ML14092A301
| ML14092A301 | |
| Person / Time | |
|---|---|
| Site: | Palisades  |
| Issue date: | 12/31/1981 |
| From: | ABB COMBUSTION ENGINEERING NUCLEAR FUEL (FORMERLY |
| To: | |
| Shared Package | |
| ML13308A045 | List: |
| References | |
| RTR-NUREG-0737, RTR-NUREG-737, TASK-2.K.2.13, TASK-TM CEN-189, TAC-46893, TAC-59981, TAC-59982, NUDOCS 8201130480 | |
| Download: ML14092A301 (155) | |
Text
CEN - 189 EVALUATION OF PRESSURIZED THERMAL SHOCK EFFECTS DUE TO SMALL BREAK LOCA'S WITH LOSS OF FEEDWATER FOR THE COMBUSTION ENGINEERING NSSS Prepared for the C-E OWNERS GROUP NUCLEAR POWER SYSTEMS DIVISION DECEMBER, 1981 IP POWER SYSTEMS 82 0 1 a y o DMBUSTION ENGINEERING, INC.
LEGAL NOTICE THIS REPORT WAS PREPARED AS AN ACCOUNT OF WORK SPONSORED BY COMBUSTION ENGINEERING, INC. NEITHER COMBUSTION ENGINEERING NOR ANY PERSON ACTING ON ITS BEHALF:
A. MAKES ANY WARRANTY OR REPRESENTATION, EXPRESS OR IMPLIED INCLUDING THE WARRANTIES OF FITNESS FOR A PARTICULAR PURPOSE OR MERCHANTABILITY, WITH 'RESPECT TO THE ACCURACY, COMPLETENESS, OR USEFULNESS OF THE INFORMATION CONTAINED IN THIS REPORT, OR THAT THE USE OF ANY INFORMATION, APPARATUS, METHOD, OR PROCESS DISCLOSED IN THIS REPORT MAY NOT INFRINGE PRIVATELY OWNED RIGHTS; OR B. ASSUMES ANY LIABILITIES WITH RESPECT TO THE USE OF, OR FOR DAMAGES RESULTING FROM THE USE OF, ANY INFORMATION, APPARATUS, METHOD OR PR9CESS DISCLOSED IN THIS REPORT.
ABSTRACT This report has been prepared in response to Item II.K.2.13 of NUREG-0737, "Effect of High-Pressure Injection on Vessel Integrity for Small-Break Loss-of-Coolant Accidents with No Auciliary Feed water". This set of events, or scenario, is not considered to be part of the plant design basis, and has not been evaluated pre viously in any licensing analyses. The body of this report des cribes the methods of analysis used to evaluate the integrity of the reactor vessel pressure boundary for the prescribed scenario.
Plant-specific results are provided in a separate Appendix for each plant. This report concludes that all C-E plants can withstand the postulated small-break loss-of-coolant accidents with extended loss of feedwater scenarios for the assumed life of the plant.
The evaluation of the prescribed scenario is not considered to expand the plant design basis to include extended loss of feedwater events.
CEN-189 TABLE OF CONTENTS TITLE PAGE CTION ABSTRACT 1 PURPOSE 1-1 2 SCOPE 1-1 3 INTRODUCTION 3-1 4 THERMAL HYDRAULICS ANALYSES 4-1 5 FLUENCE DISTRIBUTIONS 5-1 6 MATERIAL PROPERTIES 6-1 7 VESSEL INTEGRITY EVALUATIONS 7-1 8 CONCLUSIONS 8-1 PLANT SPECIFIC APPENDICES A. FORT CALHOUN B. CALVERT CLIFFS 1 & 2 C. MAINE YANKEE D. PALISADES E. MILLSTONE 2 F. ST. LUCIE 1 & 2 G. ANO-2 H. SAN ONOFRE 2 & 3 I. WATERFORD J. PALO VERDE 1,.2 &,3
1.0 PURPOSE This report provides the results of analytical evaluations of pressurized thermal shock (PTS) effects on reactor vessels in Combustion Engineering NSSS's for the case of a small break loss-of-coolant accident (SBLOCA) with the assumption of concurrent loss of all feedwater (LOFW). This report is intended to respond to Item H.K.2.13 of NUREG-0737.
2.0 SCOPE This report considers the thermal and mechanical effects of several different SBLOCA and LOFW cpses, which were chosen to be consistent with measures which might be taken to prevent core uncovery in order to assure adequate core cooling. A range of break sizes, locations and high pressure injection characteristics were evaluated for the following two scenarios:
- 1. SBLOCA + LOFW followed by opening of power operated relief valves(PORV) in time to prevent core uncovery
- 2. SBLOCA + LOFlW followed by recovery of feedwater in time to prevent core uncovery PTS cooldown transient scenarios beyond II.K.2.13 will be considered separately.
The plants considered in this report are:
- 2. Calvert Cliffs - 1 and 2 (Baltimore Gas and Electric Company)
- 3. Fort Calhoun Station (Omaha Public Power District)
- 4. Millstone - 2 (Northeast Utilities Power Service Company)
- 6. Palisades Plant (Consumers Power Company)
- 7. St. Lucie 1 and 2 (Florida Power and Light Company)
- 8. San Onofre Nuclear Generating Station Units 2 and 3 (Southern California rdi.oni CoImiptny)
- 9. Watertird S;t*in 1:I :ri S ni P and I..i ht y
- 10. Palo, Verde Units 1, 2 and 3 (Ariz ona Publ Ic Service Cmpany
3.0. INTRODUCTION This report was prepared for the C-E Owners Group for their use in responding to Item II.K.2.13 for NUREG-0737. (3-1) Each of the participating utilities is represented by a plant specific Appendix to this report.
The basic elements of the C-E Owners Group program for consideration of the pressurized thermal shock question were presented at meetings with the NRC Staff on July 30, 1981 and October 7, 1981 in Bethesda. PTS questions beyond the definition of II.K.2.13 will be considered separately.
3.1 Background Action Item II.K.2.13 of NUREG-0737, Reference (3-1), requires that "a detailed analysis shall be performed on the thermal-mechanical conditions in the reactor vessel during recovery from small breaks with an extended loss of all feedwater". The requirement "deals with the potential for thermal shock of reactor vessels resulting from cold safety injection flow". "In 'particular, demonstration shall be provided that sufficient mixing would occur of the cold high pressure injection water with reactor coolant s.o that significant thermal shock effects to the vessel are precluded."
In evaluating thermal shock effects, it is recognized that the contribution of pressure is also very important. The subject has thus come to be called Pressurized Thermal Shock, or PTS. Consideration of PTS effects in general is not new, and has always been included in the design basis for PWR vessels. What is new is an increased awareness of the frequency of cooldown transients due to several occurrences, and an awareness of the sensitivity of PTS to overly conservative assumptions, particularly with respect to mixing of the high pressure safety injection (HPSI) water.
Previous licensing evaluations of PTS effects generally assumed complete mixing of the HPSI water with the system water in the vessel annulus. This report includes justification of the amount of heatup of the HPSI water calculated during the analyses.
3-1
3.2 Format of Report The body of this report describes the methods used to perform the plant specific analyses reported in separate Appendixes for each plant. The thermal-hydraulic system transient analyses were performed on a composite reference plant basis, and the results are included in Section 4 of the report. Section 5 describes the methods used to establish the axial, azimuthal and thru-wall fluence magnitudes and spatial distributions for each plant. Section 6 describes the methods used to determine the material property changes due to irradiation for each plant. Section 7 describes the fracture mechanics methods and results for the reference plant transients. The comparison of the fracture mechanics results to material propertylimits are reported in each plant specific Appendix.
3.3 Discussion' 3.3.1 Thermal-Hydraulic Transients Action Item II.K.2.13 of NUREG-0737 requires evaluation of the pressurized thermal shock effects of small break loss-of-coolant-accident (SBLOCA) with extended loss of feedwater (LOFW). This is understood to be a loss of all feedwater, both main and auxiliary feedwater. In order to create a challenging PTS transient for the C-E NSSS design, several very conservative simultaneous assumptions are required. The most basic conservative assumption is that of a simultaneous SBLOCA and extended LOFW. The C-E NSSS is designod so that there are no operational transients which result in PORV operat ion, which might be the 'source of a SBLOCA. In the analyses reported here, simultaneous independent occurrences of complete loss of feedwater and a small break LOCA were assumed.
The C-E NSSS is designed with a reactor trip signal on low steam generator water level, with the setpoint chosen to provide sufficient steam generator inventory to ensure 15 to 30 minutes before dryout in the absence of auxiliary feedwater. This is ample time to ensure initiation of auxiliary feedwater prior to dryout, ,either by automatic systen response or manual action.
3-2
During recovery from a LOFW situation the operator is given several precautions (Reference 3-2) to avoid overcooling, including:
- 1) reestablish feedwater preferentially to the steam generator containing water if another steam generator is dry, or feed only one generator if both are dry.
- 2) avoid exceeding the Tech. Spec. cooldown rate limit based on continuous monitoring of primary temperature and pressure.
- 3) do not place auxiliary feedwater in the manual feed mode unless necessary due to a failure of the automatic system, and check the system parameters frequently if in the manual mode to prevent overfeed.
The first type of SBLOCA + LOFW cases evaluated in this report assume total loss of all feedwater without initiation of auxiliary feedwater, making use of PORV's to provide a means for removal of the decay heat. As discussed above, the large steam generator secondary side inventory provides ample time to initiate auxiliary feedwater for plant cooldown by use of the steam generator.
The second type of SBLOCA - LOFW cases evaluated in this report assume
- 1) recovery of auxiliary feedwater after los$ of feedwater has persisted long enough to allow dryout of both steam generators, and
- 2) refill of the generators at maximum auxiliary feed pump capacity in order to conservatively maximize the cooldown effect.
As discussed above, the large steam generator inventory would provide ample time to initiate feedwater prior to dryout, and operating instructions specifically direct the operation of auxiliary feedwater in a manner to minimize the cooldown effect.
There are, therefore several very conservative assumptions incorporIted in theseanalyses, nevertheless, this report concludes that all C-E NSSS reactor vessels can withstand the resultant PTS effects without crack initiation for assumed plant life. If this were not the case, additional conservatism could be eliminated as necessary to produce a more realistic set of assumptions for the scenarios which are beyond the plant design basis, in order to justify additional plant capability.
3-3
In choosing the small break LOCA's to be considered by this program, It was recognized that the concern involves the possibility of repressurization following or during a thermal shock condition.
Accordingly, break sizes considered were at the small end of the spectrum. Break sizes large enough to ensdre no repressurization might still be small enough to fit the licensing definition of a "small break" rather than a "large break" LOCA, but would not be of concern from a pressurized thermal shock standpoint. Pressure vessel integrity for LOCA break sizes large enough to ensure no repressurization has been previously evaluated in large break LOCA analyses.
The thermal-hydraulic system analyses were performed on a reference plant with parameter variations as necessary to cover the ranges for different plants. The Fort Calhoun, Maine Yankee, Millstone 2 and St. Lucie 1 plants contain thermal shields in the annulus between the core support barrel and the reactor vessel. The thermal shields typically extend from slightly above the top of the active core region down to the bottom of the core. Therefore,,fluid flow and mixing in the annulus region above a thermal shield would be relatively unaffected by the presence of a thermal shield. The presence of a thermal shield was accounted for in the fluence distribution calculations reported here in.
Loss of natural circulation was observed in the analyses of some of the cases with assumed complete loss of feedwater with no recovery of auxiliary feedwater during the time span of the analysis. The C-E NSSS is designed to ensure 15 to 30 minutes prior to complete steam generator dryout Lifter loss of feedwater. The emergency procedure strategy and training direct the operator to effect a system cooldown by means of the secondary side heat sink. Therefore, the complete and continued loss of all feedwater is an unrealistic conservatism assumed simply for the sake of this analysis.
Good mixing/heatup of the HPSI water was calculated during the loss of natural circulation associated with the scenarios evaluated, as discussed in more detail in Chapter 5.
3-4
3.3.2 Vessel Fluence The basic method used to establish fluence values was to calculate normalized fluence profiles and then scale the magnitude of the peak of the profile to the peak of the available plant specific surveillance caosule results.
Some plants used variations of this approach, depending on plant specific data availability. Fluence profiles were developed in this manner for the Fort Calhoun, Calvert Cliffs 1 & 2, Maine Yankee, Palisades and St. Lucie 1 vessels which contain more than 0.15% copper in their welds. The ANO-2, St. Lucie 2, San Onofre 2 & 3, Waterford 3 and Palo Verde 1, 2, & 3 vessels beltlines all contain less than 0.10% copper welds. These low copper vessels were evaluated using peak fluence values assumed everywhere.
The basic unit of measure used to define the capability of each vessel is the number of ,effective full-power years (EFPY) of operation during which the vessel could withstand the specified PTS conditions without violation of the pressure boundary. The neutron fluence per EFPY for each vessel is based on the cumulative average value determined from surveillance capsule evaluations. Assumed end-of-life fluence values are based on a linear extrapolation of fluence per EFPY. This assumes that plant operation and fuel'parameters result in the same cumulative average fluence per EFPY during future plant operation. This assumption is conservative for plants in which low-leakage fuel management is presently being used or is being planned in future fuel cycles. Based on a conserva:
tive performance target of .80 EFPY per calendar year, 32 EFPY is assumed to represent end-of-life conditions.
3.3.3 Vessel Materials Plant specific vessel material properties "maps" were developed for each vessel containing more than 0,15% wt. copper welds, for superposition with the fluence distributions. Vessels containing low copper welds were treated on the basis of peak fluence assumed to coincide with most sensitive beltline matorial, regardless of loation. in the vessel.
3.3.4 Mechanical Evaluations of Vessels Each vessel was evaluated on the basis of the results of generic stress calculations of the reference plant temperature-pressure transients.
Linear elastic fracture mechanics (LEFM) analysis methods were used.
3-5
The results of theseLEFM analyses for a range of assumed initial crack depths were compared to the plant specific material toughness properties for a range of different plant life conditions.
The mechanical evaluations of the accident cases prescribed by fl.K.2.13 were performed based on an acceptance criterion of no pressure boundary violation, defined to be crack initiation with subsequent arrest. A conservative upper-shelf toughness of 200 ksi 1/TE was used as the applica bility of LEFM for prediction of crack arrest. This typically resulted in a crack depth limit around one-half the wall thickness. Because of the characteristics of the pressure-temperature transients, no crack initiation was calculated.
The evaluations reported herein take credit for warm prestressing where applicable.' Warm prestressing occurs when the stress intensity K is decreasing from a condition where the <material was at a previously more ductile (warmer) state. This ,may occur during a transient when the thermal gradient is decreasing with time as the metal begins to approach the fluid temperature through the entire wall thickness.
If,while the stress intensity is decreasing, K exL.eeds the critical crack initlation stress intensity KIC then the warm prestressing aspect of this unloading sequence will prevent crack initiation.
For a situation of K1 greater than KIC, the normal cooldown minimum-pressure temperature (MPT) limits would also be violated. Guidance is provided in the C-E LOCA Guidelines,(33) and CEN-128 (34) cautioning the operator to prevent violation of the MPT limits. Since the conditions existing for a situation where warm prestress prevents crack initiation would be outside the MPT limits, the operator would be guided to bring the plant conditions within limits, primarily through depressurization to shutdown cooling conditions.
The most severe pressure-temperature transient developed for the range of parameters representing the operating plants was used to evaluate the low copper vessels in the nonoperating plants. The evaluation showed the low copper vessels to be very insensitive to pressurized thermal shock effects, even for an assumed end-of-life fluence value twice the FSAR prediction value.
3-6
3.3.4 Results Results of the composite reference plant thermal-hydraulic system transient analyses are provided in the body of this report. Also included in the body of the report are the results of linear elastic fracture mechanics evaluations of the most limiting transients.
The plant specific fluence distributions, material properties and vessel integrity evaluations are provided in separate appendices for each plant.
Each plant specific appendix concludes that the reactor vessel can..safely withstand a small break LOCA with extended LOFW for the assumed design life of the plant.without crack initiation.
3-7
3.4 REFERENCES
3-1 NUREG-0737, "Clarification to the Action Plan," October, 1980.
3-2 CEN-1,52,, "Combustion Engineering Emergency Procedure Guidelines," June, 1981.
3-3 C-E LOCA Guideline, submitted to NRC by letter November 8, 1979.
3-4 CEN-128, "Response of C-E NSSS to Transients and Accidents,"
April, 1980.
CHAPTER 4 THERMAL-HYDRAULIC ANALYSIS
4.1 INTRODUCTION
This chapter describes the thermal-hydraulic analysis performed as part of.the effort requested in Section II.K.2.13 of NUREG-0737( 4 -1 ). This NRC action item requires that all PWR operating reactors and applicants for a license complete a detailed analysis "of the thermal-mechanical conditions in the reactor vessel during recovery from small breaks with extended loss of all feedwater".
The thermal-hydraulic conditions within the reactor vessel during a small break LOCA (SBLOCA) combined with a total loss of feedwater (LOFW) event are generally as follows. Initially, the reactor coolant system (RCS) pressure will decrease as a result of the small break where coolant leaves the primary system and voids develop. The rate of pressure decrease and fluid mass loss depends primarily on the break,size. If the system depressurizes sufficiently, cold water from the high pressure safety injection (HPSI) system combined with charging pump flow will enter the system in the cold legs and then flow into the vessel downcomer. After some time following the LOFW, the steam generators will no longer remove a significant portion of the primary system energy and the primary pressure may then increase. Additionally, the :system repressurization could also result from the primary system refilling. The system will even'tually 'ref'll if HPSI and charging pump flow exceeds the break flow for a sufficient time period, assuming no operator action to throttle the flow. The II.K.2.13 pressurized thermal shock (PTS) concern deals with the possibility that the cold safety injection fluid may overcool and thermally stress the inside surface of. the reactor vessel wall. The thermal stresses, combined with the resultant stresses from the repressurization, may then initiate and propagate cracks in the vessel, if the vessel material properties have been sufficiently degraded by long-term irradiation.
4-1
The fluid conditions in the cold legs and downcomer are an important aspect for the II.K.2.13 analysis. Early in the SBLOCA and LQFW transients, the fluid flow in the cold legs will decrease to natural circulation after the coolant pumps are tripped. Cold water from the HPSI system enters the primary system at rates which vary depending on the number of pumps operating and on the system pressure. This cold water, combined with charging pump water, will flow along with the warm system fluid and enter the downcomer. The degree of mixing of the cold and warm water is calculated in this analysis and the cooler layer of water is considered to travel along the reactor vessel wall. The PTS concern must determine how this HPSI flow mixes with the hot system fluid before passing along the reactor vessel wall.
Natural circulation of fluid in the cold legs has been calculated to stop after depletion of the steam generator secondary side in some of the transients which were evaluated. The net flow in the cold legs then essentially becomes zero for the rest of the transient. In this condition, the cold leg and downcomer
. fluid temperatures will eventually decrease due to injection of cold HPSI water. The .PTS concern must deal with the rate at which this water cools in both the cold leg and downcomer.
The break location is important for establishing the timing when natural circulation is broken. The small breaks for the analysis of II.K.2.13 was located in the pressurizer. 'Single-phase natural circulation is broken after sufficient fluid leaves the system and fluid no longer passes over the top of the steam generator tubes. The flow into the cold legs then stops. The fluid in the cold leg and annulus will remain liquid because it is subcooled by HPSI water. Similar conditions would be expected for a break located in the hot leg. In comparison, a cold leg break would promote loop flow into the cold legs for-a longer time because the majority of the system inventory follows the natural circulation flow path towards the break.
4-2
The break location is also important in how natural circulation ceases.
For this analysis, natural circulation ceases first in the cold legs of the unbroken loop (the loop with no pressurizer) due to faster depletion of the secondary side inventory of this steam generator. The faster inventory depletion is due to the fact that some of the flow leaving the core and going into the broken loop goes out the break instead of being transported to the steam generator of the broken loop. Therefore, the heat removal load on the steam generator in the broken loop Is loss than the steam gonerator In the intact loop (s). This causes the secondary side inventory of the steam generator in the intact loop (s) to deplete faster than in the steam generator in the broken loop. Consequently, natural circulation stops earlier in the unbroken loop. Thus, there is a condition that can be established wherein one loop (the broken loop) still has natural circulation while in the other loop(s) natural circulation flow has ceased. Eventually, all loops will cease natural circulation.
