ML20062B268
| ML20062B268 | |
| Person / Time | |
|---|---|
| Site: | Calvert Cliffs  |
| Issue date: | 10/20/1978 |
| From: | Lundvall A BALTIMORE GAS & ELECTRIC CO. |
| To: | Reid R Office of Nuclear Reactor Regulation |
| References | |
| NUDOCS 7810260202 | |
| Download: ML20062B268 (51) | |
Text
-
BALTIMORE GAS AN D ELECTRIC COMPANY QAS AN D ELECTRIC SulLDING BALTIM OR E, MA9YLAN D 21203 October 20, 1978 ARTMum C. LUNOVALL,JR.
w u n.......,
som.
Office of Nuclear Reactor Regulation U. S. Nuclear Regulatory Co= mission Washington, D. C.
20555 Attn:
Mr. Robert W. Reid, Chief Operating Reactors Branch #h Division.of Operating Reactors f; '
j
- e..
k'.
Subject:
Calvert Cliffs Nuclear Power Plant j
Unit No. 1, Dc:ket No. 50-317
~~
Request for Amendney, Additional Information
Reference:
(a) BG&E letter dated 7/27/77 from Lundvall to-Case, Request for Amendment (b) NRC letter dated 5/2/78 from Reid to Lundvall, Request for Additional Information Gentlemen:
Reference (a) requested, among other things, an amendment to our Unit No.1 Operating License to modify the require =ents for containment tendon surveillance (Item 7 of that letter). Reference (b) informed us that additional information was required for NRC to complete its review of our l
request.
l
) contains the specific questions asked by NRC in Reference (b) and our responses thereto. Attachment 2 is a general description.
l of Bechtel Power Corporation's acceptance criteria for containment tendon lift-off force surveillance and provides the basis for several of our re-sponses.
We hope the attached information is sufficient to permit you to complete your review and approval of our July 27 amendment request.
~r,
Very truly youls, j
s Q,
4;
{
. ' Wb'8 0,Q e
cc - J. A. Biddison, Esquire G. F. Trowbridge, Esquire Messrs. E. L. Conner, Jr. - NRC J. V. Brothers - Bechtel 7HORh02*
f
- Attachment 1 A.
SPECIFIC QUESTIONS FOR CALVERT CLIFFS #1 1.
Page 2-1.
It is stated that an acceptance band (approximately 4%)
was established around the mean lif t-off curve, based on numerous factors which may not be well known:
Explain the method by which the precise figure of i 4% has a.
been arrived at.
b.
Explain why an acceptance band has been established only for the mean and not also for upper and lower bounds of the lift-off curves.
c.
It results from Fig. 5-1 to 5-3 that these lif t-off curves are applicable to isolated wires and not to the whole tendon.
Justify this approach considering that the number of effective wires in a tendon is not known, and that the use of a theoretical number of wires in the evaluation of lif t-off forces may intro-duce errors of the same order of magnitude as the width of the acceptance band.
d.
Because of the concern stated above in (c) the following state-ments (pages 2-1 & 2-2) on the normalized lift-off forces for t
any individual tendon being satisfactory cannot be accepted without further discussion.
Discuss this problem in the light of the concern expressed above and other concerns indicated in the following questions. Note that in Reg. Guide 1.35 the basic surveillance unit is the complete tendon and not the isolated wire.
RESPONSE
1.a The design criteria have required that the containment resist the loads utilizing only the effective prestress and dead load. The effective prestress is usually defined as the amount of prestress left after all losses are deducted considering a 40 year plant life.
These losses are due to friction, creep and shrinkage of concrete, elastic, shortening and wire relaxation. Some of these losses are time-dependent.
The acceptance band width of i 4 percent (approximately) was arrived considering the range of time dependent losses as follows:
I a.
Creep and Shrinkage After 40 Years 335 x 10-6 in/in Loaded 0 180 days
=
200 x 10-6 in/in Loaded 0 365 days
=
These values are based on tests conducted by the University of California for the Calvert Cliffs Nuclear Power Plant.
l
- Attachment 1 (cont'd) b.
Wire Relaxation Relaxatien G 680F 9.5%
=
Relaxation @ 950F
- 11. 07.
=
These values are based on the test conducted by the wire supplier.
Hooo Tendons Average installed seating stress before elastic shortening 168.35 kai
=
3.63 kai Average initial elastic shortening
=
Average installed seating stress 164.72 ksi after elastic shortening
=
Losses Due To Expected Losses Maximum Losses Creep & Shrinkage 5.8 kai 9.7 ksi 200x10-6x29x10 335x10-6x29x106 6
1000 1000 Relaxation 15.64 kai 18.12 kai 0.095x164.72 0.11x164.72 Total Losses 21.44 ksi 27.82 ksi 2hl= Difference between expected losses and max losses.
A L= 27.82 - 21.44 s
tiL= 6.38 ksi
- 7. difference between expected losses and max. losses
=
Zi L x100 164.72 6.36x100
=
164.72 3.8%
=
(Approxientely 4 percent)
. Attachment 1 (cont'd) 1.a Other variables such as modulus of elasticity for concrete and steel (t 3.0%), ram calibration accuracy and human error in reading dial gages (t 2%) were not accounted for in establishing an acceptance band width of 4 percent.
If these other variables are accounted for, acceptsace band of 4
percent can be expanded to i 10%.
Similar method was used to compute an acceptance band width for vertical and dome tendons.
1.b An acceptance band has been established only for the mean because the design of the containment structure is based on an average force required per tendon and an average (mean) force required per wire in a tendon.
1.c The lif t-off curves in Figures 5-1 to 5-3 are not for isolated wire.
As stated in response 1.b, the containment structure is designed for an average force per tendon. The expected lift-off curve for a wire is based on the average installed seating stresses for all tendons in a given group of tendons (hoops, vertical or dome) and numbers of effective wires in a given group of tendons.
For example, in hoop tendons, based on the average installed seating stress after friction losses is 168.35 kai. Using average time depen-dent losses of 24.29 kai, average stresses in hoop tendcn will be 144.06 kai or an average force of 636.47 kip / tendon.
For an average 90 wire tendon, this force can be reduced as 636.47/90 = 7.07 k/ wire.
As explained above for hoop tendons, for a given family of tendons (hoop, dome or vertical), average force per tendon was computed using average installed seating stresses and average time-dependent losses per tendon. This force then was reduced to force per wire for an average of 90 wires tendon. Therefore, lift-off curves in Figures 5-1 to 5-3 are not for isolated wires but for an average tendon reduced to a force per wire. It should be however noted that the I
lift-off force per tendon shown in Figures 5-1 to 5-3 was based on the number of effective wires in a given tendon.
1.d The containment structure is designed for an average prestressing force, i.e., hoop tendons are designed for an average effective pre-stressing force of 630 k/ft. This would require a total force of 90562.5 k.
The containment has 465 hoop tendons with 3 surveillance tendons (tendons with 93 wires). Not coun?.ing these surveillance tendons, there are 462 usable tendons. Based on the installation records,18 wires in total were missing or broken. Thus, there are 13854 (2(462(90)-18)) effective wires to provide 90562.5 kips.
