ML20196G525

Technical Review Rept T809, Blocked Thimble Tubes/Stuck Incore Detector

ML20196G525
Person / Time
Site:	Beaver Valley, Salem, Kewaunee, Sequoyah, Braidwood, Surry, North Anna, Ginna, 05000000, 05000465, Trojan
Issue date:	06/15/1988
From:	NRC OFFICE FOR ANALYSIS & EVALUATION OF OPERATIONAL DATA (AEOD)
To:
Shared Package
ML20196G410	List: ML20196G405 ML20196G463 ML20196G525
References
TASK-AE, TASK-T809 AEOD-T809, T809, NUDOCS 8807050550
Download: ML20196G525 (4)

Text

.

9 AEOD TECHNICAL REVIEW REPORT UNIT: Multiple TR REPORT ~N0. AE00/T809 DOCKET N0: Multiple DATE: June 15, 1988 LICENSEE: Multiple EVALUATOR / CONTACT: M. Wegner NSSS/AE: Westinghouse / Multiple

SUBJECT:

BLOCKED THIMBLE TUBES / STUCK INCORE DETFCTOR EVENT DATE: Multiple

SUMMARY

Thimble tubes through which the incore detector probes are inserted into the reactor in a Westinghouse pressurized water reactor (PWR) may become blocked by hardening of the detector cable lubricant, denting of the thimble tube caused by flow-induced vibration, or by flooding of the tube caused by a leaking thimble tube. Detectors may become stuck when the thimble tube is blocked or the withdrawal mechanism is faulty. Recovery of a stuck detector and cleaning of thimble tubes generally proceed without incident, but several events point out the potential for personnel hazard. Salem I had a thimble tube begin to leak, Sequoyah I had a thimble tube ejection, and Surry 2 had significant exposure to personnel. Causes for the problems include damage to the thimble tubes caused by flow-induced vibration aggravated by damage caused by inserting or withdrawing the detector and a combination of planning and maintenance errors.

DISCUSSION Salem 1 LER 50/272-81/028 March 11, 1981 While attempting to do a flux map during power operation, the C detector would not insert into the thimble tube. At the same time, a leak alarm on the incore detector operating console annunciated in the control room. Tube H11 had begun to leak and reactor co'olant had flooded all 6 ten-path transfer devices parti-ally or completely filling all thimble tubes. The valve on thimble tube H11 was closed isolating the leak. The leak may have been due to damage to the thimble tube caused by flow-induced vibration aggravated by detector insertion or withdrawal.

Sequoyah 1 LER 50/327-84/030 April 19, 1984 There are 58 thimble tubes in each of the Sequoyah plants (typical of Westing-house-designed PKPs) which penetrate the reactor pressure vessei bottom head and through which detector are inserted into the core region. The incore detectors measure neutron density at different locations in the core (flux mapping) for the purpose of calibrating the excore neutron measuring devices and confirming design parameters. The thimble tubes are inserted into guide tubes and the detector or probe is inserted into the thimble tube. The 1.d. of the thimble tube is 0.201 inch and the o.d. of the probe is 0.188 inch. The guide tube is reduced from 1.0 inch o.d. to 0.625 inch o.d. at the seal table l

MR7 na S$he P

l

\

and is welded to the seal table. The thimble tube passes through the guide tube at the seal table and is connected to it by a union, ferrule, and nuts, forming a reactor coolant pressure boundary. The lubricant used on the detec-tor cable, if used in excess, mixes with the corrosion products from the system hardening, lumping, and blocking the thimble tube.

Sequoyah I had problems with thimble tube blockage which remained unsolved from prestartup days. During the 1934 outage, attempts were made to clean 9 of the tubes with limited success. A total of 23 tubes were identified as being blocked. A limited scope survey provided determination that there was prece-dent for cleaning them at power. Startup proceeded while the cleaning was being done.

With the reactor at 545'F and 2250 psig in mode 1, the workers were assembled around and above the seal table performing'their assigned tasks. The cleaning tool was being inserted into a thimble tube when it became increasingly diffi-cult to turn the crank to insert the tool. Some movement of the tool was observed at the same time that a, leak occurred. The cleaning tool pulled loose from the worker's grasp. He pushed it aside as the water began to gush and exited the area. All eight men left the seal table area without being injured or contaminated. The thimble tube was later detennined to have been ejected from the core and twisted throughout the room. An unisolable RCS leak of about 30 gpm initially was caused by the ejection. Total coolant lost was estimated to be 1600 gallons.

