ML20195G578
| ML20195G578 | |
| Person / Time | |
|---|---|
| Issue date: | 08/31/1988 |
| From: | Surmeier J NRC OFFICE OF NUCLEAR MATERIAL SAFETY & SAFEGUARDS (NMSS) |
| To: | Nussbaumer D NRC OFFICE OF GOVERNMENTAL & PUBLIC AFFAIRS (GPA) |
| Shared Package | |
| ML20151C617 | List:
|
| References | |
| FOIA-88-470 NUDOCS 8811280093 | |
| Download: ML20195G578 (1) | |
Text
_ _ _ - _ _ _ _ _ _ _ _ _ - .
8
.,A, a
UNITEJ 8 TATE 8 9 e M[
^
NUCLEAR REGULATORY COMMISSION '
8 5 *
! WASHIM TON, 0. C. 20684 ..
^
...../ -
/
AUG 311933 MEMORANDUM FOR: Don Nussbaumer Assistant Director for State Agreements Program Stete, Local, and Indian Tribe Program FROM: John J. Surmeier, Chief Technical Branch Division of Low Level Waste Management and Decomissioning, hKS$
SUBJECT:
CHEM NUCLEAR SYSTEMS, INC. HDPE HIC INFORMATION Cher. Nuclear Systems, Inc. (CNSI) has supplied tha NRC with additional inforr.ation l regarding (HIC) madethewith300 Highyear structural Density stability (HDPE).of Polyethylene theirStewart Professor high integrity Silling container t
has responded to the additional inforration supplied by CNSI. We have enclosed both of these docurents for your information. If you have any questions regarding these documents, please contact me at x23439.
/
g John J. Surmeier, Chief I
Technical Branch -
Division of Low-Leve? Waste Management !
and Decomissioning, MMSS i i
Enclosures:
As Stated I
b l
hh l2f8 3 081315 i
p nesmaao mo von .
l
[
e HEM NUCLEAR SYSTEMS,1NC.
22o Stoner 49e Drtve
- Colsmtsa. South Carol.na 2321o August 8, 1988 RA-0437-8 t
Dr. Michael Tokar Section Leader. Technical Branch Division of Low-Level Waste Management
& Decomissioning, MMSS U.S. Nuclear Regulatory Cosenission .
MS 5 E-4 11555 Rockville Pike Rockville, Maryland 20852
Dear Dr. Tokar:
Chem-Nuclear Systems, Inc. resubmits our Supplementary Report on HUPE HIC Materials prepared in response to US NRC request of 7/13/88 (Revision 1). In our haste to make the July 28, 1988 submittal, soaa typos were not corrected.
Revision 1 makes editorial corrections only.
We regret any inconvenience this may cause. If you have any questions, please do not hesitate to contact me.
4 Sincerely, CHEMNUCLEARjYSTEMS,INC. ;
jf i -
Michae T. Ryan, Ph.D.$ C.H.P.
Executive Director Regulatory Affairs NTR/nhs Attachment
~ () 0 i
I (803) 25&o45o e Teles: 216S47
G e
$UFPLEMENTARY REPORT ON HDPE H!C MAftR!ALS l .
i PREPARED IN RESPONSE TO US NRC REQUEST l
0F 7/13/03 1
(REVISION 1) e ..
' A O %
l i
CHtH NUCL".n2 $YSTEMS. INC. .
l 220 STONERIDGE ORIVE COLUM8!A, SOUTH CAROLINA 29210 AUGUST 2, 1988 Q Q,r/G lc A 3 c c -
,u pr f ' t 6 p>
.\- -
L
1.0 htroduction & Susnary 1.1 Introduction This supplementary report presents Chem-Nuclear Systems Inc.
(CNSI)'s evaluation of the material properties and ' structural capability of high density polyethylene (HDPE), (cross linked) related to its appilcation for burial of low-level radioactive waste (radwaste). Specifically, the radwaste is placed within HDPE vessels known as high-integrii;y containers (HIC's) at various nuclear reactors and shipped with transport casks to the Barnwell, LLH facility for below ground disposal.
The purpose of this report is to address various concerns related to the suitability of this material for the proposed application. In so doing, the information should establish that HOPE material can be designed to safely accommodate earth burial loads. These concerns were reported by Dr. Silling of Brown University in Ref. 1. The points were addressed by CNS!
at a meeting with the Nuclear Regulatory Comission (NRC) on 7/13/88. At this meeting, CNSI was requested to prepare a written response elabe. rating the various points presented in this meeting.
The remainder of this report covers data and evaluattens -
related to HDPE HI(, performance. It is presented in 'he '
following sections:
Section 2.0 HIC Burial Data and Conditions at Barnwell Section 3.0 HIC Irradiation' Dose Evaluation Section 4.0 HOPE Embrittlement Evaluation Section $.0 HDPE-HIC Buckling Analysis Section 6.0 HDPE-HIC Creep Stress Evaluation
.2-
i 1
1.2 Background Information Cross linked HDPE has been approved for burial of radwaste in HIC's at Barnwell since 1981. This material is also used for other non-nuclear waste burial appilcations, e.g. chemical waste forms. This is due to its excellent propertiet for containment of po+entially, corrosive waste materials.
l To our knowledge. HDPE is only used for radwaste disposal at ;
the Barnwell site. Since burial depth al the other radwaste burial sites are over two times greater than at 8arnwell. the structural loads on the. vessel are also much higner. The NRC/$1111ng questions on HDPE-HIC applic?, tion focus specifically on the capability of this material to withstand ;
the structural loading imposed by burial. This loading is due J
to burial under 20 feet of soll (a 16 psi static load). The
! buria. Ioading requirements structural and duration on the HIC l
are defined in Ref. 2 by SC DHEC. One of the requirements is that the HIC must maintain its structural integrity under these burial conditions for a 300 year period. In this regard, the l HIC's are used as disposal containers for devatered ion-exchange resins discharged from nuclear power plants. ;
These resins are used in the clean-up of reactor coolant water f streams. The resins contain 8 and C types of waste (as defined
(
in 10CFR61) with certain long-lived isotopes. The genesis of . l the 300 years burial design lifetime is based on 10 half Ilves ' l of Cs-137. (Cesium 137) which is a fission product derivative and is a significant isotepic constituent in radwaste. This is
(
3 due to its covaratively long half Ilfe and that it frequently represents 10-20% of the total curie content in 8 and C westes shipped in HIC's.
