsequent runs, the water temperature in a direction normal to the surface can be taken

-89NOTE: NUMBERS q/A= 24,450 BTU/HR-FT 236 _ CORRESPOND TO DATA 7 7 POINTS OF FIGURE 33. 234 3. 4 232 w =190 RPM /g = 15. 73 218 O/g = 21.15 216 TSAT 214 210 - 234 Fir 3.=110 RPM we 55RPM Wae Sra foVruAcr16 232 5 8o/g: 5.29 12-3 a/g 10.'47 230 216 T5 214 212 5 3 8 16 210 STSAT HEATER SURFACE Z HEATER SURFACE WATER SURFACE WATER SURFACE Figure 34. Run No. B-9. Temperature Profile between Heater and Water Surface for Various Accelerations.

-90as constant up to the mid-plane between the heater and water surface and most likely closer to the water surface. When the bubble detaches from the heater surface it passes first through a locally subcooled regiony and then a more or less superheated region. Because of the small curvature at the liquid surface, the water near the surface will not remain highly superheated, The acceleration and density differences will then drive this water down toward the heating surface where it will be subcooled with respect to local conditions. A similar situation was found to exist in a study of boiling heat transfer with (Ercury42) Here it was observed that the temperature throughout the bulk of the boiling mercury was essentially at the saturation temperature of the upper surface. Jacob(2 p6 623) also shows a similar disparity between local saturation and fluid temperatures with water. Figure 35 is the plot of the differences in the water temperature measured by T5 and T6, indicating that small differences do exist in a plane parallel to the heating surfaceo The minimum values of Tw - Tsat'in Figure 33 are plotted as a function of acceleration on Figure 36, along with the temperature differences in the heater -block. It may be that the small change in water temperature across the heater surface in part causes the variation of AT(1-3) with acceleration. However, the variations of LT(2-4) cannot be accounted fot., since this temperature difference is measured 7/16 inches below the heating surface. A minus value means that the centerline temperature is lower than that toward the periphery.

-91r+ TSAT O T5 \,-TSAT@ O!0.0 q/A 24,450 BTU/HR.FT -I oJ -- 220 RPM +1 — I c 190 RPM +1 o e0 00 0~0 -! o w=155 RPM 1 LLi i, 0 0 -I: -135 RPM +1 ~ —~ w -Ico 10 RPM -I I I I! 210 211 212 213 214 215 216 217 THERMOCOUPLE - w Various85 RPM LOCATIONS -I Wm 65 RPM +1'= 0 RPM Various Accelerations and Subcooling.

-9223 RUN NO. B- 9 q22 /A =24,450 BTU/HR-FT2 22 6 NUMBERS INDICATE SEQUENCE IN WHICH DATA WERE TAKEN 21 - 0 16 0 2e06 8 14 K) 3 l 1-5-7-11 28 ] Q6 08 012 014 010 0 -I I' 3I -3-5-9-11-15 3 20 -Q6 016 82 -I 2U E~L 3-9- I I I3-ls 82 08 012 A 010 I 0 2 4 6 8 10 12 14 16 18 20 22 TOTAL ACCELERATION a/g Figure 36. Run No. B-9. Plot of Tw-Tsat and T in Heater Block vs. Acceleration with Pool Boiling.

-933. q/A' 50,000 Btu/hr-ft2. The temperature data for Run No, B-14 are given in Appendix D-4. The range of subcooling possible with the attendant variations in Tw - Tsat were so small that the values of Tw - Tsat and water temperatures T5 and T6 are plotted directly as functions of acceleration in Figure 37, The.local saturation temperatures for the water thermocouples are also included, Thermocouple T6 was located in position 3 of Figure 25 for this particular test run, The value of Tw - Tsat at a/g = 1 varied somewhat during the course of the test, and to provide a more stable reference for comparing the effects of acceleration, the average changes of Tw - Tsat in coming up to and going down from a particular acceleration are shown in the lower curve of Figure 37. The variations of the differential temperatures in the heater block with acceleration are shown in Figure 38, and Figure 39 is a plot of the heater surface and water temperature profiles for several accelerations. 4) q/A a 75,000 Btu/hr-ft2 The temperature data for Run No, B-22 are given in Appendix ]D-5. In several earlier tests at this level of rate of heat flux, attempts were made to vary the water temperature by varying the cooling water flow rate, No changes.in water temperature could be detected, but undesireably large quantities of water were lost through the atmospheric vent, Subsequently no attempts were made to control the cooling water flow rate, As an extra precaution to prevent further subcooling, however, the drip plate described earlier was installed.

-9428 r28~~ | ~~~~RUN NO. B - 14 q/A = 48,800 BTU/HR-FT 27 o 13 II i 5 25 24 NUMBERS INDICATE SEQUENCE IN WHICH DATA WERE TAKEN. 217 216 215 w TI O-TI;.2I4 T.ot ~ T5 212 = 10 I-. - I I I I I I I I I 0 2 4 6 8 10 12 14 16 18 20 22 TOTAL ACCELERATION, a/g Figure 37. Run No. B-14. Plot of Tw-Tsat and Water Temperatures vs. Acceleration with Pool Boiling.

-95+2 1 -3-7-Ql -I _ I I-15 N~ 6 0 4 2 8 q/A 48800 BTU/HR-FT2 8 NUMBERS INDICATE SEQUENCE IN WHICH DATA WERE TAKEN NU t004 <36 2J. 5 4 02 O8 012 0 - 5 1-15 0 2 4 6 8 10 12 14 16 18 20 22 TOTAL ACCELERATION a/g Figure 38. Run No. B-14. Plot of AT in Heater Block vs. Acceleration.

-96244 W r | | q/A = 48,800 BTU/HR.FT2 242 240 - 238 _= 190 RPM W = 220 RPM a/g: 15.73 a/g = 21.15 218 - 216 SAT SAT 214.. 212 0 D 210 ___ a 240 W 3o= 110 RPM i = 1U55 RPM 238 ~ - at/9gr 5.29 a/g 10.47 236 216 5T6 T6 214 TSAT SAT 212 210 HEATER SURFACE' -HEATER SURFACE_ WATER SURFACEJ WATER SURFACE Figure 39. Run No. B-14. Temperature Profile between Heater and Water Surface for Various Accelerations.

-97At this heat flux level the heater surface temperature began to oscillate with a period of between one and five seconds. The values of Tw - Tsat listed in Appendix D-5 are the mean values, and the magnitude of the variations are listed in the column following. Figure 40 is a plot of Tw - Tsat, A(Tw - Tsat) as defined previously and the differential temperatures in the heater block versus total acceleration. The entire range of accelerations were covered twice.in succession, and.the data are quite reproducible in spite of a time interval of over six hours between given accelerations, Thermocouple T6 was located in position 4 of Figure 25. An earlier test, Run No, B-19, was conducted with thermocouple T6 in position 2, The heater surface temperatures were not acceptable because a number of pitnpoint shadows were observed on the heater surface at the conclusion of the test and the resistivity of the water had decreased to a value of 700,000 -Y -cm, indicating a slight degree of contamination. However, it is felt that the water temperatures are not materially affected, and are plotted on Figure 41 together with those of Run No, B-22 as a function of acceleration, It should be noted that the water temperatures in the direction leading the direction of rotation are higher than those lagging. This is in the same direction that the Coriolis force acts, and hence is an..indication that the flow pattern is being influenced somewhat by the Coriolis acceleration, Figure 42 is a plot of the heater surface and water temperature profiles for several representative accelerations.

-98RUN NO. B —22 q/A = 73,000 BTU/HR-FT 14-24 o27 1-3 -7-9-11-15-23 <~. 26, 8 2 22 0 - 17-25 F3 25 6-18 24 NUMBERS INDICATE SEQUENCE IN WHICH DATA WERE TAKEN +2 24 L. +1 I - _ 25 0 0 9 2-~~~1 4 N 8-6 8 20 0 -2 - I 4 0 10 -, 12 o 0 2-16 4 6 8 10 14-24 6 e 1 4 -2I 0 2 4 6 8 10 12 14 16 18 20 22 TOTAL ACCELERATION ao/g Figure 40. Run No. B-22. Plot of Tw-Tsat and AT in Heater Block vs. Acceleration with Pool Boiling.

-99218,.....__.. -__.21 8 | R U N N O. B - 2 2 2175 Vq/A: 73,000 BTU/HR-FT 217 QI. 216 Tsat ( H S' 215 < LOCATO O OT =: -I,6 A RUN NO. B- 19 217 L A T6 |q/A - 73,000 BTU/HR-FT,, 21 7 5 P f "-Ts IJ 215 8 10 12 14 1 21 4 211 4 1.. A l8 fr TwNUMBERS INDICATE SEQUENCE IN WHICH DATA WERE TAKEN 210.. 0 RUN NO. e -19 T Vq/A 73,000 BTU/HR-FT 217 216 1. 215 w,. 210 0 2 4 6 8 10 12 14 16 18 20 22 TOTAL ACCELERATION, / for Two Runs identical Except for Location04

-100244 q/A=73Q000 BTU/HR-FT| 242 240 W ) 190 RPM (0:220 RPM O/g = 15.73 Oa/g21.15 238 218 216 2A TSAT SAT 214 212 210, 240 Fiur 4 =110 RPM (=155 RPM 2 38 g52/g = 10.47 236 21 6 T T 6' — T5 -TSAT 216 214 212 208 tHEATER SURFACE HEATER SURFACE WATER SURFACE WATER SURFACE Figure 42. Run No. B-22. Temperature Profile Between Heater and Water Surface for Various Accelerations.

5. q/A 100,000 Btu/hr-ft2. The temperature data for Run No, B-21 are given in Appendix D-6. The oscillation of the heater surface temperatures were greater than for the previous heat flux, but decreased in magnitude with an.increase.in acceleration. The entire range of acceleration was again covered twice. Values of Tw - Tsat, A(Tw - Tsat) and water temperatures are shown in Figure 43 as a function of acceleration. Tw - Tsat at a/g = 1 changed 0.80F during the period of the test run and except for two points, the use of A(Tw - Tsat) resulted in a better representation of the data. Figure 44 is a plot of the temperature differences within the heater block, Data points 20 and 22 of AT(1-3) are somewhat high and point 20 of Wu(2-1) is somewhat low, These indicate,respectively, that the surface temperature at the center of the heater has increased and the heat transfer rate at the center had decreased, phenomena which would result from a decrease in the heat transfer coefficient at the liquid-solid interface in the center of the heater, It may be that the nucleate boiling sites were temporarily deactivated.in the vicinity of the center of the heater. Figure 45 is a plot of the surface and water temperature profiles for several accelerations, C. Overall Results 1. Boiling In spite of the precautions taken to provide consistant heater surface conditions and the purest water available for each test,

-102RUN NO. B- 21 3 |/A 99 500 BTU/HR-FT IL 200 U 230 2 12 - -8 5I 0 282 NUMBERS INDI CATE SEQUENCE 27- 1 1 1 1 1 IN IWHICHI DATAI WERE, TAKEN oU +2 30 _ 2 4 6 0 24 12 a -r ~~ 16 18 27 I -I I I I I I I I I I NOTE: VALUES OF T APPLY ~216 X~ ONLY TO POINTS 1-14 210 I-' -I- I eso w 214 a.. AT6AT H. 0 21 0 6 2 w A24 012 CL x ~16 18 I- 421 26 4 cr A6 24 w 26' < --— ~ Tr 12 28PT 210 0 2 4 6 8 10 12 14 16 18 20 22T Temperature vs. Acceleration with Pool Boiling.

-103LL o +1 0 22 24 Q 12 14 -164 8 20 0 802,,,, 26 Q I-I 01 o 18 +1 4 0 2; 0 0 0 22 10 2-16 18 20 ( 8 926 14 < 12 I I I.. I I I I I I I, 0 2 4 6 8 10 12 14 16 18 20 22 q/A- 99,500 BTU/HR-FT2 TOTAL ANUMBERS INDICATIONTE SEQUE/CE INgWHICH DATA WERE TAKEN vs. Acceleration. 6 B,0 24 20 8 0 2 4 6 8 I0 12 14 16 18 20 22 TOTAL ACCELERATION oa/g Figure 44. Run No. B-21. Plot of AT in Heater Block vs. Acceleration.

-104248 q/A 99,500 BTU/ HR.FT 246 244 2 w -190 RPM W= 220 RPM 242 _aa/g= 15.73 a/g = 21.15 24' 218 216 TSAT TSAT 214 ou. 212 -- 12 210 -- u I I I I I II a. 244 w 10) w 6. a/g 5.29 a/g = 10.47 240 238; l!T5 TS tt 5 T6 238 214 4 212 20 210 - NOTE: NUMBERS CORRESPOND TO DATA POINTS OF FIGURE 43 208 HEATER SURFACE HEATER SURFACE WATER SURFACE- WATER SURFACE - Figure 45. Run No. B-21. Temperature Profile between Heater and Water Surface for Various Accelerations.

