ML20195G558
| ML20195G558 | |
| Person / Time | |
|---|---|
| Issue date: | 06/28/1988 |
| From: | Nussbaumer D NRC OFFICE OF GOVERNMENTAL & PUBLIC AFFAIRS (GPA) |
| To: | Ingersoll C WASHINGTON, STATE OF |
| Shared Package | |
| ML20151C617 | List:
|
| References | |
| FOIA-88-470 NUDOCS 8811280074 | |
| Download: ML20195G558 (1) | |
Text
.
' $?A*e d
[ '
JUN 2 81988 Ref: SA/ CHM Mr. Clarence E. Ingersoll, Head Radioactive Waste Program Department of Social and Health Services Mail Stop LE-13 Olympia, Washington 90504
Dear Mr. Ingersoll:
Enclosed for your review is report entitled, "Review of the Structural Design of Polyethylene High Integrity Containers." This report was developed for the NRC Division of Low-level Waste Management and Decow.issioning as a study of the scientific literature pertaining to
, the structural designs of high integrity containers for use in the burial of low-level radioactive waste. Based on the topical reports submitted to the NRC and the scientific literature available, this report concludes that it is impossible to have confidence in the ability of any
. polyethylene HIC designs to retain a sufficient degree of structural integrity for the 300 year period.
This report will be discussed at the meeti. ' the Advisory Comittee on Nuclear Waste (ACNW) scheduled for June <.. to 29, 1988. The current draft agenda for the ACNW meeting indicates tnat the report will be discussed at 8:15 am on June 28.
If you have any questions, please contact Mrs. Cardelia Maupin at 301-492-0312.
Sincerely, OriBinal signed by pot;ALD A. IN335ALG Dunald A. Nussbaumer Assistant Director for State Agreements Program State, Local and Indian Tribe Programs
Enclosure:
Distribution:
As stated SA R/F Dir R/F Identical ltr. sent to S. Marshall, NV CHMaupin Tonical opnnete filo
>FC: SLITP/SA :5LITP 5 : : : : :(f 4 AME: in/bh : DAN M baumer : : : : :
.'---- ------------------------------------------------------------------------------------I--
' ATE:6/27/88
- 6/[88 8911200074 00111b h
,
- p ut0u9 bhbh-
- k, UNITED STATES o [ g NUCLEAR REGULATORY COMMISSION
, 5 j WASHING TON, D. C. 20555
/
JR 131938 MEMORANDUM FOR: Donald A. Nussbaumer Assistant Director for State Agreements Prograe Office of Governm4ntal and Public Affafts FROM: John J. Surmeier, Chief Technical Branch Division of Low-Level Waste Management and Decomissioning, NMSS
SUBJECT:
REPORT ENTITLED, "REVIEW OF THE STRUCTURAL DESIGN OF POLYETHYLENE HIGH INTEGRITY CONTAINERS."
The report developed for the Division of Low-Level Waste Management and Decomissioning entitlid, "Review of the Structural Design of Polyethylene High Integrity Containers," is enclosed. We are transmitting it to you for your information and for transmittal to the appropriate Agreement State authorities.
The report will be discussed at the meeting of the Advisory Comittee on Nuclea'.-
Waste (ACNW)scheduledforJune27and28,1988. The current draft agenda for the ACNW meeting indicates that the report will be discussed et 8:15 AM on June 28.
We will make sure you are informed of the final date and time for the discussion of the report when the final agenda is published.
If you have any quetions, please contact Dr. Michael Tokar at x20590.
_e_- -
John J. Surmeier, Chief v-Technical Branch -
Division of Low-level Waste Management and Decomissioning, NMSS
Enclosure:
As Stated
/,
v7'n tfrW ?TT77 (l}-
~
1 l
I REVIEW OF THE STRUCTURAL DESIGNS OF POLYETHYLENE HIGH INTEGRITY CONTAINERS S. A. Silling Assistant Professor of Engineering Brown University Division of Engineering Providence, Rhode Island 02912 prepared for U.S. Nuclear Regulatory Commission Division of Lowd evel Waste hianagement and Decommissioning June 10,1988
_ D CALd7 d l O I vo yw v v r ' ' ' d{(f) y t
_1_
This report contains a review of the structural designs of high integrity ,
containers (HICs) for use in the burial of low level radioactive waste. It is based on the topical reports and supplemental material submitted to the U.S. Nuclear Regulatory Commission (NRC) by the vendors [1- 6). It is also based on a review of the scientific literature.
