ML18282A103

From kanterella
Jump to navigation Jump to search
NRC Comments on NEI 96-07, Appendix D, August 2018
ML18282A103
Person / Time
Issue date: 08/30/2018
From:
Division of Inspection and Regional Support
To:
Govan T
References
NEI 96-07
Download: ML18282A103 (67)


Text

NRC Comments on NEI 96-07, Appendix D On July 16, 2018, the Nuclear Energy Institute (NEI) submitted to the U.S. Nuclear Regulatory Commission (NRC), NEI 96-07, Appendix D, Revision 0f (ADAMS Accession No. ML18199A647). The enclosure to this cover page documents the NRC staffs comments (in comment bubbles) on this latest version of NEI 96-07, Appendix D. This document will be used to support a Category 2 public meeting with NEI to discuss these comments on August 30, 2018.

The NRC comments in this document are generally identified in two categories:

1. Correction:

(a) Comment discussed at the June 26, Category 2 public meeting with NEI.

Documentation of the meeting can be found in ADAMS under Accession No. ML18234A534. It is essential that these comments be addressed since, as written, they would result in situations of licensing uncertainty for both the licensees and inspectors.

(b) Comments that document instances in the guidance where there exists a potential incorrect interpretation of referenced guidance (e.g., NEI 96-07 or RIS 2002-22) or regulation (such as 10 CFR 50.59). It is essential that these comments be addressed since, as written, they would result in situations of licensing uncertainty for both the licensees and inspectors.

2. Enhancement - Comments that document potential points of uncertainty and could be clarified.

General Appendix D Comments:

1. As written, there is an insufficient level of incorporation of RIS 2002-22, Supplement 1 guidance in Appendix D to say that the Supplement and/or NEI 01-01 could be superseded by Appendix D.
2. The examples in Appendix D, as written, are not complete enough to guide a licensee in making its determination on screening. The example conclusions should not be written in the context of screens out but rather does not screen in for the aspect or topic within the section/subsection. This approach should be applied to all examples.

Because this document is used in conjunction with NEI 96-07, one can never reach a conclusion of screens out or does not need a license amendment but can only reach conclusions of does not screen in for this aspect or does not require a license amendment request for this aspect. For example, use of malfunction results as the basis of conclusions for the examples is not appropriate because malfunction results are not the only consideration; whether the change introduces a new malfunction is only one consideration. Yet, the examples in Appendix D use this as the sole basis for conclusion.

3. Prior to the issuance of the qualitative assessment guidance from RIS 2002-22, Supplement 1, the definition of engineering evaluations was universally understood to mean engineering/technical information supporting the change. However, since the

issuance of RIS 2002-22, Supplement 1 information, it is not clear from the NEI 96-07 draft whether engineering evaluations means engineering/technical information supporting the change or qualitative assessment This same comment is consistent for use of terms qualitative assessment and technical assessment.

4. Screening issues related to human-machine interface need to be resolved by a revision to Regulatory Guide 1.187 or a revision to NEI 96.07 rather than Appendix D.

Human-machine interface issues are not unique to digital I & C and should be addressed in the broader context of the RG 1.187 or NEI 96.07.

5. The evaluation section of Appendix D appears to introduce new expansions of or paraphrasing of general 10 CFR 50.59 guidance without a clear nexus or explanation on why this is necessary for digital-specific modification. Specifically, section 4.3.6 Discussion of "different result" are not specific to digital I & C and should be worked, if needed, as a revision to the RG 1.187/NEI 96-07 documents and not in an appendix with limited applicability. Most of this area of concern may be addressed by the qualitative assessment approach documented in RIS 2002-22, Supplement 1.

NEI 96-07, Appendix D Draft Revision 0f Nuclear Energy Institute SUPPLEMENTAL GUIDANCE FOR APPLICATION OF 10 CFR 50.59 TO DIGITAL MODIFICATIONS

July 2018 ACKNOWLEDGMENTS NEI would like to thank the NEI 01-01 Focus Team for developing this document.

Although everyone contributed to the development of this document, NEI would like to give special recognition to David Ramendick, who was instrumental in preparing this document.

NOTICE Neither NEI, nor any of its employees, members, supporting organizations, contractors, or consultants make any warranty, expressed or implied, or assume any legal responsibility for the accuracy or completeness of, or assume any liability for damages resulting from any use of, any information apparatus, methods, or process disclosed in this report or that such may not infringe privately owned rights.

EXECUTIVE

SUMMARY

NEI 96-07, Appendix D, Supplemental Guidance for Application of 10 CFR 50.59 to Digital Modifications, provides focused application of the 10 CFR 50.59 guidance contained in NEI 96-07, Revision 1, to activities involving digital modifications.

The main objective of this guidance is to provide all stakeholders a common framework and understanding of how to apply the 10 CFR 50.59 process to activities involving digital modifications.

The guidance in this appendix supersedes the 10 CFR 50.59-related guidance contained in NEI 01-01/ EPRI TR-102348, Guideline on Licensing of Digital Upgrades. Commented [A1]: Correction (a): Currently, there is an insufficient level of incorporation of Supplement 1 to RIS 2002-22 guidance into Appendix D to say that NEI 01-01 and Supplement 1 could be superseded by Appendix D. See below for more detailed comments.

Note from meeting:

NEI does not see RIS 2002-22, Suppl 1being sunset as a result of the issuance of Appendix D. Only need information from section 2.2 of the RIS to address qualitative assessment.

Context of the roadmap for digital modifications would be helpful in the executive summary.

Adding that the associated RIS and RIS Supplement can be added in an additional sentence to this paragraph.

This topic will be discussed further as we address some of the other red comments.

i

TABLE OF CONTENTS EXECUTIVE

SUMMARY

...................................................................................................................................... i 1 INTRODUCTION ........................................................................................................................................ 2

1.1 BACKGROUND

.................................................................................................. 2 1.2 PURPOSE ......................................................................................................... 3 1.3 10 CFR 50.59 PROCESS

SUMMARY

................................................................... 3 1.4 APPLICABILITY TO 10 CFR 72.48 ..................................................................... 3 1.5 CONTENT OF THIS GUIDANCE DOCUMENT ....................................................... 3 2 DEFENSE IN DEPTH DESIGN PHILOSOPHY AND 10 CFR 50.59 ................................................ 3 3 DEFINITIONS AND APPLICABILITY OF TERMS ............................................................................... 4 4 IMPLEMENTATION GUIDANCE............................................................................................................. 5 4.1 APPLICABILITY ................................................................................................ 5 4.2 SCREENING ..................................................................................................... 5 4.3 EVALUATION PROCESS................................................................................... 24 5.0 EXAMPLES ............................................................................................................................................... 62 D-1

1 INTRODUCTION There are specific considerations that should be addressed as part of the 10 CFR 50.59 process when performing 10 CFR 50.59 reviews for digital modifications. These specific considerations include different potential failure modes of digital equipment as opposed to the equipment being replaced, the effect of combining functions of previously separate devices (at the component level, at the system level, or at the "multi-system" level) into fewer devices or one device, and the potential for software common cause failure (software CCF ).

The format of this Appendix was aligned with NEI 96-07, Rev. 1 text for ease of use. As such, there will be sections where no additional guidance is provided.

1.1 BACKGROUND

Commented [A2]: Correction (a): This section should include a summary of Supplement 1 to RIS 2002-22. This version and subsequent versions of Appendix D will use the Licensees have a need to modify existing systems and components due to the guidance of Supplement 1 as a basis for guidance in this growing problems of obsolescence, difficulty in obtaining replacement parts, document. This is essential if NEI intends to eventually supersede NEI 01-01 and its supplement.

and increased maintenance costs. There also is great incentive to take advantage of modern digital technologies which offer potential performance and reliability improvements.

In 2002, a joint effort between the Electric Power Research Institute (EPRI) and the Nuclear Energy Institute (NEI) produced NEI 01-01, Revision 0 (also known as EPRI TR-102348, Revision 1), Guideline on Licensing Digital Upgrades: A Revision of EPRI TR-102348 to Reflect Changes to the 10 CFR 50.59 Rule, which was endorsed (with qualifications) by the Nuclear Regulatory Commission (NRC) in Regulatory Issue Summary (RIS) 2002-22.

Since the issuance of NEI 01-01 in 2002, digital modifications have become more prevalent. Application of the 10 CFR 50.59 guidance contained in NEI 01-01 has not been consistent or thorough across the industry, leading to NRC concern regarding uncertainty as to the effectiveness of NEI 01-01 and the need for clarity to ensure an appropriate level of rigor is being applied to a wide variety of activities involving digital modifications.

NEI 01-01 contained guidance for both the technical development and design of digital modifications as well as the application of 10 CFR 50.59 to those digital modifications. The NRC also identified this as an issue and stated that NEI could separate technical guidance from 10 CFR 50.59 related guidance.

D-2

1.2 PURPOSE Appendix D is intended to assist licensees in the performance of 10 CFR 50.59 reviews of activities involving digital modifications in a consistent and comprehensive manner. This assistance includes guidance for performing 10 Commented [A3]: Enhancement: Recommend inserting here or on Section 1.5 below to add clarity for use:

CFR 50.59 Screens and 10 CFR 50.59 Evaluations. This appendix does not NEI 96-07, Appendix D, Supplemental Guidance for include guidance regarding design requirements for digital activities. Application of 10 CFR 50.59 to Digital Modifications, does not alter, and unless explicitly noted, should not be The guidance in this appendix applies to 10 CFR 50.59 reviews for both interpreted differently than the guidance contained in NEI 96-07, Revision 1. Rather, Appendix D provides focused small-scale and large-scale digital modifications; from the simple guidance for the application of the 10 CFR 50.59 to replacement of an individual analog meter with a microprocessor-based activities involving digital modifications.

instrument, to a complete replacement of an analog reactor protection system Commented [A4]: Enhancement: Recommend inserting here or in Section 1.5 below to add clarity for use:

with an integrated digital system. Examples of activities considered to be a While this appendix does not include guidance regarding digital modification include computers, computer programs, data (and its design requirements for digital activities, as with other plant modifications, fundamental to digital presentation), embedded digital devices, software, firmware, hardware, the modifications is proper handling of key technical issues human-system interface, microprocessors and programmable digital devices during the design process before 10 CFR 50.59 is applied.

Appendix D describes three general types of failures that (e.g., Programmable Logic Controllers and Field Programmable Gate Arrays). are specific to digital equipment (e.g., software related, other failure not software related, and human-system This guidance is not limited to "stand-alone" instrumentation and control interface related) and the effects of these failures on the function of the system in which they are installed. The 10 systems. This guidance can also be applied to the digital aspects of CFR 50.59 process considers effects of a design change on a modifications or replacements of mechanical or electrical equipment if the UFSAR-described design function that have the potential to increase the likelihood of malfunctions, increase new equipment makes use of digital technology (e.g., a new HVAC design consequences, create new accidents or otherwise meet the that includes embedded microprocessors for control). 10 CFR 50.59 evaluation criteria in paragraph 50.59(c)(2).

In each section, Appendix D refers to specified 10 CFR Finally, this guidance is applicable to digital modifications involving safety- 50.59 guidance in NEI 96-07, Revision 1, and then provides guidance for applying NEI 96-07 as-written to address the related and non-safety-related systems and components and also covers effects on the design function of these general types of digital-to-digital activities (i.e., modifications or replacements of digital- failures that are specific to digital equipment.

based systems).

1.3 10 CFR 50.59 PROCESS

SUMMARY

No additional guidance is provided.

1.4 APPLICABILITY TO 10 CFR 72.48 No additional guidance is provided.

1.5 CONTENT OF THIS GUIDANCE DOCUMENT No additional guidance is provided.

2 DEFENSE IN DEPTH DESIGN PHILOSOPHY AND 10 CFR 50.59 No additional guidance is provided.

D-3

3 DEFINITIONS AND APPLICABILITY OF TERMS Definitions 3.1 through 3.14 are the same as those provided in NEI 96-07, Rev. 1. Definitions specific to this appendix are defined below.

3.15 Sufficiently Low Sufficiently low means much lower than the likelihood of failures that are considered in the UFSAR (e.g., single failures) and comparable to other common cause failures that are not considered in the UFSAR (e.g., design flaws, maintenance errors, and calibration errors).

D-4

4 IMPLEMENTATION GUIDANCE 4.1 APPLICABILITY No additional guidance is provided.

4.2 SCREENING CAUTION The guidance contained in this appendix is intended to supplement the generic Screen guidance contained in the main body in NEI 96-07, Section 4.2.

Namely, the generic Screen guidance provided in the main body of NEI 96-07 and the more-focused Screen guidance in this appendix BOTH apply to digital modifications.

Commented [A5]: Correction (b): As written the staff is not clear on the intent of the second sentence of the Introduction Caution statement. Replace second sentence of the above CAUTION with:

Unless specifically stated, the Screen guidance in this Throughout this section, references to the main body of NEI 96-07, Rev. 1 will appendix does not alter and should not be interpreted be identified as "NEI 96-07." differently than the generic Screen guidance provided in the main body of NEI 96-07. Rather, this appendix As stated in NEI 96-07, Section 4.2.1, the determination of the impact of a provides focused guidance for applying the guidance in the main body of NEI 96-07 to address the three general types proposed activity (i.e., adverse or not adverse) is based on the impact of the of failures that are specific to digital equipment (e.g.,

proposed activity on UFSAR-described design functions. To assist in software related, other failures not software related, and human-system interface related).

determining the impact of a digital modification on a UFSAR-described design function, the general guidance from NEI 96-07 will be supplemented with the digital-specific guidance in the topic areas identified below.

Digital-to-Digital Replacements and "Equivalency" In NEI 96-07, Section 4.2.1.1, equivalent replacements are discussed.

However, digital-to-digital changes may not necessarily be equivalent because the component/system behaviors, response time, failure modes, etc.

for the new component/system may be different from the old component/system. All non-equivalent digital-to-digital replacements should utilize the guidance provided in this Appendix.

Guidance Focus In the following sections and sub-sections that provide the Screen guidance unique to the application of 10 CFR 50.59 to digital modifications, each section and sub-section addresses only a specific aspect, sometimes at the deliberate exclusion of other related aspects.

This focused approach is intended to concentrate on the particular aspect of interest and does not imply that the other aspects do not apply or could not D-5

be related to the aspect being addressed. Initially, all aspects need to be considered, with the knowledge that some of them may be able to be excluded based on the actual scope of the digital modification being reviewed.

Example Focus Commented [A6]: Correction (a) - Since the examples do not address all aspects, the conclusion should not be Within this appendix, examples are provided to illustrate the guidance. screens out but rather does not screen in for the aspect or topic within the section/subsection. This approach Unless stated otherwise, a given example only addresses the aspect or topic should be applied to all examples. That is, since this within the section/sub-section in which it is included, sometimes at the document is used in conjunction with NEI 96-07, one can never reach a conclusion of screens out or does not need deliberate exclusion of other aspects or topics which, if considered, could a license amendment but can only reach conclusions of potentially change the Screen conclusion. does not screen in for this aspect or does not require a LAR for this aspect.

Human-System Interface Evaluations For NEI consideration:

(1) make it clear that the example only has one consideration that has been analyzed prior to documenting Similar to other technical evaluations (performed as part of the design the conclusion, or.

modification package), a human factors engineering (HFE) evaluation (2) make the guidance examples reflect all considerations prior to documenting a conclusion determines the impacts and outcomes of the change (e.g., personnel acts or omissions, as well as their likelihoods and effects). The licensing-based Commented [A7]: Correction (b): NEI 96-07 does not reviews (Screens and Evaluations) performed in accordance with 10 CFR contain this statement and this approach is inconsistent with NEI 96-07 and could lead to confusion for both 50.59 compare the impacts and new outcomes (i.e., post-modification) to the licensees and inspectors in implementation.

initial conditions and current outcomes (i.e., pre-modification) in order to Enhancement: In NEI 96-07, all formal Examples (i.e.,

determine the effect on design functions (in the Screen phase) and the need labeled Example 1, Example 2) reach a final decision for for a license amendment request (in the Evaluation phase). screening and evaluation (e.g., Screens In OR Screens Out; 50.59 criterion Met OR 50.59 Criterion not Met).

Otherwise, NEI 96-07 describes a specific aspect or topic 4.2.1 Is the Activity a Change to the Facility or Procedures as Described in the within the guidance that precedes the formal Example 1.

To illustrate:

UFSAR?

NEI 96-07 Section 4.1.6 states Certain malfunctions are Introduction not explicitly described in the UFSAR because their effects are bounded by other malfunctions that are described. For example, failure of a lube oil pump to supply oil to a A 10 CFR 50.59 Evaluation is required for digital modifications that component may not be explicitly described because a failure adversely affect design functions, or the methods used to perform or control of the supplied component to operate was described.

design functions. There is no regulatory requirement for a proposed activity Please delete the example focus section for the reasons involving a digital modification to default (i.e., be mandatorily "forced") to an noted above. After addressing general comment 3, this section will no longer be necessary.

adverse conclusion.

