ML050450353
| ML050450353 | |
| Person / Time | |
|---|---|
| Site: | Palo Verde  |
| Issue date: | 02/04/2005 |
| From: | Bauer S Arizona Public Service Co |
| To: | Document Control Desk, Office of Nuclear Reactor Regulation |
| References | |
| 102-05208-SAB/GAM | |
| Download: ML050450353 (28) | |
Text
Scott A. BdUL!I flepariment Leader 1 02-05208-SABIGAM February 4,2005 ATTN: Document Control Desk U.S. Nuclear Regulatory Commission Washington, DC 20555-0001
Dear Sirs:
Subject:
Palo Verde Nuclear Generating Station (PVNGS)
Units 1, 2 and 3 Docket Nos. STN 50-528, 50-529, and 50-530 Submittal of Redacted Version of Recirculation Sump Void Testing and Probabilistic Risk Assessment Preliminary Results In letter no. 102-05195, "Submittal of Recirculation Sump Void Testing and Probabilistic Risk Assessment Preliminary Results," dated December 27, 2004, Arizona Public Service Company (APS) submitted to the NRC a report that was considered to be proprietary along with a request that the report be withheld from public disclosure pursuant to 10 CFR 2.390. In response to a request from the NRC, enclosed is a redacted copy of the report.
There are no commitments in this letter. Should you have any questions, please contact Mr. Thomas N. Weber (623) 393-5764.
Sincerely, SABIGAM
Enclosure:
Redacted Version of Preliminary Safety Significance Evaluation of ECCS Containment Sump Voided Piping cc:
- 6.
S. Mallet NRC Region IV Regional Administrator (all wlenclosures)
A. T. Howell 111, G. G. Warnick M. B. Fields Director, Division of Reactor Projects, NRC Region IV NRC Senior Resident Inspector for PVNGS NRC NRR Project Manager A member of the STARS (Strategic Teaming and Resource Sharing) Alliance Callaway Comanche Peak Diablo Canyon Palo Verde m South Texas Project Wolf Creek
ENCLOSURE Redacted Version of Preliminary Safety Significance Evaluation of ECCS Containment Sump Voided Piping
REDACTED VERSION SIGNIFICANT CRDR 2726509 PRELIMINARY SAFETY SIGNIFICANCE EVALUATION OF ECCS CONTAINMENT SUMP VOIDED PIPING REDACTED VERSION
REDACTED VERSION 1.1 BackgroundlPurpose of Report In July, 2004, Engincering personnel determined that a section of Emergency Core Cooling System (ECCS) piping leading from the containment recirculation sump, in both ECCS trains in each of the three Palo Verde Units, was left in an unfilled condition during normal plant operation. The resultant air void could potentially he ingested into the ECCS pumps suction following a Recirculation Actuation Signal (RAS). A review of design hasis information determined that this condition was not consistent with the design intent of the ECCS and not consistent with the analyses that demonstrate the ability of the ECCS to perform its design hasis safety functions. Condition ReporUDisposition Request ( O R ) 2726509 was initiated to document and evaluate the condition.
Thc purpose of this report is to describe and provide the preliminary results of a comprehensive testing and analysis program performed to determine the safety significance of this condition. The results of the evaluation are intended to he used in a risk assessment to determine the safety significance of the discovered condition.
1.2 Description of Condition The Palo Verde ECCS design employs recirculation from the containment sump after the contents of the Refueling Water Tank (RWT) have been injected into the reactor vessel and containment building. Upon receipt of a RAS, automatic valve actuations result in suction of the ECCS pumps being transfcrred from the RWT to the containment sumps. Two completely redundant and separated ECCS trains are utilized. Figure 1-1 illustrates a typical ECCS suction piping and component layout.
