•
V
•
.
I
U"
I OFT
ORNLP
3319

-
-
-
-
-
01
.
•
:
::..:
1
-
.
.
.
.
.
.
:
1
...mert
.
.
W
FEEFEEEE
i
|| 1.25 11.4 11.6
MICROCOPY RESOLUTION TEST CHART
NATIONAL BUREAU OF STANDARDS - 1963
WR.
'
!
.
kerty
2
.
LT
/
191.
.1.-
1.
HT
M
-----
..
.ILLUE --.-14
ORNL of 3319
. Conf.670602--26
RECEIVED BY DTIE AUG 29 30
CESKI PRICES
Ac
= m ,
MASTER
COMMENTS ON SHIELDING STANDARDS *
E. A. Straker
Oak Ridge National Laboratory
Oak Ridge, Tennessee
*Research sponsored by the U. S. Atomic Energy Commission under contract
with the Union Carbide Corporation. Presented at the Panel Discussion on
Shielding Standards of the ANS Annual Meeting, San Diego, California, ,
June 11-15, 1967.
LEGAL NOTICE
This report was prepared as an account of Government sponsored work. Neither the United
States, nor the Commission, nor any person acting on behalf of the Commission:
A. Makes any warranty or representation, expressed or implied, with respect to the accu-
racy, completeness, or usefulness of the information contained in this report, or that the use
of any information, apparatus, method, or process disclosed in this report may not Infringe
privately owned righto; or
B. Assumes any liabilities with respect to the use of, or for damages resulting from the
use of any information, apparatus, method, or procesa disclosed in this report,
As used in the above, "person acting on behalf of the Commission" includes any em-
ployee or contractor of the Commission, or employee of such contractor, to the extent that
such employee or contractor of the Commission, or employee of such contractor prepares,
disseminates, or provides acce68 to, any information pursuant to his employment or contract
with the Commission, or his employment with such contractor.
DISTRIBUTION OF THIS DOCUMENT IS UNLIMITED
.
.
LV
."
'
"
"
"
'
"
.***. ***** ***-77793: ?VITIT
,*
!....
.....
.
.
..
.
***********
******
..
.
..
1
**************
!:**
WEIT
AYAW
W
A
.
V
...
......
..
-
.
.
.
.
"
****
* .-
-
.-
..-.
.
.
---
...
N
....
- V
7,7;
17"7
:
...
" :
..
1
i sit
"p
"
.ruar
"17
:
T
yhtih
-2-
The advisability and practicality of setting up specific configurations
of materials in geometries which are amenable to calculation and expecting
that these specific cases act as standards which all shielding codes must be
capable of calculating is somewhat questionable. The state of the art of
performing shielding calculations has progressed to the point that rigorous
or exact methods are generally available if the computer costs will fit the
budget. However, these "exact" methods in many cases may represent a much
better approximation to the solution of the transport equation than the cross-
section input justifies. Thus a standard for a rigorous or exact method must
exist for each shieldir.z material and for several source energies or spectra
if the code and its representation of cross sections are truly to be checked.
That is, the test of comparing calculated and measured results is not only a
test of the technique alone but also of the cross sections. Thus benchmark
problems can be used to intercompare codes but the results may still be far
from reality.
If the less exact kernel codes are to be checked, then many "standard"
geometries and material arrangements are needed because of the need of
supplying the properly weighted "cross sections" whether they be removal or
transport cross sections or attenuation coefficients for integral quantities.
For laminated shields these cross sections are somewhat less well defined.
Although there are many cases in which these approximate solutions have been
proven to be adequate, the extrapolation of their use to an untested domain
is indeed dangerous. Because of the wide range of types of shielding prob-
lems encountered it is probably not feasible to provide standards for each
case.
Although the above comments may sound pessimistic insofar as standards
are concerned, I do feel that a compilation of "evaluated" measurements and
calculations should be made available so that there wouid be in one place an
easy reference to results that would be of use in debugging a code or determin-
ing the adequacy of the approximations made. A compilation which contained
all the pertinent information about an experiment and/or calculations is again
probably not feasible, but a brief resume of the problem investigated with
information concerning the source, geometry, type of results measured or col-
culated and apparent restrictions would be sufficient to indicate the parti-
cular experiment that might be used to check the code.
Some "evaluation" as
to the validity of approach, applicability of techniques and complexity of the
problem would have to be performed in order to make the compilation practical.
