ML20215G653

Trojan Nuclear Plant Evaluation of Main Feed Line Seismic Restraint Failure

ML20215G653
Person / Time
Site:	Trojan File:Portland General Electric icon.png
Issue date:	06/11/1987
From:	BECHTEL GROUP, INC.
To:
Shared Package
ML20215G104	List: ML20215G101 ML20215G650 ML20215G653
References
TAC-65471, NUDOCS 8706230340
Download: ML20215G653 (8)

Text

- _ _ _ _ _ _ _ _ - _ _ _ _ _ _

t.

(- TROJAN NUCLEAR PLANT EVALUATION OF MAIN. FEED LINE SEISMIC

. RESTRAINT FAILURE PREPARED FOR

PORTLAND GENERAL ELECTRIC COMPANY i

i BECHTEL WESTERN POWER CORPORATION SAN FRANCISCO, CALIFORNIA I

f JUNE 11, 1987 Originator i &

1

)

Checker (

!1 8706230340 DR 870612 '

p ADOCK 05000344

, PDR ,

b l

h.

l TROJAN NUCLEAR PLANT 1 EVALUATION OF MAIN FEED LINE SEISMIC.

, RESTRAINT FAILURE PREPARED FOR PORTLAND GENERAL ELECTRIC COMPANY BECHTEL WESTERN POWER CORPORATION  ;

SAN FRANCISCO, CALIFORNIA l

'l JUNE 11, 1987  ;

Originator i &

1 Checker ( M

> , 4 8706230340 DR 870612 '

l p ADOCK 05000344

) PDR ,.

TROJAN NUCLEAR PLANT EVALUATION OF MAIN FEED LINE SEISMIC RESTRAINT FAILURE

1. BACKGROUND During the 1987 refueling outage, a seismic restraint (EBB-3-1-SR8) on the Loop B feedwater piping inside containment was found to have a failed structural attachment to the concrete. The restraint had been inspected during the 1986 outage and was found to be intact at that time.

EBB-3-1-SR8 is a seismic restraint located on the loop B feedwater piping inside containment. The restraint is located on the first vertical run of piping inside containment. The design loadings for SR8 are 0.642 kips (thermal), 5.246 kips (0BE) and 8.761 (SSE). The restraint is orientated

) in the east-west direction and consists of a pipe clamp, a rigid sway strut i I (Bergen-Paterson RSSA-1$ an end attachment, a supplementary steel frame structure (W6x20 vertical member, W4x13 knee brace) with two 1/2" base plates each attached to the concrete with 4-5/8" phillips red head concrete expansion anchor bolts. The failure consisted of pullout in one irregular shaped cone of the concrete under the baseplate for the knee brace and broken concrete under the vertical member baseplate. j The evaluation of the failure concluded that the load causing failure was tension in the strut assembly which produced combined compression and shear on the W6 vertical member base plate and combined tension and shear on the W4 knee brace member base plate. The equivalent static failure load was estimated to be not less than 40 kips at the baseplate.

2. IDENTIFICATION OF POSSIBLE ROOT CAUSES The following events which might have contributed to a significant load in the feedwater system were investigated: .

a. Thermal loads.

b. Steam generator restraint binding or snubber locking.

c. Feedwater piping snubber locking.

d. Containment expansion due to ILRT pressure.

e. Excessive preload on strut.

f. Low flow thermal stratification / pipe bowing.

g. Anchor bolt or concrete failure with normal loads. -

h. Water hammer from various sources.

Later in the report, the most likely events to cause failure of seismic restraint SR8 will be identified and the probable causes determined.

f

3. EVALUATION OF POSSIBLE ROOT CAUSES The above events were investigated to determine their likelihood to result in the observed failure of SR8, with results as follows:

a. Normal operation, startup, shutdown and scram thermal operating modes were reviewed and enveloping load cases were reanalyzed.- The resulting thermal loads on SR8 were small in comparison to its capacity. The maximum load for these conditions on SR8 was found to be 778 pounds.

b. The effects of hypothesized binding of steam generator restraints or locking of steam generator snubbers were analyzed. The resulting loads on SR8 were small in comparison to its capacity with a maximum load on SR8 of 442 pounds. Additionally, all steam genertor snubbers were tested after the failure of SR8 and met the drag and actuation acceleration acceptability requirements.