1.3 sumarv The following points are made by CN5! in the report
. Barnwall Aenlicatten - The HIC application and resulting data relate specifically to its uttitration at Barnwall.
This is not a eeneric situation. The actual Barnwell data regarding the nature of HIC burial, contents, radiation dose, fill level, etc. is available and can be bounded.
. Nic Radutien Done *- A study of liner actual absorbed dose'shows that the integrated 300 year absorbed dose is in the. 10 6R range vs the 108R limit. Hence, radiation dose has a very limited effect on HDPE properties.
HDPE Erbrittlement - Actual 14 year test data from Phillips (the NDPE supo11er) on Harlex CL-100 indicate that the imbrittlement (ductile-to-brittle transition) effect is not nearly as significant as predicted.
hit Wall Butkitne - Analytical modeling of HIC buckling is very difficult due to the non-linear effects of HDPE n under loading conditions and including the real world -
effects of sott-structure interaction and partial filling of the HIC's with a known material. The best vertf tcation of Mckling adequacy is a vessel coepression test simulating actual burial conditier.s. CNS! has successfully performed such tests even under overload conditions.
hic-HDPE creen stress Phenmana -The effect of material creep has been incorporated in HIC ' design. Engineering judgment was used to establish a design stress value.
For low stress levels, Harlex CL-100 exhibits a very 4-
! tsall secondary creep strain rate. The maximum allowable l l stresses were derived by extrapolation of the creep data l
to keep the accumulated creep strain to a small value, j i
) In ramary. it is CN5!'s contention - based on prior testing l l
a .d evaluation of HOPE that this material can 'be safely
{ utilitsd for the actual MIC a;pitcations. Engineering design .
j l stress parameters with realistic conservatism and reasonable l
4 judgement concerning the exte.t nature of the burial appitcation '
can and have been used to design these vesseis. '
f i i
! 2.0 marnwall conditions . l j An evaluation of a material application si.ould realistically l l consider the actual service conditions. Design guidelines l
! should be sufficiently conservattve to provide safety, but !
l should not unduly penalize a material by placing restrirstons !
j and reflecting conditions that never exist in actual practice. l To place some degree of reality to the NDPE material situation, !
j this section presents specific data ,on its use as a radwaste i container at Barnwell, f 1
i The following points relate specifically to the burial !
conditions of HOPE HIC's at Barnwell: [
l . !
^
(1) Barnwe11 oniv - At present Barnwell is the only radweste ' l site which burles HDPE HIC's.
l
] (2) Barnwall Aurtal Nethods - HIC's are buried in a 20 foot {
deep trench. They are covered by sand, clay, and vegative soll. The sand is sequentia11y compar.ted to (
I minimize the po:sibility of earth volds or arches. The l earth cap is inspected monthly for damage (e.g. evidences 1
of subsidence, erosion, etc.). ;
I
[
f -
I i
f
. I I
i
(3) HIC ouantitiet/shes - About 500 H!C's are used
, annually. A total of 627 HIC's were buried in Barnwell in 1907. A typical (HIC is nominally 74" $ X 78" High with 1/2 - 3/4 inch thick HDPE walls.
(4) !!1C Fill Rate - Barnwell site receipt criterta' rewire a minimum of 85'1 filling of the vessel with radwaste.
Economic cenditions make higher fill volumes desirable.
Fill mater!als include standard bend resins Eceder and Powdex resins. Hence, some internal support of the -
vessel under burial conditions is provided by the fill medium. The fill media can be physically c5aracterized.
(5) HIC rabrication Precast - CNSI HIC's are fabricated using Marlex CL-100 resin supplied by Philltpy. The atterial is cross-linked. The fabrication process is rotational molding and CNS! HIC's are supplied by only two fabricators. Hence, reasonable concistency in fabrication technique and "as-built" properties can be i expected and are vertfled by rigid quality control requirements implemented by CNSI.
[
(6) Radioactive Content -
Extent.tve data exists on the i radioactive materlat content in all bu.lat containers. A l survey us made of all shipser.ts to Barnwell in 1987 and - ;
special eraatnation made of 411 HIC shipments made in '
1987. Table 1 presents data on the entire 19871sotopic inventory. The isotopic content is representative of '
prior year expertence, Figure 1 shows the distributton of all 1987 HIC dose rates. In this regard, special examination was also made of the cases with highest <
contact dose (87 cases or 15% of total) to evaluate worst l case situations. These suonaries show that the absorbe6 dose of radioactivity can be established for HICs within !
certain upper limits.
{
l l
/ 00 '.
/ .
0 0
/ 0 V
1
> s -
l e n L v e
L 0 0 E l E 0 /
E W / .
0 e m e s
N V 1 M
d o
R
- A S t -
At i 0 E c
a Bm .
. 0 0
0 0
9 t
n OSEO .
5 T
o c
TD F T r
^
T e
=- N
- SCTA 0 O C
n Ts s 0 1 i
l No C ,
0 3
2 R
t u e C Et r I Muu b u g M
~
J E -
Pt E / r i F
f o
- I P 0 s -
HO -
, 0 0 I n
S HM
- t
. 0 a o -
1 i U
- M 1 t . -
EM X b
u Pm A M i Du 0 t r
s H u 0 0
6 i
- p. d p
7 y -
8 c ,
n 9 / e n 1 0 u 0 q m 0 3 r e
F
- - ~ - - -
C 0 0 0 0 0 0 0 5 0 0 5 0 4 3 3 # 2 1 1 r1M b s .
_ il ! i'
l TABLE 1 1987 Sarnwell Disposal Inventory 13 i l
3
- IS0 TOPE Cix10 Tyg (YR$) Casma Energy Fractlen l (MtV) 1 .
l
-,l 1. Fe-55. 44.0 2.6
- 0.23 .404 l
- 2. Co-60 68.9 5.3 1.17,1.33 .327 l
- 3. Mn-54 10.0 0.8 0.84 .047 [
- 4. Wi-63 9.1 100.1 Seta emitter .C13 !
- 5. Co-57 7.5 0.8 0.12 .036
- 6. Gd-153 7.1 0.7* 0.10 .034
- 7. Cs-137 5.0 30.1 0.67 .024 !,
l 8. Y-)) 5.0 02 1.21 .024 j
i
- 9. Cr-51 3.7 O.1 0.32 .018 l'
{ 10. Cs-134 3.1 2.1 0.80 All I
Cumulative .976 OTHER 5!GNIFICANT !$0'#t$
Kr-85 10.7 0 52 .0003 '
I
- K-40 1.26X10 1.46 (.0001 !
