ML19296D743
| ML19296D743 | |
| Person / Time | |
|---|---|
| Site: | Trojan File:Portland General Electric icon.png |
| Issue date: | 02/22/1980 |
| From: | Office of Nuclear Reactor Regulation |
| To: | |
| Shared Package | |
| ML19296D739 | List: |
| References | |
| TAC-12369, NUDOCS 8003130098 | |
| Download: ML19296D743 (12) | |
Text
C.
TROJAN NUCLEAR PLANT CONCRETE MASONRY WALL DESIGN CRITERIA FOR fRANSVERSE LOADINGS SAFETY EVALUATION REPORT BACKGROUND On October 19, 1979, the NRC staff was informed by Portland General Electric Company (PGE) of a problem with respect to seismic anchor SA-83 in the spent pool cooling system.
In conducting the analyses, the inspections and the surveys required by IE Bulletins 79-02 and 79-14, PGE found that SA-83 did not meet the acceptance criteria of Bulletin 79-02, and, in addition, the concrete block wall to which SA-83 was attached was structurally inadequate to resist the piping reactions that would result from an earthquake, per the seismic criteria specified for the Trojan site.
This information led PGE to an investigation as to how this occurred, and the extent of this problem throughout the plant.
PGE filed additional information on this subject in Licensee Event Report 79-15 dated November 4,1979, and Supplements 1 and 2 dated November 19 and December 4,1979, respectively.
Additional information requested by the NRC staff, resulting from a 2-day meeting on December 5 and 6, 1979, among the NRC, Bechtel, and PGE, was furnished in a letter from PGE dated December 13, 1979.
In response to concerns on this matter raised in a December 21, 1979 affidavit of Herring and Trammell of the NRC regarding the Trojan Control Building hearing proceeding, PGE supplied additional information to the NRC on this matter in a letter dated December 22, 1979.
Following intensive discussions among the NRC, Bechtel, and PGE on December 29 through 31, 1979, further information was provided to the NRC by PGE in a letter dated December 31, 1979.
The initial scope of LER 79-15 was limited to addressing the design criteria for single and double wythe masenry walls supporting safety-related piping and equipment imparting significant support reactions to the walls, i.e., excluding piping with a diameter of 2 inches or less which do not produce thermal loads and whose support reactions are generally less than 100 pounds, and equipment loads less than 100 pounds.
Composite
- walls were included at a later time.
However, review of the information submitted regarding these walls led to the scope being expanded to include the evaluation of the design criteria for
- " Composite" means a wall with a central concrete core, sandwiched by concrete block.
S0033800T
. transverse loading of single and double wythe and composite masonry walls which do not support safety-related piping and equipment, including shear walls.
DISCUSSION The problem initially identified with respect to seismic restraint SA-83 was reported by PGE to be attributable to error in engineering judgment, a lack of procedures or procedural detail and insufficient design criteria (1971-72) with respect to consideration of external loads on block walls (see supplement 2 to LER 79-15).
The concern that many structural elements may not have been designed to with-stand the required piping and equipment support reactions was addressed by PGE as follows:
All single and mortared double wythe and composite masonry walls carrying a.
significant piping and equipment support reactions were re-evaluated using the crieria described in LER 79-15, its supplements, and additional information submittals.
The re evaluation resulted in 39 cases of through-bolting and 88 cases of modifying (removal, enlargement of base plate er addition of a strut) existing supports on single and mortared double wythe masonry walls, and the removal of eight supports for the hydrogen vent system from a composite wall.
b.
A field survey of the entire plant was conducted to identify those concrete walls, floor slabs, and structural steei members supporting safety-related piping and equipment
- which could be highly loaded relative to their capacities.
Initial determinations were made as to what condition could lead to a potential overstress in light of the initial design and considering support and structural member configurations.
No concrete walls were found which required further investigation.
Five supports involving structural steel members and nine supports involving concrete floor slabs required further evaluations.
The results of these evaluations showed that of these 14 conditions, 12 were adequate in their existing configura-tions. One (1) support reviewed required supporting the top and bottom flanges of a steel beam to add torsional resistance and the remaining one (1) support reviewed carrying primarily thermal load and attached to a concrete floor slab required the addition of a column and a horizontal restraint to the pipe.
