ML20216J199

Evaluation of Long-Term Effects of Moisture Ingress in Fort St Vrain Nuclear Reactor

ML20216J199
Person / Time
Site:	Fort Saint Vrain
Issue date:	02/04/1987
From:	Gitterman M, Martin R, Scott Moore ADVANCED SCIENCE & TECHNOLOGY ASSOCIATES, INC., LOS ALAMOS NATIONAL LABORATORY
To:	Heitner K NRC
Shared Package
ML20216J189	List: ML20216J184 ML20216J192 ML20216J199
References
CON-FIN-A-7290 TAC-59936, NUDOCS 8707020240
Download: ML20216J199 (21)

Text

^ ,

ENCLOSURE 2 Evaluation of Long-Term Effects of Moisture ingress in the Fort St. Vrain Nuclear Reactor I

Los Alamos National Laboratory S. W. Moore, MEE-13 R. A. Martin, MEE-13 M.Gitterman, ASTA(Consultant)

NRC Fin No. A-7290 February 4,1987 Responsible NRC Individual and Division K. Heitner/ Operating Reactors Branch 3 Prepared for the U. S. Nuclear Regulatory Commission Washington, D. C. 20555

$$$702o240070617 p ADOCK 0500o267 PDR

s.

i

\

DISCLAIMER This 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 implied, 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.

s I

.[.

CONTENTS

1. BACKGROUND A. Fort St. Vrain Moisture ingress B. Previous issues Considered C. Los Alamos Scope of Work II.

SUMMARY

OF ASTA RESULTS  :

A. Sources of Moisture ingress B. Quantity of Water C. Water Removal 111. ASSESSMENT OF FORT ST.VRAIN MOISTURE INGRESS SAFET(ISSUES AND IDENTIFICATION OF CORE COMPONENTS REQUIRING FURTHER JUSTIFICATION A. Components Exposed to Water in the Primary Coolant Helium Flow Regions B. Components Exposed to Water in the Purge Helium Flow Regions IV. ANALYTICAL TOOLS FOR CORROSION DAMAGE EVALUATION V. CORROSION EXPERIMENTS VI. EVALUATION AND RECOMMENDATIONS A. General '

B. Operational Aspects to Reduce the Frequency and Severity of Moisture ingress incidents C. Suggestions to Minimize Corrosion Damage to Plant Components D. Water Removal E. Non-Destructive Tests F. Conclusion Vll. REFERENCES

- li -

1. BACKGROUND A. Fort St. Vrain Moisture Inoress The Fort St. Vrain (FSV) high-temperature helium gas-cooled reactor has a long history of moisture ingress events covering a time span from July 19,1974 to the present. These events have led to loss of availability of the plant to produce power and have raised a number of safety issues including Control Rod Drive Mechanism (CRDM) failures to scram, moisture induced Reserve Shutdown System (RSS) failure, material corrosion problems, and instrumeration anomalies.

B. Previous issues Considered Some of the moisture ingress safety issues have been discussed in detail elsewhere .2. These issues include the following:

1

1. CRDM and Control Rod Drive Orifice Assembly (CRDOA) failure mechanisms and refurbishment program. .

2. RSS material related failure.

3. CRDM cavity purge flow and seal replacemet;!.

4. Control rod instrumentation anomalies.

5. Effects of chloride contamination-induced stress corrosion including stress corrosion cracking (SCC) of metallic components.

On June 23,1984, following a moisture ingress event resulting in a bss of purge flow to the CRDM cavities,6 of 37 control rod pairs in FSV failed to insert on a scram signal. In July 1984, an assessment team consisting of Nuclear Regulatory Commission (NRC) personnel from Headquarters, Region ill and Region IV, and personnel from Los Alamos conducted an on-site review of the CRDM failures, overall conduct of plant operations, adequacy of technical specifications, and a review of the continued moisture ingress problem. An additional plant visitin August, 1984, reviewed CRDM instrumentation anomalies.

