Loading...
HomeMy WebLinkAbout851166.tiff rJ``Epn"E0Ny� UNITED STATES , 9� NUCLEAR REGULATORY COMMISSION 1.1 { i m REGION IV oY 1' 611 RYAN PLAZA DRIVE, SUITE 1000 •0 ARLINGTON.TEXAS 78011 JUL 0 9 1985 �� In Reply Refer To: � ' ar Docket: 50-267 ��)� �t_ic“ 4114 121g95 1� GREELY. Coco. Public Service Company of Colorado ATTN: 0. R. Lee, Vice President Electric Production P. 0. Box 840 Denver, Colorado 80201 Dear Mr. Lee: We have reviewed your June 17, 1985 (P-85212) , response to our March 11, 1985, request for resolution of our concerns over Beta radiation in liquid effluents. We understand that you are committed to: 1. Perform releases from the Reactor Building sump by a batch method, whereby the sump contents will be sampled and analyzed prior to commencing a release and at 24-hour intervals during the release (we recognized that some addition to the sump may occur during the release period); and 2. Continue the investigation into installing in-line, Beta-sensitive, effluent monitors. We find these commitments to be an acceptable, interim resolution of our concerns and will include them in the listing of commitments that will be confirmed in connection with our authorization of plant restart. If you have any questions on this subject, please contact the NRC Project Manager. Sincerely, . H. Johnson, Chief Reactor Projects Branch 1 cc: (cont. on next page) 851166 a. -,l , IC'. Public Service Company of Colorado -2- Mr. D. W. Warembourg, Manager Nuclear Engineering Division Public Service Company of Colorado P. O. Box 840 Denver, Colorado 80201 Mr. David Alberstein, 14/159A GA Technologies, Inc. P. 0. Box 85608 San Diego, California 92138 Kelley, Stansfield & O'Donnell Public Service Company Building 550 15th Street, Room 900 Denver, Colorado 80202 Chairman, Board of County Comm. of Weld County, Colorado Greeley, Colorado 80631 Regional Representative Radiation Programs Environmental Protection Agency 1860 Lincoln Street Denver, Colorado 80203 Mr. H. L. Brey, Manager Nuclear Licensing/Fuels Div. Public Service Company of Colorado P. O. Box 840 Denver, Colorado 80201 J. W. Gahm, Manager, Nuclear Production Division Fort St. Vrain Nuclear Station 16805 WCR 193 Platteville, Colorado 80651 L. Singleton, Manager, Quality Assurance Division (same address) Colorado Radiation Control Program Director °`EPa net°<qr UNITED STATES a NUCLEAR REGULATORY COMMISSION co S N 3 REGION IV 9» S-I• U 611 RYAN PLAZA DRIVE, SUITE 1000 ARLINGTON. TEXAS 76011 JUL C 3 1985 !`" 1 of ,,, P ,t„,,�., rPt Docket: 50-267 i' -l .Li t� �• r � a ‘11936 II Mr. 0. R. Lee, Vice President GREELEY. COLd, Electric Production Public Service Company of Colorado P. 0. Box 840 Denver, Colorado 80201 Dear Mr. Lee: Our September 22, 1983, letter provided the Preliminary Safety Evaluation Related to the Post Accident Sampling System (NUREG-0737 , Item II.B.3) for the Fort St. Vrain Station (FSV) . Your October 28, 1983 (P-83352) , and July 2 and 16, 1984 (P-84192 and P-84216) letters provided additional information to resolve the open items in our evaluation. We have reviewed your submittals and have determined that additional information is still required for us to resolve this issue. The results of our review are contained in the enclosed Supplemental Safety Evaluation (SSE) . We have found that you now meet eight of the nine pertinent criteria contained in Item II.B.3 of NUREG-0737 and that some revision is required in the procedures used for estimating the extent of core damage in order for us to find the procedure to be acceptable. Therefore, we request that you review the attached SSE and provide your comments on resolving our concern within 60 days of the date of this letter. If you have any questions on this matter, contact the NRC Project Manager - P. Wagner - at (817) 860-8127. Since this reporting requirement relates solely to FSV, OMB clearance is not required under P.L.96-511 . Sincerely, \\l Eric Johnson , Chief Reactor Project Section 1 Enclosure: SSE on I1.6.3 cc: (see next page) i2 _ I _ +,. n I.4/S'C Mr. 0. R. Lee, Vice President -2- Mr. D. W. Warembourg, Manager Nuclear Engineering Division Public Service Company of Colorado P. 0. Box 840 Denver, Colorado 80201 Mr. David Alberstein, 14/159A GA Technologies, Inc. P. 0. Box 85608 San Diego, California 92138 Kelley, Stansfield & O' Donnell Public Service Company Building 550 15th Street, Room 900 Denver, Colorado 80202 Chairman, Board of County Comm. of Weld County, Colorado Greeley, Colorado 80631 Regional Representative Radiation Programs Environmental Protection Agency 1860 Lincoln Street Denver, Colorado 80203 Mr. H. L. Brey, Manager Nuclear Licensing/Fuels Div. Public Service Company of Colorado P. 0. Box 840 Denver, Colorado 80201 J. W. Gahm, Manager, Nuclear Production Division Fort St. Vrain Nuclear Station 16805 WCR 191 Platteville, Colorado 80651 L. Singleton, Manager, Quality Assurance Division (same address) Colorado Radiation Control Program Director Supplemental Safety Evaluation by the Office of Nuclear Reactor Regulation Related to Operation of Fort St. Vrain Nuclear Generating Station Public Service of Colorado Docket No. 50-267 Post-Accident Sampling System (NUREG-0737,II.B.3) I . Introduction In our safety evaluation, we concluded that the licensee's proposed Post- Accident Sampling System (PASS) met six of the nine criteria in Item II .6.3 of NUREG-0737 which are relevant for a gas-cooled reactor. The three criteria which were not fully resolved were: Criterion (2) Provide a plant-specific core damage estimate procedure to include radionuclide concentrations and other physical parameters as indicators of core damage. Criteria (9) Provide information on the procedure for taking samples of and (10) highly radioactive coolant for gamma spectrometry in such a manner that the activity of the sample does not exceed the measurement capability of the spectrometer. II. Evaluation By letters dated October 28, 1983 and July 2, 1984, the licensee provided additional information. Criterion (2) : The licensee shall establish an onsite radiological and chemical analysis capability to provide, within the three-hour time frame established above, quantification of the following: -2- a) certain radionuclides in the reactor coolant and containment atmosphere that may be indicators of the degree of core damage (e. g. , noble gases, iodines and cesiums, and nonvolatile isotopes) ; b) hydrogen levels in the containment atmospheres; c) dissolved gases (e. g. , H2) , chloride (time allotted for analysis subject to discussion below), and boron concentration of liquids; d) alternatively, have in-line monitoring capabilities to perform all or part of the above analyses. The PASS provides in-line monitoring for noble gas activity, CO and moisture in the helium coolant, as well as for radioactivity in the reactor building stack gas. The PASS also provides the capability to collect grab samples of the coolant and of the reactor building atmosphere that can be transported to the radio-chemical laboratory for CO, CO2, H2, CH4, N2 and radionuclide analyses. These species are indicators of core damage in a gas-cooled reactor, and their relative magnitudes indicate core temperature, fuel particle failure, air ingress or water ingress. We find that the licensee partially meets Criterion (2) by establishing an onsite radiological and chemical analysis capability. However, the licensee should provide a procedure, consistent with the clarification of NUREG-0737, Item II.6. 3, Post-Accident Sampling System, transmitted to the licensee on July 9, 1982, to estimate the extent of core damage based on radionuclide concentrations and taking into consideration other physical parameters such as the concentrations of other gases and core temperature data. Guidance for the procedure to estimate core damage for water-cooled reactors was provided. The procedure for estimating core damage should be consistent with those portions of these recommendations which are applicable to a gas-cooled reactor. -3- The procedure for estimating core damage presented in the letter of July 2, 1984, is not acceptable because it is based solely on the Xe133 concen- tration in the coolant. An acceptable procedure should include consideration of (1) the concentrations of other volatile radionuclides such as xenons, kryptons and iodines, (2) the concentration of other gaseous species, such as H20, CO, CO2, H2, CH4 and N2, and (3) core temperature. The procedure should indicate how these additional considerations would (1) confirm the core damage estimate based on Xe133 (2) provide an estimate of core damage due to water or air ingress, and (3) provide an estimate of core temperature. Criterion (9): The licensee' s radiological and chemical sample analysis capability shall include provisions to: a) Identify and quantify the isotopes of the nuclear categories discussed above to levels corresponding to the source terms given in Regulatory Guide 1 . 3 or 1 .4 and 1.7. Where necessary and practicable, the ability to dilute samples to provide capability for measurement and reduction of personnel exposure should be provided. Sensitivity of onsite liquid sample analysis capability should be such as to permit measurement of nuclide concentrations in the range from approximately 1p Ci/g to 10 Ci/g. b) Restrict background levels of radiation in the radiological and chemical analysis facility from sources such that the sample analysis will provide results with an acceptably small error (approximately a factor of 2). This can be accomplished through the use of sufficient shielding around samples and out- side sources, and by the use of a ventilation system design which will control the presence of airborne radioactivity. -4- The radionuclides in both the helium coolant and the reactor building atmosphere samples will be identified and quantified using the onsite gamma spectrometer. By letter dated July 2, 1984, the licensee provided information on the procedure to take small low pressure samples of the highly radioactive coolant during the period of maximum activity between approximately 5 hours and 7 days after the onset of a loss-of-cooling accident. By controlling the sample size, the measurement capability of the gamma spectrometer will not be exceeded. Radiation background levels will be restricted by shielding. Ventilated radiological and chemical analysis facilities are provided to obtain results within an acceptably small error (approximately a factor of 2). We find that these provisions meet Criterion (9) and are, therefore, acceptable. Criterion (10): Accuracy, range, and sensitivity shall be adequate to provide pertinent data to the operator in order to describe the radiological and chemical status of the reactor coolant systems. The accuracy, range, and sensitivity of the PASS instruments and analytical procedures are consistent with the recommendations and the clarifications of NUREG-0737, Item II.6. 