O\ __;_;ZEOAK ""IDGE NATIONAL LABORATORY R operated by s e ""‘;:V.'UNION CARBlDE CORPORATION . NUCLEAR DIVISION s AU fo,. Ihe e | . U_S.;ATOMIC ENERGY COMMISSION | ORNL m 2953 - RECEWED BY oTiE S asimh T @ A NEW APPROACH TO THE DESIGN OF STEAM GENERATORS L i FOR MOLTEN SALT REACTOR POWER PLANTS _; - ) arres - Nfllltf Ti‘us document 'contmns information: of @ prehmmury nafure_-}-— L ‘and ‘was ‘prepared pmncnly for internal use at the Oak Ridge Nationol - = © 7 . Laboratory. lt-is -subject to revision or correction ond therefore does S - .not rePresent a final report.” - . - ISTRBUTON OF THS DOCUNET IS URLATED This report. was prepared .as an account of work sponsored by.the United | - . -~ .. o States Government. Neither the United States nor the United States Atomic | -~~~ 7 = &~ Energy Commission, nor any _df their employees, nor any of their contractors, B : C ' subcontractors, or their employees, makes any warranty, express or implied, or | ) - o . gfl assumes any legal liability or responsibility for the accuracy, completeness or | =~~~ = = - S . usefulness of any information, -apparatus, product or process disclosed, or | = - ' : RN ' represents that its use would not infringe privately owned rights, ' ) ks 4 ‘f' 4) ORNL-TM-2953 Contract No. W-TL05-eng-26 A NEW APPROACH TO THE DESIGN OF STEAM GENERATORS FOR MOLTEN SALT REACTOR POWER PLANTS A. P. Fraas © | This report was prepared as an account of work : .} sponsored by the United States Government. Neither i the United States nor the United States Atomic Energy "1 {1 Commission, nor any of their employees, nor any of I their contractors, subcontractors, or their employees, ¢ { makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, com- ‘pleteness or usefulness of any information, apparatus, product or process disclosed, or represents that its use .} would not infringe privately owned rights, - SN ~ JUNE 1971 OAK RIDGE NATIONAL LABORATORY Oak'Ridge, Tennessee Operated by UNION CARBIDE COPERATION - for the U. S ATOMIC ENERGY COMMISSION DISTRIBUTION OF THIS DOCUMENT IS UNLI . ¢ <) «j) ABSTRACT ...... tisecessieocrsnrnonrosuy e Ceeeaieiecsiacnaras v INTRODUCTION e eeesmenennanannnenns . ... REVIEW OF BOILER DESIGNS PROPOSED FOR MOLTEN SALT REACTOR ..... ‘o SYSTEMS Conventional Shell-and- Tube Heat Exchangers e eneneeann eeneoe Supercritical Pressure Unlts Ceececccsssaneannas ;...;........ Double-Walled Tubes with & Heat DA o.le.eueneen.. celiesanes Flash Boiler .....;..,....g....,........} ........ e Loeffler Boiler ....... esesienenns Cesseaanaa fi....,....; ...... ; Triple Tube Boiler ...ccevess cieseans chedsectrsrence vecassseas Reentry Tube Boiler ............ e enene s et eeraeens ' DESIGN FOR GOOD STAEBILITY AND CONTROL CHARACTERISTICS . .ovvewen... Molten Salt Temperatures at Part Load . Ceetasseseneneea ces Effects of Mode of Control on Steam Temperatures at eresssens Part Load : Heat Tragnsfer Instdbilities and Thermal Stresses ............ Flow Stability'Considérations Ceetieisietisetatrae s e Heat Transfer Analysis ,...; ................. Leetersaceasanna - Typical Calculatiofis e PR [ Heat Transfer CORFFficients .........e.ee... e Temperature Differences'and Heat FLUXES +.oovvveennnnnns | Pressure DroD v.veeeeevessoeneeeeens e ieeaaeeeens SEstlmated.Perfonmance”Characteriétics eereeientaaresaeaanne Effects of DeSign'Heat‘Load on Tube Length ....... PRI Effects of De31gn Heat Load on Temperature .......ceoe.. Dlstrlbutlon , _ Effects of Part Load Operation ........ .......;.;..;..;; Pressure Drop and ‘Pressure Distribution .,.; ........ e - Effects of the Fraction of Heat Added in the e ean e Inner Tube 7 Effects of Slze Of Heat Increment Used .o.eeveveeeeennns _ Turbine Control Considerations .;..........,....;;.; ......... ILimitations on Rates of Change in Ivad .......ccv0eveeee Startup and Rates of Change of System Possibility of Eliminating the Throttle Valve ......... . Proposed Design for a Molten Salt Reactor Plant ........ ceoae ' Reheateré cereecratasiaeaas s ......;..;..........;... General Description .......... e e ereenanaaaas ;.,. Headering Problems ....;..Q....,..................,....; Differential Therfiél Expansion {..:....' ...... ceseeene ves ‘Geometric and Performance Data .......... e eeeneenetees _ Cosf Estimate- .............. ; ............................ Conclusibns ................................................. Récomméndgtions ........................................... .o REFERENCES ..¢veeveeenncccnaasca cietsescensesassanus . » iv Load with Proposed ...... Page 51 53 54 55 ° 55 56 - 61 65 65 67 B e &) o) @’ - C _ ) ) A NEW APPROACH TO THE DESIGN OF STEAM GENERATORS - FOR MOLTEN SALT REACTOR POWER PLANTS A P, Fraas ABSTRACT | A new type of steam generator has been dev1sed to meet 'the special requirements of high-temperature liquid metal and - molten salt reactor systems, The basic design concept is such that bolling heat transfer instabilities and their attendant severe thermal stresses are avolded even for a temperature - difference of as much as 1000°F between the feedwater and the high-temperature liquid, thus giving good control characteris- tics even under startup conditions. This is accomplished by 'employing a vertical reentry tube geometry with the feedwater entering the bottom of the inner small diameter tube (~1/4 in. diam) through which it flows upward until evaporated to dry- ‘ness. The slightly superheated steam emerging from the top of the small central tube then flows back downward through the annulus between the central tube and the outer tube. A por- tion of the heat transferred from the high-temperature liquid - to the superheated steam in the annulus is in turn transferred - to the water boiling in the central tube. Design studies in- ‘dicate that this type of boiler not only avoids thermal stress and salt freezing problems but it also gives a relatively com- pact and inexpensive construction. Further, it appears to make possible a simple plant control system with exceptlonally good plant response to changes in load demand " INTRODUCTION It has been'apparenthinceVearly in the'molten salt reactor develop- ment work that the high melting point of the fluoride salt sultable as fuel for molten salt reactOrS”coupled with the thermal stress problems "inherent in high—temperature liquid systems pose some exceedingly diffi- '“cult problems in the design of steam generators.! These are compounded - by the complexities of two—phase flow and heat transfer problems under 'rboiling conditions, possible difficulties with b01ling flow instabilities, and_thevproblems of obtaining good boiler operating characteristics for a wide range of both full power design steam temperatures and pressures and under the much reduced'pressare‘and temperature conditions inherent in startup and part load operation, Many different attempts to design boilers for molten salt reactor systems hare been made,! ® but each of the approaches proposed has had some serious disadvantages. The startup and part load control problems in particular have been so formidable, in fact, that3nofattempt:hashbeec made tQVSO1Ve‘tfiem for many of the'designs that have been proposed‘—-Oniy full load'desdgn conditions-hare been con- sidered. It is belleved that the new. reentry tube concept proposed in this report w1ll yleld compact, economical b01lers ‘that can be designed | for any full power de51gn steam condltlons and yet w1ll glve good stability and control characteristlcs over: the full range from zero power to full power“condltlons,and further Wlll notppresent dlfflcult.thermal stress or salt freezing :problems.', - - | . A draft of this report substantially as it now stands was prepared and distributed July 15, 1968 to'key people. ifi the molten salt reactor project at ORNL' . They klndly rev1ewed the draft and suggested a number of addltlons to help clarlfy thls new approach By far the most serious reservatlons they had were concerned with boillng flow stabillty at low loads. Thls requlred SOme sort of a test, and the report was held pend- ing the availability of funds for a projected test rig. Because of lack of funds this rig has hot yet been built. | ' _ It happened that basically similer requirements for a steam generator arose last year in a program to develop a small isotope'power unit. These requirements made it importaotlto test a low steam pressure, short tube version of the reentry boiler. The results of the tests are now in, and good performance was obtained.” Even though these results cover a limited range and the steam generator proportions are substantially'different from - those proposed here for a large central station, the basic concept appears -to be validated for the low load portion of the operating.regime that was most open to question. As a consequence, it was decided that the-report should be issued. Q. <) g}/ REVIEW OF BOILER DESIGNS PROPOSED FOR ~ MOLTEN SALT REACTOR SYSTEMS ' Thesvarious boilers that have been considered for use with molten salt reactors might-be.grouped in the categories outlined in Table 1. The principal features and major advanteges and disadvantages of each ‘are summarized very briefly to help point up the problems and provide & framework for the subseqnent analysis of thetheatftransfer, thermal stress, and controlrproblems. Conventional Shell-and-Tube Heat Exchangers A logical first candidate is a shell-and-tube heat exchanger with the salt inside the tubes and the water boiling outside of the tubes. Two_configurations-are'included in Table 1. - The hcrizontal'U—shaped tube and casing'geometry_employedifor_the early pressurized water reactors (see Ref. 8, p. 197) has the advantage that not only is differential ther- mal expansion between the tubes end the casing readily accommodated'With- out producing serious thermal stresses, but the difference between the hot fluid'inlet-and outlet temperatures will not induce severe stresses in the header sheet as would be the case for @ simple shell-and-tube heat ex- changer with a'U—tube'configuration; _The'majorvdifficulty with this con- - ventional Ueshaped'casing_design'iS'tnat the temperature'difference be- tween the water and the walls of the tubes carrying the molten salt is . far greater than the temperature difference in the nucleate boilihg regime so that an unstable vapor blanket would form”between the liquid and the ~hot metal surfaCe~(See*Figeil) This leads'to unstable, noisy operation and severe thermal stresses 1n the tubes as a consequence of violent ir- ;regularitles in the heat transfer coefficient . Supercritical Pressure Units It has beenisuggestedsthat these difficulties might-be reduced through the use of a supercriticel water system inasmuch as this would reduce the temperature difference between the feedwater and the molten 4 Table 1. Types of Btesu Generstor that Have fleen Proposed for Use with Molten Belt Reactors Type of Beat , L "' Shell-Side Tube-Side Iten Exchanger . Gecnetry Flud Fluid Major Problems 1 Conventionsl Shell. and-tuba, U-tube Water Salt Excessive temperature difference be- tveen salt and steam gives unstable boiling and/or possible freezing of salt. Tempersture difference between salt in- “let and outlet streams causes large thermal ltrenel in hender sheet lnd : casing. Salt Excessive tem;perlture difference be- _ tween salt and steam gives unstable boiling apd/or possible freezing of salt, Water 2 Blull-md-tube vith U-tubes in & [~shaped cssing for luhcritical pressure steam 3 Shell-and-tube with U-tubes in a U- shaped casing for supercritical pressure H,0 Sinilar to above but problems less severe at full load, but still serious at part load and in startup; large ateam preheater required to add 20% of heat as preheat, A Dpouble-walled . 8imilar to Item £ but with tubes fubri- . T Severe thermal stresses in porous tubes vith a cated as in the detail shown below: Outer tube metal region would cause cracking heat dan . and indeterminately urse thermal barrier. 