In November.1984, during the required testing of a 20 weight % boron and a 40 weight % boron hopper in the RSS, only half of the RSS materialin 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 because of a boric acid crystalline structure 1

I l

on the ball surfaces. The formation of the 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 I vent / purge line by " breathing," and/or by water contamination in the helium purge line.

On July 30,1984, the NRC was informed of various control rod I instrumentation anomalies from several regions in FSV. The eleven anomalies observed included simultaneous rod-in and rod-out indications, out-limit switch lights remaining lit, indications of partial rod withdrawal, no position signals, disparity between analog and digital information, a slack-cable indication, and a short-circuited shim motor. A team of personnel from NRC and Los Alamos was sent to the plant on August 1--3,1984 to review the situation with regard to continued safe shutdown and overall plant safety.

These water ingress related problems were described in Ref.1. Also described are recommended programs of replacement, refurbishment, and future inspection and prevention. It is assumed that these recommend-ations are being followed.

After the June 1984 shutdown of FSV, SCC failuren were detected in some of the type 347 stainless steel control rod cables. . Metallurgical examinations revealed the presence of chloride on the c:tble surfaces. As a consequence the Public Service Company of Colorado contracted with GA Technologies, Inc. (GAT) to review the safety impt' cations of chloride exposure of other components bathed by the primary coolant. This study was begun in December 1984 and completed in March 1985.2 While the GAT study was underway and while reassembling helium circulator C2102, one of the high strength primary closure bolts failed during re-torquing. The cause of failure was determined to be SCC.

Examination of this bolt and others exposed to the primary coolant and randomly selected from the same circulator showed evidence of incipient SCC and surface chloride contamination levels similar to that previously observed on the control rod cables.

As stated in Ref. 2, The presence of chlorides in the primary coolant of the reactor has a number of potentially adverse complications including enhanced elevated temperature corrosion during

normal reactor operation and enhanced general corrosion, pitting corrosion, stress corrosion cracking, and hydrogen embrittlement when water is also present in the rea-tor...Thus, it is necessary to consider imprcations on a component by-component basis. Based on such a preliminary review,it was concluded that the phenomena most likely to be important in the reactor are:

1. Stress corrosion cracking of austenitic stainless steel,

2. Stress corrosion / hydrogen embrittlement of high strength steel and

3. General and pitting corrosion of carbon steel and low a!!oy stcel.

The FSV components evaluated by GAT were exposed to the reactor primary coolant, made of materials susceptible to chloride-induced J corrosion failure, and stressed and/or subject to an occasional moisture environment. These components were:

1. PCRV cavity liner

2. Plenum elements

3. Core restraint devices

4. Helium purification system filters and hydrogen getter

5. Instrument and sensor lines

6. Control rods and CRDM assembly

7. Thermal barrier attachments A metallurgical and engineering evaluation was performed for each of these components. The engineering evaluations considered the failure and safety consequences for each component and concluded,

1. The component failure modes and failure consequences of credible chloride-induced failures were found to be either similar or less severe than those failure cases which have already been considered in the FSAR.

2. The consequences of the chloride-induced failures identified would not adversely affect public health and safety, which is consistent with ,

j similar results in the FSAR.

t C. l.os Alamos Scoce of Wark j The purpose of the present Los Alamos moisture ingress task was to ]

j evaluate the long term effects of moisture ingress on core cornponents '

exposed to the primary helium coolant. Specific subtasks were to:

1. Review the licensee's data and periodic progress reports to determine the frequency, extent and mitigation of moisture ingress events since plant startup. i l

l 1

2. Identify all core components and systems exposed to the primary helium coolant that could be affected on a short and/or long term basis, by moisture ingress events, and determine the effects of moisture ingress on ,

identified core components and systems. l

3. Consult with NRR and Region IV to determine those identified core l components that require further justification as to their performance. '

Provide analytical support, as required.

