s -I‘ - OAK RIDGE NATIONAL LABORATORY operated by UNION CARBIDE CORPORATION for the U.S. ATOMIC ENERGY COMMISSION ORNL- TM- 935 - copY No. - £ ( DATE - September 11, 1964 MSRE NEUTRON SOURCE REQUIREMENTS J. R. Engel P. N. Haubenreich B. E. Prince NOTICE This document contains information of a preliminary nature and was prepared primarily for internal use ot the Ock Ridge National Laboratory. It is subject to revision or correction and therefore does not represent a final report. The information is not to be abstracted, reprinted or otherwise given public dis- semination without the approval of the ORNL patent branch, Legal and Infor- mation Control Department. LEGAL NOTICE This report was prepaered as an account of Govarnment sponsored work. Neither the United States, nor the Commission, nor any psrson acting on behalf of the Commission: A. Makes any warranty or represeatation, expressed or implied, with respect to the accuracy, completeness, or usefulness of the information contained in this report, or thot the use of any information, apparatus, method, or process disclosed in this report may not infringe privately owned rights; or B. Assumes any liabilities with respect to the use of, or for damages resulting from the use of any information, apparatus, methed, or process disclosed in this report. As used in the above, “person acting on bshalf of the Commission' includes any smployee or contractor of the Commission, or employee of such contractor, to tha extent that such employee or contractor of the Commission, or employee of such contracter prepares, disseminates, or provides access to, any information pursuant to his employment or contract with the Commission, or his employment with such contractor, 1ii CONTENTS Introduction . Internal Source Provision for External Source . . . . . . . . . . . . Neutron Detectors Flux in Subcritical Reactor Flux Due to Internal Source Flux Due to External Source Effect of keff on Flux Safety Requirements Initial Startup Experiments Normal Operational Requirements Partial and Complete Shutdown and Startup . Second Stage Startup Requirement First Stage Startup Requirement . . . . . . . . . Limiting Regulirement on Source Strength . . . . . Other Considerations Recommendations . . . . . + « .+ + & Appendix: Cglculation of Flux from an External Source Geometric Approximations Nuclear Approximations Page 10 12 16 19 ol 26 26 27 28 28 29 30 30 31 MSRE NEUTRON SOURCE REQUIREMENTS J. R. Engel, P. N. Haubenreich and B. E. Prince ABSTRACT The alpha-n source inherent in the fuel salt meets all the safety requirements for a neutron source in the MSRE. Subcritical flux distributions were calculated to determine the combination of external source strength and detector sensitivity required for monitoring the reactivity. If more sensitive detectors than the servo-driven fission chambers are installed in the instrument shaft to monitor the filling operation, the calculations indicate that the required source strength can be reduced from L x 107 n/sec to 7 x 10° n/sec. An antimony-beryllium source with an initial strength of 4 x 10% n/sec would still produce 7T x 10° n/sec one year after installation. Because there is considerable uncertainty in the calculated fluxes, the final specification of source and type should be made after preliminary flux measurements have been made in the reactor. CWPRODUCTION Some source of neutrons that is independent of the fissilon chailn reaction is essential to the safe and orderly operation of the MSEE. The primary requirement for such a source is to insure that when- ever the reactor is suberitical, the neutron population in the reactor is still high enough that in any conceivable reactivity excursion the inherent shutdown mechanisms and the action of the safety system become effective in time to prevent damaging power and temperature excursicns. Besides the safety requirements for a source, there is another, related to the convenient and orderly operation of the reactor. This is that the neutron flux at the detectors be high enough that the fission chain reaction in the core can be monitored at all times. The source strength reguired for this purpose depends on the experiment being conducted or the condition of the reactor and the location and sensitivity of the detectors. This report describes the conditions that will exist in the MSRE during the initial critical experiment and during subsequent startups, both before and after extended operation at power. The source reguire- ments for the various conditions are described and the extent to which these are satisfied by the inherent, internal sources is discussed. The requirements for an external source to supplement the inherent source are analyzed and recommendations are made for an external source and mode of startup operation that satisfy the requirements in a reasonable fashion. INTERNAL SCURCE The fuel salt itself provides a substantial source of neutrons in this reactor.t In the clean fuel the largest contribution to the internal source is from the alpha-n reactions of uranium alpha particles with the fluorine and beryllium in the salt. Neutrons from spontaneous fission of the uranium add to the internal source but this contribution is much smaller than the alpha-n contribution. The uranium in the fuel salt is a mixture of four isotopes; UEBu, UEES, U256, and U258; the proportions depend on the choice of fuel to be used in the reactor. Table 1 gives the compositions of three mixtures that have been considered, along with the isotopic composition of the uranium in each. All of the uranium isotopes undergo alpha decay and any of the uranium alphas can interact with the fluorine and beryllium in the salt to produce neutrons. The more energetic of the alpha particles can also produce neu- trons by interaction with 1ithium, but the yield is neglegible in comparison with that from fluorine and beryllium. Table 2 gives the neutron source in the core due to the various uranium isoctopes . N. Haubenreich, "Inherent Neutron Scurces in Clean MSRE Fuel Salt,” USAEC report ORNL-TM-611, Oak Ridge National Laboratory, August 27, 1963. Table 1 Compositicn of MSRE Fuel Salt Mixtures Fuel Type A B c Composition®(mole %) LiFo 70 67 65 BeFa 257 29 29.2 ZrFy 5 5.8 2 ThF4 1 O 0 UFy4 0.3 0.2 0.8 Uranium Isotopic Composition (Atom %) =34 1 1 0.3 U23Z 9% 93 35 =23 1 1 0.3 gese 5 5 6l b a ‘ oy ‘g Clean, critical condition. 