ML15337A293
| ML15337A293 | |
| Person / Time | |
|---|---|
| Site: | Indian Point  |
| Issue date: | 09/09/2015 |
| From: | Lahey R Rensselaer Polytechnic Institute |
| To: | Atomic Safety and Licensing Board Panel |
| SECY RAS | |
| References | |
| RAS 28272, ASLBP 07-858-03-LR-BD01, 50-247-LR, 50-286-LR | |
| Download: ML15337A293 (33) | |
Text
United States Nuclear Regulatory Commission Official Hearing Exhibit In the Matter of: Entergy Nuclear Operations, Inc.
(Indian Point Nuclear Generating Units 2 and 3) NYS000567 ASLBP #: 07-858-03-LR-BD01 Submitted: September 9, 2015 Docket #: 05000247 l 05000286 Exhibit #: NYS000567-PUB-00-BD01 Identified: 11/5/2015 Admitted: 11/5/2015 Withdrawn:
Rejected: Stricken:
Other:
1 UNITED STATES 2 NUCLEAR REGULATORY COMMISSION 3 BEFORE THE ATOMIC SAFETY AND LICENSING BOARD 4 -----------------------------------x 5 In re: Docket Nos. 50-247-LR; 50-286-LR 6 License Renewal Application Submitted by ASLBP No. 07-858-03-LR-BD01 7 Entergy Nuclear Indian Point 2, LLC, DPR-26, DPR-64 8 Entergy Nuclear Indian Point 3, LLC, and 9 Entergy Nuclear Operations, Inc. September 9, 2015 10 -----------------------------------x 11 PRE-FILED SUPPLEMENTAL REPLY WRITTEN TESTIMONY OF 12 Dr. RICHARD T. LAHEY, JR.
13 REGARDING CONTENTION NYS-25 14 On behalf of the State of New York (NYS or the State),
15 the Office of the Attorney General hereby submits the following 16 testimony by RICHARD T. LAHEY, JR., PhD. regarding Contention 17 NYS-25.
18 Q. Please state your full name.
19 A. Richard T. Lahey, Jr.
20 Q. By whom are you employed and what is your position?
21 A. I am retired and am currently the Edward E. Hood 22 Professor Emeritus of Engineering at Rensselaer Polytechnic 23 Institute (RPI), which is located in Troy, New York.
Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 1
1 Q. Have you previously summarized your educational and 2 professional qualifications?
3 A. Yes, my education and professional qualifications and 4 experience are described in my Curricula Vitae and previously 5 filed testimony in this proceeding.
6 Q. I show you what has been marked as Exhibit ENT000616 7 and ENT000679. Do you recognize those documents?
8 A. Yes. They are copies of the pre-filed testimony of 9 the witnesses for Entergy on Contentions NYS-25 and NYS-26B/RK-10 TC-1B that were submitted in August 2015.
11 Q. I show you what has been marked as Exhibit NRC000197 12 and NRC000168. Do you recognize those documents?
13 A. Yes. They are copies of the pre-filed testimony of 14 NRC Staff witness that were submitted in August 2015. NRC000168 15 concerns Contention NYS-26B/RK-TC-1B, and NRC000197 concerns 16 Contention NYS-25. (I note that portions of those two USNRC 17 submissions also discuss Contention NYS-38/RK-TC-5, which I will 18 discuss separately.)
19 Q. Have you had an opportunity to review ENT000616, 20 ENT000679, NRC000168, and NRC000197?
21 A. Yes.
22 Q. Has Entergys and the USNRC Staffs August pre-filed 23 testimony caused you to change the testimony and opinions that Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 2
1 you have previously submitted in this proceeding in connection 2 with Contention NYS-25 and Contention NYS-26B/RK-TC-1B?
3 A. In general, no. Entergy and the USNRC Staff have 4 failed to resolve the age-related safety concerns that I have 5 raised throughout this relicensing proceeding. They continue to 6 approach various aging mechanisms in silos, without addressing 7 the potential synergistic interactions between multiple 8 degradation mechanisms, and my related safety concerns.
9 Q. According to the USNRC Staff , the 10 Expanded Materials Degradation Assessment (EMDA) (NYS000484A-11 B) and Light Water Reactor Sustainability (LWRS) Program 12 (NYS000485) do not apply to IP2 and IP3 because they are 13 associated with subsequent license renewal from 60 to 80 years.
14 (NRC000168, at A176, A178; ). Do you agree?
15 A. Absolutely not. The EMDA and LWRS programs study 16 aging degradation mechanisms that affect all licensed nuclear 17 reactors. The effects studied in the EMDA and LWRS programs are 18 also quite relevant to the safe extended operation of nuclear 19 reactors out to 60 years. The size and cost of these research 20 programs reveals the importance placed on studying the various 21 PWR aging concerns that I have previously raised.
Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 3
1 Q. Do you agree with the USNRC Staff that 2 irradiation embrittlement can increase a components fatigue 3 life?
4 A. There are some data which show that irradiation 5 embrittlement can increase fatigue life in certain situations, 6 especially where high cycle, low amplitude, fatigue occurs.
7 However, there are other data which show that for low cycle, 8 high amplitude fatigue the effects of embrittlement decreases 9 fatigue life by reducing the number of cycles to failure (Nf).
10 As the USNRC Staff concedes, the data regarding the effects of 11 irradiation embrittlement on fatigue life are currently 12 incomplete and inconclusive. NRC000168, at A154; NRC000197, at 13 A196, A200. However, in the face of this uncertainty, Entergy 14 and the USNRC Staff simply assume that the effect can be 15 ignored, and that the Indian Point reactors can continue to be 16 operated until any synergistic degradation effects are directly 17 observed in the operating plants. NRC000197, at A204 (If 18 synergistic effects of aging mechanisms were to occur, the 19 resulting degradation will likely be found in at least one plant 20 in the fleet.) This is exactly what I mean when I say that 21 Entergy and the USNRC Staff have taken a wait-and-see 22 approach. Unfortunately, such an approach could lead to a 23 component failure and potentially serious consequences that Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 4
1 cannot be easily addressed after such a failure has occurred. A 2 much more prudent approach would be to apply an uncertainty 3 factor, or penalty, to the CUFen calculations, and to repair or 4 replace components with very high projected CUFen values. This 5 is a well-accepted engineering practice; indeed, the ASME Code 6 applies a similar uncertainty penalty to CUFen calculations 7 (i.e., a factor of 2 on stress or 20 cycles) to account for 8 uncertainties in test data, which were obtained from tests on 9 small scale, polished metal samples in air, rather than on 10 actual industrial structures, components and fittings in a 11 reactor environment (e.g., in a PWR).
12 13 14 15 A.
16 17 The operative document for management of aging 18 degradation effects on RVI components at Indian Point is the 19 Revised and Amended RVI Plan, which relies on MRP-227-A and 20 which was the subject of the Second Supplemental Safety 21 Evaluation Report for Indian Point (NUREG-1930, Supplement 2) 22 (Exh. NYS000507). I remain concerned with the adequacy of the 23 Revised and Amended RVI Plan, because it relies entirely on Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 5
1 inspections to determine the condition of RVI components. The 2 absence of detectable surface cracks does not necessarily mean 3 that the embrittled and fatigue-weakened structures, components 4 and fittings are not vulnerable to early (i.e., CUFen < 1.0) 5 failures. As MRP-227-A concedes, inspections cannot determine 6 the existence and extent of embrittlement. Accordingly, the 7 Revised and Amended RVI Plan does not account for the 8 possibility that embrittled and fatigue-weakened RVI components 9 could be subject to a shock load which would cause them to fail 10 suddenly.
11 Q. Do you agree with the USNRC Staff that 12 MRP-227-A inspections, coupled with environmentally assisted 13 fatigue calculations resulting in a CUFen of less than 1.0, are 14 adequate to manage the effects of aging on RVI components?