-105it was found that some of the values of Tw - Tsat with the system under gravitational acceleration were not as reproducible as was desired, In order to provide a representative composite graph.showing the influence of acceleration on Tw - Tsat for the entire range of heat flux, the values of Tw - Tsat for a/g = 1 were plotted versus heat flux on logarithmic coordinates in Figure 46. The data for these points are given in Appendix D-7, More test runs are included than were utilized for acceleration data because the measurements were made early in the test period while they were still considered reliable, It is noted in Appendix D-7 that even though the water had become somewhat contaminated, for the most part the measurements of Tw - Tsat at the beginning of the test period and at the end were quite reproducible. Under the influence of acceleration, however, the data became scattered, The best straight line was drawn through the points of Figure 46, and the values of Tw - Tsat from the curve were taken as the base points in plotting Tw - Tsat as a function of acceleration in Figure 47 for all the heat fluxes. Figure 47 thus shows the net influence that acceleration normal to a heated surface has upon pool boiling to saturated water, or rather water as nearly saturated as was possible, Test runs in addition to those presented in the previous section are included to show the degree of reproducibility. The original data points as well as the modified values, due to shifting the base point a/g = 1, are given in Appenfix D-8. At the low heat flux levels, an increase in acceleration causes a decrease in Tw - Tsato The natural convection contribution to the

-106I0 0o0 0 7 6 5 1o x CY 4 I I 4 O O 10 20 30 40 TW TSAT F Figure 46. Plot of qA vs. Tw-Tsat for Boiling in Standard Gravitational Field.

-10731 30 A0. A 29 q/A: 99,500 BTU/HR-FT' 0 RUN NO.B-21 I,0 Ai -20 28 27 q/A= 73,000 BTU/HR-FT RUN NO. B-22 26 25 23 0 23~ /A-48,800 BTU/HR-FT{ 0 RUN NO. B12 22 A RUN NO. B-8 2 1 q/A -24450 BTU/HR- FT. A O B-9 0~ El "t t B-I 20 19'7 q/A 10,870 BTU/HR- FT2 0{ RUN NO. B-15 2 2 1 6 15 A 14 0 2 4 6 8 10 12 14 16 18 20 22 TOTAL ACCELERATION -a/g Figure 47. Influence of Acceleration on Tw-Tsat with Pool Boiling to Saturated Water.

-o108 total heat transfer increases in magnitude, in effect depriving the boiling process of part of the heat fluxo Fewer active nucleating sites are thus needed and the heater surface temperature decreases. It is postulated that the accelerationr acts to decrease the size of the bubbles at departure from the heating surface with an attendant decrease in agitation and heat transfer rate per bubble~ At higher values of heat flux the increase in natural convection is not sufficient to counteract this decrease and the heater surface temperature rises to provide additional nucleating centers. At the nominal flux q/A = 100 000 Btu/hr-ft2, the surface temperature was observed to oscillate approximately ~ 0.35F with a period varying from one to five seconds, ostensibly due to the irregular boiling taking place. Upon acceleration, the magnitude of the oscillation should decrease since smaller bubbles would cause smaller irregularities. At a/g = 21l5 the oscillation was approximately ~ 0.10F with a period of from one to three seconds. Figure 48 is a plot of the difference between the heater surface and the water temperature, Tw - T5, corresponding to the saturated pool boiling data of Figure 47. Use of this data will be made in the next chapter in determining the contribution of natural convection to the total heat fluxo The increase in T'w - T5 with acceleration at the higher values of heat fl.ux results from the increase in subcooling. 2, Miscellaneous Observations a, Variation of AT in heater block In analyzing the data it is observed that in several cases the valu.es of the temperature differences in the heater block

-10938 36 q/A =99,500 BTU/HR-FT z 34 RUN NO B-21 32 30 24 q /A =7324,45 000 BTU/HR-FTz 20 - 18 26 q/A =48,800 BTU/HR-FT2 ~~16 -p~ ~ ~~ ORUN NO B- 14 14 12 I I I I I I I 0 2 4 6 8 10 12 14 16 18 20 22 TOTAL ACCELERATION - a/g Figure 48. Plot of Tw-T5 vs. Acceleration with Pool Boiling to Saturated Water.

-110changed somewhat with acceleration, possibly due to non-uniform boiling conditions at the heater surface. It was at first believed to be due to the flow pattern in the water resulting from the Coriolis acceleration, but the position of the thermocouples relative to the direction of rotation was the same for all test runs, and no consistent trend is present either with respect to heat flux nor acceleration. AT(1-3) provides a relative measure of the uniformity of the heater surface temperature, if not absolute because of the uncertainty -of the exact location of the thermocouples below the surface. Where test runs were duplicated, in some cases the values of AT(1-3) would be reproduced and in other cases they would not. The values of Tw - Tsatj however, were consistent. b4 Calculation of heat transfer rate from temperature gradients in the heater The values of AT(2-1), which provide an indication of the temperature gradient at the center of the heater, were averaged for the non-rotating and the rotating conditions for each level of heat flux, The rates of heat transfer were then calculated for each of these cases and are fisted in Table VII along with the percent variation from the heat flux calculated from the measurement of power input. It is recalled that the inherent uncertainty of heat flux calculated from AT(2-1) was determined to be + 15%. The larger difference at the low heat flux values may be due to the heat loss through the heater skirt, but the percentage difference decreases with rotation, whereas the heat l.oss through the skirt should increase with

TABLE VII COMPARISON OF HEAT FLUX CALCULATED FROM HEATER TEMPERATURE GRADIENT WITH THAT CALCULATED FROM POWER MEASUREMENT Non-Rotating Rotating Average q/A %. Run No. (Power) AT(2-1) q/A Difference ar(2-1) q/A Difference LT(2-1) q/A Difference B-15 10,870 1.06 7,600 -30.0 1.26 9,040 -16.8 1,16 8,190 -24,6 B-7 10,580 096 6,880 -35.0 1.17 8,400 -20.6 1.07 7,610 -28.0 B-8 24,450 2.45 17,600 -28.0 3.19 22,900 - 6.3 2.82 20,500 -16.1 B-9 24,450 3.16 22,600 - 7.5 B-11 24,451 3.14 22,500 - 8.0 B-12 48,800 5.75 41,200 -15.6 6,18 44,300 - 9.2 5.97 42,700 -12.5 B-14 48,800oo 5.90 42,300 -13.3 6,13 44,000 - 9.8 6.11 43,000 -11.9 B-22 73,000 10.27 10,23 10.25 73,600 + 0.8 B-20 99,500 13,65 97,900 - 1.6 13.33 95,500 - 440 13,49 97,000 - 2,5 B-21 99,500 13,42 96,32 - 4 95,600 - 2 123 95,000 - 45 132 95,600 _ ~ ~ ~ ~ ~ ~.-0,.1.2.5OO 33

-112increased acceleration. For Run No. B-15, it is noted in Figure 30 that A&T(2-1) changes very little with acceleration. At the low heat flux levels a small change in AT(2-1) reflects in a large change of calculated flux cn Effect of Subcooling At the heat flux levels q/A = 10,000 and 25O000 Btu/hr-ft2 the sensitivity of boiling heat transfer to subcooling under the action of acceleration was demonstrated~ At higher flux levels it was not possible to control the degree of subcooling with pool boiling in the system. In the series of tests reported here the water depth 1 was maintained constant at 2- inches, By decreasing the depth of the water, presumably the subcooling would have decreased at the higher accelerations Whether this would modify the curves of Figure 47 is still open to question* Figure 49 is a plot of the subcooling existing in the water at thermocouple T5 as a function of acceleration for the various heat flux levels, The subcooling is essentially constant for q/A > 50,000 Btu/hr-ft2 and is directly proportional to the total acceleration~ For constant water depth, increasing the acceleration increases the pressure at the heating surface, but the effect on the bollinrg characteristics is considered negligible in the acceleration range covered. d. Effect of Coriolis Forces The measurement of water temperatures at various locations within the system has aindicated that a definite non-symmetrical temperature pattern is established in the water as a result of accelerating the system by rotation. This was attributed to the action of the

-113SUBCOOLING =(TSAT. O T5)-T5 D q/A 24450 BTU/HR-FT2 0 is = 48800 of "3 I 0 0 2 4 6 8 10 12 14 16 18 20 22 TOTAL ACCELERATION, a/g Figure 49. Subcooling at Thermocouple T5 as a Function of Heat Flux and Acceleration. a (CORIOLIS) 2 o Vr Figure 50. Example of Coriolis Acceleration on a Particle Constrained to Move Radially on a Rotating Bar.

Coriolis acceleration, Figure 50 demonstrates the direction in which the acceleration acts on a particle which is constrained to move radially along a bar rotating about one end. A fluid such as water moving radially because of convection is not so constrained and the acceleration will be in the opposite direction with respect to a rotating reference line. The question of interest is; how does this effect the results obtained? Although a rigorous calculation of the component of water velocity parallel to the heating surface is not possible with the -information available, an upper limit can be approximated with restrictive assumptions, This is done in Appendix C for a convection case, Run No, C-5, with q/A = 10,220 Btu/hr-ft2 and acceleration a/g = 21o15., The maximupn velocity possible with the assumptions made is 0,19 ft/sec, and most likely is less than this by a factor of two or three, The.e'fect of this velocity is, therefore, considered negligible. The agreement of the natural convection (non-boiling) data without the flow guide in Figure 28 would appear to lend further evidence to this conclusion, With boiling, the influence of a water velocity parallel to the heating surface should be even smaller, as the intense turbulence caused by the bubbles tends to isolate the heating surface from any velocity effect0 3, Convection Correlation In Figure 28 the results of natural convection tests without the flow guide were.correlated well by Nu = 0,14 (Gr Prl/ (54)

-115However, it was noted that the flow guide caused an increase in the value of the Nusselt number for Run No. C-5. Additional convection data exist from the low heat flux boiling tests where the subcooling was increased until boiling ceased, Values of the heater surface and water temperatures were taken where Tw - T5 became independent of the water temperature, in Figure 29 for example, and the Nusselt-Giashof-Prandtl mrroduli were calculated. These are plotted in Figure 51 with the best straight line drawn through the points. With increasing acceleration, the Nusselt number for Run No. C-5 rises somewhat above the line, This is believed'due to the increase in heat loss through the heater skirt as a.result of the large subcooling existing in the water. In the remaining tests plotted, the water temperature was not far below the boiling point. The correlation which best describes the natural convective process within the flow guide.is given by the equation of the line in Figure 51: Nu =.0505 (Gr Pr)9 (55)

1000 900 0 RUN NO C-5 q/A = 10,220 BTU/HR-FT2 800 A,,, B -7 = 10,870 A. 700 - 0,,, B-15,I 10,870 -600 + " B-II I =24,450,, 600 I. 0 RPM 500 2. 110 is 5 3. 155 4 15 4 400 4. 190,, 3 5. 220,, z 300 Nu =0.0505 (Gr x Pr) 0.396 200 100 I 1O9 1.5 2 3 4 5 6 7 8 9 1010 1.5 2 3 4 Gr x Pr Figure 51. Correlation of Natural Convection with Flow Guide.

VI. ANALYSIS A. General Because of the limited quantity of data obtained and the relatively small range of conditions covered, it is believed -that it would be premature to attempt a direct correlation of the data, either with existing relationships or with ones which might be derived. Further, none of the correlations in the literature which have been examined (e.g. 12,29,30,44,47,48) appear capable of following the trends shown in Figure 47, which result from acceleration of the boiling system. This may be due in part to the inadequacy and treatment of the models used to describe heat transfer by boiling, and to the neglecting of the nonboiling convection contribution. The test of the understanding of the boiling phenomenon is the ability to predict the effect of the major variables governing the process. For many applications acceleration is not a significant variable, but it must be included to complete the description of boiling since it makes the natural convective process possible. For a given geometry, say a flat horizontal heated surface, the rate of heat transfer by nonboiling natural convection is dependent upon LG and properties of the fluid. When the temperature of the surface exceeds the saturation temperature of the fluid by a certain amount the liquid generally vaporizes in discrete bubbles at the heating surface. Accompanying this is a sharp increase in heat transfer rate for a given increase in AG. Aside from the higher AS, which will alter the properties of the fluid near the heating surface somewhat, the primary difference -117

-118between the two processes is the formation of the bubbles which serve (2) to highly agitate the fluid, as was proposed by Jacob and others. Many of the correlations proposed for nucleate boiling initially consider the action of the bubbles in promoting the heat transfer. Because of the inability of adequately describing this action, it is always necessary to revert to experiment to provide a relationship for Prediction purposes, An imposed acceleration changes the buoyant forces within the fluid being heated, and any analysis of the effect of acceleration on boiling heat transfer is really an analysis of the role of the bubbles. What will be attempted here is a presentation of the various interrelationships of bubble characteristics in boiling which have been given in the literature, together with limited extensions based on the observations made as a result of the experiments reported here. B, Bubble Relationships in Boiling (49) Hirano and Nishikawa used the simple representation of Figure 52 to point out the overall relationships involved in boiling heat transfer for the case where heat flux rate is the forcing function, as with an electrically heated surface or in a nuclear reactor. Figure 53 is a detailed extension of this, and serves to show the interdependence and complexity of the elements which are customarily considered in the boiling process. As yet little is known about the relations between the majority of these elements which must be determined before boiling can be described completely.

q/A Figure 52. Representation of Hirano and Nishikawa (9) for Boiling Heat Transfer. q/A LEVEL AND NO OF GROWTH DISTRIBUTION ~~~GROGWT~ ~IN SiV \E r -E // BOUNDARYI LAYER FLUID SURFACE PROPERTIES CONDITION ACCELE RATION; FORCED CONVENTION; Figure 53. Interdependence and Complexity of Boiling Elements.

