DCL-86-027, Forwards Rept on Anchor Darling A/D 5500 Snubbers Installed on Piping Sys,Per 860109 Request Re Part 21 Reportability. Facility Operability Unaffected by Excessive Drag Characteristic Exhibited by Snubbers

Forwards Rept on Anchor Darling A/D 5500 Snubbers Installed on Piping Sys,Per 860109 Request Re Part 21 Reportability. Facility Operability Unaffected by Excessive Drag Characteristic Exhibited by Snubbers

ML20214C817
Person / Time
Site:	Diablo Canyon
Issue date:	02/04/1986
From:	Shiffer J PACIFIC GAS & ELECTRIC CO.
To:	Martin J NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION V)
References
DCL-86-027, DCL-86-27, NUDOCS 8602210283
Download: ML20214C817 (5)

Text

,

PACIFIC GAS AND ELECTR,IC C O M PANY 9@7W3 { 77 BEALE STREET . SAN FRANCISCO, C ALIFORNI A 94106 . (415)781-4211 . TWX 910 372-6587 February 4, 1986 JAMES D. SHIFFER vics PassocENT PGandE Letter No.: DCL-86-027 Mr. John B. Martin, Regional Administrator IT; U. S. Nuclear Regulatory Commission, Region V 1450 Maria Lane, Suite 210 #1

['g Halnut Creek, CA 94596-5368 _

Re: Docket No. 50-275, OL-DPR-80

?

m m

J

3 e

';5 Docket No. 50-323, OL-DPR-82 'n

• Diablo Canyon Units 1 and 2 9 Report on Anchor Darling A/D 5500 Snubbers a co

Dear Mr. Martin:

In response to a January 9, 1986 Region V request, enclosed is a report on Diablo Canyon Units 1 and 2 Anchor Darling A/D 5500 series snubbers which exhibit an excessive drag characteristic under certain operating conditions.

This report discusses the investigation into the cause, the solution, and the safety significance of the Anchor Darling A/D 5500 series snubber excessive drag condition. During the investigation, snubbers which exhibited this drag characteristic were evaluated and replaced or restored to ensure their operability in accordance with the requirements of DCPP Units 1 and 2 Technical Specification 3.7.7.1.

As indicated in 'he enclosed report, evaluations of the Anchor Darling A/D 5500 excessive drag characteristic resulted in the conclusion that operability of Diablo Canyon Units 1 and 2 is not affected. On January 10, 1986, Anchor Darling was requested to review this snubber drag condition for reportability i

under 10 CFR 21 requirements. By letter dated January 20, 1986, Anchor Darling responded that this excessive drag condition has been classified as a product deviation, and all customers that require notification have either been notified or will be notified in the near future.

l Kindly acknowledge receipt of this material on the enclosed copy of this letter and return it in the enclosed addressed envelope.

Sincerely, l

J. . hiffer Enclosure cc: L. J. Chandler R. T. Dodds B. Norton 8602210283 860204 H. E. Schierling ADOCK 0 25 S. A. Varga {DR CPUC Diablo Distribution 0676S/0041K/DH0/204 II

r PGandE Letter No.: DCL-86-027 ENCLOSURE REPORT ON ANCHOR DARLING A/D 5500 SNUBBERS Diablo Canyon Units 1 and 2 have 127 Anchor Darling A/D 5500 series snubbers installed on piping systems. These snubbers are large units rated at 55,000 pounds and installed primarily on large diameter piping in the feedwater, component cooling water, residual heat removal, and main steam systems. A small percentage of these snubbers exhibit an excessive drag characteristic under certain operating conditions. This report provides a discussion of the issues concerning the A/D 5500 snubbers.

Backaround An A/D 5500 snubber was first observed to be restraining the pipe thermal movement during the Diablo Canyon Unit 2 post-fuel-load reactor coolant system heatup walkdowns. During the Unit 2 Power Ascension Test Program, additional A/D 5500 snubbers were observed to exhibit excessive drag. As a result of these findings, a program was developed to moniter A/D 5500 snubbers to assure that the piping was not adversely affected. This program also provided .

additional data on the significance and cause of the excessive drag.

Excessive drag is defined as a snubber that is restricting free thermal movement of the pipe in excess of the original specified drag (2% of rated load).

The first snubber to clearly exhibit excessive drag in its installed configuration was removed and shipped to Anchor Darling's facility in Kulpsville, Pennsylvania. This unit was extensively tested and in.spected by PGandE and manufacturer personnel, using Anchor Darling's standard procedures, to determine if there was a hardware problem with the snubber. All testing and inspection indicated that the snubber met its design requirements and was nct defective. A complete review was performed by PGandE and manufacturer personnel of the support designs associated with snubbers exhibiting excessive drag. This review revealed no problems.

As a result of installed A/D 5500 snubbers continuing to exhibit excessive drag, additional snubbers were tested at Diablo Canyon using PGandE's hydraulic test stand with a representative from Anchor Darling present. This effort provided the first successful attempt to produce a high drag force under test conditions. This effort showed t'nat the problem occurred when the snubber was subjected to a low travel veloc'.ty (0.5 to 1.0 inch per minute).

In all observed cases, the excessive drag load was found to relieve itself as the external force is increased or reversed. Testing of installed snubbers indicated release at 18% of rated load. However, data taken during conservative bench tests indicated that this value may be as great as 15,000 pounds, 27% of the rated load. In 75% of the bench test cases, the drag force was relieved at less than 15% of the rated load.

0676S/0041K Cause and Solution The ability to reproduce the excessive drag force on the test bench eliminated external factors (e.g., pipe support design, system function, etc.) as the cause. It also showed that high drag is usually observed only at low velocities associated with an immediate change in the direction of travel.