5 Flash Boiler large number of tube-to-header joints required. Not well suited to steanm pressures above sabout 600 psi. Tends to give large heat flux and local ealt ‘freeging near spray nozzle end. 6 loeffler Boller large stean drum required coupled with heat exchanger and steam pump mekes the equipment expensive. Operation is inherently extremely noisy and excites vibration. ‘? Triple Tube Boiler No suitable vapor separator is avail- o able, and experience indicates that it is unlikely one can be developed. Outer tube diameter is inherently large and thus requires & thick wall; this léads to low heat flux and large - walght and investment in tube material. - Recirculating Water Insulation T f Final ¥ B Buperheat Tube Sheet Bection Peedwater 8 Reentry Tube Boiler i ' - _ : Salt | Steam Concept hes not been tested at .. - - . = ) pressures above 200 peia. . O §) $5 _ - ORNL DWG. T1-5T2h ~<—— Interface evaporation ———>t<——— Bubbles >te<——— Film < : 1' — >t J] —>of—— III'—->--<—1'V——>-<- V >4=<— V] — K 8 ~|E . §sc| £ |2 1 L £EE fi£§ = B8EE| 2 S: - w | Pure convection ~ heat "?fig’ 885/ H2 \ =8 &€ |== | transferred by superheated wd=|825/ S|E EEgl = |E8 5| liquid rising to the. £3¢(=° BE® \VWSF 8 |gs2 £ liquid-vapor interface 2= 5 e = 2| where evaporation gEg> - - | takes place | SES o oo S = 8.2 o 5 B —~—— poling W — o 0110 10 100 — 100 T,-T,, °F —> : Fig. 1. Diagram Showing the Principal Pool-Boiling Regimes and Their - Relative Position on a Curve for the Heat-Transfer Coefficient Plotted as a - Function of the Film Temperature Drop (Farber and Scorah, Heat Transfer to Water Boiling Under Pressure, Trans. ASME, p. 369, Vol. 70, 1948)" | salt, reduce the fraction of heat added at low water temperatures, and reduce the sharp changes in physical properties associated‘fiith_the phase change from liguid to vapor.2°® In units of this type it is usually con- sidered best to have the molten salt outside the tubes and the super- critical watér and steam flowing inside the tubes. However, while there is no‘sharp phase change, the very rapid changes in density and othér physicalAproperties near the critical temperature under supercritical pressfire conditi@ns lead to marked changes ifi the heat.transfer perform- -~ ance and abrupt reductions in the heat transfer coefficient something like those associated with the burnout heat flux encountered at subcritical pressures.? In addition to these flow and heat transfer phenomena and the boiling flow instabilities that would be associated with them, there would be large, irregularly'fluCtuating thermal stresseé in the tube wall ~in the region near the feedwater inlet of thé.boiler unit, and these would be.likely to cause tube cracking and failure. These problems are dis- cussed later in some detail. | - Although the use of a supercritical pressure system reduces the severity of the boiling flow stability pfoblem.for operation near the de- sign point, a great deal of difficulty has been encountered in ali of the coal-fired once-through and supercritical pressure boilers in going from zero power to part load conditions of at least 10%, and often to as high as 30% power. These difficulties stem in part from the'large_density | change as the water-steam mixture. flows through the boilers and in part -from the reduced pressure drop at the lower flows which reduces the damp- ing of the oscillations by'turbulencé losses.® ‘Double-Walled Tubes with a Heat Dam ‘M. E. Lackey suggested in 1958 that one means for.reducing the tem- perature difference between the tube wall ahd the water to avoid film boiling conditions wouid be to incorporate alheat dam in the form of a sintered powder matrix between an inner tube carrying the molten salt and the outer tube in contact with the boiling water.* This approach would be quite effective at full'power where the average heat flux would be <) o) x) high, but wonld:presentfproblems_under startnpand low power conditions -becausera'high heet flux is inherently associated with & large tempera- ture drop through such a heat dam..iA,further disadvantage is that dif- ferential thermal_expansion between the inner and outer tube walls would ~induce severe-thermal stresses in the sintered buffer material, and these would lead to cracking of the sintered matrixland'unpredictable increases in the thermal res1stance.87 , _ ‘The cracking problem 1n the s1ntered materlal could be avolded hy using 1nstead an air space asra heat dem. The OD of the inner tube could ~ be knurled, for example,-and the outer tube swaged down onto it. This would have the dlsadvantage that dlfferentlal thermal expans1on between the inner and outer tubeS‘would probably loosen the swaged joint and give an 1ndeterm1nately large -probably excessively large — thermal barrier at full load. If this did not happen, thermal stresses would probably. | cause cracklng of the tubes. ‘Flash Boiler In an effort tOVavcid"the difficulties outlined above, a flash boiler was prOPOSed'in 1955, 1In a ‘unit of this type the molten salt would flow outside of the tubes and feedwater would be 1n3ected in the form of a long thin plume of,flne spray dlreoted‘along the axis of the boiler tube. '.Experience in the development of noZZleS for diesel'engines indicates that @ high penetration spray could be obteined with sharp-edged, single- “orifice nozzles_having;a'hole.diameter of about 0.020 in.,° and that no2z1es of this type would'break'the liquid:up into_droplets having ‘a . diameter of the order of 0. 005'in 11 These droplets would form a long - 8lender spray plume that’ would extend for perhaps 2 £t down the bore of a 1/2 in. tube. Droplets would impinge on the tube wall at.a very low 'fangle of 1nc1dence, and’ would tend to skltter along the wall riding on __”a'thin film of Vapor. Analyses indicated that the local thermal stresses associated with the cold. traoks left by droplets of this sort would be - well within the elastlc llmlt and should not glve dlfflcultles with thermal stresses. * A brief series of tests to investigate this concept was run by an MIT practice school group.}2 These tests showed that there was a strong tendehcy for a large fraction of the droplets to impinge on the tube wall in the region close to the 1n3ectlon nozzle, and that this led to such a ' pronounced cooling effect that a frozen film of molten salt tended to form on the outside of the tube in that reglon.la“‘The ‘tests had been initiated on the premise that they offered an attractive way to provide for emer - wgency cooling of the ART fuel dump tanks, hence the test work Was termi- nated with the demise of the ANP proéram.” No further work on the concept was carried out because it inherently requires a very large number of tubes of rather short length so that the tube-to-header joint costs tend to be- come excess1ve.' Further, the concept does not appear to lend 1tself well ‘to hlgh pressure and supercrltlcal pressure steam systems. Loeffler Boiler The Loeffler boiler concept used in @ few coal-fired steam plants = has been considered.® Systems of this sort have been built and operated for coal- fired furnaces. Operatlon apparently has been satlsfactory ex- | cept for the extremely high nozse level assoc1ated w1th the b01ler and dlfflcultles with the steam pumps required. The Loeffler concept entalls admission of saturated steam to & heat transfer matrlx heated by molten salt. The superheated steam leav1ng thls matrlx would be d1v1ded 1nto two portions, one of whlch would flow to the turbine and the other would be returned to a boiler drum where it would bubble through the water in ~the drum. ThlS approach has the dlsadvantages that 1t requlres large and expensive boiler drums, 1mplos1on of the vapor bubbles in the boiler drum makes operation extremely n01sy and 1nduces v1olent v1brat10n ex- citing forces, the relatlvely poor steam-s1de heat transfer coefflclent .and low enthalpy rise lead to a large number of tube-to-header JOlntS, and steam pumps posing tough design and rellablllty problems are required to recirculate steam through the boiler. On the other hand, the system has the advantage that it lends itself readlly to startup and part load operations, it wvirtually eliminates the possibility of salt freezing &s w ) a consequence of excessivescooling in the steam generator, and it greatly eases the thermal stress problems by substituting a salt-to-steam heat ex- changer for the salt-to-water boiler. Triple Tube Boiler _ When the writer solicited criticisms and-suggestions on‘the proposed new boiler concept, M. E. Lackey pointed out that B. Kinyon and G. D, _Whitman had proposed & somewhat similar boiler in 1960, (Ref. 6), and S. E. Beall pointed out that a variation of this approach had been tested as a means for cooling fuel dump tanks."_l3 The arrangement proposed by B Kinyon employed three concentric tubes with boiling water flowing upward ;_through the inner annulus to a vapor separator and superheated steam flow- ing down through the outer annulus The central passage would serve to return the water from the vapor separator to a boiler water recirculating | pump. This arrangement has the advantage that the superheated vapor in the annulus between the outer tube heated by the molten salt and the tube . containing the boilingpwater would act as a buffer both'to eliminate ser- ious thermal stresses and to avoid excessive'metal temperatures adjacent to the boiling water. The arrangement has the disadvantage that it re- quires a fairly large tube diameter and hence & fairly large tube wall thickness for supercritical water systems. Thus, the amount of heat trans- fer surface area required tends to be large because the principal barrier to ‘heat transfer lies in heat conduction through thick tube walls. Fur- _'ther, the arrangement requires the development of a vapor separator that - would fit within a small diameter, preferably that of the tube. Experience in vapor separator development indicates that the velocities required for good vapor-liquid separation are much lower than. those one would like to - use in the tube for heat transfer purposes, and hence a rather bulky pro- tuberance would have to be:employed at the end of each tube; this appears %o lead to a set of extremely avkward mechanical design problems. 10 Reentry Tube Boiler The boiler proposed in this report is somewhat similar to the one described above.but differs in that it makes use of only two vertical, concentric tubes in the form shown in Fig. 2. The water enters at the ‘bottom through a central tube having a dlameter of about 1/4 in. Pre- 'heatlng and b01llng occur as the water rises 1n thls tube untll evapora- tion is complete, after which there is some superheatlng. The steam emerges from the top end of the small dlameter tube, reverses direction, and flows back downward through an annulus between the 1nner small tube and an outer tube having an ID of ‘around 1/2 in. The molten salt enters ~at the bottom, flows upward around the outer tube, and out the top With this arrangement there is only one header sheet separatlng the molten salt from the atmosphere, and this header sheet is not subaect to a large pres- sure differential. Thermal sleeves would be used at the header sheet to minimize thermal stresses (see Flg. 3). There'would“bevno"high pressure header sheets in this system, the tubes for the hlgh pressure feedwater and the exit steam would be manifolded as in hlgh pressure coal fired boilers rather than run into header sheets. The_steam annulus between the inner and outer tubes would'act'as.a'buffer to isolate the relatively low- -temperature boiling water region from the high-temperature molten salt. This isolation would be so effectlve ‘that 1t would be qulte poss1ble to | heat the unit to the molten salt operatlng temperature w1th no water in the system and then slowly add water to 1n1t1ate boiling. CAs will be shown later, it should be possible to design the unit so that it would operate stably'over a wide range of condltlons from zero load to overload with no difficulties from thermal stresses or freez1ng of the salt. It might at first appear that the extra heat transfer films through Wthh heat must be transmltted from the molten salt to the boiling water might lead to a large increase in