-120'For a given system and conditions the following relationship was found to hold true over a large part of the nucleate boiling range: q/A = C(AG)n (56) where the interaction of the bubble elements in the lower part of Figure 53 is manifested by the values of' C and n. Another form often used is q/A = h - A (57) where these interactions are now represented in the value of the convective heat transfer coefficient ho The expression developed by Fritz(14) for the maximum volume of a bubble at departure from a surface under equilibrium conditions is given in Equation (3), and good agreement was found with measurements at low values of heat flux (51) Jacob(2) experimentally determined the relation between the frequency of formation and bubble diameter at departure for saturated boiling from a horizontal surface and found f Db = Constant (58) This was also found to hold true in a statistical sense with the presence of contamination and surface active agents in water(5 51) The data were obtained from high speed photographs at low values of heat flux only, out (52) of necessity. For methanol, Perkins and Westwater determined that f, Db and hence f Db remained constant up to approximately 80o of the peak heat flux, after which they increased. However, a horizontal round tube was used for the heating surface, and no information was given as to where the measurements were made, Recent experimental work(53) appears to indicate that the diameter of the bubble at departure Db decreases

-121with an increase in heat flux rate, By using the variation in index of refraction of light in water with temperature, the thickness of a "boundary layer" near the horizontal heating surface with convection and pool boiling to saturated water was (54) measured at low values of heat flux. The data presented is shown in Figure 54. With values of heat flux less than 3300 Btu/hr-ft2 it was observed that ha 1 (59) which is a characteristic of laminar flow. In the turbulent region, where boiling is taking place the boundary layer thickness decreates-imore;:,s1ow(ly with! increaSing, heat flux, Treshchov(55) measured temperatures to within 0.0025 inches of the heating surface with high rates of heat transfer to nearly saturated water with forced convection by means of a microthermocouple, The results are shown in Figure 55, It is noted that at a distance of 0.0059 inches the temperature of the water was uninfluenced by a large increase in heat flux, By varying the surface tension with surface active agents(5), Equations (60), (61), and (62) were obtained from experimental data at low values of heat flux with saturated boiling from a flat horizontal heating surface (Maximum q/A % 25,000 Btq/hr-ft2). These relations were found to be independent, and were also found to be valid at reduced (56), pressures

-1221500 FLUID -DISTILLED WATER 1000 900 o- 800 1- 700 I 600 500 I400 - 300 q/A < 3300 BTU/HR -FT 200.015.02.03.04.05.06.07.08.09. 8 -INCHES Figure 54. Thickness of Boundary Layer on a Horizontal Heating Surface as Measured by Hirano and Nishikawa(54) Using Refraction Method. 320, FLUID- SATURATED WATER PRESSURE - 1.2 ATM. 310 VELOCITY - 13.1 FT/SEC. 300 0 El q/A O0.48x 106 BTU/HR.FT2 0 A = 1.07 xIO6 " 290 V\\ = 1.19 x106 0 z 1.33 x106 280 o' 270: 260 a, 250 240 230 220 210 I I 0 I 2 3 4 5 6 7 DISTANCE FROM HEATING SURFACE,INCHESxlO Figure 55. Effect of Heat Flux on Fluid Temperature Near Heating Surface with Forced Convection Boiling. Due to Treschov(55).

-125h cu(N/A Db3f)"13 (6o) Le cL(N/A)-1/6 (q/A)2/3 (61) q/A a(N/A)l/2 (62) By substituting Equatiors (60) and (61) in Equation (57), solving for q/A and equating to Equation (62), it is observed that the quantity Db3f is predicted to be independent of heat flux. A consequence of this will be enlarged upon later. Equations (61) and (62) were also valid for different degrees of roughness of the heating surface and for different degrees of contamination of the water, the main effect of these being to change the proportionality constant. Equation (62) has been determined (53) to hold true up to heat fluxes near the burnout heat flux. As pointed out earlier, Corty and Foust(9) have shown some effects of the surface roughness on nucleate boiling, The bubble growth rate problem has been well investigated''13) and from the agreement with experiment the solutions appear to include the important governing parameters for conditions with no subcoolingo The acceleration of a boiling systmn may have a direct influence on the maximum attainable bubble size and the temperature and thickness of the superheated liquid layer near the heating surface, as indicated in Figure 53, and an indirect effect on the other elements. C. Some Observations on Boiling at Large Values of Heat Flux Equation (56) has been well established as valid over a large part of the nucleate boiling range. However, as the peak or maximum heat flux attainable with nucleate boiling is approached, distinct

_l24departures from this relationship have been observed, some examples of which are given in Figure 56~ The point of this departure has been designated the DNB point (Departure from Nucleate Boiling) by one group (57) Defining i by (q/A) DNB (63) (q/A) Peak it is noted in Figure 56 that' ranges from 0.52 to more than 0.78, depending on the conditions under which boiling is taking place. The inequalities are necessary because of the gaps between data points. The significance of the region between the DNB point and the peak heat flux has been given little consideration in the literature. As a bubble grows because of evaporation of liquid from the superheated boundary layer, the surrounding liquid is impelled away from the nucleating center. In so doing, part of the liquid is moved directly away from the heating surface, and the remainder moves more or less parallel to it for a short distance. It is the rapid movement and displacement of the liquid including ejection of the bubble from the surface and liquid inflow behind, generally termed "agitation", which results in the high heat transfer rates associated with boiling. For a completely nonwetted sur(58) face Averin found that the peak heat flux was 8% of that for a wellwetted surface, and concludes that in boiling water approximately 8% of the heat transmitted by the liquid is expended on vapor formation at the surface. Rohsenow and Clark(7) estimated that the energy transport by the bubbles as "latent heat" could only account for 10% of the boiling heat transfer.

-1253 1 REF CONDITIONS IY(EQU. 63) O 33 VORTEX FLOW > 78.3% A 53 N - SALT SOL'N |< 56.0% X 18 V 1.I ft/sec. Subc = 50~F > 53.0% *) I to I "I I " " Subc aIOOOF 2 64.0% O 63 P -o psig > 63.0% *B I* P- 50 psig 2 63.0% o10 - e BURNOUT 8 06 i- / 3. 2 co I0 20 30 40 50 607080 100 200 300 400 A 8 -~F Figure 56. Representative Plots of q/A vs AS Near Peak Heat Flux.

At low values of boiling heat flux, where the nucleating sites are relatively far apart, the heat transfer rate should be directly proportional to the number of active sites with saturated pool boiling, as was noted by Jacob.(2) Zysina-Molozhen(59) counted the number of nucleating centers with boiling water and aqueous solutions of NaCl and Glycerol for a variety of surfaces at pressures of 1-5 atmospheres, up to a maximum heat flux of 22,000 Btu/,hr-ft2. The-maximum number of active sites counted was 2780 per square foot, For water, the actual number of sites at a given flux depended on the heating surface used, but a linearity between heat flux and the number of sites existed, As the number of nucleating centers increases, interactions between bubbles will occur long before they are close enough to make contact with each other, because of the displacement of the liquid, The high speed photographs of Griffith (6) can be interpreted as indicating the type of interaction which can occur from. a single nucleating site. As a result of the interactions, the bubbles no longer are able to grow to a large size and each site becomes less effective as a fluid agitator. Hence, a proportionately larger number of sites are required, perhaps explaining the experimental observation given by Equation (62), i.e. N/A a(q/A)2 With an increase in heat flux, it is believed that a limit is reached beyond which no further agitation of the liquid by bubbles forming at discrete sites is possible, The maximum bubble size decreases with increasing heat flux because of the increasing interference from the surrounding bubbles, and it is further believed that a minimum size exists,

-127dictated by the growth dynamics. The combination of this minimum size and the hydrodynamic interaction between bubbles results in the limiting effectiveness of liquid agitation owing to bubbles originating at specific sites. This limit is tendered as being the point at which the departure from Equation (56) takes place, A minimum spacing between bubbles exists, but the surface is still far from complete vapor coverage. If an attempt is made to increase the heat flux rate beyond this point, the majority of the additional heat flux results directly in vapor generation near the heating surface, a (44) possibility mentioned by Chang and Snyder. Bubbles no longer form at preferred active sites but occur randomly and perhaps even at very short distances away from the heating surface because of the intense turbulence, The formation of bubbles away from a surface was observed in a study of cavitation64), which is recognized as similar in many aspects to boilingo (0 ) The formation of bubbles at points away from the "preferred" or active sites likewise requires proportionately higher liquid superheats0 If forming at random on the surface much higher temperatures are required to continually activate new centers, and if forming away from the surface higher superheats are required for (43) thermodynamic reasons, resulting in the bend in the plots of q/A vs AG as in Figure 56~ This picture is also in keeping with the experimental observations of Gaertner and Westwater(53) where no active sites could be discerned at heat fluxes above the DNB point. At the maximum heat flux the presence of large amounts of vapor near or at the heating surface begins to impede the transfer of heat to the liquid, resulting in a decrease in heat transfer with

-128further increase in surface temperature. A vapor coverage of 50% at this condition has been mentioned by several workers('6) although it may be less, depending.on geometry and test conditions. At the beginning of stable film boiling the vapor coverage is assumed to be 100%. The effect of subcooling and forced convection is to decrease the maximum size of the bubbles,(184162) and should extend the heat flux.at which DNB takes place because a much larger number of active sites can be sustained before the limit..of bubble interaction is reached. The value of * as defined by Equation (63) may increase or decrease, depending on the relative change in (q/A)peak due to subcooling and forced convection. On the basis of the mechanism described above it might be expected that acceleration of low magnitudes will have very little influence on the heat flux at DNBo At this point the bubbles have already attained a minimum size. The action of the acceleration is to increase the buoyant force on the bubble, which must have some finite size before this force can act. If the growth rate to the minimum size at DNB is large, sufficient time may not be available for the buoyant forces to influence the motion of the bubble. Relations have been developed(5.'3l44) which predict an increase in peak heat flux with acceleration. This appears reasonable from consideration of the peak heat flux-as the condition where the large quantity of vapor present near the heating surface impedes the movement of liquid toward the surface, An acceleration will aid

in the removal of vapor from the vicinity of the heating surface, permitting a higher heat flux. D. The Influence of Acceleration on Boiling Area With well developed pool boiling in a standard gravitational field, the contribution of nonboiling convection to the total heat transfer is so small that generally it is not separated from the total heat flux in analysis. With the application of an acceleration nonboiling convection can become quite appreciable, as was indicated in the test results. In order to describe the total heat transfer, it is necessary to determine the relative contribution of the bubbles and that not resulting from the bubbles. This was accomplished in effect by Chang(30) by considering the rate of momentum exchange in the liquid near the heating surface to consist of the sum of the molecular diffusivity and the eddy diffusivity. The eddy diffusivity provided a measure of the turbulence of nonboiling convection and agitation due to the bubbles. Another method is to consider the convection and boiling contributions as separable and to weight them in some manner. This attack will be attempted here. Accordingly, the total heat transfer rate qt is qt= q 9+ (64) qt qc qB where qc is that portion due to convection and qB is that portion due to boiling, which includes the agitation induced by the bubble action. Dividing by the total area: t/A = qc/^ +qB/^ (65)

-130. If AB represents a time-averaged area projected on the heater surface in which the action of the bubbles is felt and AC is the area in which natural convection is effective: (q/A)t = + Ac At AB At (q/A)c(l-y) + (q/A)B 7 (66) where y = AB/AT AB + Ac =At (q/A)c = qc/Ac (q/A)B = /AB AB is not the heater surface area covered by the bubbles, but is considered differently. It might be designated as an "area of influence." Figure 57 is an illustration of this concept. According to the previous section the agitation produced by bubbles formed at active sites on the heating surface attains a maximum value where departure from Equation (56) takes place - the DNB point. No surface area remains in which convection can be effective, and from Equation (66) and (67) (q/A)t = (q/A)B It is desired now to obtain values of y as a function of heat flux and acceleration. Solving Equation (66) for y (q/A)t - (q/A)c (68) (q/A)B (q/A) (q/A)t is the total heat flux measured in the experiments, but means of evaluating (q/A) and (q/A)R must be sought. (q/A)c is considered first.