15 A. No. This is a perfect example of the type of silo 16 thinking that I am concerned about. According to the USNRC 17 Staff , embrittlement and the associated aging 18 effects can be managed through the inspection-based Revised and 19 Amended RVI Plan, while fatigue is managed through a separate 20 Fatigue Management Plan (FMP), and no further consideration of 21 the interaction between multiple aging mechanisms is necessary.
22 NRC000168, at A185. For example, the USNRC Staff argues that 23 the portions of my June 2015 testimony (NYS000530) relating to Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 6
1 embrittlement are not relevant to the management of metal 2 fatigue, even though I have repeatedly argued that embrittlement 3 effects should be considered when assessing component fatigue 4 life. NRC000168, at A168, A170. As long as CUFen is calculated 5 to be less than 1.0, the USNRC Staff apparently 6 believe that no critical structures, components, and fittings 7 will exhibit fatigue-induced surface cracking and therefore the 8 effects of embrittlement need not be considered in the fatigue 9 calculation. NRC000168, at A153. Their approach, however, 10 fails to recognize certain basic realities. In particular, the 11 CUF for a structure, component or fitting is defined as the 12 number of fatigue cycles it is expected to experience (N) 13 divided by the number of cycles to failure (Nf). The Nf for a 14 new, ductile material can be significantly greater than the Nf 15 for a highly embrittled material; indeed, this is known to be 16 true for large amplitude, low cycle fatigue. See e.g. Kanaski, 17 et al., Fatigue and Stress Corrosion Cracking Behaviors of 18 Irradiated Stainless Steels in PWR Primary Water, ICONE-5, at 19 2372 (May 1997) (Exh. NRC000177); Arai, et al., Irradiation 20 Embrittlement of PWR Internals, Proceedings ASME/JSME 2d 21 International Nuclear Engineering Conference, Vol. 2, at 103 22 (1993) (Exh. NYS000564); Korth, G.E. & Harper, M.D., Effects of 23 Neutron Radiation on the Fatigue and Creep/Fatigue Behavior of Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 7
1 Type 308 Stainless Steel Weld Materials at Elevated 2 Temperatures, Proceedings of the 7th International Symposium on 3 the Effects of Radiation on Structural Materials, Gatlinburg, TN 4 (June 1974) (Exh. RIV000152). Additionally, embrittled and 5 fatigue-weakened structures may not be able to tolerate 6 significant seismic and shock loads as well as fully ductile 7 structures can. Thus, when the effects of embrittlement and 8 fatigue are considered together, there is a real risk that 9 components will fail before their calculated CUFen value reaches 10 unity.
11 Q. I show you a document marked as Exhibit [NYS000566]
12 that is entitled, Figure 1: Comparison of Limit Line and Best 13 Estimate Predictions with Embrittlement and Without 14 Embrittlement. Are you familiar with this Figure?
15 A. Yes.
16 Q. How are you familiar with this Figure?
17 I developed Figure-1 in connection with my review of 18 Entergys and the USNRCs Revised Statements of Position and 19 Revised Testimony in this proceeding.
Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 8
BEe (1)
(3) (2) 1.0
+
BEne Limit Line (LL)
CUFen Best Estimate (BE) 1017 FLUENCE (n/cm2)
EOL (PEO) TIME Figure 1: Comparison of Limit Line (LL) and Best Estimate (BE) predictions, with embrittlement (BEe) and without embrittlement (BEne).
Note:
(1) BEne predicts possible failure at end of life (EOL).
(2) BEe predicts failure (CUFen = 1.0) before end of life.
(3) BEe predicts possible failure well before end of life.
1 Q. Why did you prepare this Figure?
2 A. I prepared this Figure to visually illustrate some of 3 the concerns that I have set forth in my testimony in this 4 proceeding. Specifically, this Figure shows three important 5 things: (1) a typical WESTEMSTM Limit Line prediction, (2) a 6 Best Estimate prediction of CUFen, the cumulative usage factor 7 under reactor operating conditions, and the associated Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 9
1 uncertainty interval (), and, (3) how the neutron fluence-2 induced embrittlement of a RVI structure, component, or fitting 3 can dramatically increase the possibility of a fatigue failure, 4 as represented by the cumulative usage factor under PWR 5 operating conditions, CUFen , being equal to unity. This figure 6 applies to RVI structures, components, and fittings made of 7 stainless steel or other materials. In summary, in Figure-1, I 8 have illustrated two different approaches for calculating a 9 time-dependent CUFen, namely the Limit Line and the Best 10 Estimate calculation. For the Best Estimate line, I also 11 illustrate typical uncertainty intervals, delta (), and the 12 effects of fatigue on embrittled (e) and on non-embrittled (ne) 13 RVIs.
14 Q. Please describe what is represented by the horizontal 15 x-axis (i.e., the abscissa) in Figure 1.
16 A. Since fluence = t (where is the fast neutron flux 17 (n/cm2-s) and t is the time of operation of the reactor) the x-18 axis tracks both the fluence () and the time (t). As time 19 passes, each RVI reaches the reactors End of Life (EOL) for the 20 Period of Extended Operation (PEO). Also, as we move from left-21 to-right along the x-axis (i.e., as the time of reactor 22 operation increases), we see an increase in the number of Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 10
1 fatigue cycles (N) that the various RVIs (one of which has been 2 chosen to be displayed in Figure-1) have been subjected to.
3 Q. Please describe what is represented by the vertical y-4 axis (i.e., the ordinate) in Figure 1.
5 A. The y-axis displays CUFen, the cumulative usage factor, 6 considering PWR environmental conditions. At a nuclear power 7 plant, the maximum number of fatigue cycles that can be 8 experienced by any structure, component or fitting must always 9 result in a CUFen of less than 1.0. That is, the number of 10 actual fatigue cycles (N) should always be less than the number 11 of allowable cycles (Nf) to avoid failure of the RVI.
12 Q. In Figure-1 what does the Limit Line represent?
13 A. The Limit Line represents 14 as a supposedly-conservative 15 calculation of CUFen for a hypothetical RVI component or 16 structure.
17 18 The Limit Line shown in Figure-1 includes 19 Entergys implicit assumption that a stainless steel RVI 20 structure, component, or fitting remains perfectly ductile, and 21 thus shows that the CUFen for a RVI will increase with an 22 increasing number of fatigue cycles, but that it is not affected 23 by the fluence. The Limit Line, therefore, uniformly Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 11
1 increases in CUFen even after the fluence exceeds 1017 n/cm2, 2 where significant embrittlement begins to develop (i.e., it is 3 assumed that the number of cycles to failure are not a function 4 of fluence). Figure-1 shows that, as the RVI component 5 approaches the end of life (EOL) for the period of extended 6 operation (PEO), Limit Line will approach 7 (but still be below) CUFen = 1.0. This is depicted as location 8 (1).
9 Q. In Figure-1 what does the Best Estimate line 10 represent?
11 A. The Best Estimate plot depicts what would be, for 12 example, a WESTEMS prediction of CUFen in which conservative 13 assumptions are made and my concerns 14 15 have been 16 addressed. In addition, the Best Estimate plot in Figure-1 17 also includes a prediction of a propagation of errors type 18 analysis (i.e., an uncertainty analysis), which can, and should, 19 be done.
20 Q. What does a propagation of errors analysis include?
21 A. This type of analysis considers important parameters 22 that have some uncertainty associated with them; for example, 23 the coarseness of the computational mesh, uncertainties in Fen ,
Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 12
1 the best estimate local heat transfer coefficients, etc.;
2 uncertainties which, in turn, impact the RVIs actual fatigue 3 life (i.e., CUFen). The net uncertainty intervals () also 4 account for uncertainties in the number of transients 5 (including, for example, seismic events), and synergistic 6 effects of radiation (i.e., the fluence) and stress corrosion 7 effects on metal fatigue. Accounting for the uncertainty in 8 these parameters allows one to estimate the overall uncertainty 9 () of the Best Estimate CUFen prediction, so that we can see 10 if the Limit Line predictions are indeed 11 sufficiently conservative. That is, comparing the Best 12 Estimate results, plus , with the Limit Line predictions 13 reveals whether the Limit Line approach 14 is adequate or 15 not.
