I OFL. ORNL P 1698 FEEFEFEE 01:25 L MICROCOPY RESOLUTION TEST CHART NATIONAL BUREAU OF STANDAROS - 1963 . . . #H M 2 +. - - .... .......come 13 anos ? T . 2 .. Onnu p-1698 Conf-651009-4 MASTERS NOV 1 5 105 RESPONSE OF A HELIUM-COOLHD FAST REACTOR TO CHANGES IN COOLANT FLOW AND REACTIVITY* R. S. Carlsmith J. G. Delene Oak Ridge National Laboratory Oak Ridge, Tennessee RELEASED FOR ANNOUNCEMENT IN NUCLE:IR SCIENCE ABSTRACTS 1T1.5 .nu Abstract The ORIJL design of a helium-cooled fast reactor was studied with respect to its response to various types of accident conditions. These a- Veitad ut of Government och orod work. Neither t "yoe of sucha contractor preparu, i Heins, 'nor the Commiston, vor my person acties atbild the Coinmisation: A. Mokes my mirrunty or representation, pressed or implied, with respect to the accre , a vatabunud of the luformation contand la taula roport, or that the w paratus, method, or process disilowed in this report buy no latringe my Habilities with respect to the use of, or for dem soo renulting from the of any information, apparatus, method, or process declound in this report. As wrod ta the above, "parin acting au beball of the Commission" inoladas may ployee or contractor of the coaptadon, or employee of mucha contractor, to the extent that much employee or contractor of the Commission, or em dianntmates, or provides accome to, any information pursuant to his deploymeat or contract with the Commission, or his employment with such contractor. s petrately owned rights or i rey, complete port of any taformation, B. A w accident conditions included step and ramp reactivity insertions and loss of coolant pressure and flow. The objective was to determine whát kind of accidents this reactor could withstand before fuel melting began, without regard to questions of accident credibility. A Doppler coefficient, T dk/ar of 0.0053 and total helium loss coef- ficient of +0.0051 8k were calculated for this reactor. These coefficients proved adequate to prevent melting for a step reactivity addition of less than 29 cents, if no scram occurs, and a step change of $1.35 if a power- level scram occurs. For reasonable rates of coolant loss (half-life in the vicinity of 20 sec), the reactor would not approach prompt criticality at any time during the transient. The ability of the reference reactor to remove delayed fission product heat in the event of two loss-of- ir RA . : coolant accidents was determined. Research sponsored by the U. S. Atomic Energy Commission under contract with the Union Carbide Corporation. (For presentation at the Conference on Safety, Fuels and Core Design in Large Fast Power Reactors, to be held at Argonne National Laboratory, Argonne, Illinois, October 11-14, 1965). i Mw e m wer+ .4 .14 - 1: Tr * ** *!kama rait 12 2 Lt 7 F.." . F .-2- RESPONSE OF A HELIUM-COOLED FAST REACTOR TO CHANGES IN COOLANT FLOW AND REACTIVITY Interest in zas-cooled fast reactors as a possible alternative route to successful plutonium breeding is due in part to differences in physics characteristics between sodium-cooled and gas-cooled reactors. One would expect, for example, that the reactivity effect of the coolant would be much smaller with gas so that the core geometry could be optimized with- out regard to void or coolant-loss coefficients of reactivity. One woula also expect a harder neutron spectrum in the gas cooled case, resulting on the one hand in a higher breeding ration, but on the other hand in a lower Doppler coefficient. In this paper we point out first that the reactivity effect of a gas coolant may not in fact be negligible, and should be considered in the reactor design. Secondly we inquire what degree of protection may be provided by the smaller Doppler coefficient that does exist. Finally, we examine the power and temperature excursions to be expected in the event of loss of coolant flow and pressure. We take as our reference design a 1000 MWe helium-cooled fast reactor design prepared by ORNL in a recent study. Some of the principal features of the reactor are given in Table 1. This design was obtained