-131TIME AVERAGED PROJECTED BUBBLE AREA "AREA OF INFLUENCE" OF BUBBLES ACTIVE SITES FREE CONVECTION AREA,AC SECTION OF HEATING SURFACE Figure 57. Illustration of Area of Influence of Bubble with Boiling Heat Transfer. ILt Itlt HEATER a.WITH NO FLOW GUIDE b. WITH FLOW GUIDE Figure 58. Convective Fluid Flow Pattern in the Experimental System with and without Flow Guide.

Figure 28 demonstrated agreement between experimental nonboiling convection data under acceleration with.the correlation Equation (54) recommended by McAdams(1) for a horizontal heated plate facing upward. With the installation of a flow guide it was noted that for a given heat flux the temperature difference STc decreased, or the convective heat transfer coefficient h increased. This is attributed to increased turbulence in the boundary layer resulting from the disruptance of the gross convection pattern present without the flow guide, as illustrated in Figure 58. It is felt that the convective pattern of Figure 58b is more analogous to the type of convection which will exist in the free convection areas outside the "area of influence" of the bubbles, in Figure 57. Hence, the expression which correlates this type of data, Equation (55), is selected for determining (q/A)c in Equation (68). The turbulence induced by an individual bubble is a function of the maximum volume of the bubble and the frequency of formation. Taking the heat transfer contribution of each bubble as proportional to the turbulence and approximating the bubble shape by a sphere: qb = ClDb3f (69) If Ab is the area of a bubble projected on the heating surface, Aba Db 2, and Ab cf 2d (70) On the basis of the discussion associated with Equation (58) f Db is taken as constant for pool boiling of a saturated liquid, and is also

-133assumed to be valid under the influence of acceleration. Equation (70) then reduces to qb/A = Constant (71) for a given acceleration. It was also determined from Equations (60-62) that Db f is independent of heat fluxo For a given total heat transfer area qB = Nqb (72) If the "area iof influence" of the bubbles is proportional to their area projected on the heating surface AB = C N Ab (73) Dividing Equation (72) by Equation (73) B = 1 b (74) AB C Ab then by Equation (71) (q/A)B = Constant (75) It is necessary to determine what this constant is as a function of acceleration. Equation (67) stated that at the DNB point that (q/A)t = (q/A)Bo No expressions are available which give this value of heat flux. As a first approximation it will be taken as equal to the peak heat flux (q/A)po As discussed in the introduction, a number of workers have derived expressions for (q/A)p similar to Equation (11) showing an effect of acceleration. To the writer's knowledge these have not been confirmed by experiment for accelerating systems, With K2 = 0,146,Equation (11) appears to best represent experimental values of peak heat flux with pool boiling of saturated water. For the test conditions encountered here this reduces to:

-134(q/A)p =.02 x 106 p 1/2(a/g)/4 Btu/Hr-Ft2 (76) This is plotted in Figure 59 showing the influence of acceleration, Also included is the curve of Equation (10) showing the effect of a different exponent on acceleration. Values of y were calculated from Equation (68) with experimental values of (q/A)t, (q/A)c determined from Equation (55) with ATc taken from Figure 48, and (q/A)B calculated as (q/A)p from Equation (76), These are plotted in Figure 60 as a function of acceleration for the various total heat flux, and cross-plotted on.logarithmic coordinates in Figure 61. The data are correlated by the expression: y = e (q/A)t (77) where r = 14.38 + 0o3055 (a/g) s = 1.100 + 0o02273 (a/g) An interesting observation can be made from Figure 61, By extrapolating the best straight lines through the points of constant acceleration up to the value of y = 1, it is noted that at y = 1 the lines intersect at approximately common values of heat flux for accelerations a/g > 5, rather than attaining the peak heat flux values used for determining y, This states in effect that the heat flux at y = 1, the DNB point, will be independent of the acceleration, a conclusion mentioned in the previous section, The significance of this has not yet been evaluated~

-1351.3 T =210.5 ~F FOR a/g I SAT 1.2 1.1 LI | X EQUATION 10 / I O -/)p - F(/g.9'o.8 i I.5 I 0 2 4 6 8 10 12 14 16 18 20 22 TOTAL AC C ELE R AT ION o/g Figure 59. Effect of Acceleration on Peak Heat Flux with Pool Boiling. with Pool Boiling.

-136-.30.28 0 (q/A)T = 10,870 BTU/HR-FT A's" = 24, 450.26 E " 48,800 X " = 73,000.24 0 " = 99,500.22.20.18.16.14.12.08x.06.04.02 0 2 4 6 8 10 12 14 16 18 20 22 TOTAL ACCELERATION, a/g Figure 60. Calculated Values of 7 as a Function of Total Heat Flux- and Acceleration.

-1371.000.900.800.700.600.500 (q/A)p.400.300.200.100.090 0 o/g = 1.00.080.070 A = 5.29.060' 3 = 10.47.050. X "= 15.73.040 =~:21.15.030.020 0.010.009.008.007.006.005.004.003.002.001 o, 4 2 3 4 56 789 2 3 4 56789 i0 10. 10 2 (q/A)T- BTU/HR-FT Figure 61. Cross Plot of Figure 60.

E The Influence of Acceleration on the Number of Nucleating Sites Nucleation is a phenomenon which depends only upon the heating surface conditions and fluid properties, as can be concluded from the discussion in the introduction. Hence it should be affected by acceleration only insofar as the heating surface conditions and fluid properties are themselves influenced. For a given surface-liquid combination and with only small changes in pressure at the heating surface due to acceleration, the number of nucleating sites will be a function of the heating surface temperature onlyo If the dependence of the number of active sites on AS is known, the experimental values of A\ can be used to calculate the number of sites for the various heat fluxes and accelerations, From Equation (56), the relation between heat flux and temperature difference for a given boiling system is: 4,/A - Cl('b)n (78) Likewise, for a given system Equation (62) relates the number of active sites to heat flux: q/A = C2(N/A)1/2 (79) Equating Equations (78) and (79) and rearranging: N/A = C3() 2n (80) For the system used in the experiments reported here n = 5 from Figure 46, To determine C3, the number of active sites must be known for at least one condition of q/A and A9. Figure 62 is a view of the heating surface after boiling.at q/A = 99,500 Btu/hr-ft2 had taken place with slightly contaminated water

-139Figure 62. View of Heater Surface After Test Run With Slightly Contaminated Water at Flux q/A = 99,500 Btu/hr-ft2. (The bare portion on the left side was wiped lightly with a finger. Tw-Tsat = 295~F, N/A = 11,100 active sites per square foot.)

at standard gravitational acceleration for a considerable length of time. The contamination came from ceramic thermocouple tubes which were used for the measurement of water temperature during early test runs. This particular type of contamination appeared to have no effect on the temperature differences with normal gravity, and so it is possible the number of sites shown is representative of the number that would be present if no contamination existed. However, contamination from the ceramic tubes did result in nonreproducibility of data between various tests with the system under acceleration and all conditions otherwise identical. Stainless steel thermocouple protection tubes previously described were then substituted for water temperature measurements. The number of sites in Figure 62 was counted in the central portion of the heating surface and determined to be N/A = 11,100 per square foot with AG = 2905~F, This is quite close to the value of 9,000 per square foot taken from the mean curve of (53) Gaertner and Westwater for the same heat flux, and within the experimental scatter. Substituting these values into Equation (80) we have: N/A = 0.221 x 1010 x (AG) (81) for the system used here. Now taking values of AS from Figure 47 it is possible to determine the number of active sites as a function of heat flux and acceleration. These are plotted in Figure 63 and are similar in form to Figure 47, as is to be expected with a logarithmic plot.

-141q/A=99,500 BTU/HR-FT I0 q/A=73,000 BTU/HR-FT 3 q/A= 48,800 BTU/ HR-FT w IF l l,, | A\ r -4q/A -24,450 BTU/HR-FTl 0 1 z 2P, 10 I z q/A: 10,870 BTU/HR-FT2 10 0 2 4 6 8 10 12 14 16 18 20 22 TOTAL ACCELERATION o/g Figure 63. Number of Nucleating Sites as a Function of Heat Flux and Acceleration.

-142.At the lower values of heat flux acceleration increases the nonboiling convection effect, requiring fewer nucleating sites for a constant total heat fluxo At higher values of heat flux, however, the increased convection is insufficient to compensate for the decreased agitation presumably resulting from smaller bubble sizes. With an electrically heated surface then, the surface temperature must rise to provide the required additional nucleating sites. F. Concluding Remarks A great deal of analytic and experimental work remains before boiling heat transfer will be well understood, particularly as regards the interactions between the elements which make up the boiling process as illustrated in Figure 53. Acceleration of the boiling system directs attention to the mechanism governing the departure of a bubble from the heating surface. Equation (3), due to Fritz (14) is the only relation available showing an effect of acceleration on bubble size at departure, and must be tested experimentally. The use of fluids having liquid - vapor density ratios much different from water should prove enlightening, as should the use of higher accelerations and heat flux rates up to burnout. Other variables which might be considered with an accelerating boiling system are geometry, subcooling and forced convection0

APPENDIX A DERIVATION OF EQUATION.FOR ERROR IN MEASURING WATER TEMPERATURES Referring to Figure 64, the differential equation applied to the tube as a fin is: 2 t hc(t- t) = (82) 6x2 kA where: t = temperature of the tub~e x = distance from surface of the liquid h = coefficient of heat transfer between the fluid and.fin c = circumference of the tube k = thermal conductivity of the tube material A = cross sectional conduction area of the tube tw = Local fluid temperature two + Kx K = Saturation temperature gradient in the liquid two= Temperature of the liquid surface Let 2 hc 0 -t - two, m = hkA Substituting into Equation (82): m2Q = m2 Kx (83) whose solution is: = c1 e-a + c2 em + Kx (84)

LIQUID SURFACE!~~~~~~~~'~~~~~~~ X =O two x= Figure 64. Schematic of Thermocouple Tube Shown in Figure 21.

-145For the boundary conditions, take: (a)x=o (a)x=O = M = Constant (85) (2i>X= 7 = ( )x=6 = kA (twg - tj) = B(K - j) (86) where: R = length of tube immersed in the liquid twi - Temperature of the fluid at the end of the tube t) = Temperature of the tube at the end B = hAl/kA A1 = Cross sectional area of the closed end of the tube Almost the same numerical result is obtained if the boundary condition Gx=O = 0 is used instead of Equation (85) above. Applying the boundary conditions to Equation (84) and rearranging, the difference in temperature between the end of the tube and the local liquid temperature is-t B 1 + - M-K m+ m K t t~e- tt h B 1B + (87) Mtnh m + m(tanh mR + - m m Upon examining Equation (87) it is noted that as mE increases, the bracketed term on the right side tends toward zero, while the exponential increases. To determine the net effect, multiply the bracketed term by the exponential and replace the hyperbolic tangent by its exponential equivalent. Manipulation of this term results in the following: 2 mR -mr m B (em _ e-m) (88) m

-146For m > 2.5, emQ 7 e-m and Equation (88) can be written as 2 e myl + B ) (89) m Substituting into Equation (87) 2:(m-K)' K tw - t mem(1 + m(tanh mi +-) m m

APPENDIX B DERIVATION OF EQUATION FOR HEAT LOSS BY CONDUCTION THROUGH HEATER SKIRT The lower surface of the skirt in Figure 4 can be considered an adiabatic surface, since the convective coefficient to air is much less than to the water on the upper surface. As an approximation, a one-dimensional temperature distribution in the skirt will be assumed. Figure 65 is a schematic representation of the skirt with the adiabatic surface on the centerline and the length wo corresponding to the 0.062 inch dimension of Figure 4. Letting:.= t - tf the solution to the one-dimensional flat fin problem is given by: = C1 eNx + c e-Nx (91) where: N= h/kb (92) For the case of Figure 65, we have two related solutions. For Ox C wo 1 = C1 e1 + C2 eNl (93) For Xv wo Q2 = C3 eN2X + C4 eX (94) -147

-1480t /' X I FLUID TEMP. = tf Figure 65. Extended Surface Representing the Heater Skirt.

