ML20043D680
| ML20043D680 | |
| Person / Time | |
|---|---|
| Issue date: | 03/06/1990 |
| From: | Advisory Committee on Reactor Safeguards |
| To: | Advisory Committee on Reactor Safeguards |
| References | |
| ACRS-2688, NUDOCS 9006110061 | |
| Download: ML20043D680 (16) | |
Text
__.__ = _
..j*
yi. ?f. Q;; (h E Yl W &
h;';k 06ES-AS$f ff,
, fU4 f.Y ' i fy( J&/l95 f),' '.k hk h CERTIFIED MINUTES DATE ISSUED: March 6,1990
SUMMARY
/ MINUTES OF THE ACRS SUBCOMMITTEE ON STRUCTURAL ENGINEERING JANUARY 24-25, 1990 I
Albuquerque, New Mexico The ACRS Subcommittee on Structural Engineering met on January 24-25, 1990 at the AMFAC Hotel, in Albuquerque, NM, to discuss.
- 1) seismic resistance of concrete containments, 2) results of Sizewell B Containment Test, 3) post-test analysis of concrete containment model, 4) tests of inflatable seals and equipment hatch and 5) behavior of Category 1 structures.
Notice of the meeting was published in the Federal Register on
-January 4, 1990.
Items covered in the meeting and handouts are kept with the office copy.
There were no written or oral statements received or presented from members of the public at the meeting.
E. G.
Igne was Cognizant ACRS staff member for the meeting.
Princioal Atte* bl3 ACRS C.
P. Siess, C.d nu D. Ward, Membet C. Wylie, Membe; J.
Stevenson, ACRS Consultant M. Bender,-ACRS Consultant C. Mark, ACRS Consultant H.EQ '
J.
Costello R. Kenneally ggl Others W. von Riesemann, Sandia National Laboratories L. Lambert, Sandia National Laboratories g
9006110061 900306
.;pJ ORIGINAL PDR ACRS Certified By
[
- ir
\\
a Minutes /ACRS'Subcom. on 2
Structural Engineering 1
January 24-25, 1990 LAlbuquerque, NM
-l i
I. Ludwigsen:
M. Bohn i
B.
Spletter B. Parks D. Clauss D.JHorschel M. Amin, Sargent L Luray R. Dameron, ARAWCH hsearch Corp.
A. Tome, TENERA,.b.P.
P. Davis, PRD Consulting C.
Farrar, Los Alamos National Laboratories Hichlichts:
1.
M. Amin, Sargent & Lundy, presented an analytical study of seismic threat to containment integrity.
This study was performed during 1987-1988 under contract to Sandia National Labs.
The results of the study were documented in NUREG/CR-l 5098, dated July 1989.
The purpose of the study was to determine-the seismic capacity of four different containment in the eastern U.S.
The four containments studied were Fermi, Zion, Clinton and Sequoyah.
These containments were chosen to consider a range of containment (steel, reinforced concrete, prestressed concre.te and ice condenser), reactor (BWR, and PWR) and foundation types (rock and soil).
The study considered 16 limit states including both L
L direct and indirect.
Failure criteria for evaluation of the I
limit states were stated to be credible but conservative.
Four
' combinations of pressure and temperature with seismic loadings H
were considered for each containment studied.
l h
-i Minutes /ACRS Subcom. on 3
Structural Engineering January 24-25,-1990 Albuquerque, NM The calculation of the seismic capacity was done by time history analysis'for a prescribed horizontal peak ground-acceleration, A,.
Limit states were then evaluated using the peak seismic effects-determined at each A,.
The capacities and governing limit states were then determined by identifying the lowest value of A, that satisfies the limit states criteria.
I overall conclusions of the study indicate seismic ruggedness of containments.
The evaluated containments have seismic capacity of it least three times their design SSE.
Many of the governing limit states for Fermi, Clinton and Sequoyah are indirect limit l
states,Li.e., they are not directly related to the containment I
pressure boundary.
This indicates that when examining a containment for seismic " weak links," attention should be given to both direct and indirect limit states.
The interaction of pressure and temperature with seismic influences the seismic capacity of direct limit states.. In some cases, this influence is significant.
The containment seismic capacity could be l
increased or decreased because of this pressure-temperature interactions.
Effects associated with significant straining of concrete elements in compression or with significant beyond-L yield straining of steel elements that could cause a reduction of l-s ductility limit due to initial shock weakening of containments are not considered important.