Commented [A8]: Correction (a) - deletion of this Although there may be adverse impacts on UFSAR-described design subsection as a global comment as indicated in NRC cover page.

functions due to the following types of activities involving a digital modification, these typical activities do not default to an adverse conclusion See Comment A28 simply because of the activities themselves.

  • The introduction of software or digital devices.
  • The replacement of software and/or digital devices with other software and/or digital devices.
  • The use of a digital processor to "calculate" a numerical value or "generate" a control signal using software in place of using analog components.

D-6

  • Replacement of hard controls (i.e., pushbuttons, knobs, switches, etc.) with a touch-screen to operate or control plant equipment.

Engineering/technical information should be documented (as part of the design process) to record the impacts from digital modifications. This engineering/technical information will be used as the basis/justification for the conclusion of adverse or not adverse.

Scope of Digital Modifications Generally, a digital modification may consist of three areas of activities: (1) software-related activities, (2) hardware-related activities and (3) Human-System Interface-related activities.

NEI 96-07, Section 4.2.1.1 provides guidance for activities that involve "...an SSC design function..." or a "...method of performing or controlling a design function..." and Section 4.2.1.2 provides guidance for activities that involve

"...how SSC design functions are performed or controlled (including changes to UFSAR-described procedures, assumed operator actions and response times)."

Based on this segmentation of activities, the software and hardware portions will be assessed within the "facility" Screen consideration since these aspects involve SSCs, SSC design functions, or the method of performing or controlling a design function and the Human-System Interface portion will be assessed within the "procedures" Screen consideration since this portion involves how SSCs are operated and controlled.

4.2.1.1 Screening of Changes to the Facility as Described in the UFSAR SCOPE In the determination of potential adverse impacts, the following aspects should be addressed in the response to this Screen consideration:

(a) Use of Software and Digital Devices (b) Combination of Components/Systems and/or Functions USE OF SOFTWARE AND DIGITAL DEVICES Commented [A9]: Enhancement: Since both subsections In NEI 96-07, Section 4.2.1, sub-section titled "Screening for Adverse Effects," (a) and (b) use the criteria of a new accident or malfunction, the second paragraph contains the following guidance: this text should be discussed within the scope section above.

"...changes that would introduce a new type of accident or malfunction Commented [A10]: Correction (a): Use of malfunction would screen in." [emphasis added] results should not be the only consideration for the basis of conclusions for determining an adverse impact in the examples.

Note that this Screen guidance does NOT address the "result(s)" of a new malfunction, which is the subject of Evaluation criterion (c)(2)(vi). Whether the change introduces a new malfunction is one consideration. Yet the examples use this as the only basis for conclusion.

D-7

For applications involving SSCs with design functions, digital modifications that introduce the exact same software into redundant trains or channels to perform a design function have the potential to create a new malfunction.

The potential to create a new malfunction comes from the possibility of a Commented [A11]: Correction (b): Change The to One.

This sentence states that the one way to introduce a new software CCF that did not previously exist. malfunction is from software CCF. It is very important to imply there may be other ways to introduce new For relatively simple digital modifications, engineering evaluations may be malfunctions.

used to show that the digital modification would not adversely affect design Commented [A12]: Correction (b): Reinsert the below functions; even for digital modifications that involve redundant guidance in that was deleted from the December 2017 version of Appendix D. This guidance is essential in that it components/systems because the reliability of performing design functions is addresses screening of software. Similar wording was also not decreased and no new malfunctions are created. included in RIS 2002-22, Supplement 1.

An adverse effect may also consist of the potential To reach a screen conclusion of not adverse for relatively simple digital marginal increase in the likelihood of SSC failure due to modifications, the degree of assurance needed to make that conclusion is the introduction of software. For redundant safety systems, this marginal increase in likelihood creates a based on considerations such as the following: similar marginal increase in the likelihood of a common failure in the redundant safety systems. On this basis,

  • Physical Characteristics of the Digital Modification most digital modifications to redundant safety systems are adverse. However, for some digital modifications, The change has a limited scope (e.g., replace analog transmitter with a engineering evaluations may show that the digital digital transmitter that drives an existing instrument loop) modification contains design attributes to eliminate consideration of a software common cause failure. In such cases, even when a digital modification involves redundant Uses a relatively simple digital architecture internally (simple process systems, the digital modification would not be adverse.

of acquiring one input signal, setting one output, and performing some Alternately, the use of different software in two or more redundant SSCs is not adverse due to a software common simple diagnostic checks) cause failure because there is no mechanism to increase in the likelihood of failure due to the introduction of Has limited functionality (e.g., transmitters used to drive signals for software.

parameters monitored)

Reference May 2017 version of Appendix D, ADAMS Can be comprehensively tested (but not necessarily 100 percent of all Accession Number ML17137A020.

combinations) Commented [A13]: Correction (a): Prior to the issuance of the qualitative assessment guidance from RIS 2002-22,

  • Engineering Evaluation Assessments Supplement 1, the definition of engineering evaluations was universally understood to mean engineering/technical information supporting the change. However, since the The quality of the design processes employed issuance of RIS 2002-22, Supplement 1 information, it is not clear from the NEI 96-07 draft whether engineering Single failures of the digital device are encompassed by existing evaluations means engineering/technical information failures of the analog device (e.g., no new digital communications supporting the change or qualitative assessment Please change all uses of engineering evaluations to one of these among devices that introduce possible new failure modes involving options.

separate devices) Commented [A14]: Correction (b): NRC has previously discussed this section with NEI and made suggested edits.

Has extensive applicable operating history Please change selected text to: because software CCF is not introduced to be consistent with previous discussion on The use of different software in two or more channels, trains or loops of SSCs this language.

is not adverse due to a software CCF because there is no mechanism to create Commented [A15]: Correction (b): Simplicity is not defined by the number of inputs and output - see example a new malfunction due to the introduction of software. 4-1 Insert (e.g.) at the beginning of parenthesis.

Some specific examples of activities that have the potential to cause an adverse effect include the following activities:

  • Addition or removal of a dead-band, or D-8
  • Replacement of instantaneous readings with time-averaged readings (or vice-versa).

In each of these specific examples, the impact on a design function associated with the stated condition needs to be assessed to determine the Screen conclusion (i.e., adverse or not adverse).

Example 4-1 illustrates the application of the guidance for a relatively simple digital modification.

Example 4-1. NO ADVERSE IMPACT on a Design Function for a Relatively Simple Digital Modification Proposed Activity Description Transmitters are used to drive signals for parameters monitored by redundant ESFAS channels. The original analog transmitters are to be replaced with microprocessor-based transmitters. The change is of limited scope since the existing 4-20 mA instrument loop is maintained for each channel without any changes other than replacing the transmitter itself.

The digital transmitters are used to drive signals of monitored parameters and thus have limited functionality with respect to the Engineered Safety Features Actuation System (ESFAS) design function.

Design Function Identification The ESFAS design function is the ability to respond to plant accidents.

Screen Response The digital transmitters use a relatively simple digital architecture internally in that the firmware in the new transmitters implements a simple process of Commented [A16]: Correction (a): This example is acquiring one input signal, setting one output, and performing some simple implicitly providing guidance that simplicity is a function of diagnostic checks. the number or inputs and outputs, which is only part of the guidance above.

Failures of the new digital device are encompassed by the failures of the Simplicity is not defined by, or implied by, the number of existing analog device in that there are no new digital communications inputs and outputs.

among devices that introduce possible new failure modes involving multiple Deleting the entire in that clause would resolve this devices. The engineering evaluation of the digital device concluded that the concern.

digital system is at least as reliable as the previous system, the conclusion of Commented [A17]: Correction (a): Please delete.

which is based on the quality of the design processes employed, and the operating history of the software and hardware used. In addition, based on The inclusion of in that creates a logical flaw. There are other ways to create new failure modes besides just digital the simplicity of the device (one input and one output), it was communications.

comprehensively tested. Further, substantial operating history has Commented [A18]: Correction (a): This example is demonstrated high reliability in applications similar to the ESFAS implicitly providing guidance that simplicity is a function of the number or inputs and outputs, which is only part of the application. guidance above.

Therefore, the proposed digital modification is not adverse because the digital Simplicity is not defined by the number of inputs and outputs.

modification is relatively simple and the assessment of the considerations Deleting the parenthetical clause would resolve this part of the concern.

D-9

identified above has determined that the reliability of performing the design function is not reduced and no new malfunctions are created.

Examples 4-2 and 4-3 illustrate the application of the Use of Software and Digital Devices aspect.

Example 4-2. NO ADVERSE IMPACT on a Design Function related to use of Software and Digital Devices Proposed Activity Description Two non-safety-related trains of main feedwater heaters exist, one for each train of main feedwater. Each main feedwater train consists of six feedwater heaters, for a total of 12 heaters. Each heater possesses an analog controller to control the water level in each of the heaters. Each analog controller is physically and functionally the same.

Each of the analog controllers will be replaced with its own digital controller.

The hardware platform for each digital controller is from the same supplier and the software in each digital controller is exactly the same.

Design Function Identification There are NO design functions associated with the feedwater heater water level controllers. The only UFSAR description related to the heaters states Commented [A19]: Correction (a): NEI 96-07 States: Design functions are UFSAR-described design bases functions and other that the feedwater heater water level controllers are used to adjust the water SSC functions described in the UFSAR that support or impact levels in the heaters to optimize the thermal efficiency of the facility. design bases functions. Implicitly included within the meaning of design function are the conditions under which intended functions are required to be performed, such as equipment response times, Screen Response process conditions, equipment qualification and single failure.

Since there are no design functions associated with the feedwater heater Feedwater, at the right temperature (i.e., heated) is a water level controllers, there are no adverse impacts. design function. Heater level control, or lack thereof can adversely affect the feedwater design function. Please clarify why feedwater heaters do not provide NO design functions associated Commented [A20]: Correction (a): This sentence implies that only explicit descriptions are considered, which is Example 4-3. ADVERSE IMPACT on a Design Function related to use of inconsistent with NEI 96-07, which states design function Software and Digital Devices include SSC functions described in the UFSAR that support or impact design bases functions. As stated in the example, the level controls are described in the UFSAR, Proposed Activity Description and they support the feedwater function.

Two non-safety-related main feedwater pumps (MFWPs) exist. There are two analog control systems (one per MFWP) that are physically and functionally the same.

The two analog control systems will be replaced with two digital control systems. The hardware platform for each digital control system is from the same supplier and the software in each digital control system is exactly the same.

D-10

Design Function Identification The design function of the feedwater control systems is to automatically control and regulate feedwater flow to the steam generators.

The UFSAR identifies the following MFWP control system malfunctions:

(a) failures causing the loss of one MFWP and its associated flow to the steam generators, and (b) failures causing an increase in main feedwater flow to the maximum output from one MFWP.

Screen Response The digital modification associated with this proposed activity is not relatively simple, so the process for assessing relatively simple digital modifications could not be used. There is an adverse impact on the design function of the main feedwater control system because the use of the exact same software in both digital control systems creates a new malfunction that could impact both MFWPs due to a potential software CCF. Commented [A21]: Correction (a): This reasoning seems inconsistent with Example 4-4 below.

COMBINATION OF COMPONENTS/SYSTEMS AND/OR FUNCTIONS Commented [A22]: Enhancement: Include more examples to illustrate the combinations of different system functions besides feedwater. Currently, having one example of The UFSAR may identify the number of components/systems, how the different system combinations appears insufficient given components/systems were arranged, and/or how functions, i.e., design challenges industry has expressed in screenings for complex 50.59 digital modifications.

requirements, were allocated to those components/systems.

Note from meeting:

When replacing analog SSCs with digital SSCs, it is potentially advantageous Clarification provided -

Focus on combination of different systems.

to combine multiple components/systems and/or functions into a single device Provide another example related to a system other than or control system. However, as a result of this combination, the failure of the feedwater. Aux feedwater or the crane is a good example.

single device or control system has the potential to adversely affect the Include combination of functions and combination of design performance of design functions. functions.

The combination of previously separate components/systems and/or functions, in and of itself, does not make the Screen conclusion adverse. Only if combining the previously separate components/systems and/or functions causes an adverse impact on a design function does the combination aspect of the digital modification screen in. Commented [A23]: Correction (b): Delete Only and does. As stated previously by NEI, these sections and When comparing the existing and proposed configurations, consider how the examples only address particular aspects; therefore, these words are inappropriate.

proposed configuration affects the number and/or arrangement of components/systems and the potential impacts of the proposed arrangement on design functions.

Examples 4-4 through 4-6 illustrate the application of the Combination of Components/Systems and/or Functions aspect.

D-11

Example 4-4. Combining Components and Functions with NO ADVERSE IMPACT on a Design Function Proposed Activity Description Two non-safety-related main feedwater pumps (MFWPs) exist. There are two analog control systems (one per MFWP) that are physically and functionally the same. Each analog control system has many subcomponents.

Option #1: Within each control system, all of the analog subcomponents will be replaced with a single digital device that consolidates all of the components, sub-components and the functions associated with each component and sub-component. The components in each analog control system will be replaced with a separate digital control system.

Option #2: Instead of two separate, discreet, unconnected digital control systems being used for the feedwater control systems, only one central digital device is proposed to be used that will combine the previously separate Commented [A24]: Correction (a): The parenthetical control systems and control both main feedwater pumps. statement is incorrect. The statement is tying the combination to CCF. This reasoning is confused. The Design Function Identification combination is within a channel, but the reasoning is about the impact on both channels, and this aspect of reasoning Although the control systems and the major components are described in the conflicts with example 4-3 above. In short the reasoning is the failure is bounded by some other failure, therefore it is UFSAR, only a design function for the feedwater control systems is identified. not a new failure, which is incorrect.

The design function of the feedwater control systems is to automatically Each example in each section is only supposed to address control and regulate feedwater flow to the steam generators. one topic.

The UFSAR identifies the following MFWP control system malfunctions: Suggested correction: (i.e., the current malfunctions already consider the effect on each MFWPs).

(a) failures causing the loss of all feedwater to the steam generators, and Commented [A25]: Correction (a):

(1)A new failure mode is created, and this should screen (b) failures causing an increase in main feedwater flow to the maximum in, but this example reasoned that this new failure mode is bounded by a different failure mode, therefore it output from both MFWPs. screens out. The effect of this guidance is to move evaluation type activities into the screening portion of Screen Response the 50.59 process, which is inconsistent with the intent of NEI 96-07.

NOTE: Since the intent of this example is to illustrate the combination aspect ONLY, the software and hardware aspects will not be addressed in (2) But there is a reduction in independence which needs to be described. Previously the probability of both trains this example. failing may have assumed the failures were independent, now the failures of both trains are dependent, this has Option #1: There is no adverse impact on the design function of the main the potential to reduce reliability which needs to be evaluated.

feedwater control systems to automatically control and regulate feedwater to the steam generators due to the combination of components in each of the two Note from meeting:

Addressing A24 as a green comment will result in change of the channels because no new malfunctions are created (i.e., the current conclusion to address NRC concern. This results in a yellow malfunctions already consider the effect on both MFWPs). comment.

Option #2: Although both main feedwater pumps would be affected by the (3) There may also be a more than minimal increase of failure of the one central digital processor, the proposed activity is not an accident since another way to create it has been adverse because no new malfunctions are created (i.e., the current found; therefore, it should be adverse for another reason.

malfunctions already consider the effect on both MFWPs). Note from meeting:

Same as # 2 of this comment.

D-12

NOTE: For both options, if the malfunctions had considered the effect on only one MFWP, the Screen conclusions would have been adverse because a new malfunction would have been created.

Example 4-5. Combining Components and Functions with NO ADVERSE IMPACT on a Design Function Proposed Activity Description A temperature monitor/controller in a room provides an input to an air damper controller. If temperature gets too high, the temperature controller sends a signal to the air damper to open (if closed) to a predetermined initial position or, if already open, adjusts the position of the damper to allow increased air flow into the room.

Both analog controllers will be replaced with a single digital device that will perform in accordance with the original design requirements providing both temperature monitoring/control and air damper control.

Design Function Identification The temperature monitor/controller performs a design function to continuously monitor the temperature in the room to ensure the initial conditions are met should the emergency room coolers be needed.

The air damper controller performs a design function to continuously provide the appropriate air flow to the room to ensure the initial conditions are met should the emergency room coolers be needed.

There is no lower limit on the acceptable temperature in the room.

Screen Response An engineering evaluation has documented the following malfunctions of the analog devices:

(1) failure of the temperature monitor/controller, causing the loss of input to the air damper controller and the ability of the air damper controller to control the temperature in the room, and (2) failure of the air damper controller, causing the loss of the ability to control the temperature in the room.

Also documented in the engineering evaluation is the malfunction of the digital device, causing the loss of input to the air damper controller and the ability of the air damper controller to control the temperature in the room.