6 I
k-(cont'd) 1.d Therefore, the minimum average design force per wtre is 6.54 k/ wire (90562/13854). However, because tendons are seated at different stress levels and variations in material properties and other un-knowns, it was decided that as long as the mean lift-off force for all hoop tendon falls within the acceptable band width based on the expected average losses and average installed seating stresses, individual tendon may have lif t-off force outside the band width because the containment is designed for the average force per tendon.
Also, the acceptance criteria for average lift-off force in a* wire is based on total number of wires in hoop tendons, not on individual wire force.
(
i O
I i
I
. -. -. ~
~
5-2.
End anchorage ast.emblies were found t.a be in acceptable state (page 2-2).
However, this statement requires elaboration:
4 Explain whether the visual inspection of the assemblies is a.
1 detailed'and close enough to provide assurance that no narrow, l
deep and sharp pitting or other forms of incipient corrosion have been missed.
b.
Explain also whether it is possible for some grease in the trumpet proper also to leak out, or to be missing for any other reason such as grease thermal contraction after instsi-lation. A discussion clarifying the problem of. shrinkage of the filler when cooling is needed. What procedures have been used to check on possible voids at intermediate points in tendons, due to the cooling shrinkage?
4
RESPONSE
2.a The end anchorage assemblies were thoroughly cleaned with Viscocity i
011 Company's Industrial Solvent #16 (or equal), wire brushes and rags before inspecting visually for corrosion. Corroded areas with narrow, deep and sharp pitting or other forms of incipient corrosion g
would not have been missed.
l 2.b The filler material is installed under a pressure of about 75 psi and at a temperature of about 120*F at the pump discharge. Tempera-ture of the grease in the drum is about 170*F. The sheathing is vented 'at high points and drained at low points thus ensuring that the sheathing is filled and the wires fully coated. As the filler cools, it will shrink about 5% by volume and voids may occur.
j However, the wires removed during this surveillance for corrosion inspection were all found to be coated with grease, and as indicated in Section 2.1 of Three Year Surveillance Report, the inspected wires were all free from corrosion at intermediate points.
Although there are no provisions for physically determining the pre-sence of any voids within the sheathing, it can be concluded based
(
on the above observation that the filler material is performing its intended purpose, that of inhibiting corrosion of the tendon wires.
The possibility of grease leak out through worn out rubber gaskets have been addressed in the Surveillance Report and as stated therein (see Section 2.1), a program has been established to replace damaged rubba.r gaskets and refill the empty grease caps with new grease.
l l
i
_,_n.-
m
.,y
_,y.
y
. ~.
4
' Attachment 1 j
4.1 It is stated (page 2-3) that in general a gage length of 100 inches i
was used in testing of wires instead of 10 inches specified in ASTM A-421-65 specification.
I Define with precision the meaning of sample length (10 f t.),
a.
gage (100 inches), and gage on which the measurements have i
l been based.
j b.
Define the method used to assure that the results of the tests
- minimum ultimate strength, elongation at rupture, strain at
)
yield point, etc. satisfy the prescription of the ASTM A-421-65 specification, considering that not standard gage length has l
been used.
I c.
State whether in all other aspects the prescriptions of the ASTM A-421-65 have been followed.
1 4.2 Discuss also the following items:
j Failing inside of gripping jaws, and influence of the shape of a.
j the j aws.
b.
The influence of curvature and twisting of tendons as installed q
on the resistance of wire to removing, which may have predeformed i
the wires in a complex way.
i c.
Rate of speed of the testing machines.
j 1
d.
Ambient temperature during testing.
Possible eccentric loading by the machines, bending or twisting a.
j of the wire samples.
j f.
Influence of the type of machine.
t g.
Comparison of testing methods used by the manufacturers, and by any other organization participating in testing of wire samples.
s h.
Influence of the temperature of filler and its pressure at 1
installation.
s 4
1 1.
Influence of the average temperature of tendons in place.
i j.
Machine calibration, k.
Improper wire removal, including the use of a sheave (if any) which may deform the wires in an unpredictable way.
i f
.I B
l t
i
=
i Attachment 1.
l i
l
RESPONSE
4.1.a The tendon surveillance procedure requires removal of wires from each I
family of tendors and inspected for corresion as well as tensile and elongation test. Three samples are cut from each wire removed..The length of these namples is 10 ft.
Therefore, the sample length means the length of wite samples cut for tensile testing.- The term " Gage Length" means the distance between gripping mechanism on each and of the tensile specimen.
4.1.b & The only difference between the method used to test the wires and 4.1.c ASTM A-421-65 is gage length.
In all other aspects the prescriptions -
of the ASTM 4-421-65 have been followed.
l 4.2.a The wires tested by Pittsburgh Testing Laboratory were gripped in a i
i serrated jaws attached to movable head. Wires did not break inside i
the jaws. Any wire failure inside the gripping jaws.would indicate that the jaws caused stress concentration.
In such an event, the j
break strength will be possibly lower than the actual ultimate strength of the material.
4.2.b The force needed to remove the wire is small (approximately 25% of ultimate). Therefore, deterioration because of curvature and
(
twisting of the tendon is unlikely.
r 4.2.c In all cases, testing machine speed was in the range which is normal for steel tensile tests as specified in ASTM E8.
Therefore, rate of loading had no influence on test results.
4.2.d Although ambient temperature was not recorded, all wires were tested with ambient temperature of 65'F to 75*F.
Therefore, ambient tempera-ture has no effect on test results.
4.2.e Any eccentric loading by the PTL test machine will be precluded due to the movable attachment mounted on the machine. Also, any bending or
[
twisting of the wire would only lower the test value. The test results, i
r therefore, are conservative.
i 4.2.f The only influence which the type of testing machine could have is in the calibration. Samples were tested using Tinius Olsen Universal
\\
testing machine, PTL S/N 217, whose accuracy is traceable to the National Bureau of Standards.
4.2.g All tensile tests performed by the manufacturers were done using 10 in. gage length as required by ASTM A-421.
All tensile tests performed by Pittsburgh Testing Laboratory on the samples removed during the tendon surveillance were done with 100 in. gage length.
The longer gage length was used to' comply with requirement for the l
longest practical sample as stated in Regulatory Guide 1.35.
For I
future surveillances, tests will be performed using a standard 10 in.
)
gage length sample.
l a
i s
~
', Attachment 1 l
(cont'd) 4.2.h The filler temperature during installation is high enough to make the X
i filler quite fluid (115'F to 170*F) but not hot enough to affect the steel heat treatment. After installation, the filler temperature is in the range of 30*F to 100*F; that would have no effect on the steel heat treatment either. The filler pressure has no effect on the l'
mechanical properties of the tendon wire.
4.2.1 The average temperature of tendons in place is not large enough (30*F to 100*F) to influence the wire strength.
4.2.j Tinius Olsen Universal testiag machine used by Pittsburg Testing Laboratory has been calibrated according to good material testing laboratory practice as specified in ASTM E4 and has an accuracy traceable to the National Bureau of Standards.
4.2.k As stated in response 4.2.b, the force required to remove the wire is small. Therefore, any significant deformation affecting wire strength is unlikely.
s a
4 s
i l
r.
. Attachment 1 5.
Figures 3-1 and 3-2 show a non-uniform repartition of surveillance tendons and an accumulation of tendons at certain locations.
Discuss whether this arrangement satisfies the requirements of a.
randomness contained in Reg. Guide 1.35.
RESPONSE
5.a Reg. Guide 1.35 requires that for surveillance purposes, the tendons be chosen randomly but representative 1y from each separate group.
