ML20149H453
| ML20149H453 | |
| Person / Time | |
|---|---|
| Issue date: | 08/22/1985 |
| From: | NRC |
| To: | |
| Shared Package | |
| ML20149B626 | List: |
| References | |
| FOIA-87-728, FOIA-87-853 REGGD-01.099, REGGD-1.099, NUDOCS 8802190274 | |
| Download: ML20149H453 (28) | |
Text
.
REGULATORY ANALYSIS REGULATORY GUIDE 1.99, REVISION 2 RADIATION DAMAGE TO REACTOR VESSEL MATERIALS 1.
STATEMENT OF THE PROBLEM One obvious constraint on the operation of a reactor is prevention of fracture of the vessel. This is accomplished, in part, by warning it before pressurization, following the pressure-temperature (P-T) limits given in the Tec.hnical Specifications (Tech Specs). Neutron radiation damage to the reactor vessel is compensated for bv chifting the P-T limits up the temperature scale every few years by an amount corresponding to the shift in the Charpy test transition temperature produced ey the accumulated neutron fluence. The NRC regulates this process on the basis of Appendices G and H,10 CFR Part 50.
Paragraph V.A of Appendix G requires:
"The effects of neutron radiation...are to b6 predicted from the results of pertinent radiation effects studies...."
Since Revision 1 of Regulatory Guide 1.99 was published eight years ago, I
there has been a significant accumulation of power reactor surveillance data, which constitutes a much more pertinent basis for the Guide than was available when Revision 1 was written. Revision 2 is based entirely on the surveillance I
data, and its issuance will provide a basis for licensing decisions that con-stitutes the most pertinent results available, in conformance with the regulation.
It may be asked why the Guide is needed if plants now have surveillance data of their own. Of course, for the newer plants such data are not yet avail-able. For many older plants, unfortunately, the materials in the surveillance capsules are not the controlling satorials for that reactor according to our present day understanding. Thus, instead of using the plants' own surveillance 002 %
g;h s L
'J 80021h02749 01A PDR s
PDR cot,M,,O,R,O7-720 1
REG ANAL RG 1.99 REV 2
,, os
I l
results directly, the staff must rely on calculated values based on the chemi-cal composition of the vessel materials and the neutron fluence.
Regulatory Guide 1.99 Revision 2 upgrades and expands the calculative procedures that are acceptable to the NRC, and it describes acceptable proce-dures for using plant-specific surveillance data when they become available.
The Guide is used in any regulatory action that requires knowledge of the fracture toughness of reactor vessel beltline materials.
Three examples of such actions are: (1) setting pressure - temperature (P-T) limits for heatup and cooldown, (2) evaluating transients that threaten the integrity of the reactor vessel, such as low temperature overpressurization and pressurized themal shock events, (3) evaluating flaws found during inspection.
In any of these analyses, a key input to the calculation is the fracture toughness of the material as a function of temperature. The ASME Code gives reference values of toughness as a function of temperature relative to RTNDT, the "reference temperature nil-ductility transition" of the material. The Code also describes how to measure the initial RT for the unitradiated material. This Guide gives calculative NOT caused by neutron radiation.
procedures for ARTNDT, the adjustment of RTNOT The Guide also describes how to combine the initial and the "delta" with a suit-NOT able margin to obtain a value of RT that covers the uncertainties in both.
From analysis of the new data base, and from experience in applying the Guide, the need for certain changes became clear.
Nickel has been found to increase the Charpy shif t in the presence of copper, and should be a factor in the calculations. Thus, some reactor vessels with high nickel welds, which were made when nickel was added in the welding process, have more susceptibility to radiation than previously thought. Conversely, some early reactor vessels that were made with no deliberate alloying additions of nickel have lower sensitivity to radia-tion.
Implementation of Revision 2 will remedy these situations.
4 2
REG ANAL RG 1.99 REV 2 08/22/85
-.-n n.,
.,,.,,,._,,,_.,._,_.,,,,--._,w.._,._--.,__,_..,.,,-.g._,.n,.,,,,----,_n,,-
,-,.,,.,-,,---,.----,,.sy.
l
\\
l The effects of copper and nickel content on the sensitivity of welds to l
radiation da' mage are different than they are for base metal--so different as to require separate treatment of welds and base metal in this Guide.
The fluence function needs revision.
The calculative procedure needs to be amended to prescribe mean values instead of upper bound values and to state the margin separately.
Procedures for calculating the attenuation of radiation damage through the vessel wall need to be stated specifically.
Improved knowledge of scatter in the surdeillance data base made it necessary to rewrite the criteria for use of plant-specific surveillance data in setting P-T limits for that reactor.
2.
08JECTIVE 2.a General Cbjective of Reaulatory Guide 1.99 The objective of the guide is best described by reference to the schematic The following discussion is pressure-temperature (P-T) diagram Figure 1.
mainly applicable to PWRs.
(See paragraph 4.a.1.5) The upper-left boundary of the operating zone, the P-T limit, appears in the Technical Specifications The P-T for all plants together with certain limits on heatup/cocidown rates.
limits are based on Appendix G,10 CFR 50, which incorporates Appendix G and And, the parts of Section III of the ASME Boiler and Pressure Vessel Code.
P-T limits are affacted by Regulatory Guide 1.99 as described in the previous Section.
In the upper-left corner of Figure 1 is a region labelled "hazardous to vessel integrity" which is bounded by a set of curves (instead of one curve) to This boundary moves indicate that higher cooldown rates increase the hazards.
upscale in temperature as radiation damage accumulates during the operating li The objective of the margins added in calculating P-T limits is of the vessel.
to place the operating zone for heatup/cooldown far enough from the hazardous REG ANAL RG 1.99 REV 2 3
08/22/85
l l
region to provide the operator time to diagnose and correct system transients such as low temperature overpressurizations (LTOPS) and rapid cooldown events
~
that could threaten vessel integrity.
2.a.1 Sources of Marcin in P-T Limit Calculations A discussion of the sources of margin that are present in the calculative procedure for P-T limits given in the ASME Code and NRC Regulations is in order.
First, the postulated flaw is a sente111pse 0.25T deep by 1.5T long (2.2 in. x 13 in., typically, for a PWR).
