ML20150A665
| ML20150A665 | |
| Person / Time | |
|---|---|
| Site: | Trojan File:Portland General Electric icon.png |
| Issue date: | 09/20/1978 |
| From: | Bresler B, Holley M PORTLAND GENERAL ELECTRIC CO. |
| To: | |
| References | |
| TAC-07551, TAC-08348, TAC-11299, TAC-7551, TAC-8348, NUDOCS 7809290071 | |
| Download: ML20150A665 (16) | |
Text
_ _
i RESPONSE OF TROJAN NUCLEAR POWER PU.NT CONTROL BUILDING TO SPECIFIED SSE EVENT September 20, 1978 Hyle J. Ilolley, Jr.*
and Boris Bresler**
- 1. INTRODUCTION The writers were engaged by the firm of Lovenstein, Newman, Reis
& Axelred, outside counsel for Portland General Electric Company, on August 4, 1978, to assist in determining the capability of the Control Building of the Trojan Nuclear Power Plant to resist the specified SSE event without any structural consequences which could interfere with a safe shutdown.
In the course of studying this question over the past six weeks, the writers had many extensive discussions with engineers of the Bechtel Corporation, both at their San Francisco offices and at the offices of the Portland General Electric Company. PCE personnel participated in many of these discussions. Familiarization of the writers with the subject structure, with methods of analysis being used by the engineers to evaluate its capacity, and with pertinent results obtained from those analyses, was facilitated by the coeplete cooperation extended to them by the engineers of both organizations.
(C 7Sp?Zpyj7/ x
- 2-
~ T incipal, anIcn, Holley 6 Bl g9 ts, Inc., Catbridge, Mass.
- Principal, Wiss, Janney, Elstner & Associates, Inc., Northbrook, Ill.
I l
F The writers visited the Trojan Plant and toured the Control l
Building, personally observing its physical layout and construction :
I features. This provide'd a useful supplement to information obtained by i l
examination of the relevant design drawings and thrcugh discussions with the i engineers. [
f The writers have discussed the problem on many occasions during i the past six weeks, apart from the above-mentioned larger meetings. '-
M. J. Holley also has discussed certain aspects of the problem with his ,
i professional associates Dr. Robert J. Hansen and Mr. John M. Biggs, and i
) B. Bresler discussed the problem of wall capacities with his associate, i j
, Mr. Craig Comartin. ;
i It should be noted that the writers' efforts have not included I i
j any attempt to check the accuracy of the analyses executed by engineers of l i the Bechtel Corporation. Such checks could not have been accomplished in l
the time available. Rather, the writers focused on methods of analysis i l
used in the evaluation and upon those results of analyses with which they were j provided, j t
The report which follows is comprised of discussion of relt.vant l I
aspects of the evaluation and, finally, the writers' Suennary and Conclusions. t L
As noted above, in the course of this investigation the writers j discussed all aspects of the evaluation and are in full aCreement on the !
l results of the evaluation. In preparing this report, however, M. J. Holley [
is primarily responsible for sections dealing with dynamic analysis, ;
slab-to-vall forces, non-linear response, damping considerations, and
.2-t i
1 displacement considerations; B. Bresler is primarily responsible for the section dealing with shear wall capacities.
- 2. METHODS OF DYNAMIC ANALYSIS OT THE CONTROL-AUXILIARY-TUEL BUILDING COMPLEX of the three different models used for dynamic analyses (sticks.
TARS. STARDYNE) only STARDYNE, based on finite element modeling of the structural components, accounts directly for all aspects of the distributed stiffness and distributed masses. Therefore, the STARDYNE analysis provides the most reliable prediction of the principal mode shap e Turther. since tho STARDYNE analyses were for linear elastic response based on maximum l possible funeracked) stiffnesses, they predict useful upper-bound estimates of internal forces for the principal structural components. This follows from the fact that the fundamental period (0.15 sec) is in a region of the spectrum for which any increases in period (due to cracking-reduced stiffnesses) do not increase the accelerations.
- 3. SHEAR WALL CAPACITIES Capacity Criteria.