16 Q. Can you indicate how one might perform a propagation 17 of error analysis?
18 A. Yes, uncertainty analyses can be performed in a number 19 of ways. One of the most commonly used methods by engineers is 20 the method proposed by Kline & McClintock [ASME J. Mechanical 21 Engineering, Vol.75, No.1, 3-8, Jan. 1953] (Exh. NYS000514).
22 For the case in which the Best Estimate value of CUFen involves 23 N parameters, each having some uncertainty associated with Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 13
1 them (e.g., the Dittus-Boelter convective heat transfer, 2 has an inherent uncertainty of about 25%),
3 the net uncertainty () is given by:
4 =
5 6 This approach is widely used by engineers since it is relatively 7 easy to evaluate and it gives an acceptable approximation of the 8 error bars () in a Best Estimate evaluation.
9 Q. In Figure-1, the Best Estimate line is below the 10 Limit Line. Does this mean that everything is OK?
11 A. Not necessarily. As shown in Figure-1, a Best 12 Estimate prediction plus the uncertainty () may predict 13 possible RVI structure, components or fitting failures (i.e.,
14 CUFen = 1.0), sooner than the Limit Line would. Moreover, the 15 effect of embrittlement can make the situation much worse. The 16 problem is that no one knows how much conservatism, if any, is 17 in the Limit Line. As a consequence, we can have no 18 confidence that Limit Line results near CUFen = 1.0, are 19 sufficiently conservative to bound all the possible 20 uncertainties, as illustrated in Figure-1 by .
21 Q. In Figure-1, the Best Estimate line is shown for the 22 case in which there is no effect of embrittlement on fatigue, 23 BEne, and the case in Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 14
1 which embrittlement degrades fatigue life, BEe. Can you explain 2 what these lines mean?
3 A. Yes, the plot depicts two Best Estimate predictions 4 for a hypothetical RVI component: one under conditions of no 5 embrittlement (BEne), for which the Best Estimate CUFen line 6 continues to increase as the fatigue-inducing cycles (N) 7 accumulate with time, and the other Best Estimate prediction 8 (BEe) which includes the effect of embrittlement on the (now 9 time-dependent) maximum allowable cycles (Nf). For the latter 10 case, increases in fluence result in more embrittlement and thus 11 a more rapid increase in CUFen than in the former case. This is 12 because after a fluence of about 1017 n/cm2, significant 13 irradiation-induced damage and embrittlement begin to occur.
14 Embrittlement results in the loss of facture toughness and the 15 loss of ductility. Even though the data taken to date have been 16 inconclusive, there is ample evidence that this embrittlement 17 may reduce the number of fatigue cycles to failure (Nf) and thus 18 increase the corresponding CUFen, which is defined as CUFen = .
19 Q. What does location (1)in Figure-1 show?
20 A. Location (1) shows that, even if embrittlement is not 21 considered, RVI fatigue failure may occur by the end of life for 22 the period of extended operation. This is because the sum of a Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 15
1 Best Estimate prediction plus the uncertainty interval ()
2 associated with this prediction would exceed CUFen = 1.0. In this 3 case the Limit Line result is clearly non-conservative.
4 Q. May I now direct your attention to location (2) in 5 Figure-1. What does this location show?
6 A. Location (2) shows that embrittlement can result in 7 the physical failure of the RVI well before the end of life for 8 the period of extended operation.
9 Q. May I direct your attention to location (3) in Figure-10 1. What does this location show?
11 A. At location (3), the Best Estimate plus uncertainty 12 () for the case considering embrittlement (BEe) predicts 13 possible failure of the hypothetical RVI component well before 14 the end of life for the period of extended operation, even 15 though the Limit Line prediction indicates substantial margin 16 to failure at this point. That is, at location (3) the CUFen for 17 the Best Estimate line, accounting for embrittlement and for 18 all uncertainties (i.e., BEe + ), exceeds unity. Moreover, as I 19 have stated in my previous testimony, highly embrittled RVIs may 20 fail even sooner than by fatigue alone if significant seismic or 21 shock loads occur. That is, fatigue-weakened and embrittled 22 structures cannot tolerate large impulsive seismic and shock 23 loads like fully ductile structures can. Moreover, other Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 16
1 synergistic aging effects, such as the thermal embrittlement(TE) 2 of some CASS RVIs (the upper part of the core support columns) 3 would may move the BEe curve to the left of its position shown in 4 Figure-1 (i.e., the RVI would be predicted to reach failure, 5 CUFen = 1.0, even sooner than the cases (2) and (3) in Figure-1).
6 Q. Can any conclusions about the Limit Line and Best 7 Estimate lines be drawn from Figure 1?
8 A. Yes. Because of Entergys failure to explicitly 9 account for uncertainties, and for the effect of embrittlement 10 on fatigued RVIs, we can have no confidence in the Limit Line 11 type of CUFen predictions 12 13 14 15 16 17 18 19 20 Q. In response to Dr. Hopenfelds testimony, the USNRC 21 Staff stated that CUF or CUFen analyses are not required for 22 safety assessments of DBA events . . . because CUF and CUFen, 23 which are indicators of possible fatigue crack initiation, are Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 17
1 not a significant contributor to safety during DBA events.
2 NRC000168, at A145. How do you respond?
3 A. Again, this is a perfect example of silo thinking.
4 The USNRC Staff refuses to assess the risk that a highly 5 fatigued system, component or fitting which has been fatigue-6 weakened and embrittled as a result of neutron fluence, for 7 example, could fail, relocate, and degrade core cooling, when it 8 is subjected to a DBA LOCA or some other significant shock load.
9 This approach ignores the reality of a reactor environment, 10 where multiple aging mechanisms act simultaneously on RVIs and 11 other components, and that this age-related degradation needs to 12 be taken into account in plant safety analyses.
13 Q.
14 15 16 17 18 19 20 A. As I have stated above, ignoring the 21 effects of accident loads on aging RVI components is not at all 22 appropriate. While it is probably true that fatigued and 23 embrittled RVI components may operate normally during steady-Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 18
1 state operations, they can fail suddenly and catastrophically 2 when subjected to a significant shock load. Without considering 3 accident-induced loads, Entergy cannot provide adequate 4 assurance that the RVIs will continue to perform their functions 5 during the period of extended operation (PEO).
6 Q. Have there been RVI component failures at nuclear 7 reactors in the past?
8 A. Yes. Baffle former bolts and clevis insert bolts have 9 failed at several PWRs in the past, clearly demonstrating that 10 inspections alone will not detect aging effects prior to 11 component failure.
12 Q. With respect to the clevis insert bolt failures, do 13 you agree with the USNRC Staffs testimony that [t]he failed 14 bolts were detected visually (A290)?
15 A. Partially. According the SSER2 at 3-25 (Exh.
16 NYS000507), only 7 out of 29 damaged bolts, or about 24%, were 17 detected visually. The vast majority were not detected 18 visually. Nonetheless, Entergy has proposed to conduct only 19 visual inspections of clevis insert bolts. Exh. NYS000496, at 20 51.
21 Q.
22 23 Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 19
1 properly during steady-state operations, when a large percentage 2 of the incore bolting has failed, a significant shock load could 3 cause many of the remaining bolts to suddenly fail, resulting in 4 the relocation of core components and the possible loss of a 5 coolable core geometry.