The evaluation criteria were in accordance with the Trojan FSAR criteria for concrete and steel structures.
- Piping supports were primarily considered although some consideration was given to equipment supports.
. Also addressed was the concern that the wall should be evaluated in accordance with the following load equation, in addition to those considered by PGE, from the Trojan FSAR design criteria for concrete structures (Section 3.8.1.3.2.1),
as referenced by the Trojan Technical Specifications:
"For structural elements carrying mainly earthquake forces, such as equipment supports:
U = 1.00 + 1.0L + 1.8E + 1.0T + 1.25H
- g g
In this regard, reference was made by PGE to a paper dated August 11, 1970, entitled, " Seismic Design Criteria for Nuclear Power Plants," written for the AEC.
The only reference made to a 1.8 load factor for OBE loads is for the design of concrete equipment pedestals.
Therefore, PGE contended that the intent of the statement in the FSAR was that this load equation be applied to only equipment pedestal supports, not all equipment supports.
The design of concrete Category I equipment supports at Trojan were reviewed and only six supports were found where this load equation would be applicable; i.e., only six concrete pedestals.
These are supports for the containment spray pumps, the RHR pumps, and the RHR heat exchangers.
They all satisfy this load combi-nation and corresponding acceptance criteria.
In response to NRC staff concerns over the adequacy of the criteria used to design and evaluate single and mortared double wythe masonry and composite walls resulting from the review of LER 79-15 and its Supplements 1 and 2, additional information was supplied by PGE in the December 13, 22, and 31, 1979 submittals.
These concerns were (1) the acceptability of using a factored load approach for the design of masonry which was less conservative than the approach for the working stress design of concrete for plants of the same vintage as Trojan, (2) the transfer of local stpport reactions into the walls, and (3) the ability of the in-situ walls to behave compositely as assumed, so as not to significantly impair their transverse (out-of plane) resistance, whether or not they supported safety-related piping and equipment.
In addition to the lateral seismic loadings, lateral loadings due to tornado and portulated pipe rupture, in accordance with Trojan's specifications of these, were also considered.
The basis presented by PGE to substantiate the use of a factor of 1.5 increase in working stress allowables as an acceptance criteria for factored load combinations was based upon a ratio of ACI 318-63 (reinforced concrete code) ultimate strength to working stress allowables (see November 4,1979, LER 79-15).
However, the August 11, 1970 paper written for the AEC, and refer-enced by PGE to support the 1.8 load factor applied only for concrete pedestal supports, indicated that the acceptable method for dealing with working stress design of concrete per ACI 318-63 for earthquake loads was more conservative than the factored load approach they were using for masonry design.
As discussed later in the staff's evaluation, the AEC acceptability
" Note:
E stands for OBE loads.
. of this criteria was further verified by the NRC staff.
Concerns over this criteria were further aggravated by the higher than intended in plane stress levels in the Contrcl/ Auxiliary / Fuel Building Complex walls.
For example, a comparison of the criteria of substituting 1.5 times the allowable UBC working stresses for the ultimate strength in the Trojan factored loaj combinations for concrete to the working st ss methodology acceptable per AEC criteria is made as follows for the Safe _ atdown Earthquake (SSE):
Trojan Criteria:
- 1. 5 UBC = 1. 00 + 1. 0 L + 1. 0 E ' + 1. 0T + 1.25H + 1.0R*
g g
- 1. 5 UB C = 1. 00 + 1. 0 L + 1. 0 E ' + 1. 0TA + 1.0HA + 1.0R*
These two equations are restated in terms of the AEC criteria as follows:
1.33 UBC = 0.890 + 0.89L + 0.89E' + 0.89T
+ 1.11H + 0.89R g
g
- 1. 33 UBC = 0. 89D + 0. 89L + 0. 89E ' + 0. 89TA + 0.89HA + 0.89R In the December 31, 1979 submitt31, it was verified that all masonry and composite walls supporting signiricant safety-related piping and equipment meet the following criteria for earthquake load combinations out-of plane:
0+E+T +Hg 5 1.0 UBC g
0+E'+T +Hg 5 1.33 UBC g
These are in conformance with the early AEC criteria for dealing with working stress concrete design.
