ML20217H512
| ML20217H512 | |
| Person / Time | |
|---|---|
| Issue date: | 07/27/1997 |
| From: | Raddatz M, Sturz F, David Tang NRC OFFICE OF NUCLEAR MATERIAL SAFETY & SAFEGUARDS (NMSS) |
| To: | |
| References | |
| NUDOCS 9804290403 | |
| Download: ML20217H512 (7) | |
Text
i
- r. .
l ASME Pressure Vessels and Piping Conference Orlando, Florida, July 27 31,1997
[
i l NRC STAFF TECHNICAL APPROACH FOR SPENT FUEL STORAGE CASK DROP AND TIP0VER ACCIDENT ANALYSIS l by David T. Tang, Michael G. Raddatz, and Frederick C. Sturz
! Spent Fuel Project Office U. S. Nuclear Regulatory Commission Washington. D.C.
i
- 1. INTRODUCTION 2. STAFF ACTIONS TO DEVELOP THE APPROACH In the process of approving spent fuel storage The approach to use the steel billet test data casks. the U.S. Nuclear Regulatory Com1@n to develop cask drop and tipover analysis models staff reviews the design basis drop accidert for is based on the staff actions as follows:
the operating condition of moving casks to and on the storage pad and. to provide defense in Develoo End Droo Test Data.
depth, the non-mechanistic cask tipover accident is analyzed. Cask drop and tipover introduce In 1993. the NRC staff completed three series of the most severe mechanical loads to a cask. end-drop tests to develop low-velocity impact Concurrent with operating pressures and data suitable for benchmarking or validating temperatures. they often govern the cask design. analytical approaches for calculating storage cask deceleration loads. The tests were led by The structural design analysis of the cask drop the Lawrence Livermore National Laboratory and tipover accidents involves determining (LLNL) and Sandia National Laboratories (SNL) deceleration g loads and analyzing the with collaboration of British Nuclear Fuels associated cask stresses. Over the years. cask Ltd.. the Electric Power Research Institute vendors have been generally successful in (EPRI). Science Applications International applying the quasi-static analysis approach to Corporation. Inc., and the NUHOMS Owners Group demonstrate cask structural Integrity. (Ref. 2). The tests involved end drops of a Determining deceleration g loads. however. 3.000 kg (6.600-pound) steel billet onto various remains to be problematic in that gross configurations of scaled concrete pads and soil assumptions made for analysis often led to conditions. A 64.5 ton (142.000 pound) near-unrealistic g load calculations. which. in turn. full-scale empty Excellox 3A transport cask was would prevent stress conditions from being also tested by end dropping onto a full scale evaluated properly. As a result, the staff concrete pad. The user comunities have performed safety evaluations by using published two technical papers (Refs. 3. 4) in conservative analysis assumptions. In a few an attempt to correlate the Excellox 3A test cases, the staff 1mposed additional conditions, results to those predicted with the EPRI such as never allowing the cask to be carried monograph method (Ref. 5).
horizontally, requiring an impact limiter to be installed at the cask top, and using seismic Develoo Side Droo and Tioover Test Data.
Category I storage pads.
While the NRC staff has not performed a formal In order to develop a method to evaluate more review of the EPRI method, it 15. however, aware realistic conditions for cask drop and that, for the cask side drop and tipover, the tipover analyses the NRC initiated confimatory results based on the EPRI method could be very testing. Recently, four series of low velocity sensitive to the estimated footprint size of a impact tests were completed using either a 1/3- cask. To supplement the previous end-drop scale steel billet model or a near-full scale billet tests in February 1996, at the LLNL. the empty transport cask to collect suitable staff perfomed a series of billet side-drop and experimental data for the development of cask tipover tests. Early this year, an cnalysis analytical models. Using the billet test data, report describing the use of billet test data to the staff demonstrated that a finite element develop the finite element model of a cask-pad-analysis approach can be developed to include soil interaction system was the pad and soil flexibility for calculating public document room (PDR) Ref.(placed 6). A in NRC's cask deceleration g loads. The staff's approach companion report on the test data and the follcws the Standard Review Plan (SRP) which diskettes containing digitized raw data have provides that the analytical model should be also been docketed (Refs. 7. 8).
l- balidated (Ref. 1). This paper sumarizes the l staff actions and their bases for developing the - ,
I approach and is expected to provide sufficient insight to enable potential users to develop Utl i their own task drop and tipover analytical U
.V- l methods.
l M '
9804290403 980727 PDR ORG NOMA ,
]
DeveloD Method to Extract Ricid Body Motion. response of a structural system subject to impulsive loads. Normally, in model validation.