Only ten tendons need to be selected for the hoop tendon system. For reasons stated in the surveillance report (see Section 5.0), the sam-pie size was increased to thirty and while there were some clustering of the hoop tendons, the thirty tendons taken as a whole are both randomly and representative 1y distributed. Also, the conclusion arrived at regarding the lif t-off forces based on thirty tendons (see Section 2.1) will be equally applicable if only ten randomly but representative 1y distributed tendons are considered such as 64H71, 64H50, 64H38, 53H60, 51H42, 62H54, 62H74, 31H70, 31H22 and. 51H60 for hoop tendons.
\\
ee
____.___________-_-__________.m____._
. Attachment 1 6 '.
You indicate page 4-1 that a number of button heads were found split, discolored, partially rusted and some were off-size or epaced so close that no measurement could be made.
The original design of button heads is quite sophisticated.
It a.
is therefore a matter of concern when the button heads are not in conformance with the original design, or have deteriorated.
Indicate the provisions made to inspect the button heads of all i
tendons. Indicate also what provisions have been made to inspect the fraying surface between button head and washer.
b.
The Prescon Corporation tests are not convincing since the margin of safety obtained is only 9% and only a small number of tests have been made.
Discuss the validity of this test, especially taking into account that the safety margin should be maintained for at least 40 years.
RESPONSE
6.a "Offsize" buttonhead refers to a buttonhead which is either larger or smaller than that shown for Go-No Go GAGE. This inspection is not
\\
intended to be a basis for acceptance or rejection of the tendon but rather a recording of the as-built condition. Similarly when split were found on the buttonhead, it was compared with the installation records to check for the acceptance criteria. Buttonheads with single split greater than 45* or multiple splits equal or greater than 45*
were rejected during the initial inspection.
After a tendon is detensioned and inspected, it is retensioned to 80%
of the minimum guaranteed ultimate strength based on the total number of effective wires in the tendons. This is in accordance with STP No. M-31, " Surveillance Test Procedure." The tendon is then shimmed to a force equal to or larger than the lift-off force determined at the time of detensioning the tendon. Thus, by subjecting the tendon to a stress in excess of the required lif t-off force without a button-head failure, the adequacy of the wire anchorage has been established.
Also, for the same reason, no provisions were made to inspect the
(
fraying surface between buttonhead and washer.
6.b During the test conducted by Prescon Corporation, all the samples, regardless of the buttonhead diameter, exceeded the minimum ultimate strength (240.00 ksi) by nine (9) percent.
Based on the average design prestress, the average effective prestress in hoop tendons af ter 40 years will be 133.25 kai or 55 percent of the minimum ultimate strength of the wire. Therefore, the margin of the safety is not nine percent but 98 percent (1.09/0.55 = 1.98) when com-pared to the minimum required.
A further discussion of safety margin under various conditions is given in response to question B.2.
i i
4
. Attachment 1 B.
GENERAL QUESTIONS APPLICABLE TO CALVERT CLIFFS #1 1.
Explain the meaning of the expression " guaranteed minimum ultimate strength." What are the tolerances and permitted deviations on the
" guaranteed" values?
RESPONSE
B.1 The term " guaranteed minimum ultimate strength" means the tensile strength of Type BA wire as specified in Table 2 of ASTM A 421-65 specification. Also, ASTM A 421-65 requires that the total alonga-l tion under load shall not be less than 4 percent.
Based on the discussion with Prescon and INRYC0 by Bechtel, both companies claim to have very low rejection rates. As a consequence, they only keep low test results on file, and have not performed any statistical analyses.
In view of the lack of statistical data from the wire supplier, we cannot do a study on the significance of the deviations from ASTM A-421 of the tendon wire samples tested during the last surveillance.
Ten wires have been tested during the first two surveillances for tensile strength. Only one wire broke below the minimum guaranteed tensile strength. The reduction was only 1.5 percent. When ratested with 10 in. gage length, the average tensile strength of the same wire was 2.7 percent higher than the minimum requirement. The total elongation measurements for all samples were in accord with the ASTM A-421 criteria.
The testing with 100 in. gage length was done to obtain a larger sample. Experience had indicated that the use of the larger gage length results in a slightly lower strength than tie "dard 10 inch gage length. However, acceptance is still based on o
.r, sample if the longer sample is below the required value.
(
t
, Attachment 1 2.
Explain the significance of lower lift-off forces for the stability of the structure, and evaluate the tolerance which could be accepted a
before the stability of the structure reaches an unacceptable low safety level.
KZSPONSE B.2 The lif t-off forces measured during the surveillance test had a satisfactory average value for each group.
The prestressing system for Calvert Cliffs Nuclear Power Plant is designed for the following average force per ft.:
Hoop Tendons 630 k/ft.
=
Vertical Tendons 300 k/ft.
=
Dome Tendons 360 k/ft.
=
Based on these design forces and the number of effective wires in each group of tendons, the minimum average design force per wire is as follows:
Hoop Tendon 6.54 k/ wire
=
\\
Vertical Tendon 7.05 k/vire
=
Done Tendon 6.78 k/ wire
=
Based on the average installed seatdug stresses and average expected losses in 40 years, the expected fes es per wire will be:
Hoop Tendon 7.07 k/ wire
~
=
Vertical Tendon 7.18 k/ wire
=
Dome Tendon 7.18 k/ wire
=
The average lif t-off force for hoop tendon was within the i 4 percent acceptance band width but was slightly above the lower bound of the acceptance band width. If the lift-off forces in hoop tendons follows the lower bound of the acceptance band width, the minimum average force per wire after 40 years will be 6.78 k/ wire.
(
The internal design pressure for the containment structure at Calvert Cliffs Nuclear Power Plant is 50 psi. The containment structure has successfully proven its capacity for 1.15 loading during S.I.T.
The maximum pressure during the subsequent leak rate tests will be 1.0 P.
I The average prestressing force required to resist 1.0 P loading is:
Hoop Teodon 4.67 k/ wire
=
Vertical Tendes 5.34 k/ wire
=
Dome Tendon 2.48 k/ wire
=
l
- Attachment 1 (cont'd) l It can be concluded from the above discussion that:
i 1)
For hoop tendons, based on the actual average seating stresses and for average expected losses, safety margin is 1.51 (7.07/4.67) for 1.0 P loading.
2)
Even if the aterage lift-off forces for hoop tendons follow the lower limit of the 4 percent acceptance band width, the safety margin still be 1.45 (6.78/4.67) for 1.0 P loading.
3)
Similarly, the safety margin for done tendons is 2.74 for 1.0 P using the average installed seating stresses and average expected losses. The safety margin will be 2.63 if the lif t-off forces in the dome tendons follow the lower limit on the i 4 percent accep-tance band width.
4)
As shown in Figure 5-2 of Five Year Tendon Surveillance Raport for Unit 1 Containment Structure - Calvert Cliffs Nuclear Power Plant, average lift-off forces for vertical tendons was above the.exoected average curve.
J
(
5)
For the factored load condition of 1.5 P, the containment has an ultime.te capacity to resist the accident condition, excluding the j
effect of liner plate and reinforcing steel, of (1.51x1.00), or 0.6 2.52 times the design pressure. If the capacity of liner plate and reinforcing steel uere to be included, the factor would be i
even h'igher.