4. Report all findings and conclusions on the long term effects of l

moisture ingress on primary, auxiliary, and safe shutdown core components and systems. Include any potential problems that rnay arise as a result of moisture ingress events, and/or component degradation.

Items 1.,2., and 5 were performed by Mr. Morris Gitterman through a subcontract between Los Alamos and Advanced Science and Technology Associates (ASTA), Inc. (Address: 337 S. Cedros Ave., Suite J, Solana Beach, CA 92075. Phone: 714-755-5051). The ASTA contract number was 9-X65-M4065-9. Portions of the report3 prepared by ASTAin fulfillment of this contract have been summarized in Secs. II.,Ill., and VI.

The ASTA report 1) identified all the known moisture ingress occurrences in the FSV plant (some minor incidents were omitted),2) identified the components which were affected by these moisture ingresses, and 3) outlined the effects of these moisture ingresses on the components.

Los Alamos initiated library searches, ordered reports and reviewed the information available froni the NRC Public Documents Library. About 30 lbs of pertinent reports were selected and supplied to ASTA to support their efforts. ASTA also was able to obtain some data through GAT public information sources. Only a few reports were provided by the NRC as described in paragraph 8., NRC Furnished Materials of the " Project and Budget Proposalfor NRC Work," FIN Number 7290.

The purpose of the present report is to 1) summarize the entire effort for the moisture ingress evaluation, including the tasks performed by LANL and ASTA, and 2) provide final conclusions.

4

4

11.

SUMMARY

OF ASTARESULTS The ASTA report consists of six sections: i i

1. Introduction l

2. Water inleakage

3. Water Inteakage incidents

4. Potential Effects of Water Inteakage

5. Figures

6. Documents Reviewed and References Here we will briefly review and summarize Sec. 2. of the ASTA report.8 The term " ingress" is used interchangeably with "inleakage."

The ASTA study identified 45 incidents of moisture ingress between July 19.1974 and July 24.1985. These incidents involved about 76.300 lbs (9.155 gan of water. Some incidents involving relatively small quantities of water were ignored. Some of the effects of these ingress incidents on individual components were discussed in Sec. l. B.

A. Sources of Moisture ingress The sources of moisture ingress can be divided into actual,and potential sources. There have been oniv two actual sources of moisture incress:

1. the helium circulator auxiliary system, and

2. the economizer / evaporator /superheater (EES) section of the steam generators.

The circulator auxiliarv system was resoonsible for more than 99% of all the water that has entered the Prestressed Concrete Reactor Vessel (PCRVL From this system water has leaked into the PCRV via three paths:

1. along the shaft of a helium circulator,

2. via the buffer helium lines to the circulators, and

3. via the helium pressurization lines serving the PCRV penetrations in which the circulators are located.

Both in terms of number of events and amount of water involved, the path along the shaft of a helium circulator (path 1. above) has allowed by far

. j the most water to enter the PCRV.

Water ingress from the EES section of the steam generators occurred in two instances. Both of these pin-hole sized leaks were orders of magnitude less than the plant design basis EES tube rupture accidents.

1

/

A number of ootential sources of moisture were identified by ASTA.

These were:

)

l l

1. steam generators,

2. auxiliary boiler,

3. liner cooling system,

4. core support floor,

5. purification cooling water system, and j

6. helium storage system. 1 However none of the actual moisture incress incidents reviewed by ASTA involved these systems. Further. none of these systems exceot for S. (the

~

I curification coolina water system) is considered a credible maior source of water incress.

B. Quantity of Water There is no method for determining the quantity of water that is l leaking into the PCRV during an ingress event. An estimate of quantity can be made after the fact during primary coolant cleanup. Although the '

cleanup process can take weeks, the water removed by the Helium Purification System can be measured.

C. Water Removal By far most of the water leaking into the PCRV leaks past the seals on the circulator shafts. Except for the small amount that evaporates, most of this water flows into the circulator inlet duct and then through a gap in the ducting down onto the PCRV bottom head where it soaks the Kaowoolinsulation. Here at the lowest point most of the water is trapped against the cooled (to 110 F) liner and out of direct contact with the bulk hot helium flow.