by assuming a maximum linear heat rating of 40,000 Btu/hr-ft (12 kw/ft), a maximum cladding surface temperature of 1500°F, and a maximum pumping power of about 10% of the net electrical output. The coolant pressure was limited to 1000 psi on the basis that higher pressures would be likely to intro- duce unknown problems in component' technology. Dimensions of the fuel pin, the core and the blanket were obtained by optimizing with regard : to fuel cycle costs. -3 , Table 1. ORNL Helium-Cooled Fast Reactor Design Fuel material (Pu, 2384). Blanket material 238J02 Stainless steel 2500 993 1000 Cladding material Total thermal power, Mw Gross electrical power, Mw Coolant pressure, psi Effective core length, in. Effective core diameter, in. Radial blanket thickness, in. Axial blanket thickness, in. Fuel pin diameter, in. Cladding thickness, in. Pin spacing triangular pitch, in. 0.250 0.015 0.432 Helical fins Method of pin spacing 10 145 Average core fuel enrichment, % Average core power density, w/cm”. Average core specific power, kw/kg fissile Internal breeding ratio 0.92 1.39 Total breeding ratio Doubling time, yrs 12 We used one-dimensional multigroup-diffusion calculations for the static analysis of the core physics. We found it important to do both axial and radial calculations and to iterate the transverse buckling until consistent results were obtained. We checked one such result against a two-dimensional multigroup calculation and found excellent agreement in keep and power distribution. The 16-group cross sections were obtained by averaging, for most nuclides, over the spectrum for a composition similar to that of the reference core. The isotopic con- · tent of the fuel was determined by doing equilibrium recycle burnup calculations and choosing the average core composition during a 100,000 Mwa/T cycle as representative. The isotopic composition of the plutonium came out to be 61:31:5:3, (239Pu:240Pu: 242 pu:242 Pu). The isothermal Doppler coefficient was estimated by the use of Doppler broadened cross sections for several different temperatures in diffusion- theory calculations. The Doppler broadened cross sections were obtained with the General Electric Rapture? code and also with the General Atomic GAM-II code which uses somewhat differe:it approximations. The two codes gave essentially the same result as shown in Figure 1. We obtained the effective neutron lifetime and delayed neutron fractions by first-order perturbation theory calculations. Direct, calculations of reactivity with and without helium gave the value for the reactivity associated with helium loss. The values of these quantities for the reference reactor are shown VI itd in Table 2. . The helium-density reactivity coefficient deserves some further comment. It has two components. The transport component, which 18 always negative, depends principally on the core size and shape and on :: T the core coolant fraction and density. The downscatter component is . . . . . ... -5- Figures) Reference Reaczog Do ;Le Lex Co 1 1 : .. . 1 , - - GAMI: .... .. .. .. ..-.•**-- - EFFRES EUGENE DIETZGEN CO. MADE IN U. B. A. . .. . live- .... m y -. - -.-.-.-.-.- . --. . -. -.-.-. ? ** bine . TULLIITTI TTTTTM Lii ! mi! ? 11 11 - - generat t o maine de minimigo I IT+11 **17T d . . .- 5 .!.i...ili * mundi TIL: watoto 111 thartontti Tobot . DIETZGEN GRAP PAPER LOGARITHMIC I CYCLE XI CYCLE .. Tiilit ܙܐܐܐ - named IIIII 1IITITUTI 340 . - . - . . - - . . తం .. 500 7000 200 - 10pc 1500 2000 జుం poo Temperature (R) ::. . ca Wsca a...cine - , - - -- - -- - -- -- - - - . VRH . . 4 ? . 