The design of a HIC made of high density polyethylene (HDPE)is a unique problem in engineering. These containers are required to maintain substantial structural integrity for a period of 300 years under conditions of large stress, exposure to radiation, and possibly detrimental chemical environment. The de-sign of polymer structures for long term service is a notoriously difficult problem.
The longest term applications of HDPE are for water and gas pipelines, which are usually designed to last 50 years. The synthesis of HDPE polymers which ,
could be demonstrated to last even this long was a considerable achievement of polymer science, even in the absence of radiation damage. While improvements are occasionally made in HDPE polymers, there is no evidence that the mate-rials proposed for the HIC designs can last for 300 years under the expected ,
conditions, and there is considerable evidence that they may not.
The Space Shuttle disaster was caused by a change in the mechanical prop- l erties of a synthetic rubber 0 ring in response to a moderate change in temper - i ature. This event demonstrates the sensitivity of solid polymers to a multitude of conditions, including temperature, environment, molecular weight, processing and fabrication methods, age, and exposure to nuclear radiation and light. It is therefore essential that realistic data for the mechanical properties be used in the design of polymer structures.
The HIC vendors should have made use of the published data on how the mechanical properties of polyethylene change in response to age and radiation, both of which can lead to embrittlement and failure of a structure. (See [7,8) for reviews.) In fact, these effects were essentially ignored in the designs. The designs also neglected or inadequately considered the phenomenon of creep, even -
though it is common knowledge that creep is the predominant long term mode of deformation in most solid polymers. The sssumptions about the modes of failure to be expected were generally unrealistic.
The remainder of this report discusses some of these issues as well as creep buckling, which is shown to be virtually inevitable unless the waste itself is included as an active structural component. All of the problems discussed here j stem from the use of HDPE as the primary load bearing material.
The appendices to this report contain specific comments on the designs of each of the vendors. They are so divided because they contain proprietary data.
4 Most of the issues discussed here have been raised previously in a report by
- the Engineering Design and Testing Corporation (EDTC) [0,10), although the ;
present analysis was conducted independently. The EDTC report also discusses 4
- k
additional problems which have not been pursued here because of the limited time available for this review.
- 1. Degradation of mechanical properties of HDPE In all three HIC designs under consideration, the HDPE shell is the pri-mary structural element and is required to support the entire overburden. It is therefore necessary to consider the long term material properties of HDPE.
HDPE exhibits two general types of failure modes. Over short time inter-vals, the primary mode of failure is ductile, meaning that the material deforms extensively prior to rupture. Typically an HDPE sample in a uniaxial tension test will sustain an increase in length of a factor of at least 2 before it fails
[11,12]. Over longer time intervals, the predominant mode of failure in most types of HDPE is briffle, meaning that the material fails without extensive de-formation. In the brittle mode, failure may occur after only a 10% stretch in a uniaxial test (13].
The reasons for the transition from the ductile to brittle failure mode are not at all well understood. Fundamental research in this area is still being carried out. The long term brittle failure at low stress has been shown to occur by slow.
crack growth, and it is sensitive to the presence of microscopic defects in the material (14). It is also sensitive to temperature and to molecular weight (13).
There is additional sensitivity to the extent of cross linking, an effect which is discussed further below. There is evidence that the biaxiality of the state of stres affects brittle failure significantly (15). The kinetics of the process of brittle failure have been shown to obey an Arrhenius. type of law (13].
Small changes in chemical and physical properties as well as in fabrication methods have significant effects on the failure of polyethylene. Since most of the published experimental data are for HDPE materials other than hiarlex CL 100, the previous work cannot be used directly to establi'sh quantitative -
design limits for structures made of hiarlex CL-100. Soo has been carrying out tests specifically for this material (16- 22], and his results, some of which are preliminary, are discussed below.
The most complete set of long-term experimental data for any HDPE were obtained by Graube (23), who conducted 20 year tests on the ruptuue or pressur-ized pipe. The material was Hostalen Gh! 5010, manufactured by Hoechst. This is a linear HDPE of moderately high molecular weight with some carbon black content. The pipes were made by an extrusion process, and it is not known how this will cause the results to differ from those for a rotationally molded structure.
l Figure 1 shows data of Graube (23] for HDPE. The horizontal axis is time, and the vertical axis is stress. Each curve shows, for a specific temperature, the stress at which failure occurs in the pressurized pipes. The distinct bends in the l
\
I
,v-- - - - - .
- - ,,--,--r- --- -, - - ,- --_--- ~ ,--- -, ,
_3-curves, sometimes called "knees," indicate the transition from ductile to brittle failure. Following the transition, the failure stress drops off sharply, indicating an almost complete degradation in the mechanical properties of the material.