Natural circulation stopped for two transients of this analysis. For these cases, the cold leg and downcomer fluid temperatures were calculated to remain relatively warm for the time period of concern for PTS. The warm temperatures were due to the occurrence of a recirculation pattern that was calculated between the cold leg and annulus fluid. The recirculation pattern resulted from the interchange of cooler, more dense, cold leg fluid with the warmer, less dense annulus water. The annulus remained relatively warm due to the natural circulation still occurring in the broken loop.
The selection of important input parameters and the development of a composite plant model for the evaluation of II.K.2.13 is provided in Section 4.2. The composite plant was designed to conservatively envelope all C-E operating plants and to be adequately representative of all plants listed in Section 2 for the PTS thermal-hydrauli'c analysis. As a special case, the Maine Yankee plant was evaluated with the composite plant model utilizing conservative high-head HPSI pump injection specifications.
4-3
The analysis presented in this chapter is divided into: (1) an NSSS system or bulk thermal-hydraulic analysis and (2) a reactor coolant mixing analysis of
- HPSI water in the cold leg and vessel annulus. These analysis methods are described in Section 4.3. The system response analysis for II.K.2.13 was conducted with the CEFLASH-4AS evaluation model assuming complete mixing of the HPSI and system fluid. The CEFLASH-4AS code has been evaluated with small break LOFT Tests L3-1 and L3-6( 4 -2 ), conducted by EG&G, Idaho, Inc.,
at the Idaho Engineering National Laboratory. In parallel with the CEFLASH-4AS analysis, a method of cal,,ulating the local mixing of HPSI and RCS water in the cold legs and vessel annulus was developed.
The SBLOCA and LOFW scenarios for II.K.2.13 have been subdivided into two categories of assumed operator action. The operator actions were chosen in order to avoid core uncovery for the II.K.2.13 scenarios. In the first category, two PORV's were assumed to be opened at 10 minutes to prevent core uncovery. The opening of the PORV's iqcreases the break size and results in a cooldown by the cold safety injection. Assuming the PORV's were not opened at 10 minutes, the second category considered was the restoration-of feedwater at 30 minutes. For this thermal-hydraulic analysis, restoration of feedwater means theflow of auxiliary feedwater to the steam generators at an injection temperature of 40 F. A range of break locations, break sizes, and high pressure safety injection flow rates were evaluated for both categories. A complete description of the II.K.2.13 scenarios is provided in Section 4.4.,
Thermal-hydraulic results and conclusions are provided in Sections 4.5 and 4.6. The thermal-hydraulic results consist of reactor vessel pressures, and cold leg and downcomer fluid temperatures that were calculated for a set of eight SBLOCA andLOFW transients. The downcomer fluid temperatures were calculated next to the vessel wall at three axial locations. This information was then used to evaluate reactor vessel stresses and to determine PTS conditions for the II.K.2.13 transients.
4-4
4.2 PLANT SELECTION The thermal-hydraulic analysis presented herein is intended to apply conservatively to all Combustion Engineering operating nuclear power plants and to be adequately representative of all C-E plants listed in Section 2. In order to perform the analysis on a generic basis, the important parameters from each plant were compared. Tables 4-1 and 4-2 present plant parameters for the purpose of comparison.
Two plants were selected to represent all of the C-E PWRs; a composite" plant based on, Calvert Cliffs 1, and a composite plant based on Calvert Cliffs 1, but using high-head, high pressure safety injection pumps as in the Maine Yankee plant. Each is discussed in detail in the following subsections.
4.2.1 Composite Plant The Calvert Cliffs Plant was selected to represent the C-E plants for the PTS analysis of II.K.2.13. This plant was also used in a previous evaluation( 4 -3 of all C-E operating plants for justifying the C-E small break LOCA ECCS evaluation model. It is expected that the general behavior of the composite plant is representative of all C-E plants for this analysis.
A comparison of important plant parameters for Calvert Cliffs with the other plants is provided in Tables 4-1 and 4-2. The ratio of power level
.to system volume as well as MSSV and PORV setpoints are similar for Calvert Cliffs compared to the other plants. Similarity in these parameters yields similar thermal hydraulic response following a small break LOCA.
The analysis results are representative of both two-and three-loop designs because the important characteristics of the loop configuration, i.e.,
component relative elevations, are similar in both designs. For LOCA analysis, the most significant diffeences between these plants are the flow rates and shutoff heads of the HPSI pumps. Therefore, Calvert Cliffs.was selected to represent the operating plants with modified HPSI pump injection curves that encompass the pump injection range of operating C-E plants. As a special case, the high 4-5
head HPSI pump., conservatively representing JMairne Yankee, was separately analyzed. The HPSI pump characteristics of the other (3400 MWt and System 80) plants would produce a system thermal-hydraulic response that is adequately represented by these two cases.
4.2.1.1 HPSI Pump Flow For a small break LOCA with concurrent loss-of-feedwater, the loss of steam generator heat transfer may cause the system to repressurize with minimum HPSI flow. With maximum HPSI flow, the repressurization may be reduced or even avoided. However, the maximum HPSI flow may cause more severe cooling of the vessel wall. To investigate the effect of HPSI flow rate on the thermal shock analysis, both minimum and maximum HPSI flows were considered. The range of HPSI flow rates for the C-E operating plants is illustrated in Table 4-2 and Figure 4-1. Of the other C-E plants, not shown in Figure 4-1, the HPSI pump delivery curves for St. Lucie 2, Waterford, and San Onofre are encompassed within the maximum and minimum pump delivery curves of the operating plants.
The HPSI delivery curve for System 80 plants can be represented by the delivery curves for the operating plants since the ratio of HPSI flow-to
- volume of System 80 plants is less than the ratio of HPSI flow-to-system volume for the composite plant of this analysis.' The maximum total flow rates occur for Maine Yankee, ANO-2, and St. Lucie 1. Table 4.1 and Figure 4-1 also show that the minimum flow occurs for one Ft. Calhoun and one Millstone 2 HPSI pump. This represents the range of HPSI flows considered in this analysis. These maximum and minimum flow rates include uncertainties and are not simply a function of the number of pumps injecting (e.g. in Table 2-4 the difference between maximum and minimum HPSI flow for Ft. Calhoun is more than the difference in flow between three pumps and one pump). For maximum HPSI flow a composite delivery curve was synthesized from the three St. Lucie 1 pumps with the ANO-2 HPSI shutoff head.' One Ft. Calhoun HPSI pump, with the Millstone 2 shutoff head, was used for the minimum HPSI flow analyses. The composite maximum and minimum HPSI delivery curves are illustrated in Figure 4-2.
With a loss of offsite power and the failure of.one diesel generator to start, only one HPSI pump would be able to start injecting. However, the number of charging pumps operating could be from one to three depending on how the charging pumps were connected to the diesel generators for a particular plant.
Therefore, in order to maximize the thermal shock at high pressures three charging pumps were assumed operating even in the minimum HPSI flow analyses.
The total HPSI flow curves (HPSI plus charging) are shown in Figure 4-2.
All of the charging pumps were assumed to be actuated by pressurizer low level, which in every transient occurred before SIAS. Therefore, cold injection water starts to enter the system before the pressure falls to the SIAS setpoint.
4.2.1.2 Additional Plant Characteristics The MSSV setpoint largely determines the pressure to which the primary system will fall during the initial period of a small break transient. Because of the similarity of setpoint values, a minimum setpoint pressure of 1000 psia was selected.
The PORV flow areas are also similar for all the plants, which have PORVs so the Calvert Cliffs effective PORV flow area value of 0.0075 ft2 /valve was selected. The effective area being the flow area calculated from the manufacturer's rated flow and the CEFLASH-4AS critical flow correlation.
ANO-2, Waterford, San Onofre 2 and 3, and Palo Verde 1, 2, and 3 do not have PORVs. ANO-2 has two motor operated vent valves on its pres'surizer. Thes6 manually operated valves, mounted in series, have an effective flow area of nearly 0.05 ft2 . As described in Reference 4.3, breaks larger than 0.02 ft2 do not require the availability of feedwater to avoid repressurization.
The full opening of these ANO-2 valves would result in a rapid depressurization of the'system. With such a large break area therg would be no repressurization either by loss of steam generator heat transfer or by HPSI packing at any 4-7
Pressure of interest in a PTS evaluation. Therefore, use of the smaller Calvert Cliffs PORV area is conservative for ANO-2.
Reactor trip was assumed to occur on low steam generator water level. This occurred before low pressurizer pressure indicated SIAS. The effect on the analysis of earlier trip was to reduce the amount of heat added to the system, which is conservative for PTS considerations.
Offsite power was assumed to be lost simultaneously with reactor trip, resulting in main coolant pump coast down. In addition, the turbine admission valves and steam bypass valves are assumed to close.
In those cases where initiation of auxiliary feedwater (AFW) was assumed, the temperature of the AFW was considered to be 40oF. The AFW was also assumed to inject at the maximum rate, filling the steam generator secondary sides with subcooled water. Level control for AFW is automatic for all the plants, but was not considered in the analysis except for the high-head HPSI plant. The subcooling of the secondary sides introduces an added conservatism due to the additional cooling of the natural circulation flow.
In conclusion, the Composite plant selected for this analysis is essentially Calvert Cliffs Unit 1, with composite HPSI delivery curves selected to encompass the other operating plants and be adequately representative of other plants listed in Section 2. The basic assumptions for this analysis of the Composite plant can be summarized as follows:
- 1. 100% of nominal power (2700 MWt)
- 2. Reactor trip on low secondary level (17.0 seconds in every case)
- 3. Loss-of-Offsite power on trip .
- 4. Maximum (and minimum) HPSI for all the operating plants
- 5. HPSI and charging pump water at 40 degrees Fahrenheit
- 6. MSSV setpoint at 1000 psia
- 7. PORV flow area of 0.0075 ft2/valve
- 8. AFW temperature of 40 degrees Fahrenheit
4.2.2 High Head HPSI Pump Plant
- Maine Yankee has HPSI pump characteristics sufficiently different from the other plants to require special consideration. The HPSI pumps with their high shut-off head can inject more cold water at higher pressures. Therefore, in addition to the analyses performed for the composite plant, an analysis was also performed with HPSI pump characterisics which conservatively envelope the Maine Yankee capacity. The maximum HPSI flow delivery curve for two pumps illustrated in Figure 4-2 was used. The HPSI pumps are also used as charging pumps and were assumed to be actuated on pressurizer low level.
Maine Yankee also has greater AFW capacity than the other plants. In the analysis for Maine Yankee, the AFW was assumed to be controlled once normal operating water level was reached. This assumption was not made in the composite plant analysis.
All other basic analysis assumptions were as described for the composite plant in Section 4.2.1.
4.3 ANALYTICAL METHODS The methods used for the analysis of the thermal-hydraulic transients are divided into categories of small break thermal-hydraulics and coolant mixing.
The major system parameters (e.g. system pressure, loop flow, etc.) were calculated with a modified version of the CE Small Break LOCA Evaluation Model assuming complete mixing of the HPSI and cold leg water. More detailed calculations of the temperature of the water in contact with the vessel wall were made with a C-E developed mixing model. The modifications made to the SBLOCA Evaluation Model for this PTS evaluation, and the detailed fluid mixing model are described in the following sections.
4-9
4.3.1 Small Break Model The C-E small break LOCA evaluation model used for licensing calculations is described in References 4-4 and 4-5. The SBLOCA transient calculation is generally divided between the calculation of the system thermal-hydraulics and the core temperature calculation. Core temperature calculations were not performed in this analysis, however, because the core remained covered in every case.
4.3.1.1 Input Assumptios The basic assumptions of the evaluation model are conservative for calculating core cooling, but not necessarily conservative for calculating the possibility of thermal shock. Therefore, certain assumptions had to be modified for this PTS analysis. The basic licensing assumptions for analysis of core cooling include:
- 1. Reactor trip on low pressure
- 2. Loss of offsite power on trip
- 3. SIAS on low pressurizer pressure
- 4. Minimum safety injection flow
- 5. Maximum wall heat
- 6. 120% ANS decay heat rates
- 7. No operator action during the first hour
- 8. Maximum safety injection water source temperature
- 9. AFW initially actuated on low secondary level As will be shown below all these assumptions, except Nos. 2 and 3, were modified for the present analysis.
The accidents considered as part of this analysis assumed a complete loss of all feedwater concurrent with a small break LOCA. Therefore, the reactor trips on low secondary levAl, which proceeids the occurrence of low RCS pressure ard 4-10
initiates SIAS. This trip reduces the amount of heat added to the primary water thereby producing a conservative cooldown transient for PTS considerations.
Loss of offsite power on trip was still assumed, as well as main steam isolation valve closure and coast down of the main coolant pumps. The coast down of the main coolant pumps served to reduce the amount of safety injection mixing in the cold legs.
SIAS was still considered to occur on low pressurizer pressure, but the charging pumps were considered actuated on pressurizer low level which occurs before SIAS. This is a realistic representation of the early occurrence of cold water injection by the charging pumps.
Minimum safety injettiop flow is obviously conservative for a core cooling evaluation, but not necessarily so for pressurized thermal shock. Therefore, different amounts of injection flow were considered. In addition, the injection water temperature was taken at its minimum value.
The wall heat model used in.the core cooling licensing evaluation over estimates the amount of heating of the primary water. The wall heat model for the cold legs and downcomer was improved for this evaluation and a more realistic calculation with a lower heat rate was made.
For this evaluation, the 20% uncertainty on the ANS decay heat was removed in order to do a more realistic analysis and thereby conservatively reduce the amount of heat-up of.the system water.
In a licensing analysis no operator action is considered for the first hour after the accident, but in this study earlier operator action to prevent core uncovery is assumed. Opening PORVs at 10 minutes after a LOFW or restoring auxiliary feedwater at 30 minutes were the assumed operator actions based on previously reported studies (Reference 4-3).
4-11
The major differences between SBLOCA evaluation calculations and the SBLOCA calculations performed for the present PTS study can be summarized as:
- 1. Reactor trip on low steam generator level
- 2. 100% ANS decay heat
- 3. Maximum and minimum high pressure safety injection
- 4. Realistic wall heat
- 5. Operator action to prevent core uncovery
- 6. Minimum safety injection water temperature
- 7. Total loss of all feedwater 4.3.1.2 CEFLASH-4AS Code The reactor vessel pressure transients and the bulk fluid temperature in the cold legs (assuming 100% mixing) were calculated by CEFLASH-4AS (References 4-6 and 4-7). The CEFLASH-4AS code is a node and flowpath model of a PWR used to calculate the thermal-hydraulic response to a small break LOCA. It assumes thermodynamic equilibrium and calculates phase separation in all the control volumes and can calculate hot leg counter-current flow during a reflux boiling mode. In normal licensing calculations, safety injection is assumed to enter the lower downcomer with perfect mixing.
Verifications have been made of the CEFLASH-4AS code, most notably with the series of LOFT tests. The best estimate calculations of the L3-1 and L3-6 tests showed the ability of the CEFLASH-4AS code to model the response of a PWR to a small break LOCA. For the LOFT tests, and for standard licensing calculations, the objective is to model the general system parameters and concentrate on the core hydraulic conditions for calculations of cladding temperatures. However, for a PTS evaluation more detailed calculation of cold leg and downcomer conditions are needed for input to the mixing model. CEFLASH 4AS is acceptable for PTS thermal-hydraulic calculations providing several modifications are made to more realistically represent local fluid conditions in the vessel downcomer.
4-12
The licensing assumption of safety injection into thelower downcomer is not bulk conservative for a PTS thermal-hydraulic analysis; it over-estimates the fluid temperature in the cold legs and upper downcomer. Therefore, CEFLASH-4AS was modified to inject HPSI water directly into the cold leg. Two flow paths between the cold leg and annulus were also modeled representing the lower and upper halfs of the cold legs. This modeling provides the ability to calculate a recirculation pattern for mixing of the cold leg.and annulus fluid when natural circulation stops. FYr this analysis, the cold leg and annulus were full of subcooled water. The recirculation pattern results from the interchange of cooler, more dense, cold leg fluid with the warmer, less dense, annulus water.
The licensing evaluation model over-estimates the energy added to the primary fluid by the sensible heat of the system metal. A more detailed and realistic calculation of wall heat in the cold legs and downcomer has been made by CEFLASH-4AS for this analysis. The changes made to the CEFLASH-4AS code are:
- 1. HPSI injection directly into cold legs
- 2. Counter-current flow in cold legs
- 3. Realistic wall heat model With these modifications to both the small break evaluation model and the CEFLASH-4AS code, the pressure and local bulk fluid temperature are predicted for the PTS evaluation. More detailed calculations of downcomer fluid temperatures were made with a separate mixing model using the general cold leg and downcomer conditions calculated by CEFLASH-4AS. This mixing model is described in Section 4.3.2.
4.3.2 Fluid Mixing Analysis The degree of mixing of the cold HPSI water and warm system fluid is an important parameter for the SBLOCA and LOFW transients. A mixing model was developed to evaluate fluid mixing and to calculate fluid temperature distributions next to the reactor vessel wall in the downcomer. A description of the fluid mixing analytical method is provided in this section.
4-13
4.3.2.1 Description of Mixing Conditions, The mixing of cold injected water with the fluid flowing in the cold legs is dominated by different mechanisms when the fluid flowing in the cold legs is hot water as opposed 1o a two-phase mixture of steam and water. Since none of the transients analyzed in this report resulted in HPSI injection into a two phase or steam filled pipe, this section will only discuss the mixing of the HPSI fluid with a flowing stream of hot water. Fluid mixing for this analysis is calculated as follows. The developmental version of C-E's small break LOCA code (CEFLASH-4AS) is used to determine the overall system response. An auxiliary code (MIXUP) is then used to calculate the mixing of the cold HPSI injection water with the loop fluid in the cold leg and annulus resulting in the local fluid temperatures at the reactor.vessel wall. MIXUP uses several of the system parameters (e.g. loop flow, loop and annulus temperature) calculated by CEFLASH-4AS as inputs. The following subsections discuss the mixing model of MIXUP and present a comparison of MIXUP predictions to experimental data.
O4.3.2.2 Mixing Tests Scoping tests performed at Creare( 4 -) have shown that, when cold water Is injected into the bottom of a horizontal pipe/angulus geometry containing a slowly moving or stagnant stream of hot water, the cold (more dense) water will flow along the bottom of the pipe and spill into the annulus as shown in Figure 4-3. This stratified behavior is caused by the different densities of the two fluids. Although stratified flows are not commonly analyzed in PWR thermal/hydraulics, they are commonly studied in the areas of meteorology and oceanography. Turner(4-9) describes several commonly observed natural phenomena involving mixing due to density differences that are similar to the situation at hand.
Tests have been performe'd in the laboratory to investigate the mixing processes that occur in nature. Tests, applicable to the HPSI mixing situatIon were 4-14
performed and analyzed by Ellison and Turner . Figure 4-4 shows a schematic of their test facility. This facility, which was designed to investigate inclined plumes, consisted of a 200 x 60 x 10 cm plexiglass channel, mounted with the largest faces vertical. The channel could be tilted to any angle between horizontal and vertical. The channel was filled with fresh water, and a salt solution was supplied to an opening at the upper end of the bottom of the channel. The solution flowed in a 10 cm wide stream down the full length and was removed, together with the fresh water that it had entrained, from the bottom corner. In order to maintain a steady state, fresh water had to be added to make up for that lost by entrainment. The fresh water was supplied at the top of the channel through the long cloth cylinder to ensure a "gentle" supply of fresh water. Measurements of the salinity distribution and velocity profile were made along the channel using sampling probes and thermistors. Observations were made by injecting dye in salt solution.
At low flow rates and small slopes the plume interface was very smooth and little mixing was 'observed. .As the flow rate was increased, waves appeared on the interface which eve'ntually entrained the lighter fluid. At higher flow rates and greater slopes, the entire plume became turbulent and its thickness increased as it entrained fresh water. In other experiments, dye was added to the fresh water supply. The dyed fresh water seeped out through the cloth cylinder and flowed slowly across the channel perpendicular to the plume at all channel slopes. At the edge of the plume the fresh water was trapped by the large eddies and, as it became part of the plume, traveled at right angles to its previous direction. None of the fluid in the plume escaped; transfer was always into the turbulent plume.
By measuring the velocity of the plume at several axial locations and the flow rate of the salt solution at the bottom of the channel and comparing it to the input, Ellison and Turner were able to correlate the rate of entrainment to the layer mean velocity.