D-13

A comparison of the analog component and digital device malfunctions shows them to be the same. Therefore, although using the digital device might cause multiple design functions to not be performed, no new malfunctions are created. With no new malfunctions being created, there is no adverse impact Commented [A26]: Correction (a): This example implies guidance that is not provided explicitly above.

on the design functions due to the combination aspect. Also, there are no Furthermore, the potential guidance that a reader might indirect impacts that could affect the performance of the design functions due abstracted from this example is: even if multiple devices were postulated to fail independently, their concurrent to the combination aspect. failure (due to a change) is not a new malfunction. While in some cases this may be true, it is not true in all cases; however, no criteria are provided to distinguish between The combining of components/systems and/or functions that were previously these two set of cases.

and completely physically and/or electrically separate (i.e., not coupled) are If there is any generally rule, it should be:

of particular interest when determining the impact on design functions. (1) multiple malfunctions should screen in or (2) multiple malfunctions is a new malfunction Example 4-6 illustrates the combining of control systems from different, See general example comments mentioned above.

originally separate systems.

Delete the highlighted sentence or provide guidance in the Example 4-6. Combining Systems and Functions with an ADVERSE subsection to connect to the highlighted sentence.

IMPACT on a Design Function Proposed Activity Description Two non-safety-related analog feedwater control systems and one separate non-safety-related main turbine steam inlet valves analog control system exist.

All three analog control systems will be replaced with one digital control system that will combine the two feedwater control systems and the main turbine steam-inlet valve control system into a single digital device.

Design Function Identification The design function of the feedwater control systems is to automatically control and regulate feedwater flow to the steam generators.

The design function of the main turbine inlet valve control system is to automatically control and regulate steam flow to the main turbine.

A review of the accident analyses identifies that none of the analyses consider the simultaneous failure of the feedwater control system and the failure of the main turbine control system.

Screen Response Because new malfunctions have been introduced, there are adverse impacts Commented [A27]: Correction (a): Another reason for it being adverse is that it is a reduction in independence.

on the design function of the main feedwater control systems and the design function of the main turbine control system due to the combination of See general example comment.

Note from meeting:

components and functions from the three control systems. See Comment A24 for revision.

D-14

4.2.1.2 Screening of Changes to Procedures as Described in the UFSAR Commented [A28]: Correction (a): NRC staff is recommending the deletion of this section from Appendix D (see general comments on HFE). However, should NEI SCOPE wish to move forward with this section, the following comments will need to be resolved.

If the digital modification does not include or affect a Human-System Note from meeting:

Interface (e.g., the replacement of a stand-alone analog relay with a digital NEI will be keeping this section so comments need to be addressed.

relay that has no features involving personnel interaction and does not feed signals into any other analog or digital device), then this section does not NEI should also consider the use of the terms HFE and HSI in this guidance document.

apply and may be excluded from the Screen assessment.

Note from meeting: Staff believes that the terms should not In NEI 96-07, Section 3.11 defines procedures as follows: be used interchangeably.

"...Procedures include UFSAR descriptions of how actions related to system operation are to be performed and controls over the performance of design functions. This includes UFSAR descriptions of operator action sequencing or response times, certain descriptions...of SSC operation and operating modes, operational...controls, and similar information."

Although UFSARs do not typically describe the details of a specific Human-System Interface (HSI), UFSARs will describe any design functions associated with the HSI.

Because the HSI involves system/component operation, this portion of a digital modification is assessed in this Screen consideration. The focus of the Screen assessment is on potential adverse effects due to modifications of the interface between the human user and the technical device.

Note that the "human user" could involve Control Room Operators, other plant operators, maintenance personnel, engineering personnel, technicians, etc.

There are three "basic HSI elements" of an HSI (

Reference:

NUREG-0700):

  • Displays: the visual representation of the information personnel need to monitor and control the plant.
  • Controls: the devices through which personnel interact with the HSI and the plant.
  • User-interface interaction and management: the means by which personnel provide inputs to an interface, receive information from it, and manage the tasks associated with access and control of information.

Any user of the HSI must be able to accurately perceive, comprehend and respond to system information via the HSI to successfully complete their tasks. Specifically, nuclear power plant personnel perform "four generic primary tasks" (

Reference:

NUREG/CR-6947):

D-15

1. Monitoring and detection (extracting information from the environment and recognizing when something changes),
2. Situation assessment (evaluation of conditions),
3. Response planning (deciding upon actions to resolve the situation), and
4. Response implementation (performing an action).

D-16

Table 1 contains examples of modifications to each of the three basic HSI elements applicable to this Screen consideration.

Table 1 - Example Human-System Interface Modifications HSI Element Typical Modification Description/Example Increase/decrease in the amount of information displayed by and/or available from the HSI (e.g., combining multiple parameters into a single integrated Number of Parameters parameter, adding additional information regarding component/system performance)

Change to the type of information displayed and/or available from the HSI Displays Type of Parameters (e.g., removing information that was previously available or adding information that was previously unavailable)

Change to visual representation of information (e.g. increment of presentation Information Presentation modified)

Change to structural arrangement of data/information (e.g., information now Information Organization organized by channel/train rather than by flow-path)

Change to the type/functionality of input device (e.g., replacement of a push Control Input button with a touch screen)

Controls Change to the information sent back to the individual in response to an action Control Feedback (e.g., changing feedback from tactile to auditory)

Change in number and/or type of decisions made and/or actions taken (e.g.,

replacing an analog controller that can be manipulated in one step with a Action Sequences digital controller that must be called-up on the interface and then User-Interface manipulated)

Interaction and Changes that affect how an individual retrieves information/data (e.g.,

Information/Data Management information that was continuously displayed via an analog meter now requires Acquisition interface interaction to retrieve data from a multi-purpose display panel)

Changes from manual to automatic initiation (or vice versa) of functions (e.g.,

Function Allocation manual pump actuation to automatic pump actuation)

D-17

To determine potential adverse impacts of HSI modifications on design functions, a two-step HSI assessment must be performed, as follows: Commented [A29]: Enhancement: This process proposes a two-step HSI process and a third step (described below)

  • Step One - Identify the generic primary tasks that are involved with for performing the screening task (i.e., determine whether a change is adverse). This step contains technical guidance (i.e., potentially impacted by) the proposed activity. regarding HSI evaluations (i.e., the two-step HSI assessment), but not guidance on how to convert the technical conclusion (i.e., there is a negative impact or not)
  • Step Two - For all primary tasks involved, assess if the modification into a screening evaluation (i.e., adverse or not).

negatively impacts an individual's ability to perform the generic Commented [A30]: Enhancement: No guidance is primary task. provided on how to do this step. Furthermore, probably only an HFE professional can do this adequately.

Examples of negative impacts on an individual's performance that may Commented [A31]: Correction (a): Delete may otherwise this guidance is saying negative impacts may or may not result in adverse effects on a design function include, but are not be adverse; no guidance is provided how to distinguish limited to: these two.

increased possibility of mis-operation, Please compare with Examples 4-8a & 4-8b below. In both of these examples, an HFE review determined the HFE increased difficulty in evaluating conditions, modification to have a negative impact, but the 50.59 increased difficulty in performing an action, screening determined to be not adverse.

increased time to respond, and This guidance is inconsistent with NEI 96-07 Rev. 1 Section creation of new potential failure modes. 4.2.1, subsection: Screening for Adverse Effects which states: The screening process is not concerned with the magnitude of adverse effects that are identified. Any change that adversely affects ... is screened in. The After the two-step HSI assessment, the final step is application of the magnitude of the adverse effect (e.g., is the minimal increase standard met?) is the focus of the 10 CFR 50.59 standard Screen assessment process (i.e., identification of design functions evaluation process.

and determination of adverse or not adverse, including the justification for Furthermore NEI 96-07 Section 4.2.1.2 states: Proposed the conclusion). changes that are determined to have positive or no effect on how SSC design functions are performed or controlled may SIMPLE HUMAN-SYSTEM INTERFACE EXAMPLE be screened out.

Commented [A32]: Enhancement: This is the third step.

Example 4-7 illustrates how a digital modification with HSI considerations Commented [A33]: Correction (a): Any increase should be would be addressed. understood as adverse and this should be stated in this paragraph.

Example 4-7: Assessment of Modification with NO ADVERSE IMPACT on a Please compare with Examples 4-8a & 4-8b below. In both UFSAR-Described Design Function of these examples, an HFE review determined the HFE modification to have a negative impact, but the 50.59 Proposed Activity Description screening determined to be not adverse. Appendix D, provides no guidance for determining how much of an increase or how many types of increases result in and Currently, a knob is rotated clock-wise to open a flow control valve in 1% adverse determination.

increments and counter clock-wise to close a flow control valve in 1%

Please delete this clause. That is, the HFE professional increments. This knob will be replaced with a touch screen that has two determines whether there is a negative impact, and their separate arrows, each in its own function block. Using the touch screen, 50.59 engineer determines whether a design function is affected.

touching the "up" arrow will open the flow control valve in 1% increments and touching the "down" arrow will close the flow control valve in 1%

increments.

HSI Assessment Process STEP 1. Identification of the Generic Primary Tasks Involved:

D-18

(1) Monitoring and detection (extracting information from the environment and recognizing when something changes) - NOT INVOLVED (2) Situation assessment (evaluation of conditions) - NOT INVOLVED (3) Response planning (deciding upon actions to resolve the situation) -

NOT INVOLVED (4) Response implementation (performing an action) -INVOLVED STEP 2. Assessment of Modification Impacts on the Involved Generic Primary Tasks:

As part of the technical evaluation supporting the proposed modification, a HFE evaluation was performed.

Tasks 1, 2 and 3 were not involved, so these tasks are not impacted by the modification.

Task 4 was identified as involved; the HFE evaluation determined that the change from knob to touch screen would not have a negative impact because it does not affect the operators ability to perform the response implementation task.

Identification and Assessment of the Relevant Design Function(s)

The UFSAR states the operator can "open and close the flow control valve using manual controls located in the Main Control Room." Thus, the design function is the ability to allow the operator to manually adjust the position of the flow control valve and the UFSAR description implicitly identifies the SSC (i.e., the knob).

Using the results from the HFE evaluation and examining the replacement of the "knob" with a "touch screen," the modification is not adverse because it does not impact the ability of the operator to "open and close the flow control valve using manual controls located in the Main Control Room," maintaining satisfaction of the UFSAR-described design function.

COMPREHENSIVE HUMAN-SYSTEM INTERFACE EXAMPLES Examples 4-8a and 4-8b illustrate how a digital modification with HSI considerations would be addressed.

Although both examples use the same basic digital modification, Example 4-8a illustrates a no adverse impact case and Example 4-8b illustrates an adverse impact case by "complicating" the HSI portion of the modification.

D-19

Example 4-8a. Digital Modification Involving HSI Considerations with NO ADVERSE IMPACT on a Design Function Proposed Activity Description Analog components and controls for a redundant safety-related system are to be replaced with digital components and controls, including new digital-based HSI.

Currently, two redundant channels/trains of information and controls are provided to the operators in the Main Control Room for the redundant systems. For each channel/train, several different analog instruments present information regarding the performance of the system. The analog displays are arranged by system "flow path" to facilitate the operator's ability to monitor the system as a whole.

The existing HSI for these components is made up of redundant hard-wired switches, indicator lights and analog meters. The new HSI consolidates the information and controls onto two flat panel displays (one per train) with touch screen soft controls. The information available on the flat panels is equivalent to that provided on the current analog HSI. Each flat panel display contains only one screen that displays the information and the controls for only that train, replicating the information and controls arrangement as they are in the existing HSI.

The existing HSI requires operators to manipulate analog switches to implement a control action. To take a control action using the new HSI, the operator must (via the touch screen) select the appropriate activity (e.g.,

starting/initiating the system or changing the system line-up), select the component to be controlled (e.g., pump or valve), select the control action (e.g., start/stop or open/close), and execute the action.

HSI Assessment Process Step 1. Identification of Which Four Generic Primary Tasks are Involved:

(1) Monitoring and detection (extracting information from the environment and recognizing when something changes) - INVOLVED (2) Situation assessment (evaluation of conditions) - NOT INVOLVED (3) Response planning (deciding upon actions to resolve the situation) -

NOT INVOLVED (4) Response implementation (performing an action) - INVOLVED Step 2. Assessment of the Modification Impacts on the Involved Generic Primary Tasks:

As part of the technical evaluation supporting the proposed modification, a hsi evaluation was performed.

D-20

Task 1 is involved. Any change to information presentation has the potential to impact the operators ability to monitor and detect changes in plant parameters. Even though the modification will result in information being presented on flat panels, the information available and the organization of that information (i.e., by train) will be equivalent to the existing HSI. Due to this equivalence and additional favorable factors (e.g., appropriate sized flat panels, appropriate display brightness, clearly identified function buttons, etc.) as documented in the HFE evaluation, there is no impact to the operators ability to monitor and detect changes in plant parameters.

Tasks 2 and 3 were not involved, so these tasks are not impacted by the modification.

Task 4 is involved. The modification will require the operator to perform four actions in order to manipulate a control (i.e., 1. select the appropriate activity, 2. select the specific component to be controlled, 3. select the control action to be initiated, and 4. execute the action). Currently, the operator is able to manipulate a control in one action (e.g., turn a switch to on/off). The HFE evaluation determined that the modification negatively impacts the operators ability to respond because the modification increases the difficulty of implementing a response by requiring four actions instead of one action and the additional actions result in an increase in the operators time to respond. Commented [A34]:

Correction (a): Connected to previous comments.

Identification and Assessment of Design Functions Compare to conclusion below.

Design Function Identification This example conflicts with NEI 96-07 by stating that a negative effect on a design function is not adverse and (a) Status indications are continuously available to the operator. screens out.

(b) The operator controls the system components manually. Note from meeting:

NEI to remove any judgement (negative/positive) from the Screen Response HFE section.

Immediately connected to Comments A33, A34, and A35.

Since the information available and the organization of that information Staff noted that link between Task 3 and Task 4 still using the new HSI is equivalent to the existing HSI, the design function for needed.

Ensure assumptions are documented as it related to continuous availability of status indications is met and there is no adverse adverse impact.

impact on design function (a).

NEI will make changes that says not adverse for this Using the touch screen, the operator is still able to perform design function particular..

(b) to manipulate the control for the systems components. Therefore, there is no adverse impact on satisfaction of design function (b). Commented [A35]: Correction (a): Compare to HFE evaluation above.

This example conflicts with NEI 96-07 by stating that a negative effect on a design function is not adverse and Example 4-8b. Digital Modification Involving HSI Considerations with an screens out.

ADVERSE IMPACT on a Design Function Note from meeting:

See NEIs notes on correction to this language.

D-21

Proposed Activity Description Analog components and controls for a redundant safety-related system are to be replaced with digital components and controls, including new digital-based HSI.

Currently, two redundant channels/trains of information and controls are provided to the operators in the Main Control Room for the redundant systems. For each channel/train, several different analog instruments present information regarding the performance of the system. The analog displays are arranged by system "flow path" to facilitate the operator's ability to monitor the system as a whole.

The existing HSI for these components is made up of redundant hard-wired switches, indicator lights and analog meters. The new HSI consolidates the information and controls onto two flat panel displays (one per train) with touch screen soft controls. The information available on the flat panels is equivalent to that provided on the current analog HSI. Each flat panel display contains only one screen, which can display the information for only one train and the controls for only that train, replicating the information and controls arrangement as they are in the existing HSI. Each flat panel display can be customized to display the parameters and/or the configuration (e.g. by train, by flow path or only portions of a train or flow path) preferred by the operators. In addition, the flat panel displays provide many other display options to the user (e.g., individual component status and component/system alarms).

The existing HSI requires operators to manipulate analog switches to implement a control action. To take a control action using the new HSI, the operator must (via the touch screen) select the appropriate activity (e.g.,

starting/initiating the system or changing the system line-up), select the component to be controlled (e.g., pump or valve), select the control action (e.g., start/stop or open/close), and execute the action.

HSI Assessment Process Step 1. Identification of Which Four Generic Primary Tasks are Involved:

(1) Monitoring and detection (extracting information from the environment and recognizing when something changes) - INVOLVED (2) Situation assessment (evaluation of conditions) - INVOLVED (3) Response planning (deciding upon actions to resolve the situation) -

INVOLVED (4) Response implementation (performing an action) - INVOLVED Step 2. Assessment of the Modification Impacts on the Involved Generic Primary Tasks:

D-22

As part of the technical evaluation supporting the proposed modification, a HFE evaluation was performed.

Tasks 1, 2 and 3 are involved (emphasizing that the modification includes a change to information presentation and organization, such that the indications/instruments are now consolidated and presented on customizable flat panel displays, rather than static analog control boards). With the new displays and display options available to the operators, the operators can choose which parameters to display and the organization of that information (e.g., by train/path). The HFE evaluation concluded that this modification could result in the operator choosing not to have certain parameters displayed; thus negatively impacting their ability to monitor the plant and detect changes. In addition, altering the information displayed and the organization of the information will negatively impact the operators understanding of how the information relates to system performance. This negative impact on understanding will also negatively impact the operators ability to assess the situation and plan an appropriate response. Commented [A36]: Correction (a): Compare with conclusion (a) below.