-149The boundary conditions are: (a) (%)x=o = ~ (b) (o2)X=o =o 2 - 1 ~-)x dx X-W0 62 dx (d) (l)x=w = (2:)x=w By application of these boundary conditions to Equations (93) and (94) it can be shown that: ~1 = (GO C2) e + C2 e (95) and d1 dx-= N1 (0 - 02) eNix _ C2 N1 e-NlX (96) where (1+B)o eNlWo 02 = 2(B cosh Nlwo + sinh Nlwo) and 51 N1 61 1/2 B -= (98) Let L - width of the fin. The heat transfer by conduction at the base of the fin is: d1 dG ~dg = - k(-)x=o =-:k L ( )xo (99) dx axi

-150Substituting Equations (96) and (97) with x=O, and taking one-half of the total: (1+B) e"lWo q = -k 61 LN [1o [1- B cosN + sinh N (100) The physical dimensions from Figure 4 are: 51 =.022 inches 62 =.066 inches wo = 0.062 inches L = 0.785 ft. B = 0.577 The quantity within the bracket on the right side of Equation (100) was calculated for values of h of 50 and 1000 Btu/hr-ft2-F and was found to be -1.11 and -0,993 respectively, Since the values of h fall in this range it is taken as constant at -1. Thus q = k dl~~leo (101) Taking k = 8 Btu/hr-ft-OF for stainless steel and substituting the values given above, the expression for the heat loss through the skirt per unit area of the main heating surface is: qloss/A = 2.02 (h) 1/2O (102) The above derivation assumes that h is constant. A more realistic approach is to take h as a function of G as well as acceleration.

-151An approximation can be attained by using the convection correlation recommended for a horizontal surface by McAdams(,): u = 0.14(GrPr) (103) Solving for h: h = Kg/3 (104) where (Pf a.1/5 K=0.14kf( 2 Pr (105) Substituting Equation (104) into the equation for the fin in Figure 65: ae2 -k e = 0 (106) Ox This is a non-linear differential equation, and a solution with the step change in cross sectional area would be quite complex. By assuming a uniform cross sectional area, Equation (106) can be integrated to: dQ 6 K D7/5 + C]1/2 (107) dx [ (107)7 Applying the boundary conditions: X=0= Q)x=, = 0 dx then C!1 = 0. Therefore: dO 6 K V 7/5 1/2 (108)

-152Integrating again and applying the boundary condition _x=o = Go.1 6 = -/-6 1x. )/2 (109) 4-2kb Calculation of 0 at x = wo Of Figure 65 indicated that the increase in cross sectional area will have negligible effect on the heat loss for the conditions present. Using Equation (108), the total heat loss through the skirt per unit area of the main heating surface is: L (6 ( ) /2.7/6 (110) loss A 7 0 where L = Width of the fin A = Main heat transfer area =.0471 ft2 K = Given by Equation (105)

APPENDIX C APPROXIMATION OF VELOCITY OF FLUID AT HEATING SURFACE DUE TO CORIOLIS' FORCE Considering natural convection only the fluid is assumed to flow past the heating surface through an opening provided by the baffle shown in Figure 66, and in so doing is heated from the temperature at T5 to that at T6. For convection Run No. C-5 at a/g = 21.15: q/A = 10,220 Btu/hr-ft2 q = 471 Btu/hr. =.131 Btu/sec T6-T5 = 2 ~F The mass rate of flow through the opening is: q..131 =.0605 lbm/sec Ah 2 Any increase in the difference in temperature would decrease the mass flow rate. The velocity through the opening between the baffle and heater surface then is: m.0605 x 144 V = pA - 60 x.75 =0.19 ft/sec -1553

-154DIRECTION OF ROTATION BAFFLE —, - A SU C3 — HEATER SURFACE Figure 66. Model used to Calculate Maximum Possible Water Velocity Due to Coriolis Acceleration in Test Vessel.

APPENDIX D DATA -155

-156APPENDIX D-1 SELECTED NATURAL CONVECTION DATA Run No. q/A RPM a/g Tw T5 T6 AT(Tw - T5) AT(2-1) AT(2-4) h C-1 4700 0 1 196.3 173.0 173.0 23.1 0.3 0.1 204 110 5.38 175.5 159.0 159.3 16.7 0.2 0 277 155 10.72 166.5 151.5 152.2 15.2 0.4 0 312 190 16.05 161.3 147.6 147.8 13.8 0.3 0 344 220 21.50 155.2 142.9 143.2 12.6 0.3 0 372 C-2 9840 0 1 227.4 195.7 196.1 31.5 0.6 0.2 316 ** 110 5.38 209.5 183.6 184.2 25.9 o.6 0.1 378 155 10.72 197.1 174.5 175.1 22.4 0.5 0 437 *C-5 10,220 0 1 228.2 201.8 201.9 26.4 1.0 0.4 388 ** 110 5.29 215.3 194.5 197.1 20.8 1.0 -0.6 493 155 10.47 202.6 185.6 188.0 17.0 1.1 -0.7 600 190 15.73 195.4 180.1 182.4 15.3 1.2 -0.8 669 220 21.15 190.8 177.0 178.6 13.8 1.2 -0.8 740 * With Flow Guide ** Boiling Taking Place Note: Accelerations for C-1 & C-2 vary from remainder of test because of modification to apparatus.

APPENDIX D-2 RUN NO. B-15 q/A = 10,870 Btu/hr-ft2 Time RPM Tw - Tsat Tw - T5 Tsat Tw T5 T6 AT(1-3) AT(2-1) AT(2-4) 9:35 0 18.9 30.0 211.2 230.0 200.0 200.0 10:05 a/g= 19.1 30.3 230.3 200.0 200.0 0.9 1.1 0:27 1 19.6 26.0 230.8 204.8 204.8:41 19.6 24.5 2530.7 206.2 206.3:54 19.4 23.6 211.1 230.5 206.9 206.8 11:09 19.3 23.2 230.5 207.3 207.4:21 19.3 23.0 230.4 207.4 207.6:33 19.0 22.0 230.1 208.2 208.3:46 18.9 21.9 230.0 208.2 208.4 1.3 1.1 -0.3 12:05 18.2 19.1 229.3 210.2 210.3:14 18.5 19.1 229.6 210.5 210.6:27 18.7 19.6 229.8 210.2 210.3:42 18.5 20.1 229.6 209.5 209.7:47 18.3 19.2 229.4 210.3 210.4 1:00 18.6 18.0 229.6 211.6 211.7 *:05 19.3 18.3 230.3 212.0 212.0 *:07 19.0 18.0 \ 230.0 212.0 2:32 / 18.5 19.0 211.0 229.5 210.4 210.4:40 110 16.9 17.8 212.2 229.1 211.3 211.3:56 a/g= 16.7 17.6 228.9 211.3 211.4 3:04 5.29 15.8 15.9 227.9 212.0 212.0:11 15.7 15.1 227.9 212.8 212.8 0.7 1.3 -0.4:17 15.7 14.7 227.8 213.2 213.2:40 15.5 14.1 227.6 213.6 213.6 *:48 15.5 14.1 227.6 213.6 213.6 4:09 15.5 15.2 212.1 227.6 212.4 212.4:21 15.9 16.3 228.1 211.8 211.8: 39 16.2 16.8 228.3 211.5 211.5:57 16.9 17.7 229.0 211.3 211.4 5:13 18.1 21.2 230.3 209.1 209.2:21 18.0 21.3 230.1 208.8 208.9 5:30 17.3 21.2 229.4 208.2 208.6:38 16.5 21.1 228.6 207.4 208.0:47 15.8 20.8 227.9 207.1 207.5 **:51 16.6 20.4 228.7 208.2 208.5:57 17.7 19.8 229.8 210.0 210.0 6:01 16.1 16.6 VI 228.2 211.6 211.6:11 0 18.6 20.0 210.9 229.5 209.6 209.8 7:11 0 18.9 19.9 210.9 229.8 209.9 210.1 1.2 1.1 -0.4 ** 7:26 220 12.9 14.4 216.3 229.2 214.8 214.8:538 a/g= 11.9 14.1 228.1 214.0 213.9:51 10.0 14.1 226.3 212.2 212.1 1.2 1.2 -0.7 **:57 21.15 10.8 13.9 227.1 213.1 213.5 8:10 13.3 14.3 229.6 215.2 215.4:19 14.2 14.6 230.5 215.8 215.9:27 14.5 14.8 230.8 216. 0 216.0: 32 14.9 14.7 231.2 216.5 216.4:41 14.8 15.0 231.1 216.1 216.6:45 15.7 14.5 2352.0 217.6 217.6:50 16.4 14.6 232.7 218.1 218.2 *:59 14.7 14.5 231.0 216.6 216.5 * 9:01 15.3 14.4 231.5 217.1 216.9 *:05 14.9 14.8 231.2 216.4 216.5 *:08 15.1 14.7 231.4 216.7 216.8 *:11 15.0 14.7 231.3 216.5 216.6:16 14.6 14.7 230.9 216.2 216.3:30 13.3 14.3 229.6 215.2 215.2:40 12.1 14.3 228.4 214.1 214.4:50 11.4 14.1 227.6 213.5 213.4 *'10.01 9.4 14.5 225.6 211.1 210.9:18 4.6 14.2 220.8 206.6 206.8:23 0 19.3 22.7 210. 9 250.2 207.5 207.6:34 19.1 21.7 210.9 250.0 208.3 208.5 1.3 1.0 0.2

-158APPENDIX D-2 (CONT'D) Time RPM Tw - Tsat Tw - T5 Tsat Tw T5 T6 AT(1-3) AT(2-1) AT(2-4) 9:22 0 19.0 21.3 211.0 230.0 208,6 208.6:36 0 18.9 21.2 211.0 229.9 208.7 208.7 1.3 1.1 -0.5:58 155 12.8 17.3 21.5 226.4 209.1 209.3 10:12 a/g= 14.0 17.8 227.6 209.8 210.1 0.9 1.3 -0.6:26 10.47 14.8 18.1 228.4 210.3 210.6:37 15.1 17.5 228.7 211.2 211.7:46 15.4 17.9 229.0 211.1 211.5:56 15.2 17.7 228.8 211.1 211.3 11:06 15.1 17.7 228.7 211.0 211.3:22 15.1 16.1 228.6 212.5 212.7:30 15.2 15.9 228.7 212.8 213.0:46 15.4 15.7 228.9 213.2 213.4 12:00 16.4 15.2 229.9 214.8 214.6 *:o616.3 15.2 229.9 214.6 214.7 *:09 16.2 15.1 229.8 214.7 214.7:33 15.3 17.4 228.8 211.4 211.6:44 0 18.6 19.6 211.0 229.6 210.0 210.0 1:11 0 18.7 20.6 211.0 229.7 209.1 209.2 1.3 1.1 -0.4: 31 65 16.2 17.1 211.2 227.5 210.3 210.1:43 a/g= 16.4 17.4 227.6 210.3 209.9 1.2 1.2 -0.4 2:10 1.95 16.5 16.1 227.7 211.6 211.4:27 16.3 15.9 227.5 211.6 211.4:44 16.8 15.7 228.0 212.3 212.2:57 16.9 15.7 228.1 212.4 212.3 * 3:00 17.0 15.6 228.2 212.6 212.4 *:03 16.9 15.7 228.2 212.5 212.4:16 16.8 15.8 228.0 212.2 212.0:40 \ 16.8 17.6 228.0 210.4 210.2:50 0 18.3 19.6 211.0 229.3 209.7 209.6 4:03 0 18.5 19.6 211.0 229.4 209.9 209.9 1.3 1.0 -0.4 4:23 190 13.7 15.1 214.9 228.6 213.4 213.6:39 a/g= 14.2 14.4 229.1 214.6 214.7 0.6 1.3 -0.5:48 15.73 14.1 14.3 229.0 214.7 214.8 5:06 14.2 14.3 229.1 214.8 215.1:23 14.3 13.9 229.2 215.3 215.4 *:26 14.2 13.9 229.1 215.2 215.1 *:29 14.3 14.0 4/ 229.1 215.2 215.3:46 14.3 15.7 214.8 229.1 213.4 213.5 6:04 13.9 15.8 228.7 212.8 213.2:12 14.3 15.2 229.1 213.9 214.3:32 12.4 15.8 227.3 211.5 211.8:42 11.4 15.3 226.2 210.9 211.4:55 10.7 15.2 225.5 210.3 211.0 7:C8 6.7 15.3 221.6 206.3 206..4:17 % 5.2 15.3 220.0 204.7 205.0:26 0 19.4 23.8 211.0 230.4 206.6 206.9:34 0 18.8 21.4 211.0 229.8 208.4 208.4 1.3 1.0 -0.2 * Indicates losing steam ** Data questionable because of non-equilibrium, resistivity of water Note: Resistivity of water p(before test) = 1.5 X 106 jL -cm p(after test) = 0.9 x 106 yL -cm