At room temperature (20'C), the transition from ductile to brittle does not occur for at least 10 years. However, this transition time is very sensitive to tem- !
, perature. Therefore experiments have been run at various higher temperatures i i
at which the transitions occur sooner. The data from the higher temperatures are then extrapolated to room temperature, based on the assumption of an i Arrhenius-type process. The extrapolated portion of the curves are shown as >
dashed lines in the figure.
The data in Figure I have been widely quoted in the literature on creep rupture and similar data serve as the basis for design of water and gas pipelines
]. [25]. Other aspects of this and similar data are discussed in references [7,24,26].
Gloor [27], applying a technique due to Larson and Miller [28], presents a ,
mathematical correlation for predicting the time to failure of HDPE at various stresses. The data modeled are those of Richard [13], which are essentially the same as those in Figure 1, but with fewer data points. Gloor gives separate -
corr 0lations for the ductile and brittle regions of failure. Barton and Cherry
[29] ha ' revised this correlation with no essential difference in the result for the-present arposes. At 20*C, the ductile to brittle transition time is about 11 yr. :
A report on the structural mechanics of HICs by the Brookhaven National Laboratory [30] used the data of Soo [22] in estimating stress limits for design against creep rupture. These data, however, were based on observations of failure at high stresses (above 1500 psi) and therefore reflect only the ductile failure mode. The resulting estimate of an allowable stress may not be conservative because it neglects the possibility of brittle failure.
The BNL correlation predicts that if the failure time is 300 yr, then the stress is 1100 psi. BNL uses their correlation as a basis for a recoinmendation of
<65 psi as a maximum allowable stress, which apparently includes a 30% safety 1 factor applied to the 1100 psi figure.
l The Gloor correlation discussed above predicts a stress of 694 psi for a brittle failure time of 300 yr. (This value is found by linear interpolation of the ,
logarithms of the tabulated data in [27].) This value is not to be interpreted
! as a recommended design value; it is merely a statement of what a certain i correlation predicts. The 694 psi value includes no eafety factor. There are I many uncertainties in the data, 2ncluding much scatter. As stated above, it applies to a type of HDPE somewhat diferent from Marlex CL-100, and it is '
i for unirradiated material only.
Radiation also has a large effect on the brittle failure of HDPE. Radiation in general increases tensile strength in polymers, as the HIC vendors are quick 4
to point out. It may either reduce or increase the creep rate [22] depending on '
r I
I j t 4 . - , - _ - , _ . , _ _ . _ - _ - _ , . _ - _ _ _ _ - - .. ~
- . l 10 8 a 108 1 08 106 30 .. . . . . .. . .. . ,, 30
. I ,
x
- l'NN xx i
bx NN w l.
1,,,W%NN;', .-
,,w e.;
- l. -'SSNNNX 3
- - l+,,,7.;
i .
\ sxs ' . ., e o , ,.
\ \ \ %\ l- ,
N \ .
2 $6 '
4 s l
1.5 ' ==i .- - . . .
8 '8 10-3 1 108 10' r.. m IQ*t' soo v c I
Figure 1. Time to fe.!!ure of unirradiated HDPE (data of Graube (23), figure taken from [24)).
I
___i
i the dose rate. However, its major effect for purposes of failure analysis is to make the material more brittle.
Soo (16- 22] has conducted detailed experiments on the effects of radiation and chemical environment on the creep and failure of hfarlex CL-100. Among the techniques employed were the irradiation of U bend specimens, which are short, thin strips deformed into the shape of a U. Since the specimens are bent prior to irradiation and since no additional loads are applied, the experiments take place essentially at constant strain (as opposed to constant load). These tests are particularly relevant to HICs because a buckled container would contain similar bends at strains comparable to those in the tests.
The U bend specimens irradiated at doses expected in HICs show the ini-tiation and growth of cracks at the bends. In at least one test such a crack propagated completely through the specimen. These results are consistent with the observed loss of ductility in irradiated Afarlex CL-100 in uniaxial creep tests conducted at BNL [22), by the HIC vendors, and by others.
There is evidence that a carefully controlled amount of cross linking tends to reduce the susceptibility of HDPE to brittle failure, particularly at elevated temperature. For this reason cross linked HDPE has been proposed as a ma-terial for small diameter hot water pipes in buildings (31,32). However, this-improvement is dependent on many factors, including the type of cross linking process, the degree of cross linking, the chemical environment, and the manu-facturing process. A troublesome aspect of the data for long term brittle failure of the cross linked HDPE is that although the transition to the brittle mode is delayed, failure is much less sensitive to the stress level once the transition occurs. Thus the material may fail even for very low stresses (31).