c. The possibility of feedwater piping snubbers malfunctioning was considered. The resulting loads on SR8 were small in comparison to its capacity. These loads on SR8 were found to be a maximum of 2350 pounds. Additionally, all snubbers on the feedwater piping inside containment were tested after the failure of SR8 and met the drag and actuation acceleration acceptability requirements.

i

d. The expansion of the containment due to the Integrated Leak Rate Testing (ILRT) pressure was analyzed and the resulting load on SR8 was much less than its capacity. The maximum load on SR8 from ILRT was found to be 1743 pounds.

e. The strut was reported to have been installed in the slack condition, free of preload,

f. Low flow thermal stratification due to Auxiliary Feedwater System flow during hot standby or shutdown conditions and hotter feedwater flow into the colder feedwater line during startup was considered.

Since the intervening spring hangers did show signs of vertical displacement and the failure was to the lone rigid restraint on the system, the significance of this type of event was analyzed in more detail.

/ The actual signs of unanticipated vertical movement on the spring hangers I were as follows:

i) Spring Hanger EBB-3-1-H8 has marks on the internal shank that indicate that the pipe has moved 3 inches up and 5/8 inch down from its cold position. The maximum calculated thermal movement (without considering stratification) is 0.563 inches up during startup or shutdown and 0.163 inches up during normal operation.

ii) Spring Hanger EBB-3-1-H9 has marks that indicate the pipe has moved up 2-3/8 inches from its cold position. The maximum calculated thennal movement is 0.125 inches up during startup or shutdown and 0.14 inches up during normal operation.

iii) Spring Hanger EBB-3-1-H10 has chatter marks on the inside of the spring can that indicate that the pipe has moved 1-3/4 inches up and 5/8 inch down from its cold position. The maximum calculated thermal movement is 0.022 inches down during startup or shutdown

, and 0.288 inches up during normal operation.

l An analysis was performed with an assumed piping top-to-bottom thermal gradient of 100F for the entire 82.24 ft. run of horizontal piping inside containment. The analysis revealed that SR8, because of its close proximity to the containment flued head, acts as a couple with the flued head to restrain the thermal bowing affects. At 100F differential temperature, the piping would bow approximately 1.283 inches upward at spring hanger H8, and the loading on SR8 would be 5.5 kips. We are continuing to evaluate this condition in conjunction with a bubble collapse scenario, the details of which are outlined in 1. below.

g. A review of the failed concrete by PGE determined that it was most probably either a single static load or a series of dynamic loads that failed the anchorage. No single static load of sufficient magnitude to produce the observed failure could be identified.

) h. An event identified that could generate forces large enough to cause I the damage observed was a water hammer or other similar hydraulic transient. The effects of such an event are discussed later in this report. It was noted that design provisions to prevent water hammer due to steam condensation in the steam generator feed ring had been performed on Trojan (including J-tube installation and feed line layout adjacent to the nozzle).

1. An on-going review of plant records for the 1986-1987 cycle has not identified any unusual event that could have resulted in a significant water hammer. These records include Steam Generator level, pressure and temperature charts, feedwater flow, startups, shutdowns and trips. In addition, interviews with reactor operators have revealed no unusual events. The possibility that the event would have been the result of a " Bubble Collapse" in the proximity of the Steam Generator is being examined in more detail.

4. PIPING STRESS ANALYSIS FOR DESIGN CONDITIONS /0PERABILITY ANALYSIS WITH SR8 FAILED The feedwater piping inside containment, to both the "A" and "B" steam generators, was reviewed. The calculated noma 1, upset and faulted primary stresses and secondary stresses were well within their respective allowables.

Water hammer loads are not included in the design basis of this system and therefore were not included in the review. Resulting pipe support loads were reviewed and found to be within design allowables. The piping was analyzed to determine if the system would remain operable without SR8 acting as a restraint. The calculated stresses in the piping system were within design allowables without SR8, and other hanger loads were not significantly increased. Therefore the piping system was operable without SR8, for design conditions.

5. PIPING ANALYSIS FOR A HYPOTHETICAL WATER HAMMER

] In order to understand the piping response to a hypothetical water hammer, j a linear elastic time history analysis was perfomed. The analysis included two assumed water hammer characteristics and two restraint configurations as follows:

l RESTRAINT l l CONFIGURATIONS l l With I Without l l SR8 l SR8 l l l 100 psi dP l l l l Water hammer 1.001 second I x l x l l characteristics l 100 psi dP l l l l l .01 second l x l x l In all cases, the assumed water hammer differential pressure was 100 psi.