The basis of the'model used to predict the rate of entrainment, first established by Morton, Taylor and Turner( 4 -11 ), is that the mass flux of the 4-15
fluid being entrained is proportional to the average mass flux of the turbulent layer. The proportionality constant is called the entrainment constant (E). As mentioned above, Ellison and Turner derived the entrainment constant as a function of channel slope from their experiments. Figure 4-5 shows the results of this work. The entrainment factor was found to be very low.at small slopes, suggesting little entrainment or mixing, and to increase to a maximum when the channel was vertical. For a vertical channel the equations of mass, momentum, energy and state governing the flow of the turbulent layer are:
- U rt~il 4-1
- 7) -~'1 43 where:
f = fluid density u = fluid velocity h = layer thickness T - fluid temperature g
- gravitational acceleration E3 entrainment constant in region 3 z length (along direction of flow) 4-16
subscripts:
o = designates the lighter fluid rel = designates relative velocity between two fluids The results of Ellison and Turner suggest that, for the C-E configuration, significant mixing between the cold HPSI water and the warmer water flowing in the reactor cooling system will occur at the HPSI injection location and in the reactor vessel annulus. At these two locations, the cold water flows vertically and the entrainment mechanism is expectedito be most effective. The C-E model, therefore, analyzes the mixing processes in these two regions.
Additional mixing in the horizontal section of the cold leg is neglected.
4.3.2.3. C-E Mixing Model Figure 4-6 shows a schematic of the C-E model. Region 1 is the HPSI injection region where a falling cylindrical plume of cold HPS-I water is assumed to exist.
This falling plume is expected to be carried toward the annulus when the flow 3 in the cold leg is in that direction. C-E minimizes the amount of injection in this region by assuming that the plume falls vertically to the bottom of the pipe. In region 2, the HPSI water flows horizontally in a stratified layer along the bottom of the cold leg pipe. No mixing is assumed in this region.
In region 3, the HPSI water falls vertically as a plume. Mixing between the cold HPSI water and the hot water flowing in the cold leg and annulus is calculated in regions 1 and 3.
As shown below, the governing equations for region 1.are slightly different than equations 4-1, 2/, & 3 because of the cylindrical geometry.
-i-.-~
r .Z - r7 45 4-17
- (#eA7(~j t-/ Jr4-6 where:
= fluid density u = fluid velocity in the plume r = plume radius T = fluid temperature g = gravitational acceleration E= entrainment constant in region 1 z = distance measured vertically from the ECC injection nozzle exit
.o subscripts:
= designates lighter fluid The entrainment constant (El) for a cylindrical plume has been determined experimentally( 4 -9to have a value between 0.08 and 0.10. The C-E model uses a value of 0.08, which results in the least mixing.
Region 2 consists of a horizontal stratified flow in the cold leg pipe. No mixing is assumed to exist in this region. In reality, some heat up of the cold injection water would occur in region 2 due to heat transfer from pipe metal and the warmer water in the layer above. There would also be some turbulent mixing in region 2 during periods of high loop flow. These heat up mechanisms have been neglected. The height of the fluid in the cold leg pipe, which is used to start the calculation in region 3, is calculated assuming that 4-18
the fluid flows at the critical depth (internal Froude number equal to one).
The basis for this assumption is that the height and velocity of the stratified layer will be such that critical flow is established where the cold water W spills into the annulus, as is the case for flow over weirs and spillways (4 12)'. The expressions for critical flow in a circular pipe (see Figure 4-7) are somewhat more complicated than for a one-dimensional channel, and require iterative solution techniques. These expressions are shown below.
4-8 where:
Q = volumetric flow rate g = gravitational acceleration R = radius of cold leg 9 = density of cold fluid in layer
= density of warm fluid above layer 9 - angle defined in Figure 4-7 y
- height of layer as shown in Figure 4-7 c
- width of layer as shown in Figure 4-7 u = velocity of fluid in layer 4-19
The density of the fluid in the layer is determined from the calculations performed in region 1.
I The thickness and velocity of the layer spilling into the annulus can be calculated knowing the height (y), the width (c), and the volumetric flow rate (Q) of the fluid just before It exits the cold leg. By assuming that the fluid at any elevation in the horizontal layer accelerates due to gravity until it reaches the elevation in the annulus corresponding to the lower lip of the cold leg, one can calculate the velocity profile in the layer in the vertical direction at the point where the fluid spills into the annulus. When it is assumed that the falling plume maintains the width (c), the thickness of the layer H as it spills into the annulus and its average downward velocity (U ) can be calculated from:
- - 2, L o) 4-12 4-13 Region 3 is the vessel annulus region where a falling plume of cold water is assumed to exist.. The governing equations for this geometry are equations 4-1, 2, & 3. The entrainment coefficient used in the analysis is taken from Figure 4-5 for a vertical slope (E = .086).
The fluid temperature in the falling plume, calculated by equations 4-1, 2, &
3, is the average fluid temperature in the plume. In reality there may be a temperature variation across the plume with the hottest fluid on the outside of the plume and the coldest fluid near the wall (ignoring wall heat transfer).
Although the shape of the temperature distribution was not determined in Ellison's experiments, the density distribution was measured and was found to have a Gaussian shape. In calculating the temperature of the fluid at the 4-20
reactor vessel wall it is assumed that the temperature distribution is also Gaussian with an average temperature equal to that calculated by equations 4-1, 2, & 3. This will result in an entrance effect. That is, the fluid adjacent to the wall must travel a certain distance down the wall before its temperature increases (gigure 4-8).
4.3.2.4 Comparison of Mixing Model Predictions to Creare Tests Creare, Inc. has performed mixing experiments for the two geometries shown in Figure 4-9. The tests were performed in 1/5-scale transparent facilities (4-8, 4-13). The tests performed with the facility having the horizontal cold leg (4-8) produced only visual data. In these tests, cold water was injected into the bottom of the cold leg. There was no simqlation of loop flow. Hot water was added in the upper annulus to simulate vent valve flow typical of B&W NSSS designs. As mentioned before, photographs of the tests indicated a very stable stratified flow in the horizontal cold leg with little apparent mixing due to the lack of waves at.the layer interface. The cold water spilled into the annulus as a plume which did not spread much wider than the diameter of the cold leg. The interface at the edge of the falling plume was very wavy which probably resulted in the entrainment of hot water intothe plume. No temperature measurements have been published.
The tests performed with the sloped cold leg (4-13) produced visual data and preliminary temperature data at the exit of the cold leg and along the simulated reactor vessel wall below the cold leg. Although there was no simulation of loop flow, hot make-up water was again added to the upper annulus to simulate vent v lve flow. To permit flow visualization, some tests were run with dye injected in the cold water while other tests were run with dye injected in the hot make-up water. When the dye was in the cold water, one could see a stratified flow along the bottom of the cold leg. In the sloped portion of the cold leg, the interface was very wavy and a decrease in color intensity was observed which suggests entrainment of the clear hot water. At the junction of the sloped and horizontal portions of the cold leg a very turbulent hydraulic jump was observed. The cold water then spilled into the 4-21
annulus as a plume which did not spread much wider than the diameter of the cold leg. The interface at the edge of the falling plume was very wavy as was*
observed in the tests with the horizontal cold leg.
When the dye was in the hot make-up water, one could see the entrainment of the hot water in the sloped portion of the cold leg and in the plume in the annulus. The hot "vent valve" water flowed from the upper portions of the annulus into the cold leg until it was entrained and carried toward the annulus with the cold layer. The entrainment of the hot (less dense) liquid in the sloped portion of the cold leg and in the annulus, coincides with the observations of Ellison and Turner .
The C-E model was used to predict the preliminary results of four of the Creare tests described in Table 4-3. These comparison calculations were performed for the annulus region only, since C-E designed plants do not have sloped cold legs. The experimentally measured conditions at the inlet to the annulus were taken as initial conditions for the calculations. The data being used for comparison to the calculations are thermocouple measurements made in the "annulus" directly below the cold leg and about 1/4" away from the "vessel" wall surface and are representative of the fluid temperature at the reactor vessel wall.
As the HPSI flow increases, the temperature of the fluid entering the annulus decreases. This is a result of the mixing that occurs in the cold leg. The predicted average temperature of the plume-and the predicted temperature next to the reactor vessel" wall are compared to the temperature data in Figures 4-10, 11, 12, and 13. In all cases the data, which was taken next to the vessel wall, indicates a higher temperature than the predicted temperature at the vessel wall. However, the predicted average plume temperature is in excellent agreement with the data. It is noted that the predicted difference between the fluid temperature at the wall and the average 0
fluid temperature of the plume is on the order .of 20 F. Comparison of the fluid temperatures taken 0.25 inches from the vessel wall with the only temperature measurement taken one inches from the wall indicates that 0
the observed temperature'difference is on the order of only 5 F. The actual 4-22
temperature distribution in the plume is apparently closer to being uniform than the Gaussian distribution assumed in the C-E model.
The comparison between the C-E model and the preliminary Creare test data indicates that the assumed temperature distribution used to predict the fluid temperature at the wall results in a lower than measured temperature. Better agreement between the C-E model and the Creare data would be obtained if the assumed temperature distribution were changed from Gaussian to uniform.
It is expected that comparison of the current MIXUP methodology with the results of tests which more accurately represent theC-E configuration will both confirm the reasonableness of the current predictions and permit removal of additional conservatism.
4-23
4.4 TRANSIENT SCENARIOS In this section, the transients which were analyzed for this study are described. Action Item II.K.2.13 requires an evaluation during recovery from small breaks with extended loss-of-feedwater . Accordingly, transients were analyzed which are characterized by a small break (including a break size of zero) concurrent with a loss of feedwater and by recovery actions during the transients to avoid core uncovery. Two possible recovery actions were considered: (1) opening two power-operated relief valves (PORVs) 10 minutes after the start of the transient and (2) restoration of feedwater 30 minutes after the start of the transient. Important transient parameters were first determined in a scoping study. A final set of transients was then established based on the scoping study results and analyzed in detail.
4.4.1 Parameter Scoping Study The important parameters to be included in the II.K.2.13 analysis matrix were first evaluated by a CEFLASH-4AS scoping study. Parameters considered were:
(1) break size, (2) break location, (3) and amount of HPSI flow. Results of.
the scoping study are briefly summarized.
4.4.1.1 Break Size Evaluation The three break sizes considered in the scoping study were .001, 0.01 and 0.02 ft2 . The smallest break approximates the area of a small instrument line, the 0.01 ft2 break approximates the area of one fully-opened PORV, and the largest break corresponds to the largest area where the steam generators are required to-remove core heat. As demonstrated in CEN-114, breaks larger than 0.02 ft2 result in a relatively rapid depressurization of the RCS without significant repressurization and are not a concern from a PTS standpoint. Both minimum and maximum HPSI flows were used in these scoping study evaluations. '
The break size was found to significantly influence the NSSS system response during the SBLOCA 'and LOFW transient. The worst PTS thermal-hydraulic conditions were expected for the break which resulted in the lowest downcomer fluid temperature and highest pressure. The depressurization resulting from the small (0.001 ft2 ) break was not significant and a relatively small amount of HPSI flow occurred, unless corrective actions were taken to enhance the depressurization. The fluid temperature in the downcomer remained warm in this case compared to the other breaks. The 0.02 ft2 break did result in a rapid system depressurization and significant HPSI flow. The downcomer fluid temperature decreased rapidly, but the corresponding rapid decrease in system pressure effectively eliminaths a significant PTS 'concern. The 0.01 ft2 '
break appeared to have the worst combination of pressure and temperature conditions. In this case, the HPSI flow did significantly cool the downcomer fluid, yet the system pressure remained at moderate levels. Therefore, break sizes of 0.01 ft 2 and smaller were selected for more detailed analysis.
4.4.1.2 Break Location Evaluation Cold leg, hot leg, and pressurizer break locations were considered in the scoping study. Results of the study for cases with maximum HPSI flow indicated that the break location has a relatively minor influence on the system pressure and fluid temperature. The difference in break .conditions (fluid enthalpy and mass flow) between the three 'break locations resulted in slight differences in the system pressure response. For the majority of the transient, the system pressure of the pressurizer break was higher than the hot leg break, and the cold leg breok resulted in the lowest pressur (of the three break loc.itionq).
In comparing fluid temperatures and pressures for these three break locations, the lowest downcomer fluid temperatures and highest system pressure occurred in the case 'where the break was located in the pressurizer. Therefore, this location was chosen for more detailed analysis.
The pressurizer break location is also important for establishing conditions where natural circulation is broken. Natural circulation is broken in the cold leg after sufficient fluid leaves the system and fluid no longer passes over the top of the steam generator tubes. The fluid in the cold leg and annulus 4-25
will remain liquid because it is subcooled by HPSI water. Similar conditions would be expected for a break located in the hot leg. In comparison, a cold leg break would promote loop flow into the cold legs for a longer time because the majority of the system inventory follows the natural circulation flow path towards the break.
The break location is also important in how natural circulation ceases. for this analysis, natural circulation ceases first in the cold legs of the unbroken loop (the loop with no pressurizer) due to faster depletion of the secondary side inventory qf this steam generator. The faster inventory depletion is due to the fact that some of the flow leaving the core and going into the broken loop goes out the break instead of being transported to the steam generator of the broken loop. Therefore, the heat removal load on the steam generator in the broken loop is less than the steam generator in the intact loop. This causes the secondary side inventory of the steam generator in the intact loop to deplete faster than in the steam generator in the broken loop. Consequently, natural circulation stops earlier in the unbroken loop.
Thus, there is a condition that can be established wherein one loop (the broken loop) still has natural circulation while the other loop has ceased natural circulation flow. Eventually both loops will cease natural circulation.
4.4.1.3 HPSI Evaluation Three HPSI injection curves were evaluated in the scoping study to determine the significance of the amounts of HPSI flow. In all cases the HPSI water temperature was assumed to be 400 F. A minimum HPSI pump curve and three charging pumps were modeled which corresponds to the operation of one HPSI train. This flow capacitycould be the result of the failure of one HPSI pump.
A maximum injection curve corresponding to three charging pumps and three HPSI pumps was also modeled which assumed the maximum pump flow of the 2700 MWt class plants. Finally, a special case with a high head HPSI pump curve which conservatively envelopes the Maine Yankee plant was evaluated.
4-26
The HPSI flow rate was found to have a significant impact on the RCS pressure and downcomer fluid temperature. The minimum HPSI flow of the 2700 MWt plants resulted in the least injection of cold water and therefore higher downcomer fluid temperatures, but also higher pressures. The maximum HPSI flow resulted in significantly lower downcomer fluid temperatures, but also significantly lower system pressures. The results of the scoping study indicated that a range of HPSI flow capacities must be considered for a more detailed analysis.
4.4.2 Detailed Analysis Matrix A total of eight SBLOCA and LOFW scenarios were selected for detailed analysis. The range of parameters considered in the analysis matrix are summarized in Table 4-4. Parameters considered were operator action to prevent core uncovery, break size, and HPSI flow. The selection of the eight SBLOCA and LOFW scenarios is described as follows.
The eight SBLOCA LOFW scenarios are subdivided according to the recovery action to prevent core uncovery with four cases that assume opening of 2 PORVs at 10 minutes and four cases where restoration of feedwater is assumed.
In the first category, if a SBLOCA and LOFW, occurs the PORVs may be opened to depressurize the system and actuate the HPSI pumps.. Based on a conservative analysis described in (CEN-114)(4-3) this action when taken within 10 minutes ensures that core uncovery does not occur. An alternative action will be to restore the auxiliary feedwater to the steam generators. Based on the conservative CEN-114 analysis, the feedwater was restored by 30 minutes to avoid core uncovery. This corresponds to the time after the steam generators nearly dry-out and are no longer removing significant heat from the RCS. It should be noted that the times for recovery action in CEN-114 are conservatively short, primarily because minimum HPSI flow was assumed for conservative core cooling analysis and conservative generic rather than plant specific parameters were assumed.
4.
4-27
Break sizes of zero, 1 PORV (0.0075 ft2 ), and 0.01 ft were selected for
- the detailed analysis. The 1PORV and 0.01 ft2 break areas provide essentially equivalent thermal-hydraulic conditions for the fracture mechanics analysis of II.K.2.13. The zero break size was selected as the limiting small break to fully evaluate the two types of core cooling recovery actions for PTS. The 0.01 ft2 and 1-PORV break sizes were selected because the scoping study indicated severe PTS conditions might occur for breaks of this size. The pressurizer was also selected as the worst break location, again based on the scoping study results.
A range of HPSI pump curves were selected to complete the set of eight transients. The HPSI flow was found to have a significant impact on system pressures and temperatures in the scoping study. The minimum and maximum composite HPSI pump curves were evaluated. A description of the HPSI and charging pumps modeled for the II.K.2.13 analysis was provided in Section 4.2.
4.5 RESULTS The eight SBLOCA and LOFW transients that were evaluated for II.K.2.13 are summarized in Table 4-4. These transients were subdivided into two categories of operator action to prevent core uncovery where: (1) 2 PORVs were opened at 10minutes or (2) the feedwater was restored at 30 minutes. The break size, break location, amount of HPSI, and plant design constitute the important II.K.2.13 parameters within these categories of corrective operator action.
Fluid pressures in the reactor vessel and fluid temperatures in the vessel annulus for the eight II.K.2.13 transients are provided in Figures 4-14 through 4-29. The fluid pressures were calculated for the reactor vessel using the best-estimate CEFLASH-4AS code. The fluid temperatures provided are: (1) the CEFLASH-4AS calculated cold leg bulk fluid temperature with complete mixing, and (2) the (MIXUP) calculated fluid temperature next to the reactor vessel 4-28
wall at three downcomer axial locations (mid-way between bottom of cold leg and top-of-core, top-of-core and core midplane). The calculation for fluid temperatures next to the vessel wall accounts for incomplete mixing of the cold leg and HPSI fluid as described in Section 4.3.2. Pressure and temperature results from the two categories of analyses are individually discussed in the following subsections.
4.5.1 II.K.2.13 Transient Results with Opening of 2 PORVs at 10 Minutes Results .for the four SBLOCA and LOFW transients with 2 PORVs opened at 10 minutes are presented in Figures 4-14 through 4.21. The four cases differ in the break size and amount of HPSI. Cases 1 and 2 are essentially LOFW transients (initial break size is zero) with minimum (Case 1) and maximum (Case 2).HPSI pump curves. Cases 3 and 4 are SBLOCA and LOFW transients with an initial break size of one PORV and minimum (Case 3) and maximum (Case 4) HPSI pump curves.
The system pressure shown in Figures 4-14 and 4-16 of both Cases 1 and 2 (with a zero initial break size) initially decreased from 2250 to about 2000 psia as the reactor was tripped due to a low steam generator level following the initiating LOFW event. The pressure remained above 1950 psia until 600 seconds, at which time the 2 PORVs were opened, and the pressure rapidly decreased to about 1100 psia. After this time, the,pressure of Case 1 (minimum HPSI flow), stabilized at about 1075 psia until 4000 seconds. In comparison, the pressure of the maximum HPSI Case 2 briefly stabilized at about 1000 psia until 1400 seconds and then continued to depressurize. The more rapid depressurization in Case 2 is due to A greater proportion of the core dec.1y heat being absorbed in heating the injection water. N6 significant system repressurization was apparent in either transient.
4-29
The fluid temperatures shown in Figures 4-15 and 4-17 are calculated by: (1)
CEFLASH-4AS for the cold leg and (2) by a mixing model (MIXUP) for the fluid temperature next to the vessel wall. The downcomer fluid temperature in both cases was initially 550 F. The average cold leg temperature calculated with CEFLASH-4AS for Case 1 (minimum HPSI) remained relatively steady for the entire transient. 'Incomparison, the average cold leg temperature for Case 2 (maximum HPSI) rapidly decreased after the PORVs were opened. The temperatures next to the vessel wall, initially decreased at 66 seconds as charging pumps were started and then step-decreased after the PORVs were opened and HPSI flow was initiated. As expected, the lowest fluid temperature was calculated for the highest of the three axial positions( midway between the core and cold leg). The top-of-core temperatures were slightly higher, and the core midplane temperatures were significantly warmer than the temperatures between the core and cold leg.
The vessel pressure and cold leg and downcomer temperatures for Cases 3 and 4 are provided in Figures 4-18 through 4-21. These SBLOCA and LOFW transients have an initial break size of one PORV and differ in the HPSI flow. The system pressures of Cases 3 and 4 rapidly decreased from 2250 to about 1050 psia as both the SBLOCA and LOFW occurred simultaneously. The pressure of Case 3 (minimum HPSI) stabilized at this level until 3600 seconds as a balance between primary system input and output energies was established. The pressure of Case 3 was then calculated to gradually decrease after this time as the system gradually cooled. In comparison, a repressurization from 1050 to 1085 psia occurred with the maximum HPSI (Case 4) before 600 seconds. The repressurization occurred as the large HPSI flow refilled the primary system.