Task 4 is involved. The modification will require the operator to perform This example conflicts with NEI 96-07 by stating that a negative effect on a design function is not adverse and four actions in order to manipulate a control (i.e., 1. select the appropriate screens out.

activity, 2. select the specific component to be controlled, 3. select the control action to be initiated, and 4. execute the action). Currently, the operator is able to manipulate a control in one action (e.g., turn a switch to on/off). The HFE evaluation determined that the modification negatively impacts the operators ability to respond because the modification increases the difficulty of implementing a response by requiring four actions instead of one action and the additional actions result in an increase in the operators time to respond. Commented [A37]: Correction (a): Compare with conclusion (b) below.

Identification and Assessment of Design Functions This example conflicts with NEI 96-07 by stating that a negative effect on a design function is not adverse and Design Function Identification screens out.

(a) Status indications are continuously available to the operator.

(b) The operator controls the system components manually.

Commented [A38]: Correction (a): This conclusion only Screen Response addresses one of the negative conclusions reached in the HFE evaluation; does this mean the others do not have an The information available and the organization of that information in the adverse effect? What guidance states what does or does not have an adverse effect?

new displays are customizable based on operator preference. Critical status indications may not be continuously available to the operator, thus there is This example conflicts with NEI 96-07 by stating that a negative effect on a design function is not adverse and an adverse impact on design function (a). screens out.

Commented [A39]: Correction (a): Compare with HFE Using the touch screen, the operator is still able to perform design function evaluation above.

(b) to manipulate the control for the systems components. Therefore, there is no adverse impact on satisfaction of design function (b). This example conflicts with NEI 96-07 by stating that a negative effect on a design function is not adverse and screens out.

D-23

Since there is an adverse impact on design function (a), the overall conclusion of the Screen for this consideration would be adverse.

4.2.1.3 Screening Changes to UFSAR Methods of Evaluation By definition, a proposed activity involving a digital modification involves SSCs and how SSCs are operated and controlled, not a method of evaluation described in the UFSAR (see NEI 96-07, Section 3.10).

Methods of evaluation are analytical or numerical computer models used to determine and/or justify conclusions in the UFSAR (e.g., accident analyses that demonstrate the ability to safely shut down the reactor or prevent/limit radiological releases). These models also use "software." However, the software used in these models is separate and distinct from the software installed in the facility. The response to this Screen consideration should reflect this distinction.

A necessary revision or replacement of a method of evaluation (see NEI 96-07, Section 3.10) resulting from a digital modification is separate from the digital modification itself and the guidance in NEI 96-07, Section 4.2.1.3 applies.

4.2.2 Is the Activity a Test or Experiment Not Described in the UFSAR?

By definition, a proposed activity involving a digital modification involves SSCs and how SSCs are operated and controlled, not a test or experiment (see NEI 96-07, Section 4.2.2). The response to this Screen consideration should reflect this characterization.

A necessary test or experiment (see NEI 96-07, Section 3.14) involving a digital modification is separate from the digital modification itself and the guidance in NEI 96-07, Section 4.2.2 applies.

4.3 EVALUATION PROCESS CAUTION The guidance contained in this appendix is intended to supplement the generic Evaluation guidance contained in the main body in NEI 96-07, Section 4.3.

Namely, the generic Evaluation guidance provided in the main body of NEI 96-07 and the more-focused Evaluation guidance in this appendix BOTH apply to digital modifications.

D-24

Introduction Throughout this section, references to the main body of NEI 96-07, Rev. 1 will be identified as "NEI 96-07."

Guidance Focus Commented [A40]: Correction (a): The evaluation section of Appendix D appears to introduce new expansions of or In the following sections and sub-sections that describe the Evaluation paraphrasing of general 10 CFR 50.59 guidance without a clear nexus or explanation on why this is necessary for guidance particularly useful for the application of 10 CFR 50.59 to digital digital-specific modification. Clarifications of general 50.59 modifications, each section and sub-section describes only a specific aspect, guidance topics should be made within the main body of NEI 96-07. Appendix D should avoid unnecessary sometimes at the deliberate exclusion of other related aspects. This focused expansion of general 50.59 concepts already covered in the approach is intended to concentrate on the particular aspect of interest and main body of NEI 96-07 unless specifically demonstrated to be uniquely necessary for digital modifications, as does not imply that the other aspects do not apply or could not be related to demonstrated by Section 4.2s screening guidance on the aspect being addressed. combination of components/functions subsection. Proper incorporation of the guidance of Supplement 1 to RIS 2002-22 obviates the need for expansion, refinement or Example Focus paraphrasing of general 50.59 concepts from the main body of NEI 96-07 revision 1 for evaluation guidance.

Examples are provided to illustrate the guidance provided herein. Unless stated otherwise, a given example only addresses the specific aspect or topic within the section/sub-section in which it is included, sometimes at the deliberate exclusion of other aspects or topics that, if considered, could potentially change the Evaluation conclusion.

Qualitative Assessment Commented [A41]: Correction (a): Since a qualitative assessment is used to address more than one evaluation question, the guidance for performing it should be in one For digital I&C systems, reasonable assurance of low likelihood of failure is place, presumably in this sub-section.

derived from a qualitative assessment of factors involving the design Very little qualitative assessment guidance from attributes of the modified SSC, the quality of the design processes, and the Supplement 1 to RIS 2002-22 has been incorporated.

operating experience of the software and hardware used (i.e., product Furthermore, the restrictions on the applicability of the maturity and in-service experience). The qualitative assessment is used to qualitative assessment have not been incorporated.

record the factors and rationale and reasoning for making a determination Note from meeting:

that there is reasonable assurance that the digital I&C modification will Clarification provided that this comment is provided for exhibit a low likelihood of failure by considering the aggregate of these RPS and ESFAS as discussed in the RIS 2002-22, Supp 1.

factors.

Commented [A42]: Correction (a): Supplement 1 to RIS 2002-22 defines a particular type of qualitative assessment SSC Failure Likelihood Determination Outcomes germane to 50.59 evaluations, based upon the guidance of NEI 01-01. This paragraph appears to define the general concept of a qualitative assessment with that described in The possible outcomes of an engineering evaluation (e.g., qualitative Supplement 1. This distinction should be made when assessment), performed in accordance with applicable Industry and/or NRC introducing this concept here and made more clearly in subsequent discussion of a qualitative assessment, as guidance documents, are as follows: described in the RIS supplement to be applied to evaluation criteria.

(1) SSC failure likelihood is sufficiently low (as defined in Definition 3.15), Commented [A43]: Correction (a):

Little guidance from Supplement 1 to RIS 2002-22 has been or incorporated. Specifically information from Section 2 of the RIS has been omitted.

(2) SSC failure likelihood is not sufficiently low.

D-25

If the SSC failure likelihood is determined to be sufficiently low, then by extension, the likelihood of software CCF is also considered to be sufficiently low.

If the failure likelihood of the modified SSC is not examined as part of an engineering evaluation, then the failure likelihood of the modified SSC will be assumed to be not sufficiently low for purposes of responding to the following 10 CFR 50.59 Evaluation criteria.

These possible outcomes (i.e., sufficiently low or not sufficiently low) will be used in developing the responses to Evaluation criteria 1, 2, 5 and 6.

4.3.1 Does the Activity Result in More Than a Minimal Increase in the Frequency of Occurrence of an Accident? Commented [A44]: Correction (b): This section contains many instances that are contrary to 50.59(c)(2)(i).

INTRODUCTION Neither the 50.59 regulation nor NEI 96-07, Section 4.3.1, use the word initiator. Therefore

  • only an accident, not an accident initiator must be From NEI 96-07, Section 3.2: explicitly mentioned in the UFSAR., and
  • a more than minimal increase in the frequency of an "The term 'accidents' refers to the anticipated (or abnormal) accident, as a result of the change could be for any operational transients and postulated design basis accidents..." reason such a new, entirely different initiator that was introduced by the change.

Therefore, for purposes of 10 CFR 50.59, both Anticipated Operational See specific comments below Occurrences (AOOs) and Postulated Accidents (PAs) fall within the definition of "accident."

After applying the generic guidance in NEI 96-07, Section 4.3.1 to identify any accidents affected by the systems/components involved with the digital modification and examining the initiators of those accidents, the impact on the frequency of the initiator (and, hence, the accident itself) due to the Commented [A45]: Correction (b): Contrary to 50.59 and NEI 96-07. As described in the above comment, an digital modification can be assessed. increase in frequency can be for any reason and is not limited by examining the initiators of those accidents All accident initiators fall into one of two categories: equipment-related or because the change can introduce a new, entirely different initiators.

personnel-related. Therefore, the assessment of the impact of a digital Commented [A46]: Enhancement: To address comments modification also needs to consider both equipment-related and personnel- regarding accident initiators, the NRC staff recommends related sources. wording from NEI 96-07, Section 4.3.1, which states, In answering this question, the first step is to identify the For a digital modification, the range of possible equipment-related sources of accidents that have been evaluated in the UFSAR that are affected by the proposed activity. Then a determination accident initiators includes items unique to digital and items not unique to should be made as to whether the frequency of these digital. An example of an item unique to digital is consideration of the accidents occurring would be more than minimally increased.

impact on accident frequency due to a software CCF, which will be addressed in this guidance. An example of a potential source of common cause failure Alternate recommended wording: NEI 01-01, Section 4.3.1, that is not unique to digital is consideration of the impact on accident The first step in addressing this criterion is to identify the frequency due to the digital system's compatibility with the environment in accidents that have been evaluated in the UFSAR and that may be affected by the proposed activity. Then the change which the system is being installed, which would be addressed by applying is evaluated to determine whether the frequency of these the general guidance related to meeting applicable regulatory requirements accidents could increase as a result of the change. In answering this question for digital upgrades, the key issue and other acceptance criteria to which the licensee is committed, and is whether the digital equipment can increase the frequency of initiating events that lead to accidents D-26

departures from standards as outlined in the general design criteria, as discussed in NEI 96-07, Section 4.3.1 and Section 4.3.1, Example 2.

For a digital modification, the assessment for personnel-related sources will consider the impact due to the Human-System Interface (HSI).

Typically, numerical values quantifying an accident frequency are not available, so the qualitative approach using the attributable (i.e., causal relationship) and the negligible/discernable (i.e., magnitude) criteria from NEI 96-07, Section 4.3.1 will be examined in this guidance.

GUIDANCE Determination of Attributable (i.e., Causality)

NOTE: This guidance is not unique to digital and is the same as that provided in NEI 96-07, Section 4.3.1. This guidance is included here for completeness.

If none of the components/systems involved with the digital modification are identified as affecting an accident initiator previously identified in the UFSAR, then there is no attributable impact on the frequency of occurrence of an accident.

Alternately, if any component/system involved with the digital modification is identified as affecting an accident initiator previously identified in the UFSAR, then an impact on the frequency of occurrence of an accident can be attributed to the digital modification. If an attributable impact is identified, then further assessment to determine the magnitude of the impact will be performed.

Examples 4-9 and 4-10 will illustrate the application of the attributable criterion.

Example 4-9 illustrates a case of NO attributable impact on the frequency of occurrence of an accident.

Example 4-9. NO ATTRIBUTABLE Impact on the Frequency of Occurrence of an Accident Proposed Activity Description Two safety-related containment chillers exist. There are two analog control systems (one per chiller) that are physically and functionally the same.

Each analog control system will be replaced with a separate digital control system. The hardware platform for each digital control system is from the same supplier and the software in each digital control system is exactly the same.

D-27

Affected Accidents and Accident Initiators The review of the UFSAR accident analyses identified the Loss of Coolant Accident (LOCA) and Main Steam Line Break (MSLB) events as containing requirements related to the safety-related containment chillers. Specifically, Commented [A47]: Correction (a): This type of criteria may be applicable to malfunctions; however, what is the UFSAR states the following: "To satisfy single failure requirements, the applicable to question (i) (i.e., accidents) is the AOOs or loss of only one control system and its worst-case effect on the containment DBAs that the chiller failures and misbehaviors would create. This example appears to provide a methodology for post-accident [emphasis added] environment due to the loss of one chiller has determining whether a modification relates to an accident been considered in the LOCA and MSLB analyses." initiator. It is not clear why this method will always be conservative.

Therefore, the affected accidents are LOCA and MSLB. For example ESFAS mitigates against certain accidents.

The ESFAS is not an initiator of those accidents that it The UFSAR identified an equipment-related initiator for both accidents as mitigates. By the reasoning implied in this example, the being a pipe break. For LOCA, the pipe break occurs in a hot leg or a cold ESFAS could not impact the frequency of an accident.

However, a spurious ESFAS actuation is itself an accident.

leg. For MSLB, the pipe break occurs in the main steam line exiting the steam generator. It is not apparent that this example considers whether the failure of the components could be a different accident than Impact on Accident Frequency one that it mitigates (i.e., the chillers do not contribute to the LOCA).

In these accidents, the safety-related containment chillers are not accident Commented [A48]: Correction (a): This example considers the outcomes of an event. Clarify the relevance of the initiators (i.e., they are not pipe breaks). Furthermore, the chillers are only outcome when determining the frequency of the event.

considered as part of accident mitigation; after the accidents have already occurred. Therefore, there is NO impact on the frequency of occurrence of the accidents that can be attributed to the digital modification.

Example 4-10 illustrates a case of an attributable impact on the frequency of occurrence of an accident.

Example 4-10. ATTRIBUTABLE Impact on the Frequency of Occurrence of an Accident Proposed Activity Description Two non-safety-related main feedwater pumps (MFWPs) exist, each with its own flow control valve. There are two analog control systems (one per MFWP and flow control valve combination) that are physically and functionally the same.

Each analog control system will be replaced with a separate digital control system. The hardware platform for each digital control system is from the same supplier and the software in each digital control system is exactly the same.

Affected Accident and Accident Initiators The affected accident is the Loss of Feedwater event. The UFSAR identifies the equipment-related initiators as being the loss of one MFWP or the closure of one MFWP flow control valve.

D-28

Impact on Accident Frequency In this accident, the non-safety-related feedwater system is related to the accident initiators (i.e., loss of a MFWP and/or closure of a flow control valve).

Therefore, an impact on the frequency of occurrence of the accident can be attributed to the digital modification. (NOTE: The magnitude of the impact would be assessed next.)

Determination of Negligible/Discernable (i.e., Magnitude)

NOTE: The guidance in this sub-section applies ONLY when an attributable impact on the frequency of occurrence of an accident has been established.

For proposed activities in which there is an attributable impact on the frequency of occurrence of an accident, the negligible/discernable portion of the criteria (i.e., magnitude) also needs to be assessed.

To determine the overall effect of the digital modification on the frequency of an accident, an engineering evaluation is performed. An engineering evaluation that uses a qualitative assessment to judge the failure likelihood Commented [A49]: Enhancement - Please explain the change from software CCF to failure likelihood.

of the modified SSC should consider each factor identified below:

This same comment applies to Section 4.3.2.

  • Design attributes employed
  • Quality of the design processes, and
  • Operating experience of the software and hardware used (i.e., product maturity and in-service experience).

Negligible:

To achieve a negligible conclusion, the engineering evaluation of each factor (e.g., as documented in a qualitative assessment) would conclude that the affected SSC will exhibit a sufficiently low likelihood of failure, and by extension, that the change in the accident frequency "...is so small or the uncertainties in determining whether a change in frequency has occurred are such that it cannot be reasonably concluded that the frequency has actually changed (i.e., there is no clear trend toward increasing the frequency)"1

[emphasis added]. Commented [A50]: Correction (a): Conflicts with paragraph below. In short, this paragraph states that Therefore, if the qualitative assessment outcome is sufficiently low, there is sufficiently low means that there is no clear trend not more than a minimal increase in the frequency of occurrence of an accident previously evaluated in the UFSAR.

Discernable:

1 Refer to NEI 96-07, Section 4.3.1, Example 1.

D-29

If the examination of each factor concludes that the change in the accident frequency exhibits a clear trend towards increasing the frequency, then a discernable increase in the accident frequency would exist. In this case, the software CCF likelihood could be sufficiently low or not sufficiently low. Commented [A51]: Correction (a): Conflicts with paragraph above. In short, this paragraph states that a The engineering evaluation (e.g., the qualitative assessment) is also used to clear trend towards increasing means nothing with respect to sufficiently low of software CCF. The logic in determine if the discernible increase in the accident frequency is "more than transition between negligible and discernable is not clear minimal" or "NOT more than minimal." To achieve a conclusion of "NOT when sufficiently low or not sufficiently low is determined, more than minimal," the proposed activity must continue to meet and/or Furthermore this paragraph seems to be inconsistent with satisfy all applicable NRC requirements, as well as design, material, and the definition of sufficiently low in Section 3.15 of Appendix D.

construction standards, to which the licensee is committed. Applicable requirements and standards include those selected by the licensee for use in This same comment applies to Section 4.3.2.

the development of the proposed digital modification and documented within the design modification package. Commented [A52]: Correction (b): This criteria is necessary but not sufficient to demonstrate no more than a Examples 4-11 and 4-12 illustrate the negligible/discernable portion (i.e., minimum increase. For example, if the accident frequency is doubled but the modification still meets requirements, magnitude) of the criteria and assume the attributable portion of the criteria the criteria would allow this change to be made under has been satisfied. 50.59.