-159APPENDIX D-3 RUN NO. B-9 q/A = 24,450 Btu/hr-ft2 Time RPM Tw - Tsat Tw - T5 Tsat Tw T5 T6 AT(1-3) AT(2-1) AT(2-4) 12:45 0 22.2 22.8 211.1 253.3 210.5 210.5 1:20 a/g= 22.3 22.3 233.4 211.1 211.1:31 1 22.3 22.2 233.4 211.1 211.1 0.7 2.9 0.4:39 22.2 21.9 233.3 211.4 *:42 22.3 21.7 233.4 211.7 211.5:52 22.3 22.2 233.4 211.2 211.1:57 110 20.2 21.2 212.3 232.4 211.2 211.6 2:06 a/g= 20.3 21.5 232.6 211.1 211.4 0.9 3.2 -0.2:13 5.29 20.3 21.4 232.6 211.2 211.5 **:19 20.2 20.9 232.4 211.5 212.0:24 20.1 20.5 232.3 211.8 212.2 *:29 20.2 20.2 232.4 212.2 212.6 0.7 3.4 -0.3:36 20.0 20.7 232.2 211.5 211.8:44 20.4 21.5 232.6 211.1 211.5 **:49 20.3 20.4 232.6 212.1 212.5:51 w 20.4 232.6:55 0 22.1 22.0 211.1 233.1 211.2 211.1 3:13 22.1 22.1 233.2 211.1 211.0 0.6 3.0 0.4:22 22.2 21.9 233.3 211.4 *:24 22.4 21.7 233.4 211.7:32 220 19.6 24.2 216.4 236.0 211.8 211.8:41 a/g= 19.4 23.9 235.9 211.9 212.0:47 17.5 20.7 233.9 213.2 213.1:51 21.15 17.2 20.2 233.6 213.4 213.4 0.8 3.2 -0.5 *:55 17.0 20.0 233.5 213.4 213.4 4:01 18.9 23.0 235.3 212.2:07 j 19.7 24.5 236.1 211.7 **:13 17.9 21.3 234.3 213.1 213.0 *:17 17.2 20.1 233.6 213.5:22 0 22.2 22.0 211.1 233.3 211.3:35 0 22.2 22.2 211.1 233.3 211.1 211.0 0.7 3.1 0.3 4:41 65 21.7 21.6 211.4 233.1 211.5 212.0:47 a/g= 21.7 21.4 233.1 211.7 212.0 1.2 2.9 -0.2 *:54' 1.95 22.0 21.2 233.3 212.1 212.4 5:00 22.0 21.3 233.3 212.0 **:07 21.9 21.7 233.2 211.5:21 21.9 21.8 233.2 211.4 211.8:25 22.3 22.1 211.1 233.4 211.3 211.2:33 0 22.3 22.1 211.1 233.4 211.3 211.2 0.6 2.9 0.6:41 135 19.6 20.9 213.0 232.5 211.7 211.9:48 a/g= 19.6 20.9 232.6 211.7 211.9 1.1 3.3 -0.5:54 7.84 19.5 20.6 232.5 211.9 * 6:00 19.5 20.3 232.5 212.2 212.5:10 19.2 21.3 232.2 210.9 211.0:29 19.1 21.0 232.1 211.1 211.4:34 0 22.2 22.2 211.2 233.4 211.2 211.1:43 0 22.2 22.3 211.2 233.4 211.0 211.0 0.4 3.1 0.4:52 190 19.6 23.1 215.0 234.6 211.6:59 a/g= 18.1 20.8 233.1 212.3 212.4 1.0 3.2 -0.6 7:05 15.73 18.2 20.5 233.1 212.7 *:07 18.2 20.3 232.2 212.9 213.1:12 18.0 20.6 232.9 212.3:18 18.7 21.7 233.7 212.0 212.0:30 20.5 24.3 235.5 211.2 211.3:40 18.6 21.4 233.6 212.2 *:46 18.2 20.3 233.2 212.9 **:49 18.0 20.5 232.9 212.5:52 18.5 21.4 233.4 212.0:55 19.4 22.6 234.3 211.8:58 0 22.2 22.1 211.2 233.3 211.2 211.1 8:06 0 22.2 22.1 211.2 233.4 211.3 211.2 0.7 3.1 0.5

-160APPENDIX D-3 (CONT'D) Time RPM Tw - Tsat Tw - T5 Tsat Tw T5 T6 fIT(1-3) AT(2-1) AT(2-4) ** 8:15 155 19.5 22.1 213.6 233.1 211.0 211.0:22 a/g= 18.8 21.0 232.5 211.5 211.7 1.1 3.3 -0.6:26 10.47 18.7 20.6 232.4 211.8 **:31 19.0 20.6 232.6 212.0:33 18.8 20.1 232.4 212.3 212.6 *:35 18.7 19.9 232.4 212.5:39 18.7 20.1 23 2.3 212.3:55 19.1 21.5 232.7 211.2 211.4:59 18.8 20.8 232.4 211.6 ** 9:01 19.0 20.5 232.7 212.1:03 18.8 20.1 232.5 212.3:05 18.8 20.0 232.4 212.5 *:08 18.9 19.7 232.5 212.8 213.0:12 18.9 20.2 232.5 212.3:16 18.9 21.0 232.5 211.5:20 19.2 21.7 232.8 211.1:22 0 22.1 22.2 211.2 233.3 211.1 211.1:32 0 22.2 22.2 211.2 233.4 211.2 211.1 0.8 3.2 0.2 3:59 175 19.7 22.7 214.4 234.0 211.3 211.3:45 a/g= 20.0 23.2 234.3 211.2:53 18.7 21.2 233.0 211.8 211.9 1.1 3.2 -0.6:58 13.35 18.4 20.7 232.8 212.1 10:01 18.5 20.2 232.8 212.6 *:03 18.4 20.0 232.8 212.8 213.0:07 18.5 20.5 232.9 212.4:15 18.5 20.9 232.8 212.0:20 19.7 22.7 234.0 211.4:23 0 22.1 22.2 211.2 233.3 211.1 211.1:37 0 21.9 21.8 211.2 233.1 211.3 211.2 0.5 3.1 0.4 10:45 85 20.8 21.4 211.8 232.6 211.2 211.5:52 a/g= 20.7 21.1 232.5 211.4:57 321 20.6 20.9 232.4 211.5 211.8 0.5 3.5 -0.2 11:03 20.4 20.5 232.2 211.7:07 20.2 20.2 232.0 211.8 *:09 20.3 20.0 232.1 212.1 212.5 *:12 20.4 20.1 232.2 212.1:15 20.5 20.8 232.3 211.5:22, 20.7 21.3 232.5 211.2 211.7:27 0 21.9 21.5 211.2 233.1 211.6 211.5:31 22.0 21.6 233.2 211.6:34 22.0 21.9 233.2 211.2 211.2 0.5 3.2 0.1 *:42 21.8 21.2 233.0 211.8:46 4 21.8 22.2 233.0 210.8 210.8 * Indicates losing steam ** Data questionable because of non-equilibrium Note: Resistivity of water p(before test) = 1.4 x 106 _L -cm p(after test) = 0.8 x 106 _L -cm

APPENDIX D-4 RUN NO. B-14 q/A = 48,800 Btu/hr-ft2 Time RPM Tw - Tsat Tw - T5 Tsat Tw T5 T6 AT(1-3) AT(2-1) AT(2-4) 12:34 0 26.1 26.1 210.9 237.0 210.9 210.9 0.1 5.8 0.6:50 a/g=l 26.2 26.2 237.1 210.9 210.9 1:03 26.1 25.9 237.0 211.1 211.0:22 26.1 26.0 237.0 211.0 211.0:29 25.9 25.8 236.8 211.0 211.0 0 5.9 0.7 *:35 26.1 25.7 237.0 211.3 211.2:48 4 26.0 25.9 236.9 211.0 210.9 2:00 110 25.0 25.8 212.1 237.1 211.3:07 a/g= 24.9 25.7 237.0 211.3 0.8 6.2 -0.1 *:17 5.29 25.1 25.5 237.2 211.7:18 25.2 25.6 237.3 211.7:25 25.1 25.9 237.2 211.3 211.5:29 25.1 25.8 237.1 211.3:35 0 26.0 25.9 210.9 236.9 211.0 210.9:46 25.9 25.9 236.8 210.9 210.9 0 5.9 0.8 3:19 25.9 25.9 236.8 210.9 210.8:29 220 25.8 30.0 216.3 242.0 212.0 212.1:35 a/g= 25.7 29.9 242.0 212.1:43 21.15 25.5 29.4 241.7 212.3 212.3 0.8 6.4 -0.3:54 25.4 28.9 241.6 212.7 212.6:59 25.7 28.3 242.0 213.7 213.6 * 4:00 25.7 28.3 \4 242.0 213.7 213.6:08 25.6 29.5 216.2 241.8 212.3 212.2:16 0 26.0 25.9 210.9 236.9 211.0 210.9:25 0 26.0 26.0 210.9 236.9 210.9 210.9:33 65 25.8 25.8 211.1 236.9 211.1 211.0:43 a/g= 25.9 25.9 237.0 211.1 210.9 1.0 5.7 0.3:50 1.95 25.9 25.7 237.0 211.3 211.2 *:52 1 26.0 25.7 237.2 211.5 211.2 5:00 25.9 25.8 237.0 211.2 210.9:05 0 26.2 26.0 210.9 237.1 211.1 210.9:22 0 25.8 25.6 210.9 236.7 211.1 211.0 0 6.0 0.8 5:32 135 24.9 26.3 212.7 237.7 211.4 211.6:42 a/g= 25.0 26.2 237.7 211.5 211.7 0.9 6.2 0:52 7.84 24.9 25.7 237.6 211.9 212.1 *:55 ~ 25.3 26.0 238.0 212.0 212.2 6:05 25.1 26.3 237.8 211.5 211.7:10 0 25.9 25.8 2 0.9 236.8 211.0 210.9:39 0 26.1 26.0 210.9 237.0 211.0 210.9 0.2 5.9 0.6:51 190 25.5 28.2 214.7 240.1 211.9 211.9:57 a/g= 25.3 28.0 240.0 212.0 212.1 0.6 6.3 -0.3 7:06 15.73 25.5 27.4 240.2 212.8 212.8 *:08 25.6 27.3 240.2 212.9 212.9:12'4 25.4 28.2 214.6 240.1 211.9 212.0:17 0 26.2 26.0 210.8 237.0 211.0 210.9 10:45 0 26.2 26.1 210.8 237.1 211.0 210.9 0 6.0 0.8:57 155 25.0 26.7 213.3 238.3 211.6 211.9 11:02 a/g= 25.3 26.9 238.5 211.6:07 10.47 25.2 26.7 238.4 211.7 211.9 0.9 6.3 -0.3:14 25.0 238.3 *:16 25.4 26.3 238.7 212.4 212.6:25 24.9 26.2 238.1 211.9 212.0:28 \Y 25.0 26.6 238.3 211.7 211.9:36 0 26.2 25.9 210.9 237.0 211.1 211.0:48 0 26.4 26.2 210.9 237.3 211.1 211.0 0.4 5.8 0.6:57 85 25.7 26.0 211.4 237.1 211.1 211.2 12:05 a/g= 25.6 25.8 237.0 211.2 211.3 1.2 6.0 0 *:11 3.21 25.9 25.7 237.3 211.6 211.5:16 25.7 25.9 237.1 211.2 211.2:22 0 26.1 26.0 10.9 237.0 211.0 210.9:31 0 26.1 26.0 210.9 237.0 211.0 210.9 0.1 5.9 0.6 * Indicates losing steam Note: Resistivity of water p(before test) = 1.5 x 106 _L -cm p(after test) = 0.85 x 106 _ -cm