3, Post-Accident Sampling Capability, transmitted to the licensee on June 30, 1982. Therefore, they are adequate for describing the radiological and chemical status of the reactor. The analytical methods and instrumentation are capable of operation in the post-accident sampling environment. No additional training of chemistry personnel is required because the same systems are used for normal and post-accident sampling and analysis. The letter of July 2, 1984, describes provisions to limit sample size, enabling the onsite measurement of radionuclide concentrations in the helium coolant in the post-accident period of maximum coolant radioactivity. We find that these provisions meet Criterion (10) and are, therefore, acceptable. -5- Conclusion We conclude that the post-accident sampling system partially meets the criteria of Item II.6.3 of NUREG-0737. Two of the eleven criteria are not applicable to a gas-cooled reactor. The licensee' s proposed methods to meet eight of the remaining nine criteria are acceptable. The criterion which has not been fully resolved is: Criterion (2): Provide a core damage estimate procedure to include consideration of coolant concentrations of volatile radionuclides and gaseous chemical species together with other physical parameters as indicators or core damage. J``tio REGp49r UNITED STATES 9 F NUCLEAR REGULATORY COMMISSION a �" . . 3 REGION IV m0 � S 611 RYAN PLAZA DRIVE, SUITE 1000 r�o t ARLINGTON, TEXAS 76011 4*tXY JUL 0 7985 WELD gm--vn In Reply Refer To: EI �,1 C Docket 50-267 `L-"-- ,-,.J'1 , 5i 1 lin, 'U CREELEY Public Service Company of Colorado ATTN: 0. R. Lee, Vice President Electric Production P. 0. Box 840 Denver, Colorado 80201 Dear Mr. Lee: We have completed our review of the various submittals related to the Fort St. Vrain Prestressed Concrete Reactor Vessel (PCRV) prestressing tendon wire corrosion problems. The results of our review are contained in the attached Safety Evaluation (SE). We have concluded that sufficient assurance of Safe operation of the PCRV exists to allow the resumption of plant operations, provided the following additional items are implemented: 1. Incorporate the modified tendon surveillance program into the Technical Specifications. 2. Provide the NRC with the results of the tendon surveillance program every 6 months, including the results of inspections related to determining whether there is any evidence of anchorage stress washer failure. We have reviewed your May 20, 1985 (P-85176) , and June 14, 1985 (P-85199) letters and find that your commitments acceptably resolve the above concerns. We will include the above in the listing of the various PSC commitments which will be confirmed in connection with authorization of plant restart. We have also enclosed a copy of the evaluation report, prepared by our consultants at the Los Alamos National Laboratory, for your information and comment. 12 r n..-1-r: -r i I cG - Public service Company of Colorado -2- If you have any questions on this subject, please contact the NRC Project Manager - P. Wagner - at (817) 860-8127. Sincerely, A. E. H. Johnson, Chief Reactor Project Branch 1 Enclosures: 1. Safety Evaluation 2. LANL Evaluation cc: Mr. D. W. Warembourg, Manager Nuclear Engineering Division Public Service Company of Colorado P. 0. Box 840 Denver, Colorado 80201 Mr. David Alberstein, 14/159A GA Technologies, Inc. P. 0. Box 85608 San Diego, California 92138 Kelley, Stansfield & O'Donnell Public Service Company Building 550 15th Street, Room 900 Denver, Colorado 80202 Chairman, Board of County Comm. of Weld County, Colorado Greeley, Colorado 80631 Regional Representative Radiation Programs Environmental Protection Agency 1860 Lincoln Street Denver, Colorado 80203 Mr. H. L. Brey, Manager Nuclear Licensing/Fuels Div. Public Service Company of Colorado P. 0. Box 840 Denver, Colorado 80201 Public Service Company of Colorado -3- J. W. Gahm, Manager, Nuclear Production Division Fort St. Vrain Nuclear Station 16805 WCR 193 Platteville, Colorado 80651 L. Singleton, Manager, Quality Assurance Division (same address) 2J`EpF REDO UNITED STATES Z' �1?, NUCLEAR REGULATORY COMMISSION Cl oyy 3 REGION IV ll I y 611 RYAN PLAZA DRIVE, SUITE 1000 ARLINGTON, TEXAS 76011 SAFETY EVALUATION BY THE OFFICE OF NUCLEAR REACTOR REGULATION PUBLIC SERVICE COMPANY OF COLORADO FORT ST. VRAIN DOCKET NO. 50-267 PCRV TENDON CORROSION INVESTIGATION AND PROPOSED REMEDY PCRV Tendon Degradation - Background: In 1984 during the scheduled PCRV tendon surveillance, the licensee discovered that certain PCRV tendons had broken and corroded wires. In order to determine the extent of this problem, the licensee increased the number of visual examinations of accessible tendons ends. He also performed a number of lift-off tests. Metallurgical examinations of the tendon wires and tests on the protective grease that were performed by the licensee indicate that the corrosion is the result of microbiological attack on the original tendon grease. The licensee has proposed to halt this degradation by filling the tendon sheaths with an inert nitrogen "blanket". As a mechanism for monitoring the condition of the tendons, the licensee has proposed a surveillance program that increases the frequency of the visual inspection and lift-off tests. The surveillance program would compare an uncorroded tendon control group with a corroded tendons group to establish the effectiveness of the corrosion arresting method and the trend in the tendon wire degradation. The program would include samples of the longitudinal , circumferential , and crosshead tendons proportional to the population of the tendon types. Evaluation: 1. Monitoring: The staff evaluated the licensee' s proposed program for monitoring the PCRV tendons and finds the proposal acceptable for assuring PCRV integrity in the near term with certain modifications , as discussed below: The surveillance program, as proposed, would produce a sample of significant size to indicate the trend of the tendon wire degradation and the effectiveness of the corrosion arresting method. However, the information gathered by the licensee from the past and future surveillance activities should be integrated into a complete visual presentation covering all tendons. The purpose of this presentation format would be to provide better information of the extent and significance of the tendon degradation problem. The licensee has committed to incorporate the modified tendon surveillance requirements into the technical specifications. -2- 2. Corrosion Control : The staff evaluated the integrity of the PCRV with the degraded tendons in a safety evaluation dated May 16, 1984. The staff findings were that the reactor vessel was capable of withstanding the operating pressures with the degraded tendons as determined at that time. Since May 1984, a few additional wires have broken but the reactor vessel remains able to adequately withstand the operating pressure. The licensee plans to use a nitrogen "blanket" in the tendon sheaths to halt the degradation of tendons; however, our earlier evaluation indicated concern with this approach. Accordingly, we recommend that the licensee carefully evaluate the effectiveness of other techniques, in terms of their ability to remove oxygen and moisture, and their long term effects on tendon corrosion. 3. Corrosion Problems at Other Plants: The corrosion problem at Fort St. Vrain (FSV) appears to be different from the tendon problems recently experienced at some other nuclear plants. In the other plants, tendons are used in the containment structure which experiences ambient temperatures and the tendon sheaths are filled with grease. The tendons at Fort St. Vrain are located in the reinforced concrete reactor vessel . These FSV tendons experience higher temperatures than other plants and are in sheaths not filled with grease. The FSV tendon wires themselves are protected by a grease coating and the tendon sheath annulus is coated on the inside with a layer of grease. A failure mechanism has been identified at the other plants related to stress corrosion cracking of the tendon wire stress washers when water was present, predominately in the lower end of the vertical tendons. The stress washers are manufactured from a high strength steel which is susceptible to stress corrosion cracking when exposed to a source of hydrogen. The tendon wires at FSV appear to be corroding from the attack of formic and acetic acids generated from microbilogical sources. The corrosion and failures seen to date at FSV seem to be limited to the wires themselves with only one incident of corrosion occurring on the stress washer. No evidence of failure of the stress washers has been detected to date. The licensee has visually examined 10 of the 34 accessible bottom stress washers of the longitudinal tendons and reported the results in a letter dated June 7, 1985. No evidence of cracking was found. However, the possibility of stress washers failing from corrosion cannot be ruled out. The continued presence of moisture in the tendon tubes could lead to failure of the stress washers as seen at other plants. The licensee has proposed an intensified surveillance program which consists of visual inspection of the anchorages and lift-off tests. The licensee proposed to incorporate these inspection requirements into the plant technical -3- specifications under Section 3/4.6.4 "PCRV Integrity" . This intensified surveillance program will require a visual inspection and a report on a sample of 56 tendons at six month intervals. The surveillance program will also require 37 tendons to be lifted-off their shims to determine the amount of prestress available. A sample of 12 tendons are designated as a control set for visual inspection and 8 tendons are the control set for lift-off tests. The samples of 44 visual inspection and 25 lift off tendons will be rotated thru the tendon population. 