surface area requirements and hence in the size, weight, and cost of the unit. However, it appears quite pos- sible to design so that these disadvantages are more than offset by such features as the absence of a high-pressure header sheet and the ability to operate with high-temperature differences between the molten salt and the boiling water so that the overall size, weight, and cost of the unit 2 11 - . ORNL DWG. T1-5T25 ) Steam salt Water Fig. 2. Diagram,Showing Section-Thfohgh_?Ube: ] 0 12 5 \O 6 x o O m m o 0 B . mw . n“mw, . _ , g o pumw A mwg, g mw mm 9 & m 2 S 2 B 2 o { A T T T T T T T N T T T T L T T T U T T T TT y _ N ~ 3 WHENHIRRAR RN RNV NS . ettt A AN\ ,//Jafaafffdflflfffaidaflffaflflfflffl/l#/%f}?ffldlfllwfltfl/rr¢d1¢~ < MOLTEN SALT AIR - ,-“fi\: Section Through the Header Sheet Region Showing Two Typical Tubes with Their Thermal Sleeves and the Associated Welds. Fig. 3. » | ) 15 are at least competitive with the corresponding values for eny other de- "sign that has been proposed. DESIGN FOR GOOD STABILITY AND CONTROL CHARACTERISTICS The usual proCedure in'deVeloping a design for a steamrgenerator has‘ been to choose a geometry, establish the proportions for full povwer condi- tions, and then — sometlmes -examlne the full range of control problems. The inverse order seems at least equally logical and is followed here. The writer. has_felt from the beglnnlng that some of the most difficult condi- ‘tions to be met are those associated with initial stertup and part load op- eration. rThus, the first step'in.the evaluation work was to establish a - typical set of molten saltiand steam‘temperaturesg and from these, using bas1c heat balance cons1deratlons, deduce the ‘effects of dlfferent modes of control for the various load conditions of 1nterest This approach gives a'valuable insight into the full range of the over—all design problems. ‘Molten Salt Temperatures at Part Load Several different approaches_can'be taken to the control of a molten salt reactor steam poWer'plant. Perhaps the simplest'and most reliable ap- proach is to make use of constant speed ac motors to drive pumps in both the fuel c1rcu1t and the intermedlate salt CerUlt If this is done, the temperature rise in each qalt clrcult w1ll be. dlrectly proportlonal to the -"load so that the c1rcu1ts w1ll be 1sothermal at zero powver. The basic heat transfer relatlons are such that the temperature dlfference between the two - salt 01rcu1ts w1ll also be dlrectly proportlonal to the load and w1ll drop frto zero at zero load. If there is no control rod movement in the reactor, the zero load reactor temperature Wlll be the mean of the 1nlet and outlet fuel temperatures at full load. - These effects are shown in Flg. La for op- (*eratlon with constant: speed fuel and inert salt pumps ~ Note that both the . mlnlmum and the mean temperatures of the 1nert salt rise as the load is reduced — an undes1rable characterlstlc from the standp01nt of the design of most steam generators., - Thls 51tuatlon can be changed by holding the 14 ORNL DWG. T1-5T27 1300 1200 1100 1000 1300 Temperature (°F) 1200 1000 0 20 ko 60 8o . 100 aefQ (%) | Fig. 4. Temperature Distribution in the Fuel-to-Inert Salt Heat Exchanger for & Series of Loads for Two Control Modes, i.e., &) constant epeed pumps and a constant mean fuel temperature, &nd b) constant speed purps &nd & constant reactor fuel inlet temperature. ‘-‘m 0 . ’) 15 reactor inlet temperature constant in which case the temperatures 'in the fuel and inert salt circuits would be defined by the curves in Fig. 4b. The fuel inlet temperature’ ought not be reduced below the value shown be- cause'it is desirable to_maintain'the'fuel at least'100°F above its freez- ing point. Other effects:could be obtained by varying the speeds of the fuel and/or inert salt pumps,.but the resulting complications — particu- larly at or near zero 1oad — are quite objectionable. After analyzing a variety ‘of power plant control modes it was decided that the simplest system would be the most reliable and should be used for - the bulk of this study., The approach chosen is believed to be the simplest possible, i.e., it assumes a constant mean fuel temperature and constant speed pumps for both the fuel and the inert salt irrespective of load. " Effects of Mode of Control on Steam Temperatures at Part Load The inherent_effects'of typical control modes on the temperature dis- 'tributions“that will resultnas aiconsequence of fundamental'heat balance considerations are.shown\insFigsi 5, 6, and 7 for the full range of load conditiOns It can be seen'frOm“examination of these curves that the steam 'outlet temperature will rise as the load is reduced because the temperature difference between the two fluid streams in a heat exchanger drops off with . & reduction in the heat flux. This problem arises in the control of any steam plant that is coupled to a high temperature reactor whose mean tem- ' perature is held constant 14 po avoid damage to the turbine, the steam '-temperature could be reduced at’ part load by introduction of a desuper- heater between the steam generatcr and the turbine The -available desuper- heater units may not be: well suited to this particular application, but ‘the design of a suitable unit would be straightforward However, it is ' apparent that high steam temperatures under part load conditions would ser- iously increase the creep stress problem in the steam system if it were l-to operate with a constant b01ler discharge pressure.‘ The problem could be eased by scheduling the reactor mean temperature 80 that it would in- crease with power output. One way of doing this would be to maintain the reactor inlet temperature-constant and allow the temperature rise in the 16 ORNL DWG. T1-5728 TEMPERATURE ' (°F) 500 & 0 20 60 "8 100 AQ/Q (%) o Fig. 5. Effects of Opergtion with a Constant Mean Fuel Tempera.tm'e on the Temperature Distribution Through the Boiler for Typicel Loading Conditions, Iocal temperatures are plotted as functions of the fraction of the heat transferred to the water from the molten salt (i.e., 22/Q). For this set of curves it was assumed that the effective boiler tube - length would be varied with the load to maintein & constant temperature and pressure at the superheater outlet. x " %) 17 - ORNL DWG. 71-5_729 Temperature (°F) 60 . /e m Fig. 6Q Effects of Operation.with & Constant. Reactor Fuel Inlet - Temperature on the Temperature Distribution Through the Boiler for Typi- cal Load Conditions. Local fluid temperatures are plotted as functions of the fraction of the heat transferred from the molten salt (i.e., - M/Q). For this set of curves it was assumed that the effective boiler ~ tube length would be varied with the load at any given condition to main- tain a constant temperature and pressure at the superheater outlet. 18 o ORNL DWG. T1l-5T30 Q 1200 ' : ‘" 1100 1000 700 TEMPERATURE (%F) 300 200, 20 50 - 8 100 , ' : AQ/Q . Fig, 7. Inert Salt and Steam Temperatures as a Functicn of the | ' Fraction of the Heat Transferred from the Salt to the Steam for e Series of Design Heat Loads with the Steam Pressure Directly Proportiomsl to - o~ the Eeat lLoad, : o : | . U n i) 19 _fuel salt to increase in direct proportion to the load. If this were done, the temperatures in the inert salt and steam c1rcu1ts would vary with load as indicated in Fig. 4b, This control'schedule for the salt circuits was assumed in preparing a few .of the boiler performance estimates presented later in this report to show that this arrangement might reduce or elimin- ~ ‘ate the need for desuperheating the steam under part load conditlons at the expense of camplicating the reactor control prdblems The curves of Figs. 5 and 6 were calculeted on the basis that the boiler would be operated in'the conventional fashion with a.constant steam discharge pressure of 4000 psi, and the pressure of the steam supplied to ~ the turbine would be reduced by a throttling valve. However, there are a number of advantages aSsociated with operating a steam boiler-turbine- condenser-feed pump systEm.with no'throttle valve between the boiler and the turbine so that the steam pressure is determined by the flow rate through the critieal pressure drop orifice represented by the inlet nozzle box of the first stage of the turbine. One of the more important of these advantages in this‘instanee:is_that;'if the steam system were designed so that the boiler discherge pressure weuld be directly proportional to the load, the higher boiler temperatures would be associated with lower pres- sures, and the creep stresses in the boiler tube wall would not be exces- sive. If this were done, to a first approx1mat10n the boiler pressure will be directly proportlonal to the load, and curves for the steam temperature - as a function of the emount of heat added on the water-81de will be as in- ‘dicated in Fig. 7'for_a”typical.ease. .th‘surprisingly, calcuwlations pre- -sented later in the report show that‘the_tube-length to evaporate to dry- ness is much the same'irrespective of the boiler pressure at a given load. —However, the pressure drop through the boiler under part load conditions fis much higher 1f the pressure is directly proportional to the load than if the boiler were Operated at a constant pressure. Increasing the pres—' sure drop at part load is edvantageous in that it will much 1mprove the boiling flow stability. | Heat Transferiinstabilities and Thermal'Stresses_ One can sense intuitively that severe thermal stresses might be induced by the wide variations in the heat flux —-and hence the transfer coefficient 20 that can occur with changes in the difference in temperature between the- k&#) metal wall and the saturation‘temperature of a boiling liquid (see Fig. 1). - HoweVer, it is not obvious just how these thermal stresses may be related to fluctuations in the boiling heat transfer coefficient, how large they' may be, or why they may be more severe-at‘low loads than at the design point, and M. Rosenthal asked that this short- sectlon be .added to clerify the problem, particularly for a unit de31gned for supercrltlcal operation. Tt should be noted that for some time it was thought that this problem could be avoided by going to supercritical water pressures, but severe cracking of tubes in coal-fired supercritical boilers showed that, un- fortunatély, this is not the case.? Detailed investigations of-boiling heat transfer relations in the supercritical pressure regime have shown that large variations in heat transfer coefficient still occur, particularly at high heat fluxes, i.e., if there is a large température difference be- tween the metal wall and the bulk -free stream? (sge,Fig. 8). | To illustrate the problem, consider a short section of INOR-8 tubing with supercritical pressure water at 690°F inside and molten salt at 1150°F - ‘flowing outside the tube with a selt heat transfer coefficient of 1000 Btu/hr-ft? -°F. The thermal conductivity of the wall is about 12 Btu/hr.-ft-°F. The thickness is 0.10 in., and hence the conductance of the wall would be about 1440 Btu/hr-ft®-°F, end thus the temperature drop through the tube wall would be about 70% that through the salt film on the outer wall. Two very different operating regimes are possible. - Aésuming that the curves of Fig. 8 define the heat transfer situation on the water side, the heat flux -where the water-steam enthalpy ran 780 Btu/lb could be ~100,000 Btu/hr or 8t & nearby point downstream where the enthalpy reached 900 Btu/lb it could - be 150,000 Btu/hr The resulting film and wall temperatures and tempera- ture drops can be summarized in Table 2. Changing the radial AT through the tube wall from.70°F to 105°F at power, and to O°F at no load would lead to dlfferentlal thermal expansion between the inner and outer surfaces and hence to both. c1rcumferent1al and axial stresses that would be superimposed on the basic pressure stresses. Power cycling and changes from one heat transfer regime to the other would cause thermal strain cycling, and this could eventually.lead to cracking * and failure. The problem would be much‘worse at subcritical pressures ./ ¥) 1) 21 AR | ORNL DWG. T1-2175 1200 T T T 1 | nrs |- ~ STEAM AT 3300 PSI> - nso |- 6 = 340000 LBS/FT2-HR — | - | D= 0033 FT S - —— EXPERIMENTAL noo |- | CURVES (SHITSMAN) - —— — MACADAM'S CORRELATION (BULK PROPERTIES) — 1075 |- 1050 |- 1025 1000 975 950 925 900 875 850 WALL TEMPERATURE, °F 825 800 775 __ g i [ 600 700 800 900 1000 1100 1200 ' BULK ENTHALPY, BTO /L8 Fig 8. Deterioration in Heat Transfer Near the Critical Tempera- eeture in Supercritical Pressure Once-Through Steam Generator Tubes Operat- ‘ing at High Heat Fluxes. '(M‘~E Shitsman, Impairment of Heat TTansmissiOn 'at Supercritical Pressures, Teplofizika Vysokikh Temperatur, Vol. l, No. 3‘ p. 267, 1963) oo Table 2. Effects of Heat Flux on the Radial Temperature . Distribution Through an Element of Tube Wall in the Inlet Region of a Simple Shell-and-Tube Molten Salt Steam Generator at Supercritical Pressure Conditions Water-steam enthalpy, Btu/lb ' 785 " 900 Heat flux, Btu/hr-ft? | 100,000 150,000 Salt free steam temperature, °F 1,150 . 