\

Unfortunately, this water can only be removed from the PCRV by )

evaporation and then processing moist helium in the purification system during a very slow drying-out process. And as the drying proceeds, chloride enriched aqueous regions are formed. Chloride enriched debris are left behind when all of the water is removed. This debris is available to re-dissolve during the next water ingress.

111. ASSESSMENT OF FSV MOISTURE INGRESS SAFETY ISSUES AND IDENTIFICATION OF COMPONENTS REQUIRING FURTHER JUSTIFICATION in this section we will briefly review and summarize Sec. 4. of the ASTA report, Potential Effects of Water inleakage.8 The metallurgical effects of chlorides in the FSV primary coolant were discussed elsewhere.

A Comoonents Exoosed to Water in the Primarv Coolant Helium Flow Reaions With the exception of the Helium Purification Coolers these components are located inside the PCRV. Over the history of the reactor some components have been frequently exposed to water and some have been rarely exposed. Those components frequently exposed are:

1. PCRV bottom head liner,

2. core support floor columns,

3. Kaowoolinsulation,

4. insulation mounting studs, seal sheets, and cover sheets,

5. helium circulators,

6. steam generators,

7. helium circulator, steam Generator, and bottom access penetrations, up to the primary closure, and the

8. Helium Purification System.

In no case are the effects of corrosion of these comoonents judged to affect the health or safety of the oublic. Those failures that aooear credible have been analyzed in the Final Safety Analvsis Reoort (FSAR) for the olant. where thev were shown to result in acceotable consecuences to the oublic at larae. While no further justification of these components appears necessary, some safety related actions appear prudent and are recommended in Sec. VI.

7 W

Those components away from the PCRV bottom head liner are rarely {

exposed to moisture. The three possible transport' mechanisms within the {

J PCRV are:

1. liquid droplet entrainment,

2. moisture ingress from a source other than the circulator shaft, and

3. evaporation / condensation in the PCRV top head penetration from a helium and water vapor mixture.

Condensation is believed to have played a role in depositing chloride laden moisture on the control rod drive cables, which eventually caused stress corrosion failure of a cable.

B. Comoonents Exoosed to Water in the Purae Helium Flow Reaions Pure helium is used in the PCRV penetrations for purging and pressurizing. However, on two known occasions of equipment failure, targe amounts of water entered the purified helium distribution piping.

Rust from this source caused blockage of pressurizing lines to five of 37 refueling penetrations. Rust has been observed frequently in the circulator auxiliary system. The helium purge lines have brought water into the CRDOA mechanisms as well as into the RSS hoppers of some of the l

refueling penetrations.

IV. ANALYTICALTOOLS FOR CORROSION DAMAGE EVALUATION 2 This section has several goals. The first is to provide NRC with an understanding of the phenomena involved in moisture ingress events as we review the complexity of corrosion modeling and analysis. Next, we will indicate that because of the complexities and uncertainties in the FSV

! systems, developing a model (preferably computerized with a basis in and confirmation using experiments) "by which estimates could be made of component degradation so that any future preventive methods proposed by the licensee could be rationally evaluated"is beyond the present Los Alamos scope of work. Finally, we will present a conservative calculation of corrosion rate that illustrates some of the uncertainties and necessary assumptions specific to the FSV case. We will see that calculations such as these are instructive, but because of the uncertainties could hardly be i used to evaluate preventive methods proposed by the licensee.