4. Restart and Re-evaluation: The staff has reviewed the licensee's proposed surveillance and the commitment to incorporate the surveillance requirements into the technical specification. The staff accepts the tendon surveillance proposal and find that the proposal would lead to maintaining the structural integrity of the reactor vessel . The staff also finds that the visual inspection of the tendon wire anchorage washer is currently sufficient to determine if failure of the washers has occurred. The increased tendon surveillance would be sufficient to show any tendency of the stress washers to fail similar to those at another nuclear plant. The staff finds the plant structurally ready for restart. Principal Contributor: H. Polk, DE Date: July 2, 1985 Evaluation of Control Rod Drive Mechanism and Reserve Shutdown System Failures, and PCRV Tendon Degradation Issues Prior to Fort St. Vrain Restart NRC Fin No. A-7290 March 12, 1985 Los Alamos National Laboratory Deborah R. Bennett, Q-13 Gerald W. Fly, Q-13 L. Erik Fugelso, Q-13 Robert Reiswig, MST-6 Stan W. Moore, Q-13 Responsible NRC Individual and Division J. R. Miller/ORBS Prepared for the U.S. Nuclear Regulatory Commission Washington, D.C. 20555 DISCLAIMER Tnis report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, or any of their employees, makes any warranty, expressed or impliec, or assumes any legal liability or responsibility for any third party's use, of any information, apparatus, product or process disclosed in this report or represents that its use by such third party would not infringe privately owned rights. - i - Table of Contents 1.0 Background 1 .1 Assessment Report Restart Issues 1.2 PCRV Tendon Restart Issues 1.3 Purpose of the Technical Evaluation 2.0 Control Rod Drive and Orifice Assemblies 2.1 Failure Mechanisms 2.1.1 Motor Brake Malfunctions 2.1.2 Reduction Gear Mechanism Malfunctions 2.1.3 Motor and Motor Bearing Malfunctions 2.2 Refurbishment Program 2.2.1 CRDOA Refurbishment 2.2.2 Control Rod Cable Replacement 2.2.3 Reserve Shutdown System Material-Related Failure 2.2.4 Purge Flow and Seal Replacement 2.3 CRDM Temperature Recording and CRDM Requalification 2.4 CRDM Preventive/Predictive Maintenance and Surveillance 2.4.1 CRDM Preventive/Predictive Maintenance 2.4.2 CRDM Interim Operational Surveillance 3.0 Moisture Ingress Issues 4.0 PCRV Post-Tensioning Tendon System 4.1 Tendon Accessibility, Extent of Known Degradation and Failure Mechanism 4.2 Tendon Corrosion Corrective Measures 4.3 PCRV Tendon Interim Surveillance 4.4 PCRV Structural Calculations by Los Alamos National Laboratory 5.0 Conclusions 6.0 References - ii - Evaluation of Control Rod Drive Mechanism and Reserve Shutdown System Failures, and PCRV Tendon Degradation Issues Prior to Fort St. Vrain Restart 1.0 Background On June 23, 1984, following a moisture ingress event resulting in a . loss of purge flow to the Control Rod Drive Mechanism (CRDM) cavities, 6 of 37 control rod pairs in the Fort St. Vrain (FSV) High Temperature Gas- - Cooled reactor failed to insert on a scram signal. Subsequently, all six control rod pairs were successfully driven into the core. In July, 1984, an assessment team consisting of Nuclear Regulatory Commission (NRC) personnel from Headquarters, Region III and Region IV, and their technical consultant, Los Alamos National Laboratory, conducted an on-site review of the Control Rod Drive Mechanism failures, overall conduct of plant operations, adequacy of technical specifications and a review of the continued moisture ingress problem. An additional plant visit in August, 1984, reviewed CRDM instrumentation anomalies. 1 .1 Assessment Report Restart Issues Tne results of both assessments were reported in the "Preliminary Report Related to the Restart and Continued Operation of Fort St. Vrain Nuclear Generating Station"1, in October, 1984. Tne report concluded that Fort St. Vrain should not oe restarted until modifications and/or other corrective actions had been taken, or until all control rod drive mechanisms had been inspected and refurbished to provide reasonable as- surance that the control rods would insert automatically on receipt of a scram signal . More specifically, and as included in this technical eval- uation, Reference 1 required Public Service Co. of Colorado (PSC) to com- plete the following, prior to restart: a. The licensee must identify the CRDM failure mechanism(s) and take corrective actions, or, if the mechansm(s) cannot be posi- tively identified, take other compensatory measures to provide assurance of control rod reliability, which could reasonably include refurbishment of all CRDMs. - 1 - b. The licensee must outline and commit to periodic inspection , preventive maintenance and surveillance programs for control rod drive mechanisms and associated position instrumentation. A change in the Technical Specifications shall be proposed to implement a weekly control rod exercise surveillance program for all partially or fully withdrawn control rods. A Limiting Condition for Operation should define control rod operability, and the minimum requirements for rod position indication. c. The licensee must functionally test one-20 weight % boron and one-40 weight % boron hopper from the Reserve Shutdown System (RSS) , to assure the full availability of the RSS, prior to restart. The licensee must outline and commit to periodic in- spection, preventive maintenance and surveillance programs for Reserve Shutdown System material. A change in the Technical Specifications shall be proposed to implement the RSS surveil- lance program. A Limiting Condition for Operation should define and confirm the operability of the Reserve Snutdown System. d. The licensee should develop a procedure requiring reactor shut- down when high levels of moisture exist in the primary coolant, or when CRDM purge flow is lost. e. The licensee should implement a procedure for recording repre- sentative samples of CRDM temperatures at all operating condi- tions, until continuous recording capability is available. f. The licensee should implement procedure to prevent overdriving the control rods past the "Rod-In" limit. g. The licensee must develop a plan to implement any modifications recommended by the PSC Moisture Ingress Committee that are determined, by PSC, to have a high potential for significantly reducing the severity and frequency of moisture ingress events . 1 .2 PCRV Tendon Restart Issues As a result of previously identified tendon degradation in the Pre- stressed Concrete Reactor Vessel (PCRV) post-tensioning system, PSC must complete the following, as comfirmed by Reference 2, prior to restart: - 2 - a. The licensee should submit documentation evaluating the mechan- ism(s) causing corrosion on and failure of tne PCRV tendon wires, and corrective measures to eliminate further tendon degradation, thereby assuring the continued structural integ- rity of the PCRV and its post-tensioning system. b . The licensee should propose and implement a tendon surveillance program that determines the extent of current tendon degrada- tion in the PCRV, and that systematically monitors the rate of tendon corrosion. 1 .3 Purpose of tne Technical Evaluation This document provides a technical review of tne restart issues identified above, and the corrective measures and/or actions proposed by licensee, based on the licensee.'s January 31, 1985 submittals (References given as used in this document), and the meeting between the licensee and NRC at the FSV plant site on February 20-22, 1965, as transcribed in References 3, 4 and 5. 2.0 Control Rod Drive and Orifice Assemblies This section includes a review of CRDM failure mechanisms, Control Rod Drive and Orifice Assemblies (CRDOA) refurbishment, CRDM temperature recording and requalification testing, CRDM preventive/predictive main- tenance and surveillance. 2.1 Failure Mechanisms The failures of control rod pairs to scram, under various operating oonditions , has been documented since 1982,6'7 and are as noted in Table 1 by region, CRDOA number and CRDM purge flow subheader (total of 8 purge flow subheaders) . - 3 - Table 1. Control Rod Failures Date 2/22/82 6/23/84 1/14/85 Region 7 28 6 7 10 14 25 28 28 31 32 CRDOA # 18 44 29 18 14 25 7 44 36 17 15 CRDM Purge 1 1 6 1 7 2 5 1 1 2 3 Subheader # High moisture content in the primary coolant and loss of purge flow were common modes during the 2/22/82 and 6/23/84 events. Substantial descriptions and operating characteristics of the drive motor, friction braKe and dynamic braking, the reduction gear mechanism, the cable drum and cable, and the bearing lubricant are provided in Reference 6. Tne licensee reviewed those CRDM components that could have caused the fail- ures to scram, and postulated various failure mechanisms that could have interacted on each component, as described below. 2.1.1 Motor Brake Malfunctions During a scram, the motor brake is de-energized and released, thereby freeing the motor rotor shaft and gear train assembly to rotate under the torque applied by the weight of the control rods. In the motor brake assembly, failure of the scram contactor to de-energize dc power to the electromagnet was discounted because the operator had removed the brake fuses following the CRDM failures to insert the control rod pairs. According to the licensee, electromagnetic remanence and reduced spring constant in the brake spring plungers (due to elevated tempera- tures) were eliminated as possible failure mechanisms. Some corrosion and rust was identified on the brake disks of CRDOAs 25, 18 and 29. How- ever, the disks of a CRDM motor brake assembly with "discoloration and whatever surface variations„3, p.149 could not be made to stick in an elevated temperature helium environment with high moisture content (test T-228) . The licensee concluded that the motor brake was not instrumental in the failures to scram. Los Alamos agrees with the licensee that the motor brake assembly was probably not related to the CRDM failures. - 4 - 2.1.2 Reduction Gear Mechanism Malfunctions The reduction gear train is driven by the motor rotor shaft, and rotates the cable drum with a gear ratio of 1150 between the motor and drum. The condition of the reduction gear mechanism was postulated by the licensee to potentially contribute to a failure to scram through gear tooth or bearing damage, by the presence of large particulate matter pre- venting gear rotation, and/or the presence of particulate matter in the gears or gear bearings reducing the gear train efficiency--i.e. , the torque transmitted from the gear train to the motor rotor shaft might have been insufficient to overcome the friction of the motor bearings. The licensee stated that no major damage has been identified on sev- eral inspected reduction gear mechanisms, even though some wear and debris were observed. The licensee's analyses indicated that particulates with a size of 0.030 inches in diameter or greater, and with a comparable material composition as the reduction gear mechanism (implying comparable hardness) , would be required to inhibit gear or gear bearing rotation. Analyses of CRDOA debris8 showed the presence of rust, molybdenum di- sulfide and traces of silicon particles, which are relatively soft mate- rials. The average particle of 0.020 inches was uniform in size, and tended to be smaller than that thought to innibit rotation, even though rust particles on the order of 0.0625 to 0.125 inches were scraped off the ring gear pinion housing of CRDOA 18. However, the presence of debris in the gears and gear bearings tended to support the licensee's case of reduced gear train efficiency when sensitivity studies indicated that the motor bearings were only three times more sensitive to debris than the first pinion gear mesh of the reduction gear assembly, and 500 times more sensitive to debris than the cable drum bearings. Los Alamos agrees with the licensee that the presence of debris , especially in the first pinion gear mesh and the gear bearings, could reduce the efficiency of the reduction gear train, and thereby contribute to CRDM failures. 