Addit onally, approximately 80 to 90 percent of the masonry walls not supporting significant safety-related piping and equipment can also meet these for out-of plane seismic loadings. All other masonry walls for out of plane seismic and tornado loadings meet the following criteria:
0+E+1.25H +Tg 5 1.33 UBC g
0+E '+1. 25H +Tg 5 1.50 UBC g
D+W '+ fg 5 1.50 UBC Composite wa; 4 not supporting significant safety-related piping and equipment for out-of plane loads meet the ultimate strength design criteria described in Attachment 2-2 of the December 31, 1979 letter from PGE.
For these walls it was verified that all capacities were governed by the bending steel reaching its yield strength.
Also, at these levels of stress in the steel, the cor-responding stress in the compressive zone is well below a third of the com-pressive strength of the wall materials.
^ 0 = Dead Load, L = Live Load, E' = SSE Load, T = Normal Condition Thermal Load T
Accident Condition Thermal Load, H = Reactor Load due to Thermal Expansion of=Pipingundernormaloperation.
H Reactor Load due to Thermal Expansion of Piping under accident condition.
R g = Force or pressure on structure due to rupture of any one pipe.
W' = tornado load.
% The criteria for in plane as well as out-of plane loadings for all walls are summarized in the response to question 2 in the December 31, 1979 PGE letter.
Other details are contained in LER 79-15, including Supplements 1 and 2, and the other associated submittals.
Local loads from support reactions were considered in accordance with the criteria specified in Supplements 1 and 2 to LER 79-15.
Local block pull out was addressed in the PGE December 13, 7.979 response to question 5, along with a discussion of the acceptability of the ef fects of any cracking on concrete expansion anchor bolt capacities.
Given the limited shear friction capacities f rom the ties connecting multiple wythe walls, both mortared double block and composite, concerns arose as to the adequacy of the assumption of composite behavior for the walls.
- Also, without through bolting of supports where ;oncrete expansion anchor bolts were used to attach the supports to the wall, it was necessary to transmit tensile forces across the interface through the reliance upon tensile bond between the mortar and block, and concrete and block.
For the mortared double block masonry walls, if one wythe alone could not withstand the gross loadings of itself plus the support loading, the support was through bolted. Where groups of supports were close enough to interact in the wall loading, they were all considered together.
For the determination of wall natural frequency, composite behavior and pinned pinned end conditions were assumed, except for partial height walls where actual boundary conditions were used.
The double block walls are generally precluded from having clusters of significant support reactions on a wall.
Most walls have less than 10 supports attached, and supports are generally well dispersed.
The composite behavior of the mcrtared double block walls requires vertical shear transfer at the collar joints; i.e., the vertical mortar joint between the two block wythes.
The UBC does not address the design of this joint, the
- 3 bars at 4 feet center-to center both horizontally and vertically do not provide significant resistance to shear, and only limited testing has been performed on this type of joint.
PGE presented information in its December 22, 1979 letter which showed that mortar / block bond could be relied upon to resist at least 50 psi shear on its net area.
Additionally, in-situ tensile testing of this interface showed between about 25 and 160 psi tensile resistance on the net area of the mortar.
In-situ testing was also performed on the composite walls.
On these, PGE was relying upon tension transfer for reactions from supports attached with anchor bolts, as well as shear transfer on the interface between the concrete core and the block.
As a result of the testing, PGE proposed the use of a 40 psi principal stress allowable for all loading combinations for considering the effects of out-of plane loads.
For the determination of wall natural frequency composite behavior and pinned pinned end conditions were assumed; except for partial height walls where actual boundary conditions were used.
. Maximum stresses at the collar joint of the mortared double block walls and at the interface between the core concrete and block on the composite walls are reported in the December 4, 13, 22 and 31, 1979 PGE submittals.
A final major area of concern was that the collar joints of the mortared double block masonry walls was not filled properly by the masons during construction.