I 1

There are a number of types of corrosion, some of which are known to

have occurred in FSV. The common relevant types of corrosion include:

4 f

I

1. stress-assisted,

2. general (or uniform attack),

3. localized (including pitting and crevice),

4. hydrogen damage (including embrittlement)

5. galvanic (or two metal),

6. intergrannular, and

7. erosion Corrosion science involves a marriage between metallurgy and electrochemistry. It attempts to predict (1) whether or not a corrosion reaction will occur, and if it does, (2) the rate at which the reaction will proceed to destroy metal. Basic considerations include identifying the affected material composition, heat treatment, and shape as well as identifying environmental parameters. (The environmental parameters may be time varying and generally difficult to determine accurately.) The environmental parameters include chemical composition, temperature, pressure, and pH (acidity or concentration of H+ ions). Any wrong combination of material plus environmental conditions may lead to the formation of corrosion products.

To determine whether corrosion will occur, an equilibrium thermodynamical analysis can be performed. By combining known relationships for the free energy in a specific chemical reaction and the associated electrochemical reaction, a relationship between the cell potential, E, for a given corroding species and the pH can be derived.

Such Computer generated plots of E vs pH are called Pourbaix diagrams.5 diagrams portray regions of immunity, passivity, and corrosion. While an encyclopedia of Pourbaix diagrams is available, the entries do not cover all reactions. Careful extensions and interpretations are necessary.

The rate of reaction can be determined from electrode kinetics. Here mixed potential theory is used. It is assumed that (1) any electrochemical reaction can be divided into two or more oxidation (anodic) and reduction (cathodic) reactions and (2) that there can be no net accumulation of charge. To measure the reaction rate one can monitor either the anodic or cathodic reaction rates. Graphs of solution oxidizing power (in volts) versus corrosion rate (in mils per year, say) can be drawn. On such graphs

.g.

l

O the intersection of lines representing the specific oxidation and reduction reactions of interest will yield an estimate of corrosion rate.

We would like to have developed a computer code that would predict for an arbitrary FSV water ingress scenario the amount of corrosion, if any, on a given part. Such a code would have to take into account the above ideas. Ideally, the code would be able to handle all of the components listed in Secs.1. B. and lil, for any conceivable water ingress event.

Unfortunately, the ingress events that are known to have led to corrosion in FSV have involved a number of different materials, vastly differing geometrical configurations, a wide range of environmental conditions, and at least several differenWpes of corrosion in the presence of helium, not oxygen.

For illustrative purposes to get an idea of how such a code could be developed, we will make some simplifying assumptions. We will consider only general corrosion of smooth, flat surfaces. Now our code would consist of four major sections. These are (1) input, (2) data base, (3) calculations, and (4) output. The input section would require information on material composition and environmental conditions from the user. The data base section would include tabulations of Pourbaix-type data (E vs pH) for each of the materials of interest as well as electrochemical reaction rate data. The calculations section would do interpolations to determine if corrosion would occur and computations of reaction rates when corrosion was expected. Finally, the output section would summarize the results.

If other types of corrosion and other surface configurations were to be considered, they would have to be accounted for and the code would become more complicated. Obviously, the output predictions would only be as good as the input data, data base, and the physics used in the calculations part of the code.

There would be two major weak points in our code that would probably render it nearly useless. The first is that even for smooth flat surfaces we would not be able to find E vs pH data for all of the materials of interestin the presence of helium. This would necessitate an experimental program to obtain the required data for corrosion models to be installed in our code. The second and perhaps even more debilitating weak point is that the environmental conditions are unknown and time varying. They are unknown because local instrumentation to measure

• temperature, pH, and flow rate in the regions of interest simply does not exist. They are transient because the water ingress event itself is a transient, and as soon as the moisture levels reach 500 ppm at 700 psia an automatic reactor scram is initiated with rapid coofdown of the primary coolant. Dryout measures are initiated subsequently to remove the water as quickly as possible. Thus,in order to assess the consequences of a water ingress incident we would have to make conservative engineering estimate of the local environmental conditions. As we will see in Sec. V.,

an error in temperature of 100 F can lead to an error in corrosion rate of two orders of magnitude.

In such circumstances it is customary to perform parameter or sensitivity studies to determine which are the most important or controlling parameters. That is, given a computer code in which one has some confidence, it would be relatively easy to vary individualinput parameters or even change models in the code in a systematic way to determine the effects on the predicted results. Certainly, such numerical exper'iments would help to understand the basic phenomena involved in moisture ingress events.