1,150 Tube_Ou#er wall temperature, °F - 1,050 1;000 Tube inner wall temperature, °F 980 895 Water temperature, °F ' - 690 705 Salt film AT, °F | | 100 150 Wall AT, °F 70 105 Water film AT, °F | 20 190 where the change in heat transfer coefficient with enthalpy is both larger and more abrupt (see Fig. 9). The problem could be eased somewhat by re- ducing the molten salttemperature to 1000°F for startup and low power conditions, but it would not be eliminated. Observations of tubes transferring heat:to water at supercritical pressures have shown that both of the regimes of Table 2 will be present i. e., some sections of the tube will operate at a lower wall temperature and a high heat flux while others will operate at 8 hlgher temperature and a lower heat flux. Further, these regimes tend to shift back and forth axially along the tube with changes in water flow rate. This 1eads-to another type of thermal stress. Dilation of the hot regiou relative to the cooler region leads to bending stresses in the tube wall, end these stresses are likely to be severe because the shift from one heat transfer regime to the other tends to be abrupt and the transition zone is short .These stresses are analagous to those in thermal sleeves, and can be cal- culated in the same way (see Ref. 8, p. 125). Unfortunately, data for typical exial temperature distributions are not at hand'to provide a good _basis for estimating the'magnitude of these stresses. o ’_ At first thought it would appear that at light loads the above prob- lems would be eased because the average heat flux would be greetly reduced. - 23 ORNL DWG. T1-5731 Nucleate - Convective ' ~ Convective { boiling . boiling -t fim N Boiling =~ Liquid Poling steam A starts - ~ deficiency | High heat_ flux ~ '-'_ «—Burn-out limit [ Medium heat flux £ P T T 2 | £ [ 8 P l o ' | | 2 | | 5 a 3 l g Il | 3 N k | T 1 | o | | | s \ o=} : Burn-out N <-Subcooled water—>- -<———Steam-water—-————>- -s—Superheat—» 4 . Enthalpy . T L - Saturated - - . Saturated water e | ~steam sttance from mlet — -Fig 9, Effects of- Heat Flux on the Heat-Transfer Coefficient for .,__-a.WOnce-Through Boiler Tube (Polomik et al., Heat Transfer Coefficients ~with Annular Flow During "Once-Through' Boiling of Water to 100% Quality at 800, 1100, and 1400 psi, Jour of Heat Transfer, Trans ASME, p. 81, vol. 86-2, 1964) 24 However,rthis is not the case. The bulk of the heat transfer oéCurs at the water inlet end because the temperature difference there is inherently high irrespective of load, and it is the high—témperature'difference that gives the possibility of two drastically different temperaturé regimes. Thus réducing the heat load simply reducés the length of the regiofi in which severe thermal stresses may be induced — it does not eliminate the problem, In fact, if the pressufefis reduced, the boiling point’bf the water will drop and the témpéréfiure difference that can-be_induced in the tube wall will be increased. | - The above effects are COmplex‘and, in many respects, rathgr subtle, but they are the reason for turning from the conventional shell-and-tube heatxeXChanger geometry to the reentry tube construction proposed in this report. - | ' Flow Stability Considerations In a conventional coal-fired bbilér the pressure drop in'the boiling region is rather low in recirculating boilers, but the overall pressure drop is fairly high because the préssure drop through the superheafer is substantial. In once-through boilers, particfilarly-in supercritigal pres- sure steam plants, the bo;ler pressure drop is large, commonly 20% of the boiler inlet pressure. This stems in part from efforts to get a high heat r_transfer coefficient on the steam side, in part from the long tubes made necessary by the relatively low average heat transfer coefficient on the combustion gas side, and in part by stringent ofificing at fhe tubé inlets to assure a flow distribution across the tube bank such that it will be possible to avoid burnout in regions where the local heat flux may be high as a consequencé of irregularities in the hot gas flow on the'cbmbustion‘ gas side o£ the boiler. These irregulafities in the local hot gas tempera- 'tUre and heat flux vary substantially with the heat load onlthe boiler, the fuel used, and the peéuliarities and irregularities in the gas turbu- -lence pattern in the combustion zone. ‘Fbrtunately, in a molten salt-heated - boiler not only can the tube wall never exceed the temperature of the molten salt so that severe ofierheating of the tube wall is not a yroblem, but the molten salt temperature and flow distribution can be predicted O 0 t 25 within quite.close\limitslinstead of being subject to the vagaries in the large-scale turbulence that are characteristic of combustion zones in fur- naces. As a consequenCe,'a steam.generator for a molten salt reactor can be designed to give a much. hlgher average heat flux and yet a lower peak “heat flux than can ‘be obtalned in a conventional coal-fired boiler. This directly reduces the tube length and‘hence the pressure_drop. In addition, it reduces the.need for*orificing to control the water flow distribution through the boiler. As a consequence, it is believed that it will be pos- ;s1ble to de51gn for lower water pressure drops through steam generators for ‘molten salt reactors than are,ordinarily required for coal-fired boilers. This will reduce both the power requirements for the boiler feed pump and the tube wall thickness required for the feedwater piping. Heat Transfer Analysis Estimating the heat transfer performance of’the reentry tube steam ‘generator involves a set of implicit relations that make an explicit solu- tion out of the question and even an iterative solution surpr131ngly tricky. This stems from the wide varlety of combinations of condltlons that may occur in the boiler depending on the steam pressure, temperature, and flow '{'rate. The problems have much in common with, but are more ‘difficult than, - those of steam generatOrs for hlgh -temperature gas-cooled reactors. 14 The steam conditions were chosen to be essentially the same_as ‘those of Eddy- stone Unit No. 2, which. was . used as the basis for an earlier study 15 fThe feedvater temperature. and flow rate, the exit steam pressure, and the mol- - ten salt inlet and outlet temperatures are ordlnarlly given. From these 'data it is pos51ble to estlmate a steam outlet. temperature ‘and from heat balance considerations draw a set of curves such as those of Fig. 5 which ,show the temperatures of the molten s< and the steam as functlons of the ' fraction of heat addedrtoithersteam in the course of its transit through the boiler. Figure 7*sh¢ws a similar set of curves'for'a series of lower fpressures and lower loads, 1n thls 1nstance the pressure was taken as di- o ,rectly proportional to the load a good approx1mat10n to the natural charac=- teristics of 8 steam b01ler—turb1ne condenser-feed pump systemrln which no - throttle valve is employed. Figures 5 and 7 illustrate one of the - 26 difficulties in setting up an iterative calculation. Whereas:the.steam témperature for the supercriticél condition of Fig. 5 is uniquely defined as a function of the fraction of heat added to the steam, this is not the . case for suberitical Steam_pressure conditions where the steam temperature is essentially independent of the amount of heat addgd over a wide range ~of heat addition. In attempting an iterativersolution one can celculate stepwise upward from the bottom of the tube using as his points of depar- ture the given feedwater inlet conditions and the assumed steam outlet con- ditions. The stepwise calculétions_can be continued to a point at the top of the inner tube where evaporation would be completed or the steam super- - heated somewhat. - It is also possible to assume & set of steam conditions at the outlet of the inner tube and. make stepwise calculations frcm the top downward to the feedwater inlet end accepting the superheatér outlet temperature that results. The difference in the character of the relations between the various load conditions of Figs. 5 and 7 lead to some conver- gence problems in either case. These, in turn, make it'necessary to modify the calculational procedure somewhat depending on the steam conditions. Both methods have been used in this study, and both have been found to be not only awkward but demanding in-that they.require good engineering Judg- “ment to choose values that will'yieid convergence. However, no better ap- proach has been found in spite of samé months of effort by both MIT gradu- ate students who became interested in the problem and by the éufhors of Ref. 7. Typical Calculations The steps followed in estimating the boiler tube. length, temperature '_ distribution, and pressure drop on the water side_are summarized in Table'B, and a set of typical calculations is shown in Table 4. The first set of calculations was made for the 100% load condition. The first step in the calculations was to staft at the bottom, or salt inlet end, and assume & - decrement in the enthalpy of the salt. This was chosen to be 10% or the ' expectéd value for the boiler. For a given tube geometry the sfirface areas of the inner and outer tube walls per unit of length are defined and ~hence the area per increment of length is readily calculated. For good W 27 . " Table 3. Calculational Procedure for Establlshlng the Tube Length for a Given Set of De31gn Conditions (see nomenclature in latter portion of table) 12. Plot curves for the salt and steam temperatures as fUnctions of AQ/Q from the inlet to the outlet (e. g., Fig. 7). Specify the tube dlameters and Wall thicknesses. Compute the overall heat transfer coefflclents for the inner and outer tube walls (e.g., use Fig. 11). For convergence in the iteratlon, the heat transferred from the salt to the annulus steam near the superheater outlet must be greater than the heat transferred from the annulus steam to the water in the cen- ter tube per unit of length, i.e., ViAjAt)/I>UsA At,/L. This will probably require an inner tube liner to provide a heat dam near the bottom. To mlnlmlze the overall tube length, this liner should be terminated as soon as this can be done and stlll maintain Uy Ay 486, / L>U2A LIAD) / L. Estimate the mean temperature of the salt, annulus steam, and boiler water for the first increment of tube length using the ratio of the heat added to the steam in the outer annulus to the total heat removed - from. the salt and the curves of item 1 above (e g., Fig. 7) Using the above as the starting point, follow the calculational pro- cedure of Table 4 using constant increments in AQl, the enthalpy change in the molten salt Compute AL, that is - _ AQl UIA]_At]_ ; i Compute AQs for the value of Al found in item 7 (AQa = UgAgALfltg/L) Repeat steps 6, 7, and 8 for a better approx1mat10n if the changes in mean temperatures ylelded by these steps dlffer by more than 20% from the values estlmated in step 5. Repeat the above for new 1ncrements of length using the same 1ncre- ment in NJ; until EAQl Ql..' If there is trouble with convergence, change the value of Us by chang-~ ing the length or effectiveness of the heat dam near the feedwater in- let, It may also be advisable to change the proportions of the tubes. For part load calcalations,rbe careful to assume & small temperature - difference between the salt and the superheated steam for the first increment in tube length if the temperature distribution along the full length is desiréd, If the superheater exit temperature exceeds by more than 100°F the- value assumed for the part load condition, new curves similar to those of Fig. 7 should be prepared, If instead one wants a rough 1ndicatlon of the active tube length at part load, com- pute a case as if it were a design point and compare the resulting length with that for the 100% load condition. 