ML25002A153

DPO Case File: DPO-2022-002 - Redacted-Public

ML25002A153
Person / Time
Issue date:	01/02/2025
From:	Annie Ramirez, Andrea Veil Office of Nuclear Reactor Regulation
To:	Laura Smith Office of Nuclear Reactor Regulation
References
DPO-2022-002
Download: ML25002A153 (1)

Text

DPO Case File for DPO-2022-002 The following pdf represents a collection of documents associated with the submittal and disposition of a differing professional opinion (DPO) from an NRC employee involving the Carbon Fiber Reinforced Polymer (CRFP) installation relief request approvals for Surry, ANO, South Texas, and Brunswick.

Management Directive (MD) 10.159, NRC Differing Professional Opinions Program, describes the DPO Program. https://www.nrc.gov/docs/ML2312/ML23123A099.pdf The DPO Program is a formal process that allows employees and NRC contractors to have their differing views on established, mission-related issues considered by the highest-level managers in their organizations, i.e., Office Directors and Regional Administrators. The process also provides managers with an independent, three-person review of the issue (one person chosen by the employee).

Because the disposition of a DPO represents a multi-step process, readers should view the records as a collection. In other words, reading a document in isolation will not provide the correct context for how this issue was reviewed and considered by the NRC.

It is important to note that the DPO submittal includes the personal opinions, views, and concerns of an NRC employee. The NRCs evaluation of the concerns and the NRCs final position are included in the DPO Decision.

The records in this collection have been reviewed and approved for public dissemination.

Document 1: DPO Submittal Document 2: Memo Establishing DPO Panel Document 3: DPO Panel Report Document 4: DPO Decision

Document 1: DPO Submittal

The ASME Boiler and Pressure Vessel Code Section XI, Article IWA-4221(b) specifies that an item to be used for repair/replacement activities shall meet the Construction Code. Subarticle IWA-4221(b)(1) specifies that when replacing an existing item, the new item shall meet the Construction Code to which the original item was constructed. However, the Construction Code does not include any criteria or requirements applicable to fiber reinforced polymers. Licensees may request relief from the ASME Code under 10 CFR 50.55a(z)(1). This states that the licensee must demonstrate that the proposed alternative would provide an acceptable level of safety. Pursuant to this regulation, four licensees submitted relief requests to use fiber reinforced polymer materials to repair degraded safety-related piping.

When NRC Staff perform technical reviews of requested licensing actions (RLA), NRR Office Instruction LIC-109 first requires that an acceptance review be performed to ensure that sufficient information is available in the request. Contrary to the guidance in LIC-109, the submitted RLAs do not contain sufficient accurate technical information, material properties, and mechanical behavior information about how the material will respond under elevated temperature conditions. Some of the information presented in the RLAs and in the responses to NRC Staff requests for information is presented without adequate or rigorous technical justification or relies on unapproved guidance. See the enclosed technical paper for specific details on the specific information insufficiencies in the relief requests.

Degradation of installed fiber reinforced polymer repairs and components presents in several forms. Erosion from raw water, mechanical damage from impacts, and delamination and debonding from water intrusion or environmental temperatures above the glass transition temperature will not be detected through operational leakage monitoring, reactor coolant system technical specifications, or scheduled surveillances. In addition, inservice inspection of the carbon fiber reinforced polymer repairs will not adequately monitor degradation because the techniques for inservice inspection of fiber reinforced polymer repairs have not yet been adequately developed or shown to be effective.

Due to these reasons, I have a differing professional opinion on the approvals issued to Surry, Arkansas Nuclear One, South Texas Project, and Brunswick to install fiber reinforced polymers in their safety-related Class 3 piping systems and on the safety evaluations written to support the approvals.

1 Summary This differing professional opinion regards the approval to use carbon fiber reinforced polymer (CFRP) materials to repair safety-related piping. The specific documents that this DPO addresses are the safety evaluations for Surry Power Station Units 1 and 2, South Texas Project Units 1 and 2, Arkansas Nuclear One Units 1 and 2, and Brunswick Steam Electric Plant Units 1 and 2. In these applications, the licensees proposed to line degraded piping with CFRP such that the new CFRP pipe would assume all the structural functions of the metallic piping.

This is a considerable departure from metallic piping repairs where only the damaged portion of pipe may be repaired or replaced. The proposed CFRP installations would replace entire lengths of pipe that may be long.

CFRP is a novel and innovative material that has great potential and application in the nuclear power industry. It is much lighter than metallic materials that have similar tensile strengths, and it can be molded to fit many safety-related critical components in a nuclear power plant. In a manner analogous to rebar-reinforced concrete, CFRP consists of bundles of carbon fiber embedded in a polymer matrix. The carbon fibers provide the tensile load-carrying capacity while the polymer matrix provides the structural strength and stiffness and transfers the load between the host material and the carbon fiber. Thus, the choice of polymer matrix is a critical variable for use of CFRP material.

For the CFRP to reliably perform all the functions of the safety-related piping at all service levels, the selected polymer must withstand all applicable environmental conditions. Service level D represents the most extreme environmental conditions including the maximum temperature expected under design basis accidents. At elevated temperatures, the polymer matrix used in CFRP will soften and become a ductile, rubbery material. The key property that describes the change in material state is the glass transition temperature (Tg): the temperature at which the polymer goes from a brittle, glassy material (elastic deformation) to a ductile, rubbery material (plastic deformation). The Tg is dependent on many factors, including the chemistry of the polymer, the appropriate mixing ratio of the component parts, the saturation of the carbon fiber fabric, the handling of the saturated material in the pipe, and the curing temperature. Based on these factors, the actual Tg obtained during fabrication of the CFRP repairs in each installation may vary from theoretical Tg values. When the temperature of the system approaches Tg, the polymer will soften and may debond from the underlying metallic substrate, the layers may delaminate, or the material may deform in a variety of ways. If the service temperature approaches Tg, the CFRP material would likely peel off the pipe or delaminate and be carried with the flow of water, creating foreign material in the system and fouling downstream components. The fouled components will then be prevented from performing their safety-related function and contribute to an increase in core damage frequency.

To prevent the described failures and consequences from happening, a suitable margin between the maximum design temperature and the actual Tg of the material must be maintained. In this DPO, I will show that none of the submitted and approved relief requests offer adequate assurance that the Tg post-installation will have achieved a sufficient margin above the maximum design temperature, and that the risk associated with the potential failure is unacceptably high.

Some aspects of the relief requests and supplemental information provided by the licensees were inappropriately credited and/or inadequately considered. These aspects include but are not limited to the proposed cure temperature adjustment factor, using a higher cure temperature at the terminal ends, an inappropriate cure acceptance criterion, crediting accident condition temperatures to post-cure the material at Arkansas Nuclear One, and, for South Texas Project, lowering the maximum design temperature for the piping. Each of these factors will be discussed below.

2 Overview Four licensees have requested approval to repair degraded metallic safety-related piping with CFRP. The ASME Boiler and Pressure Vessel Code does not address using fiber reinforced polymers in either Section III or Section XI. As a result, the licensees have submitted requests for relief from the Code to use CFRP to repair the piping. The licensees are Surry, South Texas Project, Arkansas Nuclear One, and Brunswick.

In my professional opinion, the requests do not provide an adequate assurance of safety due to several issues, including:

Inadequate understanding of the thermoset polymer matrix behavior at elevated temperatures Inadequate controls for field installation conditions An unproven concept called cure temperature adjustment factors Varying cure temperatures between the terminal ends and the length of the repair resulting in varying Tg along the entire length of the repair Inaccurate acceptance criterion for cure determination Without adequate field installation controls and a more complete understanding of how thermoset polymers used in fiber reinforced polymer applications are cured and perform in all service conditions, there is no assurance of safety.

3 Glossary Degree of Cure A measure of how many crosslinking sites available at a specific temperature have been utilized Glass Transition The temperature range where the material behavior of the thermoset resin transitions from linear elastic (hard and brittle) to plastic (soft and rubbery) deformation Modulus The ratio between stress and strain used to define elastic and plastic deformation: a constant modulus value means the material is in the elastic deformation zone, while a changing modulus indicates plastic deformation

4 Technical Issues The following issues have considerable impact on determining a reasonable assurance of safety in using CFRP to repair ASME Class 2 and 3 safety related piping.

4.1 Degree of Cure All four safety evaluations accept the use of degree of cure to determine the acceptability of the final installation. Degree of cure is a measure of how many available crosslinking sites have been utilized at a specified cure temperature. When a thermoset polymer is curing, the individual polymer strands begin to link together and form a complex 3-D mesh network. The number of available crosslinking sites increases with the cure temperature, thus producing a more complex and robust mesh when cured at increased temperatures. A polymer, cured at 140°F, with 85% degree of cure will have better mechanical properties than a polymer cured at 100°F and 85% degree of cure.

Using degree of cure as an acceptance criterion does not ensure that the desired mechanical properties have been obtained, only that the mesh network has used 85% of the available crosslinking sites available at the specific cure temperature used in the fabrication of the sample.

4.2 Glass Transition Temperature The glass transition temperature (Tg) is a reliable predictor of mechanical behavior from ambient temperatures to the beginning of the glass transition range. Below the Tg, the polymer is brittle and glassy and efficiently transfers stress. In the glass transition range, the polymer changes from the brittle, glassy material to a ductile, rubbery one. The rubbery polymer plastically deforms instead of transferring the stress between the host pipe and the carbon fiber bundles.

The Tg is a single temperature that can be defined as either the midpoint of the glass transition range or as the extrapolated-onset temperature. The extrapolated-onset temperature is the temperature at which the transition from brittle to ductile begins. Defining Tg in CFRP as the extrapolated-onset temperature is the conservative approach.

4.2.1 Ultimate Tg vs Field Tg The submitted relief requests for South Texas Project, Arkansas Nuclear One, and Brunswick make a distinction between the ultimate Tg and the field Tg. They define the ultimate Tg as the maximum Tg that can be obtained for a thermoset polymer under lab-controlled conditions and high temperature curing. The lab controlled Tg is significantly higher than the field Tg; the Tg that can be obtained during installation in field conditions where the environment is essentially uncontrolled and only moderately predictable. For the polymer proposed in the relief requests, the ultimate Tg is approximately

, while the field Tg is approximately

. Qualification of the polymer based on the ultimate Tg is inadequate since it is a result that cannot be replicated in field conditions.

4.2.2 Tg Acceptance Criteria Three of the submitted relief requests address the difference between the ultimate Tg and the field Tg by adding a 10°F margin above the cure temperature. Adding a margin to the Tg to ensure that the thermoset resin remains in the linear-elastic regime is an appropriate approach to ensure an adequate assurance of safety. However, 10°F is insufficient because the two ASTM methods (E1640 and E1356) used to determine the Tg have a repeatability range of approximately +/-8°F; the difference that can be expected when running additional tests using the equipment in the same lab. Using a margin of Tg + 10°F leaves an effective margin of only 2°F when the test is off by a negative 8°F. Two degrees is an inadequate margin for an assurance of safety.

4.2.3 Elevated Tg at Terminal Ends Terminal ends are the locations of the host pipe that provides the structural foundation for the composite repair, all loads from the substrate pipe transfer to the laminate at the terminal ends.

Three of the safety evaluations accept the use of a higher cure temperature at the terminal ends than the one used along the length of the pipe to be repaired. Using a higher cure temperature at the terminal ends will result in a higher Tg at the terminal ends than the Tg of the repair between the terminal ends. This approach is inadequate because it will lead to the potential result that the terminal ends will remain in the linear-elastic region while portions of the installed repair may undergo plastic deformation with a corresponding failure to adequately transfer stress. It could also result in the middle lengths of the repair debonding from the host pipe or delaminating between layers and only the terminal ends remaining in the glassy, brittle condition. Additionally, a difference between the Tg at the terminal ends and the Tg in the middle lengths could result in unanalyzed thermal stress at the interface between Tg regions.

4.3 Witness Panels Witness panels are small, one-layer panels made at the same time and from the same materials as the CFRP repair layer. The theory is that by making a small sample at the same time the CFRP layer is installed, the sample will have the same mechanical and material properties as the installed CFRP layer.

One issue with the witness panel approach is that in order for the panel to have similar mechanical and material properties as the installed layer, the panel needs to be cured under similar environmental conditions. This approach is impractical at best, since the environmental conditions will vary along the length of the repair.

A second issue with the witness panel approach is the exothermic nature of the crosslinking reaction in the uncured resin. As the uncured, liquid resin sits, the temperature of the liquid will rise. A witness panel created from liquid resin that has aged will have a higher Tg than a witness panel created from liquid resin that is freshly mixed. Testing a witness panel that was created using aged liquid resin instead of freshly mixed liquid resin would give mechanical and material properties that are not representative of the installed repair.

4.4 Cure Effect Factors (Cure Temperature Adjustment Factors)

Cure effect factors are a ratio between a material property at elevated temperature and the same property at ambient temperature. The calculated ratio is then used to adjust a material property including tensile strength, elastic modulus, and lap shear strength. The Arkansas Nuclear One, Brunswick, and South Texas Project safety evaluations accept the use of cure effect factors to adjust tensile strength and modulus. There are several problems with the cure effect factor concept.

There is no valid technical basis for adjusting the strength or modulus of a thermoset polymer by a factor and getting a reliable prediction of elevated temperature performance.

The elastic modulus (The slope of the stress-strain curve between 0 and the yield strength) must remain constant to ensure the thermoset resin remains in the linear-elastic deformation zone. Once the elastic modulus changes, the material is no longer in the elastic deformation zone. By modifying the elastic modulus by a cure effect factor, the designers are allowing the thermoset polymer matrix to deform. This is contrary to maintaining the CFRP in the hard, brittle state necessary to maintaining the structural strength.

No supporting technical justification or equation is provided for the cure effect factor numbers included in the South Texas Project safety evaluation in either the initial relief request or in the supplemental letters dated July 15, 2020, and July 30, 2020.

The equation supplied for calculating the cure effect factors in the Arkansas Nuclear One and Brunswick relief requests is not explained, and its source is not cited.

4.5 Failure Modes The failure mode analysis in the submitted relief requests does not adequately consider deformation, delamination, and debonding of the installed CFRP due to elevated temperature (e.g., high summer temperatures of cooling pond water, abnormal operating temperatures, service level D temperature, etc.). The stated reason for discounting these failure modes is that the selection of the material is based on the ultimate Tg minus 40°F, which will not be sufficient to ensure these failure modes are prevented. This rationale neglects to consider that the ultimate Tg obtained under laboratory-controlled conditions is significantly higher than the Tg that is attainable in the field.

While in the soft, rubbery state, debonding or delamination would produce foreign material in the safety system, which would then foul downstream components including pumps, strainers, and valves. Once the temperature of the repaired system decreases below the Tg, the delaminated or debonded material will return to the glassy state. The solidified state of the material will then complicate removal, as it will have taken the shape of whatever component it has become entangled in. An example of the debonding and delaminating failure modes can be seen in the condition report from Arkansas Nuclear One, dated October 12, 2021, attached to this DPO. In the cited condition report, CFRP was used to repair a corroded circulating water return pipe. A section of CFRP laminate debonded from the pipe and moved downstream. It was found to have fouled the screen, increasing back pressure in the pipe. Some pieces were able to be

removed manually, but one piece was too large to be removed via the manway and needed to be broken apart first.

4.6 Post-Cure During Service Level D The safety evaluation for Arkansas Nuclear One Units 1 and 2 accepts the idea that the installed repair will post-cure as the temperature rises during faulted service. Post-cure completion during accident conditions is an inadequate approach because the rise in temperature will not be controlled and cannot be guaranteed to produce the required Tg in the middle of the transient. In addition, the technique assumes that a complete cure can be obtained at any time after fabrication. It is well understood and accepted in the composite industry that the window to complete the curing process closes after a period of time. After this time has elapsed, no additional curing will occur. The resulting Tg will be unknown and potentially too low to prevent failure of the CFRP repair. Accepting a low Tg under the hope that it will finish curing during an accident does not provide an adequate assurance of safety.

4.7 Lowering Maximum Design Temperature The safety evaluation for South Texas Project accepts the proposed reduction of the maximum design temperature for the specified safety system from 150°F to 134°F without any justification or evidentiary calculations. The submitted relief request states that the temperature change will be handled as an engineering change; however, that engineering change paperwork was not included with the relief request. Accepting a reduction in the maximum design temperature without justification is inappropriate and non-conservative.

5 Operating Experience 5.1 Arkansas Nuclear One In October 2021, an installed CFRP repair was found in a degraded condition. A section of the repair debonded from the pipe and moved downstream. It became entangled in the screen, causing back pressure to build in the pipe. Some sections of the topcoat were also missing, meaning foreign material entered the piping system unmonitored. The debonding failure described in this condition report is consistent with a failure due to exposure to temperatures above the Tg.