The most suspect walls were constructed during a time period within which some contractual difficulties existed between PGE and the contractor.
These were drilled to determine the extent of mortar fill and any voids found were grouted.
For all walls, review of drilling operations led PGE to conclude that about 80% mortar fill was reliably present in all the walls.
Additionally, an expedited program is in progress to determine the extent of mortar fill on any mortared double block wall where collar joint shear stresses exceed 7.E0 psi.
Finally, PGE provided information to substantiate that the inspection require-ments for the wall construction was such that the provisions for the "special inspection" allowable stresses from the UBC could be used.
EVALUATION We have reviewed the information presented by PGE in LER 79-15, including Supplements 1 and 2, and the additional information contained in the PGE submittals of December 13, 22 and 31, 1979.
The following is our evaluation of this material.
The field surveys conducted for concrete walls, floor slabs and structural steel elements, the investigations performed prior to these field surveys, and the limited corrective actions with regard to these elements resulting from these investigations provide reasonable assurance that the underdesign of these structural elements supporting safety-related piping and equipment in the plant is not a problem and they are capable of withstanding the required support reactions.
Detailed re-evaluations were made of all the significant piping and equipment supports on single and mortared double block wythe masonry walls and composite walls. This effort provides assurance that these elements would not be affected by a lack of interdisciplinary coordination in the initial phases of plant design.
In addition to the review of the paper dated August 11, 1970, entitled, " Seismic Design Criteria for Nuclear Power Plants,"* review of a November 21, 1969 letter from Roland L. Sharpe of John A. Blume and Associates to Dr. Peter A.
Morris, Director, Division of Reactor Licensing, AEC, and conversations with Robert E. Shewmaker currently in the Office of Inspection and Enforcement, who was resposible for the initial AEC structural review of the Trojan Nuclear Plant, led to the determination that the subject equation with the 1.8 load factor on the OBE loads stated in the Trojan FSAR as being applicable to concrete structural elements carrying mainly earthquake loads, such as equip-ment supports, should be applied only to equipment pedestal supports.
[ Attached to a August 27, 1970 letter from Edson G. Case, Director, Division of Reactor Standards, AEC to Agbabian - Jacobsen Associated, J. A. Blume and Assoc.,
Eng., ard N. M. Newmark Consulting Serv.]
. After review of the above documents with regard to the 1.8 load factor on the OBE loads for concrete equipment pedestal supports and further conversations with Robert E. Shewmaker, coupled with the higher in plane stress (load) levels being induced in the walls spanning between building elevations in the Control / Auxiliary / Fuel Building Complex due to the Control Building design deficiencies, concern was raised that the approach of using 1.5 times the UBC working stress allowables as the acceptance criteria in the factored concrete load equations, may not have the intended margins associated with the design of the walls to resist cut-of plane seismic loadings for a plant of that vintage.
After review of all the documentation presented by PGE, the criteria which the various walls can meet, the stress levels in the most highly loaded walls for the SSE and tornado loading conditions, and the behavior of the composite walls where ultimate strength design was used, it is concluded that there is reasonable assurance there is sufficient margin inherent, although not quantifiable, in the overall design of the existing structure to permit operatioa of Trojan until the margins inherent in the design criteria used for the evaluation of the walls to withstand out-of-plane loads are quantified by testing.
Since out-of plane tornado loads affect a limited number of walls and the gross lateral loading for the Control / Auxiliary / Fuel Building Complex is less than that due to the SSE earthquake, the higher allowable of 1.5 UBC for the tornado load combination seems reasonable.
[For example, the gross lateral loading on the Control Building from tornado is only about 25% of the
.25 g SSE loads per the May 24, 1978 PGE supplement to LER 78-13.]
With respect to local wall effects from pipe and equipment support reactions, we conclude that there is reasonable assurance that these types of wall failures are realistically precluded.
In addressing block pullout, the maximum force exerted on one block will occur when four bolts at a spacing of 6", act simul-taneously in tension on one block at their design value.
Resistance to pullout will be from friction along the bed joints, head joints and although some resistance is obtained from direct tension exerted on the backside of the block if the wall is multiple wythe, no reliance will be placed on this.
