Information Notice No. 86-102: Repeated Multiple Failures of Steam Generator Hydraulic Snubbers due to Control Valve Sensitivity

                                                            SSINS No.:  6835
                                                            IN 86-102 

                                UNITED STATES
                        NUCLEAR REGULATORY COMMISSION
                    OFFICE OF INSPECTION AND ENFORCEMENT
                            WASHINGTON, DC 20555

                              December 15, 1986

Information Notice No. 86-102:  REPEATED MULTIPLE FAILURES OF STEAM 
                                   GENERATOR HYDRAULIC SNUBBERS DUE TO 
                                   CONTROL VALVE SENSITIVITY 

Addressees: 

All nuclear power reactor facilities holding an operating license or a 
construction permit. 

Purpose: 

This notice is provided to alert recipients of a potentially significant 
safety problem pertaining to recent events in which the steam generator 
hydraulic snubbers failed to meet their bleed and lockup specifications at 
two consecutive refueling outages. The primary cause appears to be control 
valve sensitivity to low hydraulic fluid flow velocity. It is expected that 
recipients will review the information for applicability to their 
facilities. However, suggestions contained in this notice do not constitute 
NRC requirements; therefore, no specific action or written response is 
required. 

Description of Circumstances: 

On January 7, 1986, the Portland General Electric Company reported (Licensee
Event Report (LER) 85-13) multiple snubbers which failed to meet their bleed
and lockup specifications at its Trojan Nuclear Plant. The report, and its 
supplement dated April 1, 1986, identified three areas of multiple snubber 
failures that were discovered during the 1985 refueling outage that began in
May, 1985. These snubber failures were discovered as a direct result of the 
expanded inservice testing program which was instituted in accordance with a 
recent change to the plant's technical specifications. The prior inservice 
inspection program had not required the testing of these snubbers. 

The 16 steam generator hydraulic snubbers at Trojan are 900-Kip Anker-Holth 
units. Following the failure of the first 2 steam generator snubbers to meet
their bleed and lockup specifications, the remaining 14 were declared 
inoperable because of uncertainty regarding the time required to rebuild the 
snubbers following testing. All the snubbers were removed and overhauled. 
During the overhaul, the snubber seals were found to be degraded and the 
hydraulic fluid was heavily contaminated with seal material and rust. 
However, as discussed below, this was not the primary cause of the problem 
detected during the subsequent 1986 outage. 

8612100135 

.

                                                       IN 86-102 
                                                       December 15, 1986 
                                                       Page 2 of 4 

An engineering evaluation of the effect of the failed snubbers on the steam 
generators was not initiated during the 1985 refueling outage because of the
belief that they would not have restricted normal thermal growth. This 
decision was based on the snubber-testing company's judgment that the 
foreign material in the hydraulic fluid would not have affected the normal 
operation of the snubbers because of the relatively large channels through 
which fluid would flow under thermal growth conditions. However, in the case 
of a seismic or other severe dynamic event, it was determined that the 
snubbers would have activated (i.e. , locked up*) and the foreign material 
could have blocked the bleed orifice. Because of the manner in which the 
snubber control valves are hydraulically interconnected, it is the 
licensee's belief that this would have to occur to all four snubbers on one 
steam generator before they would become locked in their current position.** 
They would remain in this position until a load reversal allowed flow 
through the main valves or possibly cleared the bleed port in at least one 
snubber. 

The revised technical specifications for testing the snubbers required 
testing of each snubber that had failed its test during the previous testing 
program. Therefore, the 16 steam generator hydraulic snubbers were again 
tested during the refueling outage that began in April, 1986. The results of 
this testing indicated 12 failures--4 with excessive drag, 4 with high bleed 
rates at faulted load, 2 with no bleed rate at faulted load, 1 with 
excessive drag and high bleed rate, and 1 with high bleed in compression and 
no bleed in tension. The snubbers with no bleed rate cleared themselves upon 
load reversal. 

There also was an issue of unusual movements of the pressurizer surge line 
that was thought for a while to be related to the snubber problems. This is 
discussed in Attachment 1. 

Discussion: 

As a part of its corrective actions during the 1985 refueling outage, the 
licensee had all the steam generator snubbers overhauled. Following 
overhaul, the snubbers could not meet their safety analysis acceptance 
criteria of a 


*         Note:Common snubber nomenclature uses the term "lock up" to refer 
          to (1) that point where the main flow path is closed and all flow 
          is forced through the bleed orifice and (2) the condition where 
          all flow is stopped and the snubber becomes a rigid strut. To 
          eliminate any possibility for confusion between the two meanings, 
          the term "activated" will be used for the first definition. 

**        In their safety evaluation report (Steven A. Varga's May 30, 1986,
          letter to Bart Withers) the Office of Nuclear Reactor Regulation 
          staff concluded that "...the likelihood of full thermal lock-up  
          occurring would require that the various contributing factors 
          would have to affect three or four of the hydraulic snubbers on a 
          single steam generator." 
.