Na-22 2.6 1.28 c.000) f EU-152 12.0 1.41 c.0001 l
1 L
I k*. S The total 1987 Sarnwell isotopic inventory was 711,000 C1 's !
l 4
4 I.
4 8
3.0 HIC frradiatioq_ Dita 3.1 ceneral The Ref. I report is concerned with the degradation of the HOPT material due to irradiation of tht: material by the' contsnts.
The assumption was made that all vessels are irradiated to a highly conservative design basis limit. The CNSI Cata indicated that the ti radiation dose is much less. A methodology was developed for calculating the irradiation dose to HIC vessels based on actual loading conditions. The 1987 Barnwell data base (Ref. 3) of 627 HIC vessels was used. The guideline for the limiting absorbed dose is 108R for the 300 year burial period (Ref. 2). This radiation limit was estabitshed to preclude both physical degradation of the vessel polymer t'y irradiation and to mintalze the uneration of internal pressure in the vessel due to radiolysis of the resin material.
The key variables for Ilmiting high irradiation doses to the HDPE wall prior to vessel filling are:
(1) Hieh curie centent (or contact dose val s). This leads to high vessel contact dose rates, and hence a ht;her value of absorbed dose.
s (2) Hieh cs-137 fieteele centent. This isotope is coeparatively long lived and contributes significantly to the vessel absorbed dose over the entire 300 year period.
(3) Hieh to-Er itetente content. This isotope is the highest gama ener;y ct;ntributor, and is the dominant source leading to a high contar.t dose value.
In certain situations,12th a high dose rate and high Cs-137 content are present in the waste. In these cases, CNS! has required the waste generator to perform an assessment of HbPE dose rate and enforced these Ilmits prior to shipment 9-m
(Ref. 4). In these cases, the prelir.inary predictions of
, absorbed doses were more severe than the actual dose results calculated for the loaded HIC vessel.
3.2 Results 6
Figure 2 shows conservatively caiculated upper limits of absorbed dose for 95% of the HIC's shipments. Two conclusions can he drawn from these curves.
- A 300 yest integrated absorbed dose of < 2X10 0R is much less than the IX108 R limit for over 95% of the HIC liners being disposed of. An average HIC liner has a 300 year integrated dose in the 10 5R range.
- The integrated dose absorbed by the HIC typically reaches an asymptotic limit after 40-50 years or less. The dose buildup rate is "orders of magnitude" less than for experimental tests of HDPE sampics which are irradiated in 103 -104 R fleids.
An additionti examination was made to find a worst case. The isotopic breakdow. of 87 HIC shipments with contact values of 25R or greater were selected. The percentages of Cs-137 and Co-60 were calculated. The results of the survey are noted l below:
1 Contact Reading: 25-400 R Cs-137 Content: 0-83%; Average 15%
Co-60 Content: 1-65%; Average 22%
Co-60 + Cs-137 Content: 0-641; Average 37%
Several of the significant cases were evaluated to estimate the 300 year integrated dost. Even in these cases the value was isss than 10 7R.
1, Total absorbed dose rate over 300 years 4
C O O 3.5 /
3 g 2.5 / .-
~
.- / I
- j oE 2 t
- s av
_g 1.5 ,-
a l y . : = = =
1-- ,
0.5 0 .
O 40 80 120 160
, 200 240 280 O oil CS-137 Time 'n veors
+ ottCO-60 0 22%Co-60 15%Cs-137 ,
Figure 2 95 percentile value of 87-HIC liner sbsorbed dose
303 Methodoloav (1) The 627 HDPE shipments were categorized by contact dose as shown on Figuro 1. The spectrum of the vessel maximum contact doses were evaluated statistically:
Median Value ,
2.5R Average Value ISR 95% Percentile Value 60R (2) The Barnwell isotopic values were complied. Eight isotopes cos, prise over 971 of the Curie content and over 99% of the effective energy (See Table 1). Cobalt-60 is clearly the vajor contributor to dose rate.
(3) The 300 year integratr i dose for any isotope was calculated by:
-At D
300
- Dg ( l-e g j EQ.1 Where D 300
- 300 year integrated dose, R D
A
- HDPE liner wall absorbed dose, R/Hr
, t - 300 years x 8760 hr/yr h -
g ; Tg - isotope half life, years The absorbed dose rate was estimated assuming a Co-60
'~ '
source in a typical (14-195) liner using the Microshield computer program. (Ref. 4)
Dg -
Dj - D, EQ.2 Where Do - outer wall contact dose rate - 60R i
(35 percentile value)
Dt - inner wall dose rate - 70R (calculateu using Microshield) i 0 - 10R 4
e
i For isotopes where the half life is 5 years or less: ;
l e -At =0; D300"E/g A EQ.3 l
for multi,^1e isotope conditions:
r .x,e
- j ca.4 Daoo~Ds[7,Igtf,E,x(1-e yt Examining the 87 worst case situations and assuming that only Cobalt (22 percent) and Cesium (15 percent) are present is shown below:
t, r, t,r, e, Cs.137 0.15 0.67 0.101 .gg3/(gygD*30.3) - 2.4 x 10 s Co.60 0.22 1.13+1.37 o.530 .693/(876045.3) - 1.4e ?.10 s
~
(.55o + .101 ) 2.6 o*' , ,(.55o + . l ol) l .49 x lo**
Several of the most significant cases were specifically examined from the survey of 87 shipments using the actual isotopic values and equation 4. Even in these cases,.The highest in:2 grated dose values were all less than Ix10 7R. Hence, it is our conclusion that the effects of radiation on HDPE will be relatively small.
This situation aill continue into the future based on these additional factors:
- 1) Radwaste casks used to ship HIC liners have >
insufficient shielding to handle liners with greater than 300-400R contact dose rates..
- 2) The occurrence of waste with more than 201 Cs-137 is very rare based on the extensive .
CNSI-Barnwell data base.
, i l
! =- - ,
l >
l e
9 9
t'
4.0 HDPE Imbrit' clement The Ref. I report states that embrittlement of HDPE greatly reduces the structural capability of the material. It further states that this could occur in 30-50 years which is much less than the 300 year design life of the vessel. -
The ductile-to-brittle transitions in polymers, cited by Dr.