The use of an intentionally cross linked polymer such as hiarlex CL 100 would not be expected to help resist radiation embrittlement, since one of the pri-mary mechanisms of this embrittlement is by uncontrolled cross linking. Cross- ,
linking improves the mechanical properties of HDPE only up to a point, beyond -
which further cross-linking is detrimental. Kunert (33) and Liu et. al. [34]
have found that tensile strength and elongation of HDPE at the break point in a uniaxial test are both adversely affected by excessive cross linking. This is consistent with the observed radiation embrittlement.
In summary, while the issue of long term brittle failure of HDPE is a poorly l understood one from the point of view of basic science, the following may be concluded:
e Even without radiation, brittle failure at low stress occurs in many types of HDPE over the long term.
e Radiation causes cracking in hfarlex CL 100 at doses typical of HICs at moderate strains.
i
Because of the complexity of the problem and the sensitivities mentioned above, it may be impossible or impractical to find a dependable value for a safe design level of stress for a 300 year life.
The HIC vendors used as a failure criterion for HDPE the the tensile yield strength, which is typically quoted as 2600 psi for hfarlex CL-100 [12]. This parameter is valid only for short term loads, since it does not include the effects of creep and long term brittle failure. The use of the tensile yield strength in an analysis of failure is completely inappropriate for the design of a long term structure and will lead to a gross overestimate of the strength of the structure.
1
- 2. Uncertainty about creep properties of HDPE The previous section concerned the brittle failure of HDPE. Even ignoring the possibility of such failure, the creep properties of hiarlex CL-100 under long-term loading conditions are not well known, even without considering the effect of radiation c,d chemical emironment.
i The only published data specifically for this material are from two sources:
the Phillips Chemical Company, which manufactures the resin [11]; and Soo [16-
, 22]. The only data provided by Phillips for long times at moderate stresses are '
a single uniaxial creep curve in air at 500 psi at room temperature (page B 38' of [11]). This curve, which is for a test period of 104 br (1.1 yr), is reproduced in Figure 2 along with the curves for higher stresses. The data of Soo, which are for stresses above 1000 psi, are generally consistent with the Phillips data.
The vendors claim that creep is insignificant at the expected stresses in ,
HICs. However, the Phillips data for 500 psi plainly show that this is not the case (Figure 2). The usual way of quantifying creep test data is in terms of the secant modulus, also known as the creep modulus, defined by E,(t) =
where t is time, and er and e are the stress and strain in a uniaxial creep test '
- respectively. Referring to the 500 psi curve in Figure 2, the strain after 1 hr is 3.0%, meaning that the secant modulus at this time is 16,700 psi. As an estimate ,
of the secant modulus of hf arlex CL-100 after 300 yr at 25'C, the following value is adopted in the creep buckling analysis below:
- E,(300) = 10,000 psi.
l Obviously, this value is highly uncertain, but there is no reason to believe that it is excessively conservative.
A realistic analysis of the deformation in a HIC must account for creep :
by the use of such an estimate for E, or by some other means. However, the l
l i
7
- s. .
L o r i . i i, i i i i ,i i .; oo n ~.-
to I 3-m a.
v> -
d a-
.o n o -
. N
- .o n .
i L o
v>
m v,.
a.
..o
.e s "..
o o 1.
4 o o o
l
@ l' v) a.
.s.
. s o
o .
C t-
. o 4
. , i i i e i \. i o o o o o o o o o, o o o e e m m .o n y . m -
~
.5 'N011 YON 073
\{
Figure 2. Uniaxial creep curves for Marlex CL-100,25' O.
HIC vendors neglect creep almost completely. Their analyses generally use the Young's modulus, which characterizes the just the elastic response, and whose nlue is E = 100,000 psi. Thus creep even after I hr causes strains about 6 times as large as would be predicted by the elastic analyses performed by the designers. Furthermore, in spite of the flatness of the 500 psi curve in the interval from I hour to 1 year, it is possible that creep will accelerate later.
A perplexing aspect of the Phillips data is that it shows that at room temperature, hiarlex CL-100 creeps more than similar HDPE homopolymers and copolymers, in spite of the cross linking of CL-100. The Phillips document
[11) gives creep curves for a number of these polymers, and the creep strains for 100 hr are given in Table 2 below. It is not clear precisely what the differences are between these polymers.