Many of the existing papers would probably be eliminated due to the lack on
the part of the authors of publishing enough detail so that a new calculation
could be performed. It appears that a compilation such as this might serve
as a "standard reference" rather than a "standard."
Many of the discrepancies that have existed in t'a past between two
calculations of the same problem or between calculations and measurement are
due to either a poor treatment of the transport or a poor set of input cross
sections or both. The first of these may be eliminated for some problems by
use of so-called "exact" methods. For the second of these, it is difficult
to determine if it is really the source of the discrepancy. Quite frequently
if there is disagreement with an experiment one lays blame on the cross sec-
tions but it is frequently impossible to determine whether this is the case.
The more detailed the measurement or calculations the easier it is to pin-
point the source of the disagreement. For instance, if a measurement of dose
exists for a given problem it is next to impossible to determine what cross
section is wrong if there is a disagreement between calculations and measure-
ments. However, if spectra are measured then by comparison with the
-40
calculations it may be possible to at least determine in what energy ranges
the cross sections are suspect. New input cross sections or finer detail may
pinpoint the problem. More effort is needed in determining the effect of
cross sections on the results of a transport calculation. Some work has been
done at ORNL for specific cases, two of which are air and oxygen, but more
information is needed as to the effect of such things as proper neutron
resonance treatment and its effect, on secondary gemma production, the effect
of large group widths, the effect of different types of flux weighting, the
order of expansion needed for angular distributions, the importance of in-
elastic angular distributions, the importance of the angular distribution of
secondary gammas, the importance of photoneutrons, and the importance of
(n, 2n) and other multiple particle reactions.
To illustrate the acvantage of measuring differential instead of inte-
gral quantities I would like to discuss sore measurements made at ORNL-TSF
in the past year. Comparisons of the results of many calculations of neutron
transport in the atmosphere showed rather large differences even for the same
method of solving the transport equation. Differences in the calculated
spectra indicated that cross sections were to blame and a comparison of the
total cross sections used in some of the calculations indicated that even
those were different. An experiment was designed to measure the uncollided
dose through thick oxygen and nitrogen samples in good geometry, so that the
analysis of the experiment would involve calculating only the exponential
attenuation of the beam, thus providing a check on the total cross section.
Comparisons of calculated and measured doses showed discrepancies of the order
of a factor of 2 for a 24-in.-thick sample. With just this information, both
the experiment and the calculation were suspect. Naively, I thought that
surely the total cross sections were better than that and no other set of
.:
:..."
...'
-
"
.
. -.:.
-5-
cross sections were tried. The experiment was repeated for some other
materials and for other goometries to test the design of the experiment. We
found that the dose transmitted through carbon could be calculated accurately
as could the dose for water. The experiment was redesigned to measure the
spectra of transmitted neutrons instead of the dose.
Comparison of calculated and measured spectre for oxygen readily showed
where the problem was.
There are two energy regions in which the cross sec-
tion being used was not adequate. The comparison indicated thut as might be
expected the minima were determining the dose. Other cross-section sets were
then incorporated into the calculation and it was no problem to determine
which set was adequate; naturally it was the one with the greatest detail of
the minima.
A similar story holds for nitrogen but in the energy regions of dis-
agreement the total cross section was last measured in 1954. An unpublished
set of cross sect" ns was obtained from Foster and Glasgow at Hanford and
these cross sections gave better agreement; however, there is still some
small disagreement.
Because we were dismayed at the apparent lack of adequate total cross
sections for 2 of the 3 materials we measured, it was decided to make mea-
surement on many other shielding materials, nearly all that are commonly
used. Comparison of calculated and measured spectra indicated that in
nearly every case the most recently measured cross-section data appearing in
the literature was the best. For some elements, 1958 BNL data appears
adequate; for others, no cross-section set has been found that is adequate.
Table 1 shows the materials studied and a comment on the best cross section
found to date. For notation purposes "slight or small disagreement" between
the calculated and measured spectra is used to denote differences of less
than 20%. (usually less than 5% in the total cross section). This notation is
used only if there is a disagreement in ore energy range but agreement for
higher and lower energies. In nearly every case the calculated spectra was
low for old cross sections and if new cross sections resolved the minima the
newly calculated spectra approached our: measured spectra.