28 Table 3. Nomenclature Outer tube effective heat transfer surface area, ft2 Inner tube effective heat transfer surface area, ft2 'ID of outer tube, in. ID of inner tube, in. Friction factor for salt Friction factor for steam in superheater annulus Friction factor for steam in inner tube Mass flow rate of salt, 1b/ft2- sec _ Mass Tflow rate of steam in superheater annulus, 1b/ft2-sec Mass flow rate of steam in inner tube, 1b/ft?-sec Tube length (or distance from bottom of tube), £t Increment in tube length, ft Number of increment Steam pressure, psia Steam pressure drop in superheater annulus in increment, psi Steam pressure drop in inner tube in increment, psi Heat removed from molten salt, Btu/hr- tube Net heat added to steam in superheater annulus, Btu/hr tube - Net heat added to steam in inner tube, Btu/hr- tube Heat removed from salt in increment, Btu/hr Heat added to steam in superheater annulus in increment, Btu/hr Heat added to steam in inner tube in increment, Btu/hr . Dynamic head in superheater annulus, psi Dynamic head in central boiler tube, psi Reynolds number for salt - Reynolds number for superheated steam in annulus " Reynolds number for steam in inner tube Local temperature of molten salt, °F ‘Local temperature of superheated steam in annulus, °F Local temperature of steam in inner tube, °F Local temperature difference between salt and steam in super- ‘heater annulus, °F Toocal temperature difference between steam 1n annulus and steam in inner tube, °F . Specific volune of steam in superheater annulus, t3/lb Specific volume of water-steam mixture in inner tube, £t3/1b O ; !l‘afile‘ 4.'._:Typ1cal Calenlations] Worksheet for a Single Boiler-Superheater Tube Steam pressure, psia = 4000 ' Fraction of reference design load, % = 100 Outer tube OD, in. = .65. Total tube length, ft = 32.97 Steam temperature leaving inner tube, °F = 745 Superheater temperature out, °F = 1045 Outer tube ID, in, = .50 Pressure drop through inner tube psi = 38,3 8alt temperature in, °F = 1200 : Feedvater temperature in, °F = 560 Inner tube 0D, in. = .25 = Pressure drop through ennulus, pei = 14.4 Salt temperature out, °F = 950 Water enthalpy rise, Btu/lb = 886 Inner tube ID, in. = .23 Total preesure drop through boiler-superheater, psi = 52.7 Heat load per tube, Btu/hr =Qy = 650, Water flow per tube, lb/sec = 0,204 . (@) Fractional change in salt enthalpy in increment my/a .10 .10 .10 .10 .10 .10 .10 .10 .10 .10 - (@) Effective surface area of outer tube, £t3/ft nL JA52 0 .152 .15 L1520 L1522 .15 152 152 .152 .152 (3 Effective surtace area of inner tube, 2/t AL _ .052 .052 .0s2 .0s2 .052 .0s2 .052 ,052 .052 .0B2 ' O Oversl hea.t transfer coefficledt for outer tube, T 667.6 667.6 667.6 667.6 667.6 667.6 667.6 667.6 667.6 667.6 : @ Werall heat transrer coei'ficient for 1nner tu?be, o N s 300 - 300 300 1145.5 1145."5- 1145.5 1145.5 1145.5 - 1145.5 1145.5 -'r - ‘ - - o R SRR RE o o SR Heat transfer thraugh outer tu'be per | un:lt of length per 13U1A1/L 0. 1.4 10L.4. 1014 10L4 100L.4 - 101.4 1014 1014 10L.4 10L4 | °F, Btufir. R _ _ o _ o » L : - i ' (D Heat transfer throuh imer tube per unit of length per Uh/L S 157 157 157 59.6 - 59.6 59.6 59.6 5.6 59.6 59.6 °F, Btu/br.ft.°F . : . : . S _ . _ : _ _ : o . (8) Molten salt _temperatm-e, °F 7_(mean for 1ncrement) - ty Mg 7 1190 . 1164 1139 2 1087 1062 1038 1012 987 962 O (® stean temperature in outer tube, °F (mean for increment) t; see (D) 1030 1016 958 907 872 80 87 M2 T 752 @ Steen; tanpei-atm in 1nher tube, *P (mean for increment) ¢, see @ .585 575 605 630 660 692 708 720 728 72§ @) Temperature drop from salt to steam in snmulus, *F AT -0 160 148 181 205 215 212 n 220 217 210 @) Temperature drop from stesn in sunulus to steam in imer AT, (3) - @ 445 - 4AL 353 2 212 158 119 72 42 23 : . tube, °F ‘ T S - @ 1nerement in lemgtn, ft o & &/ (OO 4,006 4331 3.542 3.127 2,982 3.024 3.038 2.914 2,95 . 3,052 @ Distance from vottom, £t \ ‘ L.y 4,006 2337 11.879 15,006 17.988 21.012 24.050 26,964 29,918 32,970 (@ Heat ndded to steam in inner tube, Btuwhr & OO 27,991 29,988 19,628 51,623 37,672 28,474 21,547 12,503 7,395 4,184 @9 Heat added to steam 1n outer tube, Btufnr oM M - 37,009 35,002 45,372 13,377 27,328 36,526 43,453 52,497 57,605 60,816 @ mtio of incremental entbalpy change fn superheater to Ma/%. @ /& .0%9 .0539 0698 .0206 .0420 0562 .0668 .0808 .0886 .0936 total for stesm s Ratio of enthalpy change in superheater to total for . ), —1-A22/& .05%9 .1108 1806 = .2012 .2432° ,2994 .3662 .4470 .5356 6292 steam ) ©® matio of incremental enthalpy change 1n boller to total M © /a 0431 .0461 .0302 .0794 .0580 .0438 .0331 .0192 .0ll4 .0064 for asteam Ratio of enthalPy change in boiler to total for steam ), A/Q 0431 . .0892 .1194 .1988 .2568 .3006 .3337 .3529 .3643 .3707 0P00000000 90000 ® ® Mean va),lue of 2A/Q in snnulus for interval (read ta in Mg, 7 ‘ Mean V31ue of 24/Q in boller for interval (read t3 in Pig. 7 Ratio of enthalpy change in salt to total Fluld specific volume ;.n-uperhea.ter, £t? /10 Fluid specific volume in boiler, £t*/1b Mass flow rate in superheater, 1b/sec. £t? Mase flow rate in boiler, 1b/sec. £t? Dynamiec head in superheater, psi Dynemic head in boiler, psi N (a =.25) (a = .20 ana ,23) R@lds number in superheater Reynolds mumber in boiler Prictlon factor in superheater Friction factor in boiler Incremental pressure drop in superheater, psi Inc;etuenta.l pressure arop in boiler, psi Pressure drop from tube inlet, psi (cuter annulus) "Pressure drop from tube inlet, psi (inner tube) Table 4. (Continued) 1-@a- @ 2 Q.4 2 n=n - ey SR/ v @t; in Fig. 13 vy €ty in Pig. 13 64 . 204/, 00002 64 . 204/, 000236 or ,204/.00029 2 @) @26 s @ @) 2ezme Re, Ref, 8, p. 291 Res Ref. 8, p. 291 £2 Ref. 8, p. 29 2y Ref. 8, p. 294 = @D O 2 0@ O P2 Pap + 22::1; AP nooITEe 972 .022 .10 .185 . 0202 200 864 .798 1.611 179,867 238, 503 .0168 2.685 5.131 52.709 5.131 .916 . 20 .180 . 0200 200 864 772 1.660 180,871 243,578 . 0168 . 0158 2,809 5.680 50.024 10.811 +854 ’lm .30 170 .0208 200 864 .690 1,684 185,802 250,232 .0167 L0157 2,041 4.682 47,215 15,493 809 .78 729 159 .28 .279 .40 .50 .60 247 35 a2 .0215 .0230 .0260 200 200 200 703 703 703 587 522 A4 1162 1.226 1.439 191,010 192,715 192,715 243,482 253,812 294,922 L0166 .0166 .0166 L0157 L0157 .0152 1.524 1.292 1.190 2,852 2.870 3.307 45.174 43.650 42.358 18.345 21.215 24.522 .667 .%593 217,343 0 .80 116,100 0290 0318 200 200 703 703 .48 328 1.450 1,503 191,858 170,400 299,136 314,880 L0166 .0172 L0151 0150 1.054 .B22 3.326 3,285 41168 40.124 27.848 31.133 « 509 .35 » 207 1.57 148,855 317,265 .0178 0150 « 544 3,494 39,292 34,627 .368 1.00 . 0350 200 703 .160 1,599 115,628 317,265 .0189 . 0150 .461 3.660 38,748 38.287 ot 31 compatibility with both the molten salt and the water, the tubes were con- sidered to be of Hastelloy N. This is a high nickel alloy that has a rela- tively poor thermal-conductivity, an important factor in a supercritical ~ boiler because of the fairly thlck wall reqUired in the outer tube — in - this 1nstance 0.075 in., It was found advantageous to increase the thermal ‘resistance of the inner tube wall.at the lower end by making use of a rdouble-walled'tube. Indthis_regionuthe 1/4-in. OD central,tube with a wall " thickness of 0.010 in._was_assumed to be lined with a{smaller'diameter tube ,“ ofrthe same.wall thickneSs-with a radial clearance of 0. 002 in. between ',them The gap would be vented and, in view of the local metal tempera- ‘tures, would be filled with steam rather than water, The reason for using the tube liner can be seen by env1s1oning the heat transfer 51tuation at the bottom end of the reentry tube. Unless the ~heat transfer rate from the salt to the superheated steam per unit of length exceeds that from_thersuperheated_steam to the feedwater, the.tem— perature of the superheatedrsteam willractually drop as it flows toward the outlet. Inasmuch as the local temperatures of the salt' the super- I heated steam,'and the feedwater are fixed, and so, too, are the surface :_areas and the surface heat transfer coefficients, a simple solution is to interject some extra. thermal re51stance between the feedwater and the superheated steam. This region need not be very long because the steam cexit temperature is relatively insensitive to the steam temperature well above the exit. FUrther,_to keep the overall tube length down it is de- “51rable to maintain a large temperature difference between the salt and '.the superheated steam Thus one . of the places where Judgment is required when making the stepwise calculations is in ch0031ng the point at which to 'terminate the heat dam formed by the 11ner of the 1nner tube It should be noted that introducing the above liner in the inner p:tube w1ll act to 1mprove the b0111ng flow stability characteristics of rrthe reentry tube b01ler. th only will it increase the pressure drop in iithe inlet region but 1t will also increase the fluid velocity there, both "thelpful faotors., e | 32 Heat Transfer Coefficients 'Heat transfer cOefficients for the molten salt were computed'using the Dittus Boelter relation and physical propertles supplied by J. W Cooke. From these the chart shown in Fig. 10 was prepared In computing the over- all heat transfer coeff1c1ents for the 1nner and outer tube walls, the molten salt velocity was taken as constant for all of the conditions con- 51dered and hence the heat transfer coefficient for the fluid film.on the salt sxde was constant The temperature drop through the tube walls would - of course, be directly proportional to the heat load, and hence these two -represent constant conductances. The heat transfer coefficient for the superheated steam flow in the annulus varied directly with the 0.8 power of the ‘steam flow rate and hence this was the dominant factor at low flow rates. The heat transfer coefflcient for water under nucleate or annular film'boiling conditions is almost independent of heat 1oad and is very J .{high hence it had a relatively small effect on variations in the overa]l_ | heat transfer coefficient with load. ~ The calculated values for the overall heat transfer coefficients through the inner and outer walls are presented in Fig. 11 as a function of the steam flow rate. The principal resistance 'to heat transfer at the lower steam flow rates is that in the surface films in the superheated g»steam annulus, whereas at high loads the principal res1stance 1s repre- 'sented by the temperature drops through the tube walls ' An 1nherent error in the calculation procedure stems from the use of a constant heat transfer coefficient on the water side throughout the length of the boiler. Fbr the flow rates employed here, annular film boiling would prevail through the greater part of the boiler until the vapor quality reached about 90% after which the heat transfer coefficient would drop rapidly. However, this effect is small at or near supercritical ' pressures, and the heat transfer coefficient in the first portion of the superheater is about as high as in the "boiling" zonel® (see Fig. 12). This in effect compensates for the reduced hesat transfer coefficient in the last portion of the boiler. These variations have relatively small effects in the region of interest here and were neglected for the purposes of these calculations. O Heat-Transfer Coefficient, h, Btu/hr:ft3.