The pressure of the system after refilling solid was a balance between break flow and the HPSI flow. After 600 seconds, the system pressure continuously decreased as both PORVs were opened.
The system temperatures in the cold leg and the downcomer for Cases 3 and 4 are shown in Figures 4-19 and 4-21. The cold leg fluid temperatures were 4-30
calculated to gradually decrease in both transients . The rapid decrease from 1000 seconds 550 to 5000 F in the Case 3 cold leg temperature at about resulted from a break in the natural circulation of hot fluid in the cold leg.
The calculated temperatures next to the vessel wall for Case 3 by rapidly decreased from initial values of 5500F to between 400 and 480 0F 1200 seconds as cold HPSI flow was established.. In comparison, the than Case 3 temperatures for Case 4 at 1200 seconds were significantly cooler by about 100 0 F. The lowest temperature of 300OF before the repressurization at 600 seconds was calculated for Case 4 (maximum HPSI) midway between the core and cold leg.
this Natura.l circulation of fluid in the cold legs stopped for Cases 1 and 3 of were analysis. For these cases, the cold leg and downcomer fluid temperatures calculated to remain relatively warm for the time period of concern for PTS.
The warm temperatures were attributed to the occurrence of a recirculation fluid. The pattern that was calculated between the cold leg and annulus recirculation pattern resulted from the interchange of cooler, more dense, cold annulus remained leg fluid with the warmer, less dense annulus water. The in the broken relatively warm due to the natural circulation still occurring loop.
A comparison of pressures and temperatures for Cases 1 through 4 indicates that the HPSI flow significantly influenced the system response. The low HPSI flow of Cases 1 and 3 (minimum HPSI) resulted in relatively warm downcomer fluid temperatures that were greater than 350 0F for the time of interest from a PTS standpoint (about 5000 seconds). The maximum HPSI flow of Cases 2 and 4 two resulted in two different system responses. In Case 2, the opening of PORVs at 10 minutes with maximum HPSI flow resulted in a rapid system cooldown and decrease in pressure. In Case 4, the initial 0.0075 ft2 pressurizer break resulted in an initial depressurization, followed by HPSI flow that filled the system causing a slight repressurization. The lowest temperature calculated for this case before the repressurization was 300OF.
4-31
The results of Case 4 are expected to be representative if the transient had been analyzed with the high-head HPSI delivery curve. The Initial depressurization is estimated to be similar for analysis with the maximum (Case
- 4) and high-head HPSI delivery curves. The magnitude of the repressurization, which potentially could be higher for the high head HPSI pumps, would be limited (as in Case 4) by the opening of the second PORV at 10 minutes after the start of the transient.
4.5.2 II.K.2.13 Transient Results with Restoration of Feedwater at 30 Minutes Thermal-hydraulic results of the four SBLOCA and LOFW transients which assume the restoration of feedwater at 30 minutes are presented in Figures 4-22 through 4-29. The four cases differ in break size, and HPSI flow, with one case modeling the HPSI and auxiliary feedwater flow of Maine Yankee.
Results for Cases 5 and 6 are provided in Figures 4-22 through 4-25. Both cases are essentially LOFW transients (initial break size is zero) with either the maximum composite HPSI flow plants (Case 5) or the maximum conservative high-head pump flow plant (Case 6). The system pressure of Case 5 initially decreased from 2250 to about 2000 psia after the LOFW resulted in a decrease in the secondary side level and the core power.was tripped. The pressure in Case 5 then gradually decreased as the system cooled until 1700 seconds, when the steam generator heat transfer became small. Without significant steam generator heat transfer, combined with the charging pumps refilling the system, the pressure then rapidly increased in this case to the 4-32
PORV setpoints of 2400 psia. Auxiliary feedwater was restored at 1800 seconds e at the maximum feedwater rate for the composite plant and the pressure in Case 5 rapidly decreased. The feedwater flow was conservatively assumed to continue at the maximum rate without secondary side level control. This conservative assumption maximizes RCS cooldown. In reality, either manual or automatic auxiliary feedwater control would be expected to produce a less severe cooldown. The pressure decrgase resulted in a combined charging pump and HPSI flow, which refilled the primary system solid by 4200 seconds. The refilling of the primary system resulted In a significant repressurization to the PORV set point of 2400 psi. The pressure will remain at this level assuming the charging pumps continue to operate.
The system pressure of Case 6 initially decreased from 2250 to 2000 psia after the reactor was tripped. The pressure then gradually decreased to 180O psia by 500 seconds as the system cooled. In comparison to Case 5, the HPSI flow shut-off head is greater than 2400 psia and HPSI flow to the system was initiated at high pressures.' The system was calculated to refill at 500 seconds and thepressure increased to the PORV set point of 2400 psia. At 1800 secohds, feedwater was assumed to be restored at the maximum rate for Maine Yankee; however, to.prevent an unrealistic plant cooldown, the feedwater was assumed to be automatically controlled by secondary side level after the top of the steam generator tubes were covered. After '1800 seconds, the primary side momentarily depressurized due to the rapid cooling of the secondary side and the large flow rate of cold HPSI water. After the brief depressurization, the HPSI flow againfilled the system solid and the pressure increased to 2400 psia.
The system temperatures for Cases 5 and 6 are shown in Figures 4-23 and 4-25.
The fluid temperature in both cases was initially 550F. The average cold leg temperature calculated with CEFLASH-4AS for both transients remained relatively constant until the feedwater was restored. After restoring feedwater, the 4-33
temperatures rapidly decreased' as the HPSI flow became large. The temperatures for Case 5 continued to decrease as the cold feedwater continued to flow to the steam generators. The temperatures for Case 6 gradually increased after about 2200 seconds, when the steam generator tubes were covered, and the feedwater flow was controlled. The temperatures calculated next to the vessel wall initially decreased after charging pumps and/or HPSI was started and then followed the trends calculated by the bulk fluid temperature In the cold leg.
Prior to the final repressurization at 4200 seconds in Case 5, the lowest temperature calculated was 240 F. In comparison, the temperature decreased to about 2800F in Case 6 prior to the final repressurization.
The downcomer pressures and temperatures for Cases 7 and 8 are provided in Figures 4-26 through 4-29. Both cases assume an initial break size of 0.01 ft2 and differ by the amount of HPSI flow. The pressures of both transients rapidly decreased from 2250 to below 1200 psia as the SBLOCA and LOFW were simultaneously initiated. The pressure of Case 7 (minimum HPSI) then gradually decreased to about 1100 psia by 1800 seconds, and were estimated to
. continue to decrease after feedwater was restored at that time. The pressure of Case 8 (maximum HPSI) decreased to 1050 psia by 400 seconds. The maximum HPSI refilled the system by 600 seconds and the system repressurized from 1050 to 1225 psia. The pressure of the system after refilling represents a balance between break flow and HPSI flow.
The temperatures for the cold leg and adjacent to the downcomer vessel wall for Cases 7 and 8 are provided in Figures 4-27 and 4-29. The temperatures next to the vessel wall for both cases decreased rapidly by 200 seconds as cold HPSI flow was established. Before feedwater was restored at 1800 seconds, the lowest temperature of 3000F was calculated for Case 8 midway between the core and cold leg. After restoring feedwater, the pressures and temperatures of both Cases 7 and 8 continually decreased. From a.PTS standpoint, the time period of interest for these transients is to about 2000 seconds, where pressures are still relatively high.
4-34
A comparison of pressure and temperature results for Cases 5 through 8
- indicates that the worst PTS conditions for the II.K.2.13 cases occur in Cases 5 and 6. Both cases represent ,aLOFW with zero break size and maximum HPSI pump flows for the composite plant and the high head HPSI plant . In both cases, the fluid temperatures rapidly decreased as a result of restoring feedwater and the pressure increased after the HPSI pumps completely refilled the system. In comparison, Cases 7 and 8 were SBLOCA and LOFW transients with an initial break of 0.01 ft2 . The resulting pressures and temperatures for these cases are less challenging from a PTS point of view.
4-35
4.6 CONCLUSION
S FROM THERMAL-HYDRAULIC ANALYSIS According to the request of Action Item II.K.2.13, transients were analyzed where the initiating event is the simultaneous occurrence of a small break and a total loss of feed water. In addition, during these transients, actions are taken to prevent core uncovery, either by opening two PORVs 10 minutes after the accident or by restoring auxiliary feedwater 30 minutes after the accident.
A total of eight SBLOCA and LOFW transients were evaluated. These transients were expected to result in severe pressurized thermal shock conditions. The break location is at the pressurizer and the transients spa.n the range of break sizes from zero or very small breaks to 0.01 ft2 . Most of. the analyses were performed for a composite reference plant which envelopes all C-E operating plants.and is adequately representative of other plants being evaluated. In addition, one separate analysis was performed with high-head high pressure safety injection pumps which conservatively envelope the pumps of the Maine Yankee Plant. Minimum injection water temperature was assumed in all cases.
The amount of HPSI flow was found to significantly influence the thermal hydraulic response of the 'transients. Two rates of HPSI flow were evaluated for the reference plant: (1) minimum flow assuming one HPSI train and charging pumps and (2) maximum flow assuming the largest HPSI and charging pump flow of C-E 2700 MWt class plants. A comparison of downcomer fluid temperatures and systems pressures for the eight II.K.2.13 transients indicates that the lowest downcomer fluid temperatures and highest pressures were calculated for the cases of zero (or very small) initial break size, maximum HPSI flow, and a LOFW that Is restored by the operator at 30 minutes. The recovery of auxiliary feedwater was conservatively assumed to occur at'maximum flow rate at a.
feedwater temperature of 40oF.
The degree of mixing of the cold HPSI and charging pump water with the hot water in the cold leg and downcomer is an important parameter. Mixing of cold O HPSI water was evaluated based on a hot water entrainment model that was developed for the present study. The model assumes the cold liquid mixes with the hot loop flow at the injection location, moves without mixing along the
bottom of the cold leg and mixes in the downcomer with the surrounding hot fluid. The downcomer mixing prediction of the model was compared against experimental data and showed very good agreement.
4-37
REFERENCES 4-1 0. G. Eisenhut to Owners Groups, "Clarification of TMI Action Plan Requirements", NUREG-0737, November, 1980.
4-2 S. Leichtberg, "C-E Analysis of LOFT Test L3-6", Specialists Meeting on Small Break Loss of Coolant Accident Analyses in LWRs, Monterey, CA.,
August 25-27, 1981.
4-3 "Review of Small Break Transients in Combustion Engineering Nuclear Steam Supply Systems"i CEN-114, July, 1979.
4-4 CENPD-137, "Calculative Methods for the C-E Small Break LOCA Evaluation Model", August, 1974.
4-5 CENPD-137, "Calculative Methods for the C-E Small Break LOCA Evaluation Model", Supplement 1, January 1977.
4-6 CENPD-133, Supplement 1, "CEFLASH-4AS, A Computer Program for Reactor Blowdown Analysis of the Small Break Loss-of-Coolant Accident", August, 1974.
4-7 CENPD-133, Supplement 3, "CEFLASH-4AS, A Computer Program for Reactor' Blowdown Analysis of the Small Break Loss-of-Coolant Accident", January, 1977.
4-8 J. A. Block, Preintation of Slid@s of Credre Mixingj Tet, Chicago EPRI Meeting, September 30, 1981.
4-9 J. S. Turner, "Buoyancy Effects in Fluids", Cambridge at the University Press, (1973).
4-10 T. H. Ellison 'and J. S. Turner, "Turbulent Entrainment in Stratified Flows", Journal of Fluid Mechanics, Vol. 6 (1959).
4-38
4-11 B. R. Morton, G. I. Taylor, and J. S. Turner, "Turbulent Gravitational Convection From Maintained and Instantaneous Sources", Proceeding of the Royal Society, Vol. 234 (1956).
4-12 1. H. Shames, "Mechanics of Fluids", McGraw-Hill Book Company (1962).
4-13 P. H. Rothe, W. D. Marscher, and J. A. Block, "Quick Look Data Report Fluid and Thermal Mixing in a Model Cold Leg and Downcomer with Vent Valve Flows", EPRI Preliminary Report Research Project RP347-1,(1981).
4-39
TABLE 4-1 COMBUSTION ENGINEERING PLANT PARAMETERS Plant Core Primary Main Steam PORV No. of Effective PORV Power System Safety Valve Setpoint PORVs Flow Areajer Valve Mwt Volu e Setpoint psia ft ft psia Ft. Calhoun 1500 6,600 1000 2400 2 0.0057 Palisades 2530 10,600 1000 2400 2 0.0086 IMaine Yankee 2630 10,800 1000 2400 2 0.0075 Calvert Cliffs 1, 2 2700 11,000 1000 2400 2 0.0075 Millstone 2 2700 11,000 1000 2400 2 0.0075 St. Lucie 1 2700 11,000 1000 2400 2 0.0075 St. Lucie 2 2560 10,800 1000 2400 2 0.0075 Waterford 3410 11,300 1200 N/A None N/A San Onofre 3410 11,300 1200 N/A None N/A System 80 3800 14,400 1300 N/A None N/A ANO-2 2825 10,000 1100 (1) (1) (1)
(1) ANO-2 does not have PORVs, but instead has two motor operated valves which can be opened by operator action.
TABLE 4-2 COMBUSTION ENGINEERING PLANT PARAMETERS Plant Maximum Maximum HPSI Maximum Total Minjmm HPSI No. of Charging Charging No. of HPSI Pumps Pump-Shutoff HPSI Flow at Flowl) at MSSV Pumps Automatically Pump Flow Automatically Head KSSV Setpoint Setpoint Actuated Per Pump Actuated psia gpm gpm on SIAS 9PM 3 1570 870 250 3 40 Ft. Calhoun 3 1430 1150 310 .3 40 Palisades 2 2475 1100 550 (2) (2)
Maine Yankee Calvert Cliffs 1, 2 2 1440 800 340 3 44 Millstone 2 2 1410 680 280 3 44 St. Lucie 1, 2 3 1440 1170 320 3 44 ANO-2 2 1610 860 340 3 44 San Onofre 2 1610 800 310 3 44 Waterford 2 1610 800 310 3 44 System 80 2 1970 1340 460 3 44 (1) From one HPSI pump (2) The Maine Yankee HPSI pumps are used for charging.
TABLE 4-3 Test Conditions for Creare Mixing Tests Creare 1 2 3 Case No. Test No. Density Ratio Flow Ratio HPSI Flow (GPM) 1 1 .02 5 2.07 2 5 .02 5 4.14 3 .19 .02 5 6.21 4 2 .02 5 8.28
( icoldhot
- 1. Density Ratio =
~cold vent value
- 2. Flow Ratio =
- 3. HPSI Flow Volumetric flow rate of cold water into bottom of cold leg 4-42
II.K.2.13 SBLOCA + LOFW ANALYSIS MATRIX CASE OPERATOR ACTION BREAK SIZE (ft2) BREAK LOCATION HPI FLOW PLANT TO PREVENT DEGRADED CORE DESIGN 1 2 PORVs Opened at 10 min. 0 (or. Small) N/A MIN Composite 2 2 PORVs Opened at 10 Min. 0 (or Small) N/A MAX Composite 3 2nd PORV opened at 10 Min. 1-PORV Pressurizer MIN Composite 4 2nd PORV opened at 10 Min. 1-PORV Pressurizer MAX Composite 5 Restored Feedwater at 30 Min. 0 (or Small) N/A MAX Composite 6 Restored Feedwater at 30 Min. 0 (or Small) N/A MAX High-Head HPSI 7 Restored Feedeater at 30 Min. 0.01 Pressurizer MIN Composite 8 Restored Feedwater at 30 Min. 0.01 Pressurizer MAX Composite
FIGURE 4 -1 REACTOR VESSEL THER1iAL SHOCK ANALYSIS HIGH PRESSURE SAFErY INJECTION PUMP DELIVERY 3000 2500 2MIAINE YANKEE HPSI PUMPS 2000 2 ANO-2 HPSI LPM Ln C..*
8ooo I FT, 1000 CALHOU.N HPSt f .UIi PUM~P HS.UP 1 MILLSTO E-2 500 HPSI PUMP 0 1500 2500 0 500 1000 2000 HPSI FLO',! GPM 4-44
FIGURE 4-2 REACTOR VESSEL THERMAL SHOCK ANALYSIS HIGH PRESSURE INJECTION DELIVERY 3000 2500 HIGH-HEAD HPI FLOW MAXIMUM COMPOS It HPSI FLOW'
.1500 LUL MNI0NUl M MAXIMUM TOTAL 1000 COMPOSITE HPSI + CHARGING HPSI FLOW UMiPS M1INIMIUM TOTAL 50 -HPSI + CHARGING 50 PUMPS 0
0 500 1000 1500 2000 2500 HPSI.FLOW GPM 4-45
FIGURE 4-3 CREARE SCOPING TEST CONFIGURATION HOT WATER DYED COLD WATER/
SALT WATER
FIGURE 4-5 FLUID ENTRAINMENT CONSTANT AS A FUNCTION OF CHANNEL INCLINATION I I I I I I I 0.09 0 0 0.08 - 0 0
0.07 - 0 0.06 0 E 0.05 O 0 0.04 O
0.03 cm 0 0.02 - 0 0
0.01 0.00 I
- 0. 100 200 300 400 500 600 700 800 900 INCLINATION ANGLE (.)
(E BASED ON THE MEAN VELOCITY OF THE PLUME) 4-48
FIGURE 4-6 C-E MIXING MODEL COLD WATER HOT WATER FALLING PLUME - GOOD ENTRAINMENT Q HORIZONTAL GRAVITY CURRENT - POOR ENTRAINMENT (D VERTICAL GRAVITY CURRENT - GOOD ENTRAINMENT A -1
FIGURE 4-7 NOMENCLATURE FOR HORIZONTAL riRAVITY CURRENT
/0 LAYER INTERFACE DOWNCOMER COLD LEG R
20 y
C 4-50
FIGURE 4-3 TEMPERATURE DISTRIBUTION OF A FALLING PLUME T HOT WATER COLD WATER X
T CORE BARREL COLD LEG WALL rX T
REACTOR VESSEL WALL TEMPERATURE DISTRIBUTION T
EDGE OF PLUME z
4-51
FIGURE 4-9 CREARE MIXING EXPERIMENT CONFIGURATIONS HORIZONTAL COLD LEG TEST Q HOT WATER INJECTION HOT WATER DYED COLD WATER/
SALT WATER ANGLED COLD LEG TEST O2 4-52
FIGU RE 4-10o COMPARISON OF C-E MIXING MODEL WITH PRELIMINARY CREARE DATA - CASE 1 TEMPERATURE, 0F 100 120 140 160 0
I PREDICTION (AVERAGE) 5 LO
- W 10 o PREDICTION z (WALL) o 15 PLUME TEMPERATURE 20 O=WALL 1/4" FROM 4-53
FIGURE 4-11 COMPARISON OF C-E MIXING MODEL WITH PRELIMINARY CREARE DATA - CASE 2 TEMPERATURE, F 100 120 140 160 0
IPREDICTION (AVERAGE) 5 O 10
.J a1 *v' PREDICTION 15 (WALL)
- 20. g PLUME TEMPERATURE 1/4" FROM WALL 4-54
FIGURE L,-12 COMPARISON, OF C-E MIXING MODEL WITH PRELIMINARY CREARE DATA - CASE 3 TEMPERATURE, F 100 120 140 160 0
I a PLUME TEMPERATURE s 1/4" FROM WALL 5
LaJ IPREDICTION oD (WALL) a 10 -%'
F 15
\ PREDICTION (AVERAGE) 20 4-.5 4-55
FIGURE 4-13 COMPARISON OF C-E MIXING MODEL WITH PRELIMINARY CREARE DATA - CASE 4 TEMPERATURE, 0F 100 120 140 160 0 I PLUME TEMPERATURE 1/4" FROM WALL 5
-. j 0 10 PREDICTION PREDICTION (WALL) (AVERAGE) 15 .
20 4-56
FIGURE Ll-14 CASE 1 REACTOR VESSEL PRESSURE 2800,0 BREAK SIZE 0 FT2 BREAK LOCATION N/A HPI MINIMUM 20000 OPERATOR ACTION 2 PORVS AT 10 MIN 2000.0 1600.0 LUJ rsoolo ESTIMATED TREND 400.0
- 0. 2000.. 4000. 6000. 8000, 10000, TIME, SECONDS 4-57
FIGURE LI-15 CASE 1 COLD LEG AND DOWNCOMER FLUID TEMPERATURES I. I I I BULK FLUID TEMP., COLD LEG (FLASH)
FLUID TEMP. AT VESSEL WALL (Mix-UP)
-- - MID-DISTANCE, CORE AND COLD LEG
-- - TOP OF CORE
-- - MIDDLE OF CORE 600.0 500.0.
