ML16138A520

Steam Voiding in Oconee 3 on 750613-Precursor Event to TMI-2 Accident, Engineering Evaluation Rept

ML16138A520
Person / Time
Site:	05000000, Oconee
Issue date:	11/28/1983
From:	Ornstein H NRC OFFICE FOR ANALYSIS & EVALUATION OF OPERATIONAL DATA (AEOD)
To:
Shared Package
ML16138A521	List: ML16138A520 ML16299A365
References
TASK-AE, TASK-E326 AEOD-E326, NUDOCS 8312140306
Download: ML16138A520 (13)

Text

AEOD ENGINEERING EVALUATION REPO71*

UNIT:

Oconee 3 EE REPORT NO. AEOD/E326 DOCKET NUMBER: 50-287 DATE:

November 28, 1983 LICENSEE:

Duke Power Corporation EVALUATOR/CONTACT: H. L Ornstein NSSS:

Babcock and Wilcox

SUBJECT:

STEAM VOIDING IN OCONEE 3 ON JUNE 13, 1975 - A PRECURSOR EVENT TO THE TMI-2 ACCIDENT

SUMMARY

On June 13, 1975 Oconee 3 experienced a reactor trip and an overcooling transient.

Post trip analyses focused on the effect that the excessive cooldown rate had upon the fuel, the violation of the plant's technical specifications and the violation of reactor coolant pump NPSH requirements.

The analyses did not focus on the reasons for the plant's anomalous behavior (rising pressurizer level which led operators to secure HPI and reopen a block valve 15 seconds after closing it).

This report shows that primary system voiding took place during the June 13, 1975 transient (none of the reports available have indicated this phenomena had taken place). The report concludes that had an extensive review of this event been made at the time and had operators at other B&W plants been made aware of the details of this event - including primary system voiding and pressurizer level increases accompanying a stuck open PORV, it is possible that subsequent depressurization events at B&W plants (TMI-2 accident in particular) might have taken different courses.

The report clearly illustrates the need for rigorous post mortem analysis of anomalous events and the need for effective dissemination of operating data throughout the nuclear community.

This report supports ongoing AEOD and NRC activities and does not represent the position or requirements of the responsible NRC program office.

8312140306 831128 PDR ADOCK 05000287 S

PDR

0 0

-2 INTRODUCTION In the course of reviewing Duke Power Company's July 23, 1982 (Reference 1) response to I&E Bulletin 80-04 ("Main Steam Line Break With Continued Feedwater Addition"), AEOD became aware of Duke Power Company's summary of significant transients at the Oconee station (Reference 2).Duke Power Company's summary of significant transients included a brief description of the overcooling transient which occurred at Oconee Unit 3 on June 13,

1975, The June 13, 1975 event appeared to be similar to the TMl-? accident.

A check -of the ORNL precursor study for the years 1969-1979 (Reference 3) showed that ORNL considered the June 13, 1975 event to be a precursor to a potential core damage accident.

Even though the event occurred more than eight years ago, the lessons which were not learned from this event give strong support to the NRC's recent require ments for improved post trip analyses and more detailed licensee event reports.

DISCUSSION Several documents were written on the June 13, 1975 event; however, none of them appeared to recognize the fact that there was steam voiding in the primary system.

The following is a chronology, of documents which were written on the event.

a. Duke Power Company - Oconee Nuclear Station Incident Investigation Report Number B-350, circa June 19, 1975 (Reference 4).

This report has the most complete account of the event available. It also in cludes a handwritten account of the plant Station Review Committee's (SRC) recommendations.

b. Abnormal Occurrence Report AO-287/75-7, June 27, 1975 (Reference 5)

(This was a 14-day report - The equivalent of today's LERs)

(see Appendix A).

c. B&W Site Problem Report SPR-107 "Electromatic Relief Valve Mal function," June 27, 1975 (Reference 6)-Transmitted on July 13, 1975 by B&W Technical Support (Wandling) to B&W service representatives assigned to Oconee, TMI, Crystal River, ANO-1, Rancho Seco, and Davis-Besse. B&W service representatives were asked to inform the customers of the problem which resulted from the corrosion of PORV parts, and advise the customers to conduct periodic inspections of the valves.

d. IE Inspection of Oconee Station July 29 - 31, 1975 (Reference 7).

The NRC noted that during the inspection the licensee was informed that the original 14-day report appeared to be in noncompliance with NRC requirements.