99.9026% 1i7, 0.0074% Li7 Table 2 Inherent Neutron Source in Clean MSRE Fuela Fuel Type A B C o, n Source (n/sec) 34 .5 x 10° 3.1 x 10° 3.8 x 10° URs°s 9.2 x 10® 6.4 x 103 9.9 x 10% UESZ 3.2 x 10° 2.3 x 10° 2.8 x 10° U= 5 L 2.0 x 10° Spontaneous Fission Source (n/sec) 40 23 2.4 x 107 Total L.6 x 10° 3.2 x 10° 3.9 x 10° a . . "Effective" core, containing 25 £t of fuel salt of clean critical concentration. for the clean, critical locading with the three different fuels. About 97% of the alpha-n neutrons are produced by alpha particles Trom U234; thus, this source is proportional to the amount of R34 in the fuel. The most active of the available uranium isotopes from the standpoint of spontaneous fission is 7?38, As a result, Fuel C, which contains a much larger proportion of 1®3% than the other two mixtures, has a substantially larger source of neutrons from spontaneous fission. The inherent neutron source from spontaneous fission is listed in Table 2 for each of the three fuel salt mixtures. The MSRE will operate first with Fuel C, and the initial critical experiment will consist of adding fully enriched uranium to a salt already containing depleted uranium to bring the composition up to that shown in Table 1. At the beginning of the critical experiment the salt will contain 97% of the U2 but only about 0.5% of the UZ3* and US3® in the critical loading. The combined alpha-n and spontaneous fission source in the core at this point will be about 2 x 10° n/sec. After the MSRE has been operated at high power, the fuel will produce a significant number of photoneutrons from the interaction of fission-precduct decay gammas with beryllium. The threshold photon energy for this source is 1.67 Mev, so this type of source is insignificant before operation when only the uranium decay gammas are present. Since the concentrations of fission products and beryllium do not vary widely with the choice of fuel, the photoneutron source is approximately the same for all three fuels. Figures 1 and 2 show the rate of photoneutron production in the MSRE core after operation at 10 Mw for periods of 1 day, 1 week, and 1 month. The source is proportional to the power, and the source after periods of non-uniform power operation can be estimated by superposition of sources produced by equivalent blocks of steady-power operation. UNCLASSIFIED 10 ORNL DWG. 638171 SOURCE STRENGTH (neutrons/sec) 3x 10 L 8 12 16 20 2k 28 TIME AFTER SHUTDOWN (hr) Fig. 1. Photoneutron Source in MSRE Core Shortly After Various Periods at 10 Mw. UNCLASSIFIED ORNL_DWG. 628172 SOURCE STRENGTE (neutrons/sec) 10 1 10 100 TIME AFTER SHUTDOWN (days) Fig. 2. Photoneutron Source in MSRE Core After Various Periods at 10 Mw. The gamma-ray source used in the calculations is group IV of Blomeke and Todd,2 which includes all gamma rays above 1.70 Mev. The probability of one of these gamma rays producing a photoneutron was approximated by the ratio of the Be9 (7, n) cross section to the total cross section for gamma ray interaction in a homogeneous mixture with the composition of the core. A B69 (7, n) microscopic cross section of 0.5 millibarns was used,3 and the total cross section was evaluated at 2 Mev. These assump- tions lead to a conservatively low estimate of neutron source strength. PROVISION FOR EXTERNAL SOURCE For reasons which will be discussed later, it is desirable to supple- ment the inherent internal source with a removable source. Therefore a thimble is provided in the thermal shield, on the opposite side of the reactor from the nuclear instrument shaft. The thimble is a 1-1/2 inch, sch. L0 pipe of 304 stainless steel, extending vertically down to about 2 ft below the midplane of the core. It i1s mounted as close as possible to the inner surface of the thermal shield for maximum effectiveness. Location of the source thimble in the thermal shield provides water cooling and avoids the high temperatures associated with the reactor. NEUTRON DETECTORS All the permanently-installed core- neutron detecting instruments are located in the nuclear instrument shaft. This is a water-filled, 5 ft- diameter tube which slopes down to the inner surface of the thermal shield with separate inner tubes for the various chambers. Among the permanent instruments in this tube are two servo-positioned fission chambers which will be used to monitor routine startups as well as to record the entire power range of the reactor. These chambers are about 1 in. in diameter 2J. 0. Blomeke and M. F. Todd, "Uranium-235 Fission-Product Production as a Function of Thermal Neutron Flux, Irradiation Time, and Decay Time," USAEC Report ORNL-2127, Oak Ridge National Laboratory, August 1957. 