From the flaw size distribution used in the VISA code, the probability of such a flaw in the critical beltline weld of a reactor vessel is about one in 60,000. The probability may be debatable, but clearly there is margin in the 0.25T flaw assumption. Meyertheless, the use of a 0.25T flaw is an accepted feature of the Code procedure for calcula-tion of P-T limits, which we endorse. This is partly because the efficacy of the flaw detection and sizing in non-destructive examination is still debatable, and partly because some flaws are the result of esta11urgical conditions that also degrade the toughness of the adjacent material.
Second, there is a safety factor of two on the stress intensity factor due to primary membrane stress (in the belt.line, pressure stress) and a factor of one on that due to thermal stress. These factors were chosen by the Code writing bodies with the help of the basic document, WRC Su11stin 175,8 but the j
rationale for the factor of 2 on pressure is not explicitly stated therein.
A third source of mar 2in is the requirement that the toughness-temperature function used in these calculations should be the "K, curve" for crack arrest, g
called the "K curve in Appendix G, rather than the "K, curve" for static d
gg g
crack initiation.
(See Figure 4 of the Regulatory Analysis). The KIR curve is tbc lower bound of dynam!c end crack arrest toughness values for specimens some c4 vnich were full thickness.
Its adoption for Appendix G of the ASME Code "PVRC AdHoc Task Group on Toughness Requirements," PRVC Recommendations on Toughness Requirements for Ferritic Materials," Welding Research Council Bulletin No.175, August,1972.
08/22/85 4
REG ANAL RG 1.99 REV 2
-wmw--..--.c, m.,y.-,----
-,_..,--._,,y-,,,,.--.-.,~-,m-,
,.--,.r
._.,m,
~,_%.-.
.-.e-.--e,--,-a-
derived from a philosophy that was subscribed to by eeny, especially the researchers from the Naval Research Laboratory, that eeny service failures were the result of dynamic loading.
In a reactor vessel beltline, the pressure and therinal stresses are not applied dynamically in the sense intended here, not 4
even in a thermal shock situation.
However, one can postulate a case for rapid loading by postulating a small defect that is surrounded by a nugget of brittle metal that carries stress up to some critical level, then cracks open suddenly, presenting the sound metal with a running crack. Thus, the requirement to use the crack arrest toughness curve instead of the crack initiation curve is not entirely a matter of adding margin - in some unknown percentage of cases, it is the realistic thing to do. Whether based on this scenario or on simple conservatiss, the calculations required by the ASME Code for P-T limits are to be based on the K curve.
For evaluation of accident conditions, however, gg Section XI us'es the K, curve.
The temperature difference between the K,
g 3
curve and the K curve is about 65'F in the region of interest.
1c The fourth source of margin in the calculation of P-T limits is the use of lower bound toughness curves.
In the VISA code, however, toughness is simulated from the mean K curve with a distribution about the mean of i 10 per cent Ic (1 sipa). The use of a lower bound curve is consistent with Code philosophy in setting allowables, and was believed justified at the time it was drawn, because the data base was for only one heat of plate esterial. The temperature margin between the K, curve in the Code and the best-estimate curve used in j
the VISA code is about 40*F in the region of interest. (5ee Figure 5.)
The fifth source of margin is that required by Revision 2, paragraph and K, are given in terms of (T-RTET)*
C.1.a.(3). Toughness values, E g
gg
- '" ""I"
covers uncertainty in the initial RTET The margin added to RTET diated meterial as well as the uncertainty in ARTET, which is specified in Revision 2 to be 56'F for welds and 34*F for base metal. These values resulted from the regression analyses of the data and represent two-sipa upper bounds.
They are considered to cover the uncertainty caused by scatter of the data about the mean and uncertainties in the copper, nickel, and fluence. These variables j
are entered in the calculative procedure as best-estimate or mean values.
The margin also is assumed to cover uncertainty arising from possible differences between the copper and nickel contents of the weld and base metal samples and 4
I 08/22/85 5
REQ ANAL M 1.99 REV 2
{
l those of the actual vessel saterials at the location of interest in a fracture j
analysis.
In conclusion, it should be pointed out that the efforts to provide margin in the five areas described above are quite consistent with the requirements of General Design Criterion 31.
It states in part "The design shall reflect consideration of...the uncertainties in detersining (1) saterial properties, (2) the effects of irradiation on satorial properties, (3) residual, steady state and transient stresses, and (4) size of flaws."
Constraints on the amount of margin that can prudently be provided derive from the need for an operating zone of reasonable width for (fficiency in heatup/
cooldown operations and the presence of a lower limit on pressure at a given temperature based on avoidance of pump cavitation or thermal hydraulic problems caused by an approach to saturation conditions.
2.b Summary An attempt has been made to describe the objectives of Regulatory Guide 1.99 by placing it in context with the other documents that provide the basis for procedures to assure prevention of fracture of the reactor vessel and by describing the sources of margin provided in those procedures. The scope of the Guide is restricted to one part of the procedures.. to provide an acceptable basis to account for the effects of neutron radiation on the fracture toughness of reactor vessel materials. This is needed in the calculation of P-T limits, in analysis of tre.nsients that' threaten vessel integrity and in the analysis of beltline flaws.
The objective of Revision 2 is to upgrade the calculative procedures in Regulatory Guide 1.99 by basing them on the most pertinent radiation data and the best available understanding of radiation damage in reactor vessel mate-rials in accordance with 10 CFR 50, Appendix G.
The specific changes made in preparing Revision 2 are iteetzed in Section 1, above.
08/22/85 6
REG ANAL RG 1.99 REV 2
3.
At.TERNATIVES The alternatives to issuance of Revision 2 are to leave Revision 1 in place or to eliminate the Guide altogether.
The latter can be disposed of quickly: the staff reviews several P-T limits per year, plus an occasional transient and flaw indication and clearly needs a published basis for its reviews. There is at present nothing equivalent to Regulatory Guide 1.99 in the ASME Code. ASTM Standard Guide E-900-83 contains an equation relating the i
Charpy shift to copper content and fluence., but, it is out of date, and the Standard does not contain guidance on the use of plant-specific surveillance The alternative of continuing to use Revision 1 has a safety impact on data.
many plants and penalizes other plants; a detailed analysis of these conse-quences is given in the following section.
