ML20094R350

Forwards Response to RAI Re Topical Rept 093, TMI-1 Reactor Bldg Twenty Yr Tendon Surveillance (Insp Period 6)

ML20094R350
Person / Time
Site:	Three Mile Island
Issue date:	11/29/1995
From:	James Knubel GENERAL PUBLIC UTILITIES CORP.
To:	NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM)
References
C311-95-2502, NUDOCS 9512040369
Download: ML20094R350 (19)

Text

.

GPU Nuclear Corporation gg g Route 441 Soutn P.O. Box 480 Middletown, Pennsylvania 17057-0480 (717) 944-7621

< Wnter's Direct Dial Number:

(717) 948-8005 November 29, 1995 C311-95-2502 U. S. Nuclear Regulatory Commission Att: Document Control Desk Washington, DC 20555

Subject:

Three Mile Island Nuclear Station, Unit 1 (TMI-1)

Operating License No. DPR 50 Docket No. 50-289 GPU Nuclear Response to NRC Request for Additional Information Regarding the THI-120" Year Reactor Building Tendon Surveillance Test 2

Attached is the GPU Nuclear response to the NRC's August 31, 1995 request for additional information regarding Topical Report (TR) No. 093, entitled "Three Mile Island Unit 1 Reactor Building Twenty Year Tendon Surveillance (Inspection Period 6)." Our letter dated September 26, 1995 provided the GPU Nuclear schedule for responding.

Sincerely,f / .e f I

'J. Knubel

, Vice President and Director, TMI MRK Attachment cc: Region I Administrator TMI-1 Senior Project Manager TMI Senior Resident. Inspector 1

ibD.5l 9512040369 951129 hk '

PDR ADOCK 05000289 i P PDR

l C311-95-2502 Attachment A Page 1 of18 ,

GPU NUCLEAR CORPORATION  ;

THREE MILE ISLAND UNIT 1 l TENDON SURVEILLANCE PROGRAM j RESOLUTION OF NRC QUESTIONS j PURPOSE l The purpose of this repon is to address eight specific questions originated by the NRC concerning l the results of the 6th tendon surveillance represtting the 20th year of operation for the Three Mile Island Nuclear Station, Unit i Reactor Building.

1 REFERENCES

1. TMI Unit 1 Reactor Building,20th year Tendon Surveillance, inspection Period Number 6,  !

Topical Report 093, Rev. O,3/95. I

2. PSC Report #463 (2 Volumes),20th Year Physical Surveillance of the Three Mile Island Unit 1 Containment Building.

3. Vendor Manual for Inland Ryerson, GPU Nuclear Manual No. VM TM-2485, Rev. 3, 9/94.

l

4. Tendon Force Curves Calculation - Surveillances 6 through 10, DC-5390-225.01-SE,4/94.

5. TMI Unit 1 Technical Specifications, Section 4.4.1.1, Amendment #167 and Section 4.4.2.1, Amendment #187.

6. FSAR Sections 5.2,5.7, and Appendix SB.

7. Force Curves Calculation, DC-5390-225.01-SE.

8. "Tenen Surveillance Requirements - Average Tendon Force," by J. F. Fulton, Nuclear Engineering and Design, October 1982, Page 303.

9. PSC Vendor Repon for the fifth surveillance,15th Year Physical Surveillance of the Three Mile Island Unit 1 Reactor Building, Revision 1,3/90. (2 Volumes)

10. PSC Vendor Report for the fourth surveillance, loth Year Sun eillance of the Three Mile Island Unit 1 Reactor Building, Revision 1,3/90. (2 Volumes)

I1. First Period Surveillance Report, I year after SIT, GAI Report !880.

12. NRC letter from Mr. Ronald W. Hernan, NRC Senior Project Manager to Mr. James Knubel, GPU Nuclear, dated August 31,1995, a request for additional information listing 8 questions.

13. Regulatory Guide 1.35.1, Rev. O, Detennining Prestressing Forces For Inspection Of Prestressed Concrete Containments.

14. Regulatory Guide 1.35, Rev. 3, insenice Inspection Of Ungrouted Tendons in Prestressed Concrete Containments.

4 C31195-2502 Attachment A >

. Page 2 of18 NRC QUESTION NO.1 Table Vill of the Reference 2 Surveillance Report refers to the tendon force base values from VM-TM-2485. Explain how the base values are established.