1 1. Table 2. Kinetics Parameters for Helium-cooled. Reactor Dopler coefficient at 965°C, 0k/°C Reactivity gain for complete coolant loss, ok Effective neutron lifetime, sec. -0.43 x 10-5 0.0051 0.8 x 10-6 Delayed neutron yield fractions: Group 1, sec-2 Yield 0.0833 x 10-3 0 0.0129 0.0311 0.1335 0.3316 m 0.7093 0.61425 1.2517 0.5843 0.1909 f n 1.262 6 3.208 Total 3.462 x 10-3 likely to be positive unless a considerable emount of moderating material is present in the core. In our reference reactor the transport component was 0.06% ok while the downscatter component was +0.57% øk, giving a net reactivity gain of 0.51% ok for complete loss of helium. Since the trans- port component is so much smaller in magnitude than the downscatter com- ponent, it appears that changing the shape of the cori would have little effect on the net coefficient. This is in contrast to the situation of liquid metal-cooled reactors where the core shape in many of the designs - has been selected so that a large transport component of the coolant coefficient would be obtained. The downscatter component, which is the important part of the net coefficient, can vary considerably from one helium-cooled core desigr. to another. It can obviously be decreased by having a smaller core coolant fraction. It also depends very sensi- tively on the shape of the neutron energy spectrum and on the energy dependence of the absorption and fission cross sections of both the fuel and structural materials in the core. In calculations of somewhat dif- ferent helium-cooled cores we have found heliu-loss coefficients as much as a factor of three lower than in the reference core. Thus the magnitude of the net coefficient is at least partially under the control of the reactor designer. The dependence of the helium coefficient on the exact high-energy spectrum and cross sections means that there is probably considerable uncertainty in our calculations of this quantity. We have also considered co, as a fast reactor coolant. In our design we obtained the same fuel heat rating as for the helium-cooled reactor with a slightly lower coolant volume Traction (0.59 vs 0.67 for helium) and slightly higher pressure (1300 psi vs 1000 psi for helium). The downscatter and absorption component of the coolant-loss coefficient was then +0.81% ok while the transport component was -0.52% 6k. With different core design a co-cooled fast reactor in this size range could wa have a negative coolant coefficient. We have studied the response of the reference reactor to a variety of reactivity changes and to coolant flow and pressure changes. The computer program that we used. solves the point kinetics equations and the hot-channel heat transfer equations in alternating successive finite time steps in order to determine the axial and radial temperature distribu- tion in the hot channel fuel pin and the axial temperature distribution in . -8. the coolant gas. We take into account the reactivity feedback from the * fuel Doppler coefficient and the reactivity effect of any changes in coolant density. For convenience it was assumed that ek/ar varies as Tot whereas our calculations indicate an actual variation as rated An importance-weighted effective Doppler coefficient to be used with the hot-channel temperature was obtained by modifying the actual Doppler coefficient to take into account the ratio of hot-channel temperature to core-average temperature and the initial ratio of hot-channel temperature rise to core-average temperature rise. The Iission product heating was computed as a function of time by integrating the data of Shure.+ We did not include the negative reactivity feedback from fuel or structural expansion. We defined as the "failure" point for the reactor the occurrence of either a vo, temperature above 6500°F or a stainless steel cladding temperature above 2250°F. Our basis for these choices was that the UO, vapor pressure at 6500°F is high enough so that rupture of the cladding is likely,' and that this condition or the failure of the cladding at about 2250°F would cause damage to the cure. MOTOR The effect oỉ the size of the Doppler coefficient in limiting the TE excursion resulting from a step reactivity addition without control ITS Mi action is shown in Figure 2. In this figure the maximum allowable re- C activity insertion is the one which produces temperatures just short of 2 our failure point. The maximum step insertion for our reference reactor AN (without a scram) is $0.29. If the coolant loss coefficient were zero the allowable step insertion would be increased to $0.46. Hence, in this tur type of transient the positive reactivity effect associated with the de- . crease in helium density in the core is about 17 cents. . . ------*.. 