4.
CONSIOUENCES 4.a.
Costs and,8enefits of Alternatives 4
l 4.a.1.
Application to P-T Limit Calculations 4.a.1.1 Effects on P-T limits for All Plants.
In this section, it 1
will be shown that implementation of Revision 2 will mean that for about one-half of the operating reactors the P-T limits should be moved escale to higher temperatures by amounts that depend on fluence level and copper and nickel content of the beltline materials.
Referring again to Figure 1, this means that the region labelled "hazardous to vessel integrity" extends farther upscale ir. tamperature than would be pre-dicted if Revision 1 were continued to be used; hence the P-T limit should be moved also, to maintain the margin.
About one sixth of the plants will be able to operate with their present P-T limits for a longer period than previously deterstned based on Revision I.
The remaining third will be essentially unaffected.
Te deterstne the consequences of changing from Revision 1 to Revision 2 in our review of P-T limits, the first step was to calculate what differences would 08/22/85 7
REG ANAL RG 1.99 REV 2
. ~.
result if all plants should review their P-T limits this year first according to Revision 1 and than according to Revision 2.
The fluence value used for each plant was chosen, assuming that the goal was to have P-T limits that would be good for 4 or 5 years in the case of PWRs and somewhat longer for SWRs.
The importance of fluence level is shown in Figure 2 which illustrates for one saterial how the "trend curve" from Revision 1 compares to correspond-ing curves from Revision 2.
The trend curve from Revision 1 is an upper bound curve, hence for comparison, the Revision 2 values are mean-plus-margin, cal-culated as described in the Guide.
Note how the two curves from Revision 2 cross that for Revision 1 at fluences of about 3 x 10 s n/ca8 for base metal and somewhat higher for welds, indicating that plants having 0.35 Cu and 0.6 Ni and low fluences will be ratcheted whereas those having higher fluences will get some benefits. Figure 2 is drawn for 0.35 percent copper and 0.6 percent nickel.
For other compositions, the general appearance of the figure would be similar, with the crossovers occurring at different fluences.
In Figures 3 and 4, the numerical differences between shift values calcu-lated for Revision 1 and those for Reviden 2 (mean plus margin) are tabulated 2
for eip.t copper levels, three nickel levels, and seven fluence levels.
In the figures ; the boundaries between conditions for which a plant would be ratcheted (the positive values in the Table), and those under which plants would benefit (the negative values) are indicated.
Having Figures 3 and 4, the effect on each plant of a change free Revi-4 sion 1 to Revision 2 as a basis for the P-T limits was readily estimated on the basis of the copper and nickel content of its critical material and the i
fluence value described above.
For the eight plants ratcheted 50-100'F actual shift calculations were made following Revisions 1 and 2.
Our information on the limiting material and its copper and nickel content was quite good for PWR operating reactors, fairly complete for operating BWRs, but for many plants under construction it was limited to knowing whether or not the reactor vessel was bought to a low-copper specification. However, inspec-tion of Figures 3 and 4 shows that the numbers do not vary radically free box l
08/22/85 8
REG ANAL RG 1.99 REV 2
to box in most cases, and we believe the results are sufficiently accurate for discussion purposes.
Table 1 presents a summary of the changes in P-T limits that would result from a change from Revision 1 to Revision 2.
For example, the Table shows that, of the 81 operating light-water reactors (including three that are licensed only for low power testing), about 30 would find little difference (120*F) in the use of Revision 2 or Revision 1.
Eight plants (four PWRs and four 8WRs) would find that the use of Revision 2 would raise their P-T limits somewhere between 50'F and 100'F. The features that made this happen were:
high nickel welds in three cases, and low fluences in five cases.
Some 33 plants would be ratcheted 20-50'F. A total of 10 plants would bg benefitted a significant Most of these have low-nickel material (A302 8 phte and welds with asiount.
only residual nickel content).
Most owners of plants undergoing licensing would find the use of Revision 2 raises their P-T limits 20-50'F but the impact of this is small because flux reduction programs, the use of low-copper esterials, and better infomation will cause the expected end-of-life reference temperature about initial RTET to be less than 200'F even when calculated using Revision 2.
It is important to note that the fluences used in the calculations on which Table I is based correspond to about one-fourth of plant life on the average.
Inspection of Figures 3 and 4 reveals that for fluences characteristic of later life, which range from 1 to 7 x 101' n/cm8 for PWRs, and for copper levels that produce high-shift values, the ratchet effects of Revision 2 (as calculated for Table 1) disappear and become benefits during the latter half of the lifetimo.
4.a.1.2 Risk Avoided by usina Revision 2.
One source of risk of I' "**IIY continued operation with P-T limits based on Revision 1 when RTET higher (as given by Revision 2) depends upon the increased probability that a given transient will threaten the vessel. To illustrate this situation in Figure 1, consider that the hazardous region is expanded to the right, closer based to the P-T limit, because its extent really depends on a value of RTET on Revision 2.
Thus the operator will be aisled by erroneous P-T limits as to 08/22/85 9
REG ANAL. RG 1.99 REV 2
.)
1 j
the potential severity of the transient and will have less time to take corree-tive action to avoid the hazardous region.
To quantify the risk that would be avoided if the P-T limits were moved upscale to conform to Revision 2, several factors must be evaluated:
Expected frequency of transients that may violate P-T limits.
a.
b.
Expected severity of transients in tems of pressure and temperature as functions of time.
Reduction in severity caused by having proper P-T limits and therefore c.
more concern on the part of the operator and more time to take corrective action.
This factor must be evaluated as a function of the difference in P-T limits based on Revisions 1 and 2.
A quantitative 6 valuation of these factors was not undertaken, because it appeared that uncertainties, particularly in ites c, would be so large that the result might not be defensible.
Another source of risk is incurred by a heatup-cooldown operation following P-T limits that are lower than they should Le.
Consider, for example, one of the plants for which the P-T limits based on Revision 1 are 100*F below those based on Revision 2.