GPU NUCLEAR RESPONSE TO QUESTION NO.1 The base values shown on the referenced tendon force curves are the result ofindividual tendon loss calculations specifically prepared for the TMI-l Reactor Building surveillance periods. The base curve is equivalent to the prescribed lower limit as referred to in Regulatory Guide 1.35 and represents the summation of time dependent prestress losses as referred to in Regulatory Guide 1.35.1. The losses addressed in the TMI l calculations are the same as referred to in the Regulatory Guide and include:

1. Initial loss due to clastic shortening

2. Concrete shrinkage losses

3. Concrete creep losses

4. Steel wire relaxation losses The force loss due to elastic shortening was calculated specifically for individual tendons in each of the three tendon groups. Individual tendon value were calculated based on a ratio of the total clastic shortening as determined for the entire group of tendons. The clastic shortening is dependent on the total number of tendon stressing sequences and the actual stressing sequence number used for the specific tendon.

Concrete shrinkage losses were calculated based on the time since average concrete placement, the E3of the steel and the tendon area involved.

Concrete creep losses were calculated based on ie time since average concrete placement, the Es  !

i of the steel, the average stress on the transformeu area, the tendon area involved, and Poisson's ratio. l l

Steel wire relaxation losses were calculated based on stress relaxation curves available from the l wire vendor at the time ofinstallation of the wire. A wire factor was also used to account for the I effects of missing and ineffective wires on an individual tendon basis.  !

i The above losses are calculated individually and summed to detennine a total prestress loss for ]

cach individual tendon. These total losses are then deducted from the original force for the  !

individual tendon at the time ofinstallation. The resulting force-time carve is the base predicted tendon force curve for the life of the tendon. 95% and 90% base curves are then plotted on the same graph for evaluation of actual lift-off data against the acceptance criteria within Regulatory )

Guide 1.35, Revision 3. This allows for immediate action to be taken during the surveillance on adjacent tendons as required by the Regulatory Guide, based on the individual tendon lift-off results.

The curves contained in VM-TM-2485 include " lower limit" and "90% lower limit" values which are based on terminology in an earlier version of the Regulatory Guides. Recent calculations (Reference 4) were prepared to develop the force cun'es to be used for surveillances 6 through 10.

These curves show the base curve, the 95% cun e and the 90% curve plotted on the same graph.

.]

' i'

. C311-95-2502' l

• i Attachment A -

q

. Page 3 of18 This is consistent with the terminology used in Revision 3 of Regulatory Guide 1.35 and replaces ,

those presented in the VM TM 2485 document. The base curve values are the same in both j documents, unless there has been a change in the number of effective wires within the tendon. The i current criteria, methodology and design bases for the preparation of the indisidual tendon force i curves has been maintained since their original preparation, with periodic reviews during prior j surveillances. The intent of Regulatory Guide 1.35 has been maintained as it evolved from it's i proposed revision issued in 1979. However, the tendon surveillance program and specifically, the j calculation of the individual force curves may vary from the positions taken in Regulatory Guide 1

- 1.35.1, Revision 0, as formally issued in July,1990. While GPU Nuclear has not committed to l Regulatory Guide 1.35.1, Revision 0, the TMI-l tendon program has a solid basis which meets the -

intent of Regulatory Guide 1.35, Revision 3.  !

i i

1 i

I 1

l l

l

! l i l i

l l

~l

-]

i nr y 9 --m g ..54i --p + g y- T -

erww- =

• _ _ - --------

C311-95-2502

~~

Attachment A

, Page 4 of 18 NRC QUESTION NO. 2 Provide tl e lower bound tendon force curve or the tendon force as shown in Regulatory Guide 1.35.1, Revision 0, for each of the Figures (Attachments B, C, and D) in Topical Report 093.

GPU NUCLEAR RESPONSE TO QUESTION NO,2 GPU Nuclear is not committed to Regulatory Guide 1.35.1, Rev. O, and has not backfitted existing tendon calculations to determine tolerance bands for groups and subgroups of tendons.

The trending curves presented in the Topical Report as Attachments A, B and C were newly prepared for this surveillance period as a method of addressing Section 7.1.6 of Regulatory Guide 1.35, Revision 3, formally issued in 1990. The plotted data represents a trending line based on actual measured data results from the five prior surveillances, as well as the 20th year surveillance.

The bases for these trend curves are further discussed in the response provided to NRC Quevion No. 3.

The request for a plot of the lower bound tendon force to be added to the existing Attachment B, C and D curves is not fully comparable with the existing plotted data makmg up the current trending curves. He primary reasons for this are:

0 The Attachment trending curves were prepared as a method of addressing Section 7.1.6 of l Regulatory Guide 1.35, Revision 3, and were not specifically set up to address the I requirements of Section 4 and/or Figure 2 of Revision 0 of Regulatory Guide 1.35.1.

O Regulatory Guide 1.35.1 states that tolerance bands for groups and subgroups of tendons should be constructed and used for comparison of measured forces with predicted forces. The l TMl force curves are set up so that each tendon has its own unique force curve representing 1 the summation ofindividual prestress losses as discussed and reviewed in response to NRC Question No.1. The base curve and the calculatiou for the individual prestress lo: tes are based on a method and approach which has been in place for essentially the life of the plant.