                                                       IN 86-102 
                                                       December 15, 1986 
                                                       Page 3 of 4 

maximum drag force of 1,000 pounds at a minimum displacement rate of 0.025 
in./min. This was because the snubber activated each time the velocity 
approached 0.025 in./min. Following consultation with the reactor vendor, 
the acceptance criteria was revised and the snubbers tested satisfactorily. 
 Because of the reoccurring snubber failures identified during the 1986 
refueling outage, the licensee contracted for a detailed root cause analysis 
of the snubber failures. This analysis indicated that the low activation 
velocity (0.025 in./min) of the steam generator snubbers caused them to 
activate at very low fluid velocity through the main flow port. Once the 
snubber had activated, all flow was forced through the bleed port. Because 
of its extremely small size, this port acted much like a fine sieve. 
Apparently the first particle of foreign material would block the port 
causing the snubber to lock up. Thus, although contamination of the 
hydraulic fluid was a contributor to the problem, it was not the primary 
cause. 

Based on this root cause analysis, the licensee decided to continue with its
previously made plan to change out the control valves on the steam generator
snubbers. The new snubber control valves have a much higher activation 
velocity (6 to 9 in./min) which is still acceptably small compared with that 
expected during any significant seismic event. In addition, the new snubber 
control valves incorporate a widely used "self-cleaning" poppet valve design 
as opposed to the original spring-ball check valve design. In the new 
design, the bleed orifices are grooves on the main poppet valve. In this 
way, the bleed orifices tend to be self-cleaning whenever there is flow 
through the main poppet valve. 

All of the Anker-Holth steam generator snubbers were initially designed with
relatively low activation velocities. Therefore, they are suspected of 
having the same type of problems as encountered at Trojan. In addition to 
Trojan, three other utilities have modified their steam generator snubbers 
so that they have activation velocities in the 6 to 10 in./min range. 

However, since the root cause of the problem is the selection of an 
extremely low activation velocity, as opposed to a design flaw in the 
snubbers themselves, the problem may not be limited to only the facilities 
having Anker-Holth snubbers. 

Attachment 2 to this information notice describes other multiple snubber 
failures found at Trojan during the 1985 refueling outage. 

.

                                                       IN 86-102 
                                                       December 15, 1986 
                                                       Page 4 of 4 

No specific action or written response is required by this information 
notice, If you have any questions about this matter, please contact the 
Regional Administrator of the appropriate regional office or this office. 


                                   Edward L. Jordan, Director
                                   Division of Emergency Preparedness
                                     and Engineering Response 
                                   Office of Inspection and Enforcement 

Technical Contact:  Richard J. Kiessel, IE 
                    (301) 492-8119 

Attachments: 
1.   Pressurizer Surge Line Movements 
2.   Other Multiple Snubber Failures 
3.   List of Recently Issued IE Information Notices 
.

                                                       Attachment 1 
                                                       IN 86-102 
                                                       December 15, 1986 
                                                       Page 1 of 1 

                      Pressurizer Surge Line Movements

In what had been (until midway through the 1985-86 fuel cycle) a separate 
issue, the licensee had been monitoring the unusual movements of the 
pressurizer surge line since 1982. A walk-down of this line at the beginning 
of the 1985 refueling outage revealed additional movement had occurred 
during the last fuel cycle. A consultant was hired to evaluate and analyze 
these movements, and had determined that none of the previously identified 
potential causes, whether singly or combined, could have produced the 
observed movement. However, when he was advised of the possible problems 
with the steam generator snubbers, his worst case analysis (i.e., all 
snubbers on one steam generator were locked-up) indicated that locked-up 
snubbers could have produced the observed movement. This discovery delayed 
the submittal of LER 85-13, which was,being prepared at the time. 

Testing associated with the root cause analysis demonstrated that the 
snubbers on a particular steam generator would not restrict growth of that 
loop unless all four snubbers lock-up because the snubber hydraulic lines 
were connected in parallel. In addition, based on the results of the thermal 
expansion monitoring program conducted during the startup from the 1986 
refueling outage, the licensee has determined that most, if not all, of the 
observed movement of the pressurizer surge line is expected due to normal 
thermal transients experienced by this line during heatups and cooldowns. 
Based on these findings, the licensee further concluded that the most likely 
cause of the reactor coolant system thermal restraint was due solely to the 
inadequate size of the gaps between system components and associated seismic 
or pipe whip restraints.  

.

                                                       Attachment 2 
                                                       IN 86-102 
                                                       December 15, 1986 
                                                       Page 1 of 1 

                       Other Multiple Snubber Failures

In addition to the steam generator hydraulic snubber failures, the Trojan 
LER identified two other areas of multiple snubber failures. Although not 
the subject of this information notice, they are briefly discussed to assist 
in identifying all the safety-related failures discussed in the LER. 

1.   The first additional area of multiple snubber failures was a 25 percent
     overall failure rate of small mechanical snubbers (Pacific Scientific 
     models PSA-1/4 (36 percent failure rate) and PSA-1/2 (17.6 percent 
     failure rate)). 

2.   The second additional area of multiple snubber failures involved the 
     four main steam line hydraulic snubbers (two 70-Kip and two 130-Kip 
     Bergen-Paterson units). The snubbers were declared inoperable without 
     testing upon discovery of the steam generator hydraulic snubber 
     failures. 

Additional discussions of multiple snubber failures can be found in IE 
Information Notice 84-67, "Recent Snubber Inservice Testing with High 
Failure Rates," LER 84-079 for San Onofre Nuclear Generating Station Unit 2 
(dated January 25, 1985, and revised March 12, 1985), and LER 85-027 for San 
Onofre Nuclear Generating Station Unit 3 (dated May 16, 1985).