If a plant in that situation continued to operate with a P-T limit based on Revision 1, and if the operacer followed the limits closely in a heatup - cooldwn sequence, there would be grester probabi12ty of fracture of the vessel. A contract was given to Pacific Northwest Laboratories (PNL) to evaluate the increased risk, calculate the public exposure to radiation as a l
consequence of vessel failure, and calculate the costs resulting from a change to Revision 2 as the basis for P-T limits. The PNL report is Enclosure 6.
The change in the probability of fracture (Revision 1 - Revision 2) was calculated using a Monte Carlo technique and the VISA code, which originated at the NRC and is being further developed at PNL.
In this analysis, initial ET, toughness, copper content, fluence, and flaw size were treated proba-RT bilistically. Actually, in most runs, the flaw size had to be treated as I
08/22/85 10 RIG ANAL RG 1.99 REV 2
fixed at a large value (a 2 in. deep continuous flaw, probability of one) to get the Monte Carlo procedure to produce failures frequently enough to keep the required total runs to a reasonable number. Then the probability was reduced by a facter of 3500 to account for the more realistic flaw size distribution normally used, as explained in the ?NL report, The conclusion reached by PNL, based on the probabilistic fracture mechanics analysis, was that the best estimate of the increase in probability of vessel failure resulting from following P-T' limits that were 100*F too 1cw was 2.5 E-7 per heatup/cooldown cycle.
In the subsequent risk analysis, this result was used for plants ratcheted 50-100*F and half that amount for plants ratcheted 20-50'F. Using the techniques described in Enclosure 6 PNL completed tha calculation of increased public risk resulting from these vesssi failure probabilities. For. the lifetime of the present population of plants, the total mareres avoided was 1.1 E+4 man-res.
- 4. a.1. 3 Costs to Industry--PNL Value/ Impact Analysis. The cost of adopt-ir.g Revision 2 derives mainly from the cost of purchased power during delays in j
startup caused by the more restrictive operating limits.
In the PNL analysis, Enclosuro 6 the extra time involved was estimated to be 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> for the worst case of a 100*F ratchet, based on the opinions of their contacts among inspec-tors and utilities. When the P-T limit is moved upscale, the effect is to increase time spent at low pressure during heatup/cooldown, and the time was calculated by dividing 100*F by 50*F/hr, a typical heatup rate limit. The PNL estimate of industry operating cost is based on an average power cost of
$300,000 per EFPD for all plants. Thus, for the 8 plants that are expected to be ratcheted 50-100*F an estimated delay of 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> per heatup, and 6 heatup/
cooldown cycles per ycar, the cost will be $150,000 per year per plant for the next 25 years. The 77 plants, including those undergoing licensing, that are expected to be ratchett,d 20-50*F per year are assumed to incur half that amount.
The bet,t satinate present value of these industry operating costs was given by PHL as $101,000,000 and the net cost to the industry was $63,000,000. The bottoe line, best estimate, cost per man-rom saved reported by PML is "in the
$3,500-$5,600 range."
08/22/85 11 REG ANAL RG 1.99 REV 2 l
l l
4.a.1.4 Costs to Industry - NRC Cost Analysis Group. The NRC's Ccst AnalysisGroupiCAG)madeanumberofsubstantivecossnentsabouttheSeptember, 1984 draf t of the PNL report and also did a complete recalculation of the indus-try operating cost.
(See Enclosures 7 and 8).
In this work, CAG used their own plant specific power costs as well as plant specific values of the extra heatup time required as calculated by MEBR from the actual difference in P-T limits per Revision 2, ralative to those per Revision 1.
Their 1rdustry operating cost figure was $71,000,000.
If this figure is substituted for the PNL figure of $101,000,000 in the PNL cost estimate, the net cost to industry (See Sec-tion 4 of Enclosure 6) of $33,000,000 instead of the PNL figure of $63,000,000.
(Further reduction of this cost figure may be justified as discussed in para-3raph 4.a.1.6.)
This use of the CAG cost estimate reduces the botton line to
$1,800-82,900 per man-ree-saved.
There is also a paper work cost to the industry when a Tech. Spec. change is prepared and subaltted to the NRC.
PNL's test estimate of this is $2270 per plant--too small to enter into the cott benefit analysis.
4.a.1.5 Costs to Industry--Corrections Based on Cosaments Receiv e Durina Concurrence. The PNL estimate of 6 startup/ shutdowns per plant year was questioned by the staff.
It had been based on reactor scram data available to PNL. However, af ter a scram, The plant does not necessarily go to cold shut-down.
In the plant operating data given in NUREG-0020, shutdowns greater than 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br /> are tabulated for each month.
After correcting for the refueling shut-down, which extends over several months, the average over a three-year period, 1982-4 was 4.3 shutdowns greater than 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br /> per reactor year. From an ORNL analysis of plant operation for 1982, there were approximately 2.0 shutdowns per reactor year greater than 120 hours0.00139 days <br />0.0333 hours <br />1.984127e-4 weeks <br />4.566e-5 months <br />.
Based on these data, and assuming that some of the 72-hour shutdowns were not cold shutdowns, it seems reasonable to assume that there are no more than three startup/ shutdown operations per reactor year instead of six as asstaned by PNL. This correction reduces the botton line to $900-1400 per man-ree-saved.
Consideration has also been given to the estimate of 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> extra hea. tup time, caused by the P-T limits being moved upscale by 100'F. The PNL estimate was based on simple division, assuming a heatup rate of 50'F per hour.
The 08/22/85 12 REG ANAL RG 1.99 REV 2
rationale was that any restriction to the operating zone (see Figure 1) would I
From data require more care in heatup/cooldown and therefore take more time.
located by R. R. Riggs, SPEB, in EPRI NP-1139, Vol. 2 for PWRs, the average The startup time from cold shutdown to hot standby conditions is 20 hours2.314815e-4 days <br />0.00556 hours <br />3.306878e-5 weeks <br />7.61e-6 months <br />.
If it takes 34 hours3.935185e-4 days <br />0.00944 hours <br />5.621693e-5 weeks <br />1.2937e-5 months <br /> for an average shutdown time, hot to cold, is 14 hours1.62037e-4 days <br />0.00389 hours <br />2.314815e-5 weeks <br />5.327e-6 months <br />.
average startup/ shutdown operation (from cold shutdown to hot standby and back) it appears that 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> is almost "lost in the noise."