400.0 UJ 300.0 ESTIMATED TREND 200.0 100.0I
- 0. 2000. 6000.
6000. 10000.
TIME, SECONDS 4-58
FIGURE 4-16 CASE 2 REACTOR VESSEL PRESSURE 2800.0 BREAK SIZE 0 FT2 BREAK LOCATION N/A 2400.0 HPI MAXIMUM OPERATOR ACTION 2 PORVS AT 10 MIN 200010 1600.0 1200.0 800.0 400.0
- 0. 600. 1200. 1800. 2400, 3000.
TIME, SECONDS 4-59
FIGURE 4-17 CASE 2 COLD LEG AND DUWNCOMER FLUID TEMPERATURES 700.0 BULK FLUID TEMP., COLD LEG (FLASH)
FLUID TEMP, AT VESSEL WALL (MIX-up) 600.0 _ - -- MID-DISTANCE, CORE AND COLD LEG
- -- TOP OF CORE
- -MIDDLE OF CORE 500.0 r 400.0 200.0 -bI 100. 0
- 0. 600. 1200, 1800. 2400. 3000.
TIME, SECONDS 4-60
FIGURE 4-18 CASE 3 REACTOR VESSEL PRESSURE 2800.0 BREAK SIZE 1 PORV BREAK LOCATION PRESSURIZER HPI MINIMUM 2400.0 OPERATOR ACTION 1 PORV AT 10 MIN 2000.0
~ 1600.0 CO La 1200.0 80010 400.0
- 0. 1200. 2400. 3600. 4800. 6000, TIME, SECONDS 4-61
FIGURE 4-19 CASE 3 COLD LEG AND DOWNCOMER FLUID TEMPERATURES 700.0
--- BULK FLUID TEMP., COLD LEG (FLASH)
FL UID TEMP. AT VESSEL WALL (MIX-UP)
- - - MID-DISTANCES CORE AND COLD LEG 600.0 - - TOP OF CORE
- -- MIDDLE OF CORE 500.0 U
4 00.0 -...
300.0 200.0 -
100.0 I I
- 0. 1200. 2400. 3600. 4800. 6000.
TIME, SECONDS 4-62
FIGURE 4-20 CASE 4 REACTOR VESSEL PRESSURE 280010 BREAK SIZE 1 PORV BREAK LOCATION PRESSURIZER 2400.0 - HPI MAXIMUW OPERATOR ACTION .. PORV AT 10 MIN 2000.0
-cc 1600.0 C,,
1200.0 800.0 400.0 0.0 1200.0 2400.0 3600.0 4800.0 6000.0 TIME, SECONDS 4-63
FIGURE 4-21 CASE 4 COLD LEG AND DOWNCOMER FLUID TEMPERATURE 700.0 BULK FLUID TEMP., COLD LEG (FLASH)
FLUID TEMP. AT VESSEL WALL (MIX-UP1
- - - MID-DISTANCE, CORE AND COLD LEG 600.0 - TOP OF CORE
- *-MIDDLE OF CORE so I 500.0 I S400.0 200.0 100.0
- 0. 1200. 2400, 3600. 4800. 6000.
TIMiE, SECONDS 4-64
FIGURE 4-22 CASE 5 REACTOR VESSEL PRESSURE 2800.0 2400.0 2000.0 C/3 uL 1600.0 1200.0 BREAK SIZE 0 FT 2 BREAK LOCATION N/A 80010 HPI MAXIMUM OPERATOR ACTION RESTORE Fw 30MIN 400.0
- 0. 1200. 2400. 3600. 4800. 6000, TIME, SECONDS 4-65
FIGURE 4-23 CASE 5 COLD LEG AND DOWNCOMER FLUID TEiiPERATURES 700.0 .1 BULK FLUID TEMP., COLD LEG (FLASH)
FLUID TEMP. AT VESSEL WALL (MIX-UP)
- -- MID-DISTANCE, CORE AND COLD LEG 600.0 - - TOP OF CORE MIDDLE OF CORE 500.0 0 - 400.0 300.0 200,0 200.0 200.0~
TI3E, SECONDS
FIGURE 4-24 CASE 6 REACTOR VESSEL PRESSURE 2800,0 2400.0 2000.0 u 11600.0 LUJ 0
1200.0 BREAK SIZE 0 FT2 BREAK LOCATION N/A 800.0 HPI HIGH-HEAD HPSI OPERATOR ACTION RESTORE FW AT 30 MIN 400.0
- 0. 1200. 2400. 3600E 4800. 6000.
TINE, SECOND0 4-67
FIGURE 4-25 CASE. 6 COLD LEG AND DOWNCOMER FLUID TEMPERATURES 700.0 BULK FLUID TEMP., COLD LEG (FLASH)
FLUID TEMP. AT VESSEL WALL (MIX-UP)
MID-DISTANCE, CORE AND COLD LEG 600.0 - - TOP OF CORE MIDDLE OF CORE 500.0 400.0 Lu 300.0 200.0 100.0
- 0. 1200. 2400. 3600. 4800, 6000.
TIIE, SECONDS 4-68
FIGURE LI-26 CASE 7 REACTOR VESSEL PRESSURE, 2800.0 2 BREAK SIZE 0.01 FT2 BREAK LOCATION PRESSURIZER HPI MINIMUM 2400.0 OPERATOR ACTION RESTORE FW AT 30 MIN 2000.0 160010 ESTIMATED TREND 1200.0 800,0 400.0
- 0. 400. 800. 1200, 1600, 2000.
TINE, SECONDS 4-69
FIGURE 4-27 CASE 7 COLD LEG AND DOWNCOMER FLUID TEMPERATURES 700.0 600.0 500.0 400.0 ESTIMATED 300.0- TREND BULK FLUID TEMP., COLD LEG (FLASH)
Fl Iltf TFMP. AT VFSSEL WALL (MIX-UP) 200.0 - AND COLD LEG
--- MID-DISTANCE,.CORE TOP OF CORE
- *- MIDDLE OF CORE 100.01
- 0. 400. 800. 1200. 1600. 2000.
TIME, SECONDS 4-70
FIGURE 4-28 CASE 8 REACTOR VESSEL PRESSURE 2800.0 BREAK SIZE 0.01 FT2 BREAK LOCATION PRESSURIZER 2400.0 HPI MAXIMUM OPERATOR ACTION RESTORE Fw 30 MIN 2000.0 1600.0 1200.0 800 .0
- 0. 1200. 2400, 3600, 4800. 6000, TINE, SECONDS 4-71
FIGURE 4-29 CASE 8 COLD LEG AND DOWNCOMER FLUID TEMPERATURES 700.0
- BULK FLUID TEMP., COLD LEG (FLASH)
FL-I l TEMP, AT VFSSEL WALL (MIX-UP) 600.0 - -- - MID-DISTANCE, CORE AND COLD LEG
-- TOP OF CORE
- MIDDLE OF CORE 500.0 400.0 Lu 300,0 200.0 100.0
- 0. 1200. 2400. 3600. 4800. 6000.
TINE, SECONDS 4-72
Chapter 5 Fluence Distributions 5.0 Introduction for this section of the analys is The purpose of the calculations performed fluence distribution throughout the to provide the shape of the fast neutron Maine of interest in the reactor pressure vessel. For all except the region neutron fluence is aetermined from Yankee plant, the magnitude of the fast power histories applicable to erach surveillance capsule data and reactor vessel Fluence data for the Maine Yankee specific reactor pressure vessel.
The plants considered as part of this was provided by the plant owner.
Cliffs Unit 1, Calvert Cliffs Unit 2, analysis are Fort Calhoun Unit 1, Calvert 2, and St. Lucie-Unit 1.
Maine Yankee, Palisades, Millstone Point-Unit 2 and Unit 3, Waterfora-Unit 3, Arkansas Nuclear One-Unit 2, San Onofre Unit 1, 2 ana 3 dic not have fluence St. LucieUnit 2 and Palo Verde Units cistributions developed by this program.
plant was synthesized from four The fluence aistribution for a given an azimuthal shape, and a set of components, a peak value, an axial shape, Point-Unit 2 were utilized as racial shapes. Fort Calhoun and Millstone surveillance capsule reference plants because. of the availability of previous that St. Lucie 1 and Millstone Point 2
analyses. The plants were grouped such differing only in the source were treated with calculational models Calvert Cliffs Units 1 and 2 were distribution ana reactor power history.'
2 category for the azimuthal distribution placed in the Millstone Point Unit for tne determination of the axial calculation but in a separate category the thermal shield in the Calvert distribution in orcer that the absence of was grouped in the Calvert Cliffs Cliffs Units could be considered. Palisades The azimuthal Shape used in this model for the axial shape calculations.
Power. Surveillance capsule analysis for Palisades was provided by Consumers by all of the plant owners for data ana reactor power histories were supplied Plant specific fluence whom fluence distributions were calculated.
distributions are provided in the appendices.
5-1
5.1 Definitions of the The neutron fluence distribution is aefined as the integral neutron flux density over a specified time interval as shown in Equation 5.1, (ree)= (r,a,ot)dt, (5.1) where: is the fluence distribution, is the flux aensity, r is the radial coordinate,
- 2. is the axial coordinate, e is the azimuthal coorcinate t is the time variable, t is the beginning of the time interval of interest, t2 is the end of the time interval of interest.
The fast neutron flux density is defined as the portion of the neutron Tne fast flux density comprised of neutrons with energies above 1.0 MeV.
neutron flux density can also be expressea as an integral over the energy is 1.0 MeV ano the aependent flux density where the lower energy limit upper limit is unbounded, as shown in Equation 5.2.
(razet) . f ~ (r,Fe,t,E)oE, E1 =1.0 MeV (5.2)
E>1.0 MeV where:. E is the energy variable, and the other quantities are as noted in Equation 5.1.
the fast The fast fluence is obtainea by substituting the equation for flux density into Equation 5.1.
5-2
5.2.1 Normalization The magnitude of the flux distribution is determined by the normalization constant, P, as shown in Equation 5.3. The value of P for a specific as reactor vessel is based on the peak fluence on the vessel walls surveillance reported in the analysis of the reactor pressure vessel capsules. These surveillance capsule analysis reports, or their performed for this report.
equivalent, were used as input to the analysis 5.2.2 Azimuthal Distributions in The normalized azilutmal flux istribution is Cenetet as f3 ()
factor refers to tne peak Equation 5.3. In this case the normalization Esuation 5.3. The calculational fluence on the vessel wall, i.e. "P" in 5 1 -RG mocel to determine e(e) fv3 for a technique employed a DOT were calculated using reference plant ana then adjustment factors computer code.
SHADRAC(5 -2 ). a three dimensional point-kernal-metfloo in the plant specific The adjustment factors accounted for differences which were used to 52 ano reference plant time-averaged power distrioutions construct the source for the azimuthal shape calculations.
Boltzmann transport equation in DOT is a computer code that solves the discrete ordinates is used and two dimensional geometries. The method of of particles moving along balance equations are solved for the density two-dimentional spatial mesh.
discrete directiorns in each cell of a expansion. For this Anisotropic scatter~ing is treated using a Legendre order of expansion of the application afn S toR9 Quadrature and P3 model for the reference transfer cross sections were used. The geometric i.e. using radial and azimuthal plants was constructed in an "R9" mesh, through the midplane of the intervals to represent a torizontal slice
.reactor.
in an earlier verszjion as C-17.
SHADRAC mhas also been known as G30 and moments method as represented by The basic algoritnms of SHADRAC use the by the Nuclear Dev longpeot tne differetialenergy specta calculated 5-4
an. infinite Corporation of America for a point isotropic source in from the source to detector
- medium. Only the line-of-sight contribution is consioerea. Distributea sources are represented by a collection of combining point sources. The. azimuthal flux distribution is obtained by the ratio of two tne results of the fixed-source RO-DOT calculation with fixed source SHADRAC calculations as shown in Equation 5.4.
3 f SSpc(E)) (5.4) fT SREF()
where: f3 is the azimuthal flux distribution 9 D T(~~efe(O is the azimuthal angle maximum 9P is the azimuthal angle at which f is at its DRef(e) is the azimuthal flux 0istribution from the DOT calculations of the reference plant Sf is the azimutnal flux distribution from the SHADRAC calculations of the reference plant Ss (9)- is*tne azimuthal flux distribution from the SHADRAC SSPC calculations of a specific plant.
5.2.3 Radial Distributions f r).
The norm alized radial distribution is given in Equation 5.3 as by the given radial The superscript - shows the elevation represented distribution. The 40 parameter was used to treat the non-separability reactor vessel not of the flux distribution for the portion of the directly opposi.te the' active core. The :racial distributions were represented a obtained from DOT (5-1) calculations of an RE model which reference plant. Three reference RE models were employed to develop As in the azimuthal distribution plant specific. adjustments.
expansion of the calculations the DOT models employed a P3 Legenare of the DOT transfer cross section ana an S8 (RZ) quadrature. The results value of the flux on calculations were edited ana normalized to the peak the vessel inner surface, i.e. "P" in [quation 5.3.
5-5
5.2.4 Axial Distributions The normalized axial aistribution, f () of Equation 5.3 is ootainea from the same DOT-RZ calculations described in tne discussion of the racial distributions. The normalization factor is, again, "P" in Equation 5.3. In tne axial direction the models ranged from the mid-plane of the active core to the centerline of the reactor vessel nozzles. Symmetry.. about the core miaplane was assumea. I Truncated trapazoiaal power shapes that enveloped the reference plant time-averaged axial power distributions were used for the source distribution in the reactor core. The axial flux data from the DOT calculations was edited along the vessel-clad interface on the inner surface of the vessel and then normalized to form the axial aistribution, f2(s). It woulo have been possiole to parameterize f (e) as a function of radius in order to approximate the non-separability of the, flux distribution, however purpose because the fl(r), the radial aistribution was cnosen for this relative invariance of fl(r) over the active core elevations provides a more compact and concise representation of the data.
5.3 Uncertainties Tne uncertainties existing in the calculation of the normalizea axial, radial and azimuthal distributions are combined into the uncertainty quotec in the peak value that provides the magnitude of the fluence throughout the entire region of interest. It is estimated that an overall uncertainty Dand of +/-30% will cover the regions of significance, however the uncertainty may exceed +/-30% in regions of very low flux relative to the peak values.
5-6
5.2.4 Axial Distributions The normalized qxial distribution, f2 (s) of Equation 5.3 is obtainea from the same DOT-RZ calculations described in the discussion of the racial aistributions. The normalization' factor is, again, "P" in Equation 5.3. In tne axial direction the models rangeo from the mio-plane of the active core to the centerline of the reactor vessel nozzles. Symmetry about the core miaplane was assumeo. Truncated trapazoical power shapes that enveloped the reference plant time-averaged axial power distributions were used for the source distribution in the reactor core. The axial flux data from the DOT calculations was edited along the vessel-clad interface on the inner surface of the vessel and then normalized to form the axial distrioution, f2(-). It woula. nave been possiole to parameterize f2 (e) as a function of racius in orcer to approximate the non-separability of the flux distribution, however l(r), the radial aistribution was cnosen for this purpose because the relative invariance of fl(r) over the active core elevations provides a more compact and concise representation of the data.
5.3 Uncertainties Tne uncertainties existing in the calculation of the normalizec axial, radial and azimuthal distributions are combined into the uncertainty quotea in the peak value that provides the magnitude of the fluence throughout the entire region of interest. It is estimated that an overall uncertainty oand of +/-30% will cover the regions of significance, however the uncertainty may exceed +/-30% in regions of very low flux relative to the peak values.
5-6
4Flux-to-Fluence Conversion to a The flux distribution as represented in Equation 5.3 is converted fluence aistribution as given in Equation 5.1 by the choice of the limits of tne integral, tl. and .t2 ana the averaging of the flux distribution over that time interval. Typically t1 is <the time when the reactor for initially achieved' a critical state and t2 is tne specifiec time the fluence quotation. , In,this analysis tpe flux distributions were the calculated to correspond to an average up to December 31, 1981. Thus fluence is obtained by multiplying the normalization factor by the amount of time the plant has operated at a specified power level until December 31, 1981. The extrapolation of the fluence distributions to times beyond December 31, 1981, assumes that the "time averaged" flux distributions will remain the same in the intervening time pariod.
5-7
References:
5-1. Rhoades, W. A. and F. R. Mynatt, "The DOT III Two-Dimensional Discrete Ordinates Transport Code", ORNL-TM-4280, September, 1973, Oak Ridge National Laboratory, Oak Ridge, Tennessee.
5-2. SHADRAC - Shielding Heating and Dose Rate Attenuation Calculations, G30-1365, March, 1966, General Dynamics, Fort Worth, Texas.
5-8
MATERIAL PROPERTIES The integrity of the reactor vessel during a postulated system transient is a function of the localized stress intensity created by thattransient and the toughness of the vessel material in that region. Evaluation of the ma terial toughness properties requires information on the initial toughness, chemical composition,.and the operating temperature'and neutron exposure of the reactor vessel. Therefore, plant specific material property data were developed for use in the vessel integrity evaluations.
The,following subsections describe the analytical system and the assump tions used to develop the plant specific material property data which are provided in the Appendices. For the operating plants (excluding Arkansas Nuclear One, Unit 2), the neutron fluence and resultant material properties were determined for specific vessel locations. For ANO-2, Waterford 3, San Onofre 2&3,'and Palo Verde, 1,,2, & 3, the material properties were evaluate4 using peak neutron fluence and the most limiting vessel beltline material properties because these vessels were fabricated with low copper (nominally 0.10 w/o or less copper) materials.
The configuration of a typical reactor vessel designed by Combustion Engineering is shown schematically in Figure 6-1. Each vessel is fabricated using rectangular plates joined together using submerged arc weldments.
There are three shell courses in each vessel: the upper, intermediate, and lower shell courses. three formed plates are typically used for each shell course, each plate being joined to the next by a longitudinal (or vertical),
seam weld. The inlet and outlet nozzles are located in.the upper shell course.
(For System 80 vessels, the nozzles are located in the intermediate shell course and the core midplane is centered in the lower shell course.)
6.1 DATA ANALYSIS SYSTEM A computerized system was developed for determination of plant speci fic properties for the operating plants. The methodology used in this data analysis syste' is summarized in Figure 6-2. There are three data files for each plant. The first data file defines the location and dimen sions of the hot and cold leg (outlet and inlet) nozzles, plates. and welds 6-1
in the reactor vessel relative to the core mid-plane. "As-Built" drawings for each reactor vessel were used as the basis for establishing these files. The second file is a compilation of significant material properties including in formation on product form (plate or weldment), chemical composition (see Section 6.3), and initial toughness properties (reference temperature, RTNOT, as discussed in Section 6.4). The third file contains the neutron flux data for each plant including the peak neutron flux and the normalized flux pro files in the axial, azimuthal, and radial directions. The peak neutron flux values contained in the data file and used for radiation effects predictions are best estimate values accurate to within 30%. The upper bound (best es timate plus 30%) flux was not used because use of the upper bound flux would compound conservatisms already inherent in other facits of the analysis:
RTNDT values based on lower bound Charpy impact properties; shift predictions based on an upper bound enveloping curve; and lower bound stress intensity values conservatively referenced to Charpy based RTNDT data.
Included in the data analysis system is a subroutine for calculating the radiation induced chanqe in RTNDT. The calculational method is described in Section 6.2.
Material properties are defined using thfee key parameters as input to the data analysis system: the plant to be evaluated, the time period in effective full power years (EFPY), and the coordinates of the vessel locations to be evaluated (distance into the vessel wall, R; vertical distance from the nozzle centerline, Z; and aximuthal position, D). For each set of coordinates, data output includes the material identification at that location, initial RTNDT, neutron fluence adjusted to that location using the normalized flux profiles, RTNDT shift, ard adjuited RTNOT (the sum ofithe initial RTNDT and the shift).
- The adjusted RTNOT values are then used to calculate stress intensity factor.
curves (crack initiation and arrest toughness) according to the ASME Boiler and Pressure Vessel Code,Section XI, for subsequent comparison with the cal culated stress intensity factors for a system transient.
6.2 SHIFT PREDICTION METHODOLOGY The data analysis system includes a subroutine for predicting shift in the reference temperature as a function of product form (weld or plate),
' 6-2
chemical composition, and neutron fluence. The approach used is summarized in Figure 6-3. The objective of this approach is to represent the actual irradiation behavior of the vessel materials as accurately as possible.