Now we will review an engineering approximation to the amount of material degradation produced by general corrosion of the PCRV cavity liner 2. This calculation illustrates some of the uncertainties and necessary assumptions specific to the FSV case.

The PCRV liner is constructed from ASTM-A537 grade carbon steel.

Because this is a low alloy steel, SCC is not considered to be a relevant failure mode. (Low alloy steels are not susceptible to SCC unless they contain regions of very high hardness.) Only general and pitting corrosion are viewed as possible concerns.

After a water ingress, we assume that much of the water will condense and collect mostly in the bottom head of of the PCRV liner .32 . It is here that the enhanced general and pitting corrosion rates are expected to be most adverse. Unfortunately, the pH value of the condensed fluid is not easy to datermine:

}

The chlorides may be initially distributed fairly uniformly around the primary coolant j circuit. If the water vapor pressure rises to the point where condensation occurs on chloride

contaminated surfaces, the cond6ns!,ng water wilidissolve the plated-out chlorides and produce pools of chloride containing water, in addition, as the reactor is dried out following a water ingress, the chloride levels in the remaining pools of water will concentrate until finally all the water is removed to leave chloride enriched debris. During subsequent water ingress, 2

dissolution of this debris may form relatively chloride. rich water ,

. However, if the solution in question is initially acidic it is likely that reactions with the metal will quickly dissolve enough corrosion products to move the solution toward neutrality.

Because the coolant in FSV is helium, there is scant practical experience basis to quantitatively define the corrosion rates. Most industrial experience involves corrosion of metals in the presence of oxygen.

Since the PCRV liner is normally at 90--150 F, we might assume the condensed fluid will produce corrosive effects comparable to warm seawater 2. Corrosion rates of carbon steelin an environment of warm, flowing, oxygenated seawater are typically 0.005 to 0.010 in per year.

Worst case pitting rates are 0.015 to 0.020 in per year. Considering the likelihood of static conditions adjacent to the PCRV liner and the low oxygen concentration, these rates are probably quite conservative.

However, to add a further margin of safety, consider a corrosion rate of 0.030 in per year. Further, assume the. wet conditions exist for two months per year for a factor'of 1/6 = 0.167. Under these conditions:

. It is estimated that the maximum loss of effective liner materialis 0.15 in over a period of 30 years. Since the liner is fabricated from 3/4 in-thick plate, it is conduded that breaching of the liner by general corrosion is very unlikely. However, requirements have been established and employed to monitor the liner thickness in the bottom head, sidewall, and top head regions by ultrasonic testing (UT) per Technical Specification SR 5.2.14. The UT readings as of 1981 have shown no decrease in liner thickness. In the unlikely event of a liner breach occurring, radioactive release from the primary coolant will be detected by radiation monitors as described in the FSAR and corrective actions will be taken Hence, the public health and safety l 2 j would not be adversely impacted , <

i V. CORROSION EXPERIMENTS The purpose of this section is to indicate the types of experiments that would be necessary to support model development and confirmation in I

the case of FSV (see Sec. IV.). Since the actual conditions in the PCRV after a moisture ingress incident are variable and largely unknown, such experiments would necessarily encompass a range of parameters to

i i

bracket the actual conditions. It should become clear that such necessary experiments are well beyond the present Los Alamos scope of work. I Experimental techniques are commonly used to assess the corrosive effects of a specific environment on a particular material.4 The weight  ;

loss tests can be broadly classified as in-situ or laboratory simulation l tests. In summary, the idea is to carefully match the environment, select materials, prepare specimens (such as coupons suitably mounted), expose the specimens, and analyze the effects of corrosion (including weight q

loss). The more accurate and complete the simulation, the more valid the test. The data may be practically useless, for example,if complete information on the materials is not known. Surface preparation and exposure technique for the specimen may be important, as can the procedures used in post-test cleaning.