Silling's report is a phenomena observed mostly in the low molecular-weight HOPE at high temperatures (Ref. 6). The physical mechanisms causing this transition, as Dr. Silling points out, are not well-understood. The main source for the inference that Harlex CL-100 will exhibit a ductile-to-brittle transition within the life-time of the HICs (300 years), even for low sustained stresses, is based upon the curves of Figure 1 of Ref. 1. These are referred hereinafter to as the "Graube curves". -
CNSI believes that these Graube curves do not represent the material properties of Harlex CL-100 and the state of stress on the HICs accurately. Thus, they should not be used as the foundation for judging the adequacy of Harlex CL-100 material.
He have made our judgment based upon the following reasoning.
(1) fristina Phillies Test Data Ph1111ps Petroleum Company, the manufacturers of Harlex CL-100, has performd several long term creep ruptere tests, similar to the Graube tests, using 2" pipes both at room temperature and at 140' F. These continuing tests have been underway fit 14 years with no indication of failure. Ref. I describes the results of some of these tests. If the data from these tests are Hotted on the Graube curves (See Fig.3), it is clear that Harlex CL-100 exhibits markedly improved properties beyond those of the Graube curves.
4
GRAUBE CURVES REPLOTTED FOR 20'C AND 60'C .
plI -
A m bie n t .
~
4 t e rn p.
C L-- ] O m
n v
u..
.. N* w '
T s. ,
40'C E A g .
2 T
\ \
to ' o'
\
, s eo* e' g, ne l Time (h)
- - - represents a probable creep-rupture behavior
@ Phi:.ips Petroleum test data at 20'C (under progress) d Phillips Petroleum test data at 60*C (under progress)
Tensile rupture stress for Marlex CL-100 Figure 3 Stress to Ruoture Data of Harlex CL-100 S
Extrapolation of these data shows that the ,
, ductile-to-brittle transition in the material does not occur at the times indicated by the Graube curves. This is not surprising since cross-linked HDPE was developed to be suitable for high temperature applications.
Bearing in mind that the cross-linked Marlex CL ,100 is an ultra high molecular, weight HDPE, it may be argued that this transition is unlikely to occur in the material under a relatively low stata of stress at the 15'-20'C (trench temperature) within 300 years.
(2) GraubeDataIsNotAceiicable The Graube curves are based upon the results of an internal pressure loading on circular pipe with plugged ends. This gives rise to uniform circumferential and axial stresses which are both tensile in nature. The burial loading on the HICs results in comeressive stresses in the sidewall and tensile membrane stresses in the top-head. Also, the creep-rupture behavior of the HIC sidewall in compression is Intuitively different from '
the tensile-creep data of the Graube curves which is ban,d on uniform loadings. The magnitude of the HIC sidewall stress is approximately 600 psi. Even considering tensile creep, this is a small number and can be sustained by Harlex CL-100 for longer than 300 years. s P2PE HIC TWH' l
- G*t'"{0 ,
i
?
h
- . , - - - - ~ . . - - - - - , - . -. , , - - , , - , - - , . . , , - - - - - - - - - - . - - - - , -
Initially, at the time of burial, the tensile cembrane stresses in the top-head of the HIC are approx 1.nately 1,000 psi. After a period of time, the top-head, under the burial loading conditions deforms downwards - towards the contents where the total deformation is limited by the depth of the void space between the content: and the top-head. The effect of creep is to increase this deformation. Hence, af ter this time period, the stresses in the top-head will be reduced when the top-head makes contact with the contents since most of the load is transferred to compressing the fill media.
l .
f i .
e t
5.0 liDPE Bucklino Evaluation 5.1 General The Ref. I report contends that the vessel will buckle under the burial load conditions. In this regard, previous CNSI tests and calculations refute this concern. The primary concern is the degree of conservatism in the analytical modeling and the relevancy of prior tests.
The buckling of the HICs under burial load conditions is an undesirable phenomenon. Buckling is a term which refers to the phenomenon which occurs when the application of a load beyond a critical limit results in an unstable vessel geometry. For HIC's, the onset of buckling can be specifically defined ii the deformed shape of the HIC which is no longer capable of supporting an additional load over it. CNSI agrees that ouckling must be precluded in the HIC design. The point of contention is the conservatism in an analytical model. The theoretical evaluation of buckling stress of HICs is very ,
complicated due to the nonlinear properties of the material. '
The problem becomes even more complicated if soll -
HIC interaction and the time variable creep behavior of Harlex
, CL-100 HDPE is to be analyzed.
~
5.2 '
Problems with Ref. 1 Analytical Hodeling Approach In Ref. 5, the approach for evaluation of ireep buckling stress of the HICs employed closed-form / experimental results for idealized-shaped metal, not plastic, in the equations. CNSI feels that these simplified models are not realistic and hence, overly conservative. CNSI questions the following specific areas in the analytical model:
I t . _ - - _ --
(1) Hetal vs. Plastic Bucklina Modes The buckling stress formulas available in the literature have been complied using the test results of metal shells. They incorporate, among other things, the mechanical imperfections and scatter of test results.
Metal shells with dimensions comparable with typical HICs would buckle in the characteristic shape of a moderately long cylinder. The HDPE HICs, on the other hand, show a characteristic local buckling shape observed in actual HIC vessel compression tests. These buckling shapes are different from metal cylinders. These tests have bean performed under very high loads (several times higher than the postulated overload) and performed by independent test laboratories on behalf of the CNSI. The
! differencer between theoretical and experimental vessel buckling shape shows the limitations of the closed-form / experimental test formulas' and their applicability to the HOPE HICs.
(2) Inacolicable Use of Secant Modulus The use of the secant modulus in the formulas of critical stress for creep-buckling was derived by Chern .or the instability of a tube in a tubular heat exchanger which was operating at high temperatures. Its application to ,
the evaluation of the critical buckling stress for HDPE l
HICs is not considered valid and is questioned by the CNSI on the following grounds.
(a) The Ref. I calculations method employs an equation i of state approach, i.e. it is assumed that the response of the material depends explicitly on the l present state. This approach has even been
, questioned, for metals. It has been argued that the 1
only valid representati0n of creep is one that incorporates the memory of past events.
This approach is suspect when used. for viscotiastic materials, whose response is highly path dependent ;
under the large strains sustained.
(b) The Ref. I calcu14tional method embodies the
, assumption that the dependence on the material l
properties and geometric properties can t,e separated. This assumption is not valid for situations with large deflections where geometric changes give rise to stress stiffening.