The original AO report (Reference 5) "did not include an analysis and evaluation of the safety implications involved in the blowdown of the Unit-3 reactor coolant system nor did the report address the causes and corrective actions taken to prevent recurrence of the incident". During the inspection the licensee agreed to submit an updated AO report. The final inspection report 50-287/75-10 dated

August 27, 1975 indicated that the licensee's supplemental Ab normal Occurrence report, AO-287/75-7 dated August 8, 1975 (Reference 8) was adequate.

e. Abnormal Occurrence Report AO-287/75-7 supplemental report August 8, 1975 (See Appendix B) - The report is an expanded update of the licensee's original report, as noted in item d-above, this report resolved IE's concerns.

f. B&W Site Problem Report SPR - 108, "RCS Rapid-Depressurization Transient," August 20, 1975 (Reference 9).

This report analyzed the rapid cooldown and depressurization. The report indicates that the RCS cooldown rate was 174*F /hour and noted that it exceeded the plant's Technical Specification requirements The report also indicated the 174 0F/hr rate was within the limits of a rapid depressurization transient which was defined in B&W's RCS Functional Specification (240*F/hr) - and was analyzed in the vendor's ASME Section III stress reports. B&W concluded that RCS components did not suffer damage from excessive stress levels during the June 13, 1975 transient.

g. USNRC, "Current Events Power Reactors Events Selected From Reports Submitted To The United States Nuclear Regulatory Commission August September 1975," page 6, article entitled "Excessive Reactor Coolant System Cooldown Rate," October 1975 (Reference 10).

This report presented a summary of the licensee's AO reports (items b and e above).

h. NUREG-0560 "Staff Report on the Generic Assessment of Feedwater Transients in Pressurized Water Reactors Designed by the Babcock and Wilcox Company," May 1979 (Reference 11).

This report briefly described the June 13, 1975 Oconee 3 event and noted that that event and the Davis-Besse September 24, 1977 event "can be said to be precursor-type events."

The report also noted that a feedwater transient plus a PORV stuck open can result in void formation in the RCS. The report noted that voiding did take place during the Davis-Besse events however, no such statement was made about the Oconee event.

i. "Nuclear Accident and Recovery At Three Mile Island" a report prepared by the Subcommittee on Nuclear Regulation for the Committee on Environment and Public Works, U.S. Senate, June 1980 Serial No. 96-14 (Reference 12).

The report noted that the June 13, 1975 Oconee event was "quite similar to the early stages of that at TMI-2." However the report noted that "The operators diagnosed the leak and closed the block valve before the water in the primary system boiled." No mention was made of the operators reopening of the block valve 15 seconds later because the pressurizer level was still rising.

-4

j. Duke Power Letter (Parker) to NRC (Denton) dated July 23, 1980 (Reference 2) provided comments on NUREG 0667 ("Transient Response of Babcock and Wilcox Designed Reactors").

Reference 2 contained an evaluation of the Oconee station's operating experience, includ ing a brief discussion of the June 13, 1975 event.

k. NUREG/CR-2497, "Precursors To Potential Severe Core Damage Accidents 1969-1979 A Status Report," June 1982, (Reference 3).

This report presented event trees for 169 precursor events. The June 13, 1975 Oconee 3 event was analyzed using probabilistic analysis. This event sequence was quantified by ORNL as having the 82nd highest probability of leading to core damage (2.5 x 10

),

predicated on a 10% chance that the operators would not close the PORV block valve.

FINDINGS AEOD reviewed and analyzed the reports cited in the discussion section above.

AEO0's analysis revealed the following facts which appear not to have been recognized by previous reviewers.

1. During the transient there was voiding in the reactor coolant system. As shown in Figure 1, during the transient the pressure in the reactor vessel corresponded to saturation, e.g.,

State point Ptap Tcold leg Tsat (PSIG)

(OF)

(OF)

A 1000 540 546 B

900 534 534 C

800 522 520.4 D

715 502 508 It should be noted that reports which discussed the transient made mention of state point D, which corresponded to a slightly subcooled condition - whereas earlier during the transient, saturation conditions prevailed in the primary system (state points B and C).

The temperatures listed above refer to the cold leg temperature; however, at higher elevations - hot legs, candy canes, etc. - where the pressures are lower and the temperature are higher the conditions were more likely to induce voiding there than at the lower elevations.

Furthermore, if one con siders the large primary system metal mass and the high initial metal temperatures it is quite likely that voiding took place in the upper head region.

It was noted in the licensee's incident report (Reference 4) that the operators reopened the block valve 15 seconds after they had closed it.

The reason given was that pressurizer level was still continuing to rise. As shown in Figure 2 which is a plot of reactor pressure and pressurizer level with time, the pressurizer level continued to rise after the block valve was closed. However, it should be noted that

-5 the pressurizer level changes lagged reactor coolant pressure by about 20 seconds. Closing the block valve had caused the rate of pressurizer level increase to slow down; however, the operators did not wait long enough to detect such change. About 20 seconds after the block valve was reopened the rate of the pressurizer level rise had increased again as noted in Figure 2 (there was about a 20 second time delay).