Basemat uplift initiated somewhere l
l
u.-
- i Minutes /ACRS Subcom. on 4
Structural Engineering January 24-25, 1990 Albuquerque, NM after A, = 0,25g at Fermi, Zion and Sequoyah; due to its large overhang, the Clinton containment does not uplift at A, = 1.09 The effect of basemat uplift is significant for Fermi and Sequoyah containments because of hard impact between the basemat and the rock foundation.
This increases both the basemat evaluation forces and in-structure spectra in the high frequency range.
2.
D. Horschel, Sandia National Laboratories (SNL), discussed the 1:10 scale sizewell B containment tests.
This test was sponsored by the Central Electricity Generating Board (CEGB) and designed by Nuclear Design Associates (Taylor Woodrow and McAlpine).
Construction and testing was done by Taylor Woodrow.
The CEGB objectives of the model containment test are to meet NII requirements and to validate computer analyses used for ultimate load analyses.
The NRC objective is to validate analytical procedures that can be used to predict containment behavior during severe accidents.
Some major features of the 1:10 scale model are as listed below:
o Scale model of Sizewell B, o
Design pressure of 50 psig, o
1:1 tendon replacement, o
Scaled steel areas rather than 1:1 bar replacement
T Minutes /ACRS Subcom. on 5
Structural Engineering January 24-25, 1990 Albuquerque, NM e
o Hydrostatically tested, o
Thickened-flat basemat, o
Micro-concrete, and No liner (used a rubber bladder to contain pressure) o The model containment design started on October 1987 and construction started on July 1988.
A year later testing was started and initial reports of test results were released for comments on October 1989.
Some preliminary results based on transducer readings are as L
follows:
l-l' l
o Basemat uplift was the largest displacement component found, 40 mm; it occurred at a peak pressure of 112 r
- Psig, o
No indication of meridional tendon yielding occurred, 1
o Strains in the dome were generally small, o
Barrel cracking occurred at 92 psig, and o
Local concrete crushing were observed at the outer base of the wall, particularly at the base of the buttresses.
The CEGB is considering modifying / repairing the basemat and
l Minutes /ACRS Subcom. on 6
Structural Engineering January 24-25, 1990 Albuquerque, NM retesting the containment model.
The CEGB is again looking for.
financial support to continue their testing effort.
1 3.
R. Dameron, ANATECH Research Corp., discussed more in detail the results of the-1:10 scale model test of Sizewell B.
He discussed the distinguishing features of the model and itu effect on response and uncertainty in modeling.
Features discussed were the rubber liner, concrete prestressing, basemat/ foundation and the use of micro-concrete in the model.
Pre-test analysis included the identification of potential failure mode and choosing analysis procedures.
In conclusion, R. Dameron stated that satisfactory validation of analysis methods were obtained, especially global response.
In addition, he claimed that analysis methods satisfactorily predicted localized failures i.e., basemat uplift, crushing of concrete at wall-base junctions, elevated rebar strain at equipment hatch and evidence of galling.
He stated that analysis methods were found unsatisfactory in the prediction of rebar rupture at bottom of basemat (>12% strain), dome response and 3D effects.because of limitation on grid fineness.
The following conclusions based on the model test could be stated regarding the behavior of prestressed structures:
e Minutes /ACRS Subcom on 7
Structural Engineering January 24-25, 1990 Albuquerque, NM o
Extrapolation to full scale should be done through analysis only.
o No major surprises were found in-pre-stressed vs.
reinforced concrete behavior.
o Leakage failure mode has not been investigated, and no 1
comparison with reinforced structures can be made.
o Burst failure mode has not been achieved, hence ultimate capacity is yet to be determined.
4.
D. Clauss, SNL, discussed conclusions of a finite element liner tearing modelling analysis of the Sandia 1:6 scale model concrete containment test.
The conclusions of the analyses are as follows:
o The 22-inch liner tear initiated as the result of concentrated shear force transmitted by the stud
- anchors, o
Shear forces developed in the studs as they opposed slip between the liner and concrete occurs near the insert plates.
o The techniques and assumptions used in the plane stress analysis of the liner and anchorage system appear
-. - ~
b e
Minutes /ACRS Subcom. on 8
. Structural Engineering 1
cJanuary 24-25, 1990 e
Albuquerque, NM adequate for estimating the point of tear initiated in the 1:6 scale model.
o Membrane stress in the liner in addition to application of ctuds shear forces results in a liner tearing mode of failure instead of a stud shear mode of failure.
5.
M. Parks, SNL, presented the status of the containment penetration research programs.
The following programs have.been L
completed and documented as shown below:
- I o
Electrical penetration assemblies (NUREG/CR-5334),
o Personnel airlock (NUREG/CR=5118),
o Compression seals and gaskets (NUREG/CR-4944 and 5096)
- and, o
Inflatable seals (NUREG/CR-5394).