O Regulatory Guide 1.35.1 presents a method for the easy construction of tolerance bands.

However, the methods used in the preparation of the TMI-l base curves involve decisions and I assumptions made at the time of original ;tesign, which contain varying degrees of l conservatism used for most of the variables that make up the predicted prestress force line, i.e.,

the established base curve. To provide the requested lower bound plot, GPU Nuclear would have to perform an in-depth review to understand how these same variations were considered in the original design and in the preparation of the base curves, and with what degree of conservatism. Otherwise, the construction and use of the tolerance bands as noted in j Regulatory Guide 1.35.1, may provide 4r inappropriate results.

O While significant efforts have been made over the last 20 years to keep the TMI-l tendon surveillance program up to the evolving Regulatory Guide crite>ia, the tendon force curves as originally prepared for TMI l have not been reviewed in depth for coinparison with Regulatory Guide 1.35.1. issued in July,1990. As such, there has been no formal commitment by GPU Nuclear for full compliance with Revision 0 of Regulatmy Guide 1.35.1 for the TMI-l tendon surveillance program. It would require additional work to evaluate the costs and benefits of backfitting for compliance, as well as any changes necessary to the tendon surveillance program.

6 .

C311-95-2502 Attachment A

• Page 5 of18 NRC QUESTION NO. 3 For the Attachment B and other similar Figures within the Topical Report, explain how the lift-off force is represented by one value. Explain how this representative force is established and how the trend line was obtained. All the tendon lift-off forces for all surveillances should be used without any normalization and linear regression analysis should be performed to obtain the trend.

GPU NUCLEAR RESPONSE TO QUESTION NO. 3 The trending data and curves as presented in the Topical Report Attachments B, C and D were prepared new for this surveillance as a formal method of addressing Regulatory Guide 1.35, Revision 3, Section 7.1.6, formally issued in 1990. LiflofTdata was tabulated on the following Tables attached herein.

Table 3.1 Dome Tendons Surveillance Data Table 3.2 Vertical Tendons Surveillance Data Table 3.3 Hoop Tendons Surveillance Data The average lift-off force for each individual tendon, as measured (not normalized), is plotted on the individual force curve where it can be readily compared with the predetermined values of base, 95% base, and 90% base as required per the Regulatory Guide. If any measured, not normalized, lift-off force falls below the base value, the requirements of Regulatory Guide 1.35 are followed.

The lift-off forces used to construct the trending curves shown in Attachments B, C and D in the Topical Report are normalized averages calculated from the average tendon lift-off forces measured for each tendon group during that surveillance.

The NRC request to use all lift-off forces and not use the normalization factor has been reviewed in depth. It is believed that it is more conservative to review the normalized data than the measured averages. To illustrate, suppose that the small sample of tendons as selected in a random but representative manner all had very high negative normalization factors, and the normalization factors were ignored in the trending plot. Then the plot of measured lift-off values would show a non-conservative trend line plotted higher in the vertical scale than that of the normalized data. By using the normalization factors, this small sample bias is eliminated and the average value obtained is representative of the average force condition for the inspected group. In addition, the normalization factor accounts for the effects on an individual tendon's initial lock-off force value resulting from tensioning adjacent tendons (see response to Question No. 4 for additional detail). If the normalization factors are not considered, the total losses experienced by the tendons cannot be appropriately determined.

The question of normalization of the input data has been recently addressed by Parsons Power (formerly Gilbert / Commonwealth) for the Crystal River plant where it was determined that it is more appropriate to use nonnalized lift-off data. Also, the use of averages vs. the use of all tendon data points was previously studied and discussed with the NRC for Crystal River. Through comparative analyses, it was determined that this variable did not have a significant effect on the resulting regression analysis at Crystal River Similar comparative analyses of the TMI-l data has been completed in response to this question with the following results.

Based on a comparison ofindividual data pomts vs. the averages for TMI-1, it has been determined that using all data points results in the steepest, and therefore the most consen ative, slope curve.

With respect to a comparison of normalized data vs. not normalized data, the results are mixed.

i o .

, C311-95-3503 Attachment A

, Page 6 of18 For the dome tendons, using the not-normalized data results in a steeper slope. However, for the l

hoop and vertical tendons, the normalized data provides the steepest slope. None of the above comparisons show any significant difference from one presentation of the data to the other. 1 Therefore, since the NRC requested method is usually slightly less conservative, the cun es selected for any data presentation herein use all lift-o Tdata points without being normalized.  ;

With respect to the linear regression, GPU Nuclear ruiewed the tendon lift-off data points from all prior surveillances and concluded that the prestress lo ss was indeed larger in the early life of the plant, and that these losses now have leveled off signi icantly. A pure linear regression of the data points is therefore not appropriate and is not used. Tie curve selected is a best fit among all the data points considering a steeper slope in the initial y:ars, and leveling out as losses are shown to have stabilized. The curve for the hoop tendons was prepared using engineering judgment as a best fit for the data points and is shown as Attachment B in the Topical Report.