5See curve in Reactor Handbook, 2nd Edition, Vol. III B "Shielding", E. P. Blizard, Ed., p. 23 (Interscience, New York, 1962). by 6 in. long and have a rather low counting efficiency of 0.026 counts per neutron/cme. (Other chambers in the tube which have no bearing on the source requirement are 2 compensated ion chambers and 3 uncompensated safety chambers.) Two vertical thimbles, similar to the source thimble but made of 2 in. sch. 10 pipe, are installed in the thermal shield to accomodate temporary detectors. The two detector thimbles are located 120° and 150° from the source thimble, one on either side of the permanent nuclear instrument shaft. The advantage of these vertical thimbles is that they place the entire length of a chamber close to the inner surface of the thermal shield, whereas a long chamber in the sloping instrument shaft would extend back into a lower -flux region and thus be exposed to & lower average flux. In addition to these provisions, there are spare tubes in the nuclear instrument shaft which could accomodate additional detectors. WEUX TN SURCRITICAL. RzZACTCOR In planning the use of source and detectors in reactor experiments and operation, an important quantity is the ratio of counting rate to source strength under various conditions. The counting rate is the pro- duct of the counting efficiency of the chamber and the average flux to which the chamber is exposed. The flux at the chamber depends on the source—1its strength, the energy of its neutrons, and, in the case of an external source, its location. The flux also depends on the amcunt of multiplication by fissions and the shape of the neutron flux distri- bution in the core, which is determined by the location of the source and the value of k in the core. The flux distributions in and around the reactor have been calculated for several different cases to provide a basis for planning for the source and detectors. Many approximations had to be made to render the computa- tions manageable and consequently the probable error in the results 1is quite large, perhaps as much as a factor of ten. Unless specifically stated otherwise, the fluxes and source strength requirements described in this report do not contain any allowance for probable error. Flux Due to Internal Source With an internsal, distributed source of Sin n/sec in the core, the steady-state production rate will be approximately Sin/(l -k ) n/sec. el The flux at any point is then . 5 in "in ¢ (1~ keff) The factor fin for a given location depends on the shape of the flux. For a flat source and low multiplication, fin at an external detector would be somewhat higher than at high multiplication, when neutrons are, on the average, produced nearer to the center of the core. When the multiplication is high, i.e., when (1 - keff) is quite small, most of the neutrons are produced by fissions in the core, with a spatial source distribution close to the fission distribution in a critical reactor. The relation between the core power, or fission rate, in the critical core and the flux in the thermal shield was calculated in the course of the thermal shield design, using DSN,* a multigroup, transport-theory code. For the case of a thick, water-filled thermal shield, when the core power is 10 Mw, the predicted thermal neutron flux reaches a peak, 1 inch inside the water, of 1.2 x 102 n/cm®-sec. The ratio of peak flux to power is thus 1.2 x 10° n/cme—sec per watt, or 1.5 X 1078 n/cm?-sec per n/sec pro- duced in the core. It was estimated that a chamber, © in. long, at maximum insertion in the instrument shaft would be exposed to an average flux of roughly 1 x 10~° n/cm?—sec per n/sec produced. A chamber 26 in. long in the instrument shaft would see an average flux only a third as high because the shaft slopes away from the core. The flux in one of the vertical thimbles near the inner wall of the thermal shield would be about 3 x 1077 n/cm?—sec per n/sec produced. Thus as keff approaches unity, fin approaches 1 x 1077, 3 x 1078 and 3 x 1077 cm™® for a 6-in. chamber in the shaft, a 26-1in chamber in the shaft and any chamber in a thimble, respectively. *B. Carlson, C. Lee, and J. Worlton, "The DSN and TDC Neutron Trans- port Codes,” USAEC Report LAMS-2346, Los Alamos Scientific Laboratory, February 1960. 10 Flux Due to External Source If a large fraction of the neutrons come from an external source, the flux shape will differ markedly from the critical shape. Flux distributions in the suberitical reactor with & strong external source were computed by a two-group neutron diffusion method. Equipoise Burnout,5 a two-group, two-dimensional diffusion-theory program was used. The reactor was rep- resented by a model in which the cross section of the reactor and thermal shield at the midplane of the core was approximated in x-y geometry and the axial leakage was represented by an equivalent buckling. In order to make the annular gap between the reactor and thermal shield manageable by the diffusion program, the materials in the gap (electric heaters, heater thimbles, insulation, and insulation cladding) were uniformly dispersed in it. The source was represented by a localized neutron- producing region just inside the thermal shield.® Two~-group fluxes were calculated by this method for two cases — with no fuel salt in the core and with the core filled with salt containing enough U=35 4o give a keff of 0.91 (about 0.76 of the critical concentration). Although the neutron detectors respond primarily to thermal neutrons, it is enlightening to look at the fast neutron distributions because most of the thermal neutrons reach the vicinity of the detector as fast neutrons and are slowed down locally. Figure 3 shows the fast flux (normalized to one neutron from the external source) at the core midplane along a diameter which intercepts the locations of the neutron source and the fission chambers. With no fuel in the reactor, the fast flux was higher in the gap between the reactor and thermal shield, on the opposite side of the reactor from the source, than in either of the immediately adjacent regions. This implies that, under these conditions, most of the fast neutrons that reached the vicinity of the fission chambers arrived by way of the annular gap and that very few were transmitted through the core. SD. R. Vondy and T. B. Fowler, "Equipoise Burnout: A Reactor Depletion Code," USAEC Report, Oak Ridge National Laboratory, (in preparation). €A discussion of the calculational procedure is given in the appendix. 11 UNCL ASSIFED ORNL DWG. 64.82 10 i 10 $,/s 10‘“ 10”7 10-6 0 50 100 150 200 250 300 350 X {cm) Fig. 3. Fast Flux Profiles at Midplane of MSRE Core Along a Diameter Through the External Source. 