2.1.3 Motor and Motor Hearing Malfunctions During a scram, the motor is de-energized and does not directly con- tribute to the scram process, even though it operates as an induction generator. However, because 1b-20 inch-ounces of resisting torque on the - 5 - motor rotor shaft can forestall scram,9 the friction from the motor bearings can be a significant contributor to the failure to scram. Pos- sible contributions to increase the friction include debris in the bear- ing race, wear on the bearing ball or race, and changes in the lubricant properties during adverse conditions. The licensee reported that debris was observed in the bearing races of CRDOAs 7, 18 and 44, "roughness in rolling the bearing balls was noted in virtually all of the unrefurbished bearings examined",6 and minor race wear was identified. Reference 8 verified that the major debris constituents could be attributed to the motor bearing materials (which includes bearing balls, races, and other bearing components), whereas minor constituents were indicative of the motor itself. The analysis provided little evidence to support the theory that debris had been "washed" into the bearing races. The licensee also determined, because of the relatively close bearing tolerances and because rod weight alone might not produce sufficient "crushing force" to deform bearing particu- late, that bearing operation could be reduced with the presence of par- ticulate matter. Tne licensee therefore concluded that internally gener- ated wear byproducts in the CRDM motor bearings contributed significantly to tne failures to scram. Los Alamos agrees with the licensee that increased friction in the motor bearings, caused by the presence of internally generated debris, could have been a likely contributor to the failures to scram. Los Alamos also agrees with the licensee that the "wash in" theory of debris into the motor bearing races is not supported. Los Alamos contends that the loss of CRDM purge flow allowed primary coolant with high moisture content to enter the CRDM cavity. An indepen- dent literature search indicates that the dry film lubricant, molybdenum disulfide, MoS2 , experiences an increase in its coefficient of fric- tion in the presence of moisture38. Therefore, the increased frictional coefficient of the lubricant on the motor bearings, MoS2, may have also contributed to the CRDM failures by resisting motor rotor shaft rotation. 2.2 Refurbishment Program Tne cause of tne failures to scram could be attributed to several mechanisms such as reduced reduction gear train efficiency, internally - 6 - generated debris in the motor bearings causing increased friction on the motor rotor shaft, and possibly an increased frictional coefficient in the dry film lubricant in the presence of moisture. Because the CRDM failure mechanism cannot be specifically delineated, and because of CRDM cable failures, the licensee has undertaken a refurbishment program, in- volving the CRDM motors and reduction gear mechanisms, on all 37 CRDMs. The licensee reported that the CRDM refurbishment process and a testing program will ensure the ability of the control rods to scram under oper- ating conditions. In addition, the licensee has elected to replace the control rod cabling and other connecting hardware in light of recently identified stress corrosion problems, to replace the Reserve Shutdown System material due to the discovery of material "bridging" during hopper discharge, anc to install seals around certain penetrations into the CRDM cavity to mitigate the effects of primary coolant ingress by natural circulation. 2.2.1 CRDOA Refurbishment The licensee has proposed complete refurbishment of all Control Rod Drive and Orificing Assemblies to ensure that the CRDOAs will perform their intended safety functions, and to avoid potential operability prob- lems that could limit plant availability. As specified in Reference 10, the following major components are to be inspected, tested, refurbished or replaced, as necessary: 1. Control Rod Drive (200) Assembly--shim motor and brake assembly, bearings, reduction gears, limit switches/potentiometers. 2. Orifice Control Mechanism--orifice control motor, bearings, potentiometer, gears, drive shaft and nut, drive shaft housing. 3. Control rod clevis bolts. 4. Reserve Shutdown System--boron balls, rupture disks, DP switch. Design modifications include the replacement of control rod cables , cable end fittings, and cable clevis bolts, the installation of new purge seals into the CRDM cavity, and the installation of RTDs (Resistance Tem- perature Detectors) in all CRDOAs--the impact of tnese design changes will be evaluated later in this report. Each CRDOA will undergo the following series of scram tests in the refurbishment process6: a pre-refurbishment, in-core full scram test; a - 7 - pre-refurbishment full scram test in the Hot Service Facility (HSF) ; a scram test with refurbished reduction gear mechanism and unrefurbished shim motor, using dummy weights; a full scram test using a "standardized" motor, using dummy weights; a scram test with completely refurbished 200 assembly, using dummy weights ; a post-refurbishment, full scram test in the HSF; and finally, a post-refurbishment, full in-core scram test. As designated by the licensee in Reference 6, back-EMF voltage meas- urements from the shim motor will be taken for the series of scram tests conducted before, during and after refurbishment, and should define the CRDM operating characteristics. From the back-EMF voltage measurements, the licensee states that they can generate the following information-- voltage versus time, frequency versus time, voltage versus frequency, acceleration versus time, torque versus time, peak angular velocity, time to peak back-EMF and angular velocity, average torque on motor rotor dur- ing acceleration to peak velocity, maximum torque on motor rotor each 10 second interval, maximum deviation of torque values each 10 second inter- val , and gear train efficiency. The licensee has proposed a CRDOA refurbishment acceptance criterion, taking into account the results of the back-EMF voltage measurements and the resulting calculations of acceleration and torque such that6: 1. Tne minimum calculated average torque during acceleration to peak velocity will be 17.0 inch-ounces ; this value corresponds to an average acceleration to peak velocity of 98.83 radians/ second 2. The maximum torque calculated during "steady-state" will be 7.0 inch-ounces. According to the licensee, final acceptance of a refurbished CRDOA will be based upon the results of its in-core full scram test. Los Alamos agrees with the mechanical refurbishment of all CRDOAs, as the program is currently being implemented by the licensee. In par- ticular, the replacement of shim motor bearings3' pp' 174-75 is con- sidered essential to the refurbishment process. However, the current program of mechanical refurbishment alone cannot ensure CRDOA operability. - 8 - From the documentation presented by the licensee and reviewed earlier in this section, Los Alamos believes that the proposed back-EMF testing and acceptance criteria have potential in providing a data base from which control rod operability might be determined. But, an element of uncer- tainty, as to CRDOA operability based on back-EMF testing, is introduced because the test method and interpretation of its results are still in the developmental stages, and because in-core full scram testing of re- furbished CRDOAs has not yet taken place. Los Alamos recommends that the back-EMF testing method continue to be developed, that the further collection of back-EMF information be used in preparing a statistical data base for possibly defining CRDOA opera- bility, and that more attention be paid to the initial, start-up scram characteristics of the CRDOA, in developing a better understanding of break-away torque effects. In line with Region IV's increased inspection of the refurbishment process, we suggest a review, by Region IV, of all testing results pertaining to CRDOA refurbishment acceptability, after in-core testing is complete, but prior to startup. As an additional method to ensure CRDOA operability during scram, a procedure requiring control rod run-in is recommended. As a post-startup item, Los Alamos recommends that a final determina- tion be made as to the suitability and acceptability of back-EMF testing in defining CRDOA operability. 