REDACTED VERSION
Page 3 REDACTED VERSION EMERGENCY CORE COWNO AND COhlkClNtdM SPRAY 5YSTEM SUMlON PiPlNG - TRAJN A
-=--
Figure 1-1 Typical Palo Verde ECCS Suction Layout As illustrated in Figure I - I, the containment sump suction pipe contains an in-hoard and an out-hoard containment isolation valve, and a downstream check valve. Engineering personnel determined that this section of the ECCS suction piping, between the two containment isolation valves and between the out-board valve and the downstream check valve, had been routinely left in an untilled condition.
Preliminary Safety Significance Determination
&RWMETAw-REDACTED VERSION
REDACTED VERSION Page 4 In the unlikcly event of a Loss-of-Coolant Accident (LOCA), the contents of the Reactor Coolant System will leak into containment and flow into the containment sumps. Automatic ECCS actuation would occur causing the contents of the RWT to he injected into the RCS and the containment building to maintain core cooling and containment pressure and temperature control.
Ultimately the basement of the containment building, including the containment sumps, would become flooded. Once the contents of the RWT are depleted, a RAS would be automatically generated causing both containment sump isolation valves to open, resulting in closure of the RWT isolation check valve. The RAS would also cause, by design, the Low Pressure Safety Injection pumps to he turned off. ECCS pump suction, consisting of a HPSI pump and a CS pump in each train, would thus be transferred to the containment sump.
With the containment sumps flooded and the section of containment sump piping voided, air would he trapped in the piping. As flow is initiated from the sump, this air could he entrained andor transported into the ECCS suction piping and potentially into the ECCS pump inlets.
Industry literature and operating experience indicates that pump performance could be severely degraded, or even result in air binding or pump failure, if the resultant air volume fraction ingested by the pump exceeds the pumps tolerance for air ingestion. Industry literature (Ref. 1 NUREGKR 2792) indicates that a pumps tolerance for air ingestion varies by design and fluid conditions, hut at air volume fractions above approximately 3%, pump degradation can he expected.
Therefore, in order to determine the safety significance of this condition, the air volume fraction that could be ingested by the HPSI and CS pumps would need to be determined. Once the air volume fraction is determined, each pumps tolerance for the projected air ingestion can be assessed, and ultimately the impact on the ECCS safety functions.
1.3 Significance Determination Approach The assessment of voided and two-phase fluid behavior is complex. A comprehensive scale-model testing program was employed to develop a full understanding of the system response to the void and the resulting aidfluid conditions that would be delivered to the pumps suction inlet.
The impact to pump performance was then assessed via full-scale testing, given the projected aidfluid conditions.
The scale model tests were performed at Fauske and Associates, and simulated the system response during and following a RAS with the affected section of piping initially voided. The scale tests were conducted in three phases. The first phase modeled the RWT and associated piping, and the sump and associated piping down through and including the long vertical tun of pipe. The purpose of the first phase was to demonstrate the ability to simulate the transient and measure the important parameters such as void fraction, pressure, and flow rate. A series of tests were performed to test important scaling parameters to ensure the results of the test could be confidently applied lo the full scale Palo Verde units. A series of phenomenological tests (the second phase) using a larger scale model was incorporated into the test plan to verify that the Preliminary Safety Significance Determination REDACTED VERSION
Page 5 REDACTED VERSION flow regime in the vertical section of the scaledpiping configuration was representative of large pipe behavior.
The third phase extended the scale model to include the individual pump suction piping up to each pump inlet. An extensive series of tests under varying flow and pressure conditions were performed. [
scale pump perromance tests.
Full-scale pump tests were performed at Wyle Labs utilizing a spare Palo Verde High Pressure Safety Injection (HPSI) pump and a representative Containment Spray (CS) pump to determine the impact on pump performance under the projected air ingestion conditions. The HPSI pump was of the same make and model as those installed at Palo Verde. A spare CS pump of the same make and model as the Palo Verde CS pumps was not readily available; therefore a spare CS pump from a cancelled WPSS plant was utilized for the test. This pump is the same make and model as the Palo Verde LPSf pumps and is very similar in design and size to the Palo Verde CS pumps. The impact on performance for equivalent fluid conditions is expected to be representative. Tests were performed for a spectrum of flow rates and air ingestion rates based on the results of the scale-model test program. Pump performance was measured as a function of air volume fraction. A maximum degraded pump performance curve was then constructed using the test results for the tests performed at maximum air volume fractions.