12 The addition of fuel to the core increased the fast neutron source by adding fission neutrons and leakage of some of these neutrons from the core raised the fast flux near the fission chambers by a factor of 3. The fast fluxes on the source side of the reactor vessel were not af- fected by the addition of fuel at this concentration. (It was assumed that the external source was strong enough that internal non-fission sources were negligible in comparison). Figures I and 5 show parts of several thermal flux contours at the core midplane with no fuel salt in the reactor (Fig. 4) and with salt containing 0.76 of the critical UP°° concentration (Fig. 5). The contour lines are superimposed on scaled drawings of the reactor model used in the calculations and the relative positions of the external source and the neutron detectors are indicated. Table 3 giwves the ratio of the thermal neutron flux at a chamber to the external source strength. In the cases of the 120° and 150° vertical thimble locations the flux is that at the center of the thimble. For the tubes in the instrument shaft, which slope away from the reactor, the average flux seen by a chamber depends on its length. Comparison of the two figures and the numbers in the table shows quite clearly that the thermal neutron flux in the gap and in the thermal shield, for a considerable distance from the source, is highly insensitive to conditions in the core. As a result the counting rate of a chamber in the 120° thimble is a much poorer indication of changes in the core than is the counting rate of a chamber in the instrument shaft. Bffect of keff on Flux An approximate relation between the flux, or counting rate, and keff can be obtained by interpolation of the results calculated for keff of 0, 0.91 and 1.0. The manner in which the flux varies with ke can be approximated f in the following way. Represent the flux at a particular location by f S T .o, X X in in ¢_bsx+l_k+l_k (1) UNCL ASSIFIED ORNL DWG, 64-8218 150° Chamber Fission Chamber Chamber QO D o Y (cm) Thermal Flux Contours, Per Unit Source Strength, Midplane Fig. 4. of MSRE (No Fuel Salt in Reactor). 14 UNCL ASSIFIED ORNL DWG. 648219 I pritgnng 150° Chamber Y (cm) = 0.91) Fission hamber Thermal Flux Contours, Per Unit Source Strength, at . (k eff Fig. 5. Midplane of MSRE 15 Teble 3. Thermal Flux From an External Source Av. Flux/Source Strength Location Eziggir [(n/cfia-sec)/(n/se¢)] (in.) no fuel k_pe = 0.91 120° thimble any 13 x 10°® 18 x 107® 150° thimble any L x 107° 9 x 10-° Instr. Shaft (~180°) 6 2 x 107° 7 x 1078 Instr. Shaft 26 6 x 1077 1.7 x 1078 The term bSX is the flux due to neutrons bypassing the core and should be insensitive to ke . The factor fX includes the probability that T external source neutrons will get into the core and also a shape factor for the fission neutrons produced. Its value will depend on keff’ and 1s probably quite low at k = 0, reflecting the low probability that source eff neutrons will be transmitted through the core. Assume a linear increase with k . The value of f. should not change as much with k SO eff in e £’ assume that it is constant. With these assumptions ak c Sin ¢ =+ x5 tT—x% - (2) where a, b and c are constants. The value of ¢ for each chamber location can be calculated from the critical flux distributions. Values for a and b can be calculated from the two Equipcise Burnout results at k = O and k = 0.91. Values for the various proposed locations are given in Table 4. These relations were used to estimate the reactor behavior and source requirements under sub- critical conditions. 16 Table 4. Flux/Source Factors Chamber Location Length a(cm™®) b(em™2) c(em™@) (in.) 120° thimble any 5x 1077 13 x 10°® 3 x 1077 150° thimble any 5 x 1077 b x 107 3 x 1077 Instr. Shaft 6 L x 1077 2x 107 1x1077 Instr. Shaft 26 1 x 1077 6 x 1077 3 x 1078 Note: ©See text for definition of a, b and c. SAFETY REQUIREMENTS When excess reactivity is added to a reactor which is initially operating at a very low power, the fission rate must increase by several orders of magnitude before the inherent shutdown mechanism of the negative temperature coefficient of reactivity becomes effective or a rod drop is initiated by the high-level safety circuits. (There is no period scram on the MSRE.) Since this power increase takes some time, a substantial amount of excess reactivity may be added by a continuing reactivity ramp and the power may be increasing with a very short period by the time the various shutdown mechanisms begin to act. In the so-called "startup accident" so much excess reactivity is added that severe power and temper- ature transients may be produced desplite the action of the shutdown mechanisms. In such accidents, provided the fission rate follows the behavior predicted by the nuclear kinetics equations, the severity of the tempera- ture excursion is uniguely determined by the rate of reactivity increase, the characteristics of the inherent and mechanical shutdown mechanisms, and the initial power (the mean value of the initial fission rate). If, however, the initial fission rate is extremely low, statistical fluctua- tions about the mean may permit wide variations in the amount of excess reactivity which can be introduced before the power reaches a significant 17 level. The problem is described by Hurwitz et al. in a recent paper7 as follows. '"When a reactor is started up with an extremely weak source, there will be an initial period of time during which the power level Is so low that statistical fluctuations are important. FEventually the power level will rise to a sufficiently high level so that further statistical fluctuations have negligible effect. The influence of the statistical fluctuations in the early stage of the startup will, however, persist through the high level stage 1in