Utilizing the data from Topical Report Attachment B, the hoop tendon curve has been duplicated herein using a mathematical regression best fit curve as shown in the attached Figure 3.1. This curve utilizes all lift-off points without being normalized. The results of this curve show that the regression line may fall below the minimum required prestress force for the hoop group in about 6 years. While this curve differs slightly from the curve contained in the Topical Report, it still shows that the tendon force values are expected to be above the minimum required value at the next surveillance. A similar curve will be generated after the next surveillance period to determine if forces remain at acceptable levels. If the results obtained in the 25th year surveillance indicate that forces would be expected to drop below the minimum required value, GPU Nuclear would take appropriate action.

l l

1 l

l

1

~

C31195-3503 I

, A,ttachment A l Page 7 of 18 TABLE 3.1 DOME TENDON SURVEILLANCE DATA {

.J 1

AVG AVG NORMALIZED NORMALIZED AVG TENLON PERIOD DATE FORCE FORCE FACTOR FORCE NORM FORCE l KlPS KIPS KIPS l KIPS KIPS l

')101 1 May-75 1252.0 40 1292.0 ]

~

J201 ist year to 1278 0 -27 1251.0 D301 after SIT Jun-75 1269 0 37 13060 I D220 1253 0 9 1262.0 j D316 12590 -20 1239.0 D116 12590 19 1240.0 1261.7 1265.0 4

D130 2 Aug-77 1252.0 -7 1245.0 f D148 3rd year to 1226.0 10 12360 ]

D202 after SIT Oct-77 1273.0 -44 1229.0 D219 1226.0 -42 1184.0 I

D334 1247.0 -10 1237.0 0348 1226.0 21 1247.0 1241.7 1229.7 D131 3 Mer-80 1180 0 -44 1136.0 1 D147 Sth year to 1180.0 19 1161.0

D218 after SIT Aug40 1137.0 20 1157.0 0 D203 1150 0 -40 1199.0 D346 1169.0 19 1188.0 D336 1221.0 -15 1206.0 1174.3 1174 5 i D133 4 May45 1100.0 70 1170.0 0225 10th year to 1117.0 45 1162.0 D314 after SIT Jun45 1206.0 -54 1232.0 1167.7 11880 D145 5 Oct49 1220.0 -34 1186.0 D218 15th year to 1147.5 20 1167.5 D347 after SIT Nov49 1180.5 -40 1140.5 1182.7 1164.7 D141 6 Sep-94 1161.0 47 1208.0 0225 20th year to 1113.5 45 1158.5 D248 after SIT Nov-94 1188.5 9 1197.5 1154.3 1188.0 4

Minimum required prestress for Dome Group = 1140 kips O

C3II 95 3502 Attachment A  :

PAge 8 of 18 TABLE 3.2 4 VERTICAL TENDON SURVEILLANCE DATA AVG AVG NORMALtZED NORMALIZED AVG TENDON PERIOD DATE FORCE FORCE FACTOR FORCE NORM FORCE KIPS KlPS KIPS KlPS KIPS r

V16 1 May 75 1348.0 -11 1337.0

V27 1st year to 1285.0 -26 12590 V61 after SIT Jun 75 1306.0 -22 12840

< V86 1285 0 9 1294.0 j V158 1300.0 -38 12680 ,

1306.0 1288 4 V24 2 Aug-77 1283.0 -25 1258.0 V48 3rd year to 1275.0 32 1307.0 i V72 after SIT Oct-77 1258.0 9 12670 V97 1258.0 4 1262.0 V119 1200.0 15 1224.0 1256.6 1263.6 i Mar 80 -20 1254.0

)

V18 3 1274.0 V31 5th year to 1147.0 0 1147.0 i

V55 after SIT Aug-80 1211.0 -7 1204 0 4

V105 1253 0 -44 1209.0 ,

V138 1211.0 -40 1171.0 4

1219.2 1197.0 V14 4 May45 1243.0 -28 1215.0 d

V30 10th year to 1193.0 10 1183.0 V32 after SIT Jun45 1196.0 4 1188.0 V84 1203.0 -22 1181.0 V160 1192.0 7 1185.0 1205.4 1190.4 V19 5 Oct49 1186.0 -9 1177.0 V21 15th year to 1185.0 -40 1145.0 V22 after SIT Nov49 11690 -7 1162.0 V23 1175.0 17 1192.0 V50 12090 -31 1178.0 V83 1193 0 -11 1182.0 VB4 1169.0 -22 1147.0 )