However, the most important single factor in corrosion experiments can be simulating the proper temperatures.6 Corrosion rates for Type 303 stainless steelin nitric acid vary from less than 20 mils per year to several thousand mils per year as a result of increasing the temperature only 100 F. In lab tests temperatures are often controlled to within 2 F.

Other factors such as the presence of air and flow conditions can also be important. Direct or alternating current electrochemical tests in the lab can be much faster than in situ studies; however, correlation of accelerated corrosion test results with real time effects presents further unknowns.

To support the computer code development effort described in Sec. IV.

a number of difficult experiments of the type described above would have to be performed. These experiments would be designed to provide data for the FSV materials and conditions of helium exposure where a literature review revealed that data were missing. The experiments would be difficult because the FSV operating conditions range over helium pressures from 350 psia to 686 psia and temperatures up to 740 F. The experimenter would probably have to bracket the expected FSV conditions in p, T, and pH. Unfortunately, much uncertainty would occur in the utilization of these experimental results because of the unknown actual local temperatures and flow rates and the transient nature of the real problem.

VI. EVALUATION AND RECOMMENDATIONS

A. General The FSV water ingress related problems encountered to date have been very costly in terms of loss of availability of the reactor to produce power and utility clean-up and administration. Unfortunately, some down time and corrosion caused by ingressed moisture will probably continue to plague FSV until the circulator auxiliary system is fixed or replaced it is well to keep in mind that more than 99% of the estimated 9,155 gal of water that has entered the reactor has come from the circulator auxiliary system. i The FSV water ingress related safety issues seem to have emerged and are already being addressed in a conscientious and sensible way by the licensee. The most important safety issues related to the CRDM's, the RSS, and the control rod instrumentation anomalies have been reviewed 2

elsewhere.1-2 Also, a comprehensive evaluation was performed on selected components that are exposed to chloride contamination in the FSV primary coolant. In Ref.1 a program of inspection, refurbishment, and replacement of damaged components using more corrosion resistant materials was recommended. For example, we recommend oeriodic insoection. oreventive maintenance and surveillance. reolacement of Darts for the CRDM. RSS. control rod instrumentation. and any other comoonent that could conceivably have an adverse effect on oublic safety if it were damaaed by corrosion The program to study back EMF for the CRDM's should be continued. In Ref. 2 GAT concluded that the credible chloride-induced failure modes that might conceivably occur would not adversely affect public health and safety and were similar to or less severe than those already considered in the FSAR.

in Secs. IV. and V. we have considered what would be required to develop a computer code that could be used to estimate component degradation in FSV. (Similar considerations also apply to l *back-of-the-envelope" or " cookbook-type" calculations.) We indicated l

that experiments would be needed to supply missing data for use in the code. Also discussed was the idea that even with our code, major uncertainties would remain in our predictions because of the lack of information on the local environmental conditions. The cost of a research program with the goal of providing such an analytical tool along with the experiments and documentation would be many times the existing

contract, in view of this, Los Alamos does not recommend further detailed analvtical or exoerimental studies for comoonent justification. Such studies could be illuminatino and instructive from a research standooint.

btM do not aooear to be necessarv from a safety standooirit. and would be yew costly to do the job oroceriv. However, for safety assurance, some in-plant non-destructive tests are recommended in Sec. VI. E.

B. Ooerational Asoects to Reduce the Frecuency and Severity of Water Ingress incidents (See Ref. 3)

The following measures are recommended to reduce the frequency and severity of water ingress incidents:

1. Since 38% of the 45 water ingress incidents which have been identified can be attributed to improper operation (including maintenance and surveillance testing), a continuino coerational oerformance uogradg orogram is suggested. Such a program should ensure (a) adequate operating procedures that cover previous high risk water ingress occurrences, (b) adequate training for operators in the procedures of (a),

and (c) a human factors review to provide information and procedural assuranco when dealing with abnormal conditions.