(c) The Ref. I calculational method employs a critical effective strain approach where there is no stress !
redistribution. Ref. 8 states that there is a redistribution of stress from the initial state to a final state whenever the load is constant and the problem is statically indeterminate. Therefore, the HDPE HIC's under the burial loading do redistribute ;
stresses, violating this assumption.
(d) The calculational method in Ref. I assumes that r
"~ compression and tensile properties for HDPE are s
identical. Ref. 8 states that this is not always the case and one should be careful to verify it for the material under consideration. The compression croisery serem vi einfa '
j -
~~~,,,, ---
rn,. iu.a .i mal. /j# ~ ~crea.rrema
~ inia N ;
eg __ L : Ntas mieht i
ymg .i,43
,,r. . f Strole i
i
- 2) -
1
crecp curve is the mirror image of the tension creep
. curve. This certainly is the case for HDPE, where J l
compression and tens 11 properties are expected to i be different since the material has a high Poisson's ratio (0.45). The expected difference is due to cold flow of the material, since the cross section greatly increases under compressive loading. The l creep behavior in compression is clearly markedly l different in compression than in tension. I t
t (3) Effect of soil Structure Interaction Must be Considered l The experimentally obtained critical buckling stress hith l
- a constant pressure loading for a relatively flexible structure (e.g. a horizontally buried cylindrical shell, hemispherical-head caps, etc.) is conservative by -
factor of three to four. This is compared to the case [
) when soil-structure interactions are taken into account I (see Bulson, Ref. 9,. pages 160-167). The Brookhaven !
l National Laboratory Report (BNL) (Ref. 10) has l
{
conservatively derived a formula to account for the soll i t
structure interaction effact on the HDPE HIC buckling. l
- When values of the parameters corresponding to the Barnwell so11 conditions are substituted into the BNL formula, a ratio of approximately 3 is obtained between i" the critical buckling stress (when comparing the effect '
, of the soil-structure interaction to a case where there l 1s no soil structure interaction). !
i (4) HICs Are Leaded with Haterlat j 1
! Per the Barnwell site license, the liners must be a minimum of B5% filled. There are only three waste !
! material resin forms that are used to ft11 the HICs which l are (a) bead resin, (b) Powdex resin and (c) Ecodex. The ;
i 451 volume limit ensures that the entire cylindrical l l t
! t i
L-__.---,,._ _-.,._ _ _ _ _ .,__ _ _ _._ _. _ .__.,
~.
portion of the HICs along with a portion of the torospherical top will be filled with the waste form.
The cylindrical portion of the MIC vessel wall will be backed by the waste-form from the inside. However, the torospherical head will only be partially backed by the waste and may be subject to a snap-back deformation.
This snap-back deformatian occurs when a part of the convex "up" shape changes to a concave "up" shape due to soll loading conditions.
l l
. n.p.
I i This configuration will cause a tensile loading in the head. However, this is not considered to be an unstable configuration (per the buckling definition). In this configuration, the concave up (or deformed) position of the HIC will behave Itke a shallow dished-head. Such a shape is known to have small rigidity against transverse load. 7ne soll volume directly above it will have an "active ' soll arching. Therefore, 'the soll pressure on this pesition of the head will be reduced.
The Ref. I treatment of the soil "arching" phenomena is not accurate. In many cases, soil arching actually '
Teduces the earth loading on the HIC top-head. McNulty examined the effect of soil arching (Ref. 11), using a circular trap door mounted flush with the bottom of a circular soll container to evaluate the effect of door flexibility on the soil arching ratio. His results are reproduced in Figure 4. It can be seen that for active soll arching 1.e. trapdoor deflection downwards, the soll pressure is reduced to an extremely small value. Ref. 9 describe; Terraght's experiments with a long rectangular trap door and comes to a similar conclusion. (See Figure
- 5) Although these experiments are not exact duplicates I
.. i of the HIC loading conditions, they are still relevant.
These experiments clearly indicate that for a flex 4 hic horizontal member buried under a soil column, the pressure transverse would typically be an order of magnitude smaller than the pressure corresponding to the total weight of the soll column above it.
5.3 Possitie bthods of Evaluatina Bucklina Potential CNSI's reason for pointing out the variables involved and the limitations of the available theoretical / experimental data is ,
to highlight the inability of the state-of-the-art in analysis to simply model the behavior of the HDPE HICs. Simplistic analytte models with overly conservative boundary conditions will certainly lead to high vessel stresses and conclusions which we feel are overly conservative which don't reflect the actual pl.;sical situation. The modeling situation is very complicated when consideration is given that the HIC's are filled with waste and surrounded by compacted soll. The available method ', hat will best predict this behavior is a detailed nonlinear finite element analysis using a combined model of the HIC, the surrounding soll and the waste. However, this method would require characterizing and bounding the waste properties, as pointed out in Ref.1. This characterization is possible, the mechanical properties of the waste that are ,
required to perform this type of analysis are the elastic modulus E, and the Poisson's ratio, v. These properties can be established by testing the three known ~ waste forms, described earlier, by suitable methods. Confined modulus, i E,g and bulk modulos E , can be measured and the test data b
can be used to obtain the elastic modulus and the Poisson's ratto using the following two equations, j = ( ' " ")
E' -
! (1 + v)(1 - 2v) '
En 1-2v) l l
-u.
p
. IIIIIII*
T' N
J, 8U$ P* ,(v,s umMvsws HBA -
Arching rotle, p /p,
\ % ,2 ~ '
Volves of #/8 N -
30 N
\
- .. . . . * * * . ..(,g
, , , N - 25 -
. . , .. .,, N .
I % * ****..,\ -
20 *
~.;.s:y \
[ (- 15 Possive - Active .
10 Voluesof #/8 Crs-A
, , , , , 2 s : ...... , - - -. .
5 -4 3 -2 1 O' 1 2 3 4 (J/8)x1000 Figure 2.4 McNulty's c:,wriments.
- . g Figure 4 Ef f ect of Trandoor Flexibility on Soil Archina
_(Reproduced from Ref. 9) 1 1
Surface--
4 -
Compacled sond,
. 4=44*
Distribution of 3 . . verticol pressuce. .
' z/b. .
25 2 -
1 IJo arching 1 -
\ Arching 0 l' ' ' ' ' ' ' ' '
10
. Trop door, 02 04 Q.6 08 r /r; -
v vh -
A- Figure 2.1 Terzaghi's trapdoorexperiments(note: 2 m'essured upwards from trapdoor).
b = trap-door width H = height of the over burden
.'Ia, = soll stress at z rr,. - sell stress with no arching (hydrostatic stress)
Ficure 5 Terrschi's Trapdoor Experiment ResultJ: .