the sense that the early statistical fluctuations determine the initial conditions for the high level stage.” In the case of the MSRE, the inherent alpha-n source produces more than 10° neutrons/sec in the core whenever the uranlum required for criticality is present, and the fission rate is already in the high level stage (statistical fluctuations unimportant) at the outset of any startup accident. Furthermore, kinetics calculations have shown that the initial fission rate sustained by the inherent alpha-n source is high enough to make tolerable the worst credible startup accident, which is described in the following paragraphs. The meximum rate of reactivity addition that can be achieved in the MSRE results from the uncontrolled, simultaneous withdrawal of all three control rods. The rate of reactivity addition depends on the type of fuel in the reactor {(which determines the control-rod worth) and the position of the rods with respect to the differential-worth curve. The most severe rod-withdrawal accident involves fuel C and the maximum rate of reactivity addition is 0.08% ék/k per sec. (A higher reactivity ad- dition rate, 0.10% per sec, can be obtained with fuel B, but this mixture also has a larger negative temperature coefficient of reactivity so the resultant power excursion is less severe. ) For shutdown margins greater than 2% 6k/k and reactivity ramps between 0.05 and 0.1% per sec, the power level of the reactor when k = 1 is about 2 milliwatts if only the inherent alpha-n source (4 x 10° n/sec) is present. Figure 6 shows the power and temperature excursions that result with fuel C when all three control rods are moving in the region of maximum " "H. Hurwitz, Jr., D. B. MacMillan, J. H. Smith and M. L. Storm, "Kinetics of Low Source Reactor Startups. Part I, Nucl. Sci. Eng., 15, 166—186 (1963). 18 UNCLASSIFIED ORNL DWG. 63-8177 1900 soipinpied npni b L e ok 1800 - 1600 1500 1400 TEMPERATURE (°F) 1300 1200 koo 300 POWER (Mw) 8 it 0 2 4 6 8 10 12 1 16 TIME (sec) I'ig. 6. Power and Temperature Transients Produced by Uncontrolled Rod Withdrawal, Fuel C. 19 differential worth when k = 1 for this condition. In this calculation the nuclear power reached 15 Mw (the level at which the reactor safety curcuits initiate corrective action) 7.5 sec after criticality was achieved. At that time 0.6% excess reactivity had been added and, since the nuclear average temperature of the fuel (T;) had risen less than 2°F, almost none had been compensated by the temperature coefficient; the reactor pericd was 0.1 sec. In the absence of action by the safety system, intolerably high fuel temperatures would be produced by this accident, not as a result of the initial excursion but as a result of the continued rapid rod withdrawal afterwards. Figure 7 shows the results of a calculaticn of the same accident in which two of the three control rods were dropped (with a 0.l-sec delay time and an acceleration of 5 ft/sec2)8 when the power reached 15 Mw. The temperatures reached in this case would cause no damage. Thus the inherent alpha-n neutrcn source is adequate from the standpoint of reactor safety. Because the startup accident is safely limited with only the inherent source in the reactor, it is not a safety requirement that any additional source be present during startup. Nor is it necessary for safety that instrumentation capable of '"seeing" the inherent source be installed, because 1ts presence is certain and does not have to be confirmed before each startup. INITIAL, STARTUP EXPERIMENTS Although it is not a safety requirement, the presence of a source- detector combination which permits monitoring of the fission rate in the subcritical core 1s necessary for convenient and orderly experimentation and cperation. BMhese values are conservative estimates of the rod characteristics based on tests with a protolype assembly. 20 UNCLASSIFIED ORNL OWG, 430178 6 6 10 12 L 16 TIME (sec) Fig. 7. Effect of Dropping Two Control Rods at 15 Mw During Uncontrolled Rod Withdrawal, Fuel C. 21 More subcritical observations will be made during the initial nuclear startup experiments than at any other time. For these experiments it was expected that temporary neutron counting channels would be set up, using sensitive detectors. The thimbles in the thermal shield were included for this purpose, to obtain a higher average flux at the chambers than could be obtained in the nuclear instrument shaft and also to provide for installation of detectors at more than one location. The vertical thimbles will accomodate 30-in-long BFs chambers, with a counting efficiency of 14 counts per n/cm?. Even with these sensitive chambers, the inherent source in the salt (containing only the depleted uranium) before the addition of the enriched uranium is inadequate to give a significant count rate. Because it is desirable to have a refer- ence count rate at practically zero multiplication, an extraneous source is required for the critical experiment. Furthermore,in the determination of the critical point and possibly in the calibration of the control rods, it is convenient to be able to remove the major neutron source and observe the decay of the flux. For these reasons, a removable external source should be provided for these experiments. The flux at the various chamber locations, from an external source, at any value of ke can be estimated from Eg. 2 and the factors in Table 4. The inteiial source strength increases linearly with the amount of enriched uranium in the core, reaching about L x 10° n/sec at the clean critical concentration. The flux from this source can also be estimated from Eg. 2 and Table 4. The predicted variation of keff with enriched U concentration is necessary for this calculation, and this relation is shown in Fig. 8. The method of attaining the critical concentration will be to add increments whose sizes are determined by a plot of inverse count rate vs amount of enriched uranium already added. Fig. 9 is such a plot, generated from the flux calculations described above and the k vs C relationship from Fig. 8.. Neutrons from both the internal source and an external source of 10° n sec are included. The bowing of the curves in Fig. 9 reflects the contribution of neutrons which are scattered around the outside of the reactor from the source to the detectors. As would be expected, the error is largest for 22 UNCL ASSIFIED ORNL DWG, 64.8220 g al + - i al - — - 4 l - v . + - T e + y 5 v v 1 T L | & + v 1 i - 3 ' -+ . 