It is worth noting that in nearly all cases ENDF/B cross sections appear
to be the best; the two exceptions are iron and tungsten. This does not
imply that ENDF/B cross sections are adequate in all cases -- but they appear
to be at least as good as any other set that has been tried. On this basis,
the use of the ENDF/B cross sections is encouraged.
Even if correct total cross sections are used in a code it is obvious
that this does not insure that the effect of the cther cross sections 1s
smali. In fact, some work at ORNL has shown that for high energy sources
inelastic treatment is quite important and for a specific case of transport
in oxygen the proper combination of total and elastic scattering angular dis-
tribution is important.
This is because the angular distributions in the
valleys of the total cross sections are generally much more nearly isotropic
than in the resonance regions. Recently we have calculated dose from a
point fission source in infinite air with both s, and Monte Carlo. If the
same cross sections are used there is agreement within 5% at ranges up to
lý miles. For every case in which the same cross sections were used in both
the S and Monte Carlo codes the agreement has been excellent. This, I think,
shows that two techniques can calculate the same result if proper cross sec-
tions are used.
It appears that for non-engineering shielding design codes the problem
.
in shielding is not in checking à code but in checking the imput cross
-7-
sections. For engineering design codes comparison of results with those
from exact methods may be adequate and may be the best test since it is
possible to use cross section input derived from the same basic nuclear data.
Of course these codes must eventually be compared with experiment. That is,
the approximations in the code may be checked in benchmark type problems and
then the nuclear data for a particular material checked by comparison with
experiment. An example of a suitable experiment might be similar to the
ineasurements of the energy spectra of neutrons emerging at specific angles
from lead, polyethylene, and lead-polyethylene Laminated slabs which have been
made at the ORNL Tower Shielding Facility to provide check data for the two-
dimensional discrete ordinates code DOT. Detailed measurements of the beam
source characteristics were also made so that comparisons of experimental and
calculational results could be compared on an absolute besis. The experiment
was constructed so that it could be represented exactly in cylindrical r-z
geometry and thus preclude any geometric transformation. The use of a
monodirectional beam source provides a stringent test of the code's ability
to calculate the spectra at large angles to the beam.
A final remark must be made about the use of exact methods: these codes
are generally complex enough that errors in input, cross section makeup etc.,
as well as the misuse of codes as black boxes can be very disastrous. Com-
puter costs for running these codes are in many cases excessive and the need
for understanding the code and its results is frequently not considered
worth the cost of running enough calculations to fully evaluate the results.
This is truly unfortunate.
Table 1. Comparison of Calculated and Measured Spectra - Evaluation of Total Cross Sections
Thickness
(cm)
Sample
1
Total Cross Sections
Used in Calculations
Best
Cross Sections
Comments on Comparison of Spectra
(Calculation Based on Best Cross Section)
Be
27.28
053, BMI., ENDF/BC
Good agreement with all three cross sec-
tions; although there are some small
differences, no one cross section is
best
HO
20.32, 30.48
H: OSR; O: UNC,
: 05R, KAPLE
0; KAPL
Good agreement for all energies; primarily
another test of the oxygen cross sec-
tions (see Oxygen sample)
LiH
10.16, 20.32
15.24, 25.4
H: 05R; Li: 05R,
ENDFB
Good agreement for all energies; the Li
cross sections in both OSR and ENDF/B
are from the same source
30.48
05R, ENDF/B
ENDF/B
.
Slight disagreement for energies between
1.6 and 3 MeV; cross section appears to
be too high
Fe
20.32
VEN
KFK and BNL
Good agreement above 2.0 MeV but all cross
sections are too high for lower energies
Ni
15.24
05R, KFK and BNL,8
ENDF/B.
OSR, BNLS and BNL,
FZKI
05R, ENDF/B, BNL
BNLS and BNLC
Good agreement above 2.5 MeV but cross sec-
tion is too high for lower energies
Na
60.96
ENDF/B
Slight disagreement in peaks at 2.7, 3.0,
and 4.5 MeV
Ca
55.56 (chips)
C5R, BNL and ORNE
BNL and ORNL
Sample was not "thick" because of voids;
therefore the comparison is not very sen-
sitive to cross section; approximately
20% disagreement in magnitude for all
energies
·K
76.2
05R, BNL
BNL
Good agreement except for slight differences
between 1 and 2 MeV
Mg
60.96
05R, BNL, ENDF/B
BYL, ENDF/B
Good agreement for energies greater than
4 MeV; for lower energies the cross
sections are too high
Zr
19.526, 39.0
O5R, BNLS
Good agreement with both cross sections

1
S
.