,[jLtprodu ced f rom R e f . 9 ) -
However the finite element model using this approach will Ltill be debatable due to the tolerances in the assumptions and boundary conditions involved. Therefore, CNSI feels that testing under simulated Barnwell conditions r, ens to present the best means of verifying the buckling behavior.
CNSI has tested prototype HICs under the simulated Barnwell burial conditions (see Ref.12). The F'Cs were filled and the t
compression loading was applisd through a soil-like medium.
This test was continued over a period of more than 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> and the deformation was monitored. No buckling failure was observed in the container. Admittedly, this t6st shows the short-term behavior of the HIC. It does not completely evaluate the effect of creep buckling. However, BNL, using a soil-structure interaction model including the viscoelastic properties of the HDPE (Ref.10) has shown that the initial i buckling load is smaller than the creep buckling load. Thie means that if the buckling of the HIC does not occur during 'che simulated testing time, it will not occur due to the material creep. The creep behavior of the HDPE could be incorporated '
into a simulated buckling test. The duration of the testing !
would be increased to a time over which the decrease in the creep modulus of the HOPE takes place for the anticipated amount of stress. Most polymers show the onset of the secondary-creep at about 500 hours0.00579 days <br />0.139 hours <br />8.267196e-4 weeks <br />1.9025e-4 months <br /> (see for example Ref.6, page s j 410) after which the change in creep modulus is rather !
insignificant. Therefore, compression ' testing of the HICs to this period, i.e., 500 hours0.00579 days <br />0.139 hours <br />8.267196e-4 weeks <br />1.9025e-4 months <br /> (21 days) will be sufficient to 4
test for creep buckling. In such a test the period of performance and buckling acceptance criterion must be agreed l
l upon prior to testing.
i i
i j
l i
[
t
~
6.0 HDPE-HIC Creen Stress Evaluation In Ref. 1, the contention is made that thc analysis did not consider r.reep phenomenon. The cilegation by Dr. Silling that the industry ignored material creep is unfounded. The creep-rupture behavior of the harlex CL-100 material has been incornerated into the design of the HICs by establis'hing the creep-rupture stress for the 300-year life of the HICs.
The CNSI design approach has been to limit the r,*ximum stress in the HIC as selected design stress value based on examination of creep stress curves in Ref. (14). Engineering judgement has been used fn establishing the stress values and did consider creep. This approach iJ not uncommon. For example, the ASME Code Section III allows a linear analysis to incorporate effects of nonlinearities (such as plasticity) in design by analysis of nuclest power plant components.
The use of tensile creep-rupture data to limit the compressive sidewall stress is conservative. The sidewall compressive stress is approximately 600 psi. Under compression loading, the HIC sidewall cross-section increases, and the material does not exhibit the same creep behavior as 't does in tension. The top-head loading is both compressive and tens 11e. The top-head of the HIC underloading undergoes a snap-back deformation to a S - concave-down configuration. In this configuration, the loading ,_
on the top-head actually goes down due to active soil arching of the column of soll above it. (See Section 5.0 for more details of this phenomenon.)
Therefore, the stresses calculated in the top-head, without considering the effect of the soil-arching, is conservative.
CNS!'s evaluation of the stresses in the top-head region showed a maximum average stress to be less than 1.100 psi (See Ref.13).
i
I .
6.0 HDPE-HIC Creen Stress Evaluation In Ref. 1, the contention is made that the analysis did not consider creep phenomenon. The allegation by Dr. Silling that the industry ignored material creep is unfounded. The creep-rupture behavior of the Harlex CL-100 material has been incornorated into the design of the HICs by establis'hing the creep-rupture stress for the 300-year life of the HICs.
The CNSI design approach has been to limit the maximum stress in the HIC as selected design stress value based on examination of creep stress curves in Ref. (14). Engineering judgement has been used fn establishing the stress values and did consider creep. This approach is not uncommon. For example, the ASME Code Section III allows a linear analysis to incorporate effects of nonlineartties (such as plasticity) in design by analysis of nuclear power plant components.
The use of tensile creep-repture data to limit the compressive sidewall stress is conservative. The sidewai compressive stress is approximately 600 psi. Under compression loading, the HIC sidewall cross-section increases, and the material does not exhibit the same creep behavior as it does in tension. The top-head loading is both compressive and tenstle. The top-head of the HIC underloading undergots a snap-back deformation to a a- contave-down configuration. In this configuration, the loading s on the top-head actually goes down due to active soll arching of the column of soll above it. (See Section 5.0 for more
' details of this phenomenon.)
Therefore, the stresses calculated in the top-head, without considering the effect of the soil-arching, is conse.vative.
CNSI's evaluation of the stresses in the top-head region showed a maximum average stress to be Ir.ss than 1.100 psi (See Ref.13).
The basis of establishing the threshold of acceptable stress in
- the HICs was the consideration that the secondary creep rate for this stress level ought to be small. The creep strain behavior of Marlex CL-100 is different from other HDPE homopolymers and copolymers, as noted by Dr. Silling (see Table 2 of Ref.5). The primary creep rate of Marlex CL-100.is higher than the other types of HDPE, but the onsot of tertiary creep in CL-100 is a; h delayed (see Ref.14). The creep strain data of Ref.14 shows that the strain rate is essentially zero for a stress level of 500 psi. It is still small up to a stress level of 1400 psi. Extrapolation of this data to a 300-year container burial lifetime. results in a small accumulated creep strain. Hence, CNSI established a stress level of 1400 psi.
BNL, using the results of tests on Martex CL-100, has established a level of 1.100 psi (see Ref.10), which is not significantly different from the CNSI established limits.
Irrespective of these differences, both parties agree that a limit on design stress can be established.
The question of limiting the exact amount of creep deflection is immaterial when considering the actual HIC burial conditions. As described earlier, the HICs sidewalls have compressive stresses of about 600 psi. Stresses of this magnitude would not result in the sidewall compressive strain larger than 3 to 6% over the 300 years design life. The top-head deforms downwards .- towards the contents - under the '
burial loading conditions. Thus, the total deflection in the most pessimistic case is limited to the depth of the void space between the content and the top-head. Since the containers are filled to a minimum of 85% volume, this deflection is still reasonably small.
Rafarances
. (1) "Review of the Structural Design of High Integrity Containers" by S.A.
S1111ng. Report submitted to NRC, May 10, 1988.