The earthquake resistant shear valls in the Control and Auxiliary buildings fall into three categories:
- a. Masonry block walls which "sandwich" structural steel framing (columns) and a reinforced concrete core,
- b. Masonry block walls which "sandwich" a reinforced concrete core only, and 3
l i
l l
r' i
- c. Two-vythe concrete block walls without a concrete core. t Structural. behavior of "sandwich" valls (strength and load-deforma-tion characteristics) can be determined approximately from behavior of [
reinforced concrete or reinforced masonry walls, as there is no experimental j i
data on such walls. j Capacity criteria were established for the following principal ;
modes of failure !
- a. shear-compression (diagonal) failure , i
- b. flexure failure !
(
- c. shear-friction or devel action failure The lovest value is used as the capacity of a given vall.
The shear-comoression criterion has been derived by Bechtel engineers (1,2) from their evaluation of tests on reinforced casonry walls. While the test specimens do not simulate the actual vall construction encountered in the subject building, there is no other data that would better simulate this type of construction.
The interpretation of the datt in deriving the Bechtel "Basic Criteria" appears to be conservative, particula.ly as the test specimens use concrete block, mortar, i
and grout having lower e.rength values than those used in the subject building.
Also, while ,ata on the effect of cyclie loading is limited, test l
results reported by Mayes, et. al show that cyclic loading which does not l enceed a maximum shear strain of about 0.006, does not lead to redu: tion of l
strength below about 210 psi. The maximum shear stresses based on calculated l
l forces for the N-S valls all fall Selow these values.
4
- 1. Schneider, R. R., "Shear in Concrete Masonry Piers", California State Polytechnic College, Pousna, 1969.
- 2. Mayes, R. L., Clough, R. W., Hidalgo, P. L., and McNiven, H. D.."Seismic Research on Multistory Masonry Buildings". North American Masonry Conference, Boulder, Cc., August 14-16, 197S.
ta
.. .. l l
E r
l The flexure capaetty criterion is based on generally accepted l and documented formulations of ultimate moment capacity. The shear frictier !
I or dowel action is used to evaluate vertical shear transfer capacity j from the side walls to the end walls, f I
Haximum Forces on Shear Walls. ,
As noted above. STARDYNE analyses predicted upper-bound forces to be expected for linear-elastic response, assuming unlimited shear wall :
capacities. In a few of the smaller shear walls, that is, walls which contribute only a small portaou of the total story shear resistance, these
{
i predicted forces exceed the computed wall capacities. Also, the predicted !
I maximum shear force in one of the masonry walls (the West wall) exceeds the j computed (dowel) capacity. This implies that, for these particular shear t walls, maximum shear forces may equal the computed shear capacities. It
[
further implies that the other walls may experienee maximum shear forces ;
which are slightly greater than the STARDYNE elastic response values, as [
f they accept the small excesses of predicted shear forces above the capacities !
of the cited walls.
In summary, the STARDYNE-predicted wall shear forces modified as ,
I described above, represent conservative maxima, g i
Evaluation of Shear Valls. [
The cv .tical shear walls in the Trojan Control and Auxiliary i
buildings under LSE event are the N-S walls at elevations 45 and 61. Therefore, the following remarks are addressed to these walls, and while some other walls (at other elevations) may also be highly stressed, the considerations 5-l I
--m___
I described below voM d apply to the other valls as well. f In evaluating performance of these walls under a SSE event, the following approach was used: [
I Tor each wall element strength (capacity) has been calculated t i
using criteria discussed above. These capacities were compared with the f I
forces calculated using STARDYNE linear elastic analysis. As noted above, i redistribution of the forces, which would take place when some of the valls !