1 6)
A large tolerance can be applied to the lif t-off forces without affecting the stability of the structure.
i k
4 1
b I
t s
e l
.c
~.
a
- ik -
\\
1 Explain the significance of ultimate strength of wire samples lower than i
the guaranteed minimum ultimate strength for the stability of the struc-ture, and evaluate the tolerance which could be accepted before the stability of the structures reaches an unacceptable low safety level.
Items 2 and 3 above involve a study of errors inherent in such a complex operation as lif t-off force, measurement, and of tolerances which should be applied to the result of the evaluation. For instance the alongations l
are measured between the outside face of the. bearing plate and the inside face of the stressing washer. Discuss whether the precision of these measurements is sufficient to establish the adequacy and the correctness of lift-off forces, considering the deformation and displacement of the base plates.
Discuss also the ram calibration of the rams and its tolerances. In general discuss the errors, their propagation, their summation and their influence on the results of lif t-off force measurement.
RESPONSE
B.3 At the end of 40 years, all anticipated losses will have occurred and the average stresses in the tendon will be 60% of its ultimate strength.
\\
The containment is designed for 1.5 P under factored load condition.
i The containment has an ultimate capacity to resist the accident condi-i tion, excluding the effect of liner plate and reinforcing steel, of 1.51x1.00 2.52 times the design pressure.
If the capacity of liner
=
0.60 plate and reinforcing steel be included, the factor would be even higher.
Ten wires have been tested during the first two surveillances for the i
tensile strength. Only one wire broke below the minimum guaranteed tensile strength. The reduction in ultimate strength of the single wire was only 1.5 percent. When added to the total wire tested, the average ultimate strength for all ten wires is 248.81 kai, which is 3.6 percent above the guaranteed minimum tensile strength. Therefore, it can be concluded that there is no reduction in the ultimate strength l
of the wire samples to affect the stability of the structure.
(
Since lif t-off force is measured independently using a calibrated
[
hydraulic ram, the precision of elongation measurement has no effect on " adequacy and correctness" of the lift-off forces.
I Prescon rams (Numbered 4045 0090 500 08, 4045 0100 500 08, and 4045 0200 50012) were calibrated against Load Cell PCL 78-L-57 on August 24, 1976. Additionally, Prescon rams (Numbered 4045 004 0500 08 l
1 and 4045 005 0500 08) were estibrated against the same Load cell on i
October 7, 1976. Rams were individually tested by having a pulling rod link load cell with data readings taken from a rea'dout meter i
located away from the pressurized ran area.
h
~ II ~
'~
(cont'd)
Load Cell PCL 78-L-57 (make: Load Cell Transducers. Inc.) was calibrated by Texas Calibration Company at Southwest Rasearch Institute November 21, 1975. The load cell was compared against a National Bureau of Standards - calibrated proving ring (Number 01/100261) in a loading range of 100,000 pounds to 1,500,000 pounds force. The machine error was.1%.
Based upon the above factors and based upon the fact that the gauges are divided in 100 psi increments, that the accuracy of the calibrated ram-gauge combination would have been between one and two (1-2) percent.
I
\\
l l
l l
- Attachment 1 4.
Discuss che possible complication arising from the fact that several different heats for wire steel may have been used by the manufac-turer.
RESPONSE
B.4 The certified mill test reports have been reviewed for compliance with ASTM A-421 requirements for various heats for wire steel.
The physical and chemical properties for all heats used by the wire manufacturer did meet or exceed the minimum requirements specified in ASTM A-421.
Use of different heats for wire steel, therefore, should not cause any complication.
b
\\
\\
4 i
e k
\\
I i
-. =_.
! Attachment 1 5.
In Appendix D the normalizing formulae are presented. Discuss the possibility that factors neglected in these formulas may be more important than factors which have been included. Some of the neglected factors are:
First surveillance procedure does not include concrete creep and concrete shrinkage or thermal effects and concrete placing variation, also prestressing sequence effects.
Subsequent surveillance formula do not appear to include the following: Thermal effects, shrinkage, detailed effects on creep, bearing plate displacements, changes in concrete, Young's elasticity modulus and Poisson's ratio, jack orienta-tion, cracking of concrete, Indicate the tolerances which should be applied to normalizing
)
factors.
RESPONSE
B.5 The design of containment structure is based on the average force per tendon af ter considering all expected losses such as elastic shortening, creep and shrinkage, friction and steel relaxation.
Since all tendons cannot be anchored at exactly the same value and at the same time, the lift-off forces must be normalized to account for elastic losses during initial installation (a function of post-tensioning sequence), the lif t-off force deviation from the base value, and the number of effective wires in a tendon.
Thus the normalizing factor is essentially a function of prestressing sequence, the average seating stress compared to the seating stress for an individual tendon. It does not account for any other losses except elastic shortening during initial installation. Consequently, there is no need for specifying any tolerance in the normalizing factors.
Other losses such as creep and shrinkage of concrete, thermal effects,
\\
steel relaxations, variation in modulus of elasticity of steel and concrete, etc. are accounted separately. Some of these losses, i.e.,
creep and shrint. age of concrete and wire relaxation can be estimated based on actual test results. From past experience, it has been determined that approximately 70 percent of the losses due to creep and shrinkage and steel relaxation occur during the first year.
Others such as jack orientation, concrete placing variations, etc.
cannot be accounted for directly.
If all these variations, i.e.,
creep and shrinkage, steel relaxations, concrete placing variation, thermal effects, variation in modulus of elasticity for concrete and steel, Jack orientation, accuracy of calibrated rams, are to be accounted for, the acceptance band would have to be as wide as 10 percent.
i
_.. Attachment 1
- (cont'd)
The lift-off forces for the first tendon surveillance were not nor-malized at the time of that surveillance, but normalizing factors have been established for those tendons.
When reducing the data initially for this surveillance, the tech-nique similar to equations stated above was used, however, the average elastic losses were not subtracted in the numerator. Also, the value of elastic icsses, used, to compute final prestressing force (40 year life) and the wire force 1 year after post-tensioning, was smaller than the initial average elastic losses.
Once these differences were realized, it was decided to use the appropriate normalizing formula with appropriate elastic losses.
This was done not to change or modify the original criteria but to correct the original criteria.
Also, as stated in response to question A-1, variations in time dependent properties such as creep and shrinkage, wire relaxations, modulus of elasticity of steel and concrete, accuracy of ram cali-bration, human error in reading dial gages, etc. will also influence value of lift-off forces.
To account for some of these variables, it was decided to establish a conservative i 4 percent band width around the expected average lif t-off force based on expected average losses such as creep and shrinkage, wire relaxation, and average elastic losses.
Therefore, changes were made in the original criteria not to destroy the continuity of the surveillance but to make the surveillance pro-4 gram and its goal more realistic.
\\
I 4
i n
r-,
- Attachment 1.
6.
The documents furnished suggest that the acceptance values and the normalizing factors used before be modified and that an acceptance band of certain width be included. After such a revision the normalized tendon lift-off forces will exceed the required minimum values. ' This approach is questionable, since one of the goals of successive surveillance operations is to establish historic continitity in the evaluation of the safety of the structure. A modification of the basic criteria will destroy this continuity. It is imperative first to establish the significance of not satisfying 'the original criteria and normalizing procedures.
Clarify this probica.