.
T
1
:.
.
.
Table 1 (con'a.)
tori
1
**
Sample
Thickness
(cm)
Total Cross Sections
Used in Calculations
Best
Cross Sections
Comments on Comparison of Spectra
Iculation Based on Best Cross Section)
Concrete
15.24, 30.48
O: KAPL; Others: OSR
Only set tried
Good agreement except for slight difference
near 2.4 MeV, the valley regions of
oxygen and carbon; recuires further
study
.''
W
.
10.63
BNL,” ENDF/B
BNL
Good agreement except for energies less
than 2.0 MeV
.
55.56
Sand
(Si02)
Si: 05R; O: KAPL
Only set tried
. Errut
L
Slight disagreement for energies around
2.4 MeV, the valley region of oxygen
requires further study
-
IX.
2.
238
8.69
OSR, ENDF/B
: 1.22.
Very good agreement for both cross sec-
tions for all energies
L
L
WE
..).
Lead
BNL, BNLand
-
BNL and ORNL
-
ORNI,
-
2. 45, 5.08,
7.62, 10.16,
20.32, 30.48
10.64, 20.32,
30.48
Very good agreement at all energies when
Pbcos cross sections are used for
energies less than 4.3 MeV
05R, ENDF/B
ENDF/B
. -
."
Good agreement for both cross sections,
with ENDF/B slightly better around
8 MeV
i
.iii
..
.....
60.96, 91.44
05R, UNC, ORNL
ORNL
ORNL
Slight disagreement for energies between
3 and 5 MeV
.
A
.
L
60.96, 91.44,
152.4, 182.88
05R, UNC, KAPL
KAPL
.
Good agreement even for 5- and 6-ft
samples
7
I
:
Ni t. Ho
Ni: 10.16;
120: 15.24
Ni: 05R, BNLS and
BNL; H: 05R;
O: KAPL
Ni: BNIS and
BNL; 0:
KAPL
1
Primarily another test of nickel cross sec-
tion; disagreement for energies less than
2.5 MeV
.
.
.
*
.
..
...
.
e irri..
Table I -- References
Center, ORNL (1965)
b. J. R. Stehn. et al., Neutron Cross Sections, vol. I, 2d ed., Suppl. 2, BNL-325 (1964).
C. ENDF/B Tape 102, Cross Section Library, Brookhaven National Laboratory..
d. J. H. Ray, G. Grochowski, and E. S. Troubet zkoy, Neutron Cross Sections of Nitrogen, Oxygen, Aluminum,
Silicon, Iron, Deuterium, and Beryllium, UNC-5139 (1965).
e. E. L. Slaggie and J. T. Reynolds, 0-16 Fast Neutron Cross Sections and Legendre Moments, KAPL-M-6452
(1965).
f.
S. Cierjacks et al., A Novel Method for Very High Resolution Cross Section Measurements, KFK-453, Institut
fur Angewandte Kernphysik, Karlsruhe (1966).
g. M. D. Goldberg et al., Neutron Cross Sections, Vol. IIA, BNL-325, 2d ed. (1966).
h. D. J. Hughes and R . B. Swartz, Neutron Cross Sections, BNL-325, 2d ed. (1958).
i. J. J. Schmitt, Neutron Cross Sections for Fast Reactor Materials, KFK-120, V ol. I, Institute fur
Neutronenphysik und Reaktortechnik (1966).
j. P. H. Stelson, Some Recent Nuclear Cross Section Measurements at ORNL Oct. 1, 1966 - Mar. 31, 1967,
· ORNL TM-1801 (1967).
k.
J. L. Fowler, E. G. Corman, ard E. C. Campbell, p. 474 in Proceedings of the International Conference
on Nuclear Stucture, Kingston, 1960, University of Toronto Press, Canada..
1. J. W. Craven, private communication concerning evaluated cross sections submitted to ENDF/B library.
.
NR.
.
.
2.2
AR
END
2.
..... i
::
:
..:
I
- ERVA
.
.
5
.
ir
.
.
ii.
R
.
RU
.
?
DATE FILMED
9 / 29 / 67



.





"
..
.
WA
www
I
.
1
31
.
..,
TL
2.
u
i
.
+
TE
11
.
.::
sitne
0210