For blocks in composite and block masonry walls, reliance is placed on a shear resistance of 40 psi in the head joint where cell grout contacts adjacent block and no shear value for the mortar is taken.
The 40 psi value, being derived from tests addressing another concern, was obtained as a result of in-situ testing of the concrete block - core interface of composite walls in the complex at Trojan.
Cores were drilled and a tensile force applied to determine bond strength at the masonry concrete core interface.
PGE claims that valid results ranged from 143 psi to 236 psi with an average of 194 psi.
If all results are considered, a lower bound average becomes approximately 130 psi.
Thus using 40 psi as a principal ultimate stress applies a factor of safety of about 5 to their average and about 3 1/4 to the lower bound average.
Even considering the wide dispersion of values, 40 psi is less than the lowest test value except for one where the test strength was reported as 36 psi.
To attempt to account for shear-tension interaction where anchor bolt loads must be transmitted into the entire wall, 40 psi is taken as a principal ultimate stress at the interface.
Due to the existence of a reasonable factor of safety (3 1/4-5), the fact that this value is at about the lower bound of all
8-tests, and the fact that it was based on in-situ testing, the value of 40 psi ultimate principal stress seems reasonable.
Additionally, to resist block pullout, a value of 25 psi shear for grouted masonry from the UBC code is used in the mortar bed joints and a UBC value of shear strength in the cell grout infill.
As an example, a factor of safety of 3 was calculated by PGE against block pullout for this loading condition.
The validity of this calcula-tion is questionable, and considered invalid, since gross and local loading effects were neglected.
However, noting that 4 bolts all in tension is an extreme condition, one that PGE acesn't believe exists at the plant, a more realistic condition would be where 2 bolts are in tension and 2 in compression to resist a moment.
Also, testing from Testing Engineers Inc., San Diego, CA on push-out of a block with only core grout, no mortar or steel, indicates that this should not be a concern given the small crack size that may develop in the bed joints.
Additionally and realistically, should a block become loose, it most likely would not travel the distance required to lose the support but rather, the block would become wedged in place due to a twisting moment resulting from the consideration that concentric pure tension on a block is highly unlikely.
Considering other local effects, the criteria for mortared double block masonry walls of zero tension capacity at the mortar-block interface is conservative.
As stated above in determining whether a support in the mortared walls should be through-bolted, PGE first checked to see if one wythe alone can withstand the reaction.
This was brought about because of their assumption of zero tensile strength for the collar joint.
If one wythe was capable of with-standing the reaction, no through-bolting was performed.
This calculation assumes wall delamination, additional effects of which are to decrease wall frequency and increase forces.
Wall frequency calculation was based on a composite wall.
This inconsistency is not believed to be significant because should delamination occur, it most likely would be a local effect and not affect the entir7 wall.
In addition the occurrence of total delamination for walls spanning between elevations of the structure would have the effect of changing the eno conditions from pinned pinned as presently assume toward fixed-fixed and thus would be a compensating effect.
The use of values from UBC for unreinforced masonry for considering local ef fects on single and double wythe masonry walls is reasonable and conservative due to the somewhat wide spacing of rebars of 2'-0" on center for vertical steel and 4'-0" on center for horizontal steel.
The criteria for composite walls of 40 psi principal tensile stress, as discussed above, appears to be acceptable and the use of UBC values for reinforced masonry appears reasonable because of the closer rebar spacing, namely 2'-0" on center both horizontally and vertically.
Local tensile stresses on the block / core interface of composite walls from supports are calculated using an assumed stress cone.
Questioning the accuracy of this assumed stress cone, another approach was calculated by the NRC staff which used the theory of elasticity.
Calculations for a point load in an elastic medium ind' ated that this stress can be about twice as high at one point immediately t,aneath the load, however, this stress drops of f rapidly.
The two approaches yielded approximately the same values when considering the
. area over which the stress is to act.
Therefore, it appears that a local overstress may occur in a small area beneath the load, but the assumed stress cone will give a reasonable approximation of overall tension stress distribution at the interface.
For global wall effects, vertical shear between wythes must be resitted in order for the wall to act like a composite.