Table 2. Creep strains after 1C0 hr for HDPE polymers, 25*C [11).
Yolymer Stress, psi Strain, %
hiarlex BHB 5003 1250 6.5 hfarlex BHB 5012 1250 15.5 hfarlex EHB 6002 1200 7.5 hiarlex EHB 6009 1250 8.2 -
hiarlex EHB 6015 1250 6.5 Afarlex EhfB 6035 1250 5.7 hiarlex CL 100 1250 15.0 These data show that hiarlex CL-100 actually creeps more than all of the other types of HDPE except for one.
It may be possible to evaluate the long term creep properties of hiarlex CL-100 for lower stresses by conducting tests at elevated temperature. hiost polymers creep faster at higher temperature. The usual means of ascertaining the creep behavior for long times at room temperature is to test the material .
for short times at higher temperature. This method is not as reliable for crys-talline polymers such as HDPE as it is for amotphous polymers. The presence of cross. linking in hfarlex CL-100 may also influence the validity of elevated temperature testing. Other approaches to long term testing are possible, but their applicability in this case is beyond the author's area of expertise.
j 3. Creep buckling and excessive deformation This section discusses problems with the structural analysis aside from the issues surTounding the material properties.
3.1. Limits on deformation. The designers should have used explicit limits on the displacements of the containers under the expected loads. The reason for this is that the ability of the HICs to withstand large burial loads will depend i
i l
-g- 1 on the preservation of their shape. Gross changes in shape would completely !
alter the stress distribution.
3.2. Prediction ofloads. One of the troublesome phenomena which occurin the design of underground structures is arching [35]. This occurs when the structure has a stiffness different from that of the surrounding soil. If the structure is stiffer than the soil, then some of the overburden from sunounding areas is transferred to the structure. On the other hand, if the structure is less stiff, then some of the weight over the structure is transferred to the surrounding soil. For cylinders buried in a vertical position, arching has been found experimentally
[36] to influence the vertical load by up to a factor of up to 25. Prediction of the arching load depends on the type of soil, the moisture conditions, the flexibility of the container, and the means used to backfill soil following emplacement.
For HICs, the more dangerous condition would be if the structures were less stiff than the soil, because transferral of load to the surrounding soil would increase the lateral loads on the sides of the containers. This would increase the likelihood of buckling. The arching effect should therefore be considered in the design of HICs, although its impcatance may be diminished in a closely spaced array of Containes3.
There is uncertainty in the estimation of the lateral (horizontal) loads on' the containers. Assuming that the st ess in the vertical directior,in the soil near the container has been estimated, including the effect of arching, then lateral stress may be found from the conjugate ratio for the soil [37), given by g 1 - sin d 1 + sin d where K is the conjugate ratio and d is the friction angle for the soil. The value d = 30' is fairly typical, in which ease K = 1/3. The average lateral load on the side of the container is then '
pe = Kp, where ps is the lateral load (psi) and p, is the vertical load (psi). This value gives in effect the active lateral pressure, which is appropriate for walls moving away from the soil. There are other possible ways of estimating the lateral load which may be more applicable in special cases.
3.3. hsues related to buckling. There are several special aspects of the buckling problen4 in HICs which have not been fully addre. sed, and these are summarized here.
Most of the buckling analysis that has been done for HICs is based on the classical theory of linear elastic buckling as discussed by Timoshenko and
10-
, Gere (38). However, these methods are directly applicable only to linear chstic materials, such as steel at small strains. In HDPE the predominant mode of deformation is by creep, as discussed above. It is well known that creep strongly affects the ability of a structure to resist buckling because creep reduces the effective stiffness of the material. Buckling in creeping structures is known as creep buckling, and HICs should be designed to prevent this from occurring.
The consequences of the buckling of a HIC in terms of the release of radionu-clides are difBcult to assess. While buckling is not synonomous with rupture, it would appear to be a loss of stability in the sense of 10CFR61. Structural mechanics as a science is poor in its ability to predict post buckling behavior, and this is an active field of research. Buckling in a HIC would lead to strain concentrations in the form of kinks, and these might lead to accelerated material failures of the types discussed above. The integrity of the seal at the top of a container would be difficult to assure in the event of buckling. In view of previ-ously discussed observations of radiation embrittlement c,f HDPE, there is little cause for optimism about the ability of a buckled HIC to retain its structural integrity to any significant degree. Thus any prudent structural design for a HIC must resist buckling.