6 Q. When discussing aging management of the baffle-former 7 bolts, the USNRC Staff describes the number of baffle-former 8 bolts in [t]hree-loop Westinghouse design PWRs like IP2 and 9 IP3[.] NRC000197, at A243. Is it accurate to say that IP2 and 10 IP3 are three-loop PWRs?
11 A. No. IP2 and IP3 are actually four-loop Westinghouse 12 designed PWRs. It is not clear why the USNRC Staff believes the 13 IP2 and IP3 reactors use a three-loop design. Perhaps this 14 reference is a cut and paste remnant from another proceeding 15 for a different facility that USNRC Staff has reused here.
16 Q.
17 18 19 20 21 22 Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 21
1 A. Virtually all metals (embrittled or not),
2 experience a decrease in ductility as the temperature decreases.
3 However, irradiated carbon steel undergoes a very sharp decrease 4 in ductility at a fluence-dependent temperature commonly called 5 the nil ductility temperatures (NDT), while stainless steel does 6 not. Nevertheless, as I have said before, sufficiently strong 7 shock loads can lead to failure of highly fatigue-weakened and 8 embrittled stainless steel RVIs.
9 Q.
10 11 12 A.
13 14 15 16 17 18 19 20 21 22 23 Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 22
1 2
3 4
5 The USNRC 6 Staff concedes that [g]iven the variability in assumptions made 7 by different analysts, it is difficult to explicitly quantify 8 the exact overall safety margin present in fatigue 9 calculations. NRC000168, at A210.
10 11 However, given the critical 12 siting of the Indian Point plants in the New York metropolitan 13 area, it is especially appropriate to identify and verify the 14 assumptions and calculations. Indeed, as distilled by the 15 concept of trust but verify, verification precedes trust.
16 Q. Do you agree with the USNRC Staff that a 17 propagation of errors analysis (i.e., an uncertainty 18 analysis), is not needed for EAF calculations?
19 A. No. the USNRC Staff claim that an 20 uncertainty analysis is not necessary because the EAF 21 calculation is deterministic and contains adequate 22 conservatisms. NRC000168, at A171; 23 Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 23
1 2
3 4
5 6
7 8
9 10 11 Q. So far you have been talking about reactor vessel 12 internals (RVIs). Do you have some similar concerns about the 13 fatigue evaluations for primary pressure boundary components and 14 fittings?
15 A. Yes, except for the fact that radiation-induced 16 embrittlement (IE) is not normally an issue, virtually all of my 17 previously discussed concerns are the same.
18 19 20 21 22 23 Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 24
1 2
3 Q. the USNRC Staff state that Entergy will 4 compare the actual transients experienced at Indian Point with 5 the transients anticipated in the EAF calculations. Does this 6 approach address your concerns?
7 A. No. The number of fatigue cycles is obviously 8 important; however, my principal concern has never been that 9 components will experience a greater number of transient cycles 10 than anticipated in the EAF calculation. In contrast, my real 11 concern for the integrity of primary pressure boundary 12 components is that a significant shock load - caused by a 13 seismic, LOCA, or other event - could cause the fatigue-weakened 14 component to suddenly fail well before their associated CUFen 15 value reaches 1.0.
16 17 18 19 20 21 22 23 Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 25
1 A. In my opinion, an observable surface crack of 2 3mm or greater, which would be expected when CUFen reaches 1.0, 3 is quite significant. However, microscopic cracks exist and 4 propagate within the metal structures, components and fittings 5 even when CUFen < 1.0. These microscopic cracks cannot be 6 detected by non-destructive testing (NDT), but they can weaken 7 the components and make them more susceptible to failures during 8 a significant seismic or pressure/thermal shock load event, 9 especially when coupled with embrittlement due various thermal 10 embrittlement or corrosion mechanisms.
11 12 13 14 15 16 A. what I 17 mean by not reducing design and safety margins as the plant ages 18 is that component fatigue life calculations should retain the 19 original design conservatisms and that the components should be 20 repaired or replaced if they exceed acceptable design margins 21 during the period of extended operations.
22 23 Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 26
1 In my opinion, removing 2 modeling conservatisms as IP2 and IP3 exceed 40 years of 3 operation, and the components become more and more degraded, is 4 both irresponsible and dangerous.
5 Q.
6 7
8 9
10 11 12 13 14 15 16 17 18 19 20 21 22 23 Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 27
1 2
3 4
5 6
7 8
9 10 11 12 13 14 15 16 17 Q. According to the USNRC Staff, aging mechanisms acting 18 on the core support columns are properly assessed from the 19 normal, steady-state operating conditions. NRC000197, at A306.
20 How do you respond?
21 A. This approach does not account for seismic or LOCA-22 type shock loads, which could cause the columns to fail.
Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 28
1 Q. I show you what has been marked as Exhibit NYS000563.
2 Do you recognize it?
3 A. Yes, it is an excerpt from an USNRC training manual, 4 entitled Reactor Concepts Manual - PWR Systems. It was 5 prepared by the USNRC Technical Training Center.
6 Q. What information did you draw from this USNRC 7 document?
8 A. This document identifies the relative location of 9 various components in a 4-loop, Westinghouse, pressurized water 10 reactor (PWR). On page 4-25, it shows that one of the 11 accumulators uses the same nozzle as the residual heat removal 12 (RHR) system, the low pressure coolant injection (LPCI), and the 13 intermediate pressure coolant injection (IPCI) safety systems.
14 In contrast, the high pressure coolant injection (HPCI) system 15 has a separate connection on the same cold leg as the 16 RHR/Accumulator nozzle.
17 Q. Why is that relevant?
18 A.
19 20 21 22 23 the components could fail Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 29
1 if subjected to a significant seismic event or shock load. As I 2 have pointed out in some of my previous ASLB testimony, the 3 failure of this important primary pressure boundary structure 4 could lead to a LOCA event in which the Accumulator, LPCI and 5 IPCI systems are not be able to inject water into the reactors 6 cold leg, and thus into the core, which, in turn, could lead to 7 core melting.
8 Q. Do you have particular concerns with respect to any 9 other components?
10 A. Yes.
11 12 13 14 15 16 17 18 19 20 21 22 23 Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 30
1 2
3 4
5 6
7 8
9 10 11 12 13 14 the USNRC 15 Staff has indicated that if transients accumulate at a rate 16 greater than the rate assumed in the fatigue calculation, 17 Entergy would be permitted to conduct yet another more refined 18 analysis. NRC000168, at A106. In short, there is apparently 19 no limit to the number of times CUFen can be recalculated to 20 obtain a CUFen result less than unity, and there is no standard 21 which defines the amount of conservatism that must be retained 22 in these calculations.
23 Q. Does this complete your testimony?
Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 31
1 A. Yes, it does. I do, however, reserve the right to 2 supplement my testimony if new information is disclosed or 3 introduced.
4 Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 32
1 UNITED STATES 2 NUCLEAR REGULATORY COMMISSION 3 BEFORE THE ATOMIC SAFETY AND LICENSING BOARD 4 -----------------------------------x 5 In re: Docket Nos. 50-247-LR; 50-286-LR 6 License Renewal Application Submitted by ASLBP No. 07-858-03-LR-BD01 7 Entergy Nuclear Indian Point 2, LLC, DPR-26, DPR-64 8 Entergy Nuclear Indian Point 3, LLC, and 9 Entergy Nuclear Operations, Inc. September 9, 2015 10 -----------------------------------x 11 DECLARATION OF RICHARD T. LAHEY, JR.
12 I, Richard T. Lahey, Jr., do hereby declare under penalty 13 of perjury that my statements in the foregoing testimony and my 14 statement of professional qualifications are true and correct to 15 the best of my knowledge and belief.