(2) "Guide for High Integrity Container Topical Report Application" South Carolina Department of Health and Environmental Control, July,1985.
(3) CNSI Data Bank of 1988 Shipments to Barnwell, G.C. -
(4) GPU Nuclear Corporation Connunication to CNSI, No. 4230-88-078-HTC, Dated March 14, 1988.
(5) Hieroshield Version 3.1, User's Manual. Grove Engineering, Inc.,
April 4, 1988.
(6) Modern Plastics Encyclopedia, 1985-1986, a McGraw Hill Publication.
(7) Personal Communication from R.'L. Rees (Phillips 66 Company) to H. Baig (CNSI), July 12,q988.
(8) Harry Krams, Creen Analysis, a John Hiley Publication, 1980.
(9) P.S. Bulson, Buried Structures, a Chapman and Hall Publicatien, 1985.
(10) "Review of the High Integrity Cask Structural Evaluation Program (HICSEP), BNL Technical Report, April 6,1987.
(11) McNulty, J.H., "An Experimental Study of Arching of Sand". U.S. Haterway Experimentation Station, Technical Report. 1-674, 1965.
(12) "Supplement to the EnviroSAFE Polyethylene High Integrity Container Evaluation Report", Chem-Nuclear Systems, Inc. Report, September 30, 1982.
(13) "Evaluation of Stress Loadings of CNSI HDPE HICs", Report Submitted by CNSI to NRC, January 29, 1988.
(14) Phillips Chemical Company Technical Memorandum on Harlex Resins 7 TSH-243, July 1975.
l
&C s l
\
August 19, 1988 Mr. Derek A. Widmsyer, Chil Engmeer DMalon of Low Level Waste Management and Decommissioning Mail Stop 564 U.S. Nuclear Regu! story Co==l=%
Washington, DC 20555
Dear Mr. Widmayer:
This letter contains ray comments on the document "Supplementary Report on HDPE HIC Materials" (Revision 1) by Chem N6 clear Systems, Inc. (August 2,1988). 'The document refened to below as the "Structural Review" is my report "Resiew of the Stsuctural Designs of Polyethylene High Integrity Containers" (June 10, 1988).
Before proceeding with a detaGed discussion, I would like to summarize my comments. ,
CNSI's dose calculation appears to be incorrect (too low by a factor of about 1).
CNSI now agrees with one of my central points on the buckling issue, whkh is that the waste itself vpports the load, not the HIC. There is nothing wrong in principle with using the waste as a snuctural element. However, there must be adequate characterization of the waste and adenuate quality control procedures.
- While the new Phill!ps 14 year test data show improvement of the properties of unitradiated Marlex CL 100 over those of the older HDPE varieties, the rursivability of CL 100 over 300 years is still unproven, especially when irradiated.
The items below are numbered according to the wetions in 'he CNSI report.
3.0 HIC frradiation data. CNSI argues on the basis of measurements and analysis that the maximum expected dose to a HIC wGI be less than 10 7 rad, compared with the 108 rad'value presented in the Technkal Pontion.
There appears to be an error in CNSI's dose calculation. On page 12 of CNSI's report, the
- dose rate to the HIC is computed as the difference between the inner and outer dose rates. (The l
inner dose rate is found from a computer program, and the outet is found by measurement.) This procedure wems incorrect becauw gamma radiation dose (measured in rads) is defined as energy deposited per unit mass, it is not the total energy deposited in the container. The way dose is computed is from the exposure, which is measured in toentgens. A given material, when expowd to I roentgen, will absorb a certain dose, whkh is measured in rads. For water and many organic materials, the conversion factor is nearly 1, so that 1 roentgen crposure results in 1 rad dose.
Therefore exposure itself is frequently measured in rads (as in the NRC Technical Position).
Assuming that CNSI's value of 70 R at the inner wall is intended to mean an exposure dose rate of 70 roentgen /hr, this translates to an absorbed dose rate of about 70 ra4hr, not 10 as stated.
Thus CNSl's estimates for total 300 year absorbed dose should be multiplied by 7.
D P' Sk y I
ou r y)
e Mr. Derek A. Widmsyer 2 August 19,193g If all the radiation were due to Cs137, with a halflife of 30 years and a decay constant of 1 = 2.64x104 hr 4, then the total dose would be about 70 radhit 2.65x107 rad 2.64x10 6 hr*l This would be reduced if the faster decaying nuclides were accounted for. However,8consihring possible statistkal Suctuations arisong the containers and the need for conservatism, the 10 cui 8gure is not unreasc,nable.
8 Experiments at BNL have shown that at low dose @ even 10 rad signiScantly ahects the 7
elongation at break of the material. At 10 rad the material is very brittle. (See Figure 3.4 of [1).)
1herefore CNSI's argument is not comincing. .
4.0 HDPE embrittlement. L . argues that the long term resistance of unitradiated Marie CL 100 to cracking is superior to that of linear HDPE, and therefore the Graube data is inapplicable. To support this, CNSI cites recendy obtained data from the Phillips Chemkal Company on long term tests of CL 100 pipe.
On page 15 of the CNSI report, the following Watement appears, referring to the new Phillips data: "These continuing tests have been underway for 14 years with no indication of failure."
However, Bryan Roy of Hittman Nuclear, in his letter to NRC dated August 4,1988 presents a different view of the same data: "Specimens at 1586 psi and 1515 psi (hoop Wress) have failed (at approximately 70,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br />). One brittle failure and two ductile rips have occurred." Since the CNSI report is dated August 2, while the Hittman letter is dated August 4, these failures must have occurred on August 3.
Peter Soo [2] has observed failures of unitradiated Marlex CL 100 after about 1 year at stresses in the range 1200 1500 psi in air. Earlier failures were Sund ier Wre chemkal emitonments. Dr.
Soo believes that there may be a threshold stress for long term failure of Marlex CL 100. However, a reliable value for this quantity has not been found. In very long time priods surely other mechanisms of degradation will come into play. Among these is aging [3), whki. is probably not reduced by cross linking.
Incidentally, the dr.ta point for 14 year at 20*C on page 16 of the CNSI report is plotted incorrectly. It should be about 5 mm to the kft.
CNSI also argues that the loading conditions in pipe tests are different from those expected in HICat therefore the pipe test data is inapptkable. This is hardly an argument in favor of thJs material for HICs. Since the long term data are sparse, we must consider the pipe tests as valid at least qualitatively.