Concern arose about this vertical shear because the sparse tie spacing can be relied on for only approximately 1.5 to 2 psi using ACI 318 methods for shear transfer across cracked concrete and a shear friction coefficient of 1.
The Uniform Building Code, the referenced document for these masonry walls, does not specifically address the oroblem of vertical shear transfer between wythes.
In the case of double-block masonry walls in the complex, maximum interface stresses were given as 18.4 psi for a load combination including seismic.
These values can increase by as much as 3.5 psi on exterior walls to account for estimates of maximum thermal stresses and about 3.5 psi to account for estimates of maximum interstory displacement stresses.
These maximum loadings occur on walls that do not support any significant safety-related piping loads.
For mortared double block masonry walls not supporting safety-related piping, 8-12% of the walls have maximum total stresses between 13-18 psi; 70-75% have less than 10 psi.
For double block walls supporting safety-related piping, 5-10% of the walls have maximum total stresses between 12-14 psi; 80-85% have less than 7.5 psi.
These stresses can be increased as above considering interstory displacements and thermal effects on exterior walls.
Limited in-situ tensile testing was done on the mortared double block walls similar to that done on the composite walls and resulted in tension bond resistance in the range of 25-160 psi on estimated net mortar area, which indicated a good bond between the mortar and block.
Direct shear test data was also submitted for concrete block attached to a concrete slab which resulted in approximate shear strength of 200 psi (Attachment 3-1 of the December 22, 1979 PGE submittal).
Additional direct shear data submitted for stone to concrete block masonry by mortar bond indicated ranges of 50-500 psi, with 50 psi being a lower bound (Attachments 3-2, 3-3 of December 22, 1979 submittal).
Also submitted was test data obtained from a paper in the Journal of the Structural Division of the ASCE which reported direct shear and tensile stresses measured during tests of brick couplets attached by mortar.
The results indicated an average shear strength to tensile strength ratio of about 21/2 (Attachment 3-4 to December 22, 1979 submittal).
This tension-shear relation-ship was also borne out in previously discussed Attach.nent 3-1 of the December 22, 1979 PGE submittal where shear strength was approximately 3 times the tensile strength.
Although these tests were not directly applicable to the case as that at Trojan, i.e., vertical joints between 2 concrete blocks, the tests seemed to support the contention of a mortar joint supplying usable bond shear strength.
This data also showed that the shear strength tended to be greater than the tensile strength.
The lower bound of the tests (50 psi) appeared in the range of the bond strength from in-situ testing.
Limited testing has been done at the University of California at San Diego (UCSD)1 1 Hegemier, G. A., Arya, S.
K., Krishnamoorthy, G., Nachbcor W., Furgerson, R., "On the Behavior of Joints in Concrete Masonry" University of California at San Diego, LaJolla, California.
. which was not submitted by PGE wnich showed a mortar shear bond strength of 14-17 psi on a vertical head joint between concrete blocks with no normal applied forced.
In summary because of a limited number of tests at UCSD, the in-situ testing indicated good bond, and the test data submitted showing a lower bound shear strength of 50 psi, that shear strength was usually about 2 1/2 times the tensile strength, and that the maximum calculated interface stresses were substantially lower than 50 psi (or 40 psi concidering 90%
mortar fill), it was concluded that there is reasonable assurance of the adequacy of this interface even considering possible increases due to thermal and interstory displacement effects.
However, due to the lack of testing on masonry for this type of behavior and the contrary nature of the UCSD tests, PGE must perform testing on mortared double block walls similar to those at Trojan to confirm our judgments and quantify the margins inherent in the design of Trojan.
In the case of the composite walls, the value of 40 psi principal stress discussed above is being used to consider the transfer of vertical shear between core and block.
The maximum calculated shear stress is 20.7 psi in the complex considering both tornado and seismic load combinations and occurs on a wall that does not support significant piping loads.
This value can be increased to about 27 psi due to an estimate of the variation in thermal induced stresses (forces) which is verified using a more detailed analysis, and an estimate of variation in interstory displacement induced stresses (forces) which is to be verified by January 31, 1980.