1 i# + - - -» Az — + 14 iy . T T T v . - X _ ot + r } 4 il n’ + Y T 1 -+ 1t - — — + + + st ’ - - - o T -+ —— - 4 3= : : w0 -— - v - 4 i . -+ » F-TX ' - O 1= : : - 4 -+ - e - * + —————p—— - iy » ~+ » ST L - - 4 - - * ———r - 1 / : \ = d - " + i 5 i . : % IS - 3 ! > T T i 6 . - + — r pe - ] - T O — . — - —- - - Wy — - — . e | ol ——— e » -— X, » N JpS—— Jy . - e v b e e e -] -t e . — b e Fee e b s 8- +¢+ls||.’ ——— - — ‘ - —r - -4 1 — . o et ol - e - 8. _— bt e ’r - - . Iu e e e e e —_—— e . e i c e b owwe o ——— L - 4=z : —ENE =TT e - O ..... br v e e e e e e e e e ,. — — r—e Prad e e Y - - e o e e . -t - . o e ey e e s . - - - A ———— s - st - L L L L Ly LNy L Y T Y LY e e et —— 1 . 4 - e e e et ftrt o e oty b - e -———— - - e b - —— b e - Py T e e e N Tt e T . o e - st e st b fe e e § o m e ol e e bl cr-Far e e omm e g = carrpoerrmfoomadedsopremgoareatrrmitmatrde et oo - e 4 o ek = e L+ e el e s v ot baardrrr s bt ce e e ey -y pe - . v e b vt b i b b - s eas oo § vieoe Joe v d - o - - b — - - .luurlill —g e a e d e B e dm b f e b bt pae bt G oo - e fpe s e p o - ceedid e i o e ey Y et - — + - - st fem e et e ey bt i b - el i cemede rr e e e - _———d— l!t;l,-!. = ] -t bi et et M 2 e o pmasd . —e e ar s g = rn st rrrvilhrn rr e o crr e e et e R e eh g be e - . rra fr e fornr e e e f s ot - dmmee frearmre B e e b dn e B, e b B o e . cea s s b sttt et et g s - - e fe et - .- e hrmrad o s e ofm o e e e e foe e et e s ars s gr et reedie o W - - oseodpaca cbamasvf s adr-ard snn et a e mr hm s e g - - : ottt e b ee v b nnnfan o [PUSEPI USSR S + e gala - e e e 4 Lo e e b W v} oa e Bt e ad s are v e a e g . oo . . et e e i - - . e = A e - ety - - -y - R e b e ',&J..wl.lo.'al 7 e as s e —et §. + - i -+ —— e Caed e raed e . T . — + o - —f ot e - - -t : 2o fps evtoneed o b A 4o rng b ey - - - . - -5 cm e s d e et e + by e Bt - . - - - 1 + —t - trof esretnps et - e ’ . oot e B iia e 4 - —4 + - _— . b - i - Rt e . ———t . e i — oLy .- .3, . e e .- i s s srBes s besasPomrifirme e b w e b - af----4- + T N 0 " g = A et e A v s o it ot b b et f Ak e e = e e e v m v m o - —+ c/c 1011. £ with Fuel U237 Concentrat e Variation of k Fig. 8. 23 UNCL ASSIFIED ORNL DWG, 64.8221 c/c External : 107 n/sec. Inverse Count Rates vs U?3% Concentration. Fig. 9. Source Strength 2k the detector located nearest the source. Because the bowing makes ex- trapolation less accurate, the instrument shaft is the most suitable location for detectors in the approach to critical. Therefore the ex- ternal source for the critical experiment should be at least strong enough to give a convenlently high count rate at a chamber in the instru- ment shaft before any enriched uranium is added. A source of 1 x 10° n/sec would give a count rate of 10 c/sec on a chamber with a counting effi- ciency of 1k c/sec/n/cm?—sec under these conditions. The strength requirement just stated is a minimum for starting the critical experiment. The initial approach to criticality will include experimental determinations of control-rod worth and concentration coef- ficient of reactivity. These determinations are based on count-rate measurements with and without the external source present. Therefore, it must be possible to obtain a substantial difference in count rate by removing the extermal source. Since the intermal, alpha-n source in- creases in intensity with increasing uranium concentration, the external source must be strong enough to make the contribution from the internal source small by comparison when the uranium concentration is near the critical value. The flux calculations indicate that an external source of 1 x 107 n/sec would produce a flux in the instrument shaft at least 100 times that from the internal source at all points during the approach to critical. The differences in count rate which can be obtained with a source of this strength are illustrated in Fig. 10. This figure shows the reciprocals of the count rates predicted for a BFs chamber (counting efficiency of 14) in the instrument shaft as the critical point is ap- proached. The two curves are for the internal source alone (external source withdrawn) and with both the internal source and the external source. NORMAL OPERATIONAL REQUIREMENTS An external source of neutrons is required during normal operation of the reactor to permit the convenient monitoring of the reactivity during routine startups. (The safety requirements for a source are satisfied by the inherent alpha-n source.) 25 UNCL ASSIFIED ORNL DWG, 64-8222 PflEx® and produce from 1.6 x 10% to 1.6 x 107 n/sec. (The largest sources are too big to fit 29 into the MSRE source tube, however.) Although the problem of source decay is avoided, the Pu-Be source has several disadvantages. The initial cost is high and plutonium containment must be guaranteed at all times. Also, the heat generation by fission in the plutonium (about 10 kw at a reactor power of 10 Mw) would probably require that the source be re- tracted during high-power operation. RECOMMENDATIONS 1. Install in the spare tubes in the instrument shaft two additional neutron chambers with a much higher counting efficlency than the servo- driven fission chambers. Use these during the Ilnitial critical experi- ments and for monitofing the filling stage of routine startups. Change the interlock requiring a dependable count rate prior to filling from the servo-driven-fission chambers to these chambers. 