The method used for all plate materials was Regulatory Guide 1.99, Revision 1(6-1). The validity of this approach is demonstrated in Figure 6-4 which is a comparison of the Regulatory Guide 1.99 predictions to actual post-irradiation surveillance measurements. These data are particularly valid because .they are a representative sample of the plates used in the fabrication of C-E designed reactor vessels. Furthermore, surveillance plates from five of the eight operating C-E reactor vessels are represented in the Figure as indicated in Table 6-1. For shifts in excess of 150aF, the predicted shifts are up to a factor of two higher than the measured shifts, indicating that projected properties for plate materials will be conservative for system transient analyses. This con servatism could have.been reduced by further effort, but was not since the welds are generally found to be more governing.
The shift prediction method used for weld materials accounts for the known effects of nickel content on radiation sensitivity(6-2). The re lationship between nickel content and weld metal RTNDT shift is illustrated in Figure 6-5. (The data presented in the Figure are summarized in Table 6-2.) All of the surveillance weldments with less than 0.30 w/o nickel can be seen to fall near the lower bound curve of Regulatory Guide 1.99, whereas the higher nickel weldments (>.30 w/o Ni) tend to exhibit signifi cantly greater radiation sensitivity, and fall nearer to the upper bound curve. However, both the high and low nickel data sets include high copper content welds; the range of copper for the low nickel set is 0.08 to 0.30 w/o copper and the range for the high nickel set is 0.05 to 0.36 w/o copper.
The effect of nickel content on predicted'shift is illustrated in Fiqure 6-6. The measured shifts for the high nickel welds .(only C-E sur veillance weldments are shown) are matched closely by the Regulatory Guide 1.99 predictions. In contrast, the low nickel weldment shifts (all avail able surveillance data) are over-predicted by a factor approaching two.
In order to account for the nickel effect, the following modified version of the Regulatory Guide 1.99 prediction formulation was used for the low nickel (less than 0.30 w/o) materials:
6-3
ARTNOT = [90 + 600 (Cu - .24)] 1 No credit is taken for copper less than 0.24 w/o and the Regulatory Guide 1.99 fluence relationship is utilized. A comparison between Regu latory Guide 1.99 and the modified expression for low nickel welds is given in Figure 6-7. The modified expression can be seen to provide an accurate, but still conservative, alternative to Regulatory Guide 1.99 for low nickel weld material.
6.3 MATERIAL CHEMISTRY In order to predict RTNDT shift, data wgre required on the copper, For the phosphorus, and nickel content of each of the vessel materials.
plate materials, full chemistry results were generally available for the six vessel beltline plates. For a given reactor vessel, the copper content of the nozzle (upper) shell course plates was estimated, using the maximum measured value from the six beltline plates.' The copper content for these of the plates is not especially critical because the neutron flux exposure nozzle course is about a factor of 100 lower than at the core mid-plane.
For the weld materials, full chemistry data for the as-deposited weld ments of the earliest vessels were available only for the surveillance weldment. Data are more complete for the newer vessels. The "Atypical Weld Report"(6-3) was' utilized in some instances where supplemental data were reported for the proper wire heat/flux lot combinations. The.copper content for weld seams fabricated entirely from coated electrodes (E-8018 lots C-3) was based on the upper bound (0.07 w/o copper) value for 44 manufactured by Combustion Engineering between 1970 and 1974. These data are reproduced in Table 6-3. The low copper content, 0.07 w/o, is reasonable because copper was never intentionally added to the wire or flux used to fibritate the coated electrodes. Furthermore, using the upper bound copper content of the 44 lots is conservative.
Copper contents for weld seams fabricated using the submerged arc process were based primarily on analyses'of the as-deposited surveillance weldments whenever specific results were not available. Upper bound values of copper (0.35 w/o) could be justified only for Maine Yankee and Ft. Calhoun; 6-4
weld copper contents trended downward subsequent to the fabrication of the Ft. Calhoun vessel, supporting the use of a best estimate weld copper con tent in the 0.20-0.30 w/o range for Calvert Cliffs, Millstone #2, and St.
Lucie #1. Subsequent vessels were fabricated with bare weld wire (ie,not copper coated) andthe weldments of these newer vessels contain 0.10 w/o or less copper.
6_4 INITIAL REFERENCE TEMPERATURE For the majority of the plate materials, Charpy impact tests were per formed using ldngitudinalTy oriented specimens. Therefore, the 50 ft-lb and 35 mils lateral expansion temperatures for the transverse orientation were estimated using paragraph B.1.1(3)(b) of Branch Technical Position MTEB 5-2 (reference 6-4). Surveillapce test results on transversely oriented Charpy specimens were used if available. The reference temperature, RTNDT was then determined in accordance with the ASME Code,Section III, NB-2300.
In cases where no drop weight tests were performed, use of the Branch Technical Position MTEB 5-2 to determine RTNDT for the welds was found to be excessively conservative. The lowest RTNDT possible using MTEB 5-2 is +100 F, the temperature at which three Charpy qualification specimens were tested (performed to verify that the wire heat, flux lot combination used for each vessel weld seam would exhibit 30 ft-lb impact.energy). The typical RTNDT for submerged arc weldments using Linde 0091, 1092, and 123 flux is generally -600 F.
This is demonstrated by data from 82 submerged arc weldments for which RTNDT was determined in accordange with the ASME Code. These data (described in References 6-3, 6-5, and surveillance baseline test reports covering the period 1965 to 1978) are summarized in Table 6-4. The mean value for the submerged arc weldments is -560 F with a standard deviation of 170 F and a range of -80aF to -100 F. Based dn this population of representative weld data, 97.5% of the RTNOT values could be expected to be -210 F or lower for welds fabricated using the same processes and types of welding materials. In contrast, using the MTEB 5-2 criteri.a would. necessitate assuming an RTNOT of +10aF or higher which is equivalent to a value 3.7 standard deviations above the population mean.
An alternate method for establishing realistic estimates of weld seam RTNDT was developed based upon the plant specific weld qualification test 6-5
results. The general method, illustrated in Figure 6-8, uses surveillance weld Charpy impact data to benchmark the plant specific weld qualification test results. If the average of the three qualification test impact energy 0
measurements was comparable to or better than the benchmark curve at +10 F, then the surveillance.weld RTNDT and the plant specific weld seam RTNOT were judged to be equivalent (Figure 6-8a). Conversely, if the average the impact energy of the weld qualification test results was less than benchmark data (Figure 6-8b), the weld seam RTNDT was increased by an amount the benchmark equivalent to the difference between the weld seam data and Charpy transition temperature curve. In effect, the temperature value at which 50 ft-lb or better exists was determined, and the RTNDT was es tablished at a temperature 60*F below that value. The plant specific weld qualification test results and resultant RTNOT values established each reactor vessel.
using this approach are described in the Appendices for 6-6
TABLE 6-1 Surveillance Plate Irradiation Data Froi C-E Designed Reactor Vessels Reactor Vessel Capsule Fluen e Chemical Content NDTT Shift Reference (n/cm ) Cu (w/o) Phos (w/o) (ACv 30)
Maine Yankee A-25 1.3 (E19) 0.15 0.013 120OF CR-75-317, 8/75 W-263 6.8 (E18) 0.15 0.013 930 F BCL-585-21, 12/80 A-35 8.8 (E19) 0.15 0.013 195 0 F WCAP-9875, 3/81 Palisades A-240 4.4 (E19) 0.25 0.011 205 0 F BCL-585-12, 3/79 Fort Calhoun W-225 6.1 (E18) 0.10 0.009 60aF TR-0-MCM-001-1, 8/80 Calvert Cliffs #1 W-263 6.0 (E18) 0.12 0.011 60aF BCL Report, 12/80
.960F TR-N-MCM-008 (Draft), 1981 Millstone #2 W-97 3.7 (E18) 0.14 0.006 0
TABLE 6-1 Surveillance Plate Irradiation Data From C-E Designed Reactor Vessels Capsule Fluen e Chemical Content NDTT Shift Reference Reactor Vessel 4n/cm ) Cu (w/o) Phos (w/o) (ACv 30)
A-25 E.3 (E19) 0.15 0.013 120oF CR-75-317,-8/75.
Maine Yankee W-263- 6.-8 (E13) 0.15 0.013 .930 F BCL-585-21, 12/80 A-35 8.8 (E19) 0.15 0.013 195 0 F WCAP-9875, 3/81 Palisades A-240 4.4 (E19) 0.25 0.011 205 0 F BCL-585-12, 3/79 W-225 5.1 (E18) 0.10 0.009 60aF TR-O-MCM-001-1, 8/80 Fort Calhoun 6.0 (E18) 0.12 0.011 60aF BCL Report, 12/80 Calvert Cliffs #1 W-263 W-97 3.7 (E18) 0.14 0.006 960F TR-N-MCM-008 (Draft), 1981 Millstone #2 00~
- Y6-2 Surveillance Weld Irradiation Data Reactor Vessel- Capsule Fluen e Chemical Content NDTT Shift Reference (n/cm ) Cu (w/o) Phos (w/o) Ni (w/o) (AC 30)
ANO-1 E 7.3 (E17) 0.28 0.016 0.59 115 0 F BAW-1440, Apr. 1977 Surry-1 T 2.5 (E18) 0.25 0.011 0.63 165 0 F Docket 50280-462, 6/24/75 Ft. Calhoun W-225 5.1 (E18) 0.35 0.013 0.60 233 0 F TR-0-ICM-001-1, 8/80 Dresden-3 - .2 (E13) 0.35 0.013 0.73 205 0 F Capsule T3&14, 3/1/75 Maine Yankee A-25 1.3 ([19) 0.36 0.015 0.73 270oF CR-75-317, 3/75 W-263 7.0, (E18) " " 2220 F BCL-585-21, 12/80 A-35 8.3 (E19) " " 345?F WCAP-9375, 3/31 Turkey Pt.-4 T 8.6 (E13) 0.30 0.014 0.60 225 0F SWRI-02-4221, 6/14/76 Palisades A-240 4.6 (E19) 0.25 0.011 0.43 350 0 F BCL-585-12, 3/79 Zion-2 U 2.0 ([18) 0.28 0.017 0.55 128 0 F BCL-535-4, 3/78 TMI-1 E 1.1 (E13) 0.34 0.019 0.71 750 F BAW-1439, 1/77 Browns Ferry 19 1.8 (E13) 0.27 0.014 0.10 55F Nuc. Eng. & Des. II, p 393, 1970 Zion-1 L 1.8 JE18) 0.35 0.020 0.57 101 0 F BCL-535-4, 3/78 Indian Pt.-3 T 2.9 ([13) 0.15 0.019 1.02 330 F WCAP-9491, 4/79 Conn. Yankee A 2.9 ([13) 0.22 0.020 0.05 950 F BCL-50-213, 10/30/70 Pt. Beach-1 V 3.6 ([13) 0.18 0.019 0.57 110OF BCL, 6/15/73 Ginna V 6.9 ([13) 0.23 0.012 0.56 140aF FP-RA-1, 4/73
[1 1.3 ([19) "" 165 0 F WCAP-3421, 11/75 Turkey Pt.-3 T 6.4 (E13) 0.31 0.011 0.57 155 0 F WCAP-8631, 12/75 Quad Cities-1 12 7.2 ([13) 0.31 0.010 0.65 155oF Capsule 2, Basket 3, 3/1/75 Dresden-3 - 7.7 (E18) 0.20 0.011 0.30 155 0 F Capsule Basket 12&14, 3/1/75 Pt. Beach-? V 7.7 (E13) 0.25 0.014 0.59 165OF BCL, 6/10/75 R 2.0 (E19) " " 230oF WCAP-9635, 12/79 T 1.0 ([19) " " " 145 0 F Docket 50-266, 8/73
TABLE 6-2 (Cont'd)
Reactor Vessel Capsule Fluen e Chemical Content NDTT Shift Reference (n/cm ) Cu (w/o) Phos (w/o) Ni (w/o) (ACV_30)
Kewanee V 7.2 (E13) 0.20 0.016 0.77 175*F WCAP-8903, 1/77 Pt. Beach-1 S 9.5 (E1S) 0.13 '0.019 0.57 165 0 F WCAP-3739, 11/76 P 2.7 (E9) 1 " I WCAP-9357, 3/73 fluad Cities-? 13 9.0 (E18) 0.26 0.012 0.60 180OF Surv. Rep., 9/75 Surry-2 x 3.0 (E13) 0.19 0.017 0.56 95OF BCL, 9/2/75 Millstone-2 W-97 3.8 (E18) 0.30 0.015 0.06 760F TR-N-MCM-003 (Draft), 1981 4-97 3.8 (E18) 0.21 0.016 0.06 370F TR-N-MCM-003 (Draft), 1931 Lacrosse 4.4 (E13) 0.13 0.016 0.12 60OF SWRI-02-3467 Trojan U 3.8 (E18) 0.05 0.023 0.93 220 F WCAP-9469, 5/79 Calvert Cliffs-1 W-:63 6.1 (E18) 0.24 0.014 0.13 590 F BCL Report, 12/80 0
Prairie Is. -1 V 6.3 (E18) 0.13 0.017 0.07 25 F WCAP-8916, 3/77 Prairie Is. -2 V 7.2 4E18) 0.08 0.019 0.07 60oF WCAP-9212, 11/77 Quad Cities-1 P1 8.9.(E13) 0.17 0.011 0.28 300 F Surv. Rep., 9/75 0
nuad Cities-? 13 1.3 (019) 0.13 0.011 0.20 95 F Surv. Rep., 9/75
TABLE 6-3 Chemical Analysis of E-3018 C-3 Coated Electrodes (Combustion Enqineerinq Manufacture)
Size Lot Chemistry (w/o) Year
$' N --- Cu Manufactured 5/32 GBCJ .36 .003 1.07 .02 1970 1/4 HOCJ .37 .010 .93 .02 1/4 IOBJ .45 .010 .97 .02 1/4 ICJJ .42 .011 .99 .q3 1/4 JADJ .43 .009 .96 .03 3/16 JACJ-1 .32. .003 .97 .04 1/4 KOIJ .34 .010 1.0 .03 5/32 KAHJ .42 .011 1.03 .03 1/4 KBEJ .57 .011 1.04 .03 1/4 LAGJ .35 .010 1.0 .03 3/16 ABEA .36' .012 .93 .04 1971 1/4 8OIA .40 .003 .93 .02 1/4 BOLA .41 .010 .93 .02 1/4 CCJA .33 .007 .36 .02 1/4 FOAA .40 .003 1.0 .03 5/32 FOCA .39 .003 .93 .02 1/4 FAGA .33 .003 .95 '.03 1/4 FBHA .36 .006 .36 .03 1/4 HODA .038 .009 .96 .02 1/4 IAGA .34 .009 .98 .03 3/16 JBHA .50 .003 .97 .03 1/4 KOHA .44 .007 .93 .03 1/4 JOGA .43 .007 .94 .03 5/32 CAFB .39 .007 .99 .04 1972 3/16 EAIB .40 .007 .93 .02 1/3 FAB .32 .008 1 .06 .02 5/32 HAG3 .33 .006 1.04 2 1/4 KOlB .41 .006 .97 .03 3/16 LOHB .44 .008 1.03 .03 5/32 JAHB .43 .007 .97 ,.03 6-11
TABLE 6-3 (Cont'd)
Size Lot Chemistry (w/o) Year i P Ni Cu Manufactured 1/4 AAGC .3S .006 .93 .03 1973 1/4 GCJC; .37 .003 .99 .02 1/4 EOBC .39 .007 .96 .02 3/16 FFAFC .40 .006 1.04 .07 1/4 GAAHC .31 .008 1.05 .03 1/4 HAAEC .29 .006 1.03 .03 1/4 GACJC .39 .005 l.CO .03 1/4 HABJC .33 .005 1.02 .02 5/32 LAAHC .29 .003 .92 .02 3/16, . BAED .30 .007 1.00 .02 1974 1/4 BABBD .40 .006 1.04 .02 1/4 AACJO .37 .004 .97 .02 I 1/8 DABED .33 .007 .87 .03 1/4 A0Eb .45 .006 .93 .03 6-12
TABLE 6-4 Initial Reference Temperature for Representative Submerged Arc Weldments Drop Weight RTNDT Material Identification NDTT (oF) (OF)
Wire Ht. #51912, Linde 0091 Flux -50 -50 Wire Ht. #83640, Linde 0091 Flux -70 -70 Wire Ht. #83642, Linde 0091 Flux -80 -80 Wire Ht. #83653, Linde 0091 Flux -80 -80 Wire Ht. #83648, Linde 0091 Flux -80 -80 Wire Ht. #4P5174, Linde 0091 Flux -50 -50 Wire Ht. #83637, Linde 0091 Flux -50 -50 Wire Ht. #83650, Linde 0091 Flux -40 -40 Wire Ht. #5P5622, Linde 0091 Flux -40 -40 Wire Ht. #83646, Linde 0091 Flux -40 -40 Wire Ht. #2P5755, Linde.0091 Flux -50 -50 Wire Ht. #4P6052, Linde 0091 Flux -50 -50 Wire Ht. #87005, Linde 0091 Flux -10 -10 Wire Ht. #87600, Linde 0091 Flux -70 -70 Wire Ht. #88118, Linde 0091 Flux -70 -70 Wire Ht. #4P6524, Linde 0091 Flux -70 -70 Wire Ht. #87011, Linde 0091 Flux, -50 -50 Wire Ht. #86998, Linde 0091 Flux -10 -10 Wire Ht. #88112, Linde 0091 Flux -70 -70' Wire Ht. #88114, Linde 0091 Flux -70 -70 Wire Ht. #87000, Linde 0091 Flux -40 -40 Wire Ht. #V89476, Linde 0091 Flux -50 -50 Wire Ht. #30502, Linde 0091 Flux -50 -50 Wire Ht. #4P6519, Linde 0091 Flux -60 -60 Wire Ht. #90077, Linde 0091 Flux -60 60 Wire Ht. #90128, Linde 0091 Flux -60 0 Wire Ht. #90069, Linde 0091 Flux -60 -60 Wire Ht. #90146, Linde 0091 Flux -50 -50 Wire Ht. #90130, Linde 0091 Flux -60 -60 Wire Ht. #90211, Linde 0091 Flux -60 -60 Wire Ht. #90149, Linde 0091 Flux -60 -60 Wire Ht. #87003, Linde 0091 Flux -20 -20 Wire Ht. #90157, Linde 0091 Flux -60 -60 Wire Ht. #90132, Linde 0091 Flux -60 -60 Wire Ht. #90209, Linde 0091 Flux -50 -50 Wire Ht. #89204, Linde 0091 Flux -60 -60 Wire Ht. #SP7388, Linde 0091 Flux -30 -30 Wire Ht. #3P7150, Linde 0091 Flux -30 -30 Wire Ht. #89833, Linde 0091 Flux -50 -50 Wire Ht. #90159, Linde 0091 Flux -40 -40 Wire Ht. #90067, Linde 0091 Flux -70 -70 Wire Ht. #90154, Linde 0091 Flux -40 -40 Wire Ht. #89827, Linde 0091 Flux -80 -80 Wire Ht. #89823, Linde 0091 Flux -50 -50 Wire Ht. #4P7656, Linde 0091 Flux -70 -70 6-13
TABLE 6-4 (Cont'd)
Drop Weight RTNOT Material Identification NOTT (OF) e Wire Ht. #89022, Linde 0091 Flux -30 -30 Wire Ht. #90071, Linde 0091 Flux -80 -30 Wire Ht. #87603, Linde 0091 Flux -60 -60 Wire Ht. #90069, Linde 0091 Flux -70 -70 Wire Ht. #89408, Linde 124 Flux -60 -60 Wire Ht. #3P7246, Linde 124 Flux -60 -60 Wire Ht. #3P7317, Linde 124 Flux -so -s0 Wire Ht. #E56906, Linde 124 Flux -30 -30 Wire Ht. #651A708, Linde 124 Flux -80 -80 Wire Ht. #91764, Linde 124 Flux -60 -60 Wire Ht. #91762, Linde 124 Flux -50 -50 Wire Ht. #4P7927, Linde 124 Flux -80 -0 Wire Ht. #4P7869, Linde 124 Flux -70 -70 Wire Ht. 45PS366, Linde 124 Flux -80 -30 Wire Ht. #69025, Linde 124 Flux -70 -70 Wire Ht. #89833, Linde 124 Flux -60 -60 Wire Ht. #90144, Linde 124 Flux -50 -50 Wire Ht. #3P7802, Linde 124 Flux -30 -80 Wire Ht. #5P9028, Linde 124 Flux -80 -50 Wire Ht. #3P8013, Linde 124 Flux -80 -50 Wire Ht. #4P8632, Linde 124 Flux -30 -30 Wire Ht. #3PS013, Linde 124 Flux -60 -60 Wire Ht. #5P9744,' Linde 124 Flux -30 -60 Wire Ht. #LP5P9744, Linde 124 .Flux -60 -60 Calvert Cliffs #1 Surveillance Weld -30 Calvert Cliffs #2 Surveillance Weld -60 -60 Fort Calhoun Surveillance Weld -50 -50 Millstone #2 Surveillance Weld -60 -55 Forked River Surveillance Weld -70 -70 SONGS #2 Surveillance Weld -50 -50 SONGS #3 Surveillance Weld -60 -34 Waterford #3 Surveillance Weld -80 -s0 ANO-2 Surveillance Weld -10 -10 Maine Yankee Surveillance Weld -30 -30 Weld-2, L6w Cbpper Program (Ref. 6-5) -50 -50
-70 -35 Ie1d-3, Low Copper Program (Ref. 6-5) -30 -70 Weld-4, Low Cooper Program (Ref. 6-5) 6-14
FIGURE6-1 TYPICAL C-E REACTOR PRESSURE VESSEL Inlet Outlet Nozzle Nozzle Upper Shell Course tongitudinal Intermediate Circumferential Seam-.Meld Shell Course Weld Lower Shell
---Course. ---------- --
80 360 10 AZ1MiUTHAL L0CATIrON CEGREES
FIGURE 6-2 REACTOR VESSEL PRESSURIZED THERMAL SHOCK METHODOLOGY - MATERIALS EVALUATION PLANT-SPECIFIC INPUT DATA FILE - - VESSEL ID MATERIAL LOCATION - TIME PERIOD (EFPY)
VESSEL MAP WITH LOCATIONS - LOCATIONS IN VESSEL OF CRITICAL PLATES AND WELDS MATERIAL PROPERTIES OUTPUT PRODUCT FORM, CHEMISTRY AND or CRACK INITIATION AND SIINITIAL TOUGHNESS ANALYSIS ARREST TOUGHNESS AT NEUTRON EXPOSURE CRITICAL LOCATIONS IN PEAK NEUTRON FLUX AND VESSEL FOR COMPARISON NORMALIZED FLUX PROFILE WITH CALCULATED STRESS IN R,Z,0 GEOMETRY INTENSITY FACTORS RADIATION-1INDUCED SHIFT PREDICTION METHOD
FIGU'-3 TRANSITION TEMPERATURE SHIFT METHODOLOGY FOR ALL PLATE-MATERIALS
- REGULATORY GUIDE 1.99 (REV. 1):
ART NDT =[(110 + 1000 (Cu - .08) + 5000 (P - .008)
FOR ALL WELD MATERIALS REGULATORY GUIDE 1.99 (REV. 1) IF NICKEL CONTENT IS MORE THAN 0.30%
- MODIFIED REGULATORY GUIDE 1.99 IF NICKEL CONTENT LESS THAN 0.30%: 0 5 ART NDT [(90 + 600 (Cu -. 2)(8TY
300 FIGURE-6-4 COMPARISON OF RG 1.99 SHIFT PREDICTIONS WITH C-E SURVEILLANCE PLATE IRRADIATION 250 DATA PAL 200 3 M P -2 200 CAL MP-2I PLC 150 " MSMI' <
50 00I PRDITE SHFT(o PRDITE SHIF (-F
350 @ '
@350 -- - ~FIGURE 6-5 7j~
RELATION BETWEEN NICKEL. 1I CONTENT AND WELD METAL SHIFT I .