This is contrary to NEI 96-07, section 4.3.1. If the Example 4-11 illustrates a case with a negligible change to the accident qualitative assessment determines that the propose mod is frequency. not sufficiently low, then meeting applicable requirement is not sufficient to say that it is not more than minimum increase.

Example 4-11. NEGLIGIBLE Impact on the Frequency of Occurrence of an Accident This is applicable to the staffs review of Example 4-12.

Proposed Activity Description Two non-safety-related main feedwater pumps (MFWPs) exist, each with its own flow control valve. There are two analog control systems (one per MFWP and flow control valve combination) that are physically and functionally the same.

Each analog control system will be replaced with a separate digital control system. The hardware platform for each digital control system is from the same supplier and the software in each digital control system is exactly the same.

Magnitude Conclusion A qualitative assessment performed as part of the design process, considering system design attributes, quality of the design processes employed, and operating experience of the proposed equipment, concluded that the failure likelihood of the modified SSC is sufficiently low.

D-30

All applicable requirements and other acceptance criteria to which the licensee is committed, as well as applicable design, material and construction standards, continue to be met.

Therefore, the change in the frequency of occurrence of the Loss of Feedwater event is negligible due to the effect of the factors considered in the qualitative assessment.

Overall Conclusion Although an attributable impact on the frequency of occurrence of the Loss of Feedwater event was determined to exist, there was no clear trend toward increasing the frequency. With no clear trend toward increasing the frequency, there is not more than a minimal increase in the frequency of occurrence of the accident due to the digital modification.

Example 4-12 illustrates a case with a discernable increase to the accident frequency.

Example 4-12. DISCERNABLE Increase in the Frequency of Occurrence of an Accident Proposed Activity Description Same as Example 4-11.

Magnitude Conclusion Based on the qualitative assessment performed as part of the technical assessment supporting this digital modification, the likelihood of failure causing the loss of both feedwater control systems (resulting in the loss of both MFWPs) has been determined to be not sufficiently low.

All applicable requirements and other acceptance criteria to which the licensee is committed, as well as applicable design, material and construction standards, continue to be met.

The change in the frequency of occurrence of the Loss of Feedwater event is discernable due to the effect of the factors considered in the qualitative evaluation.

Overall Conclusion An attributable impact on the frequency of occurrence of the Loss of Feedwater event was determined to exist and there is a clear trend towards increasing the frequency. The clear trend toward increasing the frequency (i.e., the discernable increase) is due to the software CCF likelihood being not sufficiently low.

D-31

However, even with a clear trend towards increasing the frequency, the assessments and conclusions documented in the qualitative assessment of the considered factors and the satisfaction of applicable requirements and other acceptance criteria to which the licensee is committed, as well as applicable design, material and construction standards, there is NOT more than a minimal increase in the frequency of occurrence of the accident due to the digital modification. Commented [A53]: Correction (a): Meeting applicable requirements and standards is necessary but not sufficient.

See comment above.

HUMAN-SYSTEM INTERFACE ASSESSMENT Commented [A54]: Enhancement: This section does not contain any digital unique guidance; therefore, it can be deleted. Same comment applies to Sections 4.3.1, 4.3.2, If no personnel-based initiators involving degraded operator performance 4.3.5., and 4.3.6 (e.g., operator error) are identified among the accident initiators, then an increase in the frequency of the accident cannot occur due to the Human-System Interface portion of the digital modification. Otherwise, the application of the attributable criterion (i.e., causality) and the negligible/discernable criterion (i.e., magnitude) are assessed utilizing the guidance described in NEI 96-07, Section 4.3.1. Commented [A55]: Enhancement: There is almost no guidance related to attributable and negligible/discernable 4.3.2 Does the Activity Result in More Than a Minimal Increase in the Likelihood in NEI 96-07. Furthermore, is it necessary to say, once again, to follow the guidance in NEI 96-07?

of Occurrence of a Malfunction of an SSC Important to Safety?

This same comment applies to Sections 4.3.1, 4.3.2, 4.3.5.,

and 4.3.6.

INTRODUCTION After applying the generic guidance in NEI 96-07, Section 4.3.2 to identify any malfunctions affected by the systems/components involved with the digital modification and examining the initiators of those malfunctions, the impact on the likelihood of the initiator (and, hence, the malfunction itself) due to the digital modification can be assessed.

All malfunction initiators fall into one of two categories: equipment-related or personnel-related. Therefore, the assessment of the impact of a digital modification also needs to consider both equipment-related and personnel-related sources.

For a digital modification, the range of possible equipment-related sources of malfunction initiators includes items unique to digital and items not unique to digital. An example of an item unique to digital is consideration of the impact on malfunction likelihood due to a software CCF, which will be addressed in this guidance. An example of a potential source of common cause failure that is not unique to digital is consideration of the impact on malfunction likelihood due to the digital system's compatibility with the environment in which the system is being installed, which would be addressed by applying the general guidance related to meeting applicable regulatory requirements and other acceptance criteria to which the licensee D-32

is committed, and departures from standards as outlined in the general design criteria, as discussed in NEI 96-07, Section 4.3.2.

For a digital modification, the assessment for personnel-related sources will consider the impact due to the Human-System Interface (HSI).

Typically, numerical values quantifying a malfunction likelihood are not available, so the qualitative approach using the attributable (i.e., causal relationship) and the negligible/discernable (i.e., magnitude) criteria from NEI 96-07, Section 4.3.2 will be examined in this guidance.

GUIDANCE Commented [A56]: Enhancement: Supplement 1 to RIS 2002-22 establishes that reductions in redundancy, diversity and independence are general technical concerns Impact on Redundancy, Diversity, Separation or Independence for digital modifications and not specifically tied to any single evaluation criteria under 10 CFR 50.59. In order to As discussed in NEI 96-07, Section 4.3.2, Example 6, a proposed activity that properly align with Supplement 1, discussions regarding reductions in redundancy, diversity and independence reduces redundancy, diversity, separation or independence is considered should be stated at the beginning of the evaluation section more than a minimal increase in the likelihood of a malfunction and requires for Appendix D.

prior NRC approval. However, licensees may reduce excess redundancy, Commented [A57]: Correction (b): These four terms are used in the Principle Design Criteria of many plants, in diversity, separation or independence (if any) to the level credited in the addition, they appear in standards incorporated by safety analyses without prior NRC approval. reference into the CFR as well as in RGs which some plants state (in the UFSAR) that they comply with. In short To ensure consistent application of this guidance, each of these these terms are defined by the licensing basis of each plant, and they are not be defined in exactly the same way.

characteristics is reviewed below. Furthermore, the definitions of these terms below are inconsistent to their use in the GDCs and regulations Redundancy: incorporated by reference.

Commented [A58]: Correction (b): Suggest the revising "Redundancy" means two or more SSCs performing the same design function. the first sentence of this paragraph to: As discussed in NEI 96-07, Section 4.3.2, Example 6, a proposed activity that The introduction of the exact same software into redundant channels and the reduces redundancy, diversity, separation or independence potential creation of a software CCF has no impact on an SSCs' redundancy of the design function.

because the SSCs perform the same design function(s) before the introduction For consistency with RIS 2002-22, Supplement 1, delete the of software as they will after the introduction of software. second sentence, which is also inconsistent with the guidance in NEI 96-07.

Diversity: Note from meeting:

1...of the design function does not appear in the "Diversity" is not defined within the regulations as a stand-alone term. The reference - Staff has clarified why it is suggested to be term is defined within the context of GDC 22, as follows: added

2. NRC to determine if this is an Enhancement - Staff clarified that excess redundancy raises more questions. -

"Criterion 22 -- Protection system independence. The protection system With the correction of item 1, NEI will keep the last sentence.

shall be designed to assure that the effects of natural phenomena, and Commented [A59]: Correction (b): GDC 22 is not a of normal operating, maintenance, testing, and postulated accident definition of diversity.

conditions on redundant channels do not result in loss of the protection Suggested replacement language - used as one example of function, or shall be demonstrated to be acceptable on some other a design technique defined basis. Design techniques, such as functional diversity or Commented [A60]: Correction (b): The term such as diversity in component design and principles of operation, shall be should be understood in the same way as the term for example not like the term that is and therefore the term used to the extent practical to prevent loss of the protection function." such as implicitly means that there are other modes of

[emphasis added] diversity, not just those listed.

D-33

Therefore, "diversity" is addressed in terms of functional or component design Commented [A61]: Correction (b): Delete this paragraph.

Therefore is inappropriate based on similar reasoning to and principles of operation. that in the comment on the term such as above.

The introduction of the exact same software and the potential creation of a software CCF into single-failure proof channels, or merely redundant channels, has no impact on diversity because the channels were not initially diverse. Namely, each of the channels used the same principles of operation and they all contained identical components. Thus, the channels were identical before the introduction of software and will remain identical after the introduction of software. Commented [A62]: Correction (b): NRC staff is not clear on the intent of this paragraph. Staff recommends deleting Separation: this paragraph.

This statement is misleading and therefore will lead to regulatory uncertainty.

"Separation" refers to physical arrangement to provide missile protection, or to eliminate or minimize the detrimental impacts due to fires, floods, etc. Protection systems should probably be dealt with differently than all other systems.

The introduction of the exact same software and the potential creation of a software CCF does not impact the physical arrangement of SSCs.

Independence:

"Independence" means non-interaction of SSCs. Commented [A63]: Correction (b): This definition is inconsistent with GDC 22, Protection system Assuming that no interactions (e.g., communication between multiple independence which includes protecting redundant channels from being affected by the same cause.

applications of the software) exist, the introduction of the exact same software and the potential creation of a software CCF does not impact the Furthermore, the term independence is further elaborated in regulations and guidance.

independence of SSCs. However, the failure of such software due to a software CCF is possible and is addressed in Evaluation criterion #5 and/or This over simplification is not acceptable because it distorts the meaning.

  1. 6.

Commented [A64]: Correction (b): Software CCF is a potential common cause for equipment failure, and is Determination of Attributable (i.e., Causality) therefore potential source of reduced independence.

NOTE: This guidance is not unique to digital and is the same as that provided in NEI 96-07, Section 4.3.2. This guidance is included here for completeness.

If none of the components/systems involved with the digital modification are identified as affecting a malfunction initiator previously identified in the Commented [A65]: Correction (b): The term malfunction initiator is new and inappropriate because, in part:

UFSAR, then there is no attributable impact on the likelihood of occurrence (1) Generally only malfunctions are identified in the of a malfunction. UFSAR, not the initiators of the malfunctions.

(2) it is not defined Alternately, if any components/systems involved with the digital modification (3) if question (ii) addresses initiators in the UFSAR, the question (vi) would be for all new ones, and the outcome are identified as affecting a malfunction initiator previously identified in the threshold for question (vi) is lower than for questions (ii).

UFSAR, then an impact on the likelihood of occurrence of a malfunction can be attributed to the digital modification. If an attributable impact is identified, then further assessment to determine the magnitude of the impact will be performed.

Example 4-13 illustrates a case of an attributable impact on the likelihood of D-34

occurrence of a malfunction.

Example 4-13. ATTRIBUTABLE Impact on the Likelihood of Occurrence of a Malfunction Proposed Activity Description Two safety-related containment chillers exist. There are two analog control systems (one per chiller) that are physically and functionally the same.

Each analog control system will be replaced with a separate digital control system. The hardware platform for each digital control system is from the same supplier and the software in each digital control system is exactly the same.

Affected Malfunctions and Malfunction Initiators The affected malfunction is the failure of a safety-related containment chiller to provide its cooling design function. The UFSAR identifies three specific equipment-related initiators of a containment chiller malfunction: (1) failure of the Emergency Diesel Generator (EDG) to start (preventing the EDG from supplying electrical power to the containment chiller it powers), (2) an electrical failure associated with the chiller system (e.g., feeder breaker failure), and (3) a mechanical failure within the chiller itself (e.g., flow blockage). The UFSAR also states that the single failure criteria were satisfied because two chillers were provided and there were no common malfunction sources.

Impact on Malfunction Likelihood Although the safety-related chiller control system is not one of the three malfunction initiators identified in the UFSAR, a new common malfunction source has been introduced due to the potential for a software common cause failure from the exact same software being used in both digital control systems. A common malfunction initiator was previously considered, but was concluded to be non-existent. However, this conclusion is no longer valid.

Therefore, an impact on the likelihood of occurrence of the malfunction can be attributed to the digital modification. (NOTE: The magnitude of the impact would be assessed next.)

Determination of Negligible/Discernable (i.e., Magnitude)

NOTE: The guidance in this sub-section applies ONLY when an attributable impact on the likelihood of occurrence of a malfunction has been established.

For proposed activities in which there is an attributable impact on the likelihood of occurrence of a malfunction, the negligible/discernable portion of the criteria (i.e., magnitude) also needs to be assessed.

D-35

To determine the overall effect of the digital modification on the likelihood of a malfunction, an engineering evaluation is performed. An engineering evaluation that uses a qualitative assessment to judge the failure likelihood of the modified SSC should consider each factor identified below:

  • Design attributes employed
  • Quality of the design processes, and
  • Operating experience of the software and hardware used (i.e., product maturity and in-service experience).

Negligible:

To achieve a negligible conclusion, the engineering evaluation of each factor (e.g., as documented in a qualitative assessment) would conclude that the affected SSC will exhibit a sufficiently low likelihood of failure, and by extension, that the change in the malfunction likelihood "...is so small or the uncertainties in determining whether a change in likelihood has occurred are such that it cannot be reasonably concluded that the likelihood has actually changed (i.e., there is no clear trend toward increasing the likelihood)"2

[emphasis added].

Therefore, if the qualitative assessment outcome is sufficiently low, there is not more than a minimal increase in the likelihood of occurrence of a malfunction of an SSC important to safety previously evaluated in the UFSAR.

Discernable:

If the examination of each factor concludes that the change in the malfunction likelihood exhibits a clear trend towards increasing the likelihood, then a discernable increase in the malfunction likelihood would exist. In this case, the software CCF likelihood could be sufficiently low or not sufficiently low. Commented [A66]: Correction (b): he failure likelihood cannot be both discernably increased and still be much The engineering evaluation (e.g., the qualitative assessment) is also used to lower than...

determine if the discernible increase in the malfunction likelihood is "more This same comment applies to Section 4.3.1.

than minimal" or "NOT more than minimal." To achieve a conclusion of Commented [A67]: Correction (b): Staff suggestion "NOT more than minimal," the proposed activity must continue to meet deletion of this sentence.

and/or satisfy all applicable NRC requirements, as well as design, material, Conflicts with paragraph above. In short, this paragraph and construction standards, to which the licensee is committed. Applicable states that a clear trend towards increasing is not related to sufficiently low of software CCF.

requirements and standards include those selected by the licensee for use in the development of the proposed digital I&C design modification and Furthermore this paragraph seems to be inconsistent with the definition of sufficiently low in Section 3.15.

documented within the design modification package.

See similar comment above in Section 4.3.1. This same comment applies to Section 4.3.1 as the same logic is also used in that section for this concept.

2 Refer to NEI 96-07, Section 4.3.2, 4th paragraph.

D-36

Examples 4-14 and 4-15 illustrate the negligible/discernable portion (i.e.,

magnitude) of the criteria and assume the attributable portion of the criteria has been satisfied.

Example 4-14 illustrates a case with a negligible change to the malfunction likelihood.

Example 4-14. NEGLIGIBLE Impact in the Likelihood of Occurrence of a Malfunction Proposed Activity Description Two non-safety-related main feedwater pumps (MFWPs) exist, each with its own flow control valve. There are two analog control systems (one per MFWP and flow control valve combination) that are physically and functionally the same.

Each analog control system will be replaced with a separate digital control system. The hardware platform for each digital control system is from the same supplier and the software in each digital control system is exactly the same.

Magnitude Conclusion A qualitative assessment performed as part of the design process, considering system design attributes, quality of the design processes employed, and operating experience of the proposed equipment, concluded that the failure likelihood of the modified SSC is sufficiently low.

All applicable requirements and other acceptance criteria to which the licensee is committed, as well as applicable design, material and construction standards, continue to be met.

Therefore, the change in the likelihood of occurrence of the loss of a MFWP or the closure of a MFWP flow control valve initiated by the failure of a feedwater control system is negligible due to the effect of the factors considered in the qualitative assessment.