2. Since the second leading cause (22%) of incidents was electrical system malfunctions (frorn blown fuses to transformer fires), a crocram to imorove the reliability of the electrical oower distribution system is suggested.

3. To provide information for Sec. VI. B.1. above, a systematic and comorehensivo investigation of all water inoress incidents should be conducled as soon as oossible after each event. A systems study should be conducted to determine cause, sequence of events, preventive measures, and correlation with other occurrences. Such a study could u: cover design, training, or operating philosophy deficiencies.

4. Finally ibe utility should continue to evafuate any maior rnodifications that could lead to imoroved clant availability.

15-4

c, . '

C. Suaaestions to Minimize Corrosion Damaae to Plant Comoonents Assuming that the plant will continue to experience water ingress (hopefully reduced in frequency and severity), some suggestions are made to minimize corrosion damage. These consist of the following:

1. Give attention to the possibility of air injection by the Helium Recovery Compressors during a system upset.

2. Check for rust in, and the operability of alllines, orifices, and check valves serving PCRV penetrations.

3. Finally, investigate and take any necessary action to avoid Helium Purification Train " freeze-up," and thereby insure purge flow to the refueling penetrations. (The utility has committed to providing a backup supply of helium, which would solve this problem.)

D. Water Removal in the event of a water ingress the following operating conditions are recommended to shorten the cleanup time:

1. Raise the PCRV liner temperature from 110 F to somewhere between 120 to 150 F if a!! owed by the Plant Technical Specifications.

2. Operate only the circulator (s) that caused the intiakage. Operate at the maximum speed.

3. Maintain the PCRV pressure at the highest practicallevel to rnaximize permeation flow of helium through the Kaowoolinsulation adjacent to the liner.

4. Provide the highest allowable circulating " cold" helium temperature during dryout.

E. Non-Destructive Tests For safety assurance, some non-destructive tests are recommended:

1. Monitor the effects of corrosion using an ultrasonic test well located near the center of the PCRV bottom head liner on a more frequent schedule.

2. Consider performing pressure tests on the core support floor 16-A

^ __ ___

columns to provide assurance of their structuralintegrity.

3. Periodically inspect, test, and perform preventive or other maintenance on the RSS.

F. Conclusion We conclude that the future credible effects of moisture on FSV components will not adversely affect public health and safety and will be at worst similar to or less severe than those already considered in the FSAR. It is hoped that the rigorous programs of inspection, preventive maintenance or replacement, and improved operational performance will reduce the frequency and severity of water ingress and thereby improve the availability of the reactor and minimize corrosion damage.

i l

a g..

Vll. REFERENCES )

1. D. R. Bennett, G. W. Fly, L. E. Fugelso, R. Reiswig, and S. W. Moore,

" Evaluation of Control Rod Drive Mechanism and Reserve Shutdown System Failures, and PCRV Tendon Degradation issues Prior to Fort St. Vrain Restart, Los Alamos National Laboratory Group O-13 draft report for NRC Fin No. A-7290, March 12,1985.

2. " Evaluation of Fort St.Vrain Metallic Components Exposed to Primary Coolant Chloride Contamination," GA Technologies Document No. 907875, March 21,1985, attached to Public Service Co. of Colorado submittal P 85104.

3. M. Gitterman, Final Report for ASTA under Los Alamos contract 9-X65-M4065-9, Advanced Science and Technology Associates,Inc.

(ASTA),337 S. Cedros Ave., Suite J., Solana Beach, CA 92075, March 12, 1986.

l

4. U. R. Evans, The Corrosion and Oxidation of Metals: Scientific Princioies and Practical Acolications. St. Martin's Press, Inc., New York,1960.

5. M. Pourbaix, Attas of Electrochemical Eauilibria in Aaueous Solutions.

Pergamon Press, London,1966.

6. J. F. Bosich, Corrosion Prevention for Practicina Enaineers. Barnes and Noble, New York,1970.

4 C______._________ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ . _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _