VARIATIONS IN A y BACKGROUND IN UNDERGROUND LOW-BACKGROUND SYSTEMS

I. O. Nevinskii and T. V. Tsvetkova

UDC 539.166:550

Preliminary results have been reported [1] on the variations in background count rate in a photon detector in a passive multilayer shield in an underground laboratory [2], which have been supplemented and revised during the subsequent year of observation. For example, Fig. 1 shows the variation in background rate during the day as averaged over a year. There are substantial variations in the scale of the sinusoid and in the mean daily values from day to day, which can be related to atmospheric processes, solar-lunar tidal phenomena, and seismicity. For example, the spread in the sinusoid and the variations in the mean daily background rate increased before large world earthquakes. Figure 2 shows the variation in background rate in the annual succession. The data differ from the curve given in [1] for the previous year because of increase in seismicity (for example, the Spitak earthquake) in certain months, which has led to oscillations in the mean monthly values. The bursts mainly of hour duration indicated in [1] in excess of periodic changes have been related to the local seismieity; Fig. 3 shows the rise in background before a strong earthquake in the North Caucasus (the Garagor earthquake). The seismic data for the Caucasus region show that the bursts arise 9 +_ 1 day before all regional earthquakes in which the epicenter is not more than 200 km from the laboratory. No direct sources of such changes have so far been identified. To elucidate the causes of the bursts, we examine data from cosmic-ray counters close to the apparatus (plastic scintillators with total area 1 m2), a photon detector (Nal(TI) crystal) outside the passive shield, and a beta counter for the radon daughter products in the working volume of the shield. Peaks in the gamma detector in the low-background shield were no~ accompanied by analogous changes in count rate in all the other detectors. All the same, if we take these data as preliminaD,, it is necessary for the elucidation of the peak change in background count rate before an earthquake to observe variations in cosmic-ray flux, photons outside the low-background shield, and variations in radon in the atmosphere of the mine by independent methods (not from the y activity), e.g., by direct counting of the c~ activity in the air. In [3], there was a discussion of photon generation in the low-background shield materials due to neutrons generated by rock deformation before earthquakes. Although no such changes in background count rate were observed during the strongest earthquake, research on neutron variations under the ground, particularly in various energy ranges, may be of independent importance. Bursts have been observed in a shield with a different geometry described in [4]; also, the data are not unique or confined to the region of the laboratory [2]. An underground laboratory has been built at a depth of 130 m water equivalent in Krasnodar area in the Sakhalin mine in order to examine background variations in low-background equipments in relation to the seismic activity of the Caucasus; this showed analogous peak changes in background count rate before regional earthquakes. The background characteristics of the laboratory correspond to those described in [2]. A detector analogous to that used in [2] was placed in a lead shield 20 cm thick. Low-background underground experiments enable one to obtain unique geophysical information without adversely affecting the main measurements. For experiments requiring a particularly low and stable background, one should select underground workings in regions with minimal local seismicity. An optional design may be to use a second analogous detector in the same working volume or an analogous low-background shield, with a previously established correlation coefficient between the background count rates. Then measurements, for example on specimens for radioelement content, may be performed continuously by correcting for the background change in the auxiliary detector. If the background is known at the time of the measurement, one can improve the accuracy, in contrast to the method in which there is a cycle of background and sample measurements.

Caucasus Geocosmic Observatory, Central Geology and Geophysics Research Institute. Translated from Atomnaya Energiya, Vol. 72, No. 6, pp. 622-623, June, 1992. Original article submitted June 27, 1991. 2063-4258/92/7206-0543512.50 01993 Plenum Publishing Corporation

543

trip P counts/h ~

counts/h

60 50

i0000000000000 0 ~0 O0O00U8

+0 M i

Z0

I

[

|

l

6

M

IG

gt

l

20 O0

~, h

i

i

1

f0 IZ

i

i

Oe

I

l

f988

Fig. 1

1

I

Oq" 08

i

i

08

I

l

t0

f908

Fig. 2

Fig. 1. Changes in background count rate during the day averaged over 1988-1989 for the energy range 1.7-2 MeV (local time), error of measurement not more than 2%. Fig. 2. Variation in background count rate in the range 1.7-2 MeV from month to month.