16 Executed in Accord with 10 C.F.R. § 2.304(d) 17 18 _________________________
19 Dr. Richard T. Lahey, Jr.
20 The Edward E. Hood Professor Emeritus of Engineering 21 Rensselaer Polytechnic Institute, Troy, NY 12180 22 (518) 495-3884, laheyr@rpi.edu Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 33
United States Nuclear Regulatory Commission Official Hearing Exhibit In the Matter of: Entergy Nuclear Operations, Inc.
(Indian Point Nuclear Generating Units 2 and 3) NYS000567 ASLBP #: 07-858-03-LR-BD01 Submitted: September 9, 2015 Docket #: 05000247 l 05000286 Exhibit #: NYS000567-PUB-00-BD01 Identified: 11/5/2015 Admitted: 11/5/2015 Withdrawn:
Rejected: Stricken:
Other:
1 UNITED STATES 2 NUCLEAR REGULATORY COMMISSION 3 BEFORE THE ATOMIC SAFETY AND LICENSING BOARD 4 -----------------------------------x 5 In re: Docket Nos. 50-247-LR; 50-286-LR 6 License Renewal Application Submitted by ASLBP No. 07-858-03-LR-BD01 7 Entergy Nuclear Indian Point 2, LLC, DPR-26, DPR-64 8 Entergy Nuclear Indian Point 3, LLC, and 9 Entergy Nuclear Operations, Inc. September 9, 2015 10 -----------------------------------x 11 PRE-FILED SUPPLEMENTAL REPLY WRITTEN TESTIMONY OF 12 Dr. RICHARD T. LAHEY, JR.
13 REGARDING CONTENTION NYS-25 14 On behalf of the State of New York (NYS or the State),
15 the Office of the Attorney General hereby submits the following 16 testimony by RICHARD T. LAHEY, JR., PhD. regarding Contention 17 NYS-25.
18 Q. Please state your full name.
19 A. Richard T. Lahey, Jr.
20 Q. By whom are you employed and what is your position?
21 A. I am retired and am currently the Edward E. Hood 22 Professor Emeritus of Engineering at Rensselaer Polytechnic 23 Institute (RPI), which is located in Troy, New York.
Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 1
1 Q. Have you previously summarized your educational and 2 professional qualifications?
3 A. Yes, my education and professional qualifications and 4 experience are described in my Curricula Vitae and previously 5 filed testimony in this proceeding.
6 Q. I show you what has been marked as Exhibit ENT000616 7 and ENT000679. Do you recognize those documents?
8 A. Yes. They are copies of the pre-filed testimony of 9 the witnesses for Entergy on Contentions NYS-25 and NYS-26B/RK-10 TC-1B that were submitted in August 2015.
11 Q. I show you what has been marked as Exhibit NRC000197 12 and NRC000168. Do you recognize those documents?
13 A. Yes. They are copies of the pre-filed testimony of 14 NRC Staff witness that were submitted in August 2015. NRC000168 15 concerns Contention NYS-26B/RK-TC-1B, and NRC000197 concerns 16 Contention NYS-25. (I note that portions of those two USNRC 17 submissions also discuss Contention NYS-38/RK-TC-5, which I will 18 discuss separately.)
19 Q. Have you had an opportunity to review ENT000616, 20 ENT000679, NRC000168, and NRC000197?
21 A. Yes.
22 Q. Has Entergys and the USNRC Staffs August pre-filed 23 testimony caused you to change the testimony and opinions that Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 2
1 you have previously submitted in this proceeding in connection 2 with Contention NYS-25 and Contention NYS-26B/RK-TC-1B?
3 A. In general, no. Entergy and the USNRC Staff have 4 failed to resolve the age-related safety concerns that I have 5 raised throughout this relicensing proceeding. They continue to 6 approach various aging mechanisms in silos, without addressing 7 the potential synergistic interactions between multiple 8 degradation mechanisms, and my related safety concerns.
9 Q. According to the USNRC Staff , the 10 Expanded Materials Degradation Assessment (EMDA) (NYS000484A-11 B) and Light Water Reactor Sustainability (LWRS) Program 12 (NYS000485) do not apply to IP2 and IP3 because they are 13 associated with subsequent license renewal from 60 to 80 years.
14 (NRC000168, at A176, A178; ). Do you agree?
15 A. Absolutely not. The EMDA and LWRS programs study 16 aging degradation mechanisms that affect all licensed nuclear 17 reactors. The effects studied in the EMDA and LWRS programs are 18 also quite relevant to the safe extended operation of nuclear 19 reactors out to 60 years. The size and cost of these research 20 programs reveals the importance placed on studying the various 21 PWR aging concerns that I have previously raised.
Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 3
1 Q. Do you agree with the USNRC Staff that 2 irradiation embrittlement can increase a components fatigue 3 life?
4 A. There are some data which show that irradiation 5 embrittlement can increase fatigue life in certain situations, 6 especially where high cycle, low amplitude, fatigue occurs.
7 However, there are other data which show that for low cycle, 8 high amplitude fatigue the effects of embrittlement decreases 9 fatigue life by reducing the number of cycles to failure (Nf).
10 As the USNRC Staff concedes, the data regarding the effects of 11 irradiation embrittlement on fatigue life are currently 12 incomplete and inconclusive. NRC000168, at A154; NRC000197, at 13 A196, A200. However, in the face of this uncertainty, Entergy 14 and the USNRC Staff simply assume that the effect can be 15 ignored, and that the Indian Point reactors can continue to be 16 operated until any synergistic degradation effects are directly 17 observed in the operating plants. NRC000197, at A204 (If 18 synergistic effects of aging mechanisms were to occur, the 19 resulting degradation will likely be found in at least one plant 20 in the fleet.) This is exactly what I mean when I say that 21 Entergy and the USNRC Staff have taken a wait-and-see 22 approach. Unfortunately, such an approach could lead to a 23 component failure and potentially serious consequences that Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 4
1 cannot be easily addressed after such a failure has occurred. A 2 much more prudent approach would be to apply an uncertainty 3 factor, or penalty, to the CUFen calculations, and to repair or 4 replace components with very high projected CUFen values. This 5 is a well-accepted engineering practice; indeed, the ASME Code 6 applies a similar uncertainty penalty to CUFen calculations 7 (i.e., a factor of 2 on stress or 20 cycles) to account for 8 uncertainties in test data, which were obtained from tests on 9 small scale, polished metal samples in air, rather than on 10 actual industrial structures, components and fittings in a 11 reactor environment (e.g., in a PWR).
12 13 14 15 A.
16 17 The operative document for management of aging 18 degradation effects on RVI components at Indian Point is the 19 Revised and Amended RVI Plan, which relies on MRP-227-A and 20 which was the subject of the Second Supplemental Safety 21 Evaluation Report for Indian Point (NUREG-1930, Supplement 2) 22 (Exh. NYS000507). I remain concerned with the adequacy of the 23 Revised and Amended RVI Plan, because it relies entirely on Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 5
1 inspections to determine the condition of RVI components. The 2 absence of detectable surface cracks does not necessarily mean 3 that the embrittled and fatigue-weakened structures, components 4 and fittings are not vulnerable to early (i.e., CUFen < 1.0) 5 failures. As MRP-227-A concedes, inspections cannot determine 6 the existence and extent of embrittlement. Accordingly, the 7 Revised and Amended RVI Plan does not account for the 8 possibility that embrittled and fatigue-weakened RVI components 9 could be subject to a shock load which would cause them to fail 10 suddenly.
11 Q. Do you agree with the USNRC Staff that 12 MRP-227-A inspections, coupled with environmentally assisted 13 fatigue calculations resulting in a CUFen of less than 1.0, are 14 adequate to manage the effects of aging on RVI components?
15 A. No. This is a perfect example of the type of silo 16 thinking that I am concerned about. According to the USNRC 17 Staff , embrittlement and the associated aging 18 effects can be managed through the inspection-based Revised and 19 Amended RVI Plan, while fatigue is managed through a separate 20 Fatigue Management Plan (FMP), and no further consideration of 21 the interaction between multiple aging mechanisms is necessary.