RESPONSE
B.6 The design of containment structure is based on the average force per tendon after considering all expected losses such as elastic shortening, creep and shrinkage, friction losses & relaxation of steel wire. From past experience, it has been determined that approximately 70 percent of the losses occur during the first year and the following expression can be used to determine the average wire force which should exist af ter one year and at the end of 40 years life. If exactly the required amount of prestress has been supplied, then end of life value and the minimum required value will be the same if the surveillance acceptance criteria is the same as the original design criteria.
Wire force 1 year after post-tensioning:
Wire f
)
fAv.
)
[ Creep,
)
~
Wire Area Initial Stress Elastic
-0.70 Shrinkage &
Force
)
(loss
]
(Relaxation /.
k Losses End of life wire force:
Wire Wire
~(Initial \\ [Av. Elastic \\
f Y
~
~
_, i Creep, Shrinkage Area (Stress j (
Loss j
( & Relaxation j
Force
=
-\\
/\\
/
\\
Losses
/.
\\
Since all tendons cannot be anchored at exactly the same value and at the same tice, a correction for the initial anchorage value and the elastic loss must be made so that the value will be typical of the entire average tendon population. This is done by determining a normalizing factor by which the lift-off value is multiplied when making corrections:
" Average Initial"
" Average" Normalizing
=
Factor Anchor Force Elastic
_All Tendons
_ Loss Actual Tendon
" Actual "
Lift-Off Elastic
_ Lo ss
. Attachment 1 7.
It is questionable whether the use of mean values from a small number of surveillance tests is significant for the whole pre-stressing tendon system of the plant. Justify the use of means for this structure.
RESPONSE
B.7 The containment structure has been designed based on an average pre-stressing force for a given tendon group. One of the objectives of the surveillance program is to determine if the average lift-off force from the tests equals or exceeds this required prestress level.
l To determine the significance of the test results vis-a-vis the entire prestressing system, the following statistics were generated using the tendon surveillanca data given in the Three Year Tendon Surveil-lance Report.
Error Normalized Est. for Design Lift-Off Force Mean Prestress
- 95%
Tendon Standard
- Confidence
(
Group Mean
- Deviation Level
- Vertical 7.55 0.353 0.44 7.05 Hoop 6.97 0.26 0.10 6.54 (30 Tendons)
Hoop **
7.13 0.26 0.19 6.54 (10 Tendons)
Dome 6.97 0.14 0.15 6.78 e
1 Forces in k/ wire Original 10 hoop tendons (see Response 5)
\\
It is apparent from the above data the use of sample mean as the average value for the individual tendon group is justified in view of the rela-tively small dispersion in the test results and low error estimate derived for the mean.
j ATTACHMENT 2 CONTAINMENT TENDON SURVEILLANCE LIFT-OFF ACCEPTANCE CRITERIA
(
Prepared By:
\\
T. E. Johnson i
i i.
\\
I i
Bechtel Power Corporation l
San Francisco, California February $978 I
r
?
k
TABLE OF CONTENTS Section Page
1.0 INTRODUCTION
1 -1 2.0 CONTAINMENT TYPES AND DESCRIPTIONS 2-1 3.0 CONTAINMENT DESIGfl 3-1 4.0 GENERAL ACCEPTANCE CRITERIA 4-1 5.0 SURVEILLANCE DATA REDUCTION 5-1
- 6. 0 REGULATORY GUIDE 1.35 AND NRC REQUIREMENTS 6-1 7.0
SUMMARY
7-1
(
o O
F l
4 e
CONTAINMENT TENDON SURVEILLANCE LIFT-OFF ACCEPTANCE' CRITERIA
1.0 INTRODUCTION
i The following information has been assembled to aid in the understanding of the purpose and requirements of tendon surveillance in prestressed post-tensioned concrete nuclear containments with unbonded tendons. The major emphasis is on the determination and acceptability of tendon lift-off 4
measurements.
i It is essential when determining acceptance based on surveillance results that the reviewer consider and have knowledge of basic containment design.,
Tendon surveillance has been performed by Bechtel Power Corporation on many containments and.was first done on the Palisades Containment in 1972 based
(
on procedures developed by Bechtel.
Since that time, Regulatory Guide 1.35 has been issued and a working group under ASME Section XI has been attempting to develop standards.
Present containment desigt. and construction is covered by the ASME Section III, Division 2 Code.
l-1
2.0 CONTAINPENT TYPES AND DESCRIPTIONS The first generation o'f prestressed containment structures (Type I) used about 500 ton capacity tendons, with six buttresses in the cylinder and ellipsoidal domes. These containments had about 500 hoop,180 vertical and 165 doine tendons.
By the next generation of containments (Type II), it was found feasible to use larger and longer tendons. The tendon capacity was about 1,000 tons and three buttresses were used. With the larger tendons and three buttresses, the tendons were reduced to about 160 hoop, 90 vertical and
(
85 dome tendons. A comparison of the hoop tendon configuration is shown in Figure 2-1 for the three and six buttress containments.
In addition to the changes listed above, the miniinum prestress level was reduced from 1.5P (design pressure) to 1.2P for some containments. An approximate estimate of the amount of prestress in a containment can be obtained by using the following expression:
y, F+0 s
P where:
X = level of prestress F = effective prestress force at end of design life 0 = dead load P = force resulting from internal pressure 2-1
For the Type I and Type II containments, X > 1.5, and in the later Types II and III containments, X > 1.2.
The 1.5 criteria essentially requires the containment to resist the factored loads by the effectiva, prestress which is the initial force in the tendon at installation minus the' losses from' friction, concrete creep and shrinkage, end prestressing steel relaxation. The very conservative 1.5 criteric 4: used for the first i
containments since they were new in the U.S.
i further experierice and :
knowledge of containments ' developed, it was determined that the 1.2P could replace the 1.5P criteria and still maintain tt.a required degree of safety.
The 1.2 criteria allows membrane concrete cracking when the containment is
(
designed for combined factored loads. This cracking a,llows the tendon to j
be retensioned from the effective value until force equilibrium is satisfied.
Under these combined factored loads, which include the effects of the design accident combined with the Safe Shutdown Earthquake, the tendon is allowed to reach a stress level of 90% of the yield.
Both criteria require that the containment remain in membrane cortpression during the initial structural integrity test (SIT) since the test pressurization is equal to 1.15P. With the test pressurization of 1.15P, both the 1.2 and 1.5
,s containments would respond about the same due to the test pressurization.
The latest generation of containments (Type III) has sinplified the configuration from ellipsoidal to hemispherical domes, which eliminated the ring girder and enabled the dome and vertical tendons to be combined in a single tendon which anchors in the tendon gallery.
For contairinents with hemispherical domes and 1.2P minimum prestress level, there are about 150 l
r j
2-2
hoop and 70 combined dome-vertical tendor.s which anchor in the tendon gallery.
~
A comparison of containment tendon layout is shown in Figure 2-2.
~
The number of tendons stated previously for containments can vary due to size, magnitude of design pressure and :arthquake, and the amount of prestress loss due to concrete creep and shrinkage and prestressing wire or strand relaxation.
Figure 2-3 shows a comparison of the three types of containments that have been designed by Bechtel Power Corporation in the past and are presently under design.'