2. As soon as the installation of the reactor and equipment permits, check the flux/source ratio calculated for the core with no fuel. (This will greatly reduce the uncertainty in the source strength requirements.) Any source of 10°® n/sec or more will serve for this preliminary experiment. 3. Procure or manufacture a source which will meet the operational requirements. The choice of the source type and its strength must be based on the observed flux/source strength ratio and the considerations described in the preceding section. If the actual flux/souroe ratic is near that calculated, the choice for a source would be either a 5-curie Pu-Be source (8 x 10° n/sec) or a pair of Sb-Be sources which would pro- duce 3 to 5 x 10% n/sec (from about 125 curies of Sbl**) after an 8-week irradiation in the LITR. After one Sb-Be source had been in the MSRE for about 10 months, the other would be placed in the LITR for irradiation to be ready for exchange when required. If the Sb-Be sources are used, it will be desirable to modify the existing provisions for source insertion and removal to make the operation less time-consuming and costly. OSpeci- fically, an access port should be provided through the steel cell cover and the lower shield plug directly over the source tube. 30 APPENDIX CALCULATION OF FLUX FROM AN EXTERNAL SOURCE The neutron scurce for the MSRE will be installed in a thimble in the thermal shield, about 20 in. from the outside of the reactor vessel. The variocus neutron detectors will also be, in effect, in the thermal shield at different circumferential positions. This results in a highly unsymmetrical cylindrical geometry for conditions in which neutrons from the source contrivbute substantially to the neutron flux. The source is also short, compared to the height of the reactor, so an accurate cal- culation of the flux at the detectors resulting from the source would require the use of %-dimensional, cylindrical (r, e, z) geometry. Since there is no reactor program available for treating this problen, a number of approximations were made to reduce the problem to one which could be handled with existing programs. Geometric Approximations The program used for the flux calculations was the Equipoise Burn- out Code, a 2-group, 2-dimension neutron diffusion calculation with pro- visions for criticality search. This code uses rectangular (X-Y) geometry and is limited to 1600 mesh points. In order to treat the azimuthal non-symmetry, the calculations were made in a horizontal plane through the reactor and thermal shield at the midplane of the core. The axial dimension of the reactor was represented by a constant geometric buckling in that direction. Per- turbations caused by the neutron detectors were neglected, so the plane of the calculation had an axis of symmetry along the diameter which intercepts the positions of the source and the fission chambers. There- fore, it was necessary to describe only one-half of the plane in the mathematical model. The limitation to X-Y geometry required that the various regions be represented as collections of rectangles. The main portion of the core was made equal in cross-sectional area to the actual core with the transverse dimensions equal to core radii. These 21 two requirements determined the size of the "cutouts" at the corners of the otherwise square core. The regions surrounding the core (i.e. the peripheral reglons of the reactor, the gap between the reactor and thermal shield, and the thermal shield) were assigned transverse dimensions equal to the radial dimensions of the actual components. Figure 11 is a diagram of the resultant model. Because of the mesh point limitation in the Equipoise Burnout program, 1t was necessary to omit some physical detail in the cal- culational model. The control rod thimbles near the center of the core were neglected in this model, as were the variations in fuel fraction and graphite fraction in that region; the main portion of the core was treated as a single homogeneous mixture. The peripheral regions of the reactor, including the core can, the fuel annulus, and the reactor vessel wall, were all homogenized into a single region (reactor shell in Fig. 11). The materials in the gap between the reactor and the thermal shield (heaters, heater thimbles, insulation and insulation liner) were all homogenized and uniformly distributed throughout the gap. All the structural material inside the thermal shield was also neglected. A total of 1225 mesh points in a 49-by- 25 array were used to describe the calculational model. Nuclear Approximations Two-Group Constants The Equipoise Burnout program has provisions for calculating 2- group nuclear constants 1f the necessary microscopic cross-section data are supplied as input. In this case, however, it was more ex- pedient to calculate the Z2-group constants separately using MODRIC, a l-dimensional, 33~group calculation. Multi-group cross sections were prepared for the MODRIC program using GAM-1 and the existing cross-section library for that program. A radial criticality cal- culation was then made with MODRIC for the reactor model with fuel salt containing 0.6 of the critical concentration of U235; the presence 32 UNCLASSIFIED ORNL DWG. 64-8223 O — Fission 150° BFa Chamber '/L/ Chamber @ 4o — Thermal Shield ’ Chamber o LA b 120 ] o eactor | . Stainless (1/ Shell '/_L/;;/‘Steel g o ;: oo Core Gap Water 200 | 240 | 280 ! —~ Source 320 0 40 80 120 160 Y (cm) Fig. 11. Cross Section of Model Used in Subcritical Flux Calculations. gt 35 of the external source was neglected in this calculation. The 2-group constants generated by MODRIC were then used to calculate the flux distribution in the 2-dimensional model with the source present. MODRIC calculations were also used to estimate the 24group con- stants for the case with no fuel