Some of the material entangled in the screen was able to be removed easily; however, one piece that was too large to be taken through the manway had to be broken up before removal.

5.2 Three Mile Island In January 2014, an endcap from a distribution header in a natural draft cooling tower failed.

The endcap was crafted from a fiber reinforced polymer and should have had a glass layer to function as a corrosion barrier. The corrosion barrier was missing, and the resulting water intrusion caused the endcap to fail.

The root cause of the failure was poor control of the fabrication process. Contributing causes to the failure include over-reliance on contractor knowledge and expertise by plant personnel.

5.3 Susquehanna In December 2009, the bypass make-up line in the Unit 2 cooling tower failed. Prior to this event, the line was replaced with a fiber reinforced pipe. The failure occurred due to internal erosion in the pipe near an isolation valve.

The failure mechanism in this case was abrasive material in the raw water supply. The licensee noted in their failure analysis that licensee personnel must understand the failure mechanisms and inspection limitations of fiber reinforced polymers.

6 Risk to Safety Systems The technical issues noted above make the risk of a catastrophic failure of the installed CFRP repair a significant concern. When the CFRP fails through debonding or delamination, pieces of this foreign material will flow downstream with the water and will foul critical components like screens, pumps, and valves.

For the purposes of this DPO, I looked at the Risk Increase Ratio (RIR) to core damage frequency (CDF) and the Birnbaum number of removing the components immediately downstream from the installation locations. I reviewed the SPAR reports for each plant in order to determine this information. For any of the proposed installations, a debonding failure of the CFRP would be a common cause failure since the licensees have proposed installing it into piping header systems or common intake and outflow piping. I considered the immediate downstream components failed because the CFRP material would foul the component, preventing the performance of its safety function.

The RIR is a measure of the relative risk increase that is attributable to the specified components failing. A number greater than one indicates the risk is increasing. The higher the number, the more significant that component is to the safety of the plant.

The Birnbaum number measures of the rate of change in total risk as a result of changes to the probability of an individual basic event. In this case, it measures the rate of change in total risk if the specific components are failed. A larger number indicates a faster rate of change, meaning that the risk is increasing at a rapid rate. Smaller numbers indicate a slower rate of change.

6.1 Surry Nuclear The safety evaluation approves the use of CFRP to repair ASME Class 3 piping including the circulating water (CW) system nd service water (SW) pipe headers from the CW system The inlet piping is gravity fed Surry is installing the CFRP repair in each inlet pipe

. Each inlet pipe has several motor operated butterfly valves to control the flow of water into the service water system.

Because the CFRP is installed in all inlet piping, debonding or delamination of the layers would be a common cause failure that could affect all these valves, rendering them inoperable. The inoperable valves would prevent the service water system from performing its safety function.

6.3 Arkansas Nuclear One The safety evaluation approves the use of CFRP to repair the emergency cooling pond supply piping to the service water system. During a design basis accident or when the main ultimate heat sink is unavailable, the alternate heat sink for the service water system is the emergency cooling pond. The emergency cooling pond piping transports water from the emergency cooling pond into the basin that is the supply for the service water system. Since the specific pipe locations being repaired with CFRP are not provided in the submitted relief requests, I assume that it is the entire length of the pipe. The immediate downstream components are the service water screens and service water pumps. Because the piping to be repaired supplies all service water trains, it is a common cause failure mechanism.

The risk increase ratio and Birnbaum values are included to show the effect of a common cause failure on the core damage frequency for the service water pumps and screens. If these components are taken out of service, then all the downstream components will be unable to perform their safety function, including but not limited to the emergency feedwater system, the aux building decay heat removal coolers, and the diesel generator water jacket heat exchangers.

Component Description RIR Birnbaum 6.4 Brunswick The safety evaluation approves the use of CFRP to repair the conventional and nuclear service water headers for both units The nuclear service water header supplies the reactor building closed cooling water heat exchangers, the RHR "B" heat exchanger, all RHR pump seal cooling heat exchangers, RHR service water (RHRSW) pumps "B" and "D" and their motor cooling, and standby diesel generator cooling. The service water conventional header, when needed, supplies the RHR "A" heat exchanger and the RHRSW pumps "A" and "C" and their motor cooling.

Since the entire header pipe is being repaired with CFRP, the failure of CFRP would be a common cause failure for the downstream components and systems. The risk increase ratio and Birnbaum values are included to show the effect of a common cause failure on the core damage frequency Component Description RIR Birnbaum

6.5 Conclusion The consequences of CFRP pose a substantial risk to the ability of a nuclear power plant to remain cooled during a design basis accident. The licensees plan to install CFRP in piping that serves as a common header or supply line to a variety of safety-significant systems, all of which are at-risk when the CFRP fails from exposure to elevated temperatures.

The installation of CFRP into ASME Class safety related piping requires careful scrutiny and analysis of the downstream systems and components that will be compromised and unable to perform their safety function.

7 Proposed Resolution To date, a CFRP repair has only been installed at Surry Nuclear. The other three approved relief requests have yet to be installed.

The proposed resolution for Surry is to do additional testing on the installed repair to fully understand the mechanical properties and glass transition temperature of the installed CFRP.

The confirmatory mechanical testing should be done at the maximum design temperature of 105°F. The glass transition temperature should be determined using differential scanning calorimetry or dynamic mechanical analysis testing for the length of the repair, at the beginning, end, and a significant number of points in between. The results of the Tg testing will reveal the actual condition of the installed repair and give a reasonable prediction of performance under design basis accident conditions.

The proposed resolution for licensees who have not yet installed the CFRP is to carefully review the issued safety evaluations and the approved relief to determine the appropriate course forward, including revoking the issued approval until additional assurances of adequate safety are received.

Additional remedies to address the identified deficiencies in the safety evaluations include, but are not limited to:

Clear and rigorous justifications for the credited equations, factors, and any changes to accepted plant parameters.

Explicit selection of a thermoset resin that can produce an adequate Tg under ambient cure conditions with no post-cure exposure required to obtain the required Tg.

Detailed installation procedures specific to each licensee and plant that account for environmental factors, personnel performance, and the exothermic nature of the mixed resin prior to application.

Specific direction when making witness panels that each panel shall be made at the beginning of the shift, with fresh-mixed resin, and that no additional resin may be applied after the panel is crafted.

Ensure that sufficient witness panels are made and exposed to the same cure environment as the section of pipe with the lowest cure temperature exposure.

Independent verification of the vendor claims and processes by the licensee.

Remedies for NRC Staff include but are not limited to:

Creation of a shared resource for NRC Staff on fiber reinforced polymers and polymeric materials in general, including material and mechanical properties.

Consultation with subject matter experts in polymers when determining the acceptability of proposed requested licensing actions.

Fiber reinforced polymer technologies in general and CFRP specifically have a place in nuclear power plants; however, much more needs to be understood before they can be approved for use in safety-related systems.

8 References Surry Power Station, Unit Nos. 1 and 2 - Relief from the Requirements of the ASME Code, ML17303A037, December 20, 2017 South Texas Project, Units 1 and 2 - Proposed Alternative RR-Eng-3-24 To ASME Code Requirements for The Repair of Essential Cooling Water System Class 3 Buried Piping, ML20227A383 September 3, 2020 STP Response to Request for Additional Information Proposed Alternative to ASME Section XI Requirements for Repair/Replacement of Essential Cooling Water Class 3 Buried Piping in Accordance With 10 CFR 50.55a(Z)(1), ML20197A262, July 15, 2020 STP Response to Request for Additional Information Proposed Alternative to ASME Section XI Requirements for Repair/Replacement of Essential Cooling Water Class 3 Buried Piping in Accordance With 10 CFR 50.55a(Z)(1), ML20212L569, July 30, 2020 Arkansas Nuclear One, Units 1 and 2 - Approval of Request for Alternative from Certain Requirements of The American Society of Mechanical Engineers Boiler and Pressure Vessel Code, ML21265A255, September 20, 2021 ANO Response to Request for Additional Information (Follow-up) - Proposed Alternative to ASME Section XI Requirements for Repair/Replacement of Emergency Cooling Pond Supply Piping, ML21203A198, July 22, 2021 ANO Response to Request for Additional Information (Follow-up) - Proposed Alternative to ASME Section XI Requirements for Repair/Replacement of Emergency Cooling Pond Supply Piping Enclosure 1, ML21203A199, July 22, 2021 Brunswick Steam Electric Plant, Units 1 and 2 - Authorization and Safety Evaluation for Alternative from Certain Requirements of the American Society of Mechanical Engineers Boiler and Pressure Vessel Code of Buried Service Water Piping, ML21343A197, December 20, 2021

Brunswick Response to Request for Additional Information Regarding Proposed Alternative to ASME Boiler & Pressure Vessel Code Section XI Requirements for Repair/Replacement of Service Water (SW) System Buried Piping in Accordance with 10 CFR 50.55a(z)(1),

ML21277A306, October 4, 2021 Brunswick Response to Request for Additional Information Regarding Proposed Alternative to ASME Boiler & Pressure Vessel Code Section XI Requirements for Repair/Replacement of Service Water (SW) System Buried Piping in Accordance with 10 CFR 50.55a(z)(1) Enclosure 1, ML21277A307, October 4, 2021 Arkansas Nuclear One Unit 2 Condition Report CR-ANO-2-2021-02472, 10/12/2021 IRIS Experience Report #309745 Three Mile Island Unit 1 Cooling Tower Water Distribution Piping Failure Causes Plant Derate 2/17/2014 IRIS Experience Report #240613 Susquehanna Unit 2 Forced Down Power >20% 12/1/2009 ASTM E1356-8 Standard Test Method for Assignment of the Glass Transition Temperatures by Differential Scanning Calorimetry ASTM E1640-18 Standard Test Method for Assignment of the Glass Transition Temperature by Dynamic Mechanical Analysis Standardized Plant Analysis Risk Model for Arkansas Nuclear One Unit 1, April, 2020 Standardized Plant Analysis Risk Model for Arkansas Nuclear One Unit 2, March, 2019 Standardized Plant Analysis Risk Model for Brunswick 1, April, 2019 Standardized Plant Analysis Risk Model for Brunswick 2, September, 2020 Standardized Plant Analysis Risk Model for South Texas Project 1 & 2, May, 2019 Standardized Plant Analysis Risk Model for Surry 1, March, 2019

Document 2: Memo Establishing DPO Panel

Document 3: DPO Panel Report

MEMORANDUM TO:

Andrea Veil, Director Office of Nuclear Reactor Regulations FROM:

Robert L. Tregoning, DPO Panel Chair Office of Nuclear Regulatory Research Stephen Cumblidge, DPO Panel Member Office of Nuclear Reactor Regulations James Drake, DPO Panel Member Region IV

SUBJECT:

DIFFERING PROFESSIONAL OPINION PANEL REPORT ON INAPPROPRIATE APPROVAL OF CARBON FIBER REINFORCED POLYMER TO REPAIR DEGRADED CLASS 3 SAFETY-RELATED PIPING (DPO-2022-002)

In a memorandum dated January 12, 2023, we were appointed as members of a Differing Professional Opinion (DPO) Ad Hoc Review Panel (DPO Panel) to review a DPO regarding the approval of carbon fiber reinforced polymer to repair degraded Class 3 safety-related piping. The DPO Panel has reviewed the DPO in accordance with the guidance in Management Directive 10.159, The NRC Differing Professional Opinion Program.

The Panels review scope was limited to the issues identified in the DPO as clarified through a Summary of Issues developed by the Panel and confirmed by the DPO submitter. The Panel performed its review by collecting and reviewing documents, and conducting interviews with knowledgeable NRC staff, including the submitter. The Panel also oversaw risk evaluations performed by Office of Nuclear Regulatory Research staff in support of the Panels evaluation.

The results of the Panels evaluation of the concerns raised in the DPO are detailed in the enclosed DPO Panel Report. In summary, while the Panel believes that use of carbon fiber reinforced polymer (CFRP) is a viable repair method that holds promise for addressing large-scale or localized degradation in difficult-to-replace piping systems, there is merit in several of the concerns raised by the DPO, particularly pertaining to the following issues:

CONTACT: Robert L. Tregoning, RES/DE 301-415-2324 The enclosure to this letter contains sensitive unclassified non-safeguards information.

When separated from the enclosure, this document is DECONTROLLED.

March 4, 2024 Signed by Tregoning, Rober on 03/04/24

A. Veil 2

The initial CFRP approvals are based on temperature and stress margins that are lower than have been historically applied for commensurate metallic materials.

Common-cause failure of CFRP repairs within service water systems is a particular concern. This concern is buttressed by some very limited operating experience for use of CFRP in similar nuclear applications.

Significant uncertainty exists about the acceptability of adjustment factors to address cure temperature, environmental, and long-term degradation.

Quality assurance provisions associated with the fabrication, installation, and verification testing for the CFRP systems are complex, such that a more holistic evaluation of the quality assurance program is warranted.

The potential for performance variability within the repair is significant given both the breadth of the repair and the myriad configurations, installation orientations, and accessibility challenges that may be associated with the repair.

Particular focus is needed to ensure the representativeness and applicability of verification testing and the efficacy of further in-service evaluation.

Because risk impacts of CFRP can vary by several orders of magnitude, more rigorous risk evaluations should be explored to support both the currently approved CFRP applications and future staff evaluations.

More detailed conclusions and associated recommendations are contained in the Panel report to address several issues raised within the DPO. Recommendations are associated with either enhancing the technical adequacy of approved CFRP repairs, ensuring adequate installation procedures and quality assurance measures, or better understanding the risk associated with postulated CFRP repair failures.

It should also be noted that during its review, the Panel did not identify, nor was made aware of any significant safety issues that would require immediate regulatory action.

Please do not hesitate to contact us if you have any questions regarding the enclosed report.

Enclosure:

1. DPO Panel Report

)

ML24057A029; Memo ML24057A030 OFFICE R-IV/DRS/EB2 NRR/DNRL/NPHP RES/DE NAME JDrake SCumblidge RTregoning DATE Mar 1, 2024 Mar 1, 2024 Mar 4, 2024

Enclosure:

Differing Professional Opinion (DPO) on Inappropriate Approval of Carbon Fiber Reinforced Polymer to Repair Degraded Class 3 Safety-Related Piping (DPO-2022-002)

DPO Panel Report

______________________/RA/______________________

Robert L. Tregoning, Panel Chair

______________________/RA/______________________

Stephen Cumblidge, Panel Member

/RA/

James Drake, Panel Member March 4, 2024 This document contains proprietary information pursuant to Title 10 of Code of Federal Regulations, Section 2.390.

Proprietary information is identified by bold text enclosed within double brackets, as shown here: ((example proprietary text)).

Introduction On September 6, 2022, a U.S. Nuclear Regulatory Commission (NRC) staff member filed a Differing Professional Opinion (DPO) in accordance with NRC Management Directive 10.159, The NRC Differing Professional Opinions Program. The DPO (ML23011A255) involves the staffs approval for use of a carbon fiber reinforced polymer (CFRP) composite for repairing degraded American Society of Mechanical Engineers (ASME) Class 3 safety-related piping at the Surry, Brunswick, South Texas Project (STP), and Arkansas Nuclear One (ANO) nuclear power plants.

These licensees requested NRC approval of this repair technique in separate requests for relief from the requirement that repair/replacement activities shall meet the relevant ASME Construction Code. Because the Construction Code does not contain applicable requirements for use of CFRP, the proposed alternatives were submitted in accordance with 10 CFR 50.55a(z)(1) which requires that the proposed alternative repair provides an acceptable level of safety. The DPO submitter contends that these proposed alternatives do not provide sufficient technical justification to demonstrate an acceptable level of safety for the proposed CFRP composite systems. The submitter further contends that the risk associated with the potential failure of the CFRP composite systems is unacceptably high.

The NRCs Office of Enforcement screened and accepted the DPO on November 14, 2022, and assigned the following DPO case number: DPO-2022-002. By memorandum dated January 12, 2023 (ML23011A124), the Office of Enforcement established an Ad Hoc Review Panel (the Panel) to perform an independent review of the DPO, document the conclusions in a report, and provide any associated recommendations, if necessary. Panel responsibilities and guidance are stipulated in MD 10.159. The memorandum also articulated the DPO process milestones and timeliness goals (ML23011A252) and stressed the importance of maintaining the submitters anonymity.

The Panel developed a draft Summary of Issues (SOI) and shared it with the submitter on March 9, 2023. Based on subsequent discussions and feedback from the submitter, the DPO Panel and submitter agreed to a final SOI. The final SOI is documented in the next section.

Subsequently, the Panel performed its review by collecting and reviewing documents and conducting interviews with knowledgeable NRC staff. The Panel also oversaw risk evaluations performed by staff in RES\\DRA to support the Panels evaluation. Subsequent sections in this report describe the Summary of Issues, the Panel evaluation, Panel conclusions, and associated recommendations. A list of the documents reviewed, and the NRC staff interviewed are also provided in Appendices A and B, respectively. During its review, the Panel did not identify, nor was made aware of, any significant safety issues that would require immediate regulatory action.

Summary of Issues (SOI)

Based on a review of the DPO submittal and associated references, and after discussion with the submitter, the Panel determined that the submitters individual concerns (i.e., sub-issues)

could be grouped within the following distinct topical areas: 1) the adequacy of the technical basis supporting the CFRPs acceptable use in these piping systems, 2) installation procedures and quality assurance methods, and 3) risk of CFRP failure. This categorization allowed the Panel to focus on the submitters underlying concerns within each of these broad areas and perform an integrated evaluation in lieu of adjudicating each individual sub-issue within the submittal. Therefore, while all sub-issues have been reviewed and contribute to the reports holistic evaluation, the omission or inclusion of specific sub-issues in the report should not be construed as the Panels opinion of their validity, unless otherwise stated in this report.

The final Summary of Issues (SOI), as agreed upon by both the Panel and the submitter, are the following three principal concerns:

1.

An inadequate technical basis has been provided by licensees within the subject proposed alternatives and associated requests for additional information to demonstrate that the proposed CFRP repairs have a reasonable assurance of acceptable performance under both expected service and design-basis conditions.

2.

The licensees, in the subject proposed alternatives and associated requests for additional information, have not demonstrated that adequate installation procedures and quality assurance methods have been established to provide reasonable assurance that the installed CFRP repairs will comply with the promised material properties and performance measures.

3.

There is significant risk associated with CFRP failure within the systems approved for this repair method, especially when the potential for common-cause failures is considered.

Each of these concerns is separately addressed in the Panels evaluation. However, it is recognized that these issues are interconnected, and this connectivity is factored into the evaluation, conclusions, and recommendations that follow.

Evaluation

Background

The Panel recognizes that service water1 systems (SWSs) at some plants have experienced significant degradation, and that appropriate repair or replacement strategies are needed when degradation exceeds acceptable ASME Code design limits. Additionally, CFRP has the potential to be an acceptable repair technique that is strong, environmentally resistant, adaptable to a variety of piping locations and configurations, cost-effective, and except for the termination ends is fully independent of the original piping. This last characteristic is particularly attractive as initial proposed alternatives (PAs) have not credited the margin provided by the original pipe 1 This report uses the term service water systems to refer to piping systems that provide cooling to component cooling water plant systems (e.g., CVCS letdown heat exchanger, RHR heat exchangers). Some plants refer to service water systems using other terms such as essential cooling water systems in their FSARs.

between the CFRP termination points. Therefore, within the repair region, the original pipe only serves as a substrate during the CFRP curing process. After that, it can continue to degrade with no impact on system design margins presuming the degradation does not significantly affect the CFRP strength at either the termination points or along the remaining length of the repair, and the original pipe material thickness remains within acceptable limits in non-CFRP repaired regions.

The biggest challenge in assessing the adequacy of the CFRP technical basis is the paucity of operating experience pertaining to the use of these materials in predominantly buried nuclear power plant (NPP) SWSs. From that perspective, the subject PAs represented a first-of-a-kind application for implementing CFRP repairs. CFRP composite materials are also fundamentally different than the structural metals that comprise the bulk of NPP piping, and specifically the carbon steel and aluminum-bronze piping materials that they would be repairing in the subject PAs.

The CFRP systems used in the PAs are comprised of a number of layers, or plies. A CFRP ply is a two-part composite consisting of a unidirectional carbon fiber encased in a thermoset polymer epoxy matrix. The carbon fiber provides the tensile strength while the epoxy matrix provides the bond strength; compressive strength and buckling resistance; protection and environmental resistance; and transfers load to the carbon fibers. Each ply of the composite is manufactured on-site and the CFRP repair system consists of several plies that are hand-laid in alternating orthogonal orientations to provide sufficient hoop and axial properties. The initial layer which bonds to the metallic pipe is a glass-fiber reinforced polymer (GFRP) that serves as a dielectric layer to protect the metallic pipe from galvanic corrosion. GFRP also provides the final, watertight layer within the system. The systems approved in the PAs consist of between 5 to 7 GFRP and CFRP plies. The system is bonded to the original piping at the terminal ends and also secured using a monomer rubber seal and a stainless-steel retaining band. The terminal ends are obviously important for ensuring that the CFRP system fulfills its intended function.

Classic nuclear metallic piping materials are strongly leak-before-break tolerant. There are numerous examples throughout the operating history when this attribute has been integral in avoiding failure due to unknowns and uncertainties not considered in the initial design.

Conversely, the CFRP system is an inherently brittle composite which means that there is little resistance to crack propagation if design margins are insufficient and an existing defect or partial delamination are present. Its nature also makes it potentially susceptible to failure mechanisms not typically seen in metals such as ablation (which is a potential consideration in SWSs that intake water from an open, unfiltered source), delamination, debonding from the metal substrate, and excessive deformation at elevated temperatures. These systems are, however, expected to be resistant to corrosion in typical service water environments which is a principal reason for using them to repair systems degraded by microbiologically induced corrosion (MIC), general or pitting corrosion, or leaching.

An important CFRP physical property is the glass transition temperature, or Tg2. This is the temperature, below which, the polymer epoxy remains in its hard, glassy state. Above Tg, the epoxy starts to soften, and, as the temperature increases, the material will become rubbery, until it eventually flows. The composite also becomes increasingly susceptible to debonding from the substrate or delaminating as the temperature increases above Tg. Tg is a property that depends on the epoxy constituents, the cure time, and the cure temperature used during installation. Higher cure temperatures and longer cure times, up to a limit, generally increase the installed Tg. Individual plies are cured overnight during fabrication, but the entire system is cured in-place before the system is qualified and entered into service. This is also unique compared to most metallic piping systems which, except for joining welds during installation, are largely fabricated off-site.

The CFRP composite system has significant differences with the traditional ferritic steel that comprises the reactor pressure vessel (RPV) and pressure coolant boundary piping (PCPB) in some NPPs. The NRC has a long history and deep technical knowledge in regulating ferritic steels and this experience is instructive when considering the appropriate treatment of CFRP.

Ferritic steels have a nil-ductility reference temperature (RTNDT) which characterizes the transition between ductile and brittle fracture. Regulations exist to provide assurance that these materials operate in the ductile region above the RTNDT. Analogously, in safety-related applications, it is important that the CFRP composite system operate in the brittle regime below the Tg. Additionally, ASME has design rules and margins for dealing with ductile and brittle materials. While those rules have been based on steels, similar concepts are applicable for CFRP composites. However, the specific approach and margins associated with CFRP system should recognize the paucity of relevant nuclear operating experience and current uncertainty associated with their long-term performance in nuclear environments.

The background in the last several paragraphs is necessary to provide a basic understanding of the nature of a CFRP composite system and develop an appropriate analogy and understanding of differences with ferritic steels to provide both context and guidance on possible acceptable requirements for using these novel materials in safety-related, or safety-significant, nuclear applications. The idea is to leverage the NRCs knowledge of steel performance to develop consistent requirements for CFRP systems. Acceptable requirements are not intended to be static but should evolve as knowledge is gained through relevant nuclear operating experience and ongoing research on performance in representative nuclear environments.

Issue 1 - Adequacy of Technical Basis The most fundamental issue that the DPO raised is that the applicants in the subject PAs have not provided a sufficiently robust technical justification that CFRP will fulfill its intended function 2 While this report utilizes a single Tg value to characterize the transition region, there is a temperature range between the fully elastic and fully rubbery regions that is not adequately represented by a single Tg value. Further, there are several sources of Tg uncertainty including the uncertainty associated with measuring repeated Tg values and uncertainty due to systematic differences between the two ASTM techniques commonly used to measure Tg.

under specified operating conditions and design-basis events. The principal complications for staff in reviewing the acceptability of these PAs are, as discussed previously, that these systems represent novel applications within safety-related nuclear components, and that CFRP systems have different performance requirements and characteristics than typical metallic component repair and replacement strategies. Unfortunately, there is no existing NRC-approved ASME code requirements that can provide a foundation for staffs review and approval of the use of CFRP in these systems. Efforts to develop an applicable code case have been ongoing over the last several years but progress has been slow and sporadic. Regardless, it is imperative that NRC staff continue to proactively support the development of applicable code requirements. In the interim, staff will need to continue to comprehensively review the PAs to determine their acceptability.

Some of the more significant concerns raised in the DPO submittal which underpin this broader claim that the CFRP systems have an inadequate technical basis are as follows:

The margin between the post-installation (or field) Tg and the maximum design temperature is insufficient.

The safety factors used for analyzed loading conditions are insufficient based on current knowledge.

The basis for the proposed cure temperature adjustment factor is inadequate.

The use of a higher cure temperature at the terminal ends could lead to unacceptable performance away from the ends and additional thermal stresses.

There is uncertainty in the elevated temperature performance and durability of the CFRP system based on proposed cure conditions.

CFRP systems have unique failure modes compared to metallic materials.

Subsequently, each of these supporting concerns will be individually addressed.

Margin between field Tg and Tmax The margin between the post-installation Tg (Tgf, field Tg) and the maximum design temperature (Tmax) is an important consideration as it delineates the onset of non-elastic behavior for the epoxy that could potentially lead to debonding or delamination. The earliest PAs and associated safety evaluations (SEs) (e.g., Surry and STP) do not discuss the notion of achieving an acceptable Tgf nor do they prescribe a minimum curing temperature (Tcure). There is an implicit assumption that the ultimate Tg (Tgu, or the maximum value that can be reached for the CFRP system with a sufficiently high Tcure) will provide sufficient margin between Tmax and Tgf.

Correspondingly, there is no post-installation acceptance testing to measure the Tgf in these earliest submittals. Assuming a sufficient degree of cure, as verified during installation, the difference between the Tgu and Tgf values is largely a function of Tcure. Lower Tcure leads to a lower Tgf and a bigger discrepancy between Tgf and Tgu.

Subsequent submittals (e.g., ANO, and Brunswick, chronologically) do differentiate between Tgf and Tgu and acceptance criteria are identified for both the Tgf and Tcure.

This information is summarized in Table 1. The Tg values in this table were estimated by staff based on NRC-sponsored testing that was completed after the PAs were approved. They represent best-estimate values of the onset Tg (i.e., temperature at beginning of the elastic to rubber transition region). They do not account for uncertainty in Tg measurements nor are they necessarily representative of field installation conditions.

The are currently no ASME nuclear code requirements for CFRP so there is no readily accepted margin between the service temperature (Tmax) and Tgf for the subject CFRP systems. However, an important consideration in determining an acceptable margin is that two ASTM methods commonly used to determine Tg (ASTM E1640 and ASTM E1356) have repeatability ranges (i.e., within a given laboratory) of approximately +/- 8oF while the reproducibility limits (i.e.,

between laboratories) can be higher than +/- 20oF. Further, the two methods typically measure different Tg values with the E1356 measurements being generally more conservative. The

difference in Tg between these methods was approximately 10oF in the NRC-sponsored evaluation of representative CFRP systems.

ASME PCC-2-2019, Repair of Pressure Equipment and Piping contains Article 401 Nonmetallic Composite Repair Systems: High Risk Applications, Article 401, Table 401-3.4.2-1 specifies that Tg should be 36oF above the service temperature for non-leaking substrate conditions (which is required for CFRP application in the subject PA applications) while the acceptable margin for a leaking substrate is 54oF. Although this isnt a nuclear standard, it applies to repair of equipment, piping, pipelines, and associated ancillary equipment after they have been placed in service, and the qualitative definition of high-risk in the standard appears applicable to the risk associated with CFRP failure in the SWSs specified in the PAs. Therefore, the requirements in this standard appear germane.

NRCs requirements for RTNDT temperature margins are quantitatively similar. 10 CFR 50.61 requires that RPV materials operate above RTNDT which includes a margin term that typically varies between 40oF and 65oF. Similarly, Appendix G to 10 CFR Part 50 contains an additional temperature margin of +40oF for the calculated pressure-temperature limits based on the limiting material properties when the reactor core is critical. Further, the preservice hydrotest temperature is a minimum of 60oF above the limiting RPV RTNDT value and up to 160oF greater than the RTNDT of the highest limiting material in the closure flange region. While these requirements are for ferritic steels for components within the primary pressure boundary, and exceptions to these requirements are permitted, it is nevertheless informative to recognize the generous margins applied to metallic materials having a long regulatory history and operating experience.

Considering all these factors, a reasonable starting Tgf margin is 40oF above Tmax in applications where CFRP failure could significantly affect plant risk. As previously indicated, the transition from a glassy to rubbery material occurs over a temperature range that can be represented by several different Tg values. Use of the Tg associated with the extrapolated-onset of storage modulus is the most conservative Tg measure as it represents the first change in the lower temperature epoxy modulus. This margin of 40oF does not account for variability that may occur during installation, which may be significant, especially if the witness panels and subsequent post-installation test specimens do not adequately represent the limiting material. Differences in cure temperature alone led to Tg variation of approximately 30oF in the NRC-sponsored research. While this is a recommended margin for initial CFRP application, this margin could be revisited as additional operating experience and research information on related effects is obtained.

Effective safety factors The PA applications appear to do a thorough job of analyzing potential loading scenarios, and associated limit states (i.e., failure modes), for the CFRP systems. The PA applications all utilize the Load and Resistance Factor Design (LRFD) approach. This approach is based on target reliability indices that correspond to an acceptable probability of failure. The target reliability

index is considered separately for each failure mode, based on the nature and consequences of that failure mode. Failure modes that occur with little warning require a higher reliability index (i.e., lower probability of failure) because of the abrupt nature and potentially severe consequences. For each failure mode, the design load is increased by a load factor (LF) while the materials resistance at the limiting end condition is reduced by a resistance factor (). The value of is assigned based on the target reliability index. The CFRP system design then is based on achieving an acceptable margin for each of the limit states. The PAs evaluated either 7 (Surry) or 9 (STP, ANO, and Brunswick) limit states.

This approach differs somewhat from the ASME Section III Allowable Stress Design (ASD) approach which also requires consideration of distinct design loading scenarios. Deterministic effective factors of safety (FS) are required for each specified design condition. Some of the FS values are also a function of the service level which range from normal operation having the highest FS criteria to emergency loads with the lowest FS criteria. An effective factor of safety (FSeff) for a particular failure mode can be determined for the LFRD approach as FSeff = LF/.

FSeff values can then be used to compare ASD and the LRFD acceptance criteria used in the PAs.

The SEs comprehensively describe the evolution of the ASD FSeff criteria over time within the ASME Code. The most recent criteria for both ductile and brittle materials are summarized in

((Table 2. The table also contains the acceptance criteria used in the PAs for the various limit states and matches these criteria with the most similar ASME criteria. The actual FSeff values from the PAs are also summarized in the table.

Note that only the limiting FSeff values for the associated plant piping location and diameter are included in the table. Further, multiple CFRP system limit states may correspond to a given ASME design loading category and, in these cases, only the smallest FSeff value is provided.

Finally, it is not clear in the PAs or the associated SEs how those limit state transients for routine operating conditions would be characterized under ASME ASD rules for levels A and B transients. Therefore, such transients have simply been grouped within the normal (Level A) longitudinal stress ASD category which has a similar ductile FSeff. Further, not all ASD design loading conditions appear to be associated with CFRP limit states.

As seen in the table, there is a 1-to-1 correspondence between the PAs and ASD acceptance criteria for ductile materials. Additional PA acceptance criteria and the calculated FSeff values are also more consistent with the ASD ductile acceptance criteria. However, CFRP is an inherently brittle material, as is universally acknowledged in the PAs and SEs. The SEs also indicate that a draft ASME Code Case (under development) for CFRP, as well as a few published papers on CFRP design, utilize a FSeff of 10 for the ASD methodology, which is consistent with the design margins for brittle materials. It is more prudent to base a first-of-a-kind design for nuclear application on higher FSeff values that are more consistent with those used for brittle materials than basing the design on FSeff values associated with ductile materials, especially until additional nuclear experience is gained using these CFRP systems.

It is also important to note that there are other factors that are multiplied to the CFRP resistance factor in the SEs, namely the material adjustment factor (C), the time effect factor () and the cure effect factor (Ceff). The SEs stipulate that these additional factors increase FSeff. However, these other factors are intended to account for degraded material strength and modulus properties due to environmental exposure, time or aging effects, and effect of the cure temperature on initial properties. Therefore, they are used to determine the material state at the end of its intended design life, and it is not appropriate to consider these factors as contributing to FSeff within the LFRD approach.

Cure temperature adjustment factor (Ceff)

The properties of a CFRP system are a function of the temperature used to cure the epoxy (Tcure). This is starkly illustrated in NRC-sponsored testing performed at Engineering Mechanics Corporation of Columbus. (EMCC). Representative CFPR single-ply laminates were cured at either 85oF or 140oF. The average measured Tg values decreased by 30oF, from 160oF when Tcure = 140oF to 130oF when Tcure = 85oF. Further, the average tensile strength in air at room temperature was 240 ksi for Tcure = 85oF and 220 ksi when Tcure = 140oF. However, the

average tensile strength in air at 140oF was 120 ksi when Tcure = 85oF and 180 ksi when Tcure = 140oF.

While the lower temperature cure resulted in slightly higher strength at room temperature, the strength decreased by approximately 50% at 140oF and was only about two-thirds of the strength of the material cured at 140oF. Therefore, the Ceff factor for the tensile strength is approximately 0.67 at 140oF in air. The cure temperature had virtually no effect on the modulus measured at room temperature. The tensile modulus measured at 140oF, however, was 25%

less for the material cured at 85oF while the material cured at 140oF exhibited insignificant differences between room temperature and 140oF measurements. This infers a Ceff factor for the modulus is approximately 0.75 at 140oF in air.

These requirements will, at least partially, address the cure temperature effects if both the witness specimens and the installation samples are representative of the limiting state of the CFRP repair. This post-installation testing becomes an important consideration in demonstrating the ability of the installed system to fulfill its design requirements. This will be discussed more fully in the section of the report addressing installation requirements.

The other associated concern raised by the submitter is that the constant Ceff values associated with short-term testing in air may not be appropriate for addressing environmental and time effects, or applicable to other failure modes. The EMCC study attempted to partially address this concern by exposing specimens cured at 85oF and 140oF in salt water for up to 10800 hours at either room temperature or 140oF and then evaluating the tensile strength and modulus. The results did demonstrate that neither the tensile strength or modulus of the CFRP test specimens experienced significant environmental or temporal effects up to 10800 hours of testing.

The implications of this testing on Ceff effects were ultimately inconclusive, however, because additional curing of the Tcure = 85oF specimens occurred during higher temperature exposure, which significantly elevated the measured properties at 140oF. Additional curing is not representative of service conditions given that Tmax conditions have a relatively low likelihood of occurrence such that it is unlikely that such an event will occur before the CFRP system has completed curing at lower operating temperatures. Licensees, to date, have not tried to credit such an effect in their proposed alternative applications, which is appropriate. Similarly, there is no basis for staff to credit such an effect when evaluating proposed alternatives.

There is even less known about Ceff effects on other failure modes. The EMCC study did examine flexure strength, double-lap shear strength, and pull-off strength. Both initial performance and performance after exposure at 140oF in a salt environment for at least 9400 hours0.109 days <br />2.611 hours <br />0.0155 weeks <br />0.00358 months <br /> were studied. However, all specimens were manufactured with Tcure = 85oF such that effects of higher cure temperatures, and ultimately Ceff effects, could not be determined. Further, as with the previous EMCC testing, while initial properties typically decreased initially with elevated temperature, continued exposure caused increases in both the flexure and double-lap shear testing due to post-fabrication curing. This phenomenon makes it impossible to determine how the CFRP system will perform under more representative conditions.

Because the SEs and associated PAs do not provide a technical basis for the applicability of the constant Ceff values to other failure modes and for consideration of environmental and time effects, this is an additional area of uncertainty. However, the simple approach of using constant Ceff values for design to address all strength and modulus may provide an effective design tool, Ideally, more conclusive testing would be performed to demonstrate applicability of using constant Ceff values or to develop Ceff values that specifically address these effects as a function of Tcure. Generic testing similar to that performed by EMCC could be conclusive if additional care is taken to eliminate post-fabrication curing as well as to expand the test matrix to consider Tcure effects on other relevant failure modes. Ideally a wider range of curing temperature would be evaluated to represent expected variability during installation, but evaluating effects for bounding Tcure values could be an acceptable approach. Alternatively, verification testing could be performed on each CFRP installation to measure representative properties associated with each design limit state.

Effects of cure temperature variability The natural temperature gradient induced by the heating arrangement will result in a higher Tcure near the heaters and a minimum Tcure for those portions of the CFRP system furthest away from the heater. As indicated previously, Tgf is a function of Tcure and higher temperatures lead to higher Tgf values. Higher Tg is generally correlated with better CFRP performance as well, such that the portions of the system farthest from the heaters are expected to be the most susceptible to failure (i.e., have the lowest failure margins).

The DPO also raises a concern that variability in Tgf could result in the terminal ends remaining in the linear-elastic region while other portions of the installed repair undergo plastic deformation with a corresponding failure to adequately transfer stress during a design-basis event. Variability in Tgf could also result in the middle lengths of the repair debonding from the host pipe or delaminating between layers while the terminal ends remaining in the glassy, brittle condition.

The Panel agrees, as stated above, that CFRP system failures, assuming relatively constant intake or exhaust temperatures over the length of the CFRP system, are most likely in regions with the lowest Tgf and that failure becomes increasing likely as Tmax increases above Tgf.

The DPO also raised a concern that the variability in Tgf, due to distance from heater sources along the length of the repair, could result in unanalyzed thermal stresses at the interface between regions having different Tgf values. Because Tcure variations will be gradual over the

distance between the heater source, the Tgf variation is also expected to be gradual over the repair length and no evidence has been provided that such an abrupt interface will result from these gradual Tcure variations. The DPO more plausibly hypothesizes that additional thermal stresses could evolve from Tgf variations along the repair length. However, Tgf variations would induce thermal stresses only if it were demonstrated that the CFRP density also varies with Tgf.

While such a demonstration would be interesting to investigate as a research topic, any subsequent thermal stresses arising from expected gradual Tgf variations should be modest, and a secondary effect, at best.

Elevated temperature performance and durability of CFRP As previously discussed, the EMCC results demonstrate that Tcure can significantly affect the short-term elevated temperature tensile strength and modulus results. At low Tcure (85oF) in air, the tensile strength decreases by 50% while the tensile modulus decreases by 25% between room temperature and 140oF. At higher Tcure (140oF) in air, the tensile strength decreases only by 18% between room temperature and 140oF while the modulus doesnt significantly vary over this temperature range. Therefore, lower Tcure decreases the CFRP high-temperature system performance.

Flexure strength, flexure modulus, and terminal-end bond strength testing was also performed by EMCC at both room temperature and 140oF but only on CFRP material with a low Tcure of 85oF. The average flexure strength at room temperature was 126 ksi while it was only 11 ksi, or approximately a factor of 10 less, at 140oF. The flexure modulus was also dramatically affected by temperature, going from 9981 ksi at room temperature to 2710 ksi (73% decrease) at 140oF.

Similarly, the double-lap shear results for GFRP (i.e., the interface ply) on steel substrates for material with a low Tcure of 85oF was 1110 psi at room temperature and 350 psi (69% decrease) at 140oF. The pull-off strength of GFRP on steel substrates for material with a low Tcure of 85oF was 2260 psi at room temperature and 1380 psi (39% decrease) at 140oF. While higher Tcure (140oF) material was not tested, based on the good tensile strength and modulus property retention for this material, it is also expected that the differences in room and high-temperature flexure strength, flexure modulus, double-lap shear, and bond strength would be much less than those measured for the Tcure = 85oF material.

Recall that the measured, extrapolated-onset Tg values for the CFRP/GFRP materials tested in the EMCC study was approximately 130oF, or 10oF below the testing temperature. These results provide stark evidence of the degradation that can occur at temperatures when Tmax >

Tg. This makes it important to determine Tgf and demonstrate acceptable system performance on representative witness specimens to validate the acceptability of the repaired system.

Witness specimens specifically designed to measure other properties which determine the limit state margins, such as flexure strength and bond strength should also be considered to validate that acceptable design margins are retained in the installed system for these properties.

As indicated, the EMCC tests did combine high temperature and salt water out to 10,800 hours0.00926 days <br />0.222 hours <br />0.00132 weeks <br />3.044e-4 months <br /> and exhibited no significant degradation in either the tensile strength or modulus for specimens cured at 140oF.

The EMCC testing, as explained earlier, went further to evaluate both flexural strength and modulus and bond strength at terminal ends after saltwater exposure at 140oF for a CFRP ply with Tcure = 85oF. While post-fabrication curing makes it difficult to compare the effect of exposure with the initial properties, the flexure strength at 9400 hours0.109 days <br />2.611 hours <br />0.0155 weeks <br />0.00358 months <br /> was only 40% of the flexure strength at 1000 hours0.0116 days <br />0.278 hours <br />0.00165 weeks <br />3.805e-4 months <br />. The tensile modulus at 9400 hours0.109 days <br />2.611 hours <br />0.0155 weeks <br />0.00358 months <br /> was approximately 75% of the modulus at 1000 hours0.0116 days <br />0.278 hours <br />0.00165 weeks <br />3.805e-4 months <br />. The double-lap shear results, while exhibiting significant variability in individual results, did not change significantly between 1000 hours0.0116 days <br />0.278 hours <br />0.00165 weeks <br />3.805e-4 months <br /> and 10,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> of exposure. The measured pull-off strength was also relatively consistent between 1000 hours0.0116 days <br />0.278 hours <br />0.00165 weeks <br />3.805e-4 months <br /> and 10,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> of exposure.

This research provides some initial evidence that CFRP retains its properties in representative environments for approximately 10,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> at Tmax when Tg > Tmax while materials with Tg <

Tmax due to lower Tcure can experience significant property degradation with time. More research of field-representative CFRP systems in SWS environments is needed to reduce the uncertainty associated with the high-temperature, long-term performance of these systems. Alternatively, periodic inspection using techniques to assess hidden layer soundness should be considered in addition to the periodic visual inspection that is currently performed. Periodic surveillance could also be effective in measuring key properties of installed systems over the service life. Both enhanced inspection and surveillance could be used to demonstrate that required design margins remain acceptable in lieu of, or in addition to, conducting additional research.

Unique failure modes The DPO submittal raises the concern that the PAs and associated SEs do not adequately consider deformation, delamination, and debonding of the installed CFRP due to elevated temperatures. These failure modes are distinct from those experienced by metallic nuclear components. Deformation failures would occur if Tmax > Tgf such that the material transitions to a

rubbery state, as discussed previously. Debonding is the separation of the GFRP layer from the metal substrate. Delamination is the separation of individual CFRP or GFRP layers or plies.

There are two types of potential debonding failures, those that occur at the repair termination ends and those that could occur within the bulk of the repair. Debonding from the host pipe along the bulk of the repair is not a principal concern because the CFRP system is designed to withstand operational and design-basis loading independently, without crediting the host pipe properties. Debonding along the length of the repair could be a secondary concern if it leads to water ingress between the CFRP and the host pipe, leading to possible degradation of the host pipe at the termination ends.

Debonding at the repair termination ends is more significant because these ends are designed to transmit load to the CFRP system. Consequently, all the PAs and associated SEs address debonding of the system at these ends by analysis of this limit state as part of the systems design basis. Qualification testing has demonstrated the efficacy of this approach. A stainless-steel expansion ring with rubber gasket provides a supplemental mechanical anchorage at the termination ends to further guard against debonding, but this retaining system is not credited in the design analysis and has no safety-related function. Adhesion testing is also performed during installation to provide assurance that the substrate surface has been adequately prepared.

However, it does not appear that the effect of temperature on bond strength is addressed in the limit state analyses. As demonstrated in the EMCC testing, the bond strength could decrease at Tmax if the minimum specified Tcure is applied at the termination points. Long-term durability of the bond strength at the termination ends is also not addressed in the PAs or SEs. Further, there are no provisions to measure or sample the shear or bond strength of the installed systems or evaluate this over the intended service life of the repair. This later point is discussed in more detail in the section pertaining to installation and quality assurance.

Delamination of individual plies is predominantly a function of Tgu for the chosen epoxy materials, the nominal service temperature, Tmax, and the quality of the installation. None of the PAs consider the possibility of delamination as a limit state analysis during a design-basis event. However, excessive delamination of the original repair is addressed in the PAs and there are quality assurance measures, including visual inspection augmented by tap testing, which attempt to ensure that unacceptable delamination does not initially exist in the repair system.

Another plausible cause for delamination not discussed in the SEs is the possibility of separation if Tgf is exceeded during service.

Because delamination, debonding, and deformation due to exceeding Tgf are all unique failure modes compared to traditional nuclear metallic materials, they warrant special attention during design, installation, testing, and post-installation inspection to guard against failure under operating and design-basis conditions. Proper epoxy selection is the most important initial design consideration and this needs to be coupled with acceptable installation and curing methods to ensure that an acceptable Tgf is achieved. Acceptable installation procedures are also important to ensure an adequate bond between the initial GFRP layer and the metal substrate, and that subsequent plies also are adequately bonded. The quality assurance and testing program should be robust enough to appropriately monitor and engage in corrective action when critical variables are outside their qualified ranges, while also verifying the integrity of each ply as well as the entire repair system.

Operating experience Operating experience is typically a fundamental consideration in assessing material performance under commercial NPP conditions. However, because the subject PAs represent the first applications of CFRP in safety-related NPP systems, it can be challenging to both identify and assess relevant CFRP performance. All of PAs address operating experience in a similar manner with the only significant distinction being that the more recent submittals (i.e.,

ANO and Brunswick) updated the baseline information (i.e., in the Surry submittal) with the most recent operating experience.

Operating experience for successful internal applications in municipal water systems, fossil power plants, and, most importantly, NPPs is cited. The ranges of conditions for these applications span piping diameters from 24 to 264, operating pressures up to 300 psi, vacuum pressures down to -14.7 psi, temperatures from 32oF to 140oF, and water chemistries including salt water, raw water, raw sewage, waste-water sludge, and chemical effluent lines. These collective conditions are relevant to the PA applications. The earliest cited applications are from the early 2010s, or approximately ten years ago. The PAs indicate that multiple re-inspections of similar CFRP systems have been performed and that the systems have remained in the as-installed condition, with no change in performance. In particular, the CFRP installed in safety-related SWSs in Surry since 2018 and in STP since 2015 are cited as examples demonstrating good performance through re-inspections.

The PAs also identify some internal-piping CFRP-system failures and associated root causes, although these cases involved use of a different CFRP system and/or installation team than in the PAs. Causal failure factors include improper system design, improper material selection, lack of appropriate quality assurance or quality control measures, and inadequate installation and repair procedures. The PAs further describes mitigating factors in the design, installation, and quality assurance provisions used in the PA applications that provide assurance that similar failures will be avoided.

Issue 2 - Adequacy of Installation Procedures As mentioned previously, repairing damaged piping with CFRP can be an efficient and reliable solution but maintaining adequate quality assurance and control (QA/QC) is paramount in providing assurance that design requirements specified in the PA applications are met. The CFRP certification and repair process (i.e., system qualification, material verification, preinstallation preparation, and installation procedures) are generally well established for non-nuclear, safety-related applications. For example, carbon fiber components made for the aircraft and other industries are carefully constructed using advanced and highly automated processes conducted in dedicated production facilities that achieve very consistent material properties for the finished products.

Conversely, CFRP repairs on nuclear piping are done entirely on-site and by hand. The CFRP repairs are comprised of several ingredients combined on-site to form each individual ply within

the CFRP system. The installation is performed in the field by laying up single plies in multiple piping configurations and installation orientations within a single piping system, some of which may be challenging for the installer to access. The installation is also a lengthy process that often spans several days as individual plies are stacked during the repair to create the final multi-ply system. The final step is a complete simultaneous system cure. The nature of this manual installation process introduces a great deal of possible variability in the final material properties of the CFRP repair.

Because there is relatively little experience with correctly installing such systems in nuclear applications, the QA/QC measures required to ensure acceptable, long-term performance in nuclear applications are less proven. Chemical and physical validation testing that appropriately addresses the factors (i.e., critical variables) that can result in unacceptable performance is an essential component of the QA/QC program. Both the scope and representativeness of these validation tests are important considerations as the potential variability within the field repair can potentially lead to higher defect densities and reduce the mechanical properties compared to validation specimens prepared on flat substrate specimens under more tightly controlled installation and curing conditions. Each of the factors which potentially affect the adequacy of the installation and associated QA/QC procedures, and NRCs oversight of those procedures, will be subsequently addressed.

QA/QC adequacy The use of CFRP systems in safety-significant nuclear applications requires a robust QA/QC program. The initial impression of the DPO Panel is that the CFRP installation process and QA/QC program outlined in the PAs share more attributes with a traditional nuclear coating application than a structural application. While delving into the specifics of the program to confirm its acceptability is outside the scope of this DPO, such evaluation is warranted to both develop inspector guidance and identify potential weak programmatic aspects. Any identified potential weak points may require development of a stronger supporting technical basis or program modifications to correct any identified deficiencies. One important thrust should be a review of the qualification program(s) for installers and system inspectors. There is significant nuclear experience with the development and implementation of welder and inspection (particularly nondestructive evaluation) qualification programs. Such programs provide good models for both evaluating the appropriateness of the CFRP QA/QC programs and providing possible solutions to address any identified deficiencies.

Another focal area of this evaluation should be the control and validation of critical variables during the installation procedure. An example of a potential concern identified by the DPO Panel is the representativeness of air thermocouple readings to the cure temperatures of the CFRP plies, particularly the innermost plies. Since the installations are typically within strong-conducting metallic pipes, a large difference between the local heating temperature within the pipe and the temperature of the surrounding piping environment could lead to sharp temperature gradients within the curing CFRP system. This is just one of many potential critical variables that should be evaluated holistically during the QA/QC review.

The remainder of this section focuses on the testing used to validate the adequacy of the CFRP system. If this testing is correctly designed and performed, it can address installation uncertainties and variability and provide demonstrable proof that the system should meet its design requirements. As such, the validation testing is the most important aspect of an integrated and robust QA/QC program.

NRC inspector support The installation complexity and general lack of nuclear experience with CFRP systems challenge inspectors charged with overseeing and validating the acceptability of a CFRP installation. Once the QA/QC program review discussed above has been completed, inspector guidance should be developed to provide appropriate focus during an installation review. This guidance should contain a checklist, or summary, of the critical installation variables and their allowable ranges, and identify the most important aspects of the associated QA/QC program that an inspector should review. This summary should be supported by an assessment of the implications when allowable critical variable ranges are exceeded or when important aspects of the QA/QC program are violated. This assessment would help inspectors understand implications of any specific non-compliances issues that may arise.

The remainder of this section focuses on the testing used to validate the adequacy of the CFRP system. If this testing is correctly designed and performed, it can address installation uncertainties and variability, and provide reasonable assurance that the CFRP system meets the intended design requirements. As such, the validation testing is the most important aspect of an integrated and robust QA/QC program.

Validation testing There are both chemical and physical validation tests that are performed on the as-installed CFRP system. Chemical testing establishes that an adequate degree of cure has been reached and that, for some approved plants, the minimum required Tgf is reached. Physical testing establishes that acceptable mechanical properties have been met. There were a host of tests performed as part of the qualification (i.e., commercial dedication) process including tensile strength, tensile modulus, tensile strain, compressive strength, compressive modulus, flexural strength, flexural modulus, overlap shear strength, and shear bond strength. Mock-up tests to demonstrate watertightness of the system and demonstration of acceptable hydrostatic burst performance have also been performed. However, the principal physical tests performed on the installed system are an adhesion test of the prepared substrate surface and tensile testing of separate witness specimens.

Chemical testing

Both degree of cure and glass transition temperature sampling are essential quality assurance tests that should be performed on all installed systems. However, the implications of variability in these properties should be considered either through additional research and qualification testing or enhanced quality assurance measures in the absence of such refined knowledge.

Enhanced quality assurance measures would specify that samples be extracted from both critical locations (e.g., near the termination ends and other locations with minimum design margins) and those locations within the CFRP system that are most likely to violate the acceptance criteria (e.g., locations subjected to the lowest curing temperatures and locations with restricted accessibility or challenging installation orientations). Full-repair-thickness core samples should also be extracted from the repair at these locations so that both the surface layer and the layers closest to the metal substrate can be evaluated.

Physical testing The two most important physical tests performed in accordance with the approved QA/QC program are adhesion testing and tensile testing of a sub-population of witness specimens fabricated during the installation.

Adhesion testing The shear bond strength at the termination ends, as discussed earlier, is essential for transferring loads throughout the repair and represents a design limit state analysis. Verification that the field-installed system meets the minimum design requirements associated with the

qualification is therefore also essential. The adhesion testing specified in the QA\\QC is intended to demonstrate that the systems shear bond strength is acceptable.

The D4541 testing is a direct, or tensile pull-off test where a fixture is glued to the CFRP surface and then pulled perpendicularly to the surface either until the specified bond strength is exceeded, at which point the bond has passed, or until the bond fractures. If the stress at bond fracture exceeds the acceptance criterion, the bond also passes. The PAs are not clear on if this criterion should be exceeded for all tests, an average, or just one of the tests.

The basis and acceptability of the tensile bond strength criterion should be more clearly established and documented. If an acceptable basis does not currently exist, qualification testing should establish an appropriate basis. Alternatively, a more direct shear bond strength test could be performed on the installed system as part of the QA\\QC program to demonstrate acceptable shear bond strength. The PAs reference both ASTM D3528, Standard Test Method for Strength Properties of Double Lap Shear Adhesive Joints by Tension Loading and ASTM D7616, Standard Test Method for Determining Apparent Overlap Slice Shear Strength Properties of Wet Lay-Up Fiber-Reinforced Polymer Matrix Composites Used for Strengthening Civil Structures for determining shear bond strength. However, these standards are appropriate for laboratory testing and not tailored for field testing the installed system. It will therefore be necessary to either develop appropriate field-testing standards or conduct shear bond strength testing on representative witness specimens manufactured as part of the installation procedure if this route is pursued.

Witness specimen tensile testing Tensile testing is performed on field-prepared, single-ply witness specimens that are intended to represent the properties of the as-installed CFRP system. A minimum of two witness panels are

prepared during each day and work shift of the CFRP system installation. The witness panels are prepared by the qualified installer using the same epoxy and fiber materials which comprise the CFRP layer being installed that day. One of the two witness panels is prepared at the beginning of the work shift and the other one is prepared at the end of the shift. After preparation, the witness specimens are stored in a dry area near where they are prepared, but outside of the pipe, and typically at or near room temperature until the installation is complete.

After the installation is complete, all witness specimens receive their final cure using the same temperature and cure duration of the CFRP system repair. The final cure is performed either by placing the specimens inside the repaired piping and curing them both simultaneously or by curing the witness specimens in a separate, controlled environment using a representative temperature profile as recorded in the repaired SWS piping during the post-installation cure.

After completing the post-installation curing, a statistical sampling of witness specimens is tensile tested in accordance with ASTM D3039 in two groups: one group at the room temperature and the other group at the Tmax of the SWS piping (i.e., the design-basis faulted condition). Both the tensile strength and tensile modulus are measured at each temperature and used to confirm that the design requirements are met including, most importantly, that an appropriate cure temperature adjustment factor, Ceff, has been used in the design.

The witness specimen provides essential quality assurance that the system will satisfy the design requirements. As such, there should be high confidence that the witness specimen appropriately represents the CFRP systems properties. The witness specimens should reasonably account for differences in batch-to-batch and within-batch variability by making specimens at the beginning and end of each installation shift, assuming that only one batch is prepared during each shift. Witness specimen preparation by a qualified installer, assuming that this installer is also working the corresponding shift, also provides some measure of potential installer variability.

There are, however, other variables associated with the CFRP repair installation that raise significant uncertainty of the representativeness of the witness specimens. Variability in the as-installed properties of the CFRP repair is expected due to the length, complexity, and volume of the repair. Each witness specimen is prepared outside the pipe on an easily accessible, flat surface, in the vertical-down installation orientation. The CFRP repair installation is clearly more challenging. The fully circumferential repair will require installation at all installer orientations.

The piping size and configuration are also expected to play a role. There is inherent curvature of the work surface that increases as the pipe size decreases. Piping elbows and branch tee connections provide additional geometric discontinuities. As the pipe size decreases, accessibility is expected to become more challenging, especially further away from the systems access ports, and such restrictions may require the installer to work in different positions. The effects of worker fatigue and environmental conditions during installation should also be anticipated. These conditions are analogous to those experienced during welding and there are long-established process qualification and welder certification requirements that deal with these conditions. However, no such similar practices or requirements for CFRP system installation are readily apparent in either the SEs or the associated PAs.

Additionally, this is a multi-ply, manual installation occurring over several days and the environmental conditions within the system are likely to vary based on the location and time of the repair. The witness specimen, after its prepared, is stored in a dry location, nominally at room temperature, before the final cure. The effects of humidity and temperature variability differences between the system repair and the witness specimens before the final cure are not addressed in the SEs or associated PAs. Additional uncertainties are introduced if the final witness specimen cure is conducted separately from the final system cure, at an unspecified time and in a separate location, using the measured air temperature profile recorded during the system cure. Neither the effects of humidity differences nor the effects of differences between the air temperature and the temperature gradient through the multi-ply CFRP repair, especially at the layer adjacent to the substrate, are addressed in the SEs or associated PAs.

As discussed previously, it is well established that the locations with the lowest Tcure will have both the lowest Tgf and mechanical properties. These locations are consequently more likely failure initiation sites within the CFRP system. It is therefore essential that either the witness specimens properties are representative of these locations through appropriate fabrication, handling, storage, and testing provisions, or that property variability in CFRP systems be explicitly addressed through qualification testing or subsequent research to quantify its effects and implications. These measures are especially important because of the relatively small design margins between Tgf and Tmax and between the design limit states and the required CFRP systems properties. A ductile-type failure (i.e., Tmax > Tgf) may be constrained and limited as long as the exceedance is also limited within the repair. Initiation of a brittle-type failure (e.g.,

applied stress exceeding system strength), however, may be self-propagating such that the corresponding damage is more extensive.

Other physical testing considerations While there is a litany of other physical tests performed during qualification, appropriately representative witness specimen tensile testing can be used as a surrogate to provide reasonable assurance that these other properties will be acceptable as well. One unique property that is not well approximated by the witness tensile testing is the bond strength. The bond strength at the termination ends, as discussed earlier, is essential for transferring loads throughout the repair and represents a design limit state analysis. As indicated previously, either a stronger basis is needed to correlate current adhesion tests pull-off testing requirements with those specified for the shear bond strength within the design basis, or more appropriate tests to measure the shear bond strength in the field should be considered. Testing representative witness shear bond strength specimens manufactured during the installation procedure may provide an acceptable approach to providing quality assurance.

Issue 3 - Risk Significance As part of the risk-informed decision making process, the risk information from probabilistic risk assessments (PRAs) are used to help determine the significance of an issue for plant safety. A PRA is based around the risk triplet, which can be described as:

1.

What can go wrong.

2.

How likely is it.

3.

What are the consequences.

A PRA considers many different transients and accident scenarios, assigns frequencies to how often they may occur, and then assesses the severity of the incident if it does occur. As an example, when assessing the risk associated with a pump failing, reliability and operational data are used to determine the probability of that pump failing in a given year. This failure probability is considered the yearly initiating event frequency (IEF). Then, the PRA assesses the consequences of pump failure to determine the probability that this pump failure would lead to core damage. This probability is called the conditional core damage probability (CCDP). The IEF is multiplied by the CCDP and then the baseline risk with no assumed pump failure is subtracted to determine the change in core damage frequency (CDF).

The NRC policy for the use of risk information in making regulatory decisions is described in detail in Regulatory Guide 1.174, Revision 3 An Approach for Using Probabilistic Risk Assessment in Risk-Informed Decisions on Plant-Specific Changes to the Licensing Basis..

Regulatory Guide 1.174 describes the following five principles used at the NRC for risk-informed decision making (RIDM):

1.

The change meets current regulations unless it is specifically related to an exemption.

2.

The change is consistent with the defense-in-depth philosophy.

3.

The change maintains sufficient safety margins.

4.

Proposed changes in risk are small and consistent with the Commissions Safety Goals Policy Statement.

5.

The change is monitored using performance-measurement strategies.

Assuming the other principles of risk-informed decision making have been satisfied, the CDF for a licensing action needs to be below 10-5 or 10-6 yr-1, depending on the overall plant safety, for that action to be possibly acceptable.

It is worth noting that passive components, such as piping and vessels, present a unique challenge to PRA analysis. While many items in a NPP have a statistical basis for determining their failure frequency, passive components have been very reliable over the life of the plants such that their IEF for a significant failure is very low but difficult to accurately quantify. This historically excellent reliability exists because piping and other passive components typically were built, and are maintained, to a very high standard, such that they have a very high safety

margin. However, plant and component-specific fabrication, operation, and mitigation measures can vary in relatively small, but potentially significant ways, such that the low IEFs assigned to passive components are highly uncertain.

These attributes make it challenging to predict the effects on the IEF if the safety margin of a component decreases upon reducing construction and maintenance requirements. However, the consequences of changes in safety margin for these components can be evaluated by assessing the conditional core damage probability (CCDP) for materials issues associated with passive components. CCDP provides a conservative way to determine if changes to a particular system have the potential to lead to significant risk increases. To evaluate CCDP results, the industry and NRC often use the guidelines described in the Electric Power Research Institute (EPRI) Topical Report TR-112657 Rev. B-A Revised Risk-Informed Inservice Inspection Evaluation Procedure. This report assigns items with a CCDP of 10-4 or higher as high safety significance, items with a CCDP between 10-6 and 10-4 as medium safety significance, and items with a CCDP below 10-6 as low safety significance.

Three risk analyses were conducted to support the Panels assessment of the adequacy of the CFRP repair alternatives. Two of the risk analyses were plant-specific and focused on the plants that have SEs allowing the use of CFRP. The third analysis evaluated the impact of potential CFRP failure at 15 different plants. Each analysis used different assumptions about the likelihood and consequence of CFRP failures, and these assumptions significantly affect the predicted implications. Each of these risk analyses used the Standardized Plant Analysis Risk (SPAR) Model, which was developed by the NRC and is generally consistent with industry risk analyses.

Plant-specific analyses The risk analysis provided in the DPO submittal (DPO Analysis) covers all four NPPs that have been approved to use CFRP. A confirmatory risk analysis (Supplemental Analysis) was performed on these systems at the behest of the DPO Panel. Two plants were modeled in the Supplemental Analysis, ANO and STP, as they were considered more sensitive than Surry and Brunswick to CFRP failures.

The DPO Analysis uses the Risk Increase Ratio (RIR) and Birnbaum number as the figures of merit for the analysis. The RIR is a measure of the relative risk increase that is attributable to the specified components failing. A RIR of greater than one denotes an increase in risk while a number less than one denotes a reduction in risk. The Birnbaum number measures of the rate of change in total risk as a result of changes to the probability of an individual basic event. A larger number indicates a faster rate of change, meaning that the risk increases at a rapid rate.

Smaller numbers indicate a slower rate of change. The DPO Analysis assumes that specific component(s) have failed such that, in this case, the Birnbaum number for the entire year is equivalent to a CCDP estimate. Both of these measures are also equivalent to a CDF measure if the CDF is also predicated on the assumption that the specific component(s) fail(s).

Broader risk analysis The third risk analysis was performed to determine the risks associated with CFRP failures in a larger number of plants. This broader risk analysis is described in the report titled On Assessing the Increase In Risk Related To The Potential Failure Of The Piping Repaired With Carbon Fiber Reinforced Polymer (ADAMS ML21265A040) (Broader Risk Analysis). This analysis evaluated the implications of CFRP failures by calculating CDF for insufficient flow in the piping system caused by the pumps failing, the plugging of strainers, and plugging of heat exchangers at several selected Boiling Water Reactor (BWR) and Pressurized Water Reactor)

PWR plants.

The Broader Risk Analysis assumed that the CFRP repairs failed, increasing the chances that the downstream components would also fail. This analysis assumed that the baseline probability for insufficient pump flow due to a piping failure is 10-3 yr-1, the baseline probability of downstream strainers plugging is 10-4 yr-1, and the baseline probability of heat exchangers failing due to plugging is 10-5 yr-1. The failure probability for the affected components to account for CFRP failures were assumed to be multiplied by 10 for insufficient flow, and 100, for strainer and heat exchanger plugging, to determine the CDF due to CFRP failures in the systems. These changes to the failure probability were initially performed independently such that the assumed increased probability of one component type (e.g., pumps) did not affect the failure probability of other component types. The CDF results are given in Tables 10 and 11 for the modeled BWR and PWR plants respectively.

Table 10: CDF due to CFRP Failure for Different Scenarios in Eight Sample PWRs Scenarios PWR 1 (yr-1)

PWR 2 (yr-1)

PWR 3 (yr-1)

PWR 4 (yr-1)

PWR 5 (yr-1)

PWR 6 (yr-1)

PWR 7 (yr-1)

PWR 8 (yr-1)

No. 1 -

Insufficient Flow 3E-09 2E-07 3E-07 2E-06 2E-06 2E-06 8E-06 3E-08 No. 2 -

Downstream Strainers Plugging 5E-09 7E-08 1E-06 4E-07 4E-06 2E-06 1E-05 3E-07 No. 3 -

Downstream Heat exchangers Plugging 4E-08 5E-08 2E-08 1E-09 6E-08 8E-09 2E-07 2E-08 The resulting CDF values from this analysis vary greatly from plant to plant. The CDF values due to the assumed component failure probability increases resulting from CFRP repair failures range from 3x10-10 yr-1 to 1x10-5 yr-1, nearly four orders of magnitude of difference. Considering all three possible systems on a plant-by-plant basis, five of the eight PWRs and four of the seven BWRs have CDF increases greater than 10-6 yr-1.