°F . Temperature, °F ORNL DWG. 71-5732 equivaient diameter, 0.1 0.2 0.5 1.0 2.0 2 34 5 010 15 20 30 4 50 . 100 150 200 300 400 500 - " Flow Rate, G', lb/sec.ft? - - ' Fig. iO; Heat-Transfer Coefficients for NaEF,; at 1350°F Under Turbulent Flow Conditions. (Values for physical properties used in June 1968: cp = 0,37 Btu/1b. °F, p = 15 lb/hr--ft, k = 0,232 Btu/hr. £t. °F, density = 131 1b/£t2.) 1000 ee 34 OENL ING. T1-5733 _Overall Heat Transfer Coeffictents (Btu/hv.ft2.%F) 103 16" 10° | 108 Stean Heat Joad (Btu/hr.tube) : Fig. 11. Effects of Heat Load on the Overall Heat Transfer Coefficients for the Inner and Quter Tube Walle for an Quter Tube of 0.66 in. OD and 0.50 in. ID, an Inner Tube of 0.25 in, OD and 0.20 in. ID, & Molten Balt Mass Flow Rate of 2500 Ib/sec-ft2, and & Tube Wall Thermal Conductivity of 147 Btu/hr. £t. °F. The molten salt heat transfer coefficients were cbtained from Fig. 10 nd those for stesm from Fig. 12 or ng- H5.7’ po 320 Of Ref. 8- . ) | ) O ORNL DWG. Ti-5T34% 12— — T | ~ Pressure 4500 psi 'o—25x106|b/hr-ft2 R | o 0 =21 % 106 Ib/hreft2 | a\ b o --16x 10‘5 Ib/hr-ft? | \( By Bu/heeft2eF x 1070 > Bz o % —_° AL O 4}£%? ;aft;”‘.’Ta 4.-£f""’A _ \\- | NQ‘ \ ™ ° 'Qfif— -l - o D o gd-HrT e TN | et L L T Do o, D 3 4 5 6 7 8 9 | | L Surface temperature °Fx 102 - Fig. 12. Heat-Transfer Coefficients for Several Mass-Flow Rates 10 ) of Water in a Uniformly Heated Once-Through Boliler Tube Operating at 4500 psia (Dickinson and Welch, Ref. 16) 11 s€ - 36 Temperature Differences and Heat Fluxes The appropriate cur#es of ‘Figs. 5, 6, and 7 were used as the basis for estimating the_effective local temperature differences in any given increment of tube length. From these témperature differences, togethér with the heat transfer-coefficients and. the afiount of heat added in the increment, it is pbssible to calculate the lehgth of the first increment of outer tube length and the amount of heat transmitted écross the inner tube wall. These exchanges of heat in turn make it possible to calculate the tempersture in each fluid stream at the beginning of the next incre- ment of’the length. Note that the average local temperature differences for each increment as used in the calculations were'estimatedéand gen- erally differed & little from the values indicated by the compléted cal- culations for the increment. The'difference wés usua;ly sufficiently small so that iteration was not necessary. To facilitate the use 6f Fing 5, 6, or 7, the summations of the amounts of heat added or subtracted to each fluid stream up to and including each increment were also tébulated togetheriwith the fractions of the total heat added to the water-side stream. Pressure Drop The pressure drop across each increment of tube length was also cal- culated in Table 4. To facilitate these calculationé, Fig. 13 was prepared to show the density of the fluid on the steam side as & function of its temperature. For the subcritical pressure boiling conditions the density in any given increment of boiler tube length was estimated by considering the fractional change in density in the boiling zone as eQuai to the frac- " tion of heat ‘added to the steam fip to ‘that point in the boiling region. | No attempé_was made to allow for the effects of the two-phase flow friction féctors, but it is believed that the overall effects of these would amount to only abodt a factor of two in the boiler region, and, of course, there -would be no effect on the pressure drop in the superheater region. 37 U _ . - \ ' o ORNL DWG. T1-5735 " 10. 4,00 2,00 1.00 SPECIFIC vm (£t3/1p) 0.20 - 0,10 200 koo 600 - ‘ | - Fig.13. Specific Volume as & Function of Temperature for & Typlcal u - ~ Set of Boiler Discharge Pressures and Temperatures, 38 Estimated Performance Characteristics - The prime objective of this studyiwas_to investigate the effects of both the design heat load and the design steam tempersture"and pressure | cenditions on the boiler tube iength. Once thege effects are established for a single tube it is easy to estimate the effects of the choice of the full-power steam pressure and temperature on the size and eost of a full- scale steam generator. Further, once these effects are established for a single tube, the effects of the mode of control on the part load tempera- ture and pressure distribution can be inferred readily, and inferences can be drawn with respect to the startup and part load control characteristics and possible boiling flow stability problems. Effects of Design Heat Load on Tube Iength A study of the fine structure of the heat transfer relations indicates that; if a prime criterion is considered to be completion of boiling prior to exit from the inner tube, the.boiler tube length is'determined bj the maximun load conditions anticipated. That is, the tube length required for boiling to dryness will be reduced as the design heat load is reduced. This effect is shown in Fig. 14 for the two control modes shown in Figs. 5 and 7, that is, a constant steam generator discharge pressure ir- respective of heat load and a steam generater discharge pressure directly proportional to the load. In all cases the calculations were earried out to determine the tube length required to meet that particular design con- dition. Effects of Design Heat Load on Temperature Distribution A more detailed insight into the effects of design heat load on & reentry tube boiler is given by Fig. 15 which shows the calculated tempera- ture distribution in the boiler as a function of distance from the bottom, "and indicates the length of tube required to evaporate the water to dry- . ness in the inner tube for a wide range of steam conditions. In this in- stance the boiler exit pressure was proportionai to the load. Note how little tube length will be required for the 1% load condition. Note, too, O i* @) * 39 ORNL DWG. T1-5T36 700 500 8 TUBE LENGTH (ft) g | PRESSURE DROP (psi) 200 . . . _ 0 . 0 ' 40 -8 - 120 160 200 | DESIGN LOAD IN % os' REFERENCE DESIGN CONDITION ' Fig. 14, Effects of Ioad -on the Tube Length Required to Handle Loads ‘Iess the Design Heat Ioad of 650,000 Btu/hr per Tube, The assoclated steam- - side pressure drop is also plotted The steam generator discharge pressure - for the lower design loads was taken as & constant 4000 psia for one.set of . ;curves , &nd as directly proportional to the load for the other set. 100 TEMPERATURE (OF) ORNL DWG. T1-5737 0 | 10 DISTANCE ABOVE BASE OF BOILER (ft) Fig. 15. Effects of 100, 75%, 50%, 25%, 104 end 1% Heat Loed on the Axial Temperature Distribution for Operation Witk the Steam Pressure Directly Proportional to the Heat Load, = : : : i w 4 that the curves for the tube length andlpressure,drop in'Fig; 14 rise steeply between the lOO and 200% load conditions, thus implicitly Jjusti- fying the choice of flow rate for the nominal 100% load condition. Ac- ‘tually, the choice'of steam»flou rate for the 100% load condition in this case was arrived at by a series of preliminary calculations with just this p01nt in mind. ' - The effects on the temperature distribution of the design heat load per tube for operation-W1th the system pressure held constant irrespective of the load are shown in Fig. 16 Comparison of Figs 15 and 16 indicates that the design steam pressure has only a mild effect on the required tube - length at a given heat load, in spite of the fact that the reduced pres- sure greatly depresses the temperature in the boiling region N - Effects of Part Load'Operation o If a boiler tube length for; say, the-lOO% reference”design load.con- . dition of Fig. 15 were chosen as the basis fOr a design, it is evident that at part load the steam would be superheated to a temperature higher than - that 1ndicated for the particular cases shown in Fig. 15 because these were ecalculated on the basis that the inner tube would be termlnated at the point at which b01ling was completed The superheating effects would be . particularly large at low steam flow rates where the steam temperature in the annulus would run close to the local molten salt temperature through the upper portion of the tube until it dropped when chilled by heat extrac- ~tion for boiling the water rising in the inner tube. | . This effect is shown _fiin Fig. 17 The calculations were carrled out by recognizing that vir- | “tually all of the heat transferred from the salt to the superheated steam ‘in any given increment would flow directly into the boiling ‘water in the inner tube. Thus,,aalaffirst:approXimation,‘it vas assumed that UpA;ity = . UsApAt, which then defined tp-and t;. A check showed that this does in fact give a good approximation andrno iteration was needed in moSt'cases. | The reason for this is that the enthalpy change in the outer annulus at “the lO% load condition is only about 5% of the enthalpy rise in the inner tube. ' Tt should be noted that to get good convergence the detailed calcula- tional technique of Table A was used for the cases of Fig. 17. This mmam °r) 42 * ORNL DWG. T1-5T38 0 ' ) 10 20 30 DISTANCE ABOVE BASE OF BOILER (rt.) Fig. 16. Effects of Design Heat Load on the Axial Temperature Distribution for 0peration ¥With & Constant Steam Discharge Pressure of LOOO psia. 43 ORNL DHG. T1-5T39 » 1100 1000 800 700 RMFERATURE °F). . 500 300 b _ _ o . o 10 20 30 DISTANCE ABOVEBASEOF BOILER (£8) o _ Fig. 17. Comparison of the Axial Temperature Distributions for Operation ‘ J - &t 100% and 10% of full power with the Steam Pressure Directly Proportional to B ' the Heat Load and the Same Tube Tength in Both Cases. 44 represented an improvement over that used for Figs._lA through 16 and is the reason for a small difference in tube length that may be noted between Figs. 15 and 17. The heat transfer calculations for Figs. 14 through 16 could be repeated to give better épproximations. This was not done for this preliminary study because it was felt that the effect would be small, - and, in any 8 8 £88§°88% . _ .ao._w Im.o JSVIUONIITEV3YO34a , 1Sd =~JONVHO AUNSSIYd FILLIOUHL EXAMPLE: 80 MINUTES ARE REQUIRED TO CHANGE FROM 5% LOAD 800 PSIG-800°F TO 100% LOAD 2400 PSIG-I000°F P-68969 - Recommended Time for Changing Load and/or Steam Conditions, (Courtesy Westinghouse Electric Rorp.) Fig 22. ) » 53 pressure by 1600 psi but the temperature by only about 30°F. According to Fig 22 the turbine could tolerate making this change in only 11 min, Similarly, if the turbine were operating at full load, the load could be cut in half in as little as 20 min., i.e,, substantially faster than the - common rule of 2%7min Further; 1if the steam generator can be designed S0 that the superheater outlet temperature increases SOmewhat with a reduc-‘ tion in load, even faster rates of change can be tolerated. For example, if the full load steam pressure and temperature were . 