300 300 (SURVEILLANCE DATA ONLY) .
NIC EL NTEN~T..
250 .3fL FIGURE 6 LA..
200 RELATIONR BEWE1NCE CONTENT~~L ANeEDMEArHF I1I D IBoundn 50 1 A INet on 1luenc 2 thij'on 2 E> Me 0
1 1018 50 i 1019 KI..Ii RG1g 1020 50 7 K~~B-Bound
-- ~~Z ~ ~ ~ _n ~ 7 7--c2~E~MV_
100--*AI
.01 46151 FIGURE 6-6 COMPARISON OF REGULATORY P PAl1 GUIDE 1.99 SHIFT PREDICTIONS ....
WITH SURVEILLANCE WELD IRRADIATION DATA
.. ... .. T CL r :MY. . .: fT my 7
I.
- j
- ; .. .... ....................................................
10 . .................. ... ... . . . . . . . . .
-if .- ...... .. . . . .
w.
i::iAiI 1!_.... .. ... . . ....
. .A .. .... .. Iw
FIGURE 6-7 COMPARISON OF PREDICTION
-146 METHODS FOR LOW4 NICKEL WELDS eaurec = Pedic et I---
(b I
- O D :9!
.-- -.- - - q q: - ...... to...
I-.77'- H" G...U.ATOY.J.
. .::- I ! ; ! -- - .: -L -- . - - - .- 7- - LK 2E) C'E) H 7 ... ... . / .....
-7 0-~~~~vb 4-71tTL
~ A> hIG i Y GU/
)NIDj
FIGURE 6-8 Determination of RTNDT From Weld Qualification Test Results a) RTNOT < value for surveillance weld Surveillance 12 O Weld Z5. X Weld Qualification Results LU4
< Drop Weight C' NDTT GE (-60aF)
+10aF (Qualification Test Temperature) 0
. ) -40 0 40 80 120 TEST TEMPERATURE, a F b) RTNDT>value for Surveillance 120 surveillance weld Weld Adjusted Transition Temperature Curve C-d LUJ uJ X Weld Qualification Results a40.
-L Drop Weight N060aF),
+10aF Test Temperature
-40 0 40 8 120 TEST TEMPERATUREo 0 6-22
Chapter 7 Vessel Integrity Evaluations 7.1 Introduction The overcooling temperature/pressure transients described in Section 4 cause tensile stresses at the inner surface of the reactor vessel wall.
If flaws in the material exist, the tensile stress could tend to open and extend such flaws,' The integrity of a reactor pressure vessel is evaluated by hypothesizing that flaws or cracks are present and by determining the effect of the stresses on the extension of the hypothetical cracks.
If all credible initial cracks are found to remain stable, or if some limited extension occurs followed by no further extension, then the integrity of the vessel is assured.
The stresses resulting from the transients of Section 4 are less than yield except for local regions on the surface of the vessel wall or at the hypothetical crack tip. For conditions of limited local plasticity, such as this, linear elastic fracture mechanics has been demonstrated tb provide a sound and reliable method for evaluating the integrity of steel pressure vessels (Reference 7.1).
This section describes the linear elastic fracture mechanics (LEFM) evaluation of the effect of the overcooling transients on the integrity of the CE reactor vessels. Plant specific results are reported in the appendices.
7.2 Linear Elastic Fracture Mechanics Linear elastic fracture mechanics is based on elastic analysis and the concept of the stress intensity factor, K,, which characterizes the sitress field near the tip of a crack. The stress intensity factor is dependent on the geometry of the structure, the geometry of the crack and the loading (stresses) in the structure.
There are three basic principles of LEFM. They address crack initiation, crack blunting (or warm prestress) and crack arrest.
7-1
7.2.1 Crack Initiation Crack initiation occurs when the stress intensity factor, KI, exceeds a critical value called KIC. KIC is dependent on the material, the temperature and the irradiation fluence. This principle has been confirmed by experiments in which cracked specimens and components have been loaded to failure. (Reference 7.2) 7.2.2 Crack Blunting (Warm Prestress)
Recent experiments have demonstrated an additional requirement for crack initiation (Reference 7.3). This requirement is that KI must be increasing in time ;to cause crack initiation, that is, crack initiation will not occur during unloading. If KI becomes greater than KIC during unloading, then KIC must be decreasing from a more ductile state (for example, from a higher stresses temperature). This unloading produces compressive residual in the crack tip region which had previously yielded at the higher KI value. 'These compressive stresses prevent extension of the crack during unloading.
can also Experiments also demonstrate that some subsequent reloading take place prior to crack initiation (Reference 7.4). The amount of reloading permissible however, is not yet adequately quantified. *Therefore the the additional capability of this effect is conservatively neglected in fracture analysis reported here.
7.2.3 Crack Arrest A running crack will stop running (arrest) if the stress intensity factor KI falls below the arrest material toughness KIa. The concept of crack arrest has been verified by experiments on specimens and from components (Reference 7.5). The value of Kla has been determined irradiation dynamic experfments to be dependent on material, temperature, fluence, and to generally be somewhat less than KIC.
7.3 Description of Methods shock The determination of vessel integrity during a pressurized thermal and event requires a calculation of the time variation of temperature of thermal stress in the vessel wall during the transient. The combination the vessel if the and.pressure stresses could threaten the integrity of 7-2
level of such stresses were to be concentrated on a flawed area of the vessel, and if the properties of the vessel in the flawed region were degraded significantly by irradiation. Vessel integrity must be preserved to the extent that no flaw could propagate through the vessel wall, thus assuring pressure boundary integrity.
A vessel integrity analysis must, therefore, consider the cumulative effects of the thermal and pressure loads on the tendency for growth, and the material's resistance to growth, of a pre-existing crack in the structure. The procedure for evaluating the effect of a particular transient on reactor vessel integrity involves performing four basic calculations: (1) a thermal analysis that provides the temperature d stribution through the wall as a function of time; (2) a set of stress analyses of a cracked cylinder, considering both thermal and pressure loadings, from which the crack tip stress intensity, K1 , is computed as a function of time for various crack depths; (3) computations of K and Kla as functions of crack depth and time based on the accumulated fluence, the vessel material properties, and the temperature history; and (4) a determination of the crack depths for which K = KIC and K = KIa at various times during the transient, for construction of critical crack depth curves. Calculations (1) and (2) need only be performed once for a particular transient since these are solely a function of the loading-histories. Calculations (3) and (4), related to the material dependent parameters, are performed repeatedly for various different plant specific material properties, and plant life fluence levels. Then, based on an acceptance criterion permitting no crack initiation, or initiation with assured arrest, a given transient with a specific set of vessel material properties will enable the evaluation of the integrity of the vessel at various times in the design life.
In evaluating these results, it is recognized that warm-prestressing is an effective means of precluding crack initiation when the net stress intensity factor is decreasing with time. This may, in fact, permit a significant amount of repressurization following a thermal shock depehding on the severity of the thermal transient.
7.4 Thermal Analysis A combined conduction/convection heat transfer analysis is performed to determine the temperature history through the reactor vessel wall for 7-3
each transient. The vessel is modeled as an axisymmetric cylindrical structure as shown in Figure 7.1. Temperatures in the wall are assumed to be only a function of radial distance, r, and time, t.
Variations in temperature along the length of the wall are handled separately. This enables a one-dimensional heat transfer analysis to be performed to solve the axisymmetric heat conduction equation Tr ar a '
where K - thermal conductivity P density Cp - specific heat T = T (r,t) = temperature in the wall r = radial distance t = time Typical dimensions of the vessel are as follows:
Vessel inner radius 86. in.
Vessel outer radius 94.625 in.
Inner wall cladding thickness .21875 in.
Slight differences 'invessel geometries among the various plants considered do not significantly affect the thermal or stress analysis results.
The outer surface of the vessel is insulated, and the forced convection of the inner surface is given by
= h (Tf - Ts) where = surface heat flux h = convective film coefficient
,Tf= temperature of fluid at wa.l Ts= temperature at the surface of the wall A value of 300 Btu/hr-ft 2 0F was determined from (Reference 7.6) for i \
the film coefficie't, h, due to natural circulation cooling of the vessel. This is'not a sensitive parameter since the dominant heat transfer mechanism is conduction-limitedithrough the vessel wall.
The effects of mixing are included in the temperature, Tf, which is determined as a function of time, t, and axial distance below the nozzle as discussed in Section 4. Therefore, three separate heat transfer analyses are performed at different sections along the axial length of
the vessel whic4 are:
- 1) middle of core
- 2) top of core
- 3) halfway between top of core and inlet nozzle as shown in Figure 7.2.
Temperature around the circumference is assumed to be uniform. A 2- heat transfer analysis case was performed as discussed in Section 7.9 to evaluate this assumption.
The transient heat conduction problem is solved using the MARC finite element computer program (Ref. 7.7). The effect of the cladding is included as a separate material in the heat transfer analysis, and the properties of the base metal (SA-533 Gr.B) and cladding (304 S.S.) which were used in the analysis are given in Table 7.1.
Table 7.1 Material Properties for Thermal Analysis SA-533 Gr.B Carbon Steel Thermal conductivity, K 2.18 Btu/hr-in-OF Specific heat, Cp .12 Btu/lb-oF Density, P .283 lb/in 3 Type 304 Stainless Steel Thermal conductivity, K .78 Btu/hr-in-oF Specific heat, Cp .127 Btu/lb-oF Density, P .2865 lb/in 3 A uniform initial temperature of 550 0F was assumed, and the transient analysis was performed 6sing integration time increments of 1 minute.
Typical results of the thermal analysis are shown in Figure 7.3.
7.5 Stress Analysis The stress analysis involves evaluating the effects of combined thermal and pressure loading on a vessel containing an inside surface flaw.
The analysis is performed for a range of assumed crack depths for each specific transient. Using the method of Linear Elastic Fracture Mechanics, the crack tip stress intensity, KI, is computed as a 7-5
function of time for each crack depth. To be conservative, the vessel is assumed to be a long cylinder with an infinitely long axial inner surface flaw. This permits a two-dimensional analysis to be performed on a cross-section of the vessel at a particular axial location. A finite element model of the vessel containing an assumed crack is shown in Figure 7.4. Because of symmetry, ohly one-half of the vessel was modeled for the stress analysis. Seven separate models were used, each having a different crack depth. Cracks are assumed to be through the cladding into the base metal. The depth of the cracks in the base metal used in-the analysis are given in Table 7.2.
Table 7.2 Crack Depths Used for Stress Analysis 1.0 1.5 2.0 4.3 6.5 7.6 Depth, a(inches).5 Fractional
.06 .12 .17 .23 .5 .75 .88 Depth, a/w The stress state in the vessel is calculated elastically using the MARC finite element program. The radial temperature distribution is applied to the stress analysis model at specific times in the transient as determined from the thermal analysis. A pressure load is appl Imid to the inside vessel surface consistent with the system pressure at the on the particular transient time. Pressure loading was also applied surface of the crack. The thermal stresses develop due to non-uniform 0
cooling from the 'initial state of 550 F, and because of the differences in thermal expansion between the vessel and the cladding. The elastic properties of.the vessel and cladding are considered to be independent of temperature with conservative values chosen for the temperature range of interest as shown in Table 7.3.
Table 7.3 Elastic Material Properties SA-533 Gr.8 304 Stainiess Steel Elastic Mpdulus, E 28.3 x 106 psi 27.9 x 106 psi Thermal Expansion -6 Coefficient, 7.54 x 10 6 in/in/oF 9.96 x 10 in/in/
Poisson's Ratio,' .3 .3 For a given combination of pressure and 'rapid cooling of the vessel, the presence of an axial crack produces a non-uniform bending, or kinking, in the region of t'he flaw. This causes an additional concentration of
stress near the tipof the crack indicated by the deformed shape plot given in Figure 7.5.
The stress intensity at the crack tip is computed from the J-Integral tqchnique. J is computed for each crack depth as a function of time.
The value of J is then related to the stress'intensity factor, K1 ,
knowing that J is equivalent to the strain energy release rate. G, which is defined for Linear Elastic Fracture Mechanics (Reference 7.8).
J = G = KI - E- (for plane strain)
Therefore, KI is given by KI From this, KI is determined as a function of time for each crack depth as shown in Figure 7.6 for a typical temperature/pressure transient.
For each time in the transient,KI is also known as a function of crack depth, a/w, as shown in Figure 7.7.
A separate analysis is performed for three different sections along the axial length of the vessel consistent with the thermal'analysis. These are (1) at the middle of core, (2) at the top of'core, and (3) halfway between the top of core and the inlet nozzle. The stress analyses described here are performed only once for each pressure/temperature transient at each location since the stresses are solely a function of the loading history.,
7.6 Calculation of KIC and KIa The material toughnesg is a function of temperature and nil-ductility reference temperature, RTNDT. Since both temperature and RTNDT vary in three-dimensions, a determination of the worst locations in the vessel is 6 slective proces% based on evaluating the matoril properties diataIi for each specific plantl This incorporates the accumulation of fluent:o with Effective Full Power Years (EFPY), the variation in fluence in the azimuthal, axial, and radial direction, and the variation in chemical content and initial RTNDT.
The effect of each of these parameters is included in the calculation of the 7-7
adjusted RTNDT as discussed in Section 6. Thys, for a given set of the following parameters:
- 1) Plant
- 2) Effective Full Power Years (actual or projected)
- 3) Azimuthal Angle,
- 4) Axial Height, Z the method of Section 6: is used to determine RT NOT as a function of wall depth.
The wall temperature is also calculated as a function of wall depth (or radius) and time for a particular transient. This enables KIC and KlA to be determined using the curves from ASME Section XI, Appendix A as shown in Figure 7.8. These curves are represented by the equations:
KIC (T') - 33.2 + 2.806 . exp .020 (T'+ 100) (ksi 'in) 5 200 ksi i n and Kia (T') = 26.8 + 1.223 . exp [.0145'(T + 160)] (ksi -T) 5 200 ksi -TF where T' = T - RTNOT T = T (r,t) = temperature from thermal analysis RT =RT A RT NO, NDT NT o + NOT S~oth XKI and K Iaare presumed to have upper shelf values of 200 ksi in7.Beyond this valueflinear elastic fracture mechanics is assumed to be no longer applicab1 For a specific time, t, the material toughness values for KIC and Kla are determined as a function of wall depth, a/w, as shown in Figure 7.9. On a plant specific basis, this procedure is repeated for the three different axial locations at (1) the middle of core (2) the top of core, and (3) halfway between the inlet nozzle and the top of core. Similarly, it is repeated for each transient being considered, for a large number of times during the transient Finally, the entire process is considered repeatedly for increasing plant life in terms of accumulated fluence with EFPY. Plant specific results of this process are reported in the appendices.
7.7 Critical'Crack Depth Curves The results of the stress analysis are compared with the initiation and 7-8
arrest toughness curves to determine the critical crack depth diagram.
At selected points in time during a transient, the KI vs. crack depth curves from the stress analysis Lre compared with the K and KIa vS.
depth curves representing material toughness. Figure 7.10 shows an example of the superposition of these curves at a given time, t. The critical crack depths for initiation are determined from the intersections where KI KIC. Similarly, the critical crack depths for arrest are located at the points where KI
- KIa. By evaluating a large number of these combined plots at many times during a transient, a single critical crack depth curve is constructed as shown in Figure 7.11. This curve represents the effect of a given transient on a specific vessel. A different set of curves is calculated for the same vessel having a higher accumulated fluence. Thus, the analysis is repeated on a plant specific basis with increasing EFPY. The plant specific results are presented in the Appendices.
7.8 Evaluation of Pressure-Temperature Transients The fluid system transients described in Section 4 were screened in order to select the most severe pressure thermal shock cases for detailed analysis.
Two categories of'system transients were evaluated:
- 2. SBLOCA + LOFW - 0. Restore Feedwater The transients within each category were compared in terms of their potential for challenging vessel integrity. Transients which did not have a significant thermal shock, regardless of the pressure level were judged to be of less severity. Similarly, transients with a rapid4thermal cooldown, but a depressurization without a further rise in pressure, were also characterized to be less severe. Thus , the most severe transient in each category was chosen as the representative case for the fracture mechanics analysis.
A separate analysis was performed for Maine Yankee because of plant specific differences in system parameters., Oily one case was evaluated for Maine Yankee, that of SBLOCA + LOFW followed by a restoration of feedwater.
7.8,1 SBLOCA + LOF Open PORV's Case 4 was judged to be the most severe thermal hydraulic transient of the four cases in this group (Cases 1 through 4). The temperature 7-0
and pressure curves used are shown in Figures 7.12 and 7.13 . These curves are an early approximation to the final curves included in Section 4. The slight increase in the severity of the final calculated transient in Section 4 is not sufficient to modify the stress analysis results on the integrity evaluation. This transient corresponds to a break size of 1 PORV in area, with a second PORY being opened at 10 minutes and the maximum HPSI injection rate. The temperature of the fluid adjacent to the wall showed a drop to about 300o F within 400 seconds at the levels above the core and at the top of core. Because of mixing in the downcomer, the temperature at the mid-core level was qalculated to only drop to 4250F during the same time period. The pressure transient indicates an initial depressurigation to 1000 psi, a slight repressurization to 1200 psi, followed by a slower depressurization after opening the second PORV.