2.2.2 Control Rod Cable Replacement In September, 1984, the control rod cable on CRDOA 25 was severed in several places during an investigation of a slack cable indication.11 A subsequent metallurgical examination12 of the austenitic 347 stainless steel cable indicated that the cable surface was pitted and cracked, that the delta-like material cracks were typical of stress corrosion cracks , and that the fracture surfaces were brittle in nature. Further investi- gation revealed that the 347 SS cable material was susceptible to stress corrosion when under the existing stressed conditions, and in the presence of chlorides and moisture. Tne potential sources of the chlorides in the primary coolant'con- tributing to the chloride stress corrosion are reviewed in Reference 13. The licensee states that the chlorine occurs as two different species--HC1 - 9 - gas and a salt; the sources of the gas species include the fuel rods , H-327/H-451 graphite, PGX/HLM graphite and the Ti sponge, whereas the sources of the salt species include the ceramic insulation, concrete and water, all to varying degrees. As part of tne overall CRDOA refurbishment program, the licensee elected to replace the control rod cable with Inconel 625, which is con- sidered resistant to chloride stress corrosion, and has increased strength and fatigue properties over the former 347 SS. Cable components and con- necting hardware that were made from materials susceptible to stress cor- rosion, and are being replaced with materials more resistant to stress corrosion include: Component Material 1. Cable and rod portion Inconel 625--high strength of the ball end and resistance to oxidation 2. Anchor, set screw Martensitic steel-high strength, ability to be nitrided, resistance to oxidation 3. Spring, connecting bolt Inconel X-750--high yield strength, resistance to oxidation. Drawing numbers and material information are available in Reference 12. A safety analysis of tne material changes in tne reactor control rod drive and orificing assembly, which are classified as Class I, Safety Related and Safe Shutdown components, is included in Reference 14. Los Alamos metallurgical analyses on a sample of the corroded control rod cable15 also indicate pitting on the cable surface, ductile and brittle fracture surfaces, and to a lesser degree than the licensee, cracking indicative of stress corrosion cracking. Qualitative measure- ments confirm the presence of chlorine on fracture surfaces. Therefore, Los Alamos agrees tnat chloride stress corrosion contributed to the de- graded condition of the control rod cable. The Los Alamos analysis also observed that a certain particle removed from between the individual cable strands of the Los Alamos sample had a "shaved" appearance, and was - 10 - identified as a 7000 series aluminum alloy--the licensee noted that the control rod cable drum is constructed of 7075 aluminum alloy4' p.20, and that no excessive drum wear had been noted. Los Alamos agrees that the licensee's recommended material changes tend to improve the overall resistance of the CHDOA cable components and connecting hardware to chloride stress corrosion. However, Los Alamos also recommends a continued analysis into the sources of tne chlorine and its effects on other reactor components, especially components potentially subjected to high chlorine concentrations such as the bottom plenum or other areas where water could accumulate. 2.2.3 Reserve Shutdown System Material-Related Failure In November, 1964, during the required testing of a 20 weight % boron and a 40 weight % boron hopper in the Reserve Shutdown System, only half of tne RSS material in CRDOA 21 (40 weight % boron) was discharged. The licensee's examination of the undischarged material revealed that the B4C boronated graphite balls had "bridged" together through a crystal- line structure on the ball surfaces. Analyses on the crystalline material indicated that it was boric acid.16 The formation of tne boric acid crystals was caused by moisture reacting with residual boric oxide in the RSS material. It was concluded that the moisture had entered the RSS hopper through the CRDOA vent/purge line by "breathing", and/or by water contamination in the helium purge line. In Reference 16, the licensee proposed a threefold corrective action to the RSS material problems. First, new RSS material, manufactured by Advanced Refractory Technologies (ART) in late 1984 and early 1985, has an order of magnitude less residual boric oxide in the B4C material, and will be installed in all RSS hoppers as part of the overall CRDOA refurbishment program. No effort will be mace to use ART blended RSS material currently in stores4' p.32 unless NRC is notified. Second, an expanded RSS material surveillance program, which will be incorporated into the Technical Specification, will test one 20 weight % boron hopper and one 40 weight % boron hopper during each refueling outage, and will include visual examinations for boric acid crystal formations, chemical analyses of RSS material for boron carbide and leachable boron oxide con- tent. Tnird, efforts will be made to mitigate or eliminate the ingress - 11 - of moisture into the RSS hoppers by installing a knock-out pot, moisture elements, and a back-up helium source for the main CRDOA purge and Reserve Shutdown System purge lines 17 Each knock-out pot will be equipped with a sight glass and a high level alarm in the Control Room. Los Alamos concurs that the crystalline structures on the surface of the B4C RSS balls is meta-boric acid,18 most probably formed by mois- ture reacting with leachable boric oxide in the B4C material. In light of the new RSS material to be used, the increased surveillance efforts , and measures to mitigate the ingress of moisture in the RSS hoppers, Los Alamos agrees that the refurbished RSS should be able to reliably perform its function. 2.2.4 Purge Flow and Seal Replacement Just prior to the June 23, 1984 event when 6 of 37 control rod pairs failed to insert on a scram signal , a high moisture content in the primary coolant resulted in the loss of purge flow into the CRDM cavities. The loss of purge flow may have allowed the additional ingress of moist pri- mary coolant into the CRDM cavities, resulting in mechanisms that may have contributed to the CRDM failures. Because the exact CRDM failure mechanism has not been determined, and to alleviate the possibility of purge flow loss and/.or hign moisture content in the primary coolant con- tributing to future CRDM failures, the licensee has proposed several cor- rective measures19 as part of the overall CRDOA refurbishment program. To provide an accurate measure of the purge flow into the CRDM cavi- ties, tne licensee has proposed that new flow indicators with a range of 0-20 scfm be installed on each helium purge line, providing local indica- tion, remote indication in the Control Room, ana an alarm in the Control Room to indicate low flow conditions.20 A minimum of 8 bypass lines (one line serviced by each of the 8 purge flow subheaders) will be in- stalled prior to restart. Tne licensee intends to install the flow instrumentation4' pg 3-4 on these subheaders when the devices are available. As mentioned in section 2.2.3, to reeuce tne possibility of moisture ingress into the CRDM cavities via the helium purge lines, the licensee will install a knock-out pot, moisture elements and a back-up helium source for the main CRDOA purge and RSS purge lines, prior to criticality - 12 - following the fourth refueling outage4, pg 5. The knock-out pots will be equipped with a sight glass and a high level alarm in the Control Room. The helium trailer, which will act as the back-up source of dry helium for purge, can provide helium at a rate of 7.4 acfm (4.5 lbms/hr per pen- etration at 700 psig) for approximately 2 hours .17 To mitigate the ingress of primary coolant, which could contain moisture, into the CRDM cavity, seals will be installed on four large flow passages into the CRDM cavity--the two passages in the reserve shut- down tube holes , and the two passages over the eye bolts that penetrate the floor of the CRAM cavity.21 Cover plates with integral gaskets will also be installed on the four access openings on the lower CRDM housing. Thermal and mechanical analyses22 have determined that the seal additions will not interfere with the RSS performance under the in- fluence of mechanical , thermal or seismic loadings. Tne flow calculations in Reference 22 conclude that addition of tne mechanical seals to tne RSS pressure tubes and the lifting eyebolts will reduce naturally convective ingress of primary coolant into the CRDM cavity from a flow rate of 0.68 acfm to less than 0.006 acfm. Additional calculations have confirmed that the seals are able to withstand both a design basis slow depressuri- zation transient and a design basis rapid depressurization transient. The licensee has proposed a procedure in Reference 23 tnat basically requires reactor shutdown in the event CRDM purge flow is lost, or if high moisture content is present in tne primary coolant. Los Alamos agrees with tne efforts of tne licensee in monitoring the flow and moisture content of tne helium purge into the CRDM cavities, in restricting the ingress of moisture into the CRD cavities via the purge lines, and in providing a back-up source of helium in case of purge flow loss. From the review of the provided documentation, Los Alamos agrees that the addition of seals and coverplates with integral gaskets will indeed mitigate the ingress of primary coolant and moisture into tne CRD cavities through penetrations. In addition, Los Alamos believes tnat the procedure requiring reactor shutdown with loss of purge flow or high moisture levels in the primary coolant fulfills the requirements of the assessment report1. The licensee defines "high moisture levels" in Reference 23. - 13 - 2.3 CRDM Temperature Recording and CRDM Requalification The lack of direct measurements of CR0M temperatures curing the June 23 event, and during steady state and other transient operating condi- tions, has prompted the installation of RTDs to monitor the CRDM cavity closure plate (ambient) , orifice valve motor plate and control rod erive motor temperatures. Strip chart recorders will continuously record tne three temperatures for each CRDM,24 and will provide a CRUM operating temperature data base. The old data collection surveillance procedure25 will be modified to collect data on a continuous basis.4' p.60 The licensee intends to install the permanent recorders prior to restart.4' p'57 The licensee postulates in Reference 26 that "the maximum temperature rating of the drive mechanism which might inhibit the scram function is 272°F", and in monitoring CRDM_temperatures "the maximum temperature rating of 272°F should not be exceeded during power operation". The licensee has also proposed a CRDOA requalification testing pro- gram that is designee to establisn a temperature at which the CRDOA is qualified for operation.27 The helium test environment will be operated at 250°F, 260°F, 270°F, 2b0°F, 290°F and 300°F with a goal of qualifying all CRDOA components for 300°F operation. Results of tne requalification testing are anticipated by the end of 1985. Los Alamos agrees tnat the placement of CRDOA thermocouples, and the continuous data monitoring at all operating conditions is sufficient to provide a CRDOA temperature data base curing steady state and transient operating conditions. In addition, Los Alamos believes that tne CRDOA is currently only qualified to operate up to 215°F based on the original mechanical CRDOA qualification tests, an NRC recommendation,28 and previous Los Alamos calculations.29 The licensee's argument tnat the CRDOA is qualified for 272°F operation oased on analytical calculations4' p.49 is not sub- stantiated. Tnerefore, Los Alamos recommends tnat CRDOA operation be limited to 215°F until mechanical requalification supports a higher oper- ating temperature. - 14 - 2.4 CRDM Surveillance and Preventive/Predictive Maintenance The licensee has proposed a set of preventive/predictive mainte- nance tests and surveillance inspection procedures that are intended to monitor the performance of the CRDOAs and to determine the overall opera- bility of the CRDOAs during reactor operation. Initial development of these operating tests are considered part of the CRDOA refurbishment pro- gram, and will utilize the data base and resultant trends formulated dur- ing refurbishment. 