For composite walls not supporting significant safety-related piping, PGE estimates that approximately 5-10% of the walls have shear stresses 19-22 psi and approximately 60-65% have less than 10 psi.
For composite walls supporting significant safety-related piping loads, PGE estimates that approximately 5-10% of the walls have maximum total shear stresses of 12-14 psi and approxi-mately 70% have less than 10 psi.
PGE presents interface tensile stresses on composite walls for two cases with four categories in each due to bolt interaction ratios:
the first case considers that two base plate bolts are in tension and two in compression and the second case considers that all four bolts are in tension.
Using these tensile values and the maximum wall vertical shear-stresses (plus 7 psi allowance for variations in interstory displacements and thermal) results in the following principal stresses at the interface:
(1) for two bolts in tension, connections with the lowest tensile stress (73% of total) results in a maximum principal stress of 24 psi; connections with the highest tensile stress (11% of the total) results in a maximum principal stress of 27 psi, (2) for all four bolts in tension, connections with the lowest tensile stress (73% of the total) results in a maximum principal stress of 25 psi; connections with the highest tensile stress (11% of the total) results in a maximum princino' stress of about 30 psi.
The maximum bounding case of 30 psi does not appea tc oe excessive compared to the 40 psi proposed criteria and possible variations in tensile stresses due to the assumed stress cone or to the in-situ test results.
4 Even if the maximum presented tensile force increases by 100% due to any variations, this combination of tensile and shear stresses would result in a principal stress of about 40 psi.
In order to assess whether or not there were any significant effects of in plane loads on out-of plane capacities, interface stresses were requested for all walls where in plane stresses exceeded 50% of the shear controlled capacities (as indicated by PGE-1020 criteria).
There were no double block walls in this category and the maximum interface stress in the complex for composite walls in this category was 6.9 psi.
Due to the low interface stresses, it was judged that their effect would not be of :,ncern.
In the course of coring the double block masonry walls to test for bond strength, areas void of mortar were discovered.
This led PGE to core into the major shear walls to check for possible voids.
Upon doing so, voids were discovered in one of the major shear walls which was erected during 1972.
Additional voids were discovered in one of the diesel generator walls which was erected in mid 1972-1973 also.
After reviewing quality control records, this missing mortar (for most walls where it was missing) was attributed to a contractual dispute that occurred from May 1972 to about September 1972 because of disagreement between the blueprints not showing mortar and the specifications calling for mortar.
This led PGE to randomly core into other walls constructed in this time frame to assure quality of construction.
Additionally, PGE cored further into two of the previously mentioned shear walls and concluded that about 90% of the mortar is present.
In all cases where voids were discovered, non-shrink grout was pumped between the wythes to fill the void.
PGE's testing program and quality control record search appeared to appropriately address this problem; however, to further assure quality construction, PGE will, based upon an NRC staff request, core into double-block walls where vertical shear stress becomes greater than 7.5 psi (based on 15% mortar fill minimum and 50 psi shear on net area).
PGE will identify these walls, core 10 randomly selected holes in each and along one of the major shear walls, core 10 holes at each floor elevation.
Completion of this program will be by February 15, 1980.
If voids are found, they will be filled with grout.
In addition to coring to assure adequacy of the mortar, PGE has drilled into the mortar space for through bolting some pipe supports.
The numerous drilling and coring and the correction procedures folicwed lead PGE and the NRC staff to conclude that there is reasonable assurance of having about 20% of mortar-void areas.
This would reduce the 50 psi shear allowable to about 40 psi based on wall gross area.
Originally, all masonry and composite walls were designed for OBE and SSE loadings to meet the criteria as presented in the response to Question 2 in the December 31, 1979 submittal.
Calculations were presented for SSE loadings considering vertical shear transfer.
. Since the acceptance criteria proposed for all walls has yet to be quantified and because walls not supporting safety-related piping have not been shown to meet the proposed criteria for the OBE loadings, the.08g shutdown level should remain in effect until the confirmatory testing is performed and evaluated.
Should the OBE be exceeded the walls shall be inspected for signs of damage and delamination before resumption of power operation.