The thermal response,of the reactor vessel, wall was calculated for this transient at the three different axial levels. Following this, a set of stress analyses of a flawed vessel were performed using as applied loads the pressure and calculated temperature distribution as a function of time. The stresses in the vessel and the stress intensity factors, Ki, were computed for the seven different assumed.crack depths. The results of the stress analysis for this transient are given as KI vs. time for each of the axial hqight locations in the vessel. Figure 7.14 shows the effect of this thermal hydraulic transient on a crack located at the mid-core level of the vessel. Similarly, Figure 7.15 is for the top of core location, and Figure 7.16 shows the results for a location midway between the top of core and the inlet nozzle. Each of the K, vs. time plots exhibit a similar trend with the stress Intensity decreasing within the first 5 minutes of the transient, then increasing to a peak value at 10 minutes, followed by a gradually decreasing KI for the remainder of the transient. From this it is clear that warm prestressing would preclude crack initiation under these cond4tions beyond a transient time of 10 minutes. The rate of decrease of K becomes very small, and the K, value even fluctuates slightly about a constant value for times late in the transient. Warm prestress is known to be in effect until a significant KI increase occurs (Reference 7.9). Minor fluctuations about a constant value of Ki, therefore, do not cause a violation of the warm prestress condition . The effect of this loading condition on vessel integrity is discussed on a plant specific basis in the Appendices.
.2 SBLOCA + LOFW Restore Feedwater Of the three cases in this category (Cases 5,7, 'and 8), Case 5 is judged to be the most severe because of the fairly rapid cooldown and significant repressurization up to the safety valve setpoint of 2400 psi. The remaining case, Case 6,.is discussed in Section 7.8.3.
Case 5 is characterized by a very small break LOCA plus loss of feedwater, with a restoration of feedwater at 30 minutes. The temperature and pressure curves used are taken from Section 4. This transient is more severe than Case 4 because of the repressurization to 2400 psi. The fluid temperature at the top of core and above the core drops to about 2300F for a brief period at 4400 seconds. At the same time, a pressure spike occurs.
The thermal behavior of the vessel in response to this transient was determined at the three axial heights and the temperature distribution was calculated as a function of time. The temperatures and pressures were then applied to the stress analysis model of the vessel containing a crack to calculate the resulting strdss intensity factors. Thi produced the K( vs. time plot at the middle of core location shown in Figure 7.17 Likewise, the top of core results are given in Figure 7.18, and the above core results are presented in Figure 7.19. At the mid-core location the stress intensity for all crack depths is fairly constant up to a time of 30 minutes when the pulse in pressure appears as a slight peak in the KI values. From a time of 30 minutes to 72 minutes intthe transient, KI drops again following the pressure drop at 30 minutes, but then begins to increasedus to'the buildup of thermal stresses in the vessel from the cooldown induced by the restoration of feedwater. At a time of 72 minutes a more rapid rise is seen in KI due to the repressurization of the system.
This produces the peak value in KI to occur at 78 minutes., after which time KI gradually bjging to decrease duo to a reduction of the thrnal grdiont in the vessel wall. A similar effect is seen at the other two axial height locations on a slightly more exaggerated scale. This is because of the more pronounced effect of the thermal stresses at the higher elevations where the fluid temperature at the wall is calculated to be colder due to less mixing.
7-11
Warm prestressing for this case would not be expected prior to 78 minutes in the transient. However, because of the peak which occurs at 78 minutes due to the 'thermal and pressure loads, and continuous decrease in K'I beyond this time, warm-prestressing is considered to be effective in preventing crack initiation beyond 78 minutes into the transient.
7.8.3 High-head HPSI SBLOCA + LOFW This transient (Case,6) is similar to the previous case for a very small break LOCA plus loss of feedwater, with restoration of feedwater at 30 minutes. The temperature and pressure curves used are taken from Section
- 4. The system parameters for this transient are conservative for Maine Yankee, the particular difference being the high HPSI shutoff head. This case also is potentially severe because of the repressurization up to 2400 psi as well as the thermal stress due to fairly rapid cooling of the vessel. Very early in the transient the injection of HPSI water produces a step change in fluid temperature at the wall to 350 F at the top of core and above the core. A second thermal shock occurs following restoration of feedwater at 30 minutes in the transient.
The thermal analyses and the corresponding stress analyses were performed for the three axial heights along the vessel, and the resulting stress intensitie vs. time were calculated for assumed flaws at each axial location. At the mid-core level the K, vs. time results are given in Figure 7.20. The results at the top of the core are presented in Figure 7.21 and KI vs. time above the core'shown in Figure 7.22. Similar trends are apparent in all three cases with an initial rise in KI up to a time of 20 minutes, followed by a decrease in the stress intensity until a transient time of about 32 minutes.
The restoration of feedwater produces a significant increase in KI beyond a time of 32 minutes for all crack depths due to thermal and pressure loading until a peak is reached at a time of 40 min. Beyond this time KI steadily decreases due to the reduction of the thermal gradient while the pressure, remains constant. For this reason ,warm-prestressing would be effective in preventing crack initiation beyond a time of 40 minutes in the transient.
7.9 2-Dimensional Heat Transfer Analysis The effect of mixing in the azimuthal direction was considered in order to evaluate the conservatism associated with the assumption of uniform 7-12
temperature around the circumference (i.e. axisymmetry) for the heat transfer analysis. In essence, this analysis was performed to determine the effect of local cooling on the stress intensity of an axial crack located directly below a cold leg nozzle. It was found that the local cooling condition did not produce higher KI values than the axisymmetric case, so that the assumption of uniform temperature around the circumference is conservative. This~method could be used to perform plant specific heat transfer analyses considering the proximity of each .weld in relation to the cold leg'nozzles.
A two-dimensional heat transfer model of the vessel was used identical to that assembled for the stress analysis. The effect of temperature variations in the axial direction was considered in the same manner as in the axisymmetric case, that is, only one longitudinal section of the vessel was examined, namely the top of core. For the 2-dimensional analysis, the fluid temperature was considered to vary between the bulk temperature, TB, and the partially mixed peak temperature, T .
Directly below the nozzle, at an angle of 0 degrees, the fluid temperature was considered to be T . At some distance (i.e. angle) away from the nozzle the temperature of the fluid against the wall was taken to be the bulk temperature, T . A smooth variation between TB and Tf was accomplished by fitting a normal (Gaussian),curve using the relation Ts TB - (TB Tf) F where F = exp _.6 ( )
and Ts = temperature of fluid at the surface at angle
= angle around circumference measured from centerline of the nozzle.
A spreading paramoter
- angle of center of nozzle Because of the system geometry, three nozzles were considered at angles of 0, 60, and 1800 around the circumference. The spreading parameter was chosen such that 60 percent of the temperature variation-around the circumference occurs within the diameter of the nozzle.
7.1
The heat transfer analysis was performed using the finite element method.
The fluid temperatures TB and Tf were taken from the SBLOCA + LOFW Case 5 (top of core). This analysis calculated the temperature distribution in the wall, T, as a function of radius, r, angle, # , and time, t.
An example of the 2-dimensional temperature field is shown in Figure 7.23.
A stress analysis of the vessel was also performed for the top of core axial section. The pressure transient from Case 5 was used in loading the structure as well as the 2-dimensional variations in temperature. Various depth axial cracks were assumed to be aligned .directly below the nozzle in order to produce the maximum crack opening effect. The stress intensities were calculated as in the previous analyses, and the calculated values of K, vs. time are shown in Figure 7.24. In comparing these results with the 1-dimensional heat transfer case shown in Figure 7.18. it is noted that the 2-dimensional results are equal to or slightly below the results for the axisymmetric analysis. This indicates that the assumption of axisymmetric heat transfer is reasonable and conservative for this study.
Certainly the 1-0 would be even more conservative if the percentage of mixing were less and if a crack in a critical weld was not located directly beneath a nozzle. However, the benefits in using the 2-dimensional heat transfer method do not appear to be significant enough to justify the additional effort required in performing the thermal analysis in this manner on a plant specific basis for the transients considered here.
The additional margin made available by considering the 2-D aspects may be credited if necessary in future analyses of other transients.
7-14
REFERENCES 7.1 R. D. Cheverton, S. K. Iskander and S. E. Bolt, "Applicability of LEFM to the Analysis of PWR Vessels Under LOCA-ECC Thermal Shock Conditions",
NUREG/CR-0107 (ORNL/NUREG-40), Oak Ridge National Laboratory Oak Ridge TN (October 1978) 7.2 R. D. Cheverton "Pressure Vessel Tracture Studies Pertaining to a PWR LOCA-ECC Thermal Shock: Experiments TSE-1 and TSE-2", ORNL/NUREG/MT-31, Oak Ridge National Laboratory, Oak Ridge TN (September 1976) 7.3 R. D. Cheverton, "Experimental Verification of the Behavior of Surface Flaws in Thick Walled Cylin-ers During Thermal Shock", presented to NRC Vessel and Piping Integrity Review and Workshop Oak Ridge National Laboratory June 1, 1981 7.4 F. J. Loss et al, "Significance of Warm Prestress to Crack Initiation During Thermal Shock", Naval Research Laboratory NRL/NUREG Report 8165 (Sept. 1977) 7.5 R. D. Cheverton and S. K. Iskander "Application of Static and Dynamic Crack Arrest Theory to Thermal Shock Experiment TSE-4" NUREG/CR-0767, ORNL/NUREG-57 (June 1979) 07.6 J. J. Simon, M. W. Davis, W. H. Tuppeny, Jr., "Experimental Determination of Limiting Heat Transfer Coefficients During the Quenching of Thick Steel Plates in Water", Combustion Engineering Report A-68-10-2 (December 1968) 7.7 MAR-CDC, Non-Linear Finite Element Analysis Program, Control Data Corp.,
Minneapolis,'Minn. (1976) 7.8 J. R. Rice, "A Path Independent Integral and the Approximate Analysis of Strain Concentration by Notches and Cracks", J. Appl. Mech., Vol. 35, (1968) p. 379 7.9 G. G. Chell, "A Fracture Mechanics Approach to Predicting the Effects of Warm Prestressing and its Applications to Pressure Vessels", SMiRT 5, Paper G9/6, Berlin (West). Germany (August 1979) 7-15
.21875" 8.625" Stainless SteeT' Cladding Figure 7.1 Axisymmetric Heat Transfer Model tncludinor Stalls, Stoel Cladding 7-16
COLD LEG z
ABOVE CORE TOP OF CORE CORE MID-PLANE Figure 7.2 Location of Axial Sections in R.V.
7-17
550 Time = 0 Min.
Time = 3 min.
500 450
. Time =30 min.
400 0 1 2 3 4 5 6 7 8 1.0. 0.!
Distance Through Vessel Wall (inches)
Figure 7.3 Typical Temperature Profiles Through Vessel Wall 7-18
4I Cranck Tip Figure 7.4 Stress Analysis Model of Reactor Vessel With Axial Crack
,5 Z il
-t.-*,
Figure 7.5 Deformed Shape of Reactor Vessel Showing Crack Opening Effect Due To Thermal Shock 7-20
,9000E+03 a/w = .88/
T E
9 S .7OOOE*03
= ..a/w 75 sM
- 00+2.00E0 IoE0 y .0 0000 0*20E0 T I. ist3 Fitoure ,"a vs. Time f or Var inta Crack L ets
.9000E403
/
5 .7400003 N /
E/ T /
N /
S .60OOE+03 +
I /
E.500000 pA T .10000 01---
- //
I /
.4000E*o3 C R
) .2000E*03
- A/
cn cx0D0r(403 /
Fiur .7 K v . rrkDnt a im/
(. 7 vu ;. n o -o s-- - . avoi.mu
220 200 180
. 160 140 -KIc KiK 120 100 oz
< 80 U*
o 60 40 20
-100 -50 0 +50 +K)0 +150 +200 (F7Rf T-R )F Figure 7.8 ASME Reference Tashn-Iess Curves for Reactor Vessel Steels
STRESS INTENSITY KI VS. CHACK DEPTH
- eduled; w-- - w- - - - * - - - - - - = w- - - -- w- - - - w- - - - w- - - - - - - - ----------
I/ /
.9000E*03**
. ) /
S YO000E*03*
I /
S/ /
M / /
T / ,/
E S5000E*03*
K/ /
1 / 1 T .1000C+03
- K / /
C .4000E03 I / -
-of O.10000L400*0#600*0031000 00EU 0 0 D-EPTII A-u - -
0.. Fiure K0OOOCo7 CakeIOOnEtul O *6o0Gand0 .vCOO.o
- iqijre 7.9 KT. and 1'7 vs. Crack 9epth at Time t
STRESS INTNSIW -V49 4CCKDEPTH
.9000E403 90000E403 /
I 5* .700'oE+03 T
G.6400403.
K-,
N/
T go E .5000E+03 *. / Kla N/
Ln S/
T 0-0" S .3000E*03 e/
.1000E*03 00 .20000+00 . 4000E*00. *6000E*00 *o000000 ol000E*01 ur~e 7I1) Sunerimnosed r.read v.C ck ea at Time t
CHITICAL CRACK DEPT" VS. TIME A ------- _ - -----
.9000E*00 +
.00000100 K / /
.6000E000 +
D//
PC/ /
T z/
H .soooLoo * . Arrest if .4000510o
.aoooE*00 +Initiation*
-o o------ ----- oo ------ *--- .
______ 11__ Ti--.--. _ __ _
0, .2000E*02 , 000E402 .6000E+02 ImDLu AOOlot03 r/ r
Bulk Temp.
Middle of Core
- - -- - - Top of Core 600. - - Above Core 500 400 300 200 100 600 1200 2400 3600 4000 6000 TIME (Sec.),
Figure 7.12 Approximate Temperature Transient for Case 4 7-27
2800 2400 2000
- 1600 1200 800 400
- 0 1200 2400 3600 4300 6000 TIME ;(Sec.)
Figure 7.13 Approximate Pressure Transient for Case 4
W .TnTAL STRESS INTEMSITY 1IT K(3 T11er USING APPLJen PRE%
/ * /
.QOOuE+03 +
/
T /
F.
S *70)03F+03 +
II E .100000 --- _ _
T 5 /
Y *5000E+03
(..2 I I K.1000CI03 +
TI 7l (A ~ ~ ~ ~ ~ ---- --- ----.---- -- __ ___
11
- 0. .200O0E*O2 .4000OE+02 .600CE*02 .ff0005+02 .1000F+03 TIMEi. (Mi nal"
_____Fiqure 7.14 Case 4 - Middle of Core - K1 vs. Time
TOTAL STREEtS 1FE4SITY I(1T + KIJP$ VS TI TM USINt; APPtTMU. PqRESS 09OOE603 +
- 9000E*03 +
.1000F+0 I I Fg ____________ .40Ff0 4oo 1 I 000 *0 3
TOTAL STRESF INTPN4ITY 11
- kIp) v TIME USI'4 6 PPL1 f P $S
+
M ot4-ji I ----- --- - -----
.9000t+03+
5 .7-000003+
T 1 .I41GU-e + 1 Y *U000E03 4+
.3000E+03+
T ./ u i r n a - K v . T n /T I
FU2 *4- --
4600) P,.00 #02 *4t00 +02 I 7 C. -- me
TOTAL SIESS IIIIENSITY (KIT
- KIP) VS. TIME USING APPLIED PHESS 1co-ooI w_.........
.9000E*03
- 5 .7000*03
- IIII E G OE0*83 C, I
/ 6
.500"0 of
.1000E*#03*
Pa "II__ /
- 0. .2000E#02 .40001*0Z .6000E*02 *(sn00t*0? 000*0&
SI C I M I V i Fioure 7.17 Case 9 - liddle of Cnre - Ky vs. Time
TOTAL STRES INTENSITY IT. +..I .... TIt. . USrING Ar.t..E ES
.9000L03 4 y /
E 6
ST/ .1000E*03 E .iO000.03 Y .5000E*03 I /
- 3000E#03F
.1000E*03 0 -- A A I I - -- I- I- - - - - -
0, .2aG§E*o2 e4000E*02 .6000E#02 46000E+02 .1000*03 Finusre 7.18 Case 5 -Ton of Cre Kr vs. Tirne
TOTAL ITRESS INTENSITY IMIT t KIP) V5. TIME USING APPLIED PRESS li fl li IIIIIII I 49000E+03++
S _ _ _ _ _ _ ____
T/S @7000E+0 3 T . .. . I
$ .5000E+o3
- IF It -flflfg4 t 46A I
- 3000E+03
.1
+
7, T In Itc Timt
M4~ ;TRESS INTENSITY Iu vs. 11HE USING APPLIEP PRES)0 4
S *900@E*03*
T E
FI /
I 0
T N
151 4000
- O*@E.@3 S----I 0 10 M N3PO A.5000t*03 Ft0taECv. .3000E*Q I. 4 - 4~
- p. 000____3 I e I iT I ItI M rigure 7.2 Ct ase S I.-dilof Co rre - K 1 VS. Tille
TOTAL STRESS INTENSITY KIT
- KIP) VS. TIME USINO APPLIED PRESS
- 9000E*63 S *00085*03
- T/
E/
S /
S.4080. I S / /
I /
T Ye000qE*03
- 0* P * -/
- T .
,1000E*03 a So 2000E*02 *4000E+0? 06000E+02 .80000E*02 .1000E*03 Tf I rE - i vs. Time Fingirp 7 ?1 c~arp 6 - Ton of Corp K! vs. Time
TOTAL STRESS INTENSITY IKIT
- I V$. TIME USING APPLIED PRESS
.9000E+03 S 680000003 T /
E /
S 6000E*03 S .60001.03 4 E-T wfi0#OE-.ww* /
N /
I /
S4 Y .5000E+03
- i000(003+
I /I /
T &20400"*3 /
,1000E*03 0 .o2000E402 .4000E*02 *60OOE*02 *8000E*02 01000f#03 Fioure 7.22 Case 6 - Above Core - K, vs. Time
jw~
Nozzl Contour Temp.
Level (OF) \ *p I 2CC S 3 3 = 2585 3 4= .8"E 3 5 = 31SE 3
= . 33SS 3
= . 3 75E 3
' 33 3
. '. B 3 i 11111
- 1(11 Nozzle Nbzl Figure 7.23 Temperature Contours in Reactor Vessel for 2-Dimensional Heat Transfer Analysis
TOT#L STRESS INTENSITY IK3T
- KIP) .YW. TIME USING APPOEP PKtst
- 9000E*03 E /6 T
N/
5 .7000*03 S
K .0006003 I /
.3000E*03
/
- 1000E*03
.4000E*o2 .6000E+02 ,HO0OE+02
- 000E*03 00 *2000E+02 T I M E -- -----------
-iur 7.2 Cas r 20 ton of r ore KI vs. Time
8.0 CONCLUSION
S This report provides the results of analytical evaluations of pressurized thermal shock effects on reactor vessels in CE NSSS plants for cases of a SBLOCA + LOFW, in response to the requirements of Item-II.K.2.13 of NUREG-0737. Two different scenarios were chosen for evaluation based on remedial actions to prevent inadequate core cooling:
- 2. SBLOCA + LOFW + Aux. FW reinstated after 30 minutes The results of plant specific evaluations are provided in a separate Appendix for each vessel. For vessels containing welds with greater than 0.15% wt.
copper, neutron fluence profiles, were developed and superimposed on plant specific material property maps in order to define vessel capability. Vessels with less than 0.15% wt. copper welds were evaluated on the basis of peak fluence assumed to coincide with worst material property.
Thermal-hydraulic system transient calculations were performed on a reference-plant basis, with parameter variations over the range representing all operating plants. Four different cases were analyzed for each of the two different scenarios defined above, for a total of eight cases. The most challenging of each of the two different scenarios was analyzed using linear elastic fracture mechanics methods to determine the critical crack tip stress intensity values for comparison to plant specific materials properties at various times in plant life. The effect of the warm prestress phenomenon is identified where applicable for each transient, and credited where appropriate.
It is concluded in each of the plant specific appendices of this report that crack initiation would not occur due to the SBLOCA + LOFW transients considered, for more than 32 effective full power years of operation, which is assumed to reoresent plant life at a conservative 80% capacity factor.
8-1
t -~
4 COMBUSTION ENGINEERING, INC.
t
- 0~ *1 -~