22 NRC000168, at A185. For example, the USNRC Staff argues that 23 the portions of my June 2015 testimony (NYS000530) relating to Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 6
1 embrittlement are not relevant to the management of metal 2 fatigue, even though I have repeatedly argued that embrittlement 3 effects should be considered when assessing component fatigue 4 life. NRC000168, at A168, A170. As long as CUFen is calculated 5 to be less than 1.0, the USNRC Staff apparently 6 believe that no critical structures, components, and fittings 7 will exhibit fatigue-induced surface cracking and therefore the 8 effects of embrittlement need not be considered in the fatigue 9 calculation. NRC000168, at A153. Their approach, however, 10 fails to recognize certain basic realities. In particular, the 11 CUF for a structure, component or fitting is defined as the 12 number of fatigue cycles it is expected to experience (N) 13 divided by the number of cycles to failure (Nf). The Nf for a 14 new, ductile material can be significantly greater than the Nf 15 for a highly embrittled material; indeed, this is known to be 16 true for large amplitude, low cycle fatigue. See e.g. Kanaski, 17 et al., Fatigue and Stress Corrosion Cracking Behaviors of 18 Irradiated Stainless Steels in PWR Primary Water, ICONE-5, at 19 2372 (May 1997) (Exh. NRC000177); Arai, et al., Irradiation 20 Embrittlement of PWR Internals, Proceedings ASME/JSME 2d 21 International Nuclear Engineering Conference, Vol. 2, at 103 22 (1993) (Exh. NYS000564); Korth, G.E. & Harper, M.D., Effects of 23 Neutron Radiation on the Fatigue and Creep/Fatigue Behavior of Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 7
1 Type 308 Stainless Steel Weld Materials at Elevated 2 Temperatures, Proceedings of the 7th International Symposium on 3 the Effects of Radiation on Structural Materials, Gatlinburg, TN 4 (June 1974) (Exh. RIV000152). Additionally, embrittled and 5 fatigue-weakened structures may not be able to tolerate 6 significant seismic and shock loads as well as fully ductile 7 structures can. Thus, when the effects of embrittlement and 8 fatigue are considered together, there is a real risk that 9 components will fail before their calculated CUFen value reaches 10 unity.
11 Q. I show you a document marked as Exhibit [NYS000566]
12 that is entitled, Figure 1: Comparison of Limit Line and Best 13 Estimate Predictions with Embrittlement and Without 14 Embrittlement. Are you familiar with this Figure?
15 A. Yes.
16 Q. How are you familiar with this Figure?
17 I developed Figure-1 in connection with my review of 18 Entergys and the USNRCs Revised Statements of Position and 19 Revised Testimony in this proceeding.
Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 8
BEe (1)
(3) (2) 1.0
+
BEne Limit Line (LL)
CUFen Best Estimate (BE) 1017 FLUENCE (n/cm2)
EOL (PEO) TIME Figure 1: Comparison of Limit Line (LL) and Best Estimate (BE) predictions, with embrittlement (BEe) and without embrittlement (BEne).
Note:
(1) BEne predicts possible failure at end of life (EOL).
(2) BEe predicts failure (CUFen = 1.0) before end of life.
(3) BEe predicts possible failure well before end of life.
1 Q. Why did you prepare this Figure?
2 A. I prepared this Figure to visually illustrate some of 3 the concerns that I have set forth in my testimony in this 4 proceeding. Specifically, this Figure shows three important 5 things: (1) a typical WESTEMSTM Limit Line prediction, (2) a 6 Best Estimate prediction of CUFen, the cumulative usage factor 7 under reactor operating conditions, and the associated Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 9
1 uncertainty interval (), and, (3) how the neutron fluence-2 induced embrittlement of a RVI structure, component, or fitting 3 can dramatically increase the possibility of a fatigue failure, 4 as represented by the cumulative usage factor under PWR 5 operating conditions, CUFen , being equal to unity. This figure 6 applies to RVI structures, components, and fittings made of 7 stainless steel or other materials. In summary, in Figure-1, I 8 have illustrated two different approaches for calculating a 9 time-dependent CUFen, namely the Limit Line and the Best 10 Estimate calculation. For the Best Estimate line, I also 11 illustrate typical uncertainty intervals, delta (), and the 12 effects of fatigue on embrittled (e) and on non-embrittled (ne) 13 RVIs.
14 Q. Please describe what is represented by the horizontal 15 x-axis (i.e., the abscissa) in Figure 1.
16 A. Since fluence = t (where is the fast neutron flux 17 (n/cm2-s) and t is the time of operation of the reactor) the x-18 axis tracks both the fluence () and the time (t). As time 19 passes, each RVI reaches the reactors End of Life (EOL) for the 20 Period of Extended Operation (PEO). Also, as we move from left-21 to-right along the x-axis (i.e., as the time of reactor 22 operation increases), we see an increase in the number of Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 10
1 fatigue cycles (N) that the various RVIs (one of which has been 2 chosen to be displayed in Figure-1) have been subjected to.
3 Q. Please describe what is represented by the vertical y-4 axis (i.e., the ordinate) in Figure 1.
5 A. The y-axis displays CUFen, the cumulative usage factor, 6 considering PWR environmental conditions. At a nuclear power 7 plant, the maximum number of fatigue cycles that can be 8 experienced by any structure, component or fitting must always 9 result in a CUFen of less than 1.0. That is, the number of 10 actual fatigue cycles (N) should always be less than the number 11 of allowable cycles (Nf) to avoid failure of the RVI.
12 Q. In Figure-1 what does the Limit Line represent?
13 A. The Limit Line represents 14 as a supposedly-conservative 15 calculation of CUFen for a hypothetical RVI component or 16 structure.
17 18 The Limit Line shown in Figure-1 includes 19 Entergys implicit assumption that a stainless steel RVI 20 structure, component, or fitting remains perfectly ductile, and 21 thus shows that the CUFen for a RVI will increase with an 22 increasing number of fatigue cycles, but that it is not affected 23 by the fluence. The Limit Line, therefore, uniformly Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 11
1 increases in CUFen even after the fluence exceeds 1017 n/cm2, 2 where significant embrittlement begins to develop (i.e., it is 3 assumed that the number of cycles to failure are not a function 4 of fluence). Figure-1 shows that, as the RVI component 5 approaches the end of life (EOL) for the period of extended 6 operation (PEO), Limit Line will approach 7 (but still be below) CUFen = 1.0. This is depicted as location 8 (1).
9 Q. In Figure-1 what does the Best Estimate line 10 represent?
11 A. The Best Estimate plot depicts what would be, for 12 example, a WESTEMS prediction of CUFen in which conservative 13 assumptions are made and my concerns 14 15 have been 16 addressed. In addition, the Best Estimate plot in Figure-1 17 also includes a prediction of a propagation of errors type 18 analysis (i.e., an uncertainty analysis), which can, and should, 19 be done.
20 Q. What does a propagation of errors analysis include?
21 A. This type of analysis considers important parameters 22 that have some uncertainty associated with them; for example, 23 the coarseness of the computational mesh, uncertainties in Fen ,
Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 12
1 the best estimate local heat transfer coefficients, etc.;
2 uncertainties which, in turn, impact the RVIs actual fatigue 3 life (i.e., CUFen). The net uncertainty intervals () also 4 account for uncertainties in the number of transients 5 (including, for example, seismic events), and synergistic 6 effects of radiation (i.e., the fluence) and stress corrosion 7 effects on metal fatigue. Accounting for the uncertainty in 8 these parameters allows one to estimate the overall uncertainty 9 () of the Best Estimate CUFen prediction, so that we can see 10 if the Limit Line predictions are indeed 11 sufficiently conservative. That is, comparing the Best 12 Estimate results, plus , with the Limit Line predictions 13 reveals whether the Limit Line approach 14 is adequate or 15 not.