CNSI further argues that creep mil relieve the shell of stresses, liowever, excessht strains may take place duiing the collapse of the dome, leading to microstructural damage.
In any case, it is beyond question that inadiated Marlex CL 100 undergoss estenshe cracking, as Dr. Soo and others have shown repeatedly. Thus all the speculation about unitradiated material is perhaps academk.
l Mr. Derek A. Widmayer 3- August 19, 1988 In mammary, it is true that urdnadiated cross-linked polyethylene possesas better resir.ance to slow crack growth than linear polyethylene in 14 year tests. However, thew tests fall far hort of prodng that the material can last 300 years, c pecially when radiation is present.
5.0 Bucklina. CNSI presents a number of arguments aimed at showing that buckling will '.ot occur.
However, CNSI concedes a major point, whkh is that the dome collapses to the point where the load is being supported by the waste. Bus, from the point of view of structural mecharucs, the HIC performs a minor role. It may have some value as a chemical barrier, but it is not supporting the load. Given that the dome collapses, I sec little point in arguing about buckling in the sidewallt There is 'sothing wrong in principle with considering the waste as a atmetural element. However, '
CNSI has not done the secessary characterization of this material, nor have they included this effect in their design.
The foliowing items refer to CNSI's speciSc comments on euckling in Section 5.2.
(1) CNSI claims that buckling in metal c>iinders is somehow fundament Jy different from that in polymer c)1inders. There is no basis for this cjalm. Anyway, there is nothing in the analysis in the Structural Review whkh refers to metals. It is based on thin shell theor) . nodi 6ed to include the effects of cr:ep. The tests on HICs to whkh CNSI refers give little b, formation on the buckling of IUCs except under the most favorable conditions (full container, little creep, no change in the properties of the contents, no loss of integrity of the container).
(2) ne arguments advanced by CNSI against the secant me.dulus approach to estimation of the creep buckling load are a smokescreen. The secant modulus represents an ef'ective stiffness of the material incorporating creep straint It appears to be especially appropriate ir, this case because, as shown in the creep test data cited in Structural Resiew, the extension of Marler CL 100 in a creep test levels off over time at low stresses. It is true that estimation of creep auckling loads is highly uncertain, and there are other methods besides the secant modulus approach CNSI had the option of using one of these alternative methods, but they chose not to do this.
(3) As stated in the Structural Review, the soil will have some effect on the buckling load.
However, estimates of this effect based on the assumption of small deflections cannot be regarded as reliable. Also, the available experimental data for the buckjing of buried containers are for isolated structures, as reoosed to an array of HICs. The benefit from soil would probably be more pronounced for an isolated container, since such a container would be more able to shift load to the surrounding soil cs deflection starts In an anay of HICs, there is no pace to shift the load to except to other HICs. These considerarlons led .7e to assume a constant load boundary condition in the buckjing analysit here is nothing new in the CNSI report to refute this.
(4) As stated above, there is no problem in principe with taking credit for the waste as a structural component. However, adequate characterization of this materid must first be performed, including characterization of its long term properties Soil arching can either increase or decrease the vertical load on a container depending on the relathe stiffness of the container and the soil. However, if it decreases the vertical load on top of the container, part of the load must be shifted to the surrounding sol. This would incre.uc the lateral loads on the container sidewalls. Thus it is not clear whether this would increase or decrease the critical buckling load.
( '
' .. j
,i 1
)
, Mr. Derek A. Widma)er 4- August 19, 1988 CNSI dtes as reference 12 a test report on IUC in a soi.like medium. I do not have a copy of this report, but I note the careful wording with which the test results are described: "no buckling fa0ure was observed in the container." It is not dear what a "bucding failure" is.
In the middle of page 27 CNSI sutes that BNI. In the Pires report, condudes that."the initial buckling load is smaller than the creep buck!!ng load." Although the Pires report does contain such a
. statement or page 35, it is merely one posstle scenario. It is ba.ed on the asumption that creep prior to buckling causes the container walls to move inward, thus decreasing the presaire on the walls Pires dalms that this efect more than compensates for the loss in stiffness of the container due to creep. This implicitly asumes that the response of the soa is clastic. and this assumption is questionable for the reasons discussed above ' .d in the Structural Re',iew. Furthermore, if load is taken off the sidewalls, ,t must go somewhere else, most probably the top of the HIC This wGI tend to accelerate buckling. (Pires did not consider the contribution of the vertical loads toward buckling.)
As a Snal point, note that on page 23 OlSI was dalming that arching transfers load from the dome to the surrounding sou. Now they are in effect daiming credit for a transferral of load from the surrounding sou to the top. They cannot have it both ways.
6.0 HDPE H1C creep stress evaluation. In a!!uding to LP ' tabil.;hed practice of allowing local plasth _
flow in the design of metal structures, CNSI demt.nstrates that it does not know the diference between plastidty and creep. Plasticity is to a close approximation time independent and is of importance only above the yield stress. Creep is time dependent, and in polymers it occurs even at low stresa lewis.
This is why the design of structures with polymers for substantial loads is difficult. (CNSI is suddenly appealing to design methods for metal structures, in spite of its earlier protestations to the contra.y.)
The daim that an increase in wall thickness due to compteuive loading will sigruncantly affect the stress distribution is easily dismissed. At a tensile stress of 500 psi, the extension of a unlaxfal test specimen is 3%, including creep. This means that the crou sectional area reduction is of about the same amount, assuming near incompteuibility. Therefore !. compreuive load of 500 psi should result in an area increase of about 3% So the effect of area cl.ange is no more than about 6% at 500 psi in applying the tensile data to compteuion.
CNSI is still missing the po!nt about creep, uhkh is that it diminishes the effective stiffnen of the material. There is no way to design a structure invohing significant bending stresses when the creep strains are of the order of those expected in HICs.
[1] P. Soo et al, The effects of emironment and gamina irradiation on the mechanical properties of high density polyethyle 'e," NUREO/CR 4407 (1986).
[2] P Soo, B. S. Bowerman, J. II. Cinton, L Millan, and C L A_nderson, "L.ow level waste package and engineered barrier study, quartety progress report, October December 1987,*
Brookhaven National Laboratory report WM.32914 (Apn!,1988).
[3] E. Oraube, if Gebler, W. M611er, and C Gondro, "Creep rupture strength and aging of IIDPE pipes," Kunsts. ffe 75 (1985) 7, pp. 412 415.
Sincerely, Stewart A. Silling SAS;PCC Anistant Professor of Engineering
-