Table 11: CDF values due to CFRP Failure for Different Scenarios in Seven Sample BWRs Scenarios BWR 1 (yr-1)

BWR 2 (yr-1)

BWR 3 (yr-1)

BWR 4 (yr-1)

BWR 5 (yr-1)

BWR 6 (yr-1)

BWR 7 (yr-1)

No. 1 -

Insufficient Flow 9E-07 3E-07 3E-06 5E-09 1E-07 8E-07 1E-05 No. 2 -

Downstream Strainers Plugging 3E-06 2E-08 3E-06 1E-09 2E-08 1E-05 1E-06 No. 3 -

Downstream Heat exchangers Plugging 3E-10 1E-07 5E-08 2E-08 4E-08 2E-08 9E-09 As seen, the CDF increases from the Broader Risk Analysis are much lower than the plant-specific results discussed previously. This difference stems directly from the assumptions used in each analysis. The plant-specific analyses assumed specific, definitive component failures while the Broader Risk Analysis only increased the probability of such failures by 10 or 100. It is possible to more directly compare the results from these different studies by multiplying the initial assumed failure probabilities in the Broader Risk Analysis for insufficient flow and downstream strainers plugging by another factor of 100 while multiplying the probability for downstream heat exchanger plugging by another factor of 1000. With these additional factors, both the Broader Risk Analysis and the plant-specific analyses assume that specific components within each scenario have a failure probability of 1.

The distributions obtained from the Broader Risk Analysis results after this additional scaling are summarized in Table 12. This table contains the highest, median, and lowest scaled CDF results for the fifteen plants. This range of results implies that a size number of plants in the U.S.

Fleet could experience significant risk increases if CFRP failures occur such that they affect SWS pump, screen, or heat exchanger functionality. This range is consistent with the range of both the DPO Risk Analysis and the Supplemental Risk Analysis results.

Table 12: Scaled Broader Risk Analysis CDF Results Scenarios Highest Median Lowest CDF (yr-1)

CDF (yr-1)

CDF (yr-1)

No. 1 - Insufficient Flow 1.00E-03 8.00E-05 3.00E-07 No. 2 - Downstream Strainers Plugging 1.00E-03 1.00E-04 1.00E-07 No. 3 -Downstream Heat exchangers Plugging 2.00E-04 2.00E-05 3.00E-07

In summary, the three independent risk analyses show that CFRP failures can be risk significant if CFRP failure leads to failures of SWS components. While all three analyses used the SPAR model to conduct their evaluations, they used different assumptions and different approaches to calculate risk. Notwithstanding these differences, the three models showed relatively consistent results for the increased risk associated with SWS component failures and the median CDF for a component failure is 10-4/yr. The Broader Risk Analysis results demonstrate that the risk sensitivity of CFRP failures varies significantly from plant to plant and from system to system within a plant about this median estimate. Failures in the specific piping systems can be highly risk significant or have low-risk significance. The results also suggest that the systems modeled in the DPO Risk Analysis and Supplemental Risk Analysis are among the more highly risk sensitive plants to CFRP failures within the distribution of plants modeled in the Broader Risk Analysis.

Conclusions The subject DPO raised concerns about the acceptability of the staffs approved use of CFRP systems to repair degraded piping in service water and other similar component cooling piping systems. In particular, the DPO contends that there is insufficient technical justification to demonstrate an acceptable level of safety for the proposed CFRP composite systems and that the risk associated with the potential failure of the CFRP composite systems is unacceptably high. Specific concerns raised within the DPO were grouped within three broad categories (i.e.,

adequacy of the technical basis, adequacy of the installation procedures and quality assurance methods, and risk significance) that were analyzed individually by the DPO Panel. However, these topics are inexorably linked in determining if an acceptable safety case has been established.

The evaluation Panel believes that CFRP is a viable repair method that holds promise for use in addressing large-scale or localized degradation in difficult-to-replace piping systems. This method should continue to be explored and supported through both additional research and qualification testing along with the development of consensus codes and standards to guide the design, installation, aging management, and quality assurance provisions associated with these repairs. he Panel recognizes that epoxy material selection is a particularly important consideration within the design phase of these systems because it determines the upper limit on the achievable glass transition temperature.

Notwithstanding these sentiments, the evaluation Panel believes that there is merit in many of the concerns raised by the DPO. The initial approvals are based on temperature and stress margins that are lower than have been historically applied for commensurate metallic materials.

Lower margins increase the likelihood for potential larger-scale and common-cause failures within these systems, during either design-basis events or normal operations. Common-cause failures are a particular concern for use of CFRP in redundant service water applications designed and expected to have similar operational histories, design-basis demands, and similar installation conditions. There is some operational experience on CFRP failures within non-safety piping system which buttresses this common-cause failure concern. There is also significant

uncertainty about the acceptability of adjustment factors to address cure temperature, environmental, and long-term degradation. This uncertainty increases for operating times exceeding the current limited testing basis.

The quality assurance provisions associated with the fabrication, installation, and verification testing for the CFRP systems is complex. As such, a more holistic evaluation of the quality assurance program is warranted to identify potential weaknesses and support the development of guidance for regional inspectors charged with overseeing the installation and subsequent management of these repairs. The potential for performance variability within the repair is significant given both the breadth of the repair and the myriad configurations, installation orientations, and accessibility challenges that may be associated with the repair of any single piping system. Particular emphasis is needed to ensure the representativeness and applicability of verification testing as well as the efficacy of in-service evaluation as these measures provide the most salient indicators of the actual as-installed performance of the repair.

The risk implications of these repairs fundamentally impact these previous considerations as the strength of the technical basis and the rigor of the quality assurance and in-service evaluation programs should be commensurate with the risk. Associated requirements could be generic so that they are applicable to all plants and piping systems, including those most sensitive to potential CFRP failures. Alternatively, requirements could be graded to both address higher-risk applications while allowing appropriate relaxation as the risk decreases.

Several different risk assessments were considered in this evaluation to better understand the risk of potential CFRP failures. The highest CDF associated with assumed component failures for all three analyses is on the order of 10-3 /yr, while the median CDF in all three risk analyses is on the order of 10-4/yr. Such increases imply that CFRP failures could be highly risk significant. However, the CDF due to postulated CFRP failures in some plants is only on the order of 10-7/yr, which is of low safety significance. It is evident that the plant-specific risk can vary widely, as much as four orders of magnitude from plant to plant and from system to system within a plant. While the assumptions and approach used within the risk analysis strongly impact the results, it is clear that a CFRP failure within these systems may be risk significant, especially when considering common-cause failures.

As the risk impacts of CFRP can vary by several orders of magnitude, more rigorous risk evaluations should be explored to support both the currently approved applications and future staff evaluations of CFRP repairs. The lack of substantive operating experience and a validated failure model necessitates an evaluation of consequential effects of CFRP failure rather than a classical risk evaluation which considers the initiating event frequency associated with a CFRP failure. Failure modes and effects within these systems are also not well known which increases the uncertainty associated with even consequential evaluations. Accurately addressing common-cause failures within an assessment is an additional complexity. Notwithstanding these challenges, a more detailed risk evaluation in conjunction with appropriate sensitivity analyses can shed additional light on the risk significance of these various challenges.

More detailed recommendations associated with these conclusions follow.

Recommendations Based on the DPO Panels evaluation of the issues raised by the submitter, the following recommendations are provided to support both future staff evaluation of the proposed use of CFRP repairs within safety-significant systems and continued evaluation of the adequacy of the technical basis and the commensurate risk significance for applications that have already been approved. These later actions are intended to determine if compensatory regulatory actions are appropriate to provide reasonable assurance that these systems will continue to function as designed over their intended service life.

No NRC-approved ASME code requirements exist for use of CFRP in safety-related nuclear applications that can provide a foundation for staffs review of these systems. It is therefore imperative that NRC staff continue to proactively support the development of applicable code requirements. An acceptable code solution will allow for consistent, transparent, and stable assessments of future CFRP repairs.

Future proposed alternatives should incorporate more margin between the design-basis temperature requirements and the CFRP glass transition temperature. For risk significant applications, a margin of 40oF between the maximum service temperature and the field glass transition temperature is a recommended starting point. Smaller margins may be appropriate as the plant risk to potential CFRP failure decreases. These margins could be further reduced through additional research and assessment of the implications of CFRP system performance variability over the entire length of representative repairs, and development of enhanced installation and inspection procedures, along with requisite quality assurance measures.

More prudent load design margins should be required recognizing that these are first-of-a-kind applications and that these materials are inherently brittle and not ductile like typical metallic components. Effective factors of safety consistent with the allowable stress design margins provided in the ASME Code for brittle materials (i.e., SFeff = 10 for normal loading) are appropriate starting points that are consistent with other design standards for CFRP. As additional nuclear experience is gained on CFRP repairs, it may be possible to reduce these load margins.

Using a cure temperature adjustment factor to account for differences between the qualified, or ultimate, and field system CFRP properties is potentially a viable concept and validating these factors through witness specimen testing, as required in the more-recent SEs, is essential. However, it is currently uncertain if the same tensile strength and tensile modulus adjustment factors are appropriate for other properties such as shear, bond, or flexural strength, or flexural modulus. It is also uncertain if these factors change with exposure time within the service environment. A stronger technical basis, which could be obtained through additional research, should be provided to demonstrate the more universal applicability of this approach if future proposed alternatives utilize this concept.

While the implications of the expected gradual cure temperature variability over the length of a CFRP system are not a principal technical concern, additional research could be considered to quantify these effects and confirm this expectation.

Initial evidence demonstrates that CFRP retains its properties in representative environments for approximately 10,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> at the maximum service temperature as long as the glass transition temperature is greater than the maximum service temperature.

However, more research of field-representative CFRP systems in representative nuclear service environments should be considered to both decrease the uncertainty associated with the high-temperature, long-term (i.e., > 10,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br />) performance and durability of these systems, and evaluate representative conditions, such as abrasion from ingested particulate, that have not been previously evaluated. Alternatively, periodic inspection and surveillance testing to evaluate the performance of installed CFRP systems could be used to demonstrate that acceptable design margins remain over the intended service life.

A holistic evaluation of the entire QA\\QC program is warranted to identify the most important attributes and potential programmatic weaknesses. Any identified potential weaknesses may require development of a stronger supporting technical basis or modification of the existing QA/QC program to correct identified deficiencies. One important thrust of this evaluation should be the identification, control, and validation of critical variables as part of the installation and examination procedures. Another focal point should be a review of the qualification program(s) for installers and system inspectors.

The complexity and novelty of the CFRP installation makes it challenging for NRC inspectors to identify the critical variables that govern system performance and thus require the most scrutiny during installation. Therefore, it is strongly recommended that enhanced inspection guidance be developed for inspectors after completing the aforementioned QA\\QC program review. This guidance should contain a checklist, or summary, of the critical installation variables and their allowable ranges, and identify the most important aspects of the associated QA/QC program that an inspector should review. This summary should be supported by an assessment of the implications when allowable critical variable ranges are exceeded or when important aspects of the QA/QC program are violated. This assessment would help inspectors understand implications of any specific non-compliances issues that may arise.

The implications of expected property variability (e.g., glass transition temperature, tensile strength, degree of cure) due to installation (e.g., orientation) and environmental factors (e.g., humidity and cure temperature variability) over the extent of the repair should be addressed. One approach would be to conduct additional confirmatory research or qualification testing to address implications of these variables on CFRP system performance. Alternatively, enhanced quality assurance measures could be developed to validate that the CFRP repair at the most limiting locations remains acceptable. Enhanced quality assurance measures would require that samples be extracted from both critical locations, such as near the termination ends and other locations with minimum design margins, and locations within the CFRP system that are most likely to violate the acceptance criteria. These later locations are those subjected to the lowest curing temperatures and those with restricted accessibility or that require challenging installation orientations. Full-repair-thickness core samples should be extracted from these locations so that properties of all the plies can be evaluated.

The basis and acceptability of using the tensile bond strength criterion to serve as a surrogate for the shear bond strength design requirement should be more clearly

established. If no basis currently exists, qualification testing should establish an appropriate basis. Alternatively, a more direct shear bond strength test could be performed on the installed system as part of the quality assurance program. Appropriate test standards or tests of representative witness specimens manufactured as part of the installation procedure will be needed to support such field testing.

Witness specimen testing plays a critical role in demonstrating the acceptability of the repair, and the representativeness of these specimens to the installed repair is therefore paramount. However, a strong basis demonstrating how the witness specimen fabrication procedures are representative of potentially limiting locations within the repair has not been articulated. This basis could be provided through qualification testing or additional research to quantify the effects of allowable variations in critical variables (e.g., temperature, humidity, installer accessibility, installation orientation) and the ability of the current witness specimen fabrication procedures to discriminate between unacceptable variations in these variables.

Alternatively, witness specimen fabrication, handling, storage, and testing provisions could be altered to ensure that the witness specimens properties are representative of those installation locations which are expected to have the limiting properties.

Future CFRP proposed alternatives, should more explicitly consider plant-specific risk implications, at least qualitatively, to understand potential consequences of CFRP failures within the systems implementing this repair method. These analyses should address the possibility for both global common-cause failures which become more likely as the nominal glass transition temperature approaches the service or maximum accident temperatures and more localized failures in regions with safety factors below those classically used for brittle materials in nuclear applications.

A more rigorous plant-specific PRA evaluation is warranted for the ANO, Surry, STP, and Brunswick plants to better understand the risk implications of approved CFRP repairs. The evaluation should consider the specific piping systems and locations within those systems where CFRP will be installed and perform an initial Failure Mode Effects & Criticality Analysis to better understand possible CFRP failure implications on downstream components, and then adjust conditional failure probabilities of these components accordingly. For conservativism, volumes of CFRP material which do not have a margin of 40oF between Tg and Tmax could be assumed to be failed during an event as well as localized areas which do not have an FSeff margin of 10. Such analysis would naturally consider possible common-cause effects. The outcome of these analyses should be used to determine if any compensatory regulatory actions are warranted for these plants.

Appendices A. List of Documents Reviewed (References)

I.M. Daniel and O. Ishai, Engineering Mechanics of Composite Materials, 2nd Edition, Oxford University Press, January 2006.

W.D. Callister, Jr., Fundamentals of Materials Science and Engineering: An Integrated Approach, 2nd edition, Wiley, 2004.

J. Gotro, Practical Tips for Curing Thermosets Part Six: The Glass Transition Temperature is Your Friend, Polymer Innovation Blog, https://polymerinnovationblog.com/practical-tips-curing-thermosets-part-six-glass-transition-temperature-friend/, October 17, 2016.

L. Horath, Fundamentals of Materials Science for Technologists: Properties, Testing, and Laboratory Exercises, 2nd Edition, Waveland Press Inc., 2017.

Proposed Alternative to ASME Section XI Requirements for Repair/Replacement of Circulating and Service Water Class 3 Buried Piping in Accordance with 10 CFR 50.55a(z)(1), Enclosure 1A to letter from Sartain (Virginia Electric and Power Company), dated December 14, 2016 (ML16355A346).

Response to Request for Additional Information, Alternative Request to Use Carbon Fiber Reinforced Polymer System for the Internal Repair of Buried Circulating and Service Water Piping, Enclosure 1 17-296, Letter from Sartain (Virginia Electric and Power Company) to U.S.

Nuclear Regulatory Commission, dated August 31, 2017 (PROPRIETARY) (ML17251A902).

Safety Evaluation by the Office of Nuclear Reactor Regulation Proposed Alternative to ASME Code,Section XI, Requirements for Repair/Replacement of Buried Circulating and Service Water Piping Virginia Electric and Power Company (Dominion) Surry power Station, Units 1 and 2, Docket Nos. 50-280 and 50-281, Enclosure 1 to Letter from Markley (NRC) to Stoddard (Virginia Electric and Power Company), dated December 20, 2017 (PROPRIETARY)

(ML17303A037).

Proposed Alternative to ASME Boiler & Pressure Vessel Code Section Xl Requirements for Repair/Replacement of Essential Cooling Water (ECW) Class 3 Buried Piping in Accordance with 10CFR50.55a(z)(1), Enclosure 1 to letter from Dunn (South Texas Project Electric Generating Station) to U.S. Nuclear Regulatory Commission, dated September 26, 2019 (PROPRIETARY) (ML19274C393).