Overall Conclusion Although an attributable impact on the likelihood of occurrence of the loss of a MFWP or the closure of a MFWP flow control valve was determined to exist, there was no clear trend toward increasing the likelihood. With no clear trend toward increasing the likelihood, there is not more than a minimal increase in the likelihood of occurrence of the malfunctions due to the digital modification.

Example 4-15 illustrates a case with a discernable increase to the malfunction likelihood.

D-37

Example 4-15. DISCERNABLE Increase in the Likelihood of Occurrence of a Malfunction Proposed Activity Description Two safety-related main control room chillers exist. There are two analog control systems (one per chiller) that are physically and functionally the same.

Each analog control system will be replaced with a separate digital control system. The hardware platform for each digital control system is from the same supplier and the software in each digital control system is exactly the same.

The logic components/system and controls for the starting and operation of the safety injection pumps are located within the main control room boundary. The environmental requirements associated with the logic components/system and controls are maintained within their allowable limits by the main control room cooling system, which includes the chillers involved with this digital modification.

Affected Malfunction and Malfunction Initiator The review of the UFSAR accident analyses identified several events for which the safety injection pumps are assumed to start and operate (as reflected in the inputs and assumptions to the accident analyses).

In each of these events, the UFSAR states the following: "To satisfy single failure requirements, the loss of only one chiller control system and its worst-case effect on the event due to the loss of one chiller has been considered in the accident analyses."

Magnitude Conclusion Based on the qualitative assessment performed as part of the technical assessment supporting this digital modification, the likelihood of a failure impacting both chiller control systems has been determined to be not sufficiently low.

The change in the likelihood of occurrence of the malfunction of both safety injection pumps is discernable due to the outcome of the qualitative assessment. Specifically, single failure criteria are no longer met.

Overall Conclusion An attributable impact on the likelihood of occurrence of the malfunction of both safety injection pumps was determined to exist and there is a clear trend toward increasing the likelihood. The clear trend toward increasing the likelihood (i.e., the discernable increase) is due to the failure being not D-38

sufficiently low which, in this case, causes single failure criteria to not be satisfied.

With a clear trend toward increasing the likelihood and failure to satisfy single failure criteria, there is more than a minimal increase in the likelihood of occurrence of the malfunction of both logic components/system and controls for the starting and operation of the safety injection pumps due to the digital modification.

HUMAN-SYSTEM INTERFACE ASSESSMENT If no personnel-based initiators involving degraded operator performance (e.g., operator error) are identified among the malfunction initiators, then an increase in the likelihood of the malfunction cannot occur due to the Human-System Interface portion of the digital modification. Otherwise, the application of the attributable criterion (i.e., causality) and the negligible/discernable criterion (i.e., magnitude) are assessed utilizing the guidance described in NEI 96-07, Section 4.3.2.

4.3.3 Does the Activity Result in More Than a Minimal Increase in the Consequences of an Accident?

There is no unique guidance applicable to digital modifications for responding to this Evaluation criterion because the identification of affected accidents and dose analysis inputs and/or assumptions are not unique for a digital modification. The guidance in NEI 96-07, Section 4.3.3 applies.

4.3.4 Does the Activity Result in More Than a Minimal Increase in the Consequences of a Malfunction?

There is no unique guidance applicable to digital modifications for responding to this Evaluation criterion because the identification of the affected malfunctions and dose analysis inputs and/or assumptions are not unique for a digital modification. The guidance in NEI 96-07, Section 4.3.4 applies.

4.3.5 Does the Activity Create a Possibility for an Accident of a Different Type?

INTRODUCTION From NEI 96-07, Section 3.2:

"The term 'accidents' refers to the anticipated (or abnormal) operational transients and postulated design basis accidents..."

D-39

Therefore, for purposes of 10 CFR 50.59, both Anticipated Operational Occurrences (AOOs) and Postulated Accidents (PAs) fall within the definition of "accident."

From NEI 96-07, Section 4.3.5, the two considerations that need to be assessed when answering this Evaluation question are as likely to happen as and the impact on the accident analyses (i.e., a new analysis will be required or a revision to a current analysis is possible).

GUIDANCE Determination of "As Likely To Happen As" From NEI 96-07, Section 4.3.5:

"The possible accidents of a different type are limited to those that are as likely to happen as those previously evaluated in the UFSAR. The accident must be credible in the sense of having been created within the range of assumptions previously considered in the licensing basis (e.g., random single failure, loss of off-site power, etc.)."

If the failure likelihood of the modified SSC is determined to be sufficiently low, then the activity does not introduce any failures that are as likely to happen as those in the UFSAR that can initiate an accident of a different type. Therefore, the activity does not create a possibility for an accident of a different type than any previously evaluated in the UFSAR. Commented [A68]: Correction (a): This paragraph appears to invoke the conclusion to rationale provided in Alternately, if the failure likelihood of the modified SSC is determined to be Supplement 1 to RIS 2002-22 with regard to 10 CFR 50.59 (c)(2)(v). Without including some of the rationale prior to, not sufficiently low, then the activity does introduce failures that are as likely this paragraph appears to jump to a conclusion.

to happen as those in the UFSAR that can initiate an accident of a different Commented [A69]: Correction (b): Please explain why type. In this case, further assessment to determine the impact on the software CCF was changed to failure in this version of the document. Check redline strike version.

accident analysis is to be performed.

Commented [A70]: Enhancement: This phrase can be deleted because it is not needed and is ambiguous. That is Determination of Accident Analysis Impact as written, this sentence can be interpreted in more than one way. One such interpretation is, in short, not NOTE: This guidance is not unique to digital and is the same as that sufficiently low means there is an accident of a different provided in NEI 96-07, Section 4.3.5, as clarified in RG 1.187. type.

Commented [A71]: Enhancement: It would be more clear For the case in which the creation of a possibility for an accident of a different to simply replace this phrase with whether the accident is type is as likely to happen as those in the UFSAR, the accident analysis of a different type.

impact also needs to be assessed to determine whether the accident is, in fact, Commented [A72]: Correction (b): Clarification of Section 4.3.5 of RG 1.187, as referenced here, has not been a different type. published. In order to avoid potential confusion and/or conflicts this guidance should be removed from this There are two possible impacts on the accident analysis: document.

(1) a revision to an existing analysis is possible, or (2) a new analysis will be required because the effect on the plant is different than any previously evaluated in the UFSAR D-40

Accidents of a different type are accidents that are as likely to happen as those in the UFSAR for which a new accident analysis would be needed, not just a revision of a current accident analysis.

Example 4-16 illustrates the NO CREATION of the possibility of an accident of a different type case.

Example 4-16. NO CREATION of the Possibility of an Accident of a Different Type Proposed Activity Two non-safety-related main feedwater pumps (MFWPs) exist, each with its own flow control valve. There are two analog control systems (one per MFWP and flow control valve combination) that are physically and functionally the same.

Each analog control system will be replaced with a separate digital control system. The hardware platform for each digital control system is from the same supplier and the software in each digital control system is exactly the same.

Malfunction / Accident Initiator The malfunction/accident initiator identified in the UFSAR for the analog main feedwater control system is the loss of one main feedwater pump (out of two pumps) due to the loss of one feedwater control system.

Accident Frequency and Type The pertinent accident is the Loss of Feedwater event. The characteristics of the Loss of Feedwater event are as follows:

Type of Accident - Decrease in Heat Removal by the Secondary System Accident Category - Infrequent Incident Commented [A73]: Correction (a): The examples should follow the guidance. The guidance say the first thing you As Likely to Happen As Conclusion do is assess likelihood, and if sufficiently low, then you are done.

Based on the qualitative assessment performed as part of the technical Therefore all of the material is not needed, and is assessment supporting this digital modification, the failure likelihood of the misleading in that it implies that you do all of this first before you assess likelihood or even if any new accidents modified SSC, causing the loss of both feedwater control systems (resulting in are sufficiently low.

the loss of both MFWPs) has been determined to be sufficiently low. Commented [A74]: Correction (a): - Consistency of language. Other example refer to this term as engineering Therefore, in this case, the creation of a possibility for an accident of a evaluation.

different type is NOT as likely to happen as those in the UFSAR and there is no need to determine the accident analysis impact.

Example 4-17 illustrates the CREATION of the possibility of an accident of a different type case.

D-41

Example 4-17. CREATION of the Possibility of an Accident of a Different Type Proposed Activity Two non-safety-related analog feedwater control systems and one non-safety-related main turbine steam-inlet valves analog control system exist.

The two feedwater control systems and the one main turbine steam-inlet valves control system will be combined into a single digital control system.

Malfunction / Accident Initiator The identified feedwater control system malfunctions include (a) failures causing the loss of all feedwater to the steam generators [evaluated in the Loss of Feedwater event] and (b) failures causing an increase in main feedwater flow to the maximum output from both MFWPs [evaluated in the Excess Feedwater event].

The identified main turbine steam-inlet valve control system malfunctions include (a) all valves going fully closed causing no steam to be admitted into the turbine [evaluated in the Loss of Load event] and (b) all valves going fully open causing excess steam to be admitted into the turbine [evaluated in the Excess Steam Demand event].

Accident Frequency and Type The characteristics of the pertinent accidents are as follows:

Loss of Feedwater:

Type of Accident - Decrease in Heat Removal by the Secondary System Accident Category - Infrequent Incident Excess Feedwater:

Type of Accident - Increase in Heat Removal by the Secondary System Accident Category - Moderate Frequency Incident Loss of Load:

Type of Accident - Decrease in Heat Removal by the Secondary System Accident Category - Moderate Frequency Incident Excess Steam Demand:

D-42

Type of Accident - Increase in Heat Removal by the Secondary System Accident Category - Moderate Frequency Incident As Likely to Happen As Conclusion Commented [A75]: Correction (a): The examples should follow the guidance. The guidance say the first thing you Based on the qualitative assessment performed as part of the technical go is assess likelihood, and if not sufficiently low, then you assess the impact. Much of the above information needed assessment supporting this digital modification, the failure likelihood of the to assess the impact and therefore it should follow the modified SSC impacting both the feedwater control systems and the main likelihood evaluation.

turbine steam-inlet valves control system has been determined to be not sufficiently low.

Therefore, in this case, the following conditions are as likely to happen as those in the UFSAR, creating a possibility for several accidents:

(1) Loss of both feedwater pumps (2) Increase in main feedwater flow to the maximum output from both MFWPs.

(3) All main turbine steam-inlet valves going fully closed (4) All main turbine steam-inlet valves going fully open (5) Combination of (1) and (3)

(6) Combination of (1) and (4)

(7) Combination of (2) and (3)

(8) Combination of (2) and (4)

Accident Analysis Impact Conclusion Conditions (1) though (4) are already considered in the safety analyses, so a revision to an existing analysis is possible. Thus conditions (1) through (4) are NOT accidents of a different type.

The current set of accidents identified in the safety analyses does not consider a simultaneous Feedwater event (i.e., Loss of Feedwater or Excess Feedwater) with a Main Steam event (i.e., Excess Steam Demand or Loss of Load).

Condition (5) still causes a decrease in heat removal by the secondary system.

Condition (6) involves both a decrease and an increase in heat removal by the secondary system.

Condition (7) involves both a decrease and an increase in heat removal by the secondary system.

D-43

Condition (8) still causes an increase in heat removal by the secondary system.

Conditions (5) though (8) will require new accident analyses to be performed. As such, conditions (5) though (8) are accidents of a different type. Therefore, the proposed activity does create the possibility of accidents of a different type.

HUMAN-SYSTEM INTERFACE ASSESSMENT If no personnel-based initiators involving degraded operator performance (e.g., operator error) are identified as accident initiators, then the creation of a possibility for an accident of a different type cannot occur due to the Human-System Interface portion of the digital modification. Otherwise, the creation of a possibility for an accident of a different type is assessed utilizing the guidance described in NEI 96-07, Section 4.3.5.

Commented [A76]: Correction (a): NRC staff is recommending the deletion of this section from Appendix 4.3.6 Does the Activity Create a Possibility for a Malfunction of an SSC Important D. However, should NEI wish to move forward with this to Safety with a Different Result? section, the following comments will need to be resolved.

Note from meeting:

INTRODUCTION NEI stated that they will keep the section and address the comments.

NOTE: Due to the unique nature of digital modifications, and the inherent NEI will add to the note, to the affect, section 4.3.6 is not to complexities therein, the application of this criterion is especially be construed as generic guidance.

important. Specifically, the unique aspect of concern is the potential Commented [A77]: Correction (a): Contrary to 50.34(b),

for a software CCF to create the possibility for a malfunction with a 50.59, and NEI 96-07. 50.34(b) states "The FSAR shall include information that describes the facility, presents the different result. Therefore, rather than simply providing design bases and the limits on its operation, and presents supplemental guidance to that already included in NEI 96-07, the safety analyses of the structures, systems and components and of the facility as a whole." Appendix D Section 4.3.6, more detailed guidance will be provided in this section. incorrectly considers that different result related only to However, none of the more detailed guidance provided in this safety analysis for the facility as a whole but not to the safety analyses of the structures, systems and components.

section conflicts with that provided in NEI 96-07, Section 4.3.6, or NEI 96-07 definition for safety analyses states that also should be construed as being new, or modified from that in NEI 96- within the meaning of the definition are supporting UFSAR analyses that demonstrate SSC design functions will be 07, Section 4.3.6. accomplished as credited in the accident analyses.

Malfunctions of structures, systems and components are Review generally postulated as potential single failures.

Commented [A78]: Correction (a):

To ensure the unique aspects of digital modifications are addressed correctly Partial quotations, taken out of context, maybe and adequately, a review of selected discussions and excerpts from NEI 96- misinterpretations. It is far better to refer the reader to 07, including malfunctions, design functions, and safety analyses, is the original document, for careful study. That is, these partial quotations will lead to regulatory uncertainty.

presented first.

These partial quotations are used to support conclusions for From NEI 96-07, Section 3.9: which the staff has concerns (e.g. NEI does not provide a rationale of why the quotes are being used.)

Malfunction of SSCs important to safety means the failure of SSCs to Note from meeting:

perform their intended design functions described in the UFSAR This is a general comment that will be discussed with more specificity in comments below.

D-44

(whether or not classified as safety-related in accordance with 10 CFR 50, Appendix B). [emphasis added]

From NEI 96-07, Section 3.3:

Design functions are UFSAR-described design bases functions and other SSC functions described in the UFSAR that support or impact design bases functions... [emphasis added]

Also, Design bases functions are functions performed by systems, structures and components (SSCs) that are (1) required by, or otherwise necessary to comply with, regulations, license conditions, orders or technical specifications, or (2) credited in licensee safety analyses to meet NRC requirements. [emphasis added]

Furthermore, Design functions...include functions that, if not performed, would initiate a transient or accident that the plant is required to withstand.

[emphasis added]

Finally, As used above, credited in the safety analyses means that, if the SSC were not to perform its design bases function in the manner described, the assumed initial conditions, mitigative actions or other information in the analyses would no longer be within the range evaluated (i.e., the analysis results would be called into question). The phrase support or impact design bases functions refers both to those SSCs needed to support design bases functions (cooling, power, environmental control, etc.) and to SSCs whose operation or malfunction could adversely affect the performance of design bases functions (for instance, control systems and physical arrangements).

Thus, both safety-related and nonsafety-related SSCs may perform design functions. [emphasis added]

This definition is oriented around the definition of design bases function, which itself is defined in NEI 97-04, Appendix B, Guidelines and Examples for Identifying 10 CFR 50.2 Design Bases, endorsed by Regulatory Guide 1.186, and highlighted in bold above.

A more complete understanding of the meaning of a design basis function can be obtained by examination of NEI 97-04, Appendix B. From NEI 97-04, the three characteristics of design bases functions are summarized as follows:

1. Design bases functions are credited in the safety analyses.

D-45

2. The functions of any individual SSC are functionally below that of a design basis function.
3. Design bases functions are derived primarily from the General Design Criteria.

Repeating a portion from above to highlight the importance of identifying the design basis function and its connection to a safety analysis result, we have Commented [A79]: Correction (a): 10 CFR 50.59 is about design functions, not a subset of design functions. By the following: focusing on a design basis function and its connections to a safety analysis result an implementer may understand As used above, credited in the safety analyses means that, if this guidance as directing him to ignore the other design the SSC were not to perform its design bases function in the functions and other results.

manner described, the assumed initial conditions, mitigative The NRC made a presentation in a public meeting on actions or other information in the analyses would no longer be November 30, 2017 (ML17335A574), which expressed concern with this approach.

within the range evaluated (i.e., the analysis results would be called into question). [emphasis added] This section should be revised to provide fuller context of quotes or deleted per general comment above re: this section.

Then, from NEI 96-07, Section 3.12:

Note from meeting:

Safety analyses are analyses performed pursuant to NRC NEI believes that this comment is addressed in the requirements to demonstrate the integrity of the reactor coolant proceeding discussion with Appendix D. Intent is to augment NEI 96-07.

pressure boundary, the capability to shut down the reactor and NEI does not believe they are focusing only design basis maintain it in a safe shutdown condition, or the capability to function.

prevent or mitigate the consequences of accidents that could NRC states - Focus on what are the results of the specific result in potential offsite exposures comparable to the guidelines failure, single failure.