16 Q. Can you indicate how one might perform a propagation 17 of error analysis?
18 A. Yes, uncertainty analyses can be performed in a number 19 of ways. One of the most commonly used methods by engineers is 20 the method proposed by Kline & McClintock [ASME J. Mechanical 21 Engineering, Vol.75, No.1, 3-8, Jan. 1953] (Exh. NYS000514).
22 For the case in which the Best Estimate value of CUFen involves 23 N parameters, each having some uncertainty associated with Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 13
1 them (e.g., the Dittus-Boelter convective heat transfer, 2 has an inherent uncertainty of about 25%),
3 the net uncertainty () is given by:
4 =
5 6 This approach is widely used by engineers since it is relatively 7 easy to evaluate and it gives an acceptable approximation of the 8 error bars () in a Best Estimate evaluation.
9 Q. In Figure-1, the Best Estimate line is below the 10 Limit Line. Does this mean that everything is OK?
11 A. Not necessarily. As shown in Figure-1, a Best 12 Estimate prediction plus the uncertainty () may predict 13 possible RVI structure, components or fitting failures (i.e.,
14 CUFen = 1.0), sooner than the Limit Line would. Moreover, the 15 effect of embrittlement can make the situation much worse. The 16 problem is that no one knows how much conservatism, if any, is 17 in the Limit Line. As a consequence, we can have no 18 confidence that Limit Line results near CUFen = 1.0, are 19 sufficiently conservative to bound all the possible 20 uncertainties, as illustrated in Figure-1 by .
21 Q. In Figure-1, the Best Estimate line is shown for the 22 case in which there is no effect of embrittlement on fatigue, 23 BEne, and the case in Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 14
1 which embrittlement degrades fatigue life, BEe. Can you explain 2 what these lines mean?
3 A. Yes, the plot depicts two Best Estimate predictions 4 for a hypothetical RVI component: one under conditions of no 5 embrittlement (BEne), for which the Best Estimate CUFen line 6 continues to increase as the fatigue-inducing cycles (N) 7 accumulate with time, and the other Best Estimate prediction 8 (BEe) which includes the effect of embrittlement on the (now 9 time-dependent) maximum allowable cycles (Nf). For the latter 10 case, increases in fluence result in more embrittlement and thus 11 a more rapid increase in CUFen than in the former case. This is 12 because after a fluence of about 1017 n/cm2, significant 13 irradiation-induced damage and embrittlement begin to occur.
14 Embrittlement results in the loss of facture toughness and the 15 loss of ductility. Even though the data taken to date have been 16 inconclusive, there is ample evidence that this embrittlement 17 may reduce the number of fatigue cycles to failure (Nf) and thus 18 increase the corresponding CUFen, which is defined as CUFen = .
19 Q. What does location (1)in Figure-1 show?
20 A. Location (1) shows that, even if embrittlement is not 21 considered, RVI fatigue failure may occur by the end of life for 22 the period of extended operation. This is because the sum of a Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 15
1 Best Estimate prediction plus the uncertainty interval ()
2 associated with this prediction would exceed CUFen = 1.0. In this 3 case the Limit Line result is clearly non-conservative.
4 Q. May I now direct your attention to location (2) in 5 Figure-1. What does this location show?
6 A. Location (2) shows that embrittlement can result in 7 the physical failure of the RVI well before the end of life for 8 the period of extended operation.
9 Q. May I direct your attention to location (3) in Figure-10 1. What does this location show?
11 A. At location (3), the Best Estimate plus uncertainty 12 () for the case considering embrittlement (BEe) predicts 13 possible failure of the hypothetical RVI component well before 14 the end of life for the period of extended operation, even 15 though the Limit Line prediction indicates substantial margin 16 to failure at this point. That is, at location (3) the CUFen for 17 the Best Estimate line, accounting for embrittlement and for 18 all uncertainties (i.e., BEe + ), exceeds unity. Moreover, as I 19 have stated in my previous testimony, highly embrittled RVIs may 20 fail even sooner than by fatigue alone if significant seismic or 21 shock loads occur. That is, fatigue-weakened and embrittled 22 structures cannot tolerate large impulsive seismic and shock 23 loads like fully ductile structures can. Moreover, other Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 16
1 synergistic aging effects, such as the thermal embrittlement(TE) 2 of some CASS RVIs (the upper part of the core support columns) 3 would may move the BEe curve to the left of its position shown in 4 Figure-1 (i.e., the RVI would be predicted to reach failure, 5 CUFen = 1.0, even sooner than the cases (2) and (3) in Figure-1).
6 Q. Can any conclusions about the Limit Line and Best 7 Estimate lines be drawn from Figure 1?
8 A. Yes. Because of Entergys failure to explicitly 9 account for uncertainties, and for the effect of embrittlement 10 on fatigued RVIs, we can have no confidence in the Limit Line 11 type of CUFen predictions 12 13 14 15 16 17 18 19 20 Q. In response to Dr. Hopenfelds testimony, the USNRC 21 Staff stated that CUF or CUFen analyses are not required for 22 safety assessments of DBA events . . . because CUF and CUFen, 23 which are indicators of possible fatigue crack initiation, are Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 17
1 not a significant contributor to safety during DBA events.
2 NRC000168, at A145. How do you respond?
3 A. Again, this is a perfect example of silo thinking.
4 The USNRC Staff refuses to assess the risk that a highly 5 fatigued system, component or fitting which has been fatigue-6 weakened and embrittled as a result of neutron fluence, for 7 example, could fail, relocate, and degrade core cooling, when it 8 is subjected to a DBA LOCA or some other significant shock load.
9 This approach ignores the reality of a reactor environment, 10 where multiple aging mechanisms act simultaneously on RVIs and 11 other components, and that this age-related degradation needs to 12 be taken into account in plant safety analyses.
13 Q.
14 15 16 17 18 19 20 A. As I have stated above, ignoring the 21 effects of accident loads on aging RVI components is not at all 22 appropriate. While it is probably true that fatigued and 23 embrittled RVI components may operate normally during steady-Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 18
1 state operations, they can fail suddenly and catastrophically 2 when subjected to a significant shock load. Without considering 3 accident-induced loads, Entergy cannot provide adequate 4 assurance that the RVIs will continue to perform their functions 5 during the period of extended operation (PEO).
6 Q. Have there been RVI component failures at nuclear 7 reactors in the past?
8 A. Yes. Baffle former bolts and clevis insert bolts have 9 failed at several PWRs in the past, clearly demonstrating that 10 inspections alone will not detect aging effects prior to 11 component failure.
12 Q. With respect to the clevis insert bolt failures, do 13 you agree with the USNRC Staffs testimony that [t]he failed 14 bolts were detected visually (A290)?
15 A. Partially. According the SSER2 at 3-25 (Exh.
16 NYS000507), only 7 out of 29 damaged bolts, or about 24%, were 17 detected visually. The vast majority were not detected 18 visually. Nonetheless, Entergy has proposed to conduct only 19 visual inspections of clevis insert bolts. Exh. NYS000496, at 20 51.
21 Q.
22 23 Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 19
1 properly during steady-state operations, when a large percentage 2 of the incore bolting has failed, a significant shock load could 3 cause many of the remaining bolts to suddenly fail, resulting in 4 the relocation of core components and the possible loss of a 5 coolable core geometry.
6 Q. When discussing aging management of the baffle-former 7 bolts, the USNRC Staff describes the number of baffle-former 8 bolts in [t]hree-loop Westinghouse design PWRs like IP2 and 9 IP3[.] NRC000197, at A243. Is it accurate to say that IP2 and 10 IP3 are three-loop PWRs?