The standard production testing performed at Anchor Darling was a full-stroke drag test with a velocity of 1-1/2 to 2 inches per minute. Therefore, this problem was not detected by the manufacturer's standard testing. The Anchor Darling investigation into the cause has isolated the problem to the ball nut mechanism in the ball return tube area. Minor irregularities in the ball return tube entrance and exit cause the balls to skid instead of roll at the tube entrance, resulting in increased friction within the unit.

An engineering evaluation and testing performed by Anchor Darling have revealed that minor finish grinding of the ball return tube entrance and exit, combined with an increase in the total ball bearing vacancies in the ball nut mechanism from 5 to approximately 10, effectively eliminates the potential for excessive drag.

Safety Sionificance Evaluation A review of the information available to date indicates that A/D 5500 snubbers -

may, under certain conditions, exert a high drag force on the piping system.

The frequency of occurrence is low. Inspections, walkdowns, tests, and engineering evaluations performed to date indicate that this potential condition does not impact plant safety. The following summarizes these considerations:

1. During the present period of evaluation and corrective action, a program of increased snubber inspection remains in place. All access,1ble Unit 1 A/D 5500 snubbers affected by power changes in both Units 1 and 2 are inspected when the reactor power level changes by more than 20%. This program supplements the piping walkdowns conducted to verify that the piping systems respond as analytically predicted. To date, more than 2,000 individual A/D 5500 snubber cycles have been witnessed, with only 20 instances of excessive drag identified. This 1% rate of excessive drag demonstrates a very low rate of occurrence.

It should be noted that the original Unit I walkdown program was identical to the one presently being completed in Unit 2, and no Unit 1 A/D 5500 snubbers were observed to demonstrate excessive drag. Recently, all accessible A/D 5500 snubbers in Unit I were again inspected with no excessive drag found. All cases of excessive drag, except one, have been identified on the Unit 2 main feedwater lines.

2. For all cases where high drag of A/D 5500 snubbers was observed, the piping analysis was rerun considering the snubber as a rigid restraint.

In all but one case, the resulting piping stresses continued to meet Diablo Canyon design criteria. The one case exceeded design criteria by

)

0676S/0041K .. .- - .- .- . - , _ .

i .

4 i less than 0.7% of the code allowable based on the use of a stress '

intensification factor of 1.6, which is judged to be conservative.

Therefore, piping qualification has not been affected by any known case

, of excessive drag.

3.

An evaluation of all inaccessible A/D 5500 snubbers located inside both containments was performed to determine the effect of a maximum hypothetical drag force on the associated piping. The systems were reviewed using engineering judgment, and where judgment could not readily

indicate system adequacy, a simplified hand calculation or an analytical

! computer analysis was performed. Some. analysis included applying a maximum hypothetical drag force of 15 kips to the pipe rather than

, modeling the snubber as a rigid support.

Snubbers located on feedwater lines inside containment, if subjected to the maximum hypothetical drag force, would not cause the pipe to exceed i Diablo Canyon design criteria. For other systems, the maximum i

hypothetical drag force caused the piping stress to locally exceed 4

thermal design criteria in some cases. However, consistent with the philosophy of the ASME Code, consideration of cyclic effects shows that

-j the integrity and operability of the system 'is maintained. Based upon the present piping support systems, in no case will the operational life

]'

of the piping or associated attached equipment be adversely affected by continued operation.

4 The vast majority of the A/D 5500 snubbers inside containment, not 1

i located on the feedwater lines, do not experience cyclic thermal movement unless the normal plant operating condition is changed to residual heat removal operation.

4 It should also be noted that in only a very small percentage.of the bench test cases has the drag force been observed to approach the maximum

, hypothetical drag force. From the testing performed to date, it is known i that snubbers with a ball nut assembly problem will relieve themselves as

' the load increases or when stroke direction is changed. Therefore, relief is anticipated at low loads such that overstressing of the piping system will not occur.

} 4. This high drag condition does not affect the operability of the snubber 3

j as a seismic limiter. The snubber will function properly under seismic conditions to the full design load. Therefore, the plant seismic design and safety margins have not been affected by this condition.

i

5. The potential for high drag exists only with the A/D 5500 size snubber.

I The other two sizes of A/D snubbers (A/D 1600 and A/D 12500), which also use a ball nut mechanism, have double lead ball screws instead of the

)

i A/D 5500 single lead ball screw. Anchor Darling has stated that the A/D 5500 is the size which is most prone to exhibit high drag during

production.

Field observations on A/D 1600 and A/D 12500 snubbers did i

not indicate any incidences of high drag. Therefore, this condition does 3

not exist on the A/D 1600 or A/D 12500 snubbers.

f I

0676S/0041K  !

i

), - ,, ,- - .-

a Corrective Actions l During the current Unit 2 outage, in preparation for commercial operation, all Unit 2 A/D 5500 snubbers will be modified and retested, as required, based on the manufacturer's revised criteria.

The Unit I snubbers will be reviewed and tested based upon the results obtained from Unit 2. Resolution will be performed expeditiously, as plant operation allows, with completion prior to the end of the first refueling outage. In the interim, the inspection and evaluation program described in the " Safety Significance Evaluation" section above will continue.

This condition has been documented by a Nonconformance Report. Additionally, in a PGandE letter dated January 10, 1986, Anchor Darling was requested to review reportability under 10 CFR 21 requirements. On January 20, 1986, Anchor Darling responded that this excessive drag condition has been classified as a product deviation, and that all customers that require notification have been notifled or will be notified in the near future.

0 0676S/0041K ,