salt in the reactor. In order to get group constants for the core with only the graphite moderator present, calculations were made for the normal density of the dilute fuel and for densities that were 0.5, 0.25, and 0.1 of the normal value. The Z2-group constants obtained from these calculations were plotted as a function of fuel density and extrapolated to zero density to get constants for the reactor containing no fuel (only graphite). For several reasons, the above procedure does not lead to com- pletely accurate values for the 2-group constants. The fast-group constants in a given region depend, to some extent, on the energy distribution of the neutrons in the region. This energy distribution is different if all of the neutrons are born in the core (as was assumed in the MODRIC calculations) than if a substantial number are born in an external source region (as was the case in the 2- dimensional calculations). Some additional error is introduced by the fact that neutrons born in the reactor and those born in the external source have different energy distributions at birth. Neutrons born in the reactor are products of the fission process and have an energy distribution that corresponds to the fission distribution; the fast-group constants were calculated for this birth-energy distribution (10 to 0.011 Mev with an average of about 2 Mev in the MODRIC program used). The energy distribution of neutrons produced by a source depends on the nature of the source. For an Sb-Be source, the average neutron energy is about 34 Kev. In the Equipoise Burnout calculation the fast-group constants calculated for fission-source neutrons were applied to all the fast neutrons, regardless of their point of origin, Since absorption cross sections generally increase with decreasing neutron energy, this treatment overestimated the neutron flux at the chambers for a given neutron source. 3h Source Configuration In the Equipoise Burnout calculation, the source region was treated as a slender (2 cm by 2.6 cm) prism extending along the entire height of the reactor model. This is a consequence of applylng a constant axial buckling to all regions. As a result, this source is less efficient in terms of producing a neutron flux at the core midplane than a short source of the same total strength located near the mid- plane. No correction was applied for the higher efficiency of the short source because this underestimate tended to counteract the over- estimate inherent in the cross-section treatment. Composition of Thermal Shield The thermal shield is filled with steel balls to provide a mixture that is approximately 50% iron and 50% water. However, this mixture does not fill all portions of the thermal shield. The source thimble and the special counter thimbles are protected by half-sections of 8-in. pipe which were welded to the inside of the shield to prevent damage to the thimbles during the addition of the steel balls. As a result, each of the thimbles is surrounded by a layer of pure water. The nuclear instrument shaft, which extends to the inner wall of the thermal shield, contains no steel balls. The only material in this shaft, other than water, is the aluminum in the guide tubes for the neutron chambers. The neutron flux at the various chambers is influenced more by the water layer immediately adjacent to the thimbles than by the lron-water mixture in the rest of the thermal shield. Therefore, the presence of the steel balls was neglected in the flux calculations., This leads to highly erroncous fluxes everywnere in the thermal shield except in the immediate viecinity of the neutron chambers. Use of Diffusion Theory The Equipoise Burnout program which was used to compute the flux distributions is based on a diffusion theory treatment of the neutron transport problem. This program was used because is was the only one * - 55 Judged To be practical for an approximate calculation of the external source requirements for the MSRE. It is well known that diffusion theory has limitations which are imposed by the basic assumptions in the development of the mathematical treatment. These limitations restrict the accuracy of the theory in regions with high absorption cross sectlon, near region boundaries and in regions where the neutron mean free paths are long. ©Since all of these factors were present in the calculational model, the results of the calculations can be regarded as no more than preliminary estimates. It is likely that the calculated fluxes are at least within an order of magnitude of the correct values but it is not possible to define the limits of the probably error. 79-80. 8L. 82. 83. 8h . 85. 86-100. 1! INTERNAL DISTRIBUTION MSRP Director's Office Rm. 219, 9204-1 Adams Affel Alexander Anderson . Ball Beall Bettis lumberg Borkowski Brashear Burger Crowley Ditto Dunwoody Engel Epler . Fray . Gabbard Gallaher . Gelst . Guymon Hanauer . Harley Haubenreich Herndon Hise Holt Holz outzeel . Hudson . Kedl Krakoviak Krewson oM G2 QN }1$=?jr3$=w f:mlp g g aaEEHGgZEZ g Qg Em,eE el .Z‘.HC_iI:"!I!"UUflflzmMMMWMZWWMQWCflDJQUflmm 69- 70 fL-T72, 737k, 5=77. 78. EXTERNAL DISTRIBUTION U!K =4Hdmd -y R . Cope, Reactor Division, AEC, ORO Garrison, AREC, Washington . Philippone, Reactor Division, AEC, ORO . Roth, Division of Research and Development, AEC, ORO omalley, Reactor Division, AEC, ORO Whitman, AEC, Washington ivision of Technlcal Information Extension, AEC, ORO * = F’dfl?lfijfii?’fliw HFrP2gUOODEgwHepdoaoadsd-dQw ORNI, TM-935 Lindauer Martin MacPherson McCurdy McDonald McDuffie McGlothlan Moore . Payne Perry . Piper . Prince Redford Rlchardson . L—“‘tljbdlg’;dt“fi'flbj@fltflbd . C. Robertson . C. Roller Scott H. Shaffer . J. Skinner N. Smith Spiewak C. Steffy A. Swartout Taboada. Tallackson Thoma. Ulrich Webster Welnberg . West Za2moHEx Central Research Library Document Reference Section Reactor Division Library Laboratory Records ORNL-RC