There have been some experiments on the creep buckling of cylindrical-nietal shells, as well as various proposed means of estimating creep buckling loads (39,40). An easy method to use which gives fairly good agreement with experiment is the secant modulus method. In this method formalism of linear elastic buckling analysis is used, but the elastic Young's modulus E is replaced by the current secant modulus E,(t). Thus the buckling load is a decreasing function of time. For a 3003: design lifetime, the value of E, at this time would be evaluated and used in formulas such as those in Timoshenko and Gere (3S). It should be noted that even in elastic buckling, there is considerable uncertainty in the application of such formulas, and this uncertainty is increased by the presence of creep.
All of the analysis of HIC buckling that has been done so far has examined one of two modes: (1) buckling by axial compression loads, or (2) buckling by radial presstre on the cylinder. However, it is likely that these two modes will combine in an actual HIC to give buckling loads that are smaller than would be pradicted for each mode separately. These may be ana.lyzed according to the method discussed by Timoshenko and Gere (38], adjusted as described above to account for creep.
Evaluation of these formulas for a "Generie HIC" with thickness h = 0.5 in, radius a = 24 in, and height t = 60 la was carried out on a computer. Figure 3 shows the predicted status of any combination oflateral and verticalloads with respect to creep buckling. The secant modulus was taken to be E, = 10,000 psi.
Any combination ofloads above and to the right of the diagonalline is predicted
to result in buckling, while any combination on the other side of the line is not.
Noting the diference in the horizontal and vertical scales, it is apparent that it is the lateral loads which cause the most problem. (The expected lateral loads are typically in the range of 3 - 10 psi, depending on depth, arching, and the type of soil.)
The figure shows that the allowable lateral loads, even without considering the vertical loads, are less than 1
- psi. Even if the material really were linear elastic with a Young's modulus of 100,000 psi, buckling would probably still occur.
The buckling of an underground shell is to some extent inhibited by the surrounding soil [41). However, this efect is sensitive to the type of soil. The experiments that have been performed are for single containers in a large body of soil, and hence they do not reBect possible interaction between HICs that are buried close together. (It is not clear whether this proximity would accelerate or resist buckling.) Also, the experiments are for metal containers, and it is doubtful that they would be applicable to the more Hexible polymer materials.
It seems possible that the large wall displacements that are to be expected in HICs would induce local failures in the surrounding soil that would negate any resistance by the soil to buckling. -
Luscher (42] used the results of experiments on metal pipes to develop a .
correlation which expresses the efect of soil support on the buckling load. This correlation is sometimes used in the design of stiE pipe. However, the Luscher approach has been criticized recently for application to very Rexible pipe because it assumes that the soil acts like a linear elastic spring (43). If the deformation of the cylinder is large, as would be expected in a HIC, the reaction of the soil will be nonlinear. Under the assumption of complete nonlinearity, it would be necessary to model the soil response as being a constant pressure load as discussed previously. This is the approach taken in generating the data in Figure
- 3. '
In any case, a simple estimate of the possible eEect of soil support in the case of HICs shows that even under the best of soil conditions, based on the Luscher model, creep buckling would still occur.
The response of the waste within the HIC should provide some resistance to buckling. However, none of the vendors has characterized the mechanical properties of this material. If the waste acts like a Buid (or a low friction granular solid), it will provide little if any resistance. If it acts like a rigid solid, then it may give a great deal of resistance. It appears that the type of waste may vary between the containers. If credit for the mechanical response of the waste is to be taken, then it would be necessary to adopt operational procedures to assure that the waste in each containers has adequate properties. In the absence of a plan for such procedures, credit cannot be given for the mechanical properties
. s- ,
4 g Generic Hl0 ,
thicknese- 0.50 redive- 21.00 height- 60.00 fini i-Creep bochting t.ed.
e..
i i.. \ !
7 L
... N 2
~
N .
tlN 8/ TAILLE l s.eas
. 3 N
'J STAELE 4 \
i O.M
~
N t
, r e.. ,
I 1
I e
I e
I a
e a a 3 i
a 3 s e I
Let.ra t e spets e
[
t i
Figure 3. Critical vertical and lateral loads for creep buckling. i I
t t
?
l
of the waste.
- 4. Summary and conclusions It has been shown above that there are problems and uncertainties in the structural design of HICs, all of which result from the use of HDPE as the primary load bearing material:
e HDPE may undergo brittle failure at long times, especially after exposure to radiation.
e The creep properties of Marlex CL-100 for long times are not known ade-quately.
e The designs do not include limits on gross deformation and shape change.
. The possibility of design against creep buckling appears hopeless, at least in the absence of any reliable data on the eEect of the waste.