Unusual stresses applied to the mechanical seals by the cleaning fixture were thought to have damaged them, causing the failure. Much has been written about the lack of scope in the survey which determined that precedent existed for cleaning the tubes at power and the inadequacy of the planning and proce-dures for accomplishing the task. The licensee credits the heightened anxiety of the workmen along with luck for the fact that no one was injured or killed by the potentially lethal eruption of steam and ejection of the tube.

Foreign A August 1985 Flux probe became stuck in thimble during flux mapping. Problem believed to be denting of the thimble caused by flow-induced vibration.

Foreign B September 1985 Flux probe became stuck in a thimble tube during flux mapping due to denting of the thimble tube. The thimble leaked after recovery of the probe while cleaning the tube.

Foreignj February 1986 Flux probe became stuck in thimble during flux mapping.

Surry 2 Inspection Report 50/281-88/010 March 3, 1988 With the unit at 100% power and the containment sub-atmospheric, technicians attempted to free the A incore detector which had been lodged in the core for 24 days. With an HP (health physics) technician monitoring their efforts, two 18C (instrumentation and controls) technicians began pulling the detector cable to withdraw the detector. The polar crane wall acted as a radiation shield.

Ccmunications with the control room was established in order to determine when the detector reached the "withdrawn" position, but difficulties in comuni-cating made it necessary for the warning to be given twice. At the same time, as the detector was pulled up to the polar crane wall, the radiation levels inc* eased rapidly and the HP technician ordered work to stop and the area to be evacuated. While doses greatly exceeded the expected, they did not exceed 10 CFR 20 limits. The whole body doses were calculated to have been between 800 and 1600 mrem.

Pemote control from the control room failed to recognize the hazards of fail-ures in communications and job planning failed to allow for the possibility of exposing the highly activated detector and cable.

Foreign D March 1988 While eddy current testing the thimble tubes, one tube would not pass the probe because of blockage.

Sequoyah 2 Daily Report May 16, 1988 Incore flux detector A would not retract after being inserted for flux mapping.

Af ter replacing a fuse in the drive motor circuitry, the detector was retracted and was placed in the storage position.

_Braidwood 1 Daily Report May 24, 1988, et seq.

Since May 9,1988, the unit had experienced sticking incore detectors in 15 of 58 thimble tubes. Atterrpts to clear the blockage by running detectors through them have resulted in the replacement of two detectors. The licensee has cleared the tubes while in mode 3 by forcing the blockage using a dummy cable with a brush assembly attached driven by a mechanical drive inserted at the exit of the five-path transfer device. The NPC communicated to the licensee the concern that the potential for leaking tubes or tube ejection should be considered in the planning and procedures for cleaning in order to avoid these personnel hazards.

Other U.S. Facilities Other facilities identified as having experienced blocked thimble tubes and having cleaned them without incident are Kewaunee, Trojan, Salem, Beaver Valley 1, North Anna, and Ginna. In most cases, the cleaning was done with the reactor coolant system (RCS) depressurized and cooled.

1

o FINDINGS

1. Thimble tubes can become blocked for several reasons: plugging by foreign matter, denting, or flooding.

2. Detectors may become stuck because of blockage of a thimble tube or because the mechanism for withdrawal is faulty.

3. Unblocking of thimble tubes and recovery of detectors normally proceed without incident but the potential for life-threatening hazards to maintenance personnel and unisolable leaks from RCS exist.

CONCLUSIONS

1. Blockage of thimble tubes can be caused by denting from flow-induced vibration and may result in a small break loss-of-coolant accident from the reactor pressure vessel bottom head. Wear of tubes caused by flow-induced vibration exists in all Westinghouse PWRs. Any attempt to free a stuck detec-tor or to clear a tube blockage should consider these possibilities.

2. Cleaning of thimble tubes can be made unnecessarily complicated and hazar-dous by inadequate planning and a failure to recognize and to compensate for safety hazards. Any licensee attempting to free a stuck detector or to clear a tube blockage should become familiar with the incidents at Sequoyah in 1984 and Surry in 1988, in particular, and incorporate lessons learned into planning preparations, i s

__ _ . . . , , _ . _ _ ~ ... ..