57942..S ALO212276 Laser . . ac... 22 . - . . + . . . . . (Dollini) EUGENE DIETZGEXC MADE IN U. !i. A. - i012 ..! !! ! 1 . ' . 1 , .. .. , ...... .. 1 SUI 772 177. - Heliosis Coef 1 " * -- Nid. 3-10-14 DIETZGEN GRAPH PAPER . 2 . 83.13:177111 - - -+ --- 1 . . . ... 1 1 1 1 1 1 . : Max!17!! 17? . 1 11 :- . . 1L . 1 widodo . - til 1 16 11 . 1 1 treba dodat sve do dom .. - - werden od tak r kontener نان Talat . . -10- . DLT Figures 3, 4, and 5 are plots of the heat deposition rate and peak temperatures in the fuel and cladding as a function of time after a step reactivity insertion. Figure 3 shows the response of the reference reactor to a step insertion of slightly less than that required for fuel pin failure. Fere the peak power was approximately 1.9 times steady-state power. We repeateå this set or ca. culations but with a much higher Doppler coefficient in order to show the effect of the magnitude of the Doppler coefficient. Figure 4 shows the response with a Doppler coefficient of about four times the calculated Doppler coefficient, to a step reactivity insertion of $1.85. The prompt period caused the power deposition rate to rise over three orders of magnitude. The fuel temperature rose very: fast during the power peak. This temperature rise caused the excess reactivity to i'all due to the strong Doppler coefficient. The prompt period was terminated when the excess reactivity fell below one doilar. The power level then returned to within a factor of two or three of steady state power. The rates of change of power and temperature were more gradual after the prompt power peak was terminated. If the coolant coefficient were equal to zero the response shown in Figure 4 is changed to that shown on Figure 5. Figures 4 and 5 are virtually identical during the prompt excursion. They deviate only after sufficient time elapsed for the heat to be transferred from the fuel to the coolant gas. In this kind of transient the coolant density coefficient is a delayed effect where the Doppler coefficient is a prompt effect. The difference in the final peak temperatures was approximately 180°F. We conclude that the coolant-density coefficient becomes less important as tri Doppler coefficient is increased. RESPONSE OF REFERENCE REACTØR TO $.25 STEP ADDITION Tige ON RATE (MW) Ö 6000 5000 TEMPERATURE (DEG F.) 4000 3000 2000 1000 1073 103.70 Spoo, 101 102 * Response of Reference Reactor to a $ 1.85 Step Reactivity Addition mig. be . . ........... AllW är Ön X I dk/ar = 0.0235 ! NAI! ön ö ULIVOI ö ö ö Ö SO00 5000 TULlop 4000 Thill LIVIUI 3000 2000 0001 10-3 10-2 TIMI. c ons 10 + h 1:13. 5 Response of Reference Reactor to a ok :) Step reactivity Addition * Tdk/AT = 0.0235 Coolant loss dk = 0.0 6000 & LAPTOP TEN RATE ! I A 4 min "Como Le todo ut hr. - - - - TEMPERATURE (DEG F) . 3000 - . 0007 L 1000 nos capace poput centrs 10 102 : RE *** Viva. . . . . . . In.... ..... ... --- ------- - . -14- 1 PM Y. . . stii Visi . The maximum allowable reactivity insertion is of course significantly larger if control rod action is taken. We assumed control rod insertion starting 50 milliseconds after the step reactivity insertion, producing a linear decrease in reactivity at the rate of $173 per second for 250 milli- seconds. The assumed 50 millisecona delay is intended to include the times required for the scram signal to be given, for the rod actuation, and for an initial relatively slow rate of reactivity insertion. In this case the maximum allowable step reactivity insertion was $1.35, with excessive cladding temper- atures occurring rather than excessive fuel temperatures. Figure 6 shows the power and temperature response as a function of time for a $1.35 step re- activity insertion with control rod action. This plot illustrates the importance of the various time constants. The prorapt burst was completed eted ..: : before control action was started. The only mechanism controlling this burst was the fuel Doppler coefficient which