-162APPENDIX D-5 RUN NO. B-22 q/A = 73,000 Btu/hr-ft2 Osc. of Time RPM Tw - Tsat Tw - Tsat Tsat Tw T5 T6 LT(1-3) ) AT(2-1) AT(2-4) 2:50 0 25.7 211.0 236.7 211.2 210.9 -1.1 10.3 -0.7 3:30 0 25.6 + 0.1 211.0 236.6 211.3 211.0 -0.8 10.3 -0.8: 34 0 25.6 211.0 236.7 4:20 65 25.6 211.3 236.9 211.3 211.0 -1.0 10.2 -0.9:25 65 25.6 + 0.1 211.3 236.9:39 0 25.6 211.0 236.6 211.1 210.9 -0.8 10.3 -o.8:44 0 25.6 + 0.1 211.0 236.6 5:00 85 25.4 211.6 237.0 211.3 211.1 -0.8 10.2 -1.0:04 85 25.4 211.6 237.0:16 0 25.6 211.0 236.6 211.1 210.9 -0.7 10.2 -0.7:21 0 25.7 + 0.1 211.0 236.7:47 110 25.3 + 0.1 212.2 237.5 211.4 211.2 -0.8 10.2 -1.1:52 110 25.3 212.2 237.6 6:12 0 25.6 + 0.1 211.0 236.6 211.0 210.7 -0.6 10.3 -0.8:15 0 25.7 + 0.2 211.0 236.7:29 135 25.4 213.0 238.4 211.7 211.6 -0.8 10.2 -1.1:33 135 25.3 213.0 238.3 7:15 0 25.6 + 0.1 211.0 236.7 211.2 210.9 -0.7 10.3 -0.7:19 0 25.6 + 0.1 211.0 236.7:33 155 25.5 213.7 239.1 212.0 211.7 -0.7 10.3 -1.0:39 155 25.4 + 0.1 213.7 239.1:54 0 25.6 211.1 236.7 211.2 211.0 -0.6 10.1 -0.8:59 0 25.7 211.1 236.7 8:16 190 26.1 + 0.0 215.1 241.2 212.4 211.8 -1.2 10.4 -0.9:21 190 26.1 215.1 241.1:37 0 25.7 211.0 236.7 211.2 211.0 -0.7 10.2 -0.6:40 0 25 7 211.0 236.7:55 220 26.9 216.5 243.4 212.6 212.1 -0.7 10.2 -1.2 9:01 220 27.0 216.5 243.5:24 0 25.6 + 0.1 211.0 236.7 211.2 211.0 -0.8 10.4 -0.7:31 0 25.6 211.0 236.6:41 65 25.5 211.3 236.9 211.3 211.1 -1.0 10.2 -1.0:46 65 25.5 + 0.1 211.3 236.8 9:57 0 25.5 211.1 236.6 211.3 211.0 10:07 110 25.3 212.3 237.6 211.6 211.4:18 0 25.7 211.1 236.8 211.4 211.1:36 155 25.4 213.7 239.1 212.0 211.8:55 0 25.7 211.1 236.8 211.53 211.0 11:11 190 25.9 215.2 241.1 212.4 211.9:27 0 25.6 + 0.1 211.1 236.7 211.2 211.1:48 220 27.0 216.5 243.5 212.7 212.1 -0.8 10.2 -1.1:52 220 27.0 216.5 243.5 12:06 0 25.6 + 0.2 211.1 236.6 211.2 211.0 -0.8 10.3 -0.8:11 0 25.5 ~ 0.2 211.1 236.6 Resistivity of water p(prior to test) = 1.35 x 1o6 JL -cm p(after test) = 1.0 x 1056 L -cm

-163APPENDIX D-6 RUN NO. B-21 q/A = 99,500 Btu/hr-ft2 Osc. of Time RPM Tw - Tsat Tw - Tsat Tsat T T5 T6 AT(1-3) aT(2-1) AT(2-4) 1:53 0 29.3 + 0.3 210.5 239.7 210.4 210.1 -0.5 13.7 -0.3 2:19 29.4 + 0.2 239.8 210.5 210.3:32 29.5 240.0 210.6 210.3:53 \Y 29.5 + 0.3 239.9 210.4 210.3 -0.8 13.5 0 3:28 65 29.7 + 0.2 210.7 240.4 210.7 210.5 -0.2 13.2 -0.2:50 0 29.5 + 0.3 210.5 239.9 210.6 210.9 -0.9 13.7 0:54 0 29.5 + 0.2 210.5 239.9 4:16 85 29.7 + 0.3 211.1 240.7 210.7 211.3 -0.3 13.1 -0.1:21 85 29.6 + 0.1 211.1 240.7:35 0 29.4 + 0.2 210.5 239.9 210.6 210.8 -0.8 13.6 0:40 0 29.4 + 0.1 210.5 239.9:55 110 29.7 + 0.1 211.7 241.4 210.9 211.5 0 13.1 0.2 5:00 110 29.6 + 0.1 211.7 241.3:23 0 29.4 + 0.2 210.5 239.9 210.6 210.8 -0.8 13.6 0:28 0 29.4 + 0.2 210.5 239.9:44 135 29.0 + 0.1 212.5 241.4 211.1 211.5 -0.1 13.3 -0.5:52 *29.6 242.1 6:05 4 29.7 242.1 211.1 211.5:29 0 29.4 + 0.3 210.5 239.9 210.6 211.5 -0.9 13.7 0:35 0 29.3 ~ 0.2 210.5 239.8:49 155 29.9 + 0.1 213.2 243.0 211.4 212.6 0 13.2 -0.3:55 155 29.9 213.2 243.0 7:11 0 29.3 + 0.3 210.6 239.8 210.7 211.0 -0.7 13.7 -0.1:17 0 29.4 + 0.3 210.6 239.9:38 190 29.7 + 0.1 214.6 244.3 211.8 213.7 -0.1 13.5 -0.6:48 190 29.7 + 0.1.214.6 244.3 8:01 0 29.2 ~ 0.3 210.6 239.8 210.8 210.8 -0.9 13.7 0:07 0 29.2 + 0.3 210.6 239.8:21 220 30.5 + 0.1 216.1 246.5 212.2 213.5 0 13.8 -0.7:28 220 30.4 216.1 246.5:46 0 29.1 + 0.2 210.6 239.7 210.8 210.8 -0.9 13.6 0:55 0 29.1 210.6 239.7 9:09 65 29.0 ~ 0.3 210.9 239.9 210.9 210.8 -o.8 13.3 -0.2:16 65 29.1 210.9 240.0:31 0 29.1 + 0.3 210.7 239.8 210.8 210.9 -0.9 13.6 -0.2:37 0 29'.2 210.7 239.8:52 85 29.1 + 0.3 211.3 240.4 210.9 211.1 -0.6 13.4 -0.2:59 85 29.1 ~ 0.2 211.3 240.4 10:16 0 29.0 210.7 239.7 210.8 210.6 -0.8 13.7 -0.1:23 0 29.0 + 0.2 210.7 239.7:35 110 30.0 211.9 241.8 211.1 211.4 0.4 12.8 -0.2:43 110 30.0 + 0.1 211.9 241.9:55 0 29.0 + 0.3 210.7 239.7 210.8 210.9 -0.9 13.8 -0.1 11:02 0 29.0 + 0.3 210.7 239.7:13 135 29.4 ~ 0.1 212.6 242.0 211.3 211.8 0.2 13.2 -0.2:21 135 29.6 212.6 242.2:27 135 29.6 ~ 0.1 212.6 242.2:40 0 29.0 + 0.3 210.7 239.7 210.9 211.1 -0.8 13.6 -0.3:46 0 29.0 + 0.2 210.7 239.7:59 155 29.4 + 0.1 213.3 242.7 211.5 212.3 0.1 13.3 -0.5 12:05 155 29.4 + 0.1 213.3 242.7:20 0 29.0 ~ 0.3 210.7 239.7 210.8 210.8 -0.8 13.6 -0.2:26 0 29.0 + 0.3 210.7 239.7:37 190 29.4 ~ 0.1 214.8 244.2 212.0 213.4 -0.4 13.4 -0.5:45 190 29.4 + 0.1 214.8 244.2 1:03 0 28.7 210.7 239.4 210.8 210.7 -1.0 13.7 -0.3:09 0 28.7 + 0.3 210.7 239.4:20 220 30.2 + 0.1 216.2 246.3 212.2 213.3 -0.2 13.4 -0.9:24 220 30.1 216.2 246.3:39 0 28.8 + 0.3 210.7 239.5 210.8 210.6 -0.8 13.5 -0.2:43 0 28.8 210.7 239.5 Resistivity of water p(before test) = 1.5 x 106 /L -cm p(after test) = 1.0 x 106 _L -cm

-164APPENDIX D-7 BOILING DATA WITH a/g = 1 FOR VARIOUS TEST RUNS Run Initial Final Initial Water Final Water Final Surface q/A No. Tw - Tsat Tw - Tsat Resistivity Resistivity Condition 10,220 C-5 18.9 19.0 10,580 B-7 17.2 17.5 1.5 x 106 _ —cm 0.9 x 106 JL-cm 10,870 B-15 18.5 18.5 1.4 x 106 0.9 x 106 24,050 B-2 22.0 22.0 1.5 x 10 0.4 x 106 Light colored spots 24,050 B-3 23.3 23.2 1.5 x 106 1.0 x 106 24,050 A-16 21.9 22.5 1.5 x 106 0.4 x 106 Light colored spots 24,050 B-6 19.7 20.1 1.5 x 106 1.0 x 106 24,450 B-8 22.0 22.2 1.5 x 106 0.8 x 106 24,450 B-9 22.2 21.9 1.4 x 106 0.8 x 106 24,450 B-ll 22.4 22.5 1.4 x 106 0.9 x 106 48,800 B-12 25.3 25.2 1.4 x 106 0.8 x 106 48,800 B-14 26.0 26.1 1.5 x 106 0.9 x 106 72,700 B-17 29.5 25.7 1.5 x 106 0.6 x 106 Light colored spots 73,000 B-19 27.0 26.5 1.4 x 106 0.7 x 106 Pinpoint shadows 73,000 B-22 25.6 25.5 1.4 x 106 1.0 x 106 99,500 B-20 29.4 28.6 1.5 x 106 1.1 x 106 99,500 B-21 29.5 28.8 1.5 x 106 1.1 x 106

-165APPENDIX D-8 SUMMARY OF DATA USED IN FIGURE 47 Run Original Modified No. q/A RPM Tsat Tw - Tsat Tw - T5 Tw - Tsat Tw- T5 B-15 10,870 0 211.0 18.5 18.5 18.8 18.8 65 211.2 16.3 16.5 16.6 16.8 110 212.1 15.8 15.9 16.1 15.9 155 213.5 15.1 16.1 15.4 16.2 190 214.8 14.2 14.9 14.5 15.2 220 216.3 13.3 14.4 13.6 15.2 B-7 10,580 0 210.2 17.3 17.3 18.8 110 211.3 13.7 14.0 15.2 125 211.7 13.8 14.0 15.3 155 212.7 13.5 14.0 15.0 190 214.1 13.4 14.0 14.9 B-9 24,450 0 211.1 22.2 22.2 22.1 22.1 65 211.4 21.9 21.9 21.8 21.4 85 211.8 20.4 20.5 20.3 20.8 110 212.3 20.1 20.8 20.0 20.8 135 213.0 19.5 20.8 19.4 20.6 155 213.6 18.7 20.5 18.6 20.4 175 214.0 18.4 20.7 18.3 20.0 190 214.9 18.1 20.6 18.0 20.4 220 216.4 17.1 20.1 17.0 20.1 B-8 24,450 0 210.9 22.2 22.3 22.1 65 211.1 21.2 21.5 21.1 85 211.4 20.8 21.5 20.7 110 212.2 20.0 21.3 19.9 135 212.7 18.7 20.7 18.6 155 213.2 18.3 20.0 18.2 190 214.8 17.6 20.0 17.5 220 215.9 17.2 20.3 17.1 B-ll 24,450 0 210.7 22.4 22.4 22.1 65 210.9 22.7 22.7 22.4 155 213.2 19.0 19.9 18.7 220 215.8 17.5 20.1 17.2 B-12 48,800 0 209.5 25.3 25.3 25.4 65 209.9 24.8 24.8 24.6 110 210.7 24.5 25.1 24.7 135 211.5 24.6 25.9 24.5 190 213.5 24.5 27.1 24.7 220 214.8 25.0 28.8 25.1 B-14 48,800 0 210.9 26.1 26.1 25.4 25.4 65 211.1 25.9 25.9 25.3 25.1 85 211.4 25.7 25.9 24.9 25.0 110 212.1 25.0 25.8 24.4 25.3 135 212.7 25.0 26.3 24.4 25.5 155 213.3 25.1 26.8 24.3 26.0 190 214.7 25.4 28.1 24.7 27.4 220 216.3 25.5 29.3 25.0 29.0 B-22 73,000 0 211.0 25.6 25.3 27.7 27.4 65 211.3 25.6 25.6 27.7 27.6 85 211.6 25.4 25.7 27.5 27.8 110 212.2 25.3 26.1 27.4 28.2 135 213.0 25.4 26.7 27.5 28.7 155 213.7 25.4 27.1 27.5 29.2 190 215.1 26.0 28.7 28.0 30.8 220 216.5 27.0 30.9 29.1 33.0 B-21 99,500 0 210.5 29.5 29.5 29.5 29.5 65 210.7 29.7 29.7 29.7 29.5 85 211.1 29.6 30.0 29.6 30.0 110 211.7 29.7 30.5 29.8 30.4 135 212.5 29.6 31.0 29.8 31.2 155 213.2 29.9 31.7 30.1 31.7 190 214.6 29.7 32.5 30.0 32.9 220 216.1 30.5 34.4 30.8 34.7