There must be a number of operational factors that determine the critical path, and it is doubtful that the restriction in the operating zone is normally a critical path ites considering the number of testing and operational proce-dures involved in a startup from cold shutdown. Moreover, the ratchet effect of Revision 2 is at a maximum early in plant life when the operating zone has In effect, udoption of Revision 2 will simply mean that its greatest breadth.
anticipated restrictions in the operating zone will c: cur sooner than expected.
Consequently, among the list of uncertainties to be itemised later, there is the judgment that the 2 hour2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> 11gure, and the corresponding cost to industry of
$25,000 per startup/ shutdown, is probably an upper bound, and the lower bound is nearly zero.
Conservatism in the PNL analysis that affects both cost and risk is the assumption that the ratchets shown in Table I are constant for the remainder For the majority of those ratcheted 50-100*F, this may be trus, of plant life.
but for sees plants the ratchet steadily decreases and becomes a benefit as early as mid-life of the plant. However, the risk reductions also diminish.
It has been assumed that their ratio, the dollars per man-rea "bettes line" remains constant. Clearly, there is uncertainty in the assumption that the risks and the present value of the costa are the same function of the ratchet, but it has been made for want of better information.
4.a.1.6 Boilina Water Reactors -- Special Considerations. Boiling water reactors, four of which would be ratcheted 50-100'F according to Table I, should probably be ositted from the calculation of both cost and risk associated with heatup/cooldown operations, because saturation conditions govern the pres-sure at any temperature even during transients if the vessel is not water solid.
This considerstion would reduce the $33,000,000 industry cost estimate to about
$21,000,000, or about one-third of the original PNL cost estimate.
13 REG ANAL RG 1.99 REV 2 08/22/85
l It is during hydrotests and leak tests that 8WRs may be affected by Regula-tory Guide 1.99. -The sargins required for hydrotest are given by the ASME Code and Appendix G,10 CFR 50 just as they are for normal operation, except the factor of safety on pressure is 1.5 for,hydrotest instead of 2.0.
A number of BWRs are now at the level of radiation damage where the required metal tempera-ture at the hydrotest pressure is nearing 200'F. This causes two problems, according to verbal reports free some BWR owners representatives. Heatup with pump power apparently is slow in SWRs, compared to PWRs. Also, prolonged opera-tion of the pumps at low pressure causes extra wear of the pump seals.
In any case, it takes longer to get to the higher temperature.
The other probles arises from a requirement in the Technical Specifications that the containment drywell be closed when water temperature approaches 212*F.
Inerting is not required for 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> or more, so entry can,be made for leak test inspections, but the time required is increased. Yet, the hydrotest temperature for PWRs exceeds 200*F in many cases: hence, the problem of inspect-ing for leaks sust be manageable. These delays cccur after refueling, nonsally about once every 18 months.
Data from R. Riggs (EPRI NP 1136, Vol.1) show the aman time to conduct a hydrotest of a BWR is 28 hours3.240741e-4 days <br />0.00778 hours <br />4.62963e-5 weeks <br />1.0654e-5 months <br />, with a standard deviation of 6 hourt. Thus the impact (time / cost) on SWRs likely falls within the stand-ard deviation of the hydrotest time interval.
Risks incurred during hydrotest are probably small.
It is debatable whether or not the margin of 1.5 on pressure required by the Code, compared to a margin of 2.0 for normal operation, properly weighs the relative risks. However, any change in these values involves an amendeent to #ppendix G,10 CFR 50 and is
~
beyond the scope of Regulatory Guide 1.99.
It is co xluded that the SWR hydre-test situation represents little change in the direction of cost per aan-ree-saved ($900-$1400. as discussed in 4.a.1.5).
The effect is undoubtedly plant specific because the conclusion depends on operational details involved in conducting hydrotests and leak tests.
4.a.1.7 Effects on LTOP Limits. After a tumber of low temperature overpressurizations occurred in the 1970's a requirement was placed on PWRs to provide automatic pressure relief at low temperatures that would prevent viola-tion of the pressure-temperature limits. Some plants accomplish this by enabling 08/22/85 14 REG ANAL RG 1.99 REV 2 1 1
l a low setpoint setting on the pnssure operated relief valves (PORVs) when at low temperatures., others use the relief valves in the residual heat reeoval j
system (RHR) to provide this function.
For PWRs that would have the P-T limits moved up the temperature scale if the limits were based on Revision 2, the allowable pressure at the cold shut-f down temperature, say 140'F for example, would be slightly reduced. The amount is small, because the P-T limit curve is flat at those temperatures, especially for vessels with a large radiation shift.
Unfortunately, the system constraints are tight. The reactor coolant pumps (RCP) used in Westinghouse plants require a pressure of about 325 psig minimum for proper function of the seals.
Typical settings of the PORV for LTOP protection are 400-500 psig for temperatures up to about 350*F. This restriction on the operating zone probably accounts for most of the impact on startup time.
In his memorandum to B. D. Liaw, February 27, 1985 commenting on Aavision 2, B. J. Elliot wrote:
Thus, to begin heat @ with these limitations, the reac-ter vessel is pressurized to approximately 350 psig us-ing RHR and CVCS pumps prior to initiating flow through the RCPs. The RCPs are the main source of heat prior to core critical operations. At temperatures below 350*F, the reactor coolant pressure (RCP) must be maintained below the LT0P pressure set point of 400 psig, since actuatien of the LT0P valves would depressurize the system and heatg would have to begin again. The LTOP j
set points for pressure and temperature are derived free the pressure-temperature limits. Hence, a change in the pressure-teeperature limit requires an adjustment in the LTOP set points. The narrowing of the operating band that results free RCP sinimum pressurize and the LTOP set point, as plants age, has been recognized as a probles by the staff and NSSS vendors. NSSS vendors are reevaluating critical component limitations and operating procedures to determine whether the low ten-perature operating band say be widened. The staff, in a letter from B. W. Sheron to W. Minners dated August 1, 1984, has requested that the LTOP criteria for operat-ing plants be prioritized as a new generic issue.