Furthermore, the rate of pressurizer level increase was less than it had been prior to the initial closing of the block valve. These changes in pressurizer levels are indicative of the growth and col1apse of the voids in the primary system.

2. At about the time that operators closed and reopened the block valve, they also secured high pressure injection. This conclusion is based upon the sequence of events and the licensee's calculation of the amount of primary system coolant that was released to the sump as described in the aforementioned incident report, (Reference 4).

Reference 4 states that HPI was on for 16 minutes and 12 seconds

@300 gpm, and it also was on for 10 seconds @100 gpm. Since HPI went on at 1443:20, then one can conclude that the HPI was shut off at 1443:20 + 16:12 + 0:10 or about 1500: -

which corresponds to the time when pressurizer level was recorded to be at 367 inches (this was 6 1/2 minutes before the block valve was finally closed to end the transient).

None of the aforementioned reports written on this event explicitly acknowledge the fact that all HPI had been turned off prematurely.

This was a crucial act that was repeated at TMI-2 for apparently the same reason - rising pressurizer level.

It should be noted that the licensee's conclusion from the event (Reference 4) was that the correct operator action was to close the block valve. The licensee did not have any statement about securing High Pressure Injection.

3. Regarding cooldown rates, the licensee's original submittal to the NRC (Reference 5) contained incorrect information. Figure 3 shows the licensee's stripchart of reactor coolant system temperature.

Reference 5 stated that "The shutdown was continued with a cooldown rate of 100aF/hr as specified in Technical Specification 3.1.2.3; however, when the initial drop in temperature due to depressurization was combined with the subsequent cooldown the cooldown rate for the first hour was 1010F."

Examination of Figure 3 shows that for the first hour of the transient the average cooldown rate was 135 0 F/hr.

TStart = 585 0F 1350F/hr TStart = 450OF

+ 1 hr The licensee's update to the AO report (Reference 8) stated that "The subsequent controlled cooldown of the Reactor Coolant System, when com bined with the temperature drop during the transient, resulted in a cooldown of 101*F during the first hour when temperature was below 5300 F, contrary to the provisions of Technical Specification 3.1.2.3."

This statement is based upon the fact that the technical specification cooldown limit of 100OF/hr is for the RCS temperature range of 275 0F to 5300F.

The licensee neglected to mention the reactor vendor's August 1, 1975 letter (Reference 13).

Reference 13 informed the licensee of B&W's analysis which concluded that during the event, the cooldown rate was 174*F/hour, which was within the limits of B&W's ASME Code Stress Analyses.

TStart = 585 0F TStart + 28 minutes = 504*F 1740F/hr

4. The licensee's and the reactor vendor's post transient reviews were not complete. They did not analyze the event to the extent necessary to determine the cause of the anomalous primary system behavior. The post transient reviews focused on the transient initiator, the PORV mal function, the malfunction of the PORV position indicator, the possible violation of reactor coolant pump NPSH limits (the RCP's apparently were operating throughout the transient), violations of the fuel compression curve, reactor coolant system integrity, and cleanup of the spilled primary system coolant.

The licensee's Station Review Committee (SRC) recommended that control room operating personnel should be made aware of the June 13, 1975 event, emphasizing that immediate closure of the block valves is the proper corrective action. In addition they addressed the reasons for the load oscillations which initiated the transient, and made recommendations regarding cooldown rates during and subsequent to transients.

The licensee and reactor vendor did not appear to consider the reasons for the anomalous primary system behavior:

rising pressurizer level with the decrease in reactor coolant system pressure.

rising pressurizer level subsequent to isolation of the primary system leak.

the existence of saturation conditions and voiding within the primary system.

reactor coolant pump performance during adverse conditions (violation of NPSH operating limit, cavitation, voiding).

CONCLUSIONS It appears that had a more complete analysis of the June 13, 1975 event been made and had information about the event been transmitted to operators at other plants, subsequent events might have had significantly different results.

In particular information about:

a.

primary system voiding

b.

pressurizer level increases accompanying a stuck open PORV

c. failure of the PORV status indicator

d. premature securing of HPI due to increasing pressurizer level

e. reopening of the block valve due to increasing pressurizer level

f.

reactor coolant pump operation with inadequate NPSH, cavitation and voiding,

-7 could have had a strong bearing upon the outcome of subsequent depressurization events at other B&W plants.

If the operators at Davis-Besse and TMI-2 had been sensitized to the June 13, 1975 Oconee 3 event with the explanations for the anomalous plant behavior, they would have been better prepared to handle the stuck open PORV events that occurred at their facilities, especially when accompanied by incorrect status lights.