was assumed to act instantane- ously. Within a few milliseconds after control action began, the fuel temperature virtually stopped rising. Because of the relatively long time constants for heat transfer, the peak fuel temperature did not begin to fall until nearly a second after the control rods were inserted. As the heat floweà ouvrard from the fuel the cladding temperature started to rise, reaching its peak several hundred milliseconds after the control action began. With control roh action, the maximum allowable reactivity insertion by a ramp was somewhat larger than by a step. The reference reactor could tolerate a maximum ramp reactivity insertion of $33.00 per second until control rod action was initiated at 50 milliseconds. The total reac- tivity insertion by the ramp was thus $1.64. Figure 7 shows the power WS and temperature response as a function of time to a $35 per second ramp reactivity insertion with control rod action. The phenomenon of multiple fast bursts is illustrated here. The firsu power burst occurs when the mirova *. . .. . . . . . . FERR $1.35 STEP, CONTROL ROOS AT 50 M. SEC (AWI HEATE DEPOSITION RATE (MM) 1,01 immed 0009 operative hai GOD . 5000 TEMPERATURE IDEG Fj 4000 3000 2000 1000 10-4 10-3 10-2 TIMO ., 101 102 103 . is 103 102 101 ndesc . FERR $35/SEC RAMP, CONTROL RØDS AT 0.05 SEC 10-1 10-2 . . . , 1 - 10-3 II . 1 1 + 0009 THW 000S 000 . 000€ list 930) JUNI1y3DW31 0002 0001 IVY NO1115 AT S e e more S on instante los De Come move in armenn wir Af. . .at het tieto other member thaiming women ini...ng ... ... ... . . -17- reactivity inserted by the ramp is enough to cause the reactor to become prompt critical. The resulting swiIt temperature rise causes the Doppler coefficient to return the reactor to below prompt critical. The power level then drops and the temperature rises at a slower rate. The rate of reactivity removal by the Doppler coefficient becomes less than the addi- tion rate from the ramp. This causes the reactor to become prompt critical WISA again, initiating a second burst. After control action is taken, the temperature cehaves in the same manner as for step reactivity insertions. an In Figure 7 the limiting factor was a cladaing temperature above 2250°F. Figure 8 also illustrates the phenomenon of multiple bursts. Here we are looking at the response of the reactor with a Doppler coefficient of about four times actual size to a $100 per second ramp. The power was in the middle of its fifth burst when the fuel temperature exceeded 6500°7. The positive coolani-density coefficient implies that a sudden reduc- tion in coolant inventory would lead to a reactivity increase as well as a reduction in heat removal capability. Of course, the negative Doppler coef- ficient reduces the net reactivity increase waich would otherwise occur. In Figure 9 the maximum reactivity occurring during a coolant-loss tran- sient is plotted as a function of the coolant loss rate (assuming no control rod action). The minimum. credible half-life for coolant loss will need further study, out it appears likely that it would be in the neighborhood of 20 seconds. If so, the peais reactivity during the tran- sient would be $0.26 instead of the $1.47 associated with instantaneous coolant loss. Figure 10 shows a moze signiiicant aspect of coolant loss, namely the time available for corrective action after the coolant loss is startea, again assuming that no control rod action is taken. If the . 1 . K . -. . is boy .. '. , - -.- . HAN . . . L - . . LLYE . TEMPERATURE IDEGF) 000 2009_3000 3000 4000 5000 1000_ ION RATE ( MI) HEAT DEP is in 6000 12 10- 4 10- TAN 11 152 mesecenos! Ramp Reactivity Addition Response of Reference Heactor to S/OOT . .. Continuous Tdk/ar = 0.0235 : osowano... $ ht . Star 2 27 % 11 . ... . 21 1. Eh ... twik2 " . . .. S , .. . . AK BYC OM Sekreacz10. conos$_of_coolan_ - - word - · 1.. • . - . - - ...- - . - . . - . .. . .- - . . ..... . . . .. 1 . .-..- .- . 15 - - - - - - . - . 1 . ...... 1 MY- .. - L ' .: 11 . .:. 1 . 