The last sentence it; the quotation refers to another aspect of the LTOP issue. At Maine Yankee, the requirements for LT0P protection have been W 22/85 15 REG ANAL RG 1.99 REV 2 4
____-__,_m-_-.-_.-
broadened to mean there should be automatic protection for the P-T limits at f
high temperatures 'as well as low.
Until recently, this was provided by the normal high setpoint on the PORVs, because the temperature at which the RHR 1
system was isolated was higher than the P-T limit temperature at 2250 psig.
(See AE00 Engineering Evaluation Report No. E426, October 24, 1984, by E. V. Imbro).
Actually, Maine Yankee is not affected by Revision 2, because there are sufficient credible surveillance data. They show relatively large shift, however, and the P-T limits have been moved upscale se far that the operators elected to add a PORY setpoint at an intermediate pressure.
At present, the NRC does not require automatic protection of the P-T
~
limits at the higher temperatures.
If that becomes a requirosent,, Revision 2 may have some impact on this new aspect of LT0P requirements as well.
4.a.1.8 Costs to NRC. The immediate impact of proposed changes on MRC staff review time will be minimal, because we plan to implement Revision 2 for P-T limit calculations at the regularly scheduled times, as described in Section 6.
However, the 3-year limit on implementation will crowd the calendar f
for review of P-T limits somewhat.
4.a.1.9 Summary of Uncertainties and Conclusions re: Aeolication of Revision 2 to P-T Limits. The conclusion reached by PNL - that the costs per f
man-ree were in the range $3500-5600 tus been corrected based on more detailed power cost analysis by CAG (paragraph 4.a.1.4) and on data collected by the staff with regard to the number of cold shutdown per reactor year to yield a l
The following uncertainties have been identi-value of $900-1400 per aan-ree.
fled in this result the not offact of which is believed to reduce it signif t-i cantly, but it has not been possible to further quantify the number.
The most important uncertainty is in the risk tem.
In paragraph 4.a.1.2 and Figure 1, it was shown that the principal safety function of P-T limits is to separate the operating zone far enough free the zone that is hazardous to vessel integrity to give f.ine for corrective action to mitigata a transient.
16 REG ANAL RG 1.99 REV 2 08/22/85
^
l l
Regulatory Guide 199, being part of the bases for calculating P-T limits, affects the sargin provided.
For reasons given previously, it was not possible to quantify the risk avoided by a given correction to the P-T limits, but it is believed to be a larger source of risk than the one PNL was asked to evaluate.
The principal The second most important uncertainty is in the cost ters.
cost to industry is the cost of power not generated because of delays in startup/ shutdown operations caused by restrictions in the operating zone if the P-T limits are moved upscale. The PNL estimate of two hours per startup is believed to be high for most plants, as discussed in paragraph 4.a.1.5.
The narrowing of the operating zone resulting from adoption of Revision 2 in certain plants is not significant early 14 life when the operating zone is Moreover, it is no more severe than was expected to occur late in life.
broad.
Finally, at higher fluences the "ratchet" disappear,s, because of differences in the fluence functions given in Revisions 1 and 2.
All of these uncertainties 4
are in the direction of significantly reducing the PNL cost satinate.
4.a.2 Other Applications of Regulatory Guide 1.99 J
aust be calculated as part of Regulatory Guide 1.99 is used whenever RTET 1
an analysis of a transient that has actually occurred. The analysis provides the basis for deciding if the possibility that the vessel has been damaged is sufficiently high to warrant an inspection before returning it to service.
Such inspections are time consuming and costly in terus of power replacement.
Another application of Regulatory Guide 1.99 is in the analysis of flaws 1
A recent example l
found by inservice inspection of the reactor vessel beltline.
occurred at Indian Point 2.
First, the draft Revision 2 was used to calculate at tM inside suffete, based on the fluence and the weld and plate RTET Second, because the flaw was near the outside surface, the formula chemistry.
through the vessel wall, (a feature of Revision 2),
for attenuation of ARTET et the tiP of the flaw.
The evaluation is the was used to calculate ARTET basis for deciding if the vessel must be repaired before being put back in service.
17 REG ANAL RG 1.99 REV 2 0s/22/s5 j
l i
i In evaluations of transients and flaws there may be significant safety questions the cost impact say be high, and contention over the decisions say surface.
Clearly, it is important to have a basis for the calculation of radiation damage effects that h3s had public review and resolution of out-standing issues.
4.b.
Impact On Other Requirements 4.b.1 Impact on the Pressurized Thermal Shock Rule In the PTS rule, FEDERAL REGISTER, July 23, 1985 there is a formula for to distinguish that method from others), which calculating RTET (called RTPTS is based on an early version of the formulas that are the basis for Revision 2.
There are several differences.
For the calculation of RT (1) weld and base PTS.
metal data were analyzed as one data bcss yielding one correlation function, (2) the fluence function was of simpler fors, and (3) there was a second ega-tion which gave bounding values.
It is not intended to change the proposed PTS Rule to incorporate the (Note, however, the action reconnended below.) The latest formula for RTET.
calculation of RT required by the PTS rule is associated with screening cri-MT teria (270*F for base metal and axial welds,'300*F for circumferential welds),
l which were justified by a probabilistic ar,alysis that considered all identifi-At the time able uncertainties including those in the calculation of RTET.
the Rule was drafted, it was recognized that there would be an evolution in the calculative procedures, but there would be too many actions taken by the utili-ties in the area of flux reduction programs and other measures to permit fre-spent changes in the position of every plant relative to the screening criteria.
However, when it appears that the screening criteria will be exceeded at a I
specific plant, the PTS Rule requires an evaluation of all aspects of the PTS analysis as they apply to that plant. That re-evaluation mest include the will be re-evaluated fracture toughness of the beltline material and thus RTET too.
By that time, a number of plants will have credible surveillance data of their own. If not, and if Revision 2 of Regulatory Guide 199 were used. RTET values would be higher for a few vessels having high-nickel welds but most would be slightly lower.