(
l 1
i i
4 F
e -
9 2-3
._ ' FIGURE 2-1 l.
i CONTAINMENT HOOP TEN' DON CONT-lGURATION
.[
j$
-b s
I;
~
~
i 4
p DEVELOPED l ELEVATION HORIZONTAL WALL TENDONS (3 BUTTRESSES) g
- l l
=
I i
DEVELOPED ELEVATION
' HORIZONTAL WALL TENDONS (G BUTTRESSES) i
FIGURE 2-2 CONTAINMENT DOME TENDON CONFIGURATION
-f e
~
!)
~
.l
=
\\
^
i l
4 L i
as '.;
r LIMIT-ZONE OF VERTICAL
~1 HOOP. TENDONS DOME TENDONS (2 GROUPS) 9 s VERT. DOME
~'
d
.u-TENDONS 9
3.. )
45*
',(
PLAN - HEMISPHERICAL DOME '
SPRING LINE TYPE III S ECTION M D
4.
l
~
VERTICAL TENDONS d
m
/
DOME TENDONS fLIMIT-ZONE OF DOME
,I TENDONS (3 GROUPS) x.=.
1, h
HOOP TENDONS e
1 PLAN-SPHERE TORUS DOME SECTION M
.p TYPE II
(--/
l
i FIGURE 2-3 BECHTEL CONTAIN M ENTS l
TYPE II TYPE 11 TYPE 111 i
i i
.E.
DOME: 3 GROUPS @
DOME : 3 GROUPS @
h*-
DOME-VERT: 2 GROU TENDONS 12 0
- TENDONS 120*
TENDONS O 90*
~
SIZE :
SIZE:
SIZE-
.c HEIGHT-147' TO 220' HEIGHT-175'TO 209' HEIGHT-206' TO 2 59' L
DIAMETER-IO5'TO 130' DIAMETER -Il6'TO 130' DI AMETER - 124'To 150' TENDON: 50 T TENDON: IOOOT TENDON
- IOOO T SIZE SIZE SIZE 6
BUTTRESS.
3 BUTTRESS 3 BUTTRESS LEVEL OF:
- 1. 5 P LEVEL OF" 1.5 a I.2 P LEVEL OF :
I. 2 P PRESTRESS PRESTRESS PRESTRESS A-I.
4
[:
- d:.]
['. :.-:' :,..v...
- .
- , '. ]
h..~s.,.:.'.~.',.-
' i '..i.
[h b
O.
h b
4 9
~
3.0 CONTAINMENT DESIGN In order to set tendon surveillance acceptance criteria, the basic assumptions used in the containment design must be considered. Present containments are designed in accordance with Article CC-3000 of the ASME Section III, Division 2 Code. Prior to issuance of this code, the design was done using similar criteria based on ACI-318.
For illustration the post-tensioning system in the hoop direction of a containment will be designed; however, any interaction between the hoop and vertical directions will be neglected for simplicity. The code and
(
past criteria have required that the containment resist the operating loads and the structural integrity test pressurization utilizing only the effective prestress and dead load.
The effective prestress is usually defined as the amount of prestress left after all losses are deducted considering a 40 year plant life. The structural integrity test is performed at a pressure equal to 1.15 times the calculated accident pressure.
Another condition requires that the post-tensioning system resist factored load combinations such as accident condition of 1.5P, without exceeding the ultimate strength.
For ASTM A-421 wire, the allowable for this case is
.72f where f is the wire ultimate strength.
For this design case the pu pu concrete is allowed to have membrane concrete cracking and the tendons are allowed to increase in stress until equilibrium is reached.
3-1
l The following containment will be used to show the design process and, since it will be desirable for illustration to cover the entire tendon population, only ten tendons will be used.
I Ba' sic Design Information a)
Design accident pressure
= 60 psi b)
Inside radius
= 65 ft c)
ASTM A-421 prestressing wire:
Ultimate strength f
= 240 ksi pu Yield strength f
.8f
= 192 ksi
=
p p9 Modulus of elasticity E
= 29x103 ksi s
d)
Concrete:
(
Ultimate strength f'
= 5000 psi Modulus of elasticity E
= 5x103 ksi c
i e)
Losses: in determining the effective prestress, it is necessary to evaluate the prestressing losses.
- Elastic:
these losses occur since it is not possible to tension all the tendons at once.
For a very large number of tendons, the stress loss in the first tendon tensioned will approach the following value 1
assuming a stress in the concrete of.3f':
L,x = (~.3f' )
= (1.5) (29x
) = 8.7 ksi C
For a small sample of only ten todons, then the average _ loss of all tendons will be:
- (N-1} ' max, (10-1) 8.7 L
= 3* 9 ksi N
2 10 2
ave 3-2 un-,--
c-e
- Friction: using a curvature and wobble coefficient of friction of l
u =.14 and k =.0003 for a 240' tendon results in a friction loss of about 10 ksi.
- Wire Relaxation:
for normal relaxation wire, the loss is considered as 8% of the value at tendon seating and is therefore:
(.08)(.7)(240) = 13.4 ksi
- Concrete Creep And Shrinkage:
the::e values are determined from tests on concrete cylinders made from the design mix and are based on uniaxial compression.
500u in/in of loss at 40 years will be assumed here and the resulting loss is:
(29x103)(,0005)=14.5ksi
(
t All losses are based on testing and the interaction of losses is not considered since this is conservative.
A summary of losses af ter 40 years is:
Average elastic 3.9 ksi f.31.8ksi Relaxation 13.4 ksi Creep and Shrinkage 14.5 ksi a Friction 10.0 ksi
(
Total:
41.8 ksi Assuming that the tendon is anchored at.70fpu (168 ksi), then the effective stress in the wire at 40 years is 168 - 41.8 = 126.2 ksi.
It is permissible -
to anchor at a higher value than.70f to compensate for elastic losses; pu however, this will not be used in this example.
3-3
=
Since the containment must resist the test pressure of 1.15P, with effective prestress, then to be conservative a value of 1.20P, will be used.
Therefore, the effective membrane prestress force should be:
F,ff = 1.2 M = (1.2)
I (60)(65) = 674 k/ft A 182h" diameter wire tendon will be used here and the tendon capacity at a stress of 126.2 ksi is:
Tendon capacity = (182)(.0491)(126.2) = 1127.7 kips
(
Therefore, the average tendon spacing must be:
Tendon spacing = I 7 = 1.67 ft = 20 in Based on this spacing, the 1.5P, condition will bs checked:
- 4) (60)(65) = 842 k/ft 1.5P, = (1.5) 00
(
The tendon capacity at an allowable stress of (.9)(.8)f
= (.72)(240) = 173 p
ksi is:
l l
Tendon capacity /ft = (182)(.0491)(173)/1.67 = 926 kips /ft Since 926 > 842, then the 1.5P, load combination is satisfied.
l t
9 3-4
. ~..
l 4.0 GENERAL ACCEPTAllCE CRITERIA 4
As was previously illustrated in Section 3.0, the designer. wanted to make sure the containment would not have membrane cracking during the structural integrity test at a pressure of 1.15P,.
Therefore the effective prestress level was set at 1.20P, and all losses were conservatively considered.
l The design was also based on the average or typical tendon.
It has been conservatively assumed in the past that it was necessary to have a minimum average effective stress in the wire of 126.2 ksi or 1128 kips / tendon or 1
(126.2)(.0491) = 6.2 kips / wire.