4000 psi and 1000°F and the 50% load conditions were 2000 p81 and 1050°F, Fig. 22 indicates that the load change could be made in about 1l min. As will be shown later, it happens that the reentry tube steam generator can be proportioned to o give this characteristic and hence would make-. possible unusually rapid changes in load. Possibility of Eliminatingrthe.Throttle Valve . Steéeam throttle valves are a major source of operating trouble in steam power plants. The valves are relatively large, must operate at _ about 1000° F, they are subject to heavy loads by the high steam pressure "drop across them, and lubrication condltions are highly unfavorable Thus it is not surprising to find that throttle valve sticking is one of the | principal causes of - forced outages in steam power plants.' Throttle valves are used to make it possible to change the steam flow - rate much more rapidly than can be- accompl:l.shed by controlling the boiler : and furnace in conventional coal fired plants. The most critical consid- eration is to avoid a serious“overspeed of the turbine in the event of an " abrupt and complete loss'of~the electrical load, The inventory of super- . heated water in the . reentry tube b01ler 1s vastly less -than 'in'a conven- tional boiler; it is estimated to be only about 3 1n.3/tube at full load, and about 1in.? at 10% load on the steam generator NOte that as a con- sequence of mechanical, electrical, and fluid friction losses, ete. in the ]turbine—generator unlt, 1n the event of a complete loss of. electrical load the steam flow required to keep the turbine up to speed would be about 10% of the full power output of the steam generatOr. Thus an abrupt loss in electrical load would entail a reduction in the superheated water inventory 54 in a reentry tube boiler by about 2 in.3/tube, or 4 in.2/Mw(t). This full load steam flow rate is about 3 1b/sec for 4 Mw(t), whereas 4 in.3 of superheated water is only about 0.1 1b. Thus the full load steam flow - rate would consume the surplus water inventory in only about 1/30 sec if the feedwater supply were abruptly cut to the 10% load level. The inertis °of the massive rotor in the turbine-generator unit should be. sufficient “to - keep the‘bvershoot in rotor speed to a low value, Inasmuch as.the most diffiéult‘condition to meet is that for an abrupt ldss‘in electrical load, it appears that control of the power plant could be accomplished by con- trolling the feedwater flow rate with one or moré relatively'small valves opefating at about 650°F rather than the large, hot steam throttle valves normally employed. This should give a greatly increased reliability. It is of interest to note that this approach to the control of a Ran- kine cycle plant has been analyzed and investigated experimentally and is a very similar but much smaller system designed for a nuclear electric space power plan.t.18 The analyses and tests gave highly encouraging re- suits'fdr a system in whiéh.the thefmai ifiértig-OT the boilef wés small as _in_fhe'system proposed here. Further; the variable pressure apprbach is coming into use in conventional plants.l® Proposed Design for a Molten Salt Reactor Plant A conceptual design for a molten‘Salé7reactor'With its intermediate heat exchangers and fuel pumps integrated into & common pressure vessel is presented in-a'companion-report.20 'The net electrical output in that. study is 1000 Mw(e) and the fuel and inert salt temperatures are those - given in Fig. 4a. Plant layout studies favored-the use of six steam gen- ~erators each of which‘would be directly coupled to one of six fuel-to-NaFF, heat exchangers. The steam generators requiréd for this plant provide a 'good illustrative example for use here to show how the sihgle tube analyti- cal work presented earlier in the report can be applied to the design of =& full-scale steam generator as well as’the proportions=to be expected in a " finished unit. ‘ B O B 3 0 55 Reheaters Most modern steam-plants emploj reheaters, in part because there is a | direct increase in overall cycle efficiency of about 5%, and in part be- cause they give a further indirect increase in cycle efficlency of a few percent by eliminating the moisture churning losses 1n the lower stages of - the turbine, In addition, reheaters can be designed to eliminate the pos- sibility of turbine bucket erosion in_the.lower turbine_stages by reducing the moisture content in thet‘region to almost nothing. In conventional coal fired steam plants the length of piping required to connect the boiler to the reheater and the relatively large amount. of tube surface area re- quired because the reheater must be located in a relatively low-temperature - zone to avoid burnout difficulties have combined to make the cost of in- cluding provisions for reheat rather high for conventional coal-fired plants. However, it is believed that in a mOlten salt reactor plant the turbine can be located much closer to the steam generator?and,'because there is no danger of tube burnout from excessive locsl temperatures, & much higher average heat flux through the reheater tube walls can be main- tained so that the tube surface area requirements are quite modest., Fur- ‘ther, the design of the tube matrix for the reheater is quite straight- forward because there are no flow stability or two-phase flow problems in- volved. Thus a reheat cycle appeered_highly desirable for this design study, and the two reheat stages specified for_Eddystone'Unit No. 2 have been included in the proposed steam generator for. a mOlten:saltgreactor. . General Description The dominant consideration'in choosing the steam'generator'0verall _geometry was to make use of the statlc head in the ligquid column in the o b01ling region to help provide for flow stability in the boiler. This _'_meant that the boiler tube should be vertical with the’ feedwater entering irat the bottom While it would ‘have been desirable from the standpoint of thermal convection 1n the molten salt system to admit the molten salt at the top and have it leave at the bottom, a brief glance at ‘the temperature distribution diagrams shown in Figs. 4, 5, 6, and 7 is sufficient to show 56 that the salt sfiould move counterflow relative to the superheated steam in the outer annulus. Figure 23'presents a somewhat similar disgram including the two reheater stages and shows that'they,'tbo, should be inICOunterflow. It was thought at first that it would be desirable to make use of separate casings for the boiler and the ' two reheaters, but the additional plping, manlfolding, and associated temperature distributlon prdblems led . to the conclusion that it would be best to mount the reheater tubes in an " annulus surrounding the b01ler tubes in the conflguratlon shown schema- tically in Fig. 24 . 'fieadering Problems As mentioned earlier the headers envisioned for the boiler region would entail the thermal sleeve arrangement shbwn in Fig. 3. This avoids -the use of a high'pressureeheader sheet'ifi'eontact with the molten salt; rather, the tubes would be manifolded below the pressure vessel containing the molten salt. The same arrangement could be employed for the reheater tubes. In either case any tube could be blocked off readlly -outside the steamvgeneratorAcasing in the event of a leak or other mslfunction. Inss- much as the reheater tubes are free of the complexities associated with the ~ inner tube used for the boiler, it would also be possible to assemble them in bundles of perhaps 12 tubes with the tube bundle headers inside the casing of the steam generator to minimize the number of penetrations fhrough the shell, This looks preferable if a unit has more than 100 tubes. Provision of adequate space for header sheets is always difficult when designing heat exchangers with closely spaced tubes. Preliminary lay- out studies indicate that this is difficult but might be done with the con- figuratioh’of Fig. 25 without severely distorting the‘Outer_casing or hav- | ing a large volume of inert salt that serves no useful function but in- creases both the capital investment and the,time required for the system to respond ta control actions. The outlet'spigets from bundles of 16 tubeS'cduld be passed through a header sheet in the bottom head of the vessel with a thermal sleeve to avoid large local stresses. This would glve an adequate ligament thickness (about 0.5 in.) in the header sheet 'after allow1ng space for the thermal sleeves, This will take up most of <:;9,i ”» » " » -Tepii:eratum (°F) 57 ~ ORNL DWG. T1-5T45 '1206 1100 1000 900 = T00 600 - | f AQ/Q ® : Fig. 23, Temperature I)istribution in the Bo:ller a.nd Reheater as a Function of the Fraction of the Heat Removed from the Molten Salt Under Full Load Conditions. 58 SALT POMP ORNL DWG. T1-5Thé REHEAT STEAM OUTIET MANTFOLDS ' 4 in. dis. 5 TEEL 45 £t tall SALT IN Fig. 24. Proposed Configuration for One of & Set of Six Steam Generators for a 1000 Mu(e) Molten Salt Reactor Power Plant. Detailed data are given in Table b, O o P | ] n 59 ORNL-LR-DWG 73420 ~1),-in. TUBE - (vEg) t-in. TUBE - N7 4) %-in. TUBE Wz 4) Y5-in. TUBE_ CAN . - _ " . i . " W W W W WD LG Fig. 25. Schematic Diagram Shoving the Bifurcated Tube Header Arrangement Used as a Basis:;fbr"_ the Propbsed Tube Bu;ridle and Header ~ Sheet Configuration. . The nominal tube id (d,) is indicated for each step in the manifold to show that the flow velocity is kept constant. .60 the area of the elliposidal head in the region where its radius of curva- ture is large. The space just outside this region but inside the shell will serve as the inlet plenum for the molten salt which would enter tan- gentially.through two pipes. The 96 spigots from the tube bundles could be manifolded outside the header sheet using a bifurcated tube arrange- ‘ment similar to that of Fig. 25. Three steanm pipes-could.be coupled to each set of 32 tube bundles. The feedwafier could be suppliéd through two penetratibns in the largest bifurcated colipling in each of these three sets. _ _ : The layout envisions installation of small header drums for the re- heater tube bundles in fhe annulus just above the salt inlet plenum. ‘The annular baffle between the reheater region and the boilef-superheater re- gion fiould extend downward no farthef than the header drums for the re- heater tfibe bundles.- | _ . - The layout problems would be eased by reducing'the numbe? of boiler tubes in each steam generator unit and employing a large number of units. The ratio of tube matrix cross-sectional area to header sheet area could be increased in this way and manifolding problems of the tubes for the feedwater and superheated steam could be eased. Differential Thermal Expansion There would be no problem associated with differential thermal expan- sion between the boiler tubes and the steam generator casing beéause the . top ends of the tubés wouid be free to float axialiy. However, the re- heater tubes will tend to run at & scmewhat different temperature than the casing and hence some means for providingfieiibility shofild be employed. Probably the easiest way to accomplish this will be to install the tubes on & spiral having a stéep pitch, i.e., perhaps a total twist of 60 deg in the length of the regenerator region. This could be chosen to pyovide sufficient flexibility in the tubes so that the differential thermal ei-- pénsion between the tubes and the shell would be accommodated elastically. The pitch would be steep enough so that the lateral loads on the tubes would probably not be large enough to make it neceésary to employ addi- tional structure to resist the forces associated with the small amount of erossflow that will occur. L0 kA e LA 3l R o L Lo AR W » ol '_Geometric and Performance Data The calculations and prlncipal geometric and performance data for the complete steam generator are summarized in Table 5. Recent revisions in the physical properties of.NaBF4, particularly in the viscositj, made it desirable to change the salt massrflow rate through the steam generator from the value of 2500 1b/sec:ft? used'in_the-Calculations of'Table 4 and those for Figs. 11 and 14 through 20. This, coupled with the changes in physical properties, led to a change in the tubemlengthirequired. The heat load for the boiler-superheater and each of the reheaters was used to calculate the number of tubes for each matrix, with the molten salt streams for each sectlon treated as flowing independently of each other with no lateral heat transfer, Note in therflrst portion of Table 5 that- steam bleed-off in the turbine through seals and for feed heating re- duces the steam flow rate to the reheaters 50 that it is substantially less than the steam flow rate to the boiler. 'CoSt'EStimate A rough estimate of the cost of the steam generator of Teble 5 can be obtained by summing the estlmated costs for the three ma jor items, i.e., the shell, the tubing, and fabrication of the tube-to-header joints. This has been done and the results are summarized in Table 6. Note that the overall cost amounts to about $7.05/xw(e), which compares with around $20/kw(e) for the comparable'portion of & conventionalrcoal'fired boiler. (These cost estimates were made on the basis of unit costs as of 1968. ) - Thus it appears that the combined effects of a uniformly high heat flux, :fsmall diameter tubes with relatively thin walls, and ellmination of the need . for pressure vesselswsubject to high pressures.much.more than offsets " the high cost per pound of Hastelloy N. - The prOportions glven in Table 5 were not 1terated to mske both the 'etube length and the molten salt pressure drop for the boiler-superheater and the two reheaters the same., This could and should be done, but the ' combined effects)of uncertainties'in'the calculated heat transfer coeffi- cients and friction factors are 1arge enough to make this an unwarranted refinement at this stage. 