-166APPENDIX D-8 (CONT'D) Run Original Modified No. q/A RPM Tsat Tw - Tsat Tw - T5 Tw - Tsat Tw - T5 B-21 99,500 0 210.6 29.1 28.9 29.5 65 210.9 29.1 29.1 29.6 85 211.3 29.1 29.5 29.6 110 211.9 30.0 30.8 30.5 135 212.6 29.6 30.9 30.2 155 213.3 29.4 31.2 29.9 190 214.8 29.4 32.2 30.1 220 216.2 30.1 34.1 30.9 B-20 99,500 0 210.8 29.4 29.4 29.5 65 211.2 29.0 29.0 29.4 85 211.5 29.2 29.5 29.9 110 212.0 29.4 30.2 29.6 135 212.8 29.1 30.4 29.9 155 213.4 29.1 30.8 29.7 190 214.8 29.7 32.3 30.2 220 216.2 30.0 33.8 30.2 110 212.1 29.4 30.3 29.6 220 216.2 29.8 33.5 30.7

BIBLIOGRAPHY 1. McAdams, W. H, Heat Transmission. 3rd Edition. New York: McGraw-Hill Book Co. Inc., 1954. 2. Jacob, M. Heat Transfer. I. New York: John Wiley & Sons, In*., 1949. 3. Bosworth, R, C. L, Heat Transfer Phenomena, New York: John Wiley & Sons Inc., 1952. 40 Griffith, P. "Bubble Growth Rates in Boiling." Trans. ASME, 80, No. 3, (April, 1958), 721. 5. Zuber, N. "On the Stability of Boiling Heat Transfer." Ibid, 711. 6. Westwater, Jo W. and Santangelo, J. G. "Photographic Study of Boiling." Ind, & Eng, Chem,, 47, (August, 1955), 1605. 7, Rohsenow, W. M. and Clark, J.A. "A Study of the Mechanism of Boiling Heat Transfer," Trans. ASME, 73, No. 5, (July, 1951), 609. 8. Clark, J. A. and Rohsenow, W. M. "Local Boiling Heat Transfer to Water at Low Reynold's Numbers and High Pressures."' Trans. ASME, 76, No. 4, (1954), 553. 9. Corty, C. and Foust, A. S. "Surface Variables in Nucleate Boiling." Chem, Eng. Prog. Symp., Series No, 17, 51, 1955. 10. Clark, J. A. "Thermodynamics of Bubbles." JM.I.T., Cambridge, Mass. ]D.I.C, Technical Report-No. 7, ONR Contract N:5 or i-07827 (NR-035-267),.J!nuary 1 1956. 11. Forster, H. K. and Zuber, N. "Growth of a Vapor Bubble in a Super-,heated Liquid." J. App. Phys., 25, No. 4, (1954), 474. 12. Forster, H, K. and Zuber, N. "Dynamics of Vapor Bubbles and Boiling Heat Transfer." AIChE Journal, 1, No. 4, (December, 1955), 531, 13. Plesset, M. S. and Zwick, S. A. "The Growth of Vapor Bubbles in Superheated Liquids." J. App. Phys., 25, No. 4, (1954), 493. 140 Fritz, W. "Calculation of the Maximum Volumes of Vapor Bubbles." Physik, Zeitschr., 36, (1935), 379. 15. Kreith, F. and Summerfield, M. "Heat Transfer to Water at High Flux Densities With and Without Surface Boiling." Trans. ASME, 71, (1949), 805 -167

-16816. McAdams, W. H. et al. "Heat Transfer at High Rates to Water with Surface Boiling." Ind. & Eng. Chem., 41, No. 9, (September, 1949), 1945. 17. Pramuk, F. S. and Westwater, J. W. "Effect of Agitation on the Critical Temperature Difference for a Boiling Liquid.'" Chem. Eng. Prog. Symposium, 52, No. 18, 1956. 18. Ellion, M. "A Study of the Mechanism of Boiling Heat Transfer." J.P.L., Memo 20-88, Pasedena, Calif., March, 1954. 19. Insinger, T. H. Transmission of Heat to Boiling Liquids. PhD Dissertation, Univ. of Penn., Philadelphia, Penn., 1940. 20. Morgan, A. I., Bromley, L. A., and Wilke, C. R. "Effect of Surface Tension on Heat Transfer in Boiling." Ind. & Eng. Chem., 41, No. 12, (December, 1949), 2676. 21. Bankoff, S. G. "Ebullition from Solid Surfaces in the Absence of a Pre-existing Gaseous Phase." Trans. ASME, 79, (May, 1957), 735. 22. Wakeshima, H. and Takada, K. "Limit of Superheat." J. Appl. Phys. 29, (July, 1938), 1126. 23. Mead, B. R., Romie, F. E.,and Guibert, A. G. "Liquid Superheat and Boiling Heat Transfer." Heat Transfer & Fluid Mechanics Institute, Preprints of papers, 1951, Stanford University Press, Stanford, Calif. 24. Sabersky, R. H. and Gates, C. W. "Effect of Pressure on Start of Nucleation in Boiling Heat Transfer." Jet Propulsion, 25, No. 2, (February, 1955), 67. 25. Bashforth, F., and Adams, J. Capillary Actioni Cambridge, 1883. 26. Wark, J. W. "The Physical Chemistry of Flotation." Journal of Physical Chem., 37, (1933), 623. 27. Pike, F. R., Miller, P. D., and Beatty, K. 0. "Effect of Gas Evolution on Surface Boiling at Wire Coils." Chem. Eng. Prog. Symp. Series, 51, No. 17, 1955. 28. Larson, R. F. "Factors that Influence Heat Transfer in Boiling." Heat Transfer and Fluid Mechanics Institute, Stanford University Press, Stanford, Calif., 1953. 29. Rohsenow, W. M. "A Method of Correlating Heat Transfer Data for Surface Boiling of Liquids." Trans. ASME, 74, (1952), 969.

-16930. Chang, Y. P. "Theoretical Analysis of Heat Transfer in Natural Convection and in Boilingo" Trans. ASME, 79, (October, 1957), 1501o 31. Borishanskii, Vo M. "An Equation Generalizing Experimental Data on the Cessation of Bubble Boiling in a Large Volume of Liquid." Zhurn, Tekh,-Fiz., 26, (1956), 452. Translated in Soviet PhysicsTechnical Physics, 1, No. 2, 438, Am. Inst. of Phys., N.Y. 32. -Siegel. R. and Usiskin, C, "A Photographic Study of Boiling in the Absence of Gravity." Trans. ASME, 81, Series C, No. 3, (August, 1959), 230. 335 Gambill, W. R. and Green, N. D, "A Study of Burnout Heat Fluxes Associated with Forced-Convection, Subcooled, and Bulk Nucleate Boiling of Water in Source - Vortex Flow." Chem, Eng. Prog., 54, No. 10, (October, 1958), 68. 34. Schmidt, E. H. W. "Heat Transmission by Natural Convection at High Centrifugal Acceleration in Water-Cooled Gas-Turbine Blades." The Institution of Mech, Eng. ASME, Proceedings of the General Discussion on Heat Transfer, 11-13 September, 1951. 35. Hickman, K. C. D. "Centrifugal Boiler Compression Still." Ind.;& Eng. Chem., 49, No. 5, (May, 1957), 786. 36. Gudheim, A. R. and Donovan, Jo "Heat Transfer in Thin Film Centrifugal Processing Units," Chem. Eng. Prog., 53, No. 10, (October, 1957), 476. 37. Sparrow, E. M. and Gregg, J. L. "A Theory of Rotating Condensation." Trans. ASME, Journ. of Heat Transfer, 81, Series C, No. 2, (May, 1959), 1135 38. Clark, H. B., Strenge, P. S., and Westwater, J. W. "Active Sites for Nucleate Boiling." Preprint # 13 presented at AIChE-ASME 2nd National Conference on Heat Transfer, Chicago, Illinois, August 18-21, 1958, 39. Fultz, D. and Nakagawa, Y. "Experiments on Over-Stable Thermal Convection in Mercury." Proc. Roy. Soc., (London) A 231, (1955), 211. 40. Kline, S. J. and McClintock, F, A. "Description and Analysis of Uncertainties in Single Sample Experiments." Aerodynamics Measurements, M.IoT., Summer Session, Chapter II, Part 2, September 8-19, 1952. 41. Gunther, F. C. "Photographic Study of Surface-Boiling Heat Transfer to Water with Forced Convection." Trans. ASME, 73, No, 2, (February, 1951), 115,

-17042. Bonilla, C. F,, Busch, J. S,, Stalder, A., Shaikhinahmud, N. S., and Ramachandran, A. "Pool Boiling Heat Transfer with Mercury." Reactor Heat Transfer Conference of 1956. 43, Clark, J. A. Heat Transfer with Saturated Boiling. ScDo Thesis, Massachusetts Institute of Technology, Cambridge, Mass., May, 1953. 44. Chang, Y. and Snyder, N. W. "Heat Transfer in Saturated Boiling."' Preprint 104 presented at Third National Heat Transfer Conference, ASME-AIChE, August, 1959. 45. GriffitIi P. "The Correlation of Nucleate Boiling Burnout Data." ASME, Paper No. 57-HT-21, Presented at ASME-AIChE Heat Transfer Conference, August, 1957. 46. Handbook of Chemistry and Physics, 38th Edition, Chemical Rubber Publishing Co., Cleveland, Ohio, 1956. 47. Levy, S. "Generalized Correlation of Boiling Heat Transfer." Trans. ASME, 81, Series C, No. 1, (February, 1959), 37. 48. Forster, K. and Greif, R. "Heat Transfer to a Boiling LiquidMechanism and Correlations." Ibid, 43. 49. Hirano, F. and Nishikawa, K. "The Phenomena of Boiling - Part L'. Kikai No Kenkyu (Machinery Research), Japan, 8, No. 3, (1956), 317. 50. Hirano, F. and Nishikawa, K.'Boiling of Hard Water." Trans. SocO Mech. Eng., Japan, 19, No. 88, (1953), 33. 51. Hirano, F. and Nishikawa, K. "The Phenomena of Boiling - Part 2': Machinery Research, Japan, 8, No. 4, (1956), 431. 52. Perkins, A. S and Westwater, J. W. "Measurements of Bubbles Formed in Boiling Methanol'." AIChE Journal, 2, No. 4, December, 1956), 471. 53. Gaertner, R. F. and Westwater, J. W. "Populations of Active Sites in Nucleate Boiling Heat Transfer." Preprint 105 presented at Third National heat Transfer Conference, ASME-AIChE, August, 1959. 54. Hirano, F. and Nishikawa, K. "The Phenomena of Boiling - Part 3," Machinery Research, Japan, 8, No. 5, (1956), 521. 55. Treshchov, G. G.'"An Experimental Investigation of Heat Exchange Mechanism with Surface Boiling Water," Teploenergetika, 4, No. 5, (May, 1957), 44.

-17156, Nishikawa, K. and Urakawa, K. "Experiment of Nucleate Boiling Under Reduced Pressureo' Trans. Soc. Mech. Eng., Japan, 23, No, 136, (December, 1957), 935. 57. DeBortoli, R. A, Green, S. J., et al. "Forced-Convection Heat Transfer Burnout Studies for Water in Rectangular Channels and Round Tubes at Pressures Above 500 psia." AEC Research and Development Report WAPD-188, Westinghouse Atomic Power Division, Bettis, June, 1958. 58~ Averin, E. K. "The Effect of the Material and of the Mechanical Treatment of the Surface on the Heat Exchange in the Boiling of Water." Izv. Akad. Nauk SSSR, Otd. Tekh, Nauk, No. 3, (1954), 1168 59. Zysina-Molozhen, L, M, "Some Data on the Number of Centers of Vaporization in Boiling on Industrial Heating Surfaces."' Problems of Heat Transfer During a Change of State: A Collection of Articles, Edited by S. S. Kutateladze, Publication of State Power(:.Press, MoscowLeningrad, AEC-tr-3405, (1953), 1558 60. Griffith, PO'"Bubble Growth Rates in Boiling8" Technical Report No. 8, Mass, Inst. of Tech., Div. of Ind, Corp., Cambridge, Mass, June, 1956. 61. Donald, M. B. and Haslam, F, "The Mechanism of the Transition from Nucleate to Film Boiling." Chem. Engo Sci,, 8, (1958), 287. 62. Gunther, F, C. and Kreith, F,'Photographic Study of Bubble Formation in Heat Transfer to Subcooled Water;" Heat Transfer and Fluid Mechanics Institute, Preprints of Papers, Stanford University Press, Stanford, Califo, (1949), 1130 63. Farber, E. A. and Scorah, R, L, "Heat Transfer to Water Boiling Under Pressure,"' Trans. ASME, 70, (1948), 369~ 64, Daily, J. W. and Johnson, V, E,, Jr. "Turbulence and Boundary Layer Effects on Cavitation Inception from Gas Nuclei." ASME, Paper No. 55-A-142, November, 1955, Also, Technical Report No, 21, Hydrodynamics Laboratory, M.I.T., July,: 1955.

UNIVERSITY OF MICHIGAN 3 901 5 03483 64061111111 3 9015 03483 6406