11 A. No. IP2 and IP3 are actually four-loop Westinghouse 12 designed PWRs. It is not clear why the USNRC Staff believes the 13 IP2 and IP3 reactors use a three-loop design. Perhaps this 14 reference is a cut and paste remnant from another proceeding 15 for a different facility that USNRC Staff has reused here.
16 Q.
17 18 19 20 21 22 Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 21
1 A. Virtually all metals (embrittled or not),
2 experience a decrease in ductility as the temperature decreases.
3 However, irradiated carbon steel undergoes a very sharp decrease 4 in ductility at a fluence-dependent temperature commonly called 5 the nil ductility temperatures (NDT), while stainless steel does 6 not. Nevertheless, as I have said before, sufficiently strong 7 shock loads can lead to failure of highly fatigue-weakened and 8 embrittled stainless steel RVIs.
9 Q.
10 11 12 A.
13 14 15 16 17 18 19 20 21 22 23 Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 22
1 2
3 4
5 The USNRC 6 Staff concedes that [g]iven the variability in assumptions made 7 by different analysts, it is difficult to explicitly quantify 8 the exact overall safety margin present in fatigue 9 calculations. NRC000168, at A210.
10 11 However, given the critical 12 siting of the Indian Point plants in the New York metropolitan 13 area, it is especially appropriate to identify and verify the 14 assumptions and calculations. Indeed, as distilled by the 15 concept of trust but verify, verification precedes trust.
16 Q. Do you agree with the USNRC Staff that a 17 propagation of errors analysis (i.e., an uncertainty 18 analysis), is not needed for EAF calculations?
19 A. No. the USNRC Staff claim that an 20 uncertainty analysis is not necessary because the EAF 21 calculation is deterministic and contains adequate 22 conservatisms. NRC000168, at A171; 23 Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 23
1 2
3 4
5 6
7 8
9 10 11 Q. So far you have been talking about reactor vessel 12 internals (RVIs). Do you have some similar concerns about the 13 fatigue evaluations for primary pressure boundary components and 14 fittings?
15 A. Yes, except for the fact that radiation-induced 16 embrittlement (IE) is not normally an issue, virtually all of my 17 previously discussed concerns are the same.
18 19 20 21 22 23 Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 24
1 2
3 Q. the USNRC Staff state that Entergy will 4 compare the actual transients experienced at Indian Point with 5 the transients anticipated in the EAF calculations. Does this 6 approach address your concerns?
7 A. No. The number of fatigue cycles is obviously 8 important; however, my principal concern has never been that 9 components will experience a greater number of transient cycles 10 than anticipated in the EAF calculation. In contrast, my real 11 concern for the integrity of primary pressure boundary 12 components is that a significant shock load - caused by a 13 seismic, LOCA, or other event - could cause the fatigue-weakened 14 component to suddenly fail well before their associated CUFen 15 value reaches 1.0.
16 17 18 19 20 21 22 23 Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 25
1 A. In my opinion, an observable surface crack of 2 3mm or greater, which would be expected when CUFen reaches 1.0, 3 is quite significant. However, microscopic cracks exist and 4 propagate within the metal structures, components and fittings 5 even when CUFen < 1.0. These microscopic cracks cannot be 6 detected by non-destructive testing (NDT), but they can weaken 7 the components and make them more susceptible to failures during 8 a significant seismic or pressure/thermal shock load event, 9 especially when coupled with embrittlement due various thermal 10 embrittlement or corrosion mechanisms.
11 12 13 14 15 16 A. what I 17 mean by not reducing design and safety margins as the plant ages 18 is that component fatigue life calculations should retain the 19 original design conservatisms and that the components should be 20 repaired or replaced if they exceed acceptable design margins 21 during the period of extended operations.
22 23 Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 26
1 In my opinion, removing 2 modeling conservatisms as IP2 and IP3 exceed 40 years of 3 operation, and the components become more and more degraded, is 4 both irresponsible and dangerous.
5 Q.
6 7
8 9
10 11 12 13 14 15 16 17 18 19 20 21 22 23 Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 27
1 2
3 4
5 6
7 8
9 10 11 12 13 14 15 16 17 Q. According to the USNRC Staff, aging mechanisms acting 18 on the core support columns are properly assessed from the 19 normal, steady-state operating conditions. NRC000197, at A306.
20 How do you respond?
21 A. This approach does not account for seismic or LOCA-22 type shock loads, which could cause the columns to fail.
Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 28
1 Q. I show you what has been marked as Exhibit NYS000563.
2 Do you recognize it?
3 A. Yes, it is an excerpt from an USNRC training manual, 4 entitled Reactor Concepts Manual - PWR Systems. It was 5 prepared by the USNRC Technical Training Center.
6 Q. What information did you draw from this USNRC 7 document?
8 A. This document identifies the relative location of 9 various components in a 4-loop, Westinghouse, pressurized water 10 reactor (PWR). On page 4-25, it shows that one of the 11 accumulators uses the same nozzle as the residual heat removal 12 (RHR) system, the low pressure coolant injection (LPCI), and the 13 intermediate pressure coolant injection (IPCI) safety systems.
14 In contrast, the high pressure coolant injection (HPCI) system 15 has a separate connection on the same cold leg as the 16 RHR/Accumulator nozzle.
17 Q. Why is that relevant?
18 A.
19 20 21 22 23 the components could fail Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 29
1 if subjected to a significant seismic event or shock load. As I 2 have pointed out in some of my previous ASLB testimony, the 3 failure of this important primary pressure boundary structure 4 could lead to a LOCA event in which the Accumulator, LPCI and 5 IPCI systems are not be able to inject water into the reactors 6 cold leg, and thus into the core, which, in turn, could lead to 7 core melting.
8 Q. Do you have particular concerns with respect to any 9 other components?
10 A. Yes.
11 12 13 14 15 16 17 18 19 20 21 22 23 Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 30
1 2
3 4
5 6
7 8
9 10 11 12 13 14 the USNRC 15 Staff has indicated that if transients accumulate at a rate 16 greater than the rate assumed in the fatigue calculation, 17 Entergy would be permitted to conduct yet another more refined 18 analysis. NRC000168, at A106. In short, there is apparently 19 no limit to the number of times CUFen can be recalculated to 20 obtain a CUFen result less than unity, and there is no standard 21 which defines the amount of conservatism that must be retained 22 in these calculations.
23 Q. Does this complete your testimony?
Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 31
1 A. Yes, it does. I do, however, reserve the right to 2 supplement my testimony if new information is disclosed or 3 introduced.
4 Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 32
1 UNITED STATES 2 NUCLEAR REGULATORY COMMISSION 3 BEFORE THE ATOMIC SAFETY AND LICENSING BOARD 4 -----------------------------------x 5 In re: Docket Nos. 50-247-LR; 50-286-LR 6 License Renewal Application Submitted by ASLBP No. 07-858-03-LR-BD01 7 Entergy Nuclear Indian Point 2, LLC, DPR-26, DPR-64 8 Entergy Nuclear Indian Point 3, LLC, and 9 Entergy Nuclear Operations, Inc. September 9, 2015 10 -----------------------------------x 11 DECLARATION OF RICHARD T. LAHEY, JR.
12 I, Richard T. Lahey, Jr., do hereby declare under penalty 13 of perjury that my statements in the foregoing testimony and my 14 statement of professional qualifications are true and correct to 15 the best of my knowledge and belief.
16 Executed in Accord with 10 C.F.R. § 2.304(d) 17 18 _________________________
19 Dr. Richard T. Lahey, Jr.
20 The Edward E. Hood Professor Emeritus of Engineering 21 Rensselaer Polytechnic Institute, Troy, NY 12180 22 (518) 495-3884, laheyr@rpi.edu Pre-filed Supplemental Reply Testimony of Richard T. Lahey, Jr.
Contention NYS-25 33