2.4.1 CRDM Preventive/Predictive Maintenance Tne licensee's CRDOA preventive/predictive maintenance program is proposed in Reference 30. According to the licensee, the normal preven- tive maintenance (PM) program will be implemented on a refueling basis rotational cycle for CRDOAs that would normally be removed for refueling, unless the predictive maintenance (PDM) program indicates tne need for more frequent maintenance. The PM program woulo emphasize the mechanical examination and refurbishment of the shim motor/brake assemoly, the drive train, control rod cable, reserve shutdown system, position potentiom- eters, limit switches, orifice drive motor assembly, orifice drive lead screw, assorted seals, valves, electrical components , bolts and the ao- sorber string. On the other hand, the predictive maintenance techniques would be used to monitor the most important aspect of CRDOA performance--the "scram capability"--by determining the shim motor/brake and gear train performance. The tests proposed in the PDM program include wattage requirements, back-EMF voltages, delivered torque at the motors, scram times, rod drop rates and torques to rotate motor/brake assemblies. Cer- tain aspects of the PDM program would be implemented on a weekly basis to determine scram capability and temperature performance during power oper- ation. The licensee has also proposed that testing information be acquired during reactor shutdown for trending purposes. Los Alamos concurs with the proposed preventive maintenance pro- gram as outlined by the licensee, on the assumption that data acquired during reactor operation will show that predictive maintenance tecnniques can be used to detect a reduction in CRDOA performance. Tne PDM testinb tecnniques are closely linked to tne techniques tnat are being used for - 15 - the acceptance criteria in the refurbishment program, and will therefore be dependent on the suitability and acceptability of back-EMF testing for determining CRDOA operability, as discussed in section 2.2.1. 2.4.2 CRDM Interim Operational Surveillance Tne licensee's CRDOA interim surveillance program is proposed in Reference 31. The surveillance tests are scheduled on a weekly basis , using a 10" rod drop method on all withdrawn and partially inserted con- trol rods, except the regulating rod.5' pg 82 The surveillance tests will obtain data for analysis and long term trending, exercise the rod, test selected circuitry, verify FSAR (Final Safety Analysis Report)9 assumed scram times, and confirm control rod operability. In addition, CRDOA temperature and purge flow information will be collected. For a fully withdrawn rod, analog and digital position information will be obtained, "Rod-Out" lights will be verified on, "Rod-In" and "Slack Cable" lights will be verified off, and the rod will be dropped approximately 10" by de-energizing the brake, while back-EMF data are obtained for future trending. The "Rod-Out" light indication will be verified off, and analog and digital information will be compared, with an acceptable deviation of 10 inches between position indications. The rod will then be withdrawn to the full out position, so that analog and digital positions can again be obtained. Control rods that are par- tially or fully inserted will undergo variations of this method. Quarterly surveillance tests are intended to supplement weekly sur- veillance information, and to verify redundancy of selected control rod position limit switches. Refueling shutdown surveillance will acquire the same information as the weekly and quarterly tests, except full stroke insertion tests will be performed. The operability acceptance criteria, according to the licensee, will be based on distance and time rod drop data used to calculate a conserva- tive average full length scram time. A CRDOA will be considered inoper- able if it does not meet the maximum scram time of 160 seconds as defined in the FSAR9. Such an indication would warrant back-EMF testing in confirming scram operability. Los Alamos agrees that the basic surveillance methodology is suffi- cient to exercise the control rod, verify FSAR scram times, and to test - lb - selected circuitry. However, references to 272°F as the maximum CRDOA operating temperature are still considered inappropriate as discussed in section 2.3, and a 10 inch deviation is not considered acceptable between digital and analog position indications--such a deviation could inadver- tently lead to control rod overdrive through a misinterpretation of rod position. Also, the bacx-EMF testing methods and interpretation of the results are still in the developmental stages , and an engineering deter- mination of the suitability and acceptability of tnis testing method in determining continued CRDOA operability will need to be mace before the licensee can finalize tnis portion of tne surveillance program. 3.0 Moisture Ingress Issues The licensee has submitted32 a listing of tne issues considered, and actions taken, by the FSV Improvement Committee (formerly the FSV Moisture Ingress Committee) in significantly reducing the frequency and severity of moisture ingress events. The issues were divided into four categories: 1 . Issues currently under consideration by the Fort St. Vrain Im- provement Committee. 2. Circulator Auxiliary System modifications yet to be completed prior to startup. 3. Circulator Auxiliary System modifications to be completed prior to startup, provided material availability and schedule permits . 4. Items identified by tne Moisture Ingress Committee wnicn are installed and operational . Los Alamos believes that a listing of intended and installed modifi- cations does not provide any indication as to what any given modification really is, wny they contribute to tne reduction in potential for moisture ingress events , nor which improvements will substantially reduce the severity and frequency of moisture ingress events. Tne licensee nas com- mitted to submit a more explanatory version of the actions to mitigate moisture ingress, prior to restart4, pg 8d. 4.0 PCRV Post-Tensioning Tendon System In the spring of 1984, during scheduled PCRV tendon surveillance, tendons with corroded and broken wires were found. Since that time, the - 17 - licensee has evaluated the corrosion mechanism, has performed lift-off tests on selected tendons to determine their load-carrying capability, and proposed corrective actions and an increased surveillance procedures. 4.1 Tendon Accessibility, Extent of Known Degradation and Failure Mechanism The licensee, in determining the extent of tendon corrosion in the PCRV, determined what fraction of the tendons were available for visual examination and lift-off tests. The tendon system is subdivided into four major groups: the 90 longitudinal (vertical) tendons have 169 wires per tendon; the 210 circumferential tendons in the PCRV sidewall have 152 wires per tendon, and the 50 circumferential tendons in both the top and bottom heads have 169 wires per tendon; tne 24 bottom cross-head tendons, and 24 top cross-head tendons nave 169 wires per tendon. Of the four groups, the licensee states the following accessibility3d: Tendon Group Both Ends Acces. One End Acces. Neither End Acces. Longitudinal • Visual 20 69 1 Lift-off 0 74 16 Circumferential Visual 261 27 2 Lift-off 236 62 12 Bottom cross-head Visual 20 4 0 Lift-off lb 4 4 Top cross-nead Visual 17 7 0 Lift-off 16 6 2 The number of tendons witn known broken wires as identified in the licensee's 1904 surveillance,34 include° 10 longitudinal tendons with 1 to 22 broken wires, 2 circumferential tendons witn 2 and 15 broken wires, 8 bottom cross-head tendons with 1 to 19 broken wires, and no top crDss- head tendons with broken wires. In some cases , tne total number of cor- roded, broken wires include wires broken during lift-off tests , or during retensioning. - 18 - The results of 74 longitudinal lift-off tests35 indicated that tendons with identified broken wires generally had a slightly smaller lift-off value than intact tendons. Thirty lift-off tests on circumfer- ential tendons snowed little change in lift-off value. Some of the fif- teen bottom cross-head tendon lift-off tests showed a definite reduction in lift-off value for tendons with multiple wire breaks. The value of the lift-off test on one top cross-head tendon was nominal. All lift-off test values exceeded tne minimum limits. The licensee conducted metallurgical investigations into the cause of the corrosion, and determined that microbiological attack on the tendon NO-OX-ID CM organic grease caused the formation of formic and acetic acids.34'36 Tne acids, in conjunction with moisture in the tendon tube, vaporized and recondensed on the cooler portions of the tendons--in this case, usually toward the tendon ends. Tne acidic attack resulted in re- duced cross-sectional wire area, stress corrosion cracking, localized tensile overload and wire breakage. Los Alamos believes, based on the documentation presented by the licensee, that microbiological attack of the tendon grease and the resul- tant formation of acetic and formic acids , in the presence of moisture, is a probable cause for the currently observed tendon corrosion, and has led to the subsequent wire breakage through tensile overload. However, Los Alamos believes that the extent of known tendon corrosion, breakage and previous surveillance have not been clearly defined by the licensee. Los Alamos therefore recommends that a complete map oe mane tnat lists each tendon, its visual examinations and lift-off values, and the number and location of corroded and broken wires. An indication of tne degree of wire corrosion would also be desiraole. 