Response to Request for Additional Information, Enclosure 1 to letter from Georgeson (South Texas Project Electric Generating Station) to U.S. Nuclear Regulatory Commission, dated July 15, 2020 (PROPRIETARY) (ML20197A262).

Supplemental Information, Enclosure 1 to letter from Georgeson (South Texas Project Electric Generating Station) to U.S. Nuclear Regulatory Commission, dated July 30, 2020 (PROPRIETARY) (ML20212L569).

Safety Evaluation by the Office of Nuclear Reactor Regulation, Request RR-ENG-3-24, Alternative Repair of Essential Cooling Water Piping, South Texas Project Electric Generation Station, Units 1 and 2, STP Nuclear Operating Company, Docket Numbers 50-498 and 50-499, (PROPRIETARY) (ML20227A383).

Proposed Alternative to ASME Section XI Requirements for Repair/Replacement of ECP Supply Piping in accordance with 10 CFR 50.55a(z)(1), Enclosure 1 to letter from Gaston (Entergy) to U.S. Nuclear Regulatory Commission, dated July 15, 2020, (PROPRIETARY) (ML20218A673)

Response to Final Supplemental Request for Additional Information related to Proposed Alternative to ASME Section XI Requirements for Repair/Replacement of ECP Supply Piping in accordance with 10 CFR 50.55a(z)(1), Enclosure 1 to letter from Gaston (Entergy) to U.S.

Nuclear Regulatory Commission, dated September 14, 2021, (PROPRIETARY)

(ML21257A456).

Arkansas Nuclear One, Units 1 and 2 - Approval of Request for Alternative from Certain Requirements of the American Society of Mechanical Engineers Boiler and Pressure Vessel Code (EPID L-2020-LLR-0104), (PROPRIETARY) (ML21265A255).

Request for Relief RA-20-0353 for a Proposed Alternative to ASME Boiler & Pressure Vessel Code Section XI Requirements for Repair/Replacement of Service Water System Piping in Accordance with 10 CFR 50.55a(z)(1), Enclosure 1 to letter from Krakuszeski (Duke Energy) to U.S. Nuclear Regulatory Commission, dated February 24, 2021 (PROPRIETARY)

(ML21055A798).

Response to Request for Additional Information Proposed Alternative to ASME Section XI Requirements for Repair/Replacement of Buried Service Water Piping, Enclosure 1 to letter from Krakuszeski (Duke Energy) to U.S. Nuclear Regulatory Commission, dated October 4, 2021 (PROPRIETARY) (ML21277A307).

Brunswick Steam Electric Plant, Units 1 and 2 - Authorization and Safety Evaluation for Alternative from Certain Requirements of the American Society of Mechanical Engineers Boiler and Pressure Bessel Code of Buried Service Water Piping (EPID L-2021-LLR-0014),

(PROPRIETARY) (ML21343A197).

M. Uddin, S. Pothana, F. Orth, Y. Hioe, and P. Krishnaswamy, Technical Assistance to Review American Society of Mechanical Engineers Boiler and Pressure Vessel Code Case and Licensee Relief Requests for Carbon Fiber Repairs, Task-3: Mockup Testing and Physical Evaluation of Proposed Carbon Fiber Reinforced Polymer Repair, U.S. Nuclear Regulatory Commission Technical Letter Report, TLR-RES/DE/CIB-2021-06, 2021 (ML21126A334).

S. Pothana, F. Orth, M. Uddin, P. Krishnaswamy, and Y. Hioe, Evaluation of Factors Affecting Durability of Carbon Fiber Reinforced Polymer Composite Repairs - Single-Ply Laminate Study, U.S. Nuclear Regulatory Commission Technical Letter Report, TLR-RES/DE/REB-2022-02, April 2022 (ML22118A351).

C. Nove, M. Bisbee, and B. Dowling, White Paper: Nondestructive Examination for Carbon Fiber Reinforced Polymer Composites, December 2020 (ML21015A166).

Code of Federal Regulations, Title 10, Energy, Part 50.61, Fracture toughness requirements for protection against pressurized thermal shock events.

Code of Federal Regulations, Title 10, Energy, Appendix G to Part 50 Fracture Toughness Requirements.

ASTM E1640-18, Standard Test Method for Assignment of the Glass Transition Temperature by Dynamic Mechanical Analysis, ASTM International, West Conshohocken, PA, United States.

ASTM E1356-08 (Reapproved 2014), Standard Test Method for Assignment of the Glass Transition Temperatures by Differential Scanning Calorimetry, ASTM International, West Conshohocken, PA, United States.

ASME PCC-22015, Repair of Pressure Equipment and Piping, The American Society of Mechanical Engineers, ASME International, ASTM D3528-96(2016), Standard Test Method for Strength Properties of Double Lap Shear Adhesive Joints by Tension Loading, ASTM International, West Conshohocken, PA, United States.

ASTM D7616/D7616M-11, Standard Test Method for Determining Apparent Overlap Slice Shear Strength Properties of Wet Lay-Up Fiber-Reinforced Polymer Matrix Composites Used for Strengthening Civil Structures, ASTM International, West Conshohocken, PA, United States.

ASME PCC-22015, Repair of Pressure Equipment and Piping, The American Society of Mechanical Engineers, ASME International, Three Mile Island Unit 1, Cooling Tower Water Distribution Piping Failure Causes Plant Derate, IRIS Experience Report #309745, 2014-02-17, 5:05PM Susquehanna Unit 2, Forced Down Power >20% Scheduled <10 Days Ahead due to Failure of Fittings in Heat Rejection System Butterfly Valve 215090, IRIS Experience Report #240613 2009-12-01, 2:00 PM Entergy Condition Report CR-ANO-2-2018-02664, and attached report, 10/05/2018.

Entergy Condition Report CR-ANO-2-2020-01185, and attached report, 03/22/2020.

Entergy Condition Report CR-ANO-2-2021-02187, and attached report, 10/06/2021.

Entergy Condition Report CR-ANO-2-2021-02472, and attached report, 10/12/2021.

Entergy Condition Report CR-ANO-2-2023-00809, and attached report, 04/19/2023.

Entergy Condition Report CR-ANO-2-2023-00903, and attached report, 04/22/2023.

Regulatory Guide 1.174, Revision 3, An Approach for Using Probabilistic Risk Assessment in Risk-Informed Decisions on Plant-Specific Changes to the Licensing Basis, U.S. Nuclear Regulatory Commission, January 2018.

C. Ng, On Assessing the Increase In Risk Related To The Potential Failure Of The Piping Repaired With Carbon Fiber Reinforced Polymer, U.S. Nuclear Regulatory Commission Technical Letter Report, TLR-RES/DE/REB-2021-11, September 2021 (ML21265A040).

Revised Risk-Informed Inservice Inspection Evaluation Procedure (PWRMRP-05-A), TR-112657 Rev. B-A, Electric Power Research Institute, Palo Alto, CA., January 12, 2000 (ML013470102).

Risk Assessment of Operational Events Handbook, Volume 1 - Internal Events, Revision 2.02, December 2017 (ML17348A149).

Smith, C. L., Knudsen, J. K., and McCabe, P. H., Standardized Plant Analysis Risk (SPAR)

Model and SAPHIRE, Version 8, Common-Cause Failure User Handbook, INL/LTD-12-24727, May 2012.

B. List of People Interviewed Chakrapani Basavaraju Michelle Gonzalez Ching Ng Seung Min Ali Rezai Laura Smith John Tsao Jeffrey Wood

Document 4: DPO Decision

November 6, 2024 MEMORANDUM TO:

Laura A. Smith, General Engineer EMBARK Venture Studio Office of Nuclear Reactor Regulation FROM:

Andrea D. Veil, Director Office of Nuclear Reactor Regulation

SUBJECT:

DIFFERING PROFESSIONAL OPINION ON INAPPROPRIATE APPROVAL OF CARBON FIBER REINFORCED POLYMER TO REPAIR DEGRADED CLASS 3 SAFETY-RELATED PIPING (DPO-2022-002)

The purpose of the memorandum is to respond to your differing professional opinion (DPO) submitted on September 6, 2022, in accordance with Management Directive 10.159, The Nuclear Regulatory Commission Differing Professional Opinions Program (Agencywide Documents Access and Management System (ADAMS) Accession No. ML15132A664).

DPO-2022-002, (ADAMS Accession No. ML23011A255), which documents your concerns with the staffs approval for use of a carbon fiber reinforced polymer (CFRP) composite for repairing degraded American Society of Mechanical Engineers (ASME) Class 3 safety-related piping at the Surry (ADAMS Accession No. ML17303A037), Brunswick (ADAMS Accession No. ML21343A197), South Texas Project (STP) (ADAMS Accession No. ML20227A383), and Arkansas Nuclear One (ANO) (ADAMS Accession No. ML21265A255) nuclear power plants.

I commend you for your commitment and dedication to the Nuclear Regulatory Commissions mission. Your willingness to raise concerns with your colleagues and managers, and to ensure that your concerns were heard is admirable and vital to ensure a healthy safety culture within the Agency. I also want to thank you for the time you dedicated to put together all the information regarding the DPO submittal on an accessible SharePoint Site and for having additional discussions with me and my staff to ensure that I fully understood your perspectives on this complex issue.

My response to the DPO is described in the enclosure.

Enclosure:

Directors Decision for Differing Professional Opinion CONTACT: Annie Ramirez, NRR (301) 415-6780

Enclosure DIRECTORS DECISION FOR DIFFERING PROFESSIONAL OPINION INAPPROPRIATE APPROVAL OF CARBON FIBER REINFORCED POLYMER TO REPAIR DEGRADED CLASS 3 SAFETY-RELATED PIPING (DPO-2022-002)

=

Background===

In the Differing Professional Opinion (DPO) 2022-002, you expressed concerns with the inappropriate approval of carbon fiber reinforced polymer (CFRP) to repair degraded Class 3 safety-related piping. The specific concerns that the DPO addressed include the safety evaluations for Surry Power Station Units 1 and 21, South Texas Project Units 1 and 22, Arkansas Nuclear One Units 1 and 23, and Brunswick Steam Electric Plant Units 1 and 24. In all four applications, the license proposed to use CFRP to repair degraded piping. The intention is to use the CFRP to repair degraded pipes in order to maintain the intended design function of the pipe.

Further, in the attachment to your submittal you stated that given the considerable departure between the CFRP systems from the metallic piping repairs; and the fact that the use of CFRP applications is fairly new at nuclear power plants, the design safety function of safety related piping needs to demonstrate adequate assurance of safety on key areas. These areas include understanding of the thermoset polymer matrix behavior at elevated temperatures, adequate controls for field installation conditions, validation for the use of cure temperature adjustment factors, varying cure temperatures between the terminal ends and the length of the repair resulting in varying glass temperature (Tg) along the entire length of the repair, and acceptance criterion for cure determination. In addition, you stated that without adequate field installation controls and a more complete understanding of how thermoset polymers are cured for use in fiber reinforced polymer applications in all service conditions, there is no assurance of safety.

Specifically, you concluded that there is significant risk associated with CFRP failure within the systems approved for this repair method.

1 Surry Power Station, Unit Nos. 1 and 2 - Relief from the Requirements of the ASME Code, ML17303A037, December 20, 2017 2 South Texas Project, Units 1 and 2 - Proposed Alternative RR-Eng-3-24 To ASME Code Requirements for The Repair of Essential Cooling Water System Class 3 Buried Piping, ML20227A383 September 3, 2020 3 Arkansas Nuclear One, Units 1 and 2 - Approval of Request for Alternative from Certain Requirements of The American Society of Mechanical Engineers Boiler and Pressure Vessel Code, ML21265A255, September 20, 2021 4 Brunswick Steam Electric Plant, Units 1 and 2 - Authorization and Safety Evaluation for Alternative from Certain Requirements of the American Society of Mechanical Engineers Boiler and Pressure Vessel Code of Buried Service Water Piping, ML21343A197, December 20, 2021 The DPO review panel (the panel) issued their report to me on March 4,2024, after reviewing the applicable documents, conducting internal interviews with relevant individuals, conducting independent analysis, and completing their deliberations. On March 26, 2024, I discussed the panel report with the DPO panel Chair and the members of the panel. Further on June 11, 2024, I met with the panel to further understand the recommendations given on the report.

To inform my decision regarding your DPO, I reviewed your DPO submittal and the panels report. To better understand your concerns, I assigned a Technical Assistant (TA) from my office to assist in the evaluation and documentation of my decision. My TA gathered information through discussions with you, the DPO panel, and other knowledgeable staff and reviewed documents pertinent to your DPO submittal. Further I tasked a team of risk analysts from the Division of Risk Assessment (DRA) in NRR to perform additional independent analyses5 on the consequences of potential failures of installed CFRP in safety-related piping at the Surry, Brunswick, South Texas Project (STP), and Arkansas Nuclear One (ANO) power plants. The staff used the best available information to conduct this analysis. The analysis examined both the consequences of the failures as well as their significance based on the extent of the failure.

In addition, the analysts evaluated operational considerations to determine how these plants would mitigate such failures. The staff performed an independent evaluation using NRCs Standardized Plant Analysis Risk (SPAR) models for these facilities to develop a consequence upper bound to quantify the risk for each plant that has been approved. Further, the staff used the guidance in LIC-504 " Integrated Risk Informed Decision making Process for Emergent Issues which contains useful information to determine whether additional regulatory action is required to place or maintain the plant in a safe condition. The guidance was used to benchmark the risk level for each approval. The information collected provided independent insights and perspectives that were vital to reaching my conclusion.

Summary of Issues You and the panel agreed on a summary of issues identified in the DPO submittal that accurately addressed your concerns. This summary of issues can be found on pages 2 to 3 of the panels report. The panel determined that all the concerns could be grouped into three areas for evaluation. My assessment of the panel conclusions will follow the format delineated by the panel.

My Assessment of the Panels Conclusions The panel performed a thorough evaluation of the DPO submittal, and related technical areas.

Their report is a good source of background and reference information, it was well written, and provided thoughtful conclusions based on independent analysis. It also documented their significant and valuable effort. Based on the review of the panels report I did not identify anything that directly challenged the adequacy of the approved safety evaluations (SEs). The report identified suggested areas where we could strengthen the CFRP technical insights, conduct research, and enhance oversight, therefore, I tasked two risk assessment analysts to perform an in-depth risk analysis of the four approved SEs. The panel summarized your concerns in three categories, which integrate with each other. The next several paragraphs outline a discussion of each issue, and the basis for my decisions using both the panel report and the independent analysis.

5 Issue 1: Adequacy of Technical Basis The panel summarized Issue 1 as follows.

Based on the evaluation of your DPO, and the DPO panel report the most relevant technical criteria you include under the technical basis are the margin between the field glass temperature (Tgf) and the maximum temperature (Tmax) for the CFRP system; effective safety factors (FSeff);

validation of the cure temperature adjustment factor (Ceff); the CFRP performance at elevated temperature and its durability for an extended period of time, and the unique failures of CFRP systems compared to metallic materials.

Through the DPO panel evaluation of the first category of concern and the statements provided in your DPO submittal it is clear that the glass transition temperature (Tg) for CFRP is a key criterion to ensure desirable mechanical properties, and it is consequently affected by the cure temperature (Tcure) among other factors. However, there are no NRC-approved ASME code requirements for the use of CFRP. Efforts to develop an applicable code case are ongoing but given the complexity of the material and the lack of operational experience, progress has been slow. The DPO panel report highlighted the following points:

1) Although there is no accepted margin between Tmax and Tgf in an ASME nuclear code requirement for CFRP, there should be an acceptable margin between the two, given that it exists for metallic materials. The panel also agreed that a good starting point for this margin is a Tgf margin of 40oF above Tmax for future applications.

2) The panel also agreed on the use of FSeff for the Allowable Stress Design (ASD) methodology that is consistent with the design margins for brittle materials.

3) The panel stated that using a cure temperature adjustment factor to account for differences between the qualified, or ultimate, and field system CFRP properties is potentially a viable concept, and that this concept should be validated through witness specimen testing. However, a stronger technical basis is necessary to demonstrate the more universal applicability of this approach if future proposed alternatives utilize this concept.

4) While independent testing provides some initial evidence that CFRP retains its properties in representative environments for an extended period, more research is warranted to reduce the uncertainty associated with the high-temperature, long-term performance of these systems.

5) The panel stated that delamination, debonding, and deformation due to exceeding Tgf are all unique failure modes compared to traditional nuclear metallic materials.

These failure modes warrant special attention during design, installation, testing, and post-installation inspection to guard against failure under operating and design-basis conditions.

