ML20135H754
| ML20135H754 | |
| Person / Time | |
|---|---|
| Issue date: | 09/20/1985 |
| From: | Norelius C NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION III) |
| To: | Jordan E NRC OFFICE OF INSPECTION & ENFORCEMENT (IE) |
| References | |
| TAC-59416, NUDOCS 8509240198 | |
| Download: ML20135H754 (8) | |
Text
f'f u? g O
UNITED STATES
- g NUCLEAR REGULATORY COMMISSION YS REGloN HI j
7e f
799 ROOSEVELT RO AD ctEn Ettvu, stunOis som
+....
SEP 2 01985 MEMORANDUM FOR:
Edward L. Jordan, Director, Division of Emergency Preparedness and Engineering Response, Office of Inspection and Enforcement FROM:
Charles E. Norelius, Director, Division of Reactor Projects
SUBJECT:
PROPOSED IE BULLETIN:
MOTOR-0PERATED VALVE FAILURES DURING PLANT TRANSIENTS DUE TO IMPROPER SWITCH SETTINGS Per your request for comments on the subject draft bulletin, we have informally conveyed our comments to Mr. Henry Bailey of your staff on a number of drafts.
Below, for your further consideration, are our comments on the latest bulletin draft dated September 9, 1985. We have also provided two attachments: provides an analysis and evaluation of problems that led to or allowed the Davis-Besse valve failures used as an example in the bulletin and are provided to give a basis for our recommendations.
Ourmajorcomments are set forth below, and Attachment 2 lists some specific recommendations that we believe should be included in the final bulletin.
1.
We strongly disagree with limiting applicability of the bulletin to those valves identified in the action statements, " motor-operated valves in the high pressure coolant injection / core spray and emergency feedwater systems (RCIC for BWRs) that are required to be tested for operational readiness in accordance with 10 CFR 50.55a(g)." As indicated in, there are many reasons why the applicability should not be so limited. As one thought, if Toledo Edison removes the automatic signal from the valves used as an example in the bulletin, which we understand they may do, then the bulletin would not even apply to the example; further, as the bulletin is currently limited, the Sequoyah valve example is already excluded from coming under the bulletin action requirements. Obviously, the connection between the examples and application to valves in the action statement is not there.
2.
We feel it is necessary that the bulletin offer more specific details on how to assure valve operability while leaving the licensees an option to show that they either comply or have suitable alternatives. (See for our recommendations for licensee action.)
s 8509240190 850920 PDR 18rE MISC PDR gs\\
h
,a Edward L. Jordan 2
SEP 2 01985 Mr. Peter R. Wohld of our staff will be glad to discuss these coments further and provide other support, as necessary.
" M 3Inti f'c ' 4 7., c. r I'2,-.y t..tr Charles E. Norelius, Director Division of Reactor Projects As s'ated Attachments:
t cc w/ attachments:
F. Cherney, MEB (P-522)
J. Page, MEB (P-522)
R. Kiessel, EGCB (EWW 359)
J. Zudans, EAB (3302 MNBB)
F. Ashe, AE00 (EWS 205)
Richard W. Starostecki, Director, DRP, RI Roger D. Walker, Director, DRP, RII Richard P. Denise, Director, DRSP, RIV Dennis F. Kirsch, Director, DRSP, RV DCS/RSB(RIDS)
RyI RIII RIII RIII FA4/
.*/u G Nor[U iuf,>o y
'a Wohld/kja Guldemond
@pinz i
09/19/85 f-/f-*f 9
4 Common Mode Failure Analysis of Valve Failures at Davis-Besse Nuclear Plant and General Industry Implications Introduction The ability of the nuclear industry to properly handle Limitorque valve operator torque switches and. limit switches is seriously in question. While the problem would seem to be a simple one, motor-operated valve failure experience in the industry indicates otherwise.
An analysis and evaluation i
is given below of control measures that failed to prevent the common mode failure of the Davis-Besse Nuclear Plant auxiliary feedwater valves on June 9, 1985.
(Details of the event are provided in NUREG-1154, " Loss of Main and Auxiliary Feedwater Event at the Davis-Besse Plant on June 9,1985.")
Inspection experience in Region III indicates that many of the control measures that failed at Davis-Besse are generic to other plants as well.
Analysis and Evaluation Following is an analysis and evaluation of specific areas where better performance could have prevented or helped prevent the common mode valve failures.
1 (1) Design Control (10 CFR 50, Appendix B, Criterion III)
(a) Testing has shown that the nominal valve operator torque switch setting specified by the architect engineer was inadequate to j
assure valve opening without the corresponding, proper setting of the torque bypass switch.
However, no guidance was provided for setting the bypass switch.
(b) Out of concern for valve operator torque switch, torque switch l
bypass, and other valve limit switch settings, Toledo Edison commissioned a study by Torrey Pines which also resulted in inadequate settings.
(c) No testing was specified to provide assurance of the adequacy of the torque switch, torque bypass or other limit switch settings under differential pressure conditions.
Conclusion (a) Some motor-operator switch settings were not provided when needed.
(This is typical of other plants in Region III.)
(b) When torque switch and limit switch settings were provided, they were not reliable enough for use without additional measures to a
I assure their adequacy.
i l
11, i
I-j (2) Field Instructions and Procedures (10 CFR 50, Appendix B, Criterion V) i (a) The procedure for setting torque bypass switches, provided by j
Torrey. Pines, was confusing to use in the field and did not j
clearly make allowance for " lost motion" between valve motor j
. start and valve disc motion.
(b) The procedures did not require a mandatory independent verification j.
or Quality Control holdpoint for switch settings or other valve t
work for which, if done improperly, would not be discovered by the l
normal valve testing under passive conditions.
~(c) Procedures were not available to perform meaningful post-maintenance operability testing.
Conclusion-r While procedures were inadequate, this was not the primary cause of i
the failures. However,'the lack of procedures to ensure-adequate
~
post-maintenance operability testing in the past resulted in a failure to reveal the operator. switch setting problems and prevent the recent common mode failure.
(The lack of adequate post-maintenance test procedures is a generic problem currently being addressed by NRC j
1 (3) Maintenance (Appendix B, Introduction) s (a) A scheduled and proceduralized preventive maintenance program i
for motor-operated valves has not been in effect. Maintenance I
was more a reactive. response to otherwise identified degraded conditions or total failure.
(b) While valve failure data was available onsite, there was no
{
effort to ' trend the data and factor it into the maintenance-j program.
1 (c) Post-maintenance testing techniques had not been developed onsite t
j that were adequate to assure valve operability.
i
?
I (d) Field observations indicate that the maintenance technicians were poorly trained.
i
(
(e) The "as found" conditions of the valves indicated a number of I
previous maintenance errors.
However, it was concluded that the valves would have failed against differential pressure with or j
without the errors.
Conclusion i
i While the connection between the final-valve failures and maintenance is j
weak, a maintenance department with qualified technicians and management i
r i
1
support for meaningful testing could have discovered the valve deficiencies much earlier.
(4) Test Control'(10 CFR 50, Appendix B, Criterion XI)
Testing performed to assure valve operability' included preoperational testing, periodic surveillance testing, and post-maintenance testing.
None of these' tests were done at design differential pressure.
- Further, test techniques were not adequate to allow an extrapolation of the test data to account for the difference between actual test and design basis conditions. When the valves experienced differential pressure conditions in service, they failed to open.
Conclusion Testing should be performed to assure adequate operation at differential pressure conditions, either by testing under design conditions or by prototype testing.
Sophisticated techniques that demonstrate valve capability (such &s by using load cell measurements as with M0 VATS, etc.)
should be used to assure continued operability during plant life.
(None of this is currently required.
Most licensees depend on valve stroke timing to ensure valve operability; however, this testing is meaningless as currently applied to motor-operated valves.)
(5) Corrective Action (10 CFR 50, Appendix B, Criterion XVI)
One of the same valves failed to open when needed in 1984.
Corrective action in response to that event failed to adequately address the torque switch bypass problem which existed since plant construction and continued to exist until the June 9, 1985, event.
(Toledo Edison failed to i
duplicate the problem.
Corrective action consisted of lowering the valve closure torque switch setting so that less torque would be required initially on opening. This had no impact on the actual problem which 3
only revealed itself under differential pressure conditions.)
Conclusion Adequate corrective action in response to the 1984 problem would have prevented the common mode failure.
6)
Industry standards are lacking in the areas of motor operated valve surveillance and post-maintenance testing.
(See 7 below.)
7)
NRC Requirements (a) 10 CFR 50.55a(g) and Standard Technical Specifications endorse the ASME,Section XI, Inservice Testing requirements which are substantially deficient in assuring motor-operated valve operability.
Inspection for performance of Code testing at plants in Region III showed that not one valve was ever found to have exceeded the Code specified stroke time criteria prior to total failure to stroke.
3
t (b) Specific requirements for valve preventive maintenance do not exist.
Post-maintenance testing is required by the ASME Code, which NRC endorses, but which is not meaningful as noted above.
(c) Our qualification requirements for maintenance mechanics are extremely weak.
Conclusion While the ASME Code test requirements are beneficial for testing air-operated valves and pumps, they are inappropriate for motor-operated valves and we mislead both licensees and ourselves by endorsing it for this purpose.
Meaningful test requirements need to be developed as well as requirements for preventive maintenance and maintenance mechanic qualification.
(8) Other - It is interesting to reflect that Toledo Edison was already aware of the human engineering weakness on the control panel that contributed to the human error which precipitated the June 9, 1985, event.
(See NUREG-1154, Page 6-2, last paragraph.) While inadvertent actuation is considered in plant design, this covers only one actuation per event.
When the plant operator defeated both trains of auxiliary feedwater, this exceeded the plant design requirement to recover from such a condition; e.g., there was no design requirement for the auxiliary feedwater valves to reopen after being so closed. One of the questions this raises is about the basic plant design philosophy and the wisdom in not providing better operability assurance for equipment readily available for licensed plant operator use or misuse.
Concluding Statement Many of the factors leading to or allowing the Davis-Basse valve failures appear common to the nuclear industry.
This is based on staff experience which includes onsite inspecticn activities at operating plants for pump and valve inservice testing in Region III.
In overall summary, it appears that the industry as a whole has adopted a valve motor-operator for which it has failed to properly specify setup, maintenance, or testing.
Power plant sites each learn independently, by rote, to address obvious deficiencies as they are experienced through valve failures. The situation is unacceptable, requiring immediate, interim corrective measures as well as longer term action.
k YY YM/97 Peter R. Wohld, P.E.
Inspector, Region III 4
Recommendations for Changes to " Actions for All Holders of Operating Licenses or Construction Permits" 1.
Item a: Change the associated note at the bottom of the page to read,
" Recovery from inadvertent valve movement to an undesired position should be considered in the design basis unless the motor-operator is protected from inadvertent operation in the main control room."
2.
Item b: Add the following statement, " Include control of the settings in a design control program to assure the proper review of any changes that may be proposed."
3.
Item c: We do not feel strongly about design differential pressure testing, except perhaps initially on prototypes or typical samples.
It should be fully adequate to subsequently perform instrumented testing, such as with MOVATS, to show proper mechanical operation and capability.
4.
Item d:
Delete the re.ference to Generic Letter 83-28 or tie it in with post-maintenance testing, as this is what the letter is limited to in this case.
5.
Item d:
Change the ftrst sentence to read, " Prepare or revise procedures to ensure that correct switch settings are utilized and that they are adequate for the function intended by performing meaningful valve operability testing throughout the life of the plant."
6.
Item d:
Concurrent with the change in 5, add the following as an acceptable program to meet Item d, "While a graded approach may be appropriate, based on a safety evaluation of valve functional requirements, the following will be considered an acceptable program for any safety related, motor-operated valve:"
d.1 Perform stroke timing, both open and closed (quarterly if possible), without initial "prestroking," by measuring the motor "on time" for a full stroke.
Times should be measured to 10.5 percent accuracy, compared to an acceptable baseline, and further evaluated when a stroke open or closed changes by 12.0 percent.
(The ranges may be increased or reduced based on additional experience and knowledge.)
d.2 Valve motor current traces during valve stroking in both the open and closed directions should be taken at least once every other refueling outage and evaluated for abnormalities.
Concurrent with this cycle, valve stroking should be observed for proper operation and the valve externally inspected for general ~ condition.
e e
d.3 In depth test and evaluation for proper operation (including evaluation of valve packing loads, torque switch operation
'at proper stem thrust loads, and limit switch settings) should be done at least once every other refueling outage.
(Alternate with Item 2 every refueling outage.)
d.4 During the conduct of 3, perform both an internal and external ~ inspection of general valve mechanical condition (electrical connections, oil leaks, bolting, rust, packing leaks,etc.)
d.5 Include any or all of 1 through 4 above, as appropriate, in post-maintenance test instructions.
7.
Additional recommendation:
Specify that all licensees develop adequate training and qualification requirements for maintenance personnel involved in setting up, maintaining, and testing valve motor-operators.
Y
- k. W N Y 9l/9/$ f Peter R. Wohld, P.E.
//
Inspector, Region III 4
f 2
. -.