They also would have better understood primary system voiding and its effect upon pressurizer level.

The June 13, 1975 transient at Oconee 3, and the subsequent depressuri zation events that occurred at other B&W plants (Davis-Besse and TMI-2) emphasize the need for rigorous post mortem analysis of anomalous events, and the need for effective dissemination of operating data.

It should be noted that the NRC has taken many steps in this direction.

Subsequent to the TMI-2 accident, the Commission concluded that reactor safety would be enhanced by improved evaluation and dissemination of operational data.

As noted in the TMI Action Plan (NUREG-0660), numerous actions were taken in this regard, including the establishment of the Office for Analysis and Evaluation of Operational Data (AEOD). In recent months the NRC has taken additional actions which will improve the analysis and dissemination of operational data, i.e., NRC Generic Letter 83-28 ("Required Actions Based on Generic Implications of Salem ATWS Events," dated July 8, 1983) requires each licensee to implement a program to assure that the causes for unscheduled shutdowns, and the responses of safety-related equipment are fully understood prior to plant restart.

Further more, the new LER rule (10 CFR 50.73, effective January 1, 1984) modifies the existing Licensee Event Reporting System to require licensees to provide detailed information which will improve the capability of the NRC and industry to perform engineering studies of operational anomalies, and to do trends and patterns analyses of operational occurrences.

- 8 REFERENCES

1. Letter from W. Parker, Duke Power Company, to H. Denton,

NRC, Oconee Nuclear Station Docket Numbers 50-269, 270, 287, July 23, 1982.

2.

Letter from -W. Parker, Duke Power Company, to H. Denton, NRC, Oconee Nuclear Station Docket Numbers 50-269, 270, 187, July 23, 1980.

3. USNRC, NUREG/CR 2497, "Precursors to Potential Severe Core Damage Accidents 1969-1979 A Status Report," June 1982, (written by ORNL/SAI).

4. Duke Power Company - Oconee Nuclear Station Incident Investigation Report Number B-350, Circa June 19, 1975.

5. Letter from W. Parker, Duke Power Company, to N. Moseley, NRC, transmitting Abnormal Occurrence Report No. AO-287/75-7, June 27, 1975.

6. Babcock and Wilcox Site Problem Report SPR-107 "Electromatic Relief Valve Malfunction, June 27, 1975, transmitted by Babcock and Wilcox Technical Support (K. Wandling) to B&W Service Representatives at Oconee, TMI, Crystal River, ANO-1, Rancho Seco and Davis-Besse, on July 13, 1975.

7. Letter from N. Moseley, NRC, to W. Parker, Duke Power Company, August 27, 1975. Transmitting IE Inspection Reports 50-269/75-9, 50-270/75-10, and 50-287/75-10.

8. Letter from W. Parker, Duke Power Company, to N. Mosely, NRC, transmitting an update of Abnormal Occurrence Report No. AO-287/75-7, August 8, 1975.

9. Babcock and Wilcox Site Problem Report SPR-108, "RCS Rapid Depressurization Transient," August 20, 1975.

10.

USNRC, "Current Events Power Reactors, Events Selected from Reports Submitted To The United States Nuclear Regulatory Commission August - September 1975,"

page 6 article entitled "Excessive Reactor Coolant System Cooldown Rate,"

October 1975.

11.

USNRC, NUREG 0560 "Staff Report On The Generic Assessment Of Feedwater Transients In Pressurized Water Reactors Designed By The Babcock and Wilcox Company," May 1979.

12. "Nuclear Accident And Recovery At Three Mile Island" - a report prepared by the Subcommittee on Nuclear Regulation for the Committee on Environment and Public Works, U.S. Senate, - Serial No. 96-14, June 1980.

13.

Letter from S. Hellman, Babcock and Wilcox, to K. Canady, Duke Power Company, "Oconee Nuclear Station Unit III Cooldown Transient Master Service Task

1. 47" August 1, 1975.

09 FIGURES

0 0

1i000 PA 900 TSA+ TO B

700 -

D 450 500 5501 TEMPERATURE, OFf Figure 1 Reactor Coolant System Pressure vs Reactor Coolant System Temperature

RCS PRESSURE

~w0 CW 0

00 00 00 0

a

<)

0.<

0 CD (c

a, CO (D

CD I

CD n rI N

C PRESSURIZER LEVEL!

Trip 600 ni 101.F/hr I c:

500 F

400 350 1foo m

3001 2PM 3PM 4PM 5P TIME Figure 3 Reactor Coolant System Temperature vs Time

0 0

Appendix A Abnormal Occurrence Report AO-287/75-7 June 27, 1975