Ongoing tests include pressure unseating of equipment hatch.
seals.
A proposed program to test bellows was also discussed.
Results of the inf?atable seals program used to prevent leakage around personnel and escape lock doors for thirteen nuclear power plants (PWR or Mark-II type containments) were presented.
The following are summary of the test results:
J
\\
3 Minutes /ACRS Subcom. on 9
Structural Engineering EJanuary 24-25, 1990 1
Albuquerque, NM o
Regardless of test conditions, significant leakage did not occur until the chamber or containment pressure exceeded the initial seal pressure, then leakage increased rapidly for small increases in chamber pressure.
- t For temperatures up to 350' F, there were no o
indications of degradation of the seal material.
However, between 350' F and 400* F (max. test temp.),
signs of a breakdown in the composite seal material began to occur.
o Test validated methods have been developed to-predict i
the containment pressure for a given inflatable seal pressure and temperature, at which' leakage past the t.
seals can be expected to occur.
[
An analytical method has already been developed to evaluate 1
the leakage potential of pressure tending to unseat the equipment hatches and drywell heads.
Tests are being conducted on one of l'
the equipment hatches in the 1:6 scale model to validate this analytical approach.
parame.ters being varied in this test program include gasket material, aging history, aggregate bolt l'"
preload, and aggregate bolt stiffness. Four ambient temperature j;
a
- ~' "
W V
M T
Minutes /ACRS Subcom. on 10 Structural' Engineering
-January 24-25, 1990 Albuquerque,1NM tests have already been completed.
Preliminary test results are as follows:
o In three of the four tests, significant leakage first occurred when the separation displacement was within-one standard deviation of the mean available springback.
The mean available springback parameter is a reasonably accurate measure of gasket performance.
Leakage is very sensitive to the available springback.
o o
Presently, the method for calculating leak rate significantly overestimates the actual leakage.
Bellows are used primarily in steel containments to minimize piping loads imposed on the containment shell.
It was stated that because world-wide search for applicable bellows test data and analytical methods of bellow performance are not available, containment: penetration bellows cannot be eliminated as a possible mode of failure during a severe accident, and severe accident testing of containment bellow is essential.
It was proposed that severe accident testing of representative containment bellows be performed.
Based on test results, analytical methods can be developed to predict bellows ultimate pressure and deformation capacity.
The subcommittee did not object to the proposed bellows test program.
.. : _: s.;
Minutes /ACRS Subcom. on.
11 Structural Engineering January 24-25, 1990 Albuquerque, NM 6.
M. Parks, SNL, discussed. future plans of the containment integrity programs.
The planned test programs are as follows:
o Perform separate effects test to understand liner tearing mechanisms, o
Retest of 1:6 scale model o
Perform penetration tests on 1:6 scale model in order h
to understand the unseating of equipment hatch bellows.
l M. Parks discussed a potential cooperative agreement with Nuclear Power Engineering Test Center (NUPEC).
Primary NRC interest in this test program would be to obtain additional' data on prestressed containment while NUPEC interest would be primarily in data.on behavior of BWR vessel heads and flanges.
7.
W. von Riesmann, SNL, briefly discussed the assessment of analytical methods that was preparac by D. Clauss.
In summary, the following conclusions were offered:
o Computer codes for calculating structural response are well established, o
The challenge.is in identifying potential failure modes, developing and verifying evaluation criteria and designing consistent models.
m Minutes /ACRS Subcom. on-12 a-Structural-Engineering i
January 24-25, 1990 Albuquerque, NM o
Understanding failure is the most reliable method for.
determining the maximum pressure at which integrity is preserved, and is also useful for risk assessments and other issues associated with severe accidents.
~
8.
C.
Farrar, Los Alamos National _ Laboratories, reviewed the seismic category I structures program.
This program was originally established as part of NRC's Margins to Failure Program whose objective was to. investigate the dynamic response of seismic Category I reinforced concrete structures subjected to seismic loads beyond their design basis. Early testing (FY 1980-l L
1984) employed testing 1:30 scale isolated shear wall structures
.both statically and dynamically.
Next, scale model diesel generator buildings and auxiliary buildings (structures with 1 and 3 inch thick walls) were subjected to simulated seismic inputs.
Early test results indicate that prototype structures are expected to withstand earthquakes in excess of 2g's L
L horizontal ground acceleration.
This implies a significant L
reserve margin.
In addition, early test results indicate that stiffness, measured directly in static tests and determined l
[
indirectly from frequency measurements in dynamic tests, were as much as a factor of four below values that industry would use in the design process.
Scalability between different size I
L l
- 1..
.4 Minutes /ACRS Subcom. on 13 Structural Engineering January 24-25, 1990 Albuquerque, NM microconcrete models was demonstrated in the elastic and inelastic response regions.- The bottom line of the program at the end of FY 1984 indicates that the structures have significant reserve margin despite the reduced stiffness.
?
A technical review group of recognized experts in this-area focused its concerns on the reduced stiffness issue.
The reduced stiffness would, in general, shift the resonant frequency of the structure into the frequency range where an earthquake has its peak energy, causing plant equipment to be designed to an i
inappropriate (unconservative) response spectra if the reduced stiffness is not considered.
Possible sources of reduced stiffness were identified as follows:
o Microconcrete responding in a different manner than conventional concrete, o
Structures damaged prior to testing, and i
o Boundary conditions not properly assumed during the
- test, Results of further model tests (TRG Test Series) indicates that the reduced stiffness does not appear to be as large as initially thought.
Reductions in stiffness of four (from theory) were probably related to damage (cracks) prior to testing and to
l 1
Minutes /ACRS Subcom. on 14 Structural Engineering January 24-25, 1990 Albuquerque, NM incorrectly assumed boundary conditions.
Currently, tests show that stiffness at OBE levels is 70% of theory at worst.
Scalability of microconcrete response to conventional concrete was demonstrated in the elastic range.
No cumulative damage effects were noted.
Currently, two ASCE committee working groups (both groups are part of the Structural Division - Dynamic Analysis Subcommittee l
of the Nuclear Structures and Materials Committee) are involved j
in drafting positions on shear wall stiffness reduction and structural capacity and failure issues.
The working group cut shear wall stiffness recommends that at nominal stress levels l
below 100 psi, the NRC's-response spectra broadening (plus or minus 15 per cent in frequency) will account for reduced 1'
l stiffness.
Above 100 psi, designs should examine two stiffness values:
100% of theory and 50% of theory.
The seismic Category I program is now complete except for documentation.
Special topical reports on stiffness, damping, floor response spectra, etc. and a final summary report will be forthcoming later.
9.
M. Bohn, SNL, discussed the impact of structural response with reduced stiffness (softening) on plant risk applied to the
1
?
Minutes /ACRS Subcom. on 15 Structural Engineering January 24-25, 1990 Albuquerque, NM Peach Bottom (BWR) Nuclear Power Plant.
The program objectives were to assess the impact of decreased natural frequencies of concrete' shear wall structures on deterministic design calculations and an overall seismic plant risk.
The general-approach (best estimate) to assessing probabilistic impact on risk were given as follows:
o Choose existing seismic PRA(s) as base case (s),
o Re-compute structural response with reduced fixed-base natural frequencies using best estimate SSI calculations.
o Re-evaluate capacity of structures with new median loads and uncertainty distributions.
o
.Re-compute floor spectra for critical components, o
Re-evaluate critical component fragilities, o
compute accident sequence probabilistics with new structure and component fragilities and compare with
(
original PRA results.
I Preliminary analyses for the re-wvaluation of the Peach B;ttom nuclear power plant seismic PRA, which was originally performed in the NUREG-1150 program, have just been completed.
For this analysis, complete re-evaluations of the structural responses of I
all critical structures were performed at two different values of stiffness reduction (40% and 80%).
Significant shifts in l
l
1; r
J Minutes /ACRS Subcom. on 16 i
b, Structural Engineering
)
January 24-25, 1990 L
Albuquerque, NM spectral peaks and amplification of the spectral peaks resulted when reduced frequencies were considered.
It was found that the mean core damage frequency (CDF) was increased by about 60% (from 7.66'E-05 to 1.24 E-04 per year).
The increase in CDF is due primarily to the increased values of response in the Circulating Water Pump Structure and the Emergency Cooling Water Towers.
These building responses have been significantly increased over i
the case with no stiffness reduction and significantly affected the Emergency Cooling Water pump and the Emergency Service Water pumps.
These pumps play a critical role in providing cooling to the diesel generators in the event of loss of off-site power.
Further, they are very significant contributors to the base case (NUREG-1150) CDF.
Similar evaluation is being performed on the Zion nuclear power plant which could be more severe because the plant is on a soil foundation.
Results of this investigation will be presented to the subcommittee when completed.
NOTE:
A transcript of the meeting is available at the NRC Public Document Room, Gelman Bldg. 2120 "L"
Street, NW., Washington, D.C. Telephone (202) 634-3383 or can be purchased from Ann Riley &
Associates, LTD., 1612 K St. NW., Suite 300, Washington, D.C. 20006, (202) 293-3950 N.W.,
i-
,