06/22/85 18 REG ANAL RG 1.99 REV 2 i
=. _ _ _. - _ - _ _ _ _ _ - _. _ -. _.... - -.,,.___.-..---..,_,_.-_ - _.--_ _,__,_-,_ _.-_--, -...,. -_,.-,. - _ - _
l' If the calculative procedures of Revision 2 were substituted for those of the PTS Rule, and if the present information about copper, niemel, and fluence values is confirmed by the submittals required by paragraph (b) of Section 50.61 of the PTS Rule, it appears that eight plants would reach the screening cri-terion before E0L. Of the eight plants, four would reach the scusning cri-Based on terion in the years 1993 to 1996 and four in the year 2000 or later.
these findings, it is recommended that the copies of proposed Revision 2 sent to utilities and others be accompanied by a letter containing the following statement of the relationship of the proposed Guide to the PTS rule.
The calculative procedures given in paragraph C.I.a f this draft. Guf de are not the same as those given in the Pressurized TtEarsal Shock rule
- for calculating RTPTS, the reference t**perature that is to be compared to the screening criterion given in the rule.
Issuance of Regulatory Guide 1.99, Revision 2 for public comment in no way affects the recently promulgated PTS rule.
Licensees and the technical community are requested to comment on the technical merits of this proposal, including its effect on their plants for non - PTS purposes, chiefly as the basis for calcula-tion of pressure temperature limits as required by Appendix G,10 CFR l
Part 50. Licensees may also consider and comment on the proposed change's effect on the calculated PTS risk at their p13nt, assimita tha' Revision 2.
correlation, if justified, would at some future. ties replaie tho' RTPTS correlation in the PTS rule. Following resolution si comments, aisd once general agroceent is reachad regarding the best weg to calculate RT2T '
then it will be appropriate to re-evaluate th6 everall conservatism of the PTS rule.
i 4.b.2 Iepact on Material Selection l
For plarets in the very early construction phase before the reactor vessel materials have been ordered, for which the provisions of Position C.3 of the I
a%Y Asve some Guide are applicable, the new procedure for calculating RTET effect on the limit specified for copper content.
(Position C.L itself is t
i "Fracture Toughness Requirements for Protection Against Pressurized Thermal l
Shock Events," Federal Reaister, July 23,1985, pp.19937-29945.
' 19 REG ANAL RG 1.99 REV 2 08/22/45
unchanged from Revision 1). However, sample calculations indicate that there should not be many instances of a negative effect, because they are made for end-of-life fluences.
For plants for which the surveillance materials have not been selected, the changes from Revision 1 to Revision 2 may affect the decision as to which beltline materials will be controlling and therefore which weld and base metal should be represented in the surveillance capsules. However, because we are not changing this requirement, there should not be any extra cost involved.
4.c.
Constraints We have not identified any constraints such as scheduling or enforceabil-ity that affect the implementation of Revision 2.
5.
DECISION RATIONALE Based on the foregoing analyses of the safety issues, system impacts and i
costs it is recommended that Revision 2 be issued for public comment. The analysis has shown that the Guide is needed, because it provides part of the basis for ensuring safe operation of reactors during startup and shutdown and for the evaluation of transients and flaws found in service.
Periodic updating of the Guide is consistent with the requirements of Appendix G,10 CFR Part 50.
Adoption of Revision 2 will raise the P-T limits for about half of the operating reactors that now use Revision 1 as the basis for these limits.
In
~
preparing the value/ impact analysis, based on contract work by PML, one source of avoided risk by going to Revision 2 w6s quantified. After making corrections based on staff comments, the cost-benefit ratio was in the range 8900-1400 per man-ree avoided. However, there is another, more significant source of risk which was not quantifiable. As discussed in Section 4.a.1.2, the principal safety impact of operating with lower P-T limits (i.e., contiadng to use limits based on Revision 1) will occur during a teensient, because the operator (a) will not have accurate inforsatter, of the potential hazard to vessel integrity, and (b) will have less time to take corrective action.
In paragraphs 4.a.1.4 and 4.a.1.5, reasons have been given to believe that the amount of replacement power.
08/22/85 20 m ANAL M IM M 2 L
i which is the principal cost impact of implementing Revision 2, will be lower for In summary, most plants than the values determined in the PNL cost estimates.
r it is believed that the cost benefit ratio is in the range of a few hundred Because there is a significut safety benefit, and dollars per san-res avoided.
because it is cost effective, the implementation of Revision 2 is recommended.
Finally, we believe that industry will be receptive to the proposed "trend curves," as they are cosmonly called.
Based on industry response to presenta-tions to ASTM Cosmittee E-10, the Metal Properties Council, and ASME Boiler Code Section XI working groups concerned with this subject, there appears to At this time we know of no be general agreement on the need for new curves.
There may serious objection to the calculative procedures given in Revision 2.
be disagreements about the cheelstry factor or fluence factor near the edges of the data base, where extrapolation is required, and about the margin to be We intend to push for adoption in ASTM Standard Guide E900 as well as added.
in Section XI of the ASME Boiler Code, but this will take at least a year or If implementation of Revision 2 is delayed, there will most Itkely be longer.
more negative ballots by those who expect to be impacted by the new trend curves.
There will continue to be a need for Regulatory Guide 1.99 to provide acceptable treatment of the question of margin, the treatment of plant specific surveil-lar,ce data, and the calculation of attenuation of damage through the vessel wall.
In Revision 2 Further revision of Regulatory Guide 1.99 is to be expected.
Position C.1.b. and Figure 2, which presents a "trend curve" for the percent decrease in upper shelf energy as a function of fluence and copper content, That needs to be upgraded when the basis becomes available in a year er so.
effort will not affect P-T limits and it will affect only relatively few Further adjustment of plants--those having lod upper shelf energy initially.
the calculative procedure for reference temperature may be indicated in a few years when more surveillance data have accumulated or when our understanding l
of embrittlement mechanises and the role of copper, nickel and other elements improves.
d I
s REG ANAL RG 1.99 REV 2 21 08/22/85 l
l l
6.
IMPLEMENTATION Paragraph D.1. of the Implementation Section of Revision 2 reads much as it did for Revision 1, the key sentence being:
...the methods described in regulatory position C.1 and C.2 will be used in evaluating all predictions of radiation damage needed to implement General Design Criterion 31 or as called for in Appendices G and H to 10 CFR Part 50 submitted after the effective date i
of publication of the Guide." This means that plants will be allowed to con-tinue to follow the present schedule for updating their P-T limits, but only within a 3 year period.