"0'¢~,counts/h

tOO

I I

25.07 tO gO Z6.07~ h

Fig. 3, Variation in the Y background in the range 1.7-2 MeV before the earthquake of 03 August 1989.

As there are monthly variations in background count rate, we used a parallel background channel in an experiment on double/3 decay for 58Ni (preliminary measurements with a two-crystal gamma spectrometer and subsequently on an apparatus analogous to that in [4]). A simple design of low-background system is a small NaI(TI) crystal in a lead shield, with data output through discriminators to a scalar with timer, which can be recommended as counters for radon as most readily automated in order to evaluate radiation hazards in mine workings and obtain parallel information on the seismic setting in the region. It is necessary to examine the variations in background count rate in other energy and time ranges (down to minutes) by means of large crystals in order to obtain statistically reliable data. LITERATURE CITED

1,

G. I. Buachidze, I. O. Nevinskii, S. Yu. Red'ko, and T. V. Tsvetkova, "Gamma-background variations in an underground

2.

laboratory," At. l~nerg., 69, No. 2, 93-94 (1990). G. I. Buachidze, I. O. Nevinskii, T. V. Tsvetkova, and S. Yu. Red'ko, "Background characteristics of the S G I G underground laboratory, Georgian Academy of Sciences," ibid., 66, No. 5, 335-336 (1989).

544

I. O. Nevinskii, T. V. Tsvetkova, and M. A. Yaroslavskii, "Radiation underground associated with the runup to an earthquake," in: Abstracts for the 1 lth All-Union Symposium on Solid-State Mechanochemistry and Mechanoemission, September 1990 [in Russian], Chernigov (1990). 4.

I. O. Nevinskii and T. V. Tsvetkova, "A low-background chamber with large working volume," At. t~nerg., 70, No. 2, 2122 (1991).

OPTIMIZING

FUEL-PIN

ASSEMBLY

LOCATION

IN A

RESEARCH REACTOR

UDC 621.039.5

Yu. P. Malers

An algorithm has been considered [1] for optimizing fuel-pin assembly location in a research reactor, which is based on an analytical relation between the power distribution and the fuel concentration, together with sequential linearization and partial integer programming. Implementation has revealed problems, which have required a better formulation of the algorithm and certain changes. The relation between the fuel concentration and the specific power production is [1] N

ui=Qil E21k(kef)Q P

i= 1,2 ..... N,

(1)

k=l

in which u i and Qi are the fuel concentration and specific energy production in cell i in the finite-difference net correspondingly, N is the number of controlled zones (cells) in the core, kef the effective neutron multiplication coefficient, and 2ik is a certain matrix that is independent of u and Q and has a linear dependence on 1/kef. A similar relation can be derived for a two-group diffusion model for a light-water reactor. The fuel characteristic here is taken as the effective multiplication coefficient in the fast group with a first-order correction for the thermal-neutron leakage

[2, 31. Consider the following optimization: locate N 1 assemblies with fuel concentration Yl, N2 assemblies with concentration m

Y2, "', and N m assemblies with concentration y,, (~] Nk = N) in N cells to minimize a target function F subject to k=l N

N

Ui = Q i 1 2 2 i k ( k e f ) Q k, i = 1 , 2 . . . . . N; Q <- Qi -< Q m ~ ' i = l , 2 . . . . . N , P = 2 VIcQk <- Pmax' k=l k=l

in which V k is the volume of zone k and P the total reactor power. Here we take a finite-difference net with one node per assembly and assume that there is no need to restrict the changes in the effective neutron multiplication coefficient. We write u i as N

N

Ui = 2 ~ik'Yk' i = I , 2 . . . . . N , w h e r e k.=1

m

0_
..... . , , ~ ¢ ~ = l , i = t , 2

k=l

..... N;

k=l

with all the ~ik integers. We express the variation in u i in terms of the variations in ~ik and Qi: N

N

k=l

k=l

(2) Nuclear Research Institute, Ukrainian Academy of Sciences. Translated from Atomnaya l~nergiya, Vol. 72, No. 6, pp. 624-626, June, 1992. Original article submitted November 18, 1991. 2063-4258/92/7206-0545512.50

"1993 Plenum Publishing Corporation

545