4 .2 Tendon Corrosion Corrective Measures The licensee evaluated several methods for arresting the corrosion process,34,36 including the use of ozone as a biocide to kill the micro- organisms, the use of an alkaline grease wnich should not be conducive to microbiological growth, and tne use of an inert blanket consisting of nitrogen gas. The licensee's consultants found that the nitrogen atmo- sphere arrested the growth of the microbes in the NO-OX-ID CM organic - 19 - grease34,35, and eliminated the oxygen which is necessary for the cor- rosion process to continue. eased on these results, and as a snort term action, the licensee has proposed that nitrogen blankets be establisheu on the longitudinal and bottom cross-head tendons. Long term actions would include further investigations into the corrosion process and ar- resting techniques, and the possible installation of additional load cells in monitoring the PCRV behavior. Los Alamos believes that the use of a nitrogen blanket to halt tne corrosion process may be suitable, but difficult to implement as proposed. The tendon tubes are not likely to be leaktight, and maintaining an inert gas atmosphere at a set over-pressure may prove difficult. Consideration might be given to maintaining an intermittent or continuous purge flow through the tendon tubes, as needed, rather than to maintaining a speci- fied overpressure. However, Los Alamos recommends that initially the nitrogen be purged through the individual tendon tubes to remove as much moisture as possible, and that gas samples be used to monitor moisture and oxygen reduction. Further investigation into the long term effects of a nitrogen blanket on tendons, tne corrosion process and currently available corrosion acids are also recommended. 4 .3 PCRV Tencon Interim Surveillance Because the total extent of tendon corrosion in tne PCRV is unknown, because the rate of existing corrosion is unknown, and because tne use of a nitrogen blanket as an arrest to the corrosion process is an unknown, the licensee has proposed an interim surveillance program designed to address eacn of these issues. ' pp•16k-7' The interim tendon surveil- lance program would include increased visual and lift-off surveillance for tnree years, or until effective corrosion control has been estab- lished. Two populations of tendons would be inspected--a population of tendons that have not been previously identified as being corroded, and a control population with known corrosion. On a six-month frequency, visual surveillance of both tendon ends, when accessible, would include: - 20 - Tendon Group No. of New Tendons No. of Control Tendons Longitudinal 24 6 Circumferential 13 3 Bottom cross-head 6 2 Top cross-head 1 1 Lift-off tests would be performed on two frequencies--an to month frequency for the population of new tendons, and a b month frequency for the control population. The number of tendons for lift-off will include: Tendon Group No. of New Tendons No. of Control Tendons Longitudinal 12 3 Circumferential 13 3 Bottom cross-head • 3 1 Top cross-head 1 1 As an acceptance criteria, the licensee proposed that, based on vis- ual examinations, a mandatory engineering evaluation be conducted on any tendon that has 20% of its wires broken. For any tendon that has only one accessible end, the mandatory engineering evaluation will be con- ducted when any tendon has lob; of its wires broken. The control tendon population will include those tendons with the worst known corrosion with ready accessibility. Los Alamos agrees that the increased tendon surveillance program of the nature proposed by the licensee will provide more information on the extent of corrosion in the PCRV by inspecting new tendons each surveil- lance, and at the same time, monitor the rate of corrosion with the con- trol tendon population. Tne increased surveillance should also determine the effectiveness of the nitrogen blanket in arresting corrosion, or any other corrective measure the licensee may propose. Los Alamos recommends that the licensee submit an outline of the intended mandatory engineering evaluation, which should include all lift-off, load cell and relaxation data incorporated into a safety evaluation. The licensee should define the extent of the visual and lift-off testing procedures, and could use US/NRC Regulatory Guide 1.3537 for guidance. - 21 - 4.4 PCRV Structural Calculations by Los Alamos National Laboratory The PCRV tendons are intended to apply sufficient compression in the concrete to balance or exceed the circumferential and vertical tension in the concrete that results from the internal pressure. A combined analyt- ical and numerical study39 was undertaken by Los Alamos National Laboratory to evaluate the evolution of these stresses, both to the ini- tial prestressing and to subsequent partial and total rupture of tnese tendons. At the stress levels anticipated in the concrete, and for the anticipated operating life span of tne PCRV, the concrete benavior was modeled as a linear viscoelastic solid with the creep strain varying pro- portionally with the logarithm of time at constant stress tnrougnout the projected reactor lifetime. A one-dimensional model of a long concrete column of rectangular cross-section, with an embedded prestressing tendon along the length, was used to evaluate the concrete and steel stresses as well as tne hold-down and lift-off forces. Tnese were evaluated for the intact tendons and the degraded tendons. The degree of tendon degradation is described through the ratio of the number of unbroken strands to the original number of strands . Initial time of rupture was varied from the time of initial prestressing to 400 days after emplacement. Tne formulation led to an integral equation, which was solved numerically. The hold-down forces decayed approximately with the logarithm of time and for both the extreme observed degradation (21 broken strands) and for a more extreme case (40 broken strands) , the hold-down force still exceeded the minimum safety design requirements. In addition, several finite element calculations, using the finite element code N0NSAP-C, were made to evaluate complete tendon failure in a 60° sector of the Fort St. Vrain PCRV. This code has an extensive material library of constitutive relations to model the various properties of concrete, together with a specialized element model to simulate pre- stressing tendons. Two rows of vertical and an arc row of circumferential tendons were incorporated in the model as a baseline calculation. Tne tendons were prestressed to 70% of the ultimate and an internal pressure of 775 psi was applied (tnis pressure is the internal pressure of tne helium coolant in the HTGR) and the creep of the concrete and slow decay of tne tendon stresses were evaluated out to 30,000 days. Then, three - 22 - cases wherein one tendon was removed at one day were evaluated. First the middle vertical tendon in the outer row and in line with the outer buttress was removed. Second, an inner vertical tendon opposite the thinnest portion of the PCRV wall was removed. Finally, an inner layer circumferential tendon at midheight was removed. Stress redistributions at 300 days after ruptures were calculated and shifts of tne remaining tendon loads to accommodate the broken tendon were calculated. Regions of local tensile and snear stress in the concrete portion of tne PCRV were identified and related to overall structural integrity. With all tendons present, the mean vertical stress was about -7b0 psi, the radial stress decreased from the applied internal pressure of -705 to about -1200 psi at the ring of circumferential tendons and tne tangential stress ranged from -2400 psi at tne inner wall to about -2200 psi at the same place. Removal of a vertical tendon reduced the mean axial stress by about +40 psi, the local tangential stress by -10 psi and did not materially affect the radial stress. Removal of a circumferential tendon reduced the mean tangential stress by +30 psi and the local axial stress by -80 psi. The vertical hold-down force from zero days tnrough 30,000 days decreased linearly and remained above the prescribed safety limit, as did the circumferential hold-down force. Comparison of the analytical solution and a small finite element proolem simulating the analytical problem was made to verify the visco- elastic creep models and the tendon element in the NUNSAP-C code. Excel- lent agreement for stresses, strains and nold-down forces was ootained. 5.0 Conclusions Los Alamos concludes that the licensee, Public Service Co. of Colorado, has made a conscientious effort to address all of the restart issues listed in the assessment report.1 The refurbishment program on all CRLOAs provides confidence in CRDOA operability during reactor opera- tion and the ability to scram, even if the exact "failure to scram" mech- anism has not been defined. Questions concerning the reliability of the back-EMF testing procedure on the shim motor/brake assembly in determining control rod operational acceptability still exist, but further method development, more experience with result interpretation, and in-core - 23 - testing may alleviate the questions. Until CRDOA operability can defin- itely be ascertained with these methods, we recommend that the licensee have backup measures such as rod run-in following scram. Control rod cable and connecting hardware material replacement, along with replacement of the Reserve Shutdown System material, serve to rectify the material problems brought on by corrosive mechanisms. In light of chloride stress corrosion problems, Los Alamos also recommends that all reactor components exposed to the primary coolant be reviewed for susceptibility to chloride attack, especially the PCRV liner. Review should continue into the source of cnlorine and methods to elimi- nate its generation and presence. The effects of purge flow loss have not been determined to be in- strumental in CRDOA failures to scram, yet the licensee has committed to maintaining purge flow by external means, and to reducing the effects of primary coolant naturally convecting into the CRDOA cavity with extra seal installation. Even though current qualified CRDOA operating temperatures are very much in question, the licensee is in the process of requalifing the mech- anism for temperatures more in line with those anticipated during reactor operation. From a mechanical standpoint, CHDOA preventive/predictive mainte- nance procedures are certainly reasonable, but like the proposed surveil- lance program, they are dependent on back-EMF testing methods wnicn are still in the developmental stages. Evaluation of moisture ingress corrective measures was difficult due to the lack of information with which to understand the measures taken. The licensee has committed to submit a more explanatory version of the actions to mitigate moisture ingress prior to restart. The extent of PCRV tendon degradation is not well known, even if the licensee may nave determined the cause of the corrosion. Further investi- gation into arresting measures is definitely required, especially because the nitrogen blanket technique may be so difficult to employ. However, the interim surveillance program should provide information on the degree and rate of corrosion, in addition to establishing a tendon wire loss acceptance criteria. The tendon acceptance criteria should ensure PCRV margins to safety. - 24 - 6.0 References 1 . "Preliminary Report Related to the Restart and Continued Operation of Fort St. Vrain Nuclear Generating Station, " Docket No. 50-267, Public Service Co. of Colorado, October, 19d4. 