1. . . . . !.." . . . . j - 1 .. et -- . . * ** , - 1 - . - .. .. . . - (suerroa) hanZovga 55anx7 238 - . . T . informiered on the action instruirano mesthim in a man womente ristruttore in . - . : . . . i . .- livraisi I...riisi iro --.. . 1 . . . MILLIMETER 3.13H117710 y -- T - - A 1 11 ... -- Til: * . I. . . i.!, -...:: : . . 17 + 1 . . ... hadi III . 1.11 VTT . 1'!'' 1 S- T . L det. 1.11111111 . - -.- " ? 1. WI y . - S i , .. - 4.1 II" . . . 1 ! .. Ital. . .: LILII . be .. . . L es .... . des T ." . . + . 1 1. 1. C . i*** 1 180 70 20 30 Half beim ei fo ja cochaine prescu i sert 16( seconds) **** : A 1.. -.. * W in. . . . manji Time llast Failure of Fuel Pin SSA DO 1 .. . . . 1 0 . . 1 . . . 11 1 . * . .. . 1 . . . - 1. - C f. . 5. . . U. . . . . .. HAUL 1! : * C. 41 * . -- - . ti i minda... . U t.. 1. • 4 1 1 1 . 1 ji! .:. ini di . 1 . Time Uroll Faller (seconds) c1' 117: TITIT. - . 2 . .... . ***.. . .....- - - TI . TITILI . I!"" MILLIMETER . 4 . 1 .. .1: 1 in the I- main - 1 T . 11 IT .. 1. 7 : 1 iai - 1 de... i .. 1 erit . ... . .... .... . mo 1 0 A . . . . ( 1 RE 1 1 1 !!! . . 1 - TO i. . - - ! 1. . . . ! T . . !LUI . . II . I k . . ... . . . . 11 ' I DO & Life for pressure and Flow Lessese concis) . . . ..... .. ... -21- coolant-loss half-life is no less than 20 seconds, then the time available before the fuel cladding starts to melt is about 10 seconds. The implica- tion is that a relatively low type of emergency cooling system would be effective. We have studied two situations in which coolant pressure and flow loss was followed by control rod action. In the first on these situations SU we assumed pressure equalization between the primary and secondary con- tainment vessels, which reduces the coolant pressure to 10% oi nomnal. It was assumed also that, in this acciäent, three of the four blowers woulů stop. He allowed the pressure equalization and blower coastdown to be completed in 20 seconds, and the control rod insertion to start 1.0 secona after the start of the transient. The power and temperature response to those conditions is shown in Figure 11. Immediately following the scram the heat production decreased faster than the heat removal, allowing the temperatures to fall. As the coolut flom ara pressure continued to de- crease, the temperatures went up again until the afterheat production fell below the final. heat removal capability of the coolant. Maximum allowable temperatures were not exceeded at any time during the transient. In Figure 12 we have the more severe case of the coolant pressure being reduced to atmospheric with stoppage of three blowers on a 20 second half-life. In this instance, also, the maximum allowable tempera- tures were not exceedeu. Further analysis of such situations is required with particular emphasis on cases involving stoppage of all four blowers. The analysis here would involve the investigation of one or more types of emergency cooling system. 102 101. - * FARR, . 1 NORMAL PRESSURE - 3 BLOWERS STOP IN 20 SEC T?! ordons to CAN 0005 ?! ! NOTIISIT - Tv 0009 000S 4711 0000 000€ 7!!NIV?!?!-071. 0002 1000 . 4 ► . . . - seseis 1 : 11 HALF LIFE FOR PRESSURE 4 FLOW LOSS 20 SEC, SCRAM . . is (90 - ... - - 0 1 20 Juni 2010 1000 10-3 10-2 emom 10 . TIME 103 -24- Zezerences i. 2. to do Gas-Cooled Reactor Project Staff, "Gas-cooled Fast Reactor Concepts,". USAEC Report ORVL-3642, Oak Ridge National Laboratory, September 1964. J. #. Ferziger et al., "Resonance Integral Calculations for Evaluation 01 Doppler Cosizicients - The Rapture Coäe," USAEC Report GEAP-3923, Vallecitos Atomic Laboratory, General Electric Company, July 1962. G. D. Joanou azā J. S. Dudek, "GAM-II: A B Code for the Calculation of fast-Neutron Spectra anà Associated Multigroup Constants," USAEC Report GA-4265, General Atonic, September 16, 1963. K. Enure, "Tission-Product Decay Energy," USAEC Report WAPD-BT-214, 11 3. 4. pp 1-17, Bettis Atomic Power Laboratory, Westinghouse Electric 5. Corporation, December 1961. M. J. Nelly, "Liquid Metal Fast Breeder Reactor Design Study," USAEC Report GEAP-4418, Vallecitos Atomic Laboratory, General Electric Company, Jánuary 1964. . . . . . . . 1 END , * DATE FILMED 12/8 765 - .