It has also been shown that, by using this amount of effective prestress, the. factored load condition of 1.5P,
(
is well under'the tendon ultimate strength.
If the minimum wire force is 6.2 kips, then the force at the end anchor must be:
(136.2)(.0491) = 6.64 kips / wire due to the absence of friction loss.
The amount of prestress loss is sanewhat arbitrary since it does not really affect the ultimate strength of the containment.
It was.important to -
supply a sufficient prestress level so that the containment would not
'have membrane cracking during the structural integrity test.
However, when surveillance is being performed, the SIT has been completed and this require-
~
ment is no longer necessary assuming that future tests will only be made up to a pressure level of about P, for leak rate determination.
1 l
l L
4-1 1
~. -.
The effective required prestress in the actual structure determined by surveillance measurements should be based on the conditions that the containment will really experience during its lifetime.
These will be test conditions, operating conditions, and possibly the occurrence of the Operating Basis Earthquake (OBE).
Therefore it is not necessary to have the same acceptance criteria for effective prestress when performing surveillance as it was when sizing the system initially. A good criteria would be to require an average level of about 10% higher than the expected future test pressurization, provided-this is higher than the OBE.
For the previous example, the required force per wire would be:
h(6.69)=6.13 kips / wire Past experience has shown that about two-thirds of the losses occur in the first year after post-tensioning.
Therefore the average force per wire will be:
\\
(.0491) = 7.21 kips 168
.667(31.8)
Figure 4-1 illustrates the average tendon wire force versus time. The expected loss curve is shown together with some minimum values.
The shape of the expected loss curve is approximate and only time will tell a more exact shape based on actual measurements.
~
6 4-2
[
.1:, -
The minimum value line of either 6.69 or 6.13 kips / wire is the most important value on the curve since the average prestress for a group of, tendons must be above or equal to this value.
If a spacing of 18" were j
used instead of 20" for the same design conditions, then the containme'nc would be slightly over-designed and the following minimum value could be used:
fh (6.13) = 5.52 kips / wire h
The only purpose of the expected loss curve is to predict the general trend in losse' to predict future conditions.
This curve should not be used in s
i the acceptance criteria.
3 f
.W g
k f
i i
t 4-3 i
~
r-o
o-a'
- 5
~
i, j
l __ _ __.. _.. _1_....
1 4
.[
I a
l.
i i
.I i.
i l-i-
.iI e i e
i..
l.
' 8. 0 -
i I i g i
e I
.I i
l t
9,q O
7.21 Expected loss curve CL
+
t "a
n l
a P
7.0 -
~
2 L..
EL e
-6.69 2
Minimum based on 1.2P a
o La.
6.13
- 6. 0 -
Minimum based ois 1.1Pa 1
2 3
4 5
6 10 20 40 Years after completion of post-tensioning I
FIGURE 4-1 AVERAGE N NDON WIRE FORCE Veasus T' ~
.\\..
.v-
a
~'
5.0 URVEILLANCE DATA REDUCTION i
T As shown in Section 3.0, the containment is designed based on the average tendon for the group.
During tendon installation and tensioning it is not possible to anchor all the tendons at the same force level since it is only practical to use preselected sets of incremental shims. Also as illustrated in Section 3.0, there will be elastic losses and the tendons tensioned first will have lower force levels at completion of tensioning than those tensioned last. These two effects can be corrected for, since they are well defined.
These corrections must be made so that the measured values will be typical of average conditions and also the general loss trend can then
(
be established from one surveillance to the next.
To illustrate the concept of correcting (normalizing) lift-off readings, a containment which has only a total of 10 hoop tendons will be used. With this assumption the entire population will then be known (and is defined in Table 5-1).
The first column lists the tendon number and also shows when it was tensioned in the sequence. The second column shows the initial. lift-off values which are within 30 kips with an average of 1500 kips.
The third column 1
shows the tendon forces after elastic losses which were determined by:
Tendon Force Actual N-(n+1 )
t After Completion N
max
=
Lift-Off Of Tensioning where:
L
= elastic loss based on maximum elastic strain in the concrete (kips / tendon) max
~ total number of tendons N
=
n
= number of tendons tensioned prior to the one being considered e
r 5-1
t o
e e
From Section 3.0, the maximum elastic loss was 8.7 ksi; therefore:
l L
= (8.7)(.0491)(182) = 78 kips max 2
for a 182 wire tendon with a wire area of.0491 in,
The sixth tendon will be used for illustration:
Tendon forc'e = 1530 10-
- 1) 78.= 1499
]
/
The values in the fourth column were determined by assuming that 70% of the I
total wire relaxation and 55% of the concrete creep and shrinkage had occurred the first year. Using the sixth tendon, the relaxation is:
1st year relaxation loss = (.70)(.08)(initial lift-off)
= (.056)(1530) = 86 kips The first year concrete creep and shrinkage loss is:
(.55)(14.5)(.0491)(182) = 71 kips Total losses are: 86 + 71 = 157 kips
. Subtracting this value from column 3 for tendon 6 results in a first year predicted value of:
1 (1499 - 157) = 1342 kips 5-2
Column 4 simulates a sample that might be found in a real. containment-if the sample were not affected by other items such as a ram calibration.
Column 5 shows the nonnalizing factors which can be used to correct the actual tendon lift-off values to the average value.
The normalizing factors were determined by:
t Average E
N-1 max Lift-Off N
2 All Tendons t
NF =.-
Actual N-(n+1)
Tendon L
N max Lift-Off t
Using tendon 4 for illustration:
(
10-l' 78
~
~
10 T
1500 NF = 7
=h=.988
~
10-(3+1) 1530 10 78 i
Column 6 was determined by multiplying column 4 by column 5.
Again, using tendon 4 for illustration:
(1326)(.988) = 1310 kips t
[
All values of column 6 should have been 1310, but this would lead to i
normalizing equations which are extremely complicated and not justified.
Column 6 illustrates that by normalizing any tendon, then it will be essentially typical of the average tendon in the containment and this is a necessary condition. The bandwidths for columns 4 and 6 are:
Column 4 - unnormalized bandwidth (1365-1255)l00
= +4.2""
i (1365+1255) 5-3'
~,
Column 6 - normalized bandwidth (1314-1306)100 = +.31""
(1314+1306)
Significant errors may result if surveillance lift-off readings are not corrected for initial lift-off and elastic losses.
Table 5-2 shows the columns 4 and 6 information from Table 5-1 changed to -
force per wire. Lift-off can be stated in terms of tendon force, wire force or wire stress. Tendon force is not very good, since some tendons may not have the full amount of wire due to initial installation breakage and wire removal from previous surveillances. Wire force is more convenient than
(
stress since it is directly obtained by dividing the tendon force by the number of wires.
Based on past experience, Bechtel has observed a bandwidth of about 5%
after the lift-off readings were normalized; however, this may increase in future surveillances. Figure 5-1 shows a condition which may exist in many containments after years of operation.
If the expected loss curve and the minimum curve have the same value at end of life, then half the lift-off valueswill be above and the other half below the curve at end of li fe. Therefore, if each time a tendon falls below the curve, and if one or two additional tendon lift-offs are required, eventually all tendons r
will be needed for lift-off. Obviously this requirement would be ridiculous, since containments are designed for average tendon conditions.