The water hammer was assumed to propagate through the piping system at sonic velocity. The 100 psi value was chosen because of the ease of scaling the results to other values of differential pressure. The two pressure rise times of .001 seconds and .01 seconds were chosen to evaluate the sensitivity of the piping to the frequency characteristics of a water hammer. A longer rise time would result in lower water hammer forces in many piping segments but would also lower the frequency content of the force such that it may be more "in-tune" with the piping system frequencies and therefore result in higher piping system response. The piping was evaluated with and without SR8 active to determine the effect SR8 failure would have on the system's capability to resist water hammer type loadings.

6. HYDRAULIC ANALYSIS A hydraulic model was prepared of the affected feedwater line from the l common header upstream of the FW control valve to the steam generator.

An analysis was performed to establish water hammer forces resulting from a normal plant trip with realistic closure data for control valve, isolation valve and check valve. Additional runs were made to determine peak forces resulting from plant trips with a stuck open check valve, or with inadvertent delays in control valve and isolation valve closures.

Further analyses are being perfomed to develop water hammer forces resulting ,

from the condensation of postulated steam bubbles in the feedwater line or  !

steam generator feed ring.

i

7. RESULTS TO DATE  !

{

The results of the analyses to date can be summarized as follows:

a. No static or nomal design load event was identified as contributing to the failure of SR8.

b. The system was functional and operable for all design conditions with SR8 failed.

c. Thermal bowing may exist during system startup, shutdown and hot standby that may result in a static tension load on SR8. This loading, although quasi-transient, could coexist with a postulated water hammer loading.

This load alone would not be of sufficient magnitude to cause the observed damage.

d. The dynamic response of the piping system was similar with and without SR8 active. The maximum stress values were similar for both cases,

with and without SR8 active.

e. Restraint loadings were generally higher with the faster (.001 seconds) assumed rise time,

f. The support loadings and pipe stresses for an assumed water hamer of 100 psi differential pressure were reasonably low. The maximum calculated support loading (at SR8) is 6.4 kips and the maximum calculated pipe stress was 3,800 psi.

g. If a water hammer were assumed, SR8 would be the first restraint to indicate signs of unanticipated water hamer type loadings. The analysis showed that the other restraints on the piping would be subjected to similar loadings. Since SR8 has a lower anchorage capacity and is the .

only rigid support in the system, it would be expected to be the first to fail. Further, SR8 is the only support that could see additive thermal stratification loadings.

h. Although the piping system was not explicitly designed for water hammer loadings, the overall scheme of the dynamic restraints on the piping adequately restrains piping dynamic response due to water hammer type loadings which may have caused failure of SR8. The capacity of the dynamic restraints is such that this type moderate of water hamer would be effectively contained.

1. The relatively low frequency of the feedwater system provides significant isolation from adverse water hamer impacts.

j. The failure of SR8 indicates that a moderate level water hammer did occur.

The parametric analysis performed and the observed limited extent of damage support the position that a water hamer did occur. An event of this type would not have yielded the piping. The magnitude of the postulated water hammer loading could be significantly smaller if the loading coexisted with a thermal stratification in-the system. A water hamer of a much greater magnitude, sufficient to cause more widespread damage, was not considered credible as:

1) no physical signs of such an event exist ii) the damage was limited to anchor bolt pullout on one restraint that was clearly the weakest as well as the only rigid restraint in the system.

k. Results of hydraulic modelling as described in paragraph 6, of feedwater response to two system trips experienced in the last year yielded-loads' i on SR8 which were small in comparison to its capacity. Preliminary  !

results reveal that the maximum load on SR8 would be less than 3 kips.

~

8. CONCLUSI0t2

a. TM damage to SR8 likely occurred as a result of a water hammer, I wall within the capability of the piping system. The anchorage of SR8 was the limiting component of the system.

i

_ a

b. A water hammer of large magnitude likely did not occur based upon the physical evidence.

c. The damage to SR8 did not affect the integrity of the feedwater system.

d. SR8 may be subjected to loadings due to thennal stratification during hot standby and shutdown. These loadings could coexist with a postulated water hammer.

9. RECOMMENDATIONS

a. Restraint SR8, although not specifically required for system operability, should be returned to service.

b. A monitoring program should be implemented to confirm that the loadings acting on the system are indeed within acceptable limits and to aid in determining the root cause of the SR8 failure.

\

(

_ _ _ _ _ _ _ _ _ _ _ _ __