An examination of the modal properties and it may seem obvious to require all three typical Fourier amplitude spectra of the billet parameters to be rigorously matched between the tests suggests that the rigid body motion can be calculated and test results. Considering the obtained by removing all vibratory motions from various approximations involved in determining l the total response. This was based on the the dynamic load factors for quasi-static loads. '
observation that the dynamic response of a steel hcwever, the staff determines that it is only billet or cask is comprised of two parts: the practical to use primarily the rigid body rigid body and the vibratory components. Tne response amplitude as the basis for judging the staff pursued the approach by simply low pass cdequacy of an analytical model. As such. the filtering the total response at a lower validated billet-pad-soil analytical model frequency than the lowest mode vibration should predict a pulse amplitude slightly higher frequency of interest. The plots in Figures 1 than the recorded one. The calculated pulse thru 3 demonstrate the process and confirm that duration and shape, however, need not be the approach is practical and theoretically compared rigorously with, but are expected to sound. look similar to, those recorded.
Figure 1 compares two acceleration time AdaDt Billet Test Model for Cask Pad Soil histories, unfiltered and filtered at 850 Hz. System.
for a .91 m (36 in) billet side-drop test at LLNL. The filtered data displays a An analytical approach is adequately validated predominantly single-frequency waveform for cask response analysis if it has been oscillating at about 690 Hz, which corresponds validated for modeling the billet-pad-soll test j to the free-free flexural mode of vibration of configuration. On this basis, adapting the !
the steel billet of .514 m (20.25 in) in billet-pad soll model for the cask-pad-soil !
diameter and 1.83 m (72 in) in length. Figure 2 system is a straightforward process. Following j compares the unfiltered time history with the the typical modeling practice the finite q rigid body response obtained by low-pass element scheme of the billet is replaced with 1 filtering the total recorded time history at 450 that of the cask. and the use of site specific '
Hz, which is lower than 80 percent of the lowest pad and soil properties in the cask pad soil frequency of the billet free-free vibration at model requires no further model validation.
691 Hz. Figure 3 displays three more comparison f plots. It is worth noting that all filtered 3. APPLICATION EXAMPLE time histories display either a single- or a double-hump large initial pulse characteristic On the basis of the above development, the staff of the rigid body response of a billet-pad-soil recently applied the approach in performing interaction system. Phenomenologically, the storage cask safety evaluations. Salient points second response hump, which is related to billet of an application example for determining the rebound. can be considered test anomaly of not basket quasi-static g loads are discussed below.
being able to execute a perfect billet side drop. Basis for Acceptino the Billet Test Model.
Develoo Billet Pad Soil Analysis Model . Figure 4 presents the recorded and calculated tipover billet deceleration g loads. unfiltered Recent advancement in finite element analysis and filtered at 450 Hz. By comparing the rigid-techniques allcws the cask drop or tiDover body response time histories. the staff accident to be modeled with general purpose concludes that the DYNA 3D billet-pad-soil model computer codes available in the public domain. is validated because: (1) the calculated rigid-l To demonstrate that an analytical model, based body deceleration amplitude of 233 g exceeds the on the DYhA3D code (Ref. 9), can be used to recorded deceleration of 213 g and (2) the shape calculate cask rigid-body deceleration loads, and duration of the calculated rigid-body the staff validated the billet-pad-soil test response are similar to those recorded.
analytical model from which a cask-pad-soil analysis model would be derived. Model Cask Pad Soil AnalY5iS Mcdel.
l validation was accomplished by correlating the calculated response of the billet-pad-soil in developing the cask-pad-soil model. the system to the test results. This became billet test modeling features are retained for possible only after suitable billet test data the soil medium and for the billet-to-pad and was generated. The billet test data is ideal pad-to-soil sliding surfaces. The billet pad-for this purpose since the billet modal soil test model was modified in two respects:
properties can readily be defined to allow (1) the billet test pad was replaced with a extraction of the rigid body response from the site-specific pad of .91 m (36 in) in depth and i total response. 4.05 m (160 in) by 5.08 m (200 in) in tributary
! area and (2) the finite element model of the l Select Model Validation Acceptance Criteria. billet was replaced with that of the cask and I
its basket and spent fuel weighing a total of l The pulse amplitude, shape, and duration are the 105.5 ton (232.000 lbs).
l three major parameters that govern the dynamic 2
I Ricid. Body Cask Deceleration Loads. Facilities.* Proceedings. 5th International Conference on Nuclear Figure 5 displays the calculated cask Engineering. Nice. France. May 1997.
deceleration time histories unfiltered and filtered at 350 Hz. The 350 Hz low-pass filter 4. Stokley, J. R. and Williamson. D. H..
was selected to show the effects of removing the " Structural Integrity of Spent Nuclear high-frequency vibratory components of the total Fuel Storage Casks Subjected to Drop."
response. Since the filtered response still Nuclear Technoloav. Vol. 114, April retains some components associated with cask 1996.
vibration, the " apparent" peak response of 66.7 g is conservative. 5. Rashid. Y. R., Nickell. R. E. and James.
R. J., " Structural Design of Concrete Basket Quasi Static Deceleration Loads. Storage Pads for Spent-Fuel Casks.*
Electric Power Research Institute Report An inspection of Figure 5 suggests that, for the NP-7551, Palo Alto. California. April purpose of computing quasi-static loads on the 1993, basket, the cask rigid body deceleration time history can be characterized with an isosceles 6. Witte M.. et al.