VB5 1179 0 4 1183.0 1183.1 1170.8 V32 6 Sep-94 1204 0 4 11960 V78 20th year to 12890 -35 1254.0 i V126 after SIT Nov-94 12050 19 12240 1232.7 1224 7 i

Minimum requirec prestress for Vertical Group = 1010 kips I

4 C311-95-2503 Attachment A -

Page 9 Of 18 l TABLE 3.3 l 1

HORIZONTAL TENDON SURVEILLANCE DATA AVG AVG NORMALIZED NORMALIZED AVG f' FOROE FORCE FACTOR FORCE NORM FORCE TENDON PERIOD DATE KIPS KIPS KIPS K!PS KlPS May 75 1260 0 50 1310 0 M13-46 i ist year to 1273 0 31 13040 H13-34 Jun-75 1253 0 13 1266 0 H6216 after S17 1272 0 -30 12420 H62-10 1259 0 8 12670 H35-10 '

1282 0 -6 12760 H35 28 1261 0 29 1290 0 H13 28 1293 0 13 13060 H51 12 1267 0 41 13060 H2&21 1280 0 76 13560 H24-47 1270 0 1292 5 2 Aug-77 11050 19 11240 H24-19 3rd year to 1194 0 -22 11720 H24-48 Oct 77 1242 0 52 11900 i H35-11 after Sli ,

1219 0 43 12620 H35-29 1225 0 -4 1221 0 H46 24 H46-26 1206 0 7 12130 1217 0 -46 1171 0 H51 13 '

1163 0 62 12250 H62-11 1113 0 84 11970 H62 47 1177.0 64 1241 0 H62-53 1186 1 1201 6 3 Mar 40 1222 0 50 12720 H62 57 '

Sth year to 1191 0 35 1226 0 H2649 '

Aug40 1253 0 -25 1228 0 M46-32 after SIT H46-30 1243 0 -13 12300 1253 0 -8 12450 I H2620 1243 0 -20 12230 H2628 H62-28 1243 0 -16 1227.0 H35-16 1221.0 0 1221 0 H6210 1253 0 30 12230 H51 11 1243 0 -57 11860 1236 5 12281 H13-35 4 May45 1184 0 80 11240 H13-36 10th year to 10640 15 10790 H13-37 after SIT Jun45 1175 0 -45 11300 H24 26 1172 0 -24 1148 0 H35-26 1153 0 17 11700 ,

1138 0 2 1140 0 H62 26 11460 4 1150 0 M62 30 11474 11344 H2629 5 Oct-89 1086 0 41 11270 I 15th year to 1135 5 -36 1099 5 H24-30 d H2431 after SIT Nov-89 1108 5 31 1139 5 H24-51 1139 5 73 1212 5 H46 34 11720 27 11450 H6213 10670 59 11460 H62 26 1122 $ 2 1124 5 1121 6 1142 0 H24-40 6 Sep.94 11280 50 11230 H35-23 20th year to 1184 0 -34 0 1150 0 H35-47 after sit Nov-94 1182 0 39 0 11430

$ H62-26 11460 20 11480 H62-49 11450 47 0 1192 0 1157 0 ' 1151 2 Minimum required prestress for Hoop Group = 1121 klps

Three Mile Island Unit #1 Tendon Surveillance Program Min. Re ired Av g. Utfoff Force 0 Hoop Group Trend of Losses Tendon Yorces Not flormofized gy -> y 5I5 o9h FIGURE 3.1 2 oc >~ w 8 1500 ..

1400 -- _

ui ~ ^

n_

n _

~ 1300 - .

tij _ , -

u ,

(Y _

2 }\ + _

\; . -

9 1200 l N s -

i n;i rips c g o

~

1100

• i i i ia i i 8 i 8 ' ' ' ' ' ' ' ' ' ' ' ' ' ' '

1000 - T iiiiiiii i i i i i i i i i 40.0 50.0 10.0 20.0 30.0 O.0 SCHEDULED SURVEILLANCE PERIOD (YEARS AFTER SIT)

HiREND7.GRF l

' C311 95-2502 Attachment A

. Page 11 of18 NRC QUESTION NO. 4 in Table IX of the Reference 2 document, the normalizing factor and normalizing lift-off forces are shown. Explain how GPU Nuclear obtained the normalizing factors expressed in terms of forces.

GPU NUCLEAR RESPONSE TO QUESTION NO. 4 Table IX of the Reference 2 PSC Report shows the average lifl-off values corrected with the normalization factor and then averaged for comparison with the minimum required prestress force for that group of tendons. This computation has been performed for ali arveillances to date.

Normalization factors are a function of the origim I stressing sequence, the lock off force at the original stressing, the average original tendon lock off force, the wire stress relaxation percentage and the calculated clastic shortening loss, as well as other specific considerations for TMI-1.

Normalization factors are calculated at the time the. individual tendon force loss cunes are generated and are documented in the Reference 4 cahulation. The values are included within the current revision of the surveillance procedure and are shown on every individual tendon force curve. The expression for the basic normalization factor used for TMI-l is as follows:

NFj (t) = [{F (0)- F i(0)} * { l - Sr (t)} + ES * {(N - 2n + 1) / 2N)]

Where:

NF i(t) is the normalization force of a particular tendon (kips).

F (0) is the Average Tendon Force for le Group (kips).

Fj (0) is the Original Average Force for the Tendon (kips).

Sr (t) is the stress relaxation (% /100) for the wire at time t, from test data.

ES is the Total Elastic Shortening Loss (kips) for that group of tendons.

N is the total number of stressing sequences for the tendon group.

n is the stressing sequence for the particular tendon.

The basis for this expression was derived and is documented in the paper, " Tendon Surveillance Requirements - Average Tendon Force, by J. F. Fulton, Nuclear Engineering and Design, October 1982, Page 303 (Reference 8). Mr. Fulton was actively involved with the subject of tendon surveillance and nonnalization factors for many years and contributed to the development of the current Regulatory Guides.

l l

1 l

C311-95-2502 Attachment A

. Page 12 of18 NRC QUESTION NO. 5 Provide force-clongation information in the requested format for each detensioned and retensioned tendon at the required 1/3 and 2/3 increments as required by Regulatory Guide 1.35. Prepare graphs showing the linear force-elongation relationship for each tendon. Explain and discuss the implications if not linear.

GPU NUCLEAR RESPONSE TO QUESTION NO. 5 Tables 5.1,5.2 and 5.3 are attached and represent the force-elongation data collected in accordance with the requested NRC fonnat for each of the detensioned and retensioned tendons. This includes data and graphs for three tendons; tendon D248, V-78, and H35-47 which were detensioned and retensioned in the surveillance. A legend and list of&fmitions used for the data gathered for these tables is provided as follows:

1xgend:

PTF Force : The PTF force is used as the basis for the elongation measurement and is the pretensioned force necessary to bring the tendon into a lightly stressed condition to remove slack and seat the buttonheads. For TMI-1, a force equivalent to 1000 psi ram pressure was used as the basis for the PTF.

1/3 Increment: The 1/3 increment is the tendon force determined at a gauge pressure equivalent to 1/3 the overstress force pressure as required by Regulatory Guide 1.35, Subsection 4.2 2/3 Increment: The 2/3 increment is the tendon force determined at a gauge pressure equivalent to 2/3 the overstress force pressure as required by Regulatory Guide 1.35, Subsection 4.2.

LOF Force: The lock-off force is that force at which the tendon load is transferred to the shim stack from the ram, and is the force left in the tendon after retensioning.

OSF Force: The overstress force is that force where maximum elongation is determined, typically at a value of 80% of the Guaranteed Ultimate Tensile Strength (GUTS) for the tendon.

Figures 5.1,5.2 and 5.3 were also prepared as requested and are attached. These figures represent the plotted force-clongation relationship for the three retensioned tendons.

The resulting force-elongation plots show very good linearity as the tendon is retensioned through the 1/3 and 2/3 increment points and up to the OSF. The graphs of all three tendons show excellent linear correlation between the measured incremental elongation and the force in the tendon during the retensioning process. For lioop Tendon H35-47, the LOF is somewhat off the linear slope. This condition occurs because the shims have a finite thickness and can compress under load. This results in the LOF bemg a step function rather than a linear function. The LOF point is also taken after reversing the tensioning direction dovmward from the OSF, rather than during the retensioning process as all other points are established.

]

1 l

l i

1

C31195 2502 GPU Nuclear Corperation

, Attachment A P' age 13 of Ig Three Mile Island Jnit 1 20th Year Tendon Sarveillance Force and Elongation Measurement Data for Retensioned Tendon Table 5.1 TENDON: D248 PRESSURE REF. ELONGATION REF.

FORCE KIPS REF. ,

(k) (psi) (in)

Ref.1 1130 Ref.1 4.10 Ref.1 PTF 210 Row 1 Row 6 Row 7 Ref.1 2625 Ref.1 5.65 Ref.1 Step 1 495 Row 10 Row 10 Row 11 1/3 increment Avg.

5210 Ref.1 8.20 Ref. 1 Step 2 989 Ref.1 Row 12 Row 12 Row 13 2/3 Increment 1225 Table X1 6448 Table X1 9.60 Data Sht. 3 LOF Page 26 Page 26 Shim Thickness 7 Avg. Avg.