It is worth noting that each application approval has evolved since approval of the first submittal, and with the collected operational experience and knowledge, they have each provided more information on the key points highlighted above. There are notable changes in the level of detail provided in the first approval (Surry) as compared to the last approval (Brunswick).

testing, training and qualification of CFRP systems installers and inspectors on site) which may be important to incorporate into our inspection program guidance. This guidance could provide the appropriate focus during a CFRP installation and identify some of the critical steps/variables that inspectors should focus on during the installation and testing of CFRP materials. Therefore, I am tasking the NRR staff to conduct an inspection program assessment.

Further assessment of the oversight program in combination with further research on CFRP system (Issue 1) could result in enhanced guidance, more consistency in oversight of installations, and may also provide insights on other physical testing considerations to qualify the repairs.

Issue 3: Risk Significance The panel summarized the third issue as follows: There is significant risk associated with CFRP failure within the systems approved for this repair method, especially when the potential for common-cause failures is considered.

Through the evaluation of Issue 3 in the DPO report and the submittal, I have noted the concern associated with risk significance of CFRP system repair and the importance of its reliability for service water and component cooling water systems. Both the DPO submitter analysis and the panel analysis identified a high risk associated with any of CFRP repairs failing. In addition, the panel conducted a supplemental analysis for plant specific cases and broader risk analysis to determine the risk associated with CFRP failures in many plants6. Given that the variability on the results could be due to the different assumptions used to conduct each risk analysis and the panel recommendation #12 to perform a more rigorous plant-specific probabilistic risk analysis (PRA) evaluation for the ANO, Surry, STP, and Brunswick plants to better understand the risk implications of approved CFRP repairs, I tasked an independent group of risk analysts from NRR/DRA to perform a detailed risk analysis for ANO, Surry, STP, and Brunswick.

The DPO studied the impact of failing certain components in each plant using PRA risk importance measures such as Birnbaum and Risk Increase Ratio (RIR). The panel studied the conditional core damage probability (CCDP) of certain components and converted them to a delta core damage frequency (CDF). While both approaches have their merits for identifying the relative importance of components, they are not a substitute for a more detailed risk analysis.

The independent DRA risk analysts used the SAPHIRE PRA software and the agencys plant-specific SPAR models to evaluate the potential plant impact using realistic operational considerations. Additionally, the analysts used plant drawings, system descriptions, and abnormal operating procedures (AOPs) for each plant to understand how specific site design specifications could mitigate potential failures.

The independent risk analysts found that using risk importance metrics such as Birnbaum and RIR without additional context does not correctly characterize the risk for the site. For instance, a metric like Birnbaum is one of the measures to characterize the importance of a component and is approximately the estimate of the annual change in CDF risk if the component was removed from service for a year. The Birnbaum value calculated for a component assumes that 6 This broader risk analysis is described in the report titled On Assessing the Increase In Risk Related To The Potential Failure Of The Piping Repaired With Carbon Fiber Reinforced Polymer (ADAMS ML21265A040).

Response to Recommendations I appreciate the panels thorough assessment of the concerns raised in the DPO as well as your perspectives. The panel report highlighted a total of 12 recommendations for consideration by NRC management. The recommendations and my responses to the panels recommendations are provided below.

Panel Recommendation #1 The panel recommends that future proposed alternatives should incorporate more margin between the design-basis temperature requirements and the CFRP glass transition temperature and for risk significant applications, a margin of 40 oF between the maximum service temperature and the field glass transition temperature is a recommended starting point.

I disagree with the panel recommendation #1 that future proposed alternatives should incorporate a set margin between the design-basis temperature requirements and the CFRP glass transition temperature for high and medium risk significant applications. Given that there are currently no ASME nuclear code requirements for CFRP, no consensus exists yet on accepted margin between the maximum service temperature and the glass temperature for the subject CFRP system. While 40 oF between the maximum service temperature and the field glass transition temperature could be a good starting point, it may create unnecessary conservatism for regulatory purposes. Plant-specific margins will vary depending on pipe locations where carbon fiber is applied and what systems may be impacted. The independent risk analysis shows that while all four approved SEs had different margins and risk estimates varied across the plants, it is prudent to say that a quantitative and qualitative risk assessment provides added assurance that the system in question will perform its intended safety function without undue risk to the public. That said, I recognize uncertainties in the risk analysis exist and support further work as reflected in this decision. Nevertheless, I found that the detailed NRC risk analysis of the sites provides more comprehensive understanding of the potential safety significance than assigning an overall average temperature delta. Consistent with NRCs risk-informed, performance based regulatory policies, the panels recommendation #11 is a more appropriate approach than setting a single fixed value for margin as a starting point. Other risk-informed and performance-based frameworks may be potentially suitable for assessing future CFRP repair requests such as ASME Code Case N-752 that have evolved since the first request to use CFRP at the Surry plant. However, the staff would need to assess such approaches if proposed by a licensee.

Panel Recommendation #2 The panel stated that effective factors of safety consistent with the allowable stress design margins provided in the ASME Code for brittle materials (i.e., FSeff = 10 for normal loading) are appropriate starting points that are consistent with other design standards for CFRP systems.

However, as additional nuclear experience is gained on CFRP repairs, it may be possible to reduce these load margins.

I disagree with panel recommendation #2 to use FSeff = 10 as a starting point for an effective factor of safety for future applications while it is consistent with the allowable stress design margins provided in the ASME Code for brittle material, there still some nuances on the validation of these values. Conversely, as we gather additional experience on CFRP repairs via additional research and/or relevant nuclear operating experience we can supplement and validate the actual desirable value of these effective factor of safety margins. Research can be incorporated into the statement of work for the user need mentioned on the response to recommendation #4. In the meantime, confirmatory PRA using the NRC SPAR model for a specific site could support decision making or other risk informed request such as ASME Code Case N-752 could be used for future reviews pending NRC assessment of the user need request.

Panel Recommendation #3 The panel stated that using a cure temperature adjustment factor to account for differences between the qualified or ultimate, and field system CFRP properties is potentially a viable concept and validating these factors through witness specimen testing, as required in the more-recent SEs, is essential. A stronger technical basis should be provided to demonstrate the more universal applicability of this approach if future proposed alternatives utilize this concept.

I agree with the panels conclusion regarding recommendation #3. Specifically, using a cure temperature adjustment factor to account for differences between the qualified, or ultimate, and field system CFRP properties is potentially a viable concept. I also agree with the submitter that there is not enough information currently to support this methodology. Therefore, we can rely on the validation of this methodology via witness specimen testing and relevant operational experience to date. It is worth noting that three out of the four approved SEs used the cure temperature adjustment factor. While it might be an adequate factor to consider, other aspects can be further examined and validated. All except for Surry accounted for this approach but note that Surrys core damage frequency per year remained below 1x10-3 which is consistent with what was observed on the other sites. Hence, in accordance with LIC-504 guidance, no immediate regulatory action for Surry nor any of the other sites is warranted. However, research for the validation of these temperature factors can be incorporated into the statement of work for the user need mentioned on the response to recommendation #4; such information can be used to further refine plant specific PRA SPAR models.

Panel Recommendation #4 The panel recommended that additional research could be considered to identify the effects of Tcure over time (i.e., aging, environmental degradation).

I agree with the panel recommendation #4. It is my understanding that there is a draft user need request which proposes research on CFRP properties and the suitability of repairing ASME Class 3 piping. Pending review of the details, I support this user need request and recommend that the scope of work for the user need incorporate the appropriate areas that the panel recommended in their report.

Panel Recommendation #5 The panel recommended that alternatively, periodic inspection and surveillance testing to evaluate the performance of installed CFRP systems could be used to demonstrate that acceptable design margins remain over the intended service life.

I agree in part with panel recommendation #5. While additional surveillance testing to evaluate the performance of installed CFRP systems can be useful to demonstrate that design margins remain over the intended service life, more operating experience combined with research could provide insights on acceptable design margins therefore these insights can be gathered by additional research. There is no need to assign any additional inspections or surveillance testing to evaluate the performance of the installed CFRP. As determined by the independent PRA qualitative and quantitative analysis any disintegration or deterioration of CFRP installations that result in a small amount of debris caught in any of the downstream systems can be easily detected in advance and mitigative actions can be taken which can further inform operational experience.

Panel Recommendation #6 The panel recommends a holistic evaluation of the entire QA\\QC program to identify the most important attributes and potential programmatic weaknesses.

I partially agree with recommendation #6 to examine if there are any unique aspects of CFRP repairs (e.g., validation testing and witness specimen testing, training and qualification of CFRP systems installers and inspectors on site) which may be important to incorporate into our oversight program guidance. For this reason, I am tasking the NRC staff to perform an inspection program assessment that includes NRC oversight aspects on use of the CFRP at nuclear power plants.

Panel Recommendation #7 Given the complexity and novelty of the CFRP installations it could be challenging for an NRC inspector to identify the critical variables that govern the system performance. Therefore, the panel recommended that enhanced guidance should be developed for inspectors. The panel also noted that the guidance should contain a checklist, or summary, of the critical installation variables and their allowable ranges, and should identify the most important aspects of the associated QA/QC program that an inspector should review. The panel concluded that this summary should be supported by an assessment of the implications when allowable critical variable ranges are exceeded or when important aspects of the QA/QC program are not met.

This assessment would help inspectors understand implications of any specific non-compliance issues that may arise.

I agree with the panel recommendation #7. The applicability of guidance enhancement should be considered under the inspection program assessment discussed in recommendation #6. Any critical areas, variables and installation parameters identified as relevant for the use of polymers (i.e. CFRP) on nuclear applications should be captured and used to enhance existing or new guidance that will support inspectors during the installation or oversight of systems that rely on the polymeric mechanical properties. This guidance should consider the use of checklists, or summaries of the critical installation variables and their allowable ranges. The guidance should also identify the most important aspects of the associated QA/QC program that an inspector should review during the installation of CFRP systems.

Panel Recommendation #8 The panel stated that the implications of expected material properties vary (i.e., Tg) due to the installation (e.g., orientation), installation environmental factors (e.g., humidity and cure temperature variability) over the extent of the repair should be addressed. The panel recommended that one approach could be to conduct additional confirmatory research or qualification testing to address implications of these variables on CFRP system performance.

They also alternatively recommended that enhanced quality assurance measures could be developed to validate that the CFRP repairs at the most limiting locations remain acceptable.

I agree in part with the panel recommendation #8. Therefore, pending review of the details, I am supportive of an appropriately scoped user need request to conduct additional research on CFRP systems (see recommendation # 4). Further I am also tasking DRO to assess the inspection program and to identify any enhancements needed to support the inspectors during the CFRP systems installations at nuclear power plants (see recommendations #6 and #10).

The user need findings should inform the inspection assessments efforts.

Panel Recommendation #9 The panel stated that a clearer basis for acceptability of using the tensile bond strength criterion to serve as a surrogate for the shear bond strength design requirement should be developed.

They recommended that appropriate test standards or tests of representative witness specimens manufactured as part of the installation procedure will be needed to support such field testing.

I agree with the panel recommendation #9. The proposed user needs request statement of work that has been mentioned in this Directors Decision should include a section to evaluate and identify the basis and acceptability for the most appropriate test standard and the importance of representative specimens manufactured to support field installation and testing. Any research should be used to support guidance enhancements and to inform future installations from approved SEs and future reviews.

Panel Recommendation #10 The panel stated that witness specimen testing plays a critical role in demonstrating the acceptability of the repair. They emphasized that a strong basis demonstrating how the witness specimen fabrication procedures are representative of potentially limiting locations within the repair has not been articulated. The panel recommended that this basis could be provided through qualification testing or additional research to quantify the effects of allowable variations in critical variables (e.g., temperature, humidity, installer accessibility, installation orientation) and the ability of the current witness specimen fabrication procedures to discriminate between unacceptable variations in these variables.

I agree with the panel recommendation #10. I believe this recommendation can be included as part of the proposed user need request mentioned in the response to recommendation #4. The statement of work for the user need could include a task that will provide additional information on acceptable testing and a testing technique that demonstrates the adequacy of the repair by using a representative specimen when compared to what was installed. Acceptable testing parameters should include, but not be limited to, representative specimen fabrication processes to better understand the effect of critical variables (e.g., temperature, humidity, installer accessibility, installation orientation) and its effect on the mechanical properties of the repair.

This information could also support inspection guidance enhancements (see recommendation

1. 6) to identify parameters that are deviations from what is acceptable for the CFRP installation process.

Panel Recommendation #11 The panel recommended that future CFRP proposed alternatives, should more explicitly consider plant-specific risk implications, at least qualitatively, to understand potential consequences of CFRP failures within the systems implementing this repair method. The panel recommended that the analyses should address the possibility for both global common-cause failures which become more likely as the nominal glass transition temperature approaches the service or maximum accident temperatures and more localized failures in regions with safety factors below those classically used for brittle materials in nuclear applications.

I agree with the panel recommendation #11 that further analysis should be implemented to refine risk assessments for future applications. Specifically, staff should engage with stakeholders and licensees on CFRP applications to adequately identify and describe the risk implications for CFRP installation. Further, the staff should employ the use of risk-informed and performance-based frameworks that may be potentially suitable for assessing future CFRP repair requests such as those employed on ASME Code Case N-752 which have evolved since the first request to use CFRP at the Surry plant. For context, Code Case N-752 provides for risk-informed categorization as high safety significance (HSS) or low safety significance (LSS) consistent with the guidance in Regulatory Guide (RG) 1.174, Revision 3, An Approach for Using Probabilistic Risk Assessment in Risk-Informed Decisions on Plant-Specific Changes to the Licensing Basis, (ML17317A256). Further as stated on the section titled Issue 3 Risk Significance, I tasked NRR staff to find the most appropriate communication vehicle to convey NRC insights on the issue including those gained through the independent risk analysis. The stakeholder interaction is to provide clarity to all licensees who have been approved to installed CFRP or are considering requesting NRC approval to install CFRP. This effort may include, if appropriate, an NRC Information Notice since INs are an efficient means to communicate operating experiences and would be one of several ways to convey experience gained from industry implementation of new materials and techniques such as CFRP. As such, I expect the staffs engagement to (1) identify the key insights and observations learned from NRCs assessment of this case including risk insights from the independent analysis; and (2) point to any relevant CFRP operational experience. I expect this effort to be completed in a timely manner, and it may involve multiple Divisions (i.e. DNRL and DRA). As appropriate, these efforts may capture other relevant operating experience on code repairs to ASME Class 3 components. Specific tickets with schedules will be developed once we have more information on relevant and related activities such as the user need request and code repairs/activities.

Panel Recommendation #12 (Issue #3)

The panel recommended that a more rigorous plant-specific PRA evaluation is warranted for the ANO, Surry, STP, and Brunswick plants to better understand the risk implications of approved CFRP repairs. The panel recommended that evaluation should consider the specific piping systems and locations within those systems where CFRP will be installed and perform an initial Failure Mode Effects & Criticality Analysis to better understand possible CFRP failure implications on downstream components, and then adjust conditional failure probabilities of these components accordingly.

I agree with the panel recommendation #12 given that the results from the two-plant specific risk analysis performed to support the panels assessment of the adequacy of the CFRP repair alternatives, indicated that there was potential high risk associated with CFRP systems failing in a service water system. Therefore, as stated before, I tasked independent NRR DRA risk analysts to conduct specific, detailed qualitative and quantitative analyses for ANO, Surry, STP, and Brunswick using NRC SPAR models to better understand the risk implications associated with the CFRP repairs at these sites. The individual risk for each plant is described in detail in the section labeled Issue 3. While uncertainties remain regarding the performance of CFRP materials in nuclear power plants, none of the four sites were found to have a risk increase that would warrant any immediate actions, in accordance with LIC-504 guidance. However, it is evident that further information on CFRP performance could help refine future results and reduce the level of uncertainty. Findings from future research and CFRP operating experience should be used to refine plant specific SPAR models for NRC staff use in regulatory activities.

Concluding Remarks Your concerns and the details of your assessment were of notable technical merit and were well documented in the submittal. I want to thank you for raising this DPO and for your active participation in the process. I also want to thank the panel for their thoughtful assessment of the concerns and the additional evaluations conducted. Further I would like to thank the independent NRR DRA risk analysts that conducted the plant specific SPAR model quantitative and qualitative analysis to examine the level of risk for the approved SEs.

I have concluded that there is not an immediate safety concern with the approval for use of CFRP for repairing degraded ASME Class 3 safety-related piping at Surry, STP, ANO and Brunswick. However, I do agree with both the panel and submitter that additional research and operational experience can help enhance existing licensing and inspection programs and refine current risk analyses to further support innovative materials in a risk informed and performance-based manner. My decision is informed by Commission direction to apply risk informed principles to any regulatory decision, and the NRC Principles of Good Regulation. I commend you for your willingness to raise concerns, and for your dedication to the NRC Mission. The use of the DPO process is vital to foster a healthy, positive safety culture within our agency.

A summary of the DPO will be included in the Weekly Information Report (when the case is closed) to advise employees of the outcome.