[
reach thea. apacities, results in somewhat increased shear forces in those j I
valls which have substantial reserve capacities. Tht; sum of the computed i capacities of these shear valls is well above the sue of the predicted maximum f
I shear forces. Consequently, the group of walls resisting a given story shear j t
can withstand forces imposed during a SSE event without interfering with the ,
safe shut down. ;
I,
- i. . SLAB y0RCES ANI) 51).B-TO-k'ALL. PORCES i
l \
1 The focus of this report is primarily upon the behavior of the !
masonry shear valls. It is obvious that the capacities of the floor and roof slabs to resist the in-plane forces imposed by the SSE condition is I
equally important. However, these components are more conventional than I the shear walls, and evaluation ef their capacities is sonevhat more i
straight forward. The STARDYNE-predicted forces in the slabs represent reasonable upper-bound estimates. It is the writers' understanding that the slab capabilities for resisting these forces have been thoroughly investigated.
~6- ,
1 l
t
It is further understood that these investigations have shown the capacities of the slabs, and the slab-to-vall connections, are adequato to resist the STARDYNE-predicted forces.
- 5. NON-LINEAR SHEAR k'ALL RESP 0NSE As noted above, the STARDYNE linear analysis, based on maximum (uneracked concrete) stiffnesses provides an adequate crediction of the maximum force imposed on the shear wall group and on the slabs by the SSE condition. However, some of the walls are subjected to shear loads equal to their computed capacities, and the other walls are subjected to shear loads equal to substantial fractions of their computed capacities. This implies that the valin vill be distinctly non-tinear in their response.
l Under cycled loading their no-load stif fnc6.e4 eill be less than the initial no-load (uneracked) stiffnesses used in the STARDYNE analyses. Moreover, l
within each loading cycle the stiffness must decrease with increasing load.
l The slabs vill experience some cra. king and this cracking will reduce slab stif fnesses below (ancracked) stif fnesses assuned in the STARDYNE analyses. Hewever, the effects of cracking on elab stiffnesses should be much less than the corresponding effects in the masonry walls. For this reason, the non-linearity of overall structural response is dominated by the wall non-Itnear responses.
The engineets have recognized that the STARDYNE analyses underestimate displacements, for the above described reasons. They have estimated a
l i
maximum displacement (at roof level of the control building) of about 0 9",
in contrast to the corresponding STARDYNE-predicted displacement of 0.15".
Their estimate is based op shear strains developed in shear wall test i
specimens, when such specimens reached their shear capacities.
The writers have used a different method to estimate maximum 4
displacements. This involved approximating the non-linear force-displacement function by a linear system of greatly-reduced stiffness. If the structure with this reduced linear stiffness characteristic is assumed capable of developing the shear resisting forces predicted by STARDYNE the corresponding increased displacements can readily be predicted from the specified SSE response spectrum.
By this approach, an assumed stiffness reduction to about one-fifth of the stiffness used in the STARDYNE analysis would lead to the same six-fold increase in displacements implied by the engineers' displacement estimate.
This 80 percent reduction in linear stiffness appears to the writers to be a conservative representation of thc degraded non-linear stiffness of the actual walls subjected to cycled loading. It may be noted that for a 90, percent stiffness reduction (1. e., effective stiffness only one-tenth of that used in the STARDYNE analysis) would lead to an eight-fold increase in the STARDYNE-predicted displacement, that is, a maximum displacement of about 1.0 inch.
To further illustrate the inherently-limited nature of the maximun displacement the writersmodified the above-described reduced-linear-stif fness analysis as follows. The reduced-linear stiffness model was assumed to yield at a force level only 50 p.rcent of the STARDYNE-predicted imposed force.
In the range of the response spectrum of interest this would lead to only a 8
25, percent additional increase in the maximum displacement.
In summa y, it is the writers' judgment that, based on the STARDYNE-predicted displacement of 0.15", the actual maximum cisplacement, l
at roof level, should be little, if any, in excess of 1.0 inch. It is )
obvious that the corresponding maximum floor-to-floor relative displacements would be less than this value. ,
1 1
l l
- 6. DAMPING CONSIDERATIONS i 1
In all their analyses of forces in the structural components, and displacements, the engineers have assumed 5, percent damping for responses to the SSE condition. The writers similarly have assumed only 5 percent dampir . This is reasonabic for the reinforced concrete floor slabs for which it is reported that in-plane capacities have been found ec be well in excess of the imposed forces. However, for the masonry shear walls, subjected to substantial fractions of their computed capacities, substantially larger damping percentage would be appropriate. Such larger damping would lead to smaller predicted forces and displacements. The use of these, more realistic, larger damping percentages is not pernitted by the NRC, but they represent an un-accounted-for conservatism in all the analyses.
l l
- 7. DISPLACEMENT-DEP?NDENT CONSIDERATIONS I
l it should be emphasized that judgments regarding the capacity of the Control Building to resist the SSE event without impairment of its safety
_9
functions are almost entirely dependent on the anticipated maximum displacements. Thus, magnitudes of anticipated internal forces are not nearly as important in themselves as in the basis they provide for evaluating the maximum horizontal displacements. This would not be the case in an evaluation of the capacity of the strue.ure to resist a sustained, static, horizontal loading. For response to earthquake ground motion, a range of assumed strengths all may lead to displacements which are acceptable.
It is the writers' understanding that the structural steel framework of the. Control Building was designed to carry the entire gravity loading of the building. This is of considerable importance. It implies that collapse of the building, or any of its slabs, could only occur if the anticipated horizontal displacements were extremelv large; i. e., large enough to totally eliminate lateral stability afforded by the walls, or large enough to fail a beam-to-column connection. Neither of these effects could be realized unless the maximum horizontal displacements were many times as large as the maximum displacements that actually are anticipated. In short, co11 arse of the building, or one of its slabs, does not appear to be a possib12 consequence of the specified SSE event.
More realistic displacement-related considerations which required i address are:
- a. capability of equipment to accommodate relative displacements (e. g., floor-to-floor, or building-to-building) without loss ;
et function.
i i
L
- b. implications of Control Building displacements on the forces transmitted to the hold-up tank enclosure and to the spent fuel pool structure.
- c. possible contacts of the Control Building with craneway columns of the Turbine Building as these two buildings respond, independently, to the SSE event.
It is the writers' understanding that the capabilities of all safety-related components ( e. g., equipment, piping, electrical runs) to accept large relative displacements have been determined by an exhaustive on-site survey. This survey, executed by a team of engineers skilled in the appropriate disciplines, and including supplementary analyses where required, is reported to have shovn that relative displacements in excess of those anticipated can be accommodated by all safety-related components.
1he survey is also reported to have disclosed that possible impingement of a piece of spalled masonry block on any of the components, (however unlikely), would not cause such component to fail to perform its in*. ended function.
The engineers' analyses show that the largest horizontal displacements of the Control Building, in response to the specified SSE event, would occur in the N-S direction. The STARDYNE analysis for this condition leads to forces in the reinforced concrete structural components of the hold-up tank enclosure and spent fuel pool (connected, by floor and roof slabs, tnrough the Auxiliary Building to the Control Building) which are well below the capacities of these structural components. However, as has been noted, the
STARDYNE analyses do not account for the increased flexibility of the Control Building, arising from the reduced stiffness of the masonry walls under cycled loadings to large fractions of their calculated capacities.
The engineers have evaluated the maximum forces which can be imposed on the hold-up tank enclosure and spent fuel pool by the slabs.
These imposed forces are limited by the in-plane bending moment capacities of the slabs at their junctures with the fuel building. The slab moment capacities are limited by yielding of their E-W veinforcing steel. The results of these analyses are reported to demonstrate that the maximi;m forces that can be imposed on the hold-up tank enclosure and the spent fuel pool, regardless of'the magnitude of N-S displacements of the Control Building, are well below the capacities of these fuel building structures.
At roof elevation of the Control bu.4 ding, where the anticipated N-S horizontal displacements are maximum, a 3" clearance has been provided between the Control Building and the craneway colu=nc of the Turbine Building.
The engineers have executed an updated analysis of the Turbine Building response to the SSE event. This analysis is reported to predict a maximum N-S displacement of the Turbine Building columns of 1.5". Since, as noted in an earlier Section of this report, the maximum anticipated Control Building deflection is, in the writers' judgment, only about 1.0" (0.9" by the engineers' estimate), the worst possible relative displacement of the two buildings, 2.5",
is within the clearance, 3.0", which has been provided. However, it is conservative to base a prediction of the maximum building-to-building relative displacement by direct addition of their individual maximum displacements.