1484 Ref.1 7805 Ref.1 11.10 Ref.1 OSF Row 15 Row 9 Row 14 Avg.

I 1

Total Elongation (actual) = (LOF PTF) = 9.6 4.1 = 5.5" 12 SF )

10 '

LOF

~8 i 2/3 I ' ]

$ l j6 I 1/3 W 4 PTF ,

2

?

O 600 800 1000 1200 1400 1600 O 200 400 Force (kips)

Figure 5.1 Tendon D248 : Force-Elongation relationship during retensioning j YE AR20.XLS 1

l

. . . - . . .. . . - - . . --. - -- -- - - - - - - - - ~ ---

.i o.'  : ,

3 j  : C311-95-3503 l GPU Nuclear Corporation

' Attachment A  ;

~

P&ge 14 of18 Three Mile leland Unit 1 :i'

- 20th Year Tendon Surveillance ,

r

!: Force and Elongation Measurement Data l

for Setensioned Tendon ,

s-l 3

. Table 5.2 j TENDON: V78 i FORCE KIPS REF. PRESSURE REF. ELONGATION REF.

4 l I (k) (psil (in)-  !

j' PTF 210 Ref.1 1000- Ref.1 4.50 Ref.1  :

I' Row 1 Row 6 Row 7 l J Step 1, 493 Ref.1 2340 Ref.1 7.25 Ref.1  ;

1/3 increment Row 10 Row 10 Row 11 i

): Avg. 1 Step 2 739 3500 Ref.1 9.40 Ref.1 [

. Ref.1 Row 13 p- 2/3 increment Row 12 Row 12 j L Avg.  ;

LOF 1347 Table X1 6383 Table X1 15.00 Data Sht. 3 ,

} '

L Page 26 Page 26 Shim Thickness i N.

OSF 1478 Ref.1 7000 Ref.1 16.30 . Ref.1

[ Row 14 1 Row 15 Row 15 Total Elongation (actuatl = (LOF - PTF) = 15.0 - 4.5 = 10,5" -

4 l

. l 1 l

j. 18 I I

i l l l *OSF I 16 14 i [ g,i

, If i 8

~ 6 i 4 V PTF i i I l

2

^

l  ! l l l  ! l 4

0 0 200 400 600 800 1000 1200 1400 1600 Force (kips)

Figure 5.2

]

Tendon V78 : Force Elongation relationship during retensioning i

YEAR 20.XLS

- - . .~. _.

1 s

C31195 2502 GPU Nuclear Corporation

.- Attachment A [

Psge 15 of 18 Three Mile island Unit 1 Lf

- 20th Year Tendon Surveiller.co .  !

Force and Elongation Measurement Data for Retensioned Tendon  ;

i Table 5.3 1 r

TENDON: H35-47 i I

FORCE KlPS REF. PRESSURE REF. ELONGATION REF.

(k) (psi) (in) l' PTF 210 Ref.1 1130 Ref.1 7.30 Ref.1 Row 1 Row 6 Row 7 Step 1 521: Ref.1 2755 Ref.1 9.60 Ref.1 l 1/3 increment Row 10 Row 10 Row 11 Avg. j Step 2 1042 Ref.1 5490 Ref.1 13.35 Ref.1 2/3 Increment Row 12 Row 12 Row 13 l Avg.  ;

LOF 1219.5 Table X1 6419 Data Sht. 2 15.90 Data Sht. 3  !

Page 26 Page 25 Shim Thickness Avg. Avg.

OSF 1564 Ref.1 8225 Ref.1 17.45 Ref.1 i Row 15 Row 9 Row 14 --,

Avg. ,

l Total Elongation (actual) = (LOF - PTF) = 15.9 - 7.3 = 8.6" l

I 18 16

- 12 I

to i / 1 i,6 w

F I l

l  !

2 '

l l  ! I O

C 200 400 e00 800 1000 1200 1400 1600 Force (kips)

Figure 5.3 j Tendon H35 47 : Force-Elor.gation relationship during retensioning YEAR 20.XLS

C311-95-3502 l Attachment A

, Page 16 of 18 .;

NRC QUESTION NO. 6 Based on the informa: ion in Table II, and in the Summary of Data Sheets SQ 6.1 contained in the PSC report, you inspected only the grease caps of the upper ends of the vertical tendons. Pro ide ,

q ' assurance and bases that there is no water in the lower grease caps of the vertical tendons.

GPU NUCLEAR RESPONSE TO QUESTION NO. 6

i

^

The referenced Table II (Reference 2) is a summary of numerous inspections made for the presence of water in the tendon conduit. The shop end of the vertical tendons is at the top of the tendon as ,

designated in the Table. The field end is the bottom of the vertical tendons and the results of the inspection of the field end are also tabulated in the same Table. Thus, the lower grease caps of the vertical tendons were inspected. The results of the various inspections determined that none of the j vertical tendons had any water present during this surveillance period.