62 Table 5. Summary of Design Calculations for Steam Generator Units for a 1000 Mw(e) Molten Salt Reactor Plant | u Steam System Superheater outlet temperature, °F Superheater outlet pressure, psia 'Reheater No. 1 outlet temperature, °F " Reheater No. 1 inlet temperature, °F Reheater No. 1 outlet pressure, psia Reheater No. 2 outlet temperature, °F Reheater No. 2 inlet temperature, °F Reheater No. outlet pressure, psia Generator output, Mw Auxiliary power requirements, Mw Boiler feed pumps and boiler auxiliaries, Mw - Fuel pumps, Mw. ' NaBF, pumps, Mw Miscellaneous, Mw : Net electrical output, Mw Overall thermal efficiency, % Fraction of heat load to boiler-superheater Fraction of heat load to reheater No. 1 Fraction to heat load to reheater No. 2 Fraction of weight flow to boiler-superheater Fraction of weight flow to reheater No. 1 Fraction of weight flow to reheater No. 2 Heat added to steam flow in boiler superheater, Btu/Ib Heat added to steam flow in reheater No. 1, Btu/1b Heat added to steam flow in reheater No. 2, Btu/lb Inert Salt (NaBF;) . Entrance temperature, °F Exit temperature, °F Entrance pressure, psi Exit pressure, psi Flow rate, lb/hr Flow rate, lb/sec Flow rate, ft2/sec NaBF,; physical properties at lO75°F Specific heat, Btu/lb : Thermal conductivity, Btu/hr.ft.°F Viscosity, 1b/hr.ft Density, 1b/ft3 Melting point, °F MONOHH Shell Number of units - Shell diameter, in. Shell overall height, ft Shell wall thickness, in. 1050 4000 1050 786 1043 1050 705 251 1043 66 50 5.4 10 1 977 i 4 0.775 0.118 0.107 1.000 0. 84 0.68 886.0 161.0 - 180.2 950 1200 180 51 13.92 x 106 3860 33.6 0.36 0.27 3.4 115 725 45 40 0. 50 e e B o 63 Table 5 (Continued) Boiler Boiler heat load, Btu/hr Tube OD, in, Tube ID in. , Tube length, ft ' Tube pitch (equilateral triangular), in Salt mass flow rate, lb/sec 52 Dynamic head, psi S - Reynolds number - Friction factor - Pressure drop, psi - - Shell-side flow passage area/total Cross- sectional area ~ Shell-side flow passage equivalent diameter, in. - Number of tubes per square inch, in.” Number of tubes per unit - Shell-side surface area, ft? Rehegter No. 1 Tube OD, in. Tube ID, in. | Tube pitch (equilateral triangular), in. ' Log mean temperature difference, °F Heat load per steam génerator, Btu/hr Steam mass flow rate, 1b/sec:ft? Overall heat transfer coefficient, Btu/hr-ft2 °F Surface area per steam generator, ft? SR Surface area, ft°/ft of tube - Number of tubes - Tube length, ft o Salt flow rate, ft3/sec 'Salt mass flow rate, lb/sec £t2 Salt flow passage area, £t2 Total cross-sectional area in tubes, ft° Total cross-sectional ares in tube matrix, ft2 Shell-side flow passage/total matrix cross- sectisnaln . area , _Shell-side flow passage eqUivalent diameter, 1n -Steam pressure drop, ps1l _ : ‘Salt pressure drop, psi Reheater No. 2" - Tube OD in. Tube ID, in., ’ Tube pitch (equilateral triangular), in. Log mean temperature difference, °F Heat load per steam generator, Btu/hr 9.73 x 108 0,65 - - 0.50 . 35,0 0. 775 . .. 1350 1.5 ' 60,000 - .0.025 45 , 0,366 0.465 191 - 1493, 8900 1.00 0.80. - 1,035 150 1,478 X 108 200 - 380 | - 2390 0.236 365 33.6 33.6 11350 o127 S2.328 L W145 .18 20 84 1.00 0.90 1.01 - - 200 1.34 x 108 64 Table 5. (Continued) Reheater No. 2 (continued) Steam mass flow rate, lb/hr.ft? _ Overall heat transfer coefficient, Btu/hr ft2 °F o215 70 Surface area per steam generator,: 2 3300 - Number of tubes 632 Tube length ft 22 Salt flow rate, ft3/sec 3.6 Salt mass flow rate, 1b/sec-ft? 1000 Salt flow passage area, ft2 0.414 Total cross-sectional area in tubes, ft2 - . 3.63 Total cross-sectional area in tube matrix, f£t2 - 4,04 Shell-side flow passage/total matrix cross-sectional = 0.10 area Shell-side flow passage equ1valent diameter, in - 0,11 'Steam pressure drop, psi - : - g Table 6. Rough Cost Estimate for the Steam Generator of Table 5 . Ttem flAmount ) :M'i Unit Cost = Cost Shell (fabricated) 10,000 Ib $8/1b° ¢ 80,000 Baffle - 2,000 1b .. $/m ' 16,000 Boiler tubes 60,000 ft, 0.25 in. diam $3/ft = 180,000 - o 60,000 ft, 0.65 in. diem $7/ft =~ 420,000 Reheater tubes 25,000 ft, 1.0 in. diam - $8/ft 200,000 Material Cost | . . $ 896,000 Tube installation 2490 tubes © - $110/tube 273,900 (including welding - o ‘ of tube-to-header Joints and inspec- tion Total Shop Cost © $1,169,900 O ad e 65 Conclusions The proposed“reentry tube boiler appears to offer a long?sought solu- tion that satisfies all of the magor requirements for the steam generators of 11quid metal and molten salt reactor power plants It appears to be well suited to the generation of steam at any desired pressure and tempera~ ture condition w1th good stability and control characteristics throughout | the range from zero to full power, .It_lends,itself_to_designs in which the thermal stresses throughout the unit can be‘kept within the elastic range under all operating conditions. The heat flux is uniformly high so that the inventory of'structural metal or liquid metal and molten salt is - near minimal, the number of tube-to-header joints is relativelycsmall there is no high pressure header sheet in contact with the high-temperature liquid, there appears to be no unusual fabrication problems, and the capi- tal cost appears to be competitive. The only apparent disadvantage is that the concept is novel and has”not'been'tested. The performance characteristics of reentry tube boilers presented in this report represent rough preliminary approximations based on a set of simplifying assumptions. Test experience~with'a unit'employing at least three full-scale tubes would provide a firm foundation for the design of a large unit. ‘ -~ Recommendations A test unit consisting of one or & few tubes should be built and tested In the test program particular attention shOuld'be given to the -1nvest1gation of possible boiling flow instabilities under startup, near- zero power, and part load operating conditions to see to what extent ori- -ficing may be required at the feedwater inlet. The test program should also 1nclude an investigation of the overall stability and control charac- teristics under steam conditions ranging’from 100 psia to 4000 psia. Most of the problems and uncertainties associated with the reentry .tube boiler could be investigated with a relatively S1mple single tube unit fitted with clamshell heaters, Such a unit could be built and tested ex- peditiously and inexpensively, and would lend itself nicely to a detailed 66 investigation of the temperature distribution along the tube under a wide range-of conditions as well as facilitate modifications such gs the inser- " tion of inlet orifices. - | _ , | An electrically heated single tube test will probgbly be very much 1ess convincing to most people than a long endurancetestiof'aunit'havihg 3 to 19 tubes heated by a molten salt. Such a test unit should.give a good demonstration of the freedom of the system from éorfosidn,_mass trans- fer, thermal stress, and boiling flow stability probléms, f O » 10. 11, 13. 67 -':References A. P. Fraas, Flash Boilers for Fluoride Fuel Reactors, USAEC Report ORNL~-CF-55- 6~79, Oak Rldge Natlonal Laboratory, June 16 1955. P. R. Kasten et al., De81gn Studies of lOOO-MW(e) Mblten—Salt ‘Breeder Reactors, USAEC Report ORNL-3996, Oak Rldge National Laboratory, ,Ausust 1966 TP 77-87. D. Scott and A. G. Gr1nde11 CQmponents and - Systems Development for Molten-Salt Breeder Reactors, USAEC Report ORNL~TM;1855, Oak Ridge - Natlonal Laboratory, June 30, 1967, pp. 21-22. Oak Ridge Natlonal Laboratory, MSRP Quarterly’Progress Report _ Jan. 31, 1959, ORNL-2684, p. 5h. , L. G. Alexander et al., Molten Salt Converter Reactor Design Study and Power Cost Estimates for a 1000 MW(e) Station, USAEC Report GRNLTM—lOGO September 1965 - B. W. Klnyon and G D. Whitman, Steam Generator-Superheater for ‘Molten Salt Power Reactor, ASME Publication 614WA-228 Nov. 26- - Dec. 1, 1961 | , R.. S. Holcomb and M. E. Lackey, Performance Characterlstics of a Short Reentry Tube Steam Generator at Low Steam Output, USAEC Report ORNL-TM-3236, Oak Ridge National Laboratoiy (in press). A. P. Fraas and M. N. Ozisik, Heat Exchanger Design, Wiley & Sons, Inc., New York, 1965. | ~ B. 8. Shlralkar and P Grlffith The Deterioration in Heat Transfer to Fluids at Supercritical Pressure and High Heat Fluxes, Rept. No. 70332-51, Dept. of Mechanical Engineering, Engineering’ Progect Lab, Massachusetts Instltute of Technology (Mar. 1, 1968) D. W. Lee and R. C. Spencer, Photomicrographic Studles of FUel Sprays NACA Technlcal Report h5h 1933. D. W. Lee, The Effect of Nozzle Design and Qperatlng Conditions on the Atomization and Dlstrlbutlon of Fuel Sprays NACA Technical Report h25, 1932. ' P. J. Birbara, J. E. 'éo#les'and'v C. A. Vaughen, Flash Boiler Test II, Memo No., EPS-X-296, Massachusetts Institute of Technology, 'Englneerlng Eractlce School, Union Carbide Corp .5 Jan. 18, 1957. Oak Ridge Natlonal Laboratory, MSRP Semiann. Progr. Rept. Jan. 31, 1964, USAEC Report 0RNL-3626, P- gh 14. 15. 16. 17. 18. 19. - 20. 68 A. P. Fraas and M. N. Ozisik, Steam Generators for High-Temperature ' Gas-Cooled Reactors, USAEC Report ORNL-3208, Osk Ridge National Laboratory, April 8, 1963. | W. R. Chambers, A. P. Fraas and W. N. Ozisik, A Potassiun-Stean Binary Vapor Cycle for Nuclear Power FPlants, USAEC Report 0RNL—3584 Oak Ridge Natlonal Laboratory, May 1964 N. L. Dickinson and C. P. Welch, Heat Transfer to Supercritical Water, Trans. ASME, Vol 80, p- 746 1958 Westinghouse Turbine Operating Manual I. L. 1250- 32184, 300-350 Mm(e), 300 psi Hydraulic Governoring System, 1967. M. M. Yarosh and P. Gnadt, Use of a Cavitating Pump for Control of a Potassium Rankine Cycle System, paper presented at the Intersociety Energy Conversion Engineering Conference,_Las Vegas, September 1970. R. J. Bender, U,S. Welcomes Variable-Pressure Concept, p. 56, Power, August 1970. A. P. Fraas, Conceptual Design of & Molten Salt Reactor with Its Intermediate Heat Exchangers and Fuel Pumps Integrated in a Common’ Pressure Vessel, USAEC Report ORNL-TM-2954, Oak Ridge National Ieboratory (to be published). O r b & C 89-90. 38.' ,'I'O. 4. ho, . 86. 87. 88. al. 92. 93-94, 96-98. 99-100. Je S. M. c. E. E. Co H. R. C. d W. J. F. J - S. We -J. D. A. L. Ww. A, W. P, H. P. Jd. M. 69 INTERNAL DISTRIBUTION L. Anderson F. E. Bauman Beall, Jr. Bender E. S. G. Je I. B. W. W. B. L. L. R' Jde. P. R.. E. P. D. R. G. 0. N. E. W. R. ._J. I. Bettis Bettis Bohlmann Borkowskil Bowers Briggs Collins Cooke Cottrell Crowley Culler DiStefano Ditto Fatherly Engel Ferguson Fraas Fuller Grimes Grindell Harms Haubenrecih - Helms Hoffman Kasten Keyes, Jr. Lundin - 43, b, L6. b7, 48, kg, 50. 5. 52. 53. L .55 56-65. 66. 67. 68 69. 70, 1. 2. 3. 7)4'0 T5. 76. TTe 8. | T9e 80-81. 82. 83-8L. 85. 'EXTERNAL_DISTRIEUTION David Elias, AIE-Washlngton - R. Jones, AEC-Washington Kermit Laughon, AEC-OSR T. W. McIntosh, AEC-Washington M. Shaw, AEC—Washington | W. L. Smalley, AEC-ORO ~ ORNL-TM-2953 .R. N. Lyon --H., G. MacPherson R. E. MacPherson H. E. McCoy "H. C. McCurdy H. A. Mclain L. E. McNeese J. R. McWherter A. J. Miller R. L. Moore E. L. Nicholson A. M. Perry R. C. Robertson M. W. Rosenthal J. P. Sanders A. W. Savolainen Dunlap Scott ‘Mo Jo Skimler I. Spiewak D, A. Sundberg R. E. Thoma D. B. Trauger A. M. Weinberg J. R. Weir M. E. Whatley G. D. Whitman L. V. Wilson Central Research Library Document Reference Section Laboratory Records Laboratory Records (LRD-BC) Division of Technical Information Extension (DTIE) Iaboratory and University Division, ORO Director, Dlvision -of Reactor Licensing, AEC, Wash Director, Dlvision of Reactor Standards, AEC, Wash,