Expecting broader guidance.

in 10 CFR 50.34(a)(1) or 10 CFR 100.11...and include, but are Staff reiterated the position that this is an example of the not limited to, the accident analyses typically presented in type of comment that will take time, beyond Dec 2018, to resolve. The concern has been identified that this section is Chapter 15 of the UFSAR. [emphasis added] not digital specific.

And from the first sentence of the associated discussion: NEI may provide a preamble in the note that although this guidance can be used in digital, it should not also be used Safety analyses are those analyses or evaluations that with non-digital guidance.

demonstrate that acceptance criteria for the facilitys capability to withstand or respond to postulated events are met.

[emphasis added]

Failure Modes and Effects Analysis (FMEA)

NEI 96-07, Section 4.3.6 recognizes that the effect of a proposed modification must be assessed. This assessment may require the use of a failure modes and effects analysis (FMEA), including the possible creation of a new FMEA.

From NEI 96-07, Section 4.3.6:

In evaluating a proposed activity against this criterion, the types and results of failure modes of SSCs that have previously been evaluated in the UFSAR and that are affected by the proposed activity should be identified. This evaluation should be performed consistent with any failure modes and effects analysis (FMEA) described in the UFSAR, D-46

recognizing that certain proposed activities may require a new FMEA to be performed. [emphasis added]

Overall Perspective NEI 96-07, Section 4.3.6 provides the overall perspective on this Evaluation criterion with its first sentence, which states:

Malfunctions of SSCs are generally postulated as potential single failures to evaluate plant performance with the focus being on the result of the malfunction rather than the cause or type of malfunction.

Expanding upon this foundation, the following conclusion is reached, which is based upon discussion from 63 FR 56106:

Unless the equipment would fail in a way not already evaluated in the safety analysis, there can be no malfunction of an SSC important to safety with a different result. [emphasis added]

From NEI 96-07, Section 4.3.6, there are two considerations that need to be assessed when answering this criterion: as likely to happen as and impact on the safety analysis result.

GUIDANCE Determination of "As Likely to Happen As" From NEI 96-07, Section 4.3.6:

The possible malfunctions with a different result are limited to those that are as likely to happen as those described in the UFSARa proposed change or activity that increases the likelihood of a malfunction previously thought to be incredible to the point where it becomes as likely as the malfunctions assumed in the UFSAR could create a possible malfunction with a different result. [emphasis added]

If the failure likelihood of the modified SSC is determined to be sufficiently low, then the activity does not introduce any failures that are as likely to happen as those in the UFSAR. Therefore, the activity does not create a possibility for a malfunction of an SSC important to safety with a different result from any other previously evaluated in the UFSAR.

Alternately, if the failure likelihood of the modified SSC is determined to be not sufficiently low, then the activity does introduce failures that are as likely to happen as those in the UFSAR. In this case, further assessment to determine the impact of the malfunction on the safety analysis result is to be performed.

D-47

Example 4-18 illustrates the NO CREATION of the possibility for a malfunction with a different result due to applying the as likely to happen as consideration.

Example 4-18. NO CREATION of the Possibility for a Malfunction with a Different Result Proposed Activity A large number of analog transmitters are being replaced with digital transmitters. These transmitters perform a variety of functions, including controlling the automatic actuation of devices (e.g., valve stroking) that are credited in a safety analysis.

Conclusion Based on the qualitative assessment, the failure likelihood of the modified SSC has been determined to be sufficiently low.

Therefore, a malfunction with a different result is NOT as likely to happen as those described in the UFSAR and there is no need to determine the impact of the malfunction on the safety analysis result. Thus, the proposed activity does NOT create the possibility for a malfunction of an SSC important to safety with a different result.

Determination of Safety Analysis Result Impact Commented [A80]: Correction (a): This is an open issue documented in the meeting summary from the 11/30/2017 The generic process to determine the impact of a malfunction of an public meeting (ML17331A485), which states:

SSC important to safety on the safety analyses, i.e., a comparison of The NRC staff pointed out that Title 10 to the Code of the safety analyses results to identify any different results, consists of Federal Regulations (10 CFR), Section 50.59, (10 CFR 50.59) Changes, test and experiments, used the term multiple steps, as summarized next. final safety analysis report (as updated) while NEI 96-07, Appendix D, Section 4.3.6 used the terms safety analysis Step 1: Identify the functions directly or indirectly related to the and accident analysis. The NRC staff said that it could be understood that accident analysis is a subset of safety proposed modification. analysis which is a subset of final safety analysis report (as updated). Using more restrictive terms, it could be Considering the scope of the proposed digital modification, identify the understood that the evaluation guidance only addressed a functions that are directly or indirectly related to the proposed activity. subset of any [malfunction] previously evaluated in the final safety analysis report (as updated).

Step 2: Identify which of the functions from Step 1 are Design Note from meeting:

Functions and/or Design Bases Functions. Need to revisit this item in a separate discussion.

Staff to identify specifically what has been left out of this Utilizing NEI 96-07, Section 3.3, classify the functions from Step 1. If discussion and is necessary to resolve the comment.

Example of missing information - 3rd bullet of NEI 96-07 no design functions are identified, then the proposed activity does NOT Section 3.12 create the possibility for a malfunction of an SSC important to safety with a different result.

Utilizing NEI 96-07, Section 3.3, along with Appendix B to NEI 97-04, as needed, identify which design functions are design bases functions, which design functions support or impact design bases functions, and which design functions are not involved with design bases functions, D-48

but are functions that if not performed would initiate a transient or Commented [A81]: Correction (a): The meeting summary accident that the plant is required to withstand. If no design basis from the 11/30/2017 public meeting (ML17331A485), states:

The NRC staff noted that Step 2 did point to the functions are involved, proceed to Step 5. appropriate guidance for identifying design bases functions, more specific guidance used the 10 CFR General Design The process for determining if a design function is a design basis Criteria (GDC) for identification purposes. The NRC staff said that a subset of design bases function can be tied to function is aided by identifying the associated General Design Criteria the GDC. NEI stated that the vast majority of design basis (GDC) to which a design bases function applies. Each design function functions are derived from the GDCs.

can then be related to the requirements discussed within the GDC to The NRC made a presentation in the public meeting on determine if that design function is directly involved with the design November 30, 2017 (ML17335A574), which expressed concern with Using GDCs to Identify Design Basis basis function itself or if the design function supports or impacts the Functions Vs. UFSAR to Identify Design Functions related design basis function. If the design function is found to directly Based on these interactions, the NRC and NEI are still at involve the GDC requirement, then that design function is a design an impasse over this issue.

basis function. If the design function supports or impacts the GDC Note from meeting:

requirement, then it is not a design basis function, but is still credited NEIs intent is not to narrow the scope of the guidance.

in the safety analysis.

NEI to find a way to bridge GDC to principle design criteria. NEI 96-04, Section 3.3 as a reference.

Step 3: Determine if a new FMEA needs to be generated.

Commented [A82]: Correction (a): #82 If the impact on the design basis function involved is readily apparent, NEI 96-07, Revision 1, Section 4.3.2, Example 4, provides guidance on new operator actions as follows with regards to no new FMEA needs to be generated, skip this step and go to Step 4. interdependence:

For example, there is no reason to contemplate the generation of a new The change involves a new or modified operator action that FMEA if the impact of the failure on the design bases functions is supports a design function credited in safety analyses recognized as being immediate. Otherwise, generate the new FMEA to [would not require prior NRC approval] provided:

describe the connection of the proposed activity, or failures due to the *The action (including required completion time) is proposed activity, to an impact on the design bases functions. reflected in plant procedures and operator training programs As part of the process for generating the new FMEA, presume

  • The licensee has demonstrated that the action can be compliance with pre-existing/interdependent, modification-related completed in the time required considering the aggregate procedures and utilization of existing equipment to determine if affects, such as workload or environmental conditions, expected to exist when the action is required.

adequate options exist to mitigate potential detrimental impacts on o To illustrate this point - Common cause failures in design functions. redundant systems compromise safety if the failures are concurrent failures, that is, failures which occur over a time interval during which it is not plausible Interdependence is discussed in NEI 96-07, Sections 4.2 and 4.3. An ... [1]

example of an interdependent procedure change would be the Commented [A83]: #83 Interdependence paragraph modifications to an existing procedure to reflect operation of the new Correction (a):

Please clarify the concept to adequate options. NEI 01-01 digital equipment and controls, including any new features such as a specifically references adequate backups with regard the control system restart option. determination of potential new outcomes and effects on design functions being preserved or not. Its not clear how Step 4: Determine if each design basis function continues to be a system restart or reboot constitutes an adequate backup to the potentially failed SSC.

performed/satisfied.

As oppose to using the term adequate options should be If all design basis functions continue to be performed/satisfied, and replaced with specific guidance that described acceptable there are no other design functions involved, then the proposed activity options, which are already provided in NEI 96-07, Revision 1, Section 4.3.2, Example 4.

does NOT create the possibility for a malfunction of an SSC important to safety with a different result. Specific compensatory measure, such as opening doors, turning on fans, or installing temporary fans, is not acceptable because those actions are not interdependent with the modifications, in that the actions were added with the purpose of avoiding a license amendment.

... [2]

D-49

For any design basis function that does not continue to be performed/satisfied, or other design functions that are involved, continue to Step 5.

Step 5: Identify all safety analyses involved. Commented [A84]: Correction (a): Contrary to NEI 96-07 and 10 CFR 50.59.

Identify all safety analyses that rely directly or indirectly on the design basis functions performance/satisfaction. Also, identify all safety NEI 96-07, section 4.3.6.2, states, The safety analysis assume certain design functions of SSCs in demonstrating analyses related to any other design function that could impact either the adequacy of the design. Thus, certain design functions, the accidents initiation or the events initial conditions, i.e., design while not specifically identified in the safety analysis, are credited in a direct sense.

functions that, if not performed, would initiate a transient or accident that the plant is required to withstand. Note from meeting:

Correction - NEI96-07, Section 4.3.2 If there are no safety analyses involved, then there has been no change Commented [A85]: Correction (a):

Questions (vi) and (vi) are meant to address new or in the result of a safety analysis and the proposed activity does NOT different accidents and malfunctions; there will never be create the possibility for a malfunction of an SSC important to safety any pre-existing safety analysis for new types of events created by a change. Based upon the reasoning stated with a different result. here, it could be potentially be understood that malfunctions such as CCF would not be considered a Step 6: For each safety analysis involved, compare the different result if not previously analyzed. This would be projected/postulated results with the previously evaluated results. contrary to Questions (vi) under 50.59.

(a)Note from meeting:

NEI 96-07, Section 4.3.6 provides the following guidance regarding the Correction: Question (v) and (vi) identification of failure modes and effects: NEI does not agree that Criteria (vi) is valid for the highlighted information.

Once the malfunctions previously evaluated in the UFSAR and NRC suggested adding words - all should be caveated the results of these malfunctions have been determined, then with new or existing malfunctions. Make sure that the the types and results of failure modes that the proposed activity language includes the words in the FSAR. NRC action to provide additional context to this item by way of example.

could create are identified.

Not all design functions are credited in the safety analysis.

If any of the identified safety analyses have become invalid due to their Some design functions are required for other reasons (e.g.,

basic assumptions no longer being valid (e.g., single failure assumption to meet regulations).

is not maintained), or if the numerical result(s) of any safety analysis (b)Note from meeting:

would no longer satisfy the acceptance criteria, then the proposed NEI agrees, but request NRC suggestion.

activity DOES create the possibility for a malfunction of an SSC Some design functions exist to prevent the possibility of a important to safety with a different result. DBA or AOO (e.g., control rod withdrawal prohibit or permissive); therefore no safety analysis exists for said DBA or AOO.

As part of the response and determining if the safety analyses acceptance criteria continue to be satisfied, include the impact on the (c)Note from meeting:

NEI agrees, but request NRC suggestion.

severity of the initiating conditions and the impact on the initial conditions assumed in the safety analysis. Specifically, consider any Example 4-21 points to one of the concerns with the reasoning embodied in this step. One of the ideas is that a design functions that, if not performed, would initiate a transient or change should not more than minimally impact the accident that the plant is required to withstand. consequences which in 50.59 means dose.

Consequences is a subset of results. Radiation monitors Examples 4-19 through 4-24 illustrate cases in which the failure likelihood of are used, in part, to limit dose, so a misbehavior in that system could adversely impact dose.

the modified SSC is determined to be not sufficiently low and, by extension, the likelihood of a software CCF is also determined to be not sufficiently low, (d)Note from meeting:

NEI - Consequences go in step 6 of 4.3.6 but rad monitors i.e., a malfunction is as likely to happen as those described in the UFSAR. In are not credited in the safety analysis.

these cases, the multi-step process applying the safety analysis result D-50

impact consideration is performed to determine the impact of the malfunction on the safety analysis result.

Examples 4-19 through 4-23 illustrate some cases of NO CREATION of a malfunction with a different result.

Example 4-19. NO CREATION of a Malfunction with a Different Result Proposed Activity A feedwater control system is being upgraded from an analog system to a digital system. New components are being added that could fail in ways other than the components in the original design. Now, as a result of this change, all four feedwater flow control valves could simultaneously fail closed following a software CCF.

Safety Analysis Result Impact Consideration Step 1:

The identified function is to establish and maintain steam generator water level within predetermined physical limits during normal operating conditions.

Step 2:

The function is classified as a design function due to its ability to initiate a transient or accident that the plant is required to withstand. However, the design function is not a design basis function. With no design basis functions involved, proceed to Step 5.

Step 3:

Not applicable Step 4:

Not applicable Step 5:

The pertinent safety analysis is the accident analysis for Loss of Feedwater. The feedwater control system has a direct impact on the accident analysis assumptions and modeling.

Step 6:

The severity of the initiating failure for the Loss of Feedwater is unchanged. The event already assumes a total loss of feedwater flow.

The newly created failure modes are determined to have no effect on this assumption. The manner in which feedwater flow is lost has no impact on the initial conditions of the event.

D-51

Conclusion Although the software CCF likelihood was determined to be not sufficiently low (i.e., a malfunction is as likely to happen as those described in the UFSAR), the initiation severity of the Loss of Feedwater event, the newly created failure modes and the manner in which feedwater flow was lost do not change the result of the safety analysis. Thus, the proposed activity does NOT create the possibility for a malfunction of an SSC important to safety with a different result.

Example 4-20. NO CREATION of a Malfunction with a Different Result Proposed Activity A feedwater control system is being upgraded from an analog system to a digital system. Previously, only one of four feedwater flow control valves was assumed to fail open as part of the initiation of the Excess Feedwater event. Now, as a result of this change, all four feedwater flow control valves could simultaneously fail open following a software CCF.

Safety Analysis Result Impact Consideration Step 1:

The identified function is to establish and maintain steam generator water level within predetermined physical limits during normal operating conditions.

Step 2:

The function is classified as a design function due to its ability to initiate a transient or accident that the plant is required to withstand. However, the design function is not a design basis function. With no design basis functions involved, proceed to Step 5.

Step 3:

Not applicable Step 4:

Not applicable Step 5:

D-52

The pertinent safety analysis is the accident analysis for Excess Feedwater. The feedwater control system has a direct impact on the accident analysis assumptions and modeling.

Step 6:

The severity of the initiating failure has increased due to four valves supplying flow as compared to one valve prior to the change.

The minimum allowed departure from nucleate boiling ratio (DNBR) to satisfy the accident analysis acceptance limit is 1.30. The current minimum DNBR result is 1.42. After using an increased value for the new feedwater flow (to represent the increase in feedwater flow caused by the opening of the four feedwater flow control valves) in a revision to the Excess Feedwater accident analysis, the new minimum DNBR result is 1.33.

Conclusion Although the software CCF likelihood was determined to be not sufficiently low (i.e., a malfunction is as likely to happen as those described in the UFSAR) and the severity of the initiating failure has increased, the new minimum DNBR result continues to satisfy the accident analysis acceptance limit, which does not change the result of the safety analysis. Therefore, the Commented [A86]: Correction (a): Contrary to 10 CFR 50.59 regulation and NEI 96-07 which states that the proposed activity does NOT create the possibility for a malfunction of an SSC results must be bounded by the previous analysis, not important to safety with a different result. accident analysis acceptance criteria.

Example 4-21. NO CREATION of a Malfunction with a Different Result Proposed Activity A complete system upgrade to the area radiation monitors that monitor a variety of containment compartments that could be subject to radioactive releases during a LOCA is proposed. The outdated analog-based radiation monitors are being replaced by digitally-based monitors. The hardware platform for each area radiation monitor is from the same supplier and the software in each area radiation monitor is exactly the same.

Safety Analysis Result Impact Consideration Step 1:

The functions include the monitoring of the various compartments, rooms and areas that may be subject to an increase in radiation during the recirculation phase of a LOCA.

D-53

Step 2:

In this case, whether the function is a design bases function is not readily determined, so the associated GDC will be identified and examined.

Criterion 64 -- Monitoring radioactivity releases. Means shall be provided for monitoring the reactor containment atmosphere, spaces containing components for recirculation of loss-of-coolant accident fluids, effluent discharge paths, and the plant environs for radioactivity that may be released from normal operations, including anticipated operational occurrences, and from postulated accidents.

[emphasis added]

The area radiation monitors perform a function that is necessary to comply with a requirement specified in GDC 64. Therefore, the radiation monitor's function is directly involved with a design basis function.

Step 3:

No new FMEA needs to be generated. The effect of a postulated software CCF on the design basis function involved is readily apparent.

Step 4:

If a software CCF occurs, the area radiation monitors will not perform their design function that supports or impacts a design basis function.

Thus, the design basis function will not be performed/satisfied.

Step 5:

There are no safety analyses that directly or indirectly credit this design basis function. That is, there are no considerations of malfunctions of single or multiple radiation monitors in any safety analysis.

Step 6:

Not applicable Conclusion The cited GDC does not contain any reference to single failure protection, so there is no distinction between a failure of a single radiation monitor or multiple radiation monitors.

Although the software CCF likelihood was determined to be not sufficiently low (i.e., a malfunction is as likely to happen as those described in the UFSAR), there are no safety analyses that directly or indirectly credit the design basis function. Thus, there cannot be a D-54

different result when comparing to a pre-existing safety analysis since none exists. Commented [A87]: Correction (a):

This example points one of the concerns with the reasoning Therefore, the proposed activity does NOT create the possibility of a embodied in the multi-step process. One of the ideas is that a change should not more than minimally impact the malfunction of an SSC important to safety with a different result. consequences which in 50.59 means dose.

Consequences is a subset of results. The radiation NOTE: The acceptability of these new area radiation monitors will be monitors are used, in part, to limit dose, so a misbehavior in that system could adversely impact dose.

dictated by their reliability, which is assessed as part of Criterion #2, not Criterion #6. Questions (vi) and (vi) are meant to address new or different accidents and malfunctions; there will never be any pre-existing safety analysis for new types of events created by a change.

Note from meeting:

Example 4-22. NO CREATION of a Malfunction with a Different Result NEI believes that consequences are address through other criteria. The last 1/2 of 1st paragraph of this comment does Proposed Activity not follow the 50.59 regulation. See comment A85 from NEI tracking sheet.

Two chillers that cool the Main Control Room Ventilation System (MCRVS) are being upgraded. As part of the upgrade, each analog control system will be replaced with a digital control system. Each digital control system maintains all of the operational features (e.g.,

auto/manual start/stop, setpoints and alarms) as the analog control systems. The hardware platform for each chiller control system is from the same supplier and the software in each chiller control system is exactly the same.

Safety Analysis Result Impact Consideration Step 1:

The MCRVS also cools the Relay Room that is adjacent to the main control room. The Relay Room contains multiple instrument racks that control both Reactor Protection and Safeguards actuation signals.

The air flow path from the Main Control Room to the Relay Room is described in the UFSAR, along with a function to maintain the Relay Rooms temperature less than or equal to 120 ºF.

Step 2:

In this case, whether the function is a design bases function is not readily determined, so the associated GDC will be identified and examined.

Criterion 20 -- Protection system functions. The protection system shall be designed (1) to initiate automatically the operation of appropriate systems including the reactivity control systems, to assure that specified acceptable fuel design limits are not exceeded as a result of anticipated operational occurrences and (2) to sense accident conditions and to initiate the operation of systems and components important to safety. [emphasis added]

D-55

The chillers and the chiller control systems perform a function that supports or impacts the design basis function specified in GDC 20. Therefore, the chillers and the chillers control systems functions are design functions credited in the safety analysis.

Step 3:

The impact of a software CCF on the design bases functions is not readily apparent, so a new FMEA was generated.

Step 4:

The new FMEA concluded that compliance with pre-existing procedures will result in the restoration of at least one chiller well before the Relay Room cooling becomes inadequate. Specifically, compliance with existing procedures will direct the recognition of the problem and the restoration of the chillers function prior to the impairment of the associated design basis functions. In addition, an interdependent procedure change involved the use of the control system restart feature to reinitialize the control system, which would clear any software faults, allowing the chiller functions to be restored well before the Relay Room cooling becomes inadequate. Commented [A88]: Correction (a):

See comment above on Interdependence. Its not clear that, Step 5: based upon the guidance on Interdependence above and the content presented in this example, that the example adequately captures the necessary aspects of Although none of the safety analyses specifically identify assumptions interdependence procedural changes. Also reflects staffs or inputs related to the MCRVS, the Relay Room or the components concerns with partial quotations from NEI 96-07 out of context that potentially leads to misinterpretations and therein, several accident analyses assume correct and timely actuation ultimately, regulatory uncertainty.

of the Reactor Protection and Safeguards features. As determined in Note from meeting:

Step 2 above, the chillers operation is considered to be credited in the 1st and 3rd sentences are related.

safety analysis since they support or impact the design bases The middle sentence is the substance for the correction needed. The other sentences were added for context.

functions associated with GDC 20. As demonstrated as part of Step 4, all design basis functions are preserved.

Step 6:

As determined in Step 4, all design basis functions are preserved.

Therefore, all of the safety analyses identified in Step 5 remain valid and there is no change in any safety analysis result.

Conclusion Although the software CCF likelihood was determined to be not sufficiently low (i.e., a malfunction is as likely to happen as those described in the UFSAR), the design bases functions will continue to be performed/satisfied and the safety analyses (and all of the results from these analyses) are unaffected. Therefore, the proposed activity does D-56

NOT create the possibility of a malfunction of an SSC important to safety with a different result.

Example 4-23. NO CREATION of a Malfunction with a Different Result Proposed Activity Currently, the feedwater control system and the pressurizer pressure control system are separate analog control systems.

The feedwater control system is being upgraded from an analog to a digital system. Previously, only one of four feedwater flow control valves was assumed to fail open as part of the initiation of the Excess Feedwater event. Now, as a result of this change, all four feedwater flow control valves could simultaneously fail open following a software CCF.

The pressurizer pressure control system is being upgraded from an analog to a digital system.

As part of this modification, the two previously separate control systems will be combined within the same digital controller in a distributed control system (DCS) with the same software controlling all feedwater and pressurizer functions.

Safety Analysis Result Impact Consideration Step 1:

Feedwater - The identified function is to establish and maintain steam generator water level within predetermined physical limits during normal operating conditions.

Pressurizer - The identified function is control of the pressurizer sprays and heaters.

Step 2:

Feedwater - The function is classified as a design function due to its ability to initiate a transient or accident that the plant is required to withstand.

However, the design function is not a design bases function.

Pressurizer - In this case, whether the function is a design bases function is not readily determined, so the associated GDC will be identified and examined.

Criterion 10 -- Reactor design. The reactor core and associated coolant, control, and protection systems shall be designed with D-57

appropriate margin to assure that specified acceptable fuel design limits are not exceeded during any condition of normal operation, including the effects of anticipated operational occurrences. [emphasis added]

The pressurizer control system performs a function that supports or impacts a design basis function specified in GDC 10. Therefore, the pressurizer control systems function is a design function and is credited in the safety analysis.

Step 3:

The effect on the feedwater and pressurizer control systems is clear and understood, having a direct impact on the accident analysis assumptions and modeling. There is no reason to contemplate the generation of a new FMEA since the impact of the software CCF on the accident analysis is readily apparent (i.e., clear and understood).

Step 4:

If a software CCF occurs, the pressurizer pressure control function, which supports or impacts the GDC 10 design basis function, will not continue to be performed/satisfied. Commented [A89]: Correction (a):

Step 5: Please clarify in this step why only the impact on pressurizer pressure controls is considered affected by a software CCF when step 1 clearly states that pressurizer The pertinent safety analysis is the accident analysis for Excess and feedwater controls are integrated into the same DCS.

Feedwater. Typically, in Chapter 15 accident analyses control system What is the basis for only pressurizer controls being affected?

action is considered only if that action results in more severe accident results. The feedwater and pressurizer control systems have a direct See GDC comment on page D-47 of this document.

impact on the accident analysis assumptions and modeling. Note from meeting:

Staff does not believe that this step is representative the Step 6: level of complexity of mods that are being conducted in the field.

In the Excess Feedwater accident analysis, the initial conditions NEI requesting NRC to suggest a more complex example.

already assume abnormally low pressure and/or DNBR. Since the pressurizer pressure control system would mitigate the results of the GC and PDC item will discuss as previously stated in prior applicable comment.

accident, no credit is taken for operation of the pressurizer pressure control system. Therefore, a malfunction of the control system would Suggested language - If multiple design functions are identified then each might follow a separate a path or have no effect on this event and no effect on the safety analysis result. propagate each through the process.

The severity of the initiating failure is not affected due to the combination of the two control systems. The minimum allowed DNBR to satisfy the accident analysis acceptance limit is 1.30. The current minimum DNBR result is 1.42. After using an increased value for the new feedwater flow (to represent the increase in feedwater flow caused by the opening of the four feedwater flow control valves) and adjusting the appropriate inputs to reflect the new detrimental pressurizer D-58

heater and spray conditions in a revision to the Excess Feedwater accident analysis, the new minimum DNBR result is 1.33.

Conclusion Commented [A90]: Correction (a):

This example and conclusion appears to presume that, a With the software CCF likelihood determined to be not sufficiently low (i.e., a DCS with combined functionality of different SSC design functions will still fail in a preexisting or known way such malfunction is as likely to happen as those described in the UFSAR), the as the pressurizer sprays, heaters, etc. would still fail in a severity of the initiating failure has increased. The impairment of the previously anticipated way even though a different technology is now controlling its functionality and is no pressurizer pressure control function is already incorporated in the safety longer a separate function. Its not clear that this example analysis modeling. The new minimum DNBR result continues to satisfy the takes into account spurious actuation of various components of the system or whether the example consider accident analysis acceptance limit, which does not change the result of the the Chapter 15 analysis as the still most limiting failure for safety analysis. Therefore, the proposed activity does NOT create the the DCS.

possibility of a malfunction of an SSC important to safety with a different Note from meeting:

result. NEI agrees with the technical content of this comment.

Staff agrees with the conclusion but have concern with how the conclusion was reached.

Example 4-24 illustrates a case in which there is the CREATION of a malfunction with a different result. NEI to add... for this specific aspect needed to resolve this comment Example 4-24. CREATION of a Malfunction with a Different Result Commented [A91]: Correction (a): See comments previously stated in example 4-20.

Proposed Activity An upgrade to the analog-based reactor protection system with a digital-based reactor protection system is proposed. This proposed modification involves replacement of all the solid state cards that control the detection of anticipated operational occurrences and the actuation of the required reactor trip signals. Redundant channels contain these cards in satisfaction of single failure criteria.

Safety Analysis Result Impact Consideration Step 1:

The number of involved functions is large, all of which involve the detection of the occurrence of anticipated operational occurrences, the processing of those signals, and the generation of the appropriate reactor trip signals.

Step 2:

In this case, whether the functions are design bases function is not readily determined, so the associated GDCs will be identified and examined.

Criterion 20 -- Protection system functions. The protection system shall be designed (1) to initiate automatically the operation of appropriate systems including the reactivity control systems, to assure that specified acceptable fuel design limits are not exceeded as a result of anticipated operational occurrences and (2) to sense accident D-59

conditions and to initiate the operation of systems and components important to safety. [emphasis added]

Criterion 21 -- Protection system reliability and testability. The protection system shall be designed for high functional reliability and inservice testability commensurate with the safety functions to be performed. Redundancy and independence designed into the protection system shall be sufficient to assure that (1) no single failure results in loss of the protection function and (2) removal from service of any component or channel does not result in loss of the required minimum redundancy unless the acceptable reliability of operation of the protection system can be otherwise demonstrated. The protection system shall be designed to permit periodic testing of its functioning when the reactor is in operation, including a capability to test channels independently to determine failures and losses of redundancy that may have occurred. [emphasis added]

Criterion 22 -- Protection system independence. The protection system shall be designed to assure that the effects of natural phenomena, and of normal operating, maintenance, testing, and postulated accident conditions on redundant channels do not result in loss of the protection function, or shall be demonstrated to be acceptable on some other defined basis. Design techniques, such as functional diversity or diversity in component design and principles of operation, shall be used to the extent practical to prevent loss of the protection function.

[emphasis added]

The components perform functions that support or impact design bases functions specified in GDCs 20, 21, and 22. Thus, these functions are design functions and are credited in the safety analysis.

Step 3:

The effect on the detection, processing and generation of signals is clear and understood, having a direct impact on the safety analysis assumptions. There is no reason to contemplate the generation of a new FMEA since the impact of the software CCF on the design bases functions is readily apparent (i.e., clear and understood).

Step 4:

Performance/satisfaction of the design bases functions related to the GDC 21 and 22 requirements regarding single failure criteria and redundant channels will not continue to be performed/satisfied.

Step 5:

D-60

Numerous safety analyses contain implicit assumptions regarding the performance and/or expectation of the minimum number of system/components and/or trains/channels that are expected to perform their function, which satisfy the applicable redundancy requirements and/or single failure criteria.

Step 6:

In all cases for each safety analysis, the inability to satisfy the performance and/or expectation of the minimum number of systems/components and/or trains/channels violates an assumption upon which the safety analysis results are based.

In these instances, a simple review of the safety analyses and their structure will quickly identify that the results will exceed the associated acceptance criteria.

Conclusion With the software CCF likelihood determined to be not sufficiently low (i.e., a malfunction is as likely to happen as those described in the UFSAR), the assumptions regarding satisfaction of single failure criteria are invalidated.

Therefore, the proposed activity DOES create the possibility of a malfunction of an SSC important to safety with a different result.

HUMAN-SYSTEM INTERFACE ASSESSMENT If no personnel-based initiators involving degraded operator performance (e.g., operator error) are identified as malfunction initiators, then the creation of a possibility for a malfunction of an SCC important to safety with a different result cannot occur due to the Human-System Interface portion of the digital modification. Otherwise, the creation of a possibility for a malfunction of an SSC important to safety with a different result is assessed utilizing the guidance described in NEI 96-07, Section 4.3.6.

4.3.7 Does the Activity Result in a Design Basis Limit for a Fission Product Barrier Being Exceeded or Altered?

There is no unique guidance applicable to digital modifications for responding to this Evaluation question because the identification of possible design basis limits for fission product barriers and the process for determination of "exceeded" or "altered" are not unique for a digital modification. The guidance in NEI 96-07, Section 4.3.7 applies.

D-61

4.3.8 Does the Activity Result in a Departure from a Method of Evaluation Described in the UFSAR Used in Establishing the Design Bases or in the Safety Analyses?

There is no unique guidance applicable to digital modifications for responding to this Evaluation criterion because activities involving methods of evaluation do not involve SSCs. The guidance in NEI 96-07, Section 4.3.8 applies.

5.0 EXAMPLES

[LATER]

D-62

Page 49: [1] Commented [A82] Author Correction (a): #82 NEI 96-07, Revision 1, Section 4.3.2, Example 4, provides guidance on new operator actions as follows with regards to interdependence:

The change involves a new or modified operator action that supports a design function credited in safety analyses [would not require prior NRC approval] provided:

  • The action (including required completion time) is reflected in plant procedures and operator training programs
  • The licensee has demonstrated that the action can be completed in the time required considering the aggregate affects, such as workload or environmental conditions, expected to exist when the action is required.

o To illustrate this point - Common cause failures in redundant systems compromise safety if the failures are concurrent failures, that is, failures which occur over a time interval during which it is not plausible that the failures would be corrected.

  • The evaluation of the change considers the ability to recover from credible errors in performance of manual actions and the expected time required to make such a recovery.
  • The evaluation considers the effect of the change on plant systems.

It is not clear that the examples is this section appropriately consider the above elements.

Note from meeting:

NEI believes this information belongs to Criteria 2.

NEI will look into using 4.3.2 concepts in this step - including the definition of a FMEA.

Page 49: [2] Commented [A83] Author

  1. 83 Interdependence paragraph Correction (a):

Please clarify the concept to adequate options. NEI 01-01 specifically references adequate backups with regard the determination of potential new outcomes and effects on design functions being preserved or not. Its not clear how a system restart or reboot constitutes an adequate backup to the potentially failed SSC.

As oppose to using the term adequate options should be replaced with specific guidance that described acceptable options, which are already provided in NEI 96-07, Revision 1, Section 4.3.2, Example 4.

Specific compensatory measure, such as opening doors, turning on fans, or installing temporary fans, is not acceptable because those actions are not interdependent with the modifications, in that the actions were added with the purpose of avoiding a license amendment.

Note from meeting:

Adequate options is ambiguous. The 3rd paragraph could be incorrectly read to include compensatory measure.

Please see comment below on page D-55 for Example 4-22. This example does not appear to reference an adequate back up option.