The surveillance procedure requires that' individual tendon data inspection sheets,6.1, be

, completed for each tendon. These sheets were completed for each venical tendon end and are presented in Appendix A, pages A3 to A8 of the PSC report. The inspection results indicate no water was detected at either the upper or lower ends of all vertical tendons inspected. In fact, j examination of all data sheets from pages A3 to A24 indicates that no water was found in any tendon as a result of this surveillance inspection.

]

Similar results, data sheets and conclusions are presented for other tendons in Tables XV and XVI in the PSC Report (Reference 2). Note that Table I of the PSC Report also documents the results oflaboratory analyses of the sheathing filler for chlorides, sulfides, nitrati, and % water content.

J Results for the water content % indicate an insignificant amount of water (< 0.10 %) was found in i all samples in the Table.

i 1

I

. r 4

4 k

s J

~ .

C311-95-2502 Attachment A

,. Page 17 of18 NRC QUESTION NO. 7 What significance does the anchorage thread measurement data, as presented in the Summary Report and Data Sheets, have on the results of the tendon lift-off.

GPU NUCLEAR RESPONSE TO QUESTION NO. 7 Here is no significance of the measurement of anchorage thread diameters on the results of tendon j lift-off testing. De purpose of PSC Procedure SQ 7.1, Data Sheet 8 is to assure that the  !

anchorage and specific stressing adapter meet the minimum strength requirement of 110% of the l minimum Guaranteed Ultimate Tensile Strength of a tendon. He results of this inspection as - i performed for this surveillance are presented in Table III of the PSC Report.

?

The procedures and related text within the' surveillance procedure provide for the complete  ;

assurance of thread engagement during the lift-off, retensioning and detensioning process. The l procedures and data sheets were further discussed with Mr. Ron Hough of PSC Corporation. Mr.  ;

Hough explained that he had first-hand knowledge of earlier problems with the inryco system  ;

threads because the Military Specification bases for the threads was in error. As a result of a 1991 l failure ofinternal threads on one of the tendon anchor heads at Oconee 1, the NRC issued i information Notice 91-80. Computer analyses have been performed and completed on thread '

designs. The PSC procedures, as presented in the GPU Nuclear procedure assure that this will not occur at TMI-l during any surveillance. He only affect these procedures and data measurement l activities have on the lift-off process is that it assures a safe working environment and prevents  !

inadvertent damage to the tendon anchorage components. j 1

l l

]

. . - . . - m - ,. r ., i. - . n i w --r .. - .. _ _ _ . . , , , _ , - - , . --. _ ,

. . . . - ~ . . - .. . . . -. .

I' -

~

6

.,; C311-95-2502 .  ;

Attachment A -

Ppge 18 of 18 -

NRC QUESTION NO. 8 -

i Reference Table IV, Summary of Data Sheets' 1,2,3 and 4. Corrosion conditions as presented for the shims indicate nearly all shims have visible oxidation. Explain why such conditions exist j although the shims are enclosed in grease cans.

i

~ GPU NUCLEAR RESPONSE TO QUESTION NO. 8 ,

+

The data presented on corrosion in the reference Table IV indicate that level 2 corrosion is j commonly noted for the shims inspected in this survei!%ce. Specific data sheets which were used  !

to record the inspection data are included on pages A45, A46 and A47 in the Reference 2 PSC  !

report. Corrosion levels are defmed in Table 2 of Enclosure 6 in Surveillance Procedure 1301-9.1. l Corrosion level I is defined as " Bright metal, no visible oxidation, and corrosion level 2 is defined j

' as " Metal reddish brown, no pitting ,

To investigate this condition, previous surveillance reports were reviewed from the fifth, fourth and .l Jfirst surveillance periods. Data sheets 6,7 and 8 of the Reference 9 and 10 Reports record the ' -l corrosion levels of shims and were reviewed. Data sheets 1,2, and 3 of the Reference 11 Report t record shim corrosion levels and were also reviewed. It was determined that these same corrosion  !

levels existed for all the tendons listed in each of these past surveillance reports. Furthermore, it I was determined that these same corrosion levels are commonly noted for other anchorage l components inside the cap, such as the stressing washer and the buttonheads.

Based on the above findings and the knowledge of the field conditions and long duration it took to j typically install the prestressing system, it is cc ncluded that this level of corrosion has been present  !

since the time of original installation. There have been no incidents in the past where prior j surveillances have uncovered unusual conditions related to the components inside the tendon caps. l Also, there has been no increase in corrosion levels or problem wire test results. Therefore, it is concluded that these low corrosion levels have existed for some time and have not propagated. The j effects on the integrity of the anchorage or tendon as a whole are negligible. j i

l 1

i.  !

. i e

I

j. -

l

_ __J