5-4
A rational lift-off criteria is stated below:
1)
All lift-off values must be corrected 'or initial installation condi-tions (actual anchorage force and elastic losses during initial post-tensioning and any other significint effects) so that the value is indicative of the average level of'prestress.
2)
The average of all corrected lift-off values shall be equal to or above the minimum required prestress.
3)
Litt-off values shall be obtained on adjacent tendons for any tendon which is below 90% of the minimum required prestress.
(
I
\\
O e
i l
l 5-5
Table 5-1 End Anchor Forces 1
2 3
4 5
6 hbrmalizing Initial Anchor Forces Anchor Force Force Tensioning Lif t-Off After Elastic After First Normalizing After First Sequence Value(kips)
Loss (kips)
Year (kips)
Factor Year (kips) 1 1500 1430 1275 1.025 1307 2
1470 1408 1255 1.041 1306 3
1470 1415 1262 1.035 1306 4'
1530 1483 1326
.988 1310
(
5 1500 1461 1306 1.003 1310 6
1530 1499 1342
.977 1311 7
1470 1447 1294 1.013 1311 L
8 1500 1484 1329
.987 1312 9
1530 1522 1365
. 96 2 1313 10 1500 1500 1345
.977 1314 Average 1500 1465 1310 1.000 1310 i
6
Table 5-2 Wire Forces Normalized Wire Force Wire Force After First After First Tendon Year (kips)
Year (kips) 1 7.00 7.18 2
6.90 7.18 3
6.93 7.18 4
7.29 7.20 5
7.18 7.20 1
6 7.37 7.20 7
7.11 7.20 8
7.30 7.21 9
7.50 7.21 10 7.39 7.22 I
l l
e o
l#.'32 IINUeYa'_## $'S '. Yid'.*. k". I'. ' ' '' #^ (
46 S4(?J s
i,
(
-_a e.,
t.s e.'s c.
f.- r w
- J t.'
4 (P
.e i,
i
- q--;-j-r ;
i i
I
, i I ;i l
I !
e i
l i
i :
i 1
l 8.
t
- ^
l t
i e
r 8.0 -
1 i
l I
Nonnalized lift-off value i:
i A
g verage expected loss curve I
X wm x
o.
X
-x x
X X
K t
X E
X 5
7.0 -
x x
M X
g X
o' w
X i
a n
w g
n
. O
- - E M
o X
u.
X Average minimum x
6.0 -
i 1
2 3
4 5
10 20 30 40 YEARS AFTER COMPLETION,0F POST-TEN &IONING
?
- i :..
l..
FIGURE 5-1 Av gcE TENDON WIRE FORCE Ve-TIME
.. i f
~
l 6.0 REGULATORY GUIDE 1.35 AND NRC REQUIREMENTS As mentioned in Section 1.0, Bechtel submitted coments to the NRC on Regulatory Guide 1.35 and the NRC responded.
This correspondence is contained in Appendix A.
The requirements for lift-off acceptance from Regulatory Guide 1.35 are summarized as follows:
a)
"4.2 The maximum test liftoff force should be greater than the maximum in-service prestressing force."
b)
"7.1 The prestress force measured for each tendon in the tests described in Regulatory Position C.4 should be within the limits predicted for the time of the test."
c)
"7.2 There should be no more than one defective tendon in the total sample population.
If one sample tendon is defective, an adjacent tendon on each side of the defective tendon should also be checked."
Coments on items a), b) and c) are as follows:
(
There does not appear to be a clear definition of the intent of item a).
Item b) requires that tendon acceptance be based on a predicted value at l
the time of testing.
This requirement is meaningless, since it now makes acceptance and the amount of tendons in the sample a function of the designer's ability to predict very small loss values. Losses are originally, conserva-tively based on testing, and using these losses, the minimum amount of j
prestress in the average tendon is 6 termined. Therefore, the only thing impo.rtant to public safety is the minitum required value.
I i
6-1 l
)
As another example of the unimportance of the expected value, consider a containment which is over-designed.and has more prestress than is absolutely necessary. This condition can be considered simulated in Figure 4-1.
As shown, the acceptance should not be based on the expected loss curve, but must be based on the minimum value such as 1.lP,i Also for a tendon to be acceptable, it must be within limits predicted at the time of test (surveillance).
However these limits are not clearly defined and, in fact, it may not be possible to define these limits. If they are too loose, then the criteria may be meaningless, and if they are too tight, the surveillance sample may.
increase by a large amount and, as stated in Section 5.0, significant problems may be coming by the end of life, t
j Regulatory Guide Section 7.2 infers that if a tendon lift-off does not I
meet the requirements of 7.1, then it is defective.
This certainly appears to be a poor chotce of terminology. A tendon may just fall slightly below some chosen value and be in very good : condition.
Also, if a low lift-off is 4
determined due to excessive concrete creep and shrinkage, the tendon is j
certainly not defective, and if there is no evidence of cbrrosion, the ultimate strength has not changed; and therefore the word " defective" should be replaced with something more appropriate.
The valid and fullydef.ined technique of determining containment surveillance lift-off acceptability summarized in Sections 4.0 and 5.0 will eliminate most of the problems previously defined. This technique was previously documented in the Bechtel letter to the NRC (see Appendix A).
e 6-2
The following sumarizes the NRC response:
a)
"It is the intent of the guide that the applicants construct a tolerance band bounded by two concurrent prestressing force curves against, i.e.,... "
b)
"The practice indicated in your letter is not acceptable as it mixes up the initially determinable parameters (i.e., initial anchorage force, loss due to elastic shortening) with the parameters which cannot be detemined so accurately and are time-dependent...."
c)
"However, the criteria that you have suggested may be modified as follows:
- 1) All curves of predicted tolerance bands should be corrected for initial. installation conditions such as actual anchorage force and loss due to elastic shortening.
- 2) All lift-off values should be within the corrected tolerance bands of respective tendons. Lift-off shall be obtained on adjacent tendons for any tendon lift-off value which is outside the limits predicted for the time of the test.
- 3) The average of the lift-off values from all tendons in a particular region shall be equal to or above the minimum required prestress."
Item a) requires a band per tendon.
Item b) states that the Bechtel technique is "not acceptable" since it mixes up parameters. The l
technique does not mix parameters as has been illustrated in this document.
Item c) then states that curves should be corrected for initial installation conditions.
If done correctly, this is 6-3 l
y 4
J essentially the same as Bechtel has proposed, except the curves are adjusted instead of the measurements.
However, as.was previously pointed out, the Bechtel technique is much better for determining the condition of the entire group of tendons since they can be directly i
compared with each previous measurement and the average required value.
This is not possible with the NRC technique.
Item c)2) is a major problem since it used predicted values to set acceptability and this is incorrect.
l 1
4 l
i J
d 1
4 f
I 9
6-4
..s
/,
7.0
SUMMARY
This document has attempted to give an understanding of the design and requirements for a prestressed concrete containment post-tensioning system.
Also, realistic acceptance criteria and methods of data presentation are defined. The rajor points are summarized below:
1)
The required minimum level of prestress during surveillance may not be the same value as was used in the original design.
2)
Surveillance lift-off measurements should be normalized to eliminate elastic loss and initial anchorage effects.
3)
Acceptance should be based on the average lift-off of the sample since the or'iginal design was based on the average tendon.
4)
Acceptance must be based on the minimum effective force required and not some arbitrary predicted value.
[
l e
7-1
-