If, for example, the individual building displacements were combined on an SSRS basis, or if all the modal contributions to the displacements of the two buildings were combined on an SRSS basis, the implied clearance margin would be greater. In summary, it appears that contact between tha two buildings is not to be anticipated.
- 8. StM1ARY AND CONCLUSIONS
- a. Because of the presence of a steel frame designed to support all gravity loads and becaese of the small magnitude of displacements, neither collapse of the Control Buildings nor collapse of any of the structural members supporting its floors ,
can occur in the specified SSE event.
- b. In judging the ability of the Control Building to resist the specified SSE event without consequences which might interfere with safe shutdown, predicted maximum displacements are the aspect of response of paramount significance. Coc.puted internal force 3 are of significance only to the extent that they form a basis for judging the displacements, u-
- c. A necessary first step in the process of computing displacement maxima is a dynamic analysis of response in the linear elastic r-range. The STARDYNE program is an effective tool for this purpose because it permits a good representation of the mass distributions and stiffnesses throughout the structure. The designers properly utilized the STARDYNE analytical results ,
rather than those obtained from the stick-models or TABS analysis,
- d. Because of the region of the response spectrum in which the significant natural frequencies are located, be ause of the linear-elastic behavior assumed, and because of the conservative damping ratio used, the STARDYNE determined story shears (for the group of resisting walls) represent conservative, upper-bound, values of these group forces. It is significant to note that there is a conservative margin between these group forces and the corresponding group strengths of the resisting shear walls,
- e. The individoal shear wall forces in some walls, particularly the smaller walls.have computed capacities less than the STARDYNE predicted values of imposed forces. This simply nicans that forces in these particular walls may reach their computed capacities, and that forces in other, less severely loaded, I walls may develop forces larger than those predicted by the l STARDYNE linear-elastic analysis. However, re-distribution analyses by l the Bechtel engineers show that the other major walls would L
l s - - _ _ - _ _ - - _ - - - - - - _ _ _ - _ _ _ _
m develop forces which are still well below their capacities.
Consequently, the group of walls resisting a given story shear can withstand forces imposed during a SSE event without interfering with safe shut-down.
- f. Based upon wall shear strains at their capacity leads, reflecting the non-linear response of shear wall test specimens under cycled loading, the engineers have estimated a maximum horizontal deflection of 0.9". This deflection (at roof level, at the west end of the Control Building) is six times the displacement outputted by the STARDYNE analysis. The latter estimate assumed stiffnesses based en uncracked concrete, i. e., stiffnesses which are representative at very low levels of loading,
- g. By an independent approach, approximating the non-linear characteristics of shear walls with reduced-stiffness linear characteristics, the writers satisfied themselves that Bechtel engineers' estimate of maximum horizontal displacement is conservative. However, for purposes of evaluating the consequences of displacements, they prefer the slightly more conservative conclusion that the maximum deflection will not exceed about 1.0".
^
F. .
- h. Relative displacements (e. g. floor-to-floor) will, of course, be less than the maximum value cited above. However, a survey carried out by Bechtel engineers has disclosed that all safety-related components can accommodate relative displacements well in excess of 1.0" without loss of function. The floor slabs which connect Control Building. Auxiliary Buildings, and Fuel Building are limited in their capacity to impose forces on the Hold-Up Tank enclosure and Spent Fuel structures.
Analyses by the Bechtel engineers show that these limited forces, which are independent of the magnitude of Control Building displacements, are well within the capacities of the structures in the Fuel Building.
- i. Up-dated dynamic analyses of the Turbine Building response to the SSE event show that the displacements of this structure combined with Control Building displacements are not sufficient to cause contact between the two buildings.
- j. From all the above it is the writers' j udgment that the Trojan Control Building, in its as-built condition, can withstand the specified SSE event with no conscquences that could in:erfere with safe shutdown.