2. "Review of Dallas Meeting (1/15/85) and Restart Committments" , letter from Martin, NRC/Reg IV, to Lee, PSC, 1/17/85. 3. "Fort St. Vrain Meeting, NRC-PSC, February 20, 1985," Volumes I, II and III, recorded and transcribed by Koenig & Patterson, Inc. 4. "Fort St. Vrain Meeting, NRC-PSC, February 21, 1985," Volumes I and II, recorded and transcribed by Koenig & Patterson, Inc. 5. "Fort St. Vrain Meeting, NRC-PSC, February 22, 1985," Volumes I and II, recorded and transcribed by Koenig & Patterson, Inc. 6. "Engineering Report on CRDOA Failures to Scram-Control Rod Drive and Orifice Assemblies, " PSC suomittal P-85037, 1/31/85. 7. "Failure of Three CRDOAs to SCRAM, " PSC submittal P-85029, 1/28/85 . 8. "bearing Debris Analysis, " PSC submittal P-85017, 1/18/85. 9. "Fort St. Vrain Nuclear Generating Station, Updated Final Safety Analysis Report, " Public Service Co. of Colorado. 10. "CRDOA Refurbishment Program Report, " PSC submittal P-85040-2, 1/31/85. 11. "Control Rod Drive Cable Replacement, " PSC submittal P-85032-2, 1/20/85. 12. "Control Rod Drive Cable Replacement Report, " GA Technologies Document 907822, Attachment 1 to PSC submittal P-85032-2, 1/31/85 . 13. "Investigations into Sources of Chloride in FSV Primary Circuit," PSC submittal P-86036, 1/31/85. 14. "Safety Analysis Report--Change in Material of the FSC Control Rod and Orifice Assemblies, " Attachment 2 to PSC submittal P-85032-2, 1/31/85. 15. "FSV Control Rod Cable Metallurgical Examinations," draft report from Los Alamos National Laboratory, 3/85. 16. "Report on Reserve Shutdown Absorber Material ," PSC submittal P-85027, 1/28/85. 17. "Moisture Control in CRDOA Purge Lines, " PSC submittal P-85032-9, 1/20/85. - 25 - 18. "FSV Reserve Snutaown System Material Metallurgical Examinations, " draft report from Los Alamos National Laboratory, 3/85. 19. "CRDOA Moisture/Purge Flow, " PSC submittal P-85032-6, 1/20/85. 20. "Modifications to CRDOA Helium Purge Supply, " PSC submittal P-85032-8, 1/20/65. 21. "Control Rod Drive Cavity Seals, " PSC submittal P-85032-7, 1/20/85. 22. "FSV CRD Cavity Seals Design Report, " GA Tecnnologies Document 907604, Attachment 1 to PSC submittal P-65032-7, 1/20/85. 23. "Operations Order No. 84-17 Describing Operator Actions Upon a Loss of Purge Flow and or Detection of Hign Moisture Levels in Primary Coolant, " PSC submittal P-85040-8, 1/31/85. 24. "CED Temperature and Helium Purge Flow Recorders, " PSC submittal P-85032-3, 1/20/65. 25. "Current CRD Temperature Data Collection Procedure Which Requires Station Manager Notification Upon Discovery of a Measured CRD Temperature in Excess of 250°F, " PSC submittal P-85040-9, 1/31/85. 2b. "Control Rod System Operability Evaluation Report, " PSC submittal P-85040-1 , 1/31/85. 27. "CRDOA Mechanism Temperatures Environmental Requalification, " P.C submittal P-85032-1, 1/20/85. 26. Letter from Robert A. Clark, Chief, ORd3, to 0. R. Lee, PSCo. , December 2, 1982. 29. Meier, K. , "Fort St. Vrain Reactor Control Roo Drive Mechanism Over- Temperature Problem, " Los Alamos National Laboratory, 1962. 30. "CRDOA Proposed Preventive/Predictive Maintenance Program Report, " PSC submittal P-85040-3, 1/31/05. 31. "CRDOA Interim Surveillance Program Report, " PSC submittal P-85040-5, 1/31/85. 32. "FSV Improvement Committee Actions," PSC submittal P-85022, 1/24/85. 33. "Tendon Accessibility Report," PSCo. letter from Warembourg, PSC, to Jonnson, NEC/Reg IV, PSC submittal P-84523, 12/14/64. 34. "Lab Report No. 52--Examination of Failed Wires from Fort St. 'drain Unit No. 1," PSC submittal P-o4543-4, 1/24/65. 35. "Liftoff Tests, " Attacnment 1 to "Engineering Report on Fort St . Vrain Tenaons, " PSC submittal P-84543, 12/31/84. - 26 - 36. Thurgood, Roberts and Epstein, "Evaluation of the Causes of Corrosion in the Fort St. Vrain Post-Tensioning Tendon Wires, " GA Tecnnologies , PSC submittal P-84543-5, 1/24.85. 37. US/NRC Regulatory Guide, Rev. 2, January 1976 . 38. Clauss, F. J. , "Solid Lubricants ano Self Lubricating Solids", Academic Press, 1972. 39. Fugelso, E. and Anderson, C. , "Evaluation of Concrete Crrep and Stress Redistribution in the Fort st. Vrain PCRV Following Rupture of Prestressing Tendons", Los Alamos National Laboratory, October 31 , 1984. - 27 - SSINS NO. : 6835 IN 85-49 UNITED STATES WELD C0""iv rc"1MlSS!RNERS NUCLEAR REGULATORY COMMISSION OFFICE OF INSPECTION AND ENFORCEMENT 5, n WASHINGTON, D.C. 20555 DJ5. SC��?S�? 'i� JUL 1 01985 July 1, 1985 i IE INFORMATION NOTICE NO: 85-49: RELAY CALIBRATION PROBLEM Addressees: All nuclear power reactor facilities holding an operating license (OL) or a construction permit (CP). Purpose: This information notice provides information on a potentially significant problem pertaining to relay orientation during calibration and operation. It is expected that recipients will review the information for applicability to their facilities and consider actions, if appropriate, to preclude a similar problem occurring at their facilities. However, suggestions contained in this information notice do not constitute NRC requirements; therefore, no specific action or written response is required. Discussion: South Carolina Electric and Gas Company, at the Summer nuclear power plant, recently discovered that there was a significant error in the calibration of several E-7000 series Agastat time-delay relays. The licensee' s procedure for replacement of Agastat timing relays did not require calibration of these relays in the installed position. During relay replacement, these time-delay relays were "bench" calibrated in a horizontal position, then field mounted in a vertical orientation. Subsequent time-delay measurements determined that the delay times of the installed devices were as much as 30% greater than that established during "bench" calibration. Licensee investigation of this anomaly determined that the manufacturer' s data sheet for installation and operation of these relays identifies a potential for the observed calibration error. This data sheet states that a dial calibration error may result if the device is mounted horizontally without the manufacturer' s supplied horizontal operation options. Because these relays were mounted vertically at this facility, the potential for error as a result of horizontal "bench" calibration was overlooked. Because of this oversight, the licensee' s relay replacement procedure did not specify device orientation during calibration. Several of these relays were functionally arrayed (sequentially or in parallel) in safety-related applications, such as in the emergency diesel generator under- voltage start circuitry. The above identified time-delay error went undetected for a period of approximately 4 months for some of these relays. This occurred because the licensee' s replacement procedure specified only an operational or functional postmaintenance test of the system containing the relays. This method of testing did not detect the individual relay time-delay calibration errors. 8506260654 1,a ,r-+4 rhs18 IN 85-49 July 1, 1985 Page 2 of 2 The licensee has subsequently revised their relay calibration procedures to calibrate or check the calibration of relays after mounting in place, where practical . The procedures were also revised to ensure "bench" calibrations are performed in the same orientation as mounted, where applicable. Personnel involved in relay calibration have received training on the revised procedures. 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 NRC Regional Office or this office. 4gEdwar Je Director Divisio of Emergency Preparedness and Engineering Response Office of Inspection and Enforcement Technical Contact: H. Bailey, IE (301) 492-9006 Attachment: List of Recently Issued IE Information Notices Attachment 1 IN 85-49 July 1, 1985 LIST OF RECENTLY ISSUED IE INFORMATION NOTICES Information Date of Notice No. Subject Issue Issued to 85-48 Respirator Users Notice: 6/19/85 All power reactor Defective Self-Contained facilities holding Breathing Apparatus Air an OL or CP, research, Cylinders and test reactor, fuel cycle and Priority 1 material licensees 85-47 Potential Effect Of Line- 6/18/85 All power reactor Induced Vibration On Certain facilities holding Target Rock Solenoid-Operated an OL or CP Valves 85-46 Clarification Of Several 6/10/85 All power reactor Aspects Of Removable Radio- facilities holding active Surface Contamination an OL Limits For Transport Packages 85-45 Potential Seismic Interaction 6/6/85 All power reactor Involving The Movable In-Core facilities holding Flux Mapping System Used In an OL or CP Westinghouse Designed Plants 85-44 Emergency Communication 5/30/85 All power reactor System Monthly Test facilities holding an OL 85-43 Radiography Events At Power 5/30/85 All power reactor Reactors facilities holding an OL or CP 85-42 Loose Phosphor In Panasonic 5/29/85 All power reactor 800 Series Badge Thermo- facilities holding luminescent Dosimeter (TLD) an OL or CP Elements 85-41 Scheduling Of Pre-Licensing 5/24/85 All power reactor Emergency Preparedness facilities holding Exercises a CP OL = Operating License CP = Construction Permit Hello