Paragraph D.2. requires all operating plants to review the limits within a 3 year period after the Guide becomes effective and to revise them if necessary. The decision to forego prompt implementation across the board is based on the existence of significant margins, as discussed in i
Section 2.a.1 and elsewhere in this analysis.
In the staff's judgment, the risk of allowing some plants to operate three more years with present F T limits was justified by the reduced impact on the Industry and the NRC achieved by -
not roquiring a complete review of all P-T limits in a 6-12 month period.
No staff actions will be required to implement the Guide other than making sure that each facility gets a copy.
7.
OTHER IMPACTS J
There are no other actions, systems or prior analyses known to need reassessment as a result of publication of Revision 2.
This Guide does not add to the reporting or information collection require-sents of licensees, nor does it affect small entities as defined in the Regula-tory Flexibility Act.
j 1
i 08/22/45 22 REG ANAL RG 1.99 REV 2 i
L_
TABLE 1.
SUPHARY OF THE CHANGES IN PRESSURE-TEMPERATURE LIMITS EXPECTED To RESULT FROM A CHANGE FROM REVISION 1 TO REVISION 2 0F REGULATORY GUIDE 1.99 Effect of Change Operating Reactors Plants Undergoine Licensina from Rev. I to Rev. 2 PWR 8WR Total PWR 8WR Total Ratchet 50-100' 4*
4 8
Ratchet 20-50' 16 17 33 32 12 44 No Change (120')
23 7
30 3
0 3
)
Benefit 20-50'
- /
1 8
8enefit 50-100*
1 0
1 Benefit 100-150' 1
0 1
Totals 52 29 81 35 12 47
- Values in the table are number of plants.
i l
I O
M/22/85 23 REG ANAL RG 1.99 REY 2
l
~
'\\
/
HAZAltD005
'.T0 VE550.
INTIGit!TY W 1' s
COX.,0WN W
3 i
L p
a i
i j nminum 4 6
~~
TDFERATUltE 1
j FIGuitE 1 SCHDETIC PRE 15UltE-TDFERATUltE DIAGRAM i
4 t
l The thred curves represent different cooling rates.
- Nota:
Curve 1 being the slowest.
\\
d.4
~
. :. ~ -
. =.
~
- g, e.
,.u
=.
~
~.
- f 4
ER
- r e
~
o.
.r.
B
~
E A
.k g.
g
,=
- 9,A n
4 8
g
=
y
~w 4
=
-j
=
n tt 1
=
k a
4.
S-E4 l
N
-t 2g
=
t
}E
~
n.
gg
=
g l
t t
a
~
A
-.g' i
-...i. I s s.: s s-a i
a e
l l
t i
I HAZARD 00$
'70 ytssEL INTEGRITY s RAPI' COOL;0WN 1
l 3
../ L v
(A l
h f
>n i
4 i i
\\
)
TDFIRATWLE FIGURE 1 $CHDETIC PRES $URE-TDFERATURE DIAGRAM i
- Nota: The thre4 curves represent different cooltag rates.
Curve 1 being the slowest.
1 e
2.4
- g, p.
g m.
1
.r.
3
~
E g
A g
h sQ n
i.
]3
-4 s R
E s
8 3.2
=
g
~w n.
.es b,
t 4
i
- W n
lb'
~
g 1
e 1a
=
~ r S-E:.
44 t.i 9
,e t
a
~
2 g
I g
..h 5y,
=
[
~
[
e m.
i
- h n,
l l
11 1
l I
. g )l
.s s s t s-a
.. i.: E K
1 i
0%
P*
,a h
e a em Yt si n
i e : als l4 iii t a
- eit,
.s'g I%
D E
E D
3 w
~
w s
,e
.n 4
4 h
E 4 4 w
R D
I E
5 k
D N
N V
k D
s a
R 0
W T1 s
}
N9 T $$
Y i
lk t
Jf t a
SL V 9
5 t
4%
G, i 94 e
3 t
9 5
4
- }
B
?
Nt a
E M
(d(w t
u e a
t t
's
$'$=
$s t
e u m
a :
s y
%k
%k k
5 4
a t
4 3 ki3 I I N
Qh
@(
1 4
4 5
4 1 9 4
I 4 nt :
4 4
?
?
4 9
L b
( Na s
a 9
g 9
5 9
tit h et a s t
5 4
9 hl s{f w
g
- t ta e
t at e
g:
5!q
'e t t;
t E
4 4m 1
w o
4 o
w w
9 M
4 4
A.
e+5 o
g D
4 6
6 4
4 4 dy
=
y
)of g
es
.-y ig 9
9 g> 9 9
9 y; r
14 44 9
4 i fs n
s s vt
=,
3 m gty ~,
.s a
=
=
e E Et*
w a
t t
t n
-w 1.,l t=t 1+
x c
d t
a a
m Ns a
e e
e a
e a
1 N5 4
4 4
?
i s t
1x 1
t J#
4 4 9
9 4
5 9
n av
~i t
e
- e. m. a. a.
4 s
9
" I,
~
g[~ "
i g
t t
4
?
i8 9
(d y
$'t fe e
i g p
1*
w x
a t * *-
- t 2
47 W.h
\\
V 9 5
6
% )
J,8 gg.
l gf 1 1 4
?
6 4 4 V
i I 4 t n t ? 4 8!
9 4 ? ?
L p
ib t0=
=
5 y 9
i 9
a x
s
%$g
%w s
,o.
t t
=.
v s
o w
k I
Y
- )k; s
C 2 t t
6
%4
'e k
- b. 2 4 BE I
l Ci t a
p gp 27
~-
,, s.n.,
unms xi-nn: n.s :
iw i... s.
E
~
+
ye b
i a x l5 l
R A
5 E
J
- ii x
g
'ss '#
- d
[
l2 55 N s ft s
E 5
\\
Ea W
$ !!!l
\\
\\
g lal
)-
N
\\
\\
\\
I J
l I' E
\\\\
XI k '1 I
.1 8
I I
I I-5 I
a a
e n
=
s m 'e e
n n s m n,
&8
-