- Evaluation of Low-triangle pulse for a duration of 3 msec and an Velocity impact Tests of Solid Steel amplitude of 66.7 9 For the basket spent fuel Billet onto Concrete Pads and compartment vibrating at the lowest mode Application to Generic ISFSI Storage frequency of 137 Hz. the corresponding dynamic Cask for Tipover and Side Drop."
load factor of 1.12 provides a peak quasi static Lawrence Livermore National Laboratory, load of 75 g. which is well below the 88 g UCRL-ID,126295. Livermore. California, bounding tipover deceleration load analyzed to March 1997.
Qualify the basket.
- 7. Witte. M.. et al.,
- Low-Velocity impact
- 4. CONCLUSIONS Testing of Solid Steel Billet onto Concrete Pads." Lawrence Livermore Using the billet test data, the staff developed National Laboratory UCRL-]D-126274 a more realistic finite element modeling Livermore California. March 1997, approach for evaluating storage task side-drop and tipover accidents. The side-drop and 8. Witte. M., Lawrence Livermore National tipover test data and the analysis report have Laboratory, letter forwarding data been placed in the NRC PDR. As a result, the diskettes containing the drop and staff is beginning to discuss the approach with tipover tests. NTFS97-76 N .
cask vendors in license pre-application June 4. 1997.
meetings. The two technical papers in the conference. the one by the LLNL and the one by 9. Whirley. R. G.,
- DYNA 3D. A Nonlinear, the staff, surnarize the analysis steps and Explicit. Three-Dimensional Finite results. To further aid potential users in the Element Code for Solid and Structural interpretation and use of the billet test data Mechanics - User Manual." Lawrence and to develop analytical models more Livermore National Laboratory. UCRL-MA-effectively, the staff is sponsoring LLNL to 107254. Rev. 1, 1993. ;
prepare a report documenting the detailed bases i for selecting modeling parameters and acceptance '
criteria for this approach. The report is scheduled to be completed by the end of the year and will include the development of cask end-drop analytical models.
- 5. REFERENCES
- 1. U.S. Nuclear Regulatory Connission.
" Standard Review Plan for Dry Cask !
Storage Systems." NUREG-1536 January '
1997.
- 2. McConnell P., et al.. " Test Report. I Drop Tests Onto Concrete Pads for Benchmarking Response of Interim Spent Fuel Storage Installations." Sandia National Laboratories. Albuquerque. New .
Mexico. September 1993. l
- 3. Rashid. Y. R., James. R. J. and Ozer.
0.. " Validation of EPRI Methodology f0f Analysis of Cask Drop and Tipover Accidents at Spent Fuel Storage j 3
i
^ -
300 200 g
/ J l _ -,
jr [ k '
)
i g 0 g 100 0.016 Time (sec)
FIGURE 1 TEST #4, BILLET SIDE DROP, TEST DATA FROM CHANNEL A3:
UNFILTERED AND FILTERED AT 850 HZ 300 200 g
/ si n f
j
^
v ,,
] l M i 9
<I d
~
j L nr [
]
i i l I l l l J't Il f 100 200
- 0. M 0.016 Time (sec)
FIGURE 2 TEST #4. BILLET SIDE DROP. TEST DATA FROM CHANNEL A3:
UNFILTERED AND FILTERED AT 450 HZ 4
300<
2M.
g Test #3 Side Drop, 18 in (0.46 m) b i 100<
/) l j
fa i 300-Test M Side Drop. 36 in (0.91 m)
} t i 1 j 100 i i
p fI hl i
8 j E Test M Side Drop. 72 in (1.83 m) 8 j
5 *
} 2%- e i
- \
- ,
I 'fdl
( 11 i it,.h ago l t ljr pip'urr 1
200-0.000 0.010 0.020 Time (sec)
FIGURE 3 TESTS #3. 8. AND 9. BILLET SIDE DROP, TEST DATA FROM CHANNEL A3:
UNFILTERED AND FILTERED AT 450 HZ 5
600-400-3 h
$ 200 rp '
d!
,- h 11o . i. Ii hlk !
'v i I l
ff !P I jl
\
200-TEST RESULT: MAXIMUM ACCELERATION = 213 G l
0.00,0 0.010 0.020 Time (sec) 600 400 S
' ..Ms..
g .
v th"U"W N
200 ANALYSIS RESULT: MAXIMUM ACCELERATION = 233 G 0.000 0.010 0.020 Time (sec)
FIGURE 4 COMPARISON OF BILLET TIPOVER ANALYSIS AND TEST RESULTS, UNFILTERED AND FILTERED AT 450 HZ
1 C. ~
20 -
a _
}"\
j
- l
_ / f 1 9 * \ r ,
5 s
(/ \
g* ) l {
l 8 \ \
M e
- c. coo c.cos 0.010 0.01s o.020 c.ou c.oso c.on c.o4o a.oss time (sec) f FIGURE 5 CASK TIPOVER ANALYSIS RESULTS, UNFILTERED AND FILTERED AT 350 HZ MAXIMUM ACCELERATION = 66.7 G 7
1