Because of the problems and uncertainties discussed above, it is impossible to have confidence in the ability of any of the polyethylene HIC designs to retain a sufficient degree of structuralintegrity for the 300 year period. Simply put, HDPE is a poor choice of material from the point of view of structural.
mechanics. The problems are of a fundamental nature, and it seems unlikely that minor design changes will remove the concerns.
Ifit is decided to purme the current HIC designs with the aim of modifying them or resclving the uncertainties, then it will be necessary to institute a comprehensive program of testing related to the long term structural propersies of HDPE containers. This program should include the following:
e Assessment of the long term brittle failure ofirradiated HDPE. This could be done at elevated temperatures over a period of a few years. Fracture toughness data would also be helpful. -
e Assessment of the long term creep behavior of Marlex CL-100. The usual means of conducting these tests is at elevated temperature, which speeds up the creep process in a fairly predictable way, at least in linear polymers.
A complication may be the cross linked nature of this particular polymer, which tends to decrease creep at high temperatures. It is not clear how it afects creep at moderate temperatures, since, as discussed in section 2 above, Marlex CL-100 creeps considerably at room temperature. These tests would include the erect of radiation.
e Ebli scale structural tests. As noted above, there is great uncertainty with regard to the role of large deformation and the onset of buckling. These could be assessed by doing full scale tests of the buried irradiated HICs at elevated temperatures. Such tests would give information on the actual failure modes of the structures and the degree of soil resistance to buckling.
References
[1] Chem Nuclear Systems, Inc., "Topical Report, High Integrity Containers,"
CNSI HIC-1457101 P (December 23,1983)
[2] Chem Nuclear Systems, Inc., "Evaluation of Stress Loadings of CNSI HDPE HICs" (January 29,1988) .
[3] TFC Nuclear Associates, Inc., "Design Report of High Integrity Containers NUHIC 55 and NUHIC-120," TFC-TR 84 (June 26,1984)
[4] TFC Nuclear Associates, Inc., Amendment No. 2 to TFC-TR 84 (February 22, 1988)
[5] Westinghouse Hittman Nuclear, Inc., "Topical Report for Radlok High In-tegrity Containers," STD R 03-008-P (June,1984)
[6] Westinghouse Hittman Nuclear, Inc., letter report from Bryan A. Roy (Hittman) to hiichael Tokar (NRC), (February 29,1988)
[7] A.J. Kinloch and R.J. Young, Practure Behaviour of Polymers, London:
Applied Science Publishers (1983) 212 -
[8] R.O. Bolt and J.G. Carroll, Radiation Efects on Organie Materials, New York: Academic Press (1963) Ch. 5,6
[9] T.A. Jur, "An assessment of polyethylene as a material for use in high-integrity containers," Engineering Design and Testing Corp., Columbia, S.C.,828 660.010 (July, IfiS6)
[10) T. A. Jur and W.hi. Popliti, "A critical review of materials selected for high integrity containers," pre:ented at Waste Management 'BJ, Tucson, AZ (1987) ,
[11] Phillips Chemical Company, "Engineering Properties of hfarlex Resins,"
Bartlesville, Oklahoma, TShi-243 (1975)
[12) Phillips Chemical Company, "hfulex Crosslinkable High Density Polyethy.
lene Resins for Rotational hiolding," Bartlesville, Oklahoma, TShi-244 (1977)
[13) K Richard, E. Gaube, and G. Ludrich, Plastics 23 (1958) 444
[14] N. Brown, J. Donofrio, and X. Lu, "The transition betwenn ductile and slow crack growth failure in polytethylene," Polymer 28 (1987) 1326-1330 (15) J.L. Cooney, "EEect of morphology on biaxial stress rupture of polyethy-lene," Journal of Applied Polymer Science 8 (1964) 18891901
[16] P. Soo, B.S. Bowerman, J.H.Clinton, and C.I. Anderson, "Low-level waste package and engineered barrier study, quarterly progress report, June-July 1986," Brookhaven National Laboratory, Upton, New York, report Whi-3291-1 (1986)
[17) P. Soo, B.S. Bowerman, D.R. hiacKem:le, J.H.Clinton, and C.I. Anderson, "Low level waste package and* engineered barrier study, quarterly progress report, October-December 1986," Brookhaven National Laboratory, Upton, New York, report Whi-3291-2 (1987)
[18] P. Soo, B.S. Bowerman, J.H.Clinton, and C.I. Anderson, "Low level waste package and engineered barrier study, quarterly progress report, January-h! arch 1987," Brookhaven National Laboratory, Upton, New York, report Whi-3291-3 (1987)
[19] P. Soo, B.S. Bowerman, J.H.Clinton, and C.I. Anderson, "Low level waste package and engineered barrier study, quarterly progress report, April-June 19S7," Brookhaven National Laboratory, Upton, New York, report Whi-3291-4 (1987)
[20) P. Soo, B.S. Bowerman, L.W. hiilian, J.H.Clinton, S. Cowdery, and C.I. An -
derson, "Low level-waste package and engineered barrier study, quarterly progress report, July-September 1987," Brookhaven National Laboratory, Upton, New York, report Whi-3291-5 (1987)
(21] P. Soo, B.S. Bowerman, J.H.Clinton, L. hiilian, and C.I. Anderson, "Low-level waste package and engineered barrier study, quarterly progress re-port, July-September 1987," Brookhaven National Laboratory, Upton, New York, report Wh!-3291-5 (1987)
[22) P. Soo, H. Arora, K.J. Swyler, W. Becker, and E. Sobef, "The effects of environment and gamma irradiation on the mechanical properties of -
high density polyethylene," U.S. Nuclear Regulatory Commission report NUREG/CR-4007 (1986)
[23] E. Graube, G. Diedrich, and W. h!Giler, "Thermoplastic pipes: experience of 20 years of creep ruptum testing," Kunstsfoge 66 (1976) 2-8
[24) P.C. Powell, Engineering with Polymers, London: Chapman and Hall (1933) 140
[25] R.H. Wix, "Plastics pipe design," in Mechania.1 Performance and Design in Polymers, O. Delatycki, ed., New York: Wiley (1971)
[26) K. Richard, E. Gaube, and G. Diedrich, "Trinkwassermhre aus niederdruck-polysthylen," KunstJoge 49 (1959) 516-525
[27] W.E. Gloor, "Application of Larson-h! iller correlation to service test data on high-density polyethylene," Modern Plastics 36 (1958) 144
[28] F.R. Larson and J. hiiller, "A time-temperature relationship for rupture and creep stresses," Transactions ASME 74 (1952) 765-775
[29] S.J. Barton and B.W. Cherry,
- Predicting the creep rupture life of polyethy-lene pipe," Polym. Eng. Sci. 19 (1979) 590-595 (30) J. Pires, "Review of the high integrity cask structural evaluation program (HICSEP)," Brookhaven National Laboratory, Upton, NY, Nuclear Waste hianagement Division draft report (1987)
[31] hi. Ifwarsson and P. Eriksson, "Experience from 12 years evaluation of cross linked polyethylene," in Sixth International Conference on Plastics Pipes, March 1985, University of York p
[32] G. Rctschmann, "Long term behavior of flexible plastics pipes for district
- heating lines," Kunststof 75 (1985) 164-167
[33] K.A. Kunert, "Comparison of static mechanical properties of crosslinked polypropylene and polyethylene," J. Polym. Sci. Polym. Lef. 19 (1981)'
479-482
[34] Liu Dongyuan, Zhang Lianshui, Wang Yaqi, and Chen Wenxiu, "The ef-feet of radiation cross linking on the mechanical properties of polyethylene sheets," Radiat. Phys. Chem. 29 (1987) 175-177
[35] P.S. Bulson, Suried Structures, London: Chapman and Hall (19S5) p. 6
[36] P.A. Abbott, "Nonlinear static arching for vertically buried prismatic struc-tures," U.S. Air Force Weapons Lab. report AFWL-TR 160 (1966)
[37] h!.G. Spangler and R.L. Handy, Soil Engineering, 4th ed., Ne,v York:
Harper and Row (1982) p. 477
[3S) S. Timoshenko and J.hi. Gere, Theory of Elastic Sta6ility, 2nd ed., New York: hicGraw Hill (1961) p. 495
[39] J.hi. Chern, "A simplified approach to the prediction of creep buckl:-
time in structures," in Simplified Methods in Prenure Vessel Analysis, R Barsoum, ed., AShiE publication PVP PD 029 (1978) p.99-127
[40] H. Kraus, Creep Analysis, New York: Wiley (1950) p.137
[41] P.S. Bulson, op. cit., p.162 l [42) U. Luscher, "Buckling of Loll-surrounded tubes," ASCE J. Soil Mech and Found. 92 Sh!6 (1966) 211-228 l
l l
l l
[43) R.C. Prevost and K. Kienow, "Design of non pressure very flexible plastic pipe - I," Pipes and Pipelines Int. 30 (Nov.1985) 12-17 e
9 e
_