1 These results established the inlet conditions for the subsequent full-1 1
For those conditions which do not exceed the degraded pump performance capability, continued degraded ECCS delivery (Le. continued pump flow) is assumed until the air inventory available for ingestion into the pump is consumed, at which time restoration of full pump performance is assumed. For these conditions, maintenance of the ECCS safety function is assured.
Preliminary Safety Significance Determination REDACTED VERSION
REDACTED VERSION 2.1 Phase 1 Test Program and Results 2.1.1 Experimental Obiectives and Phvsical Arraneement The objective of the Phase 1 testing was to investigate the potential for the air initially resident in the horizontal piping section from the containment sump to be forced into the vertical downward piping section. Phase 1 tests included the transient effects of switching the supply from the simulated RWT to the simulated containment sump by sirnultancously opening the sump suction isolation valves. Clear piping was used for the horizontal and vertical segments of the simulated suction line to observe and record the flow pattern and the behavior of the initial air filled void.
The test facility that was used was comprised of two tanks with water inventories, a centrifugal pump, piping, and valves and associated instrumentation. The piping and valves used to establish and visualize the flow pattern development from the initial location between the valves and into the downcomer piping were all 4 inch in diameter. Clear plastic piping facilitated observation of the initial air inventory behavior during the opening of the motor operated valves. The vertical segment was also clear plastic piping that allowed for the observation [
I in the downward vertical flow. [
I 2.1.2 Scaline Considerations As indicated, 4 inch diameter piping was used to simulate the sump horizontal and vertical downward sections of piping. Since actual Palo Verde piping is 24 inch in diameter, this results in a 1/6Ih geometric scaling factor. This geometric (lengths and diameters) scaling factor was maintained through out the Phase 1 tests to the extent possible.
Flow rates were scaled in the Phase 1 tests so as to maintain the same dimensionless Froude Number parameter as would exist in the Palo Verde units. Previous tests and experiments described in the literature have demonstrated that maintenance of the Froude number, particularly for horizontal flow regimes, will result in prototypical behavior in scaled experiments.
2.1.3 Phase 1 Results and Observations A series of twelve tests were performed with varied [
1 1
REDACTED VERSION
REDACTED VERSION Page 7 2.2 Phenomenological Testing Program 2.2.1 Experimental Objective and Physical Arrangement Design reviews conducted before and after the Phase 1 tests and an independent review by [
] resulted in the identification of several phenomenological investigations that could be performed to provide additional perspective and assurance on proper scaling of the full plant condition to the 1/61h scale model. The phenomenological investigations included:
Preliminary Safety Significance Determination REDACTED VERSION
REDACTED VERSION Page 8 The test arrangement also provided the opportunity to observe the flow patterns and influence of the HPSl and CS branch connections off the lower header piping.
2.2.2 An extensive series of tests using the [
from these tests were Phenomenological Testing Results and Observations
] scale test apparatus were performed. Key observations Preliminary Safety Significance Determination P
K u
w REDACTED VERSION
REDACTED VERSION 2.3 Phase 2 Test Program and Results 2.3.1 ExDcrimcntal Ohiectives and Phvsical Arranecment The test facility for Phasc 2 was similar to that of Phasc 1, [
E Preliminary Safety Significance Determination Page 9 1
REDACTED VERSION
REDACTED VERSION Page 10 1
4
] In the plant system under accident condilions, air transported through the HI'S1 hne would influence the pump performance and cause a decrease in the flow rate being pumped. Reduced flow rate would cause a corresponding reduction in the rate of air ingestion. Thus, the air intrusion rate deduced from these scaled experiments provides a conservative representation of the plant response.
The test instrumentation is also illustrated in Figure 2-1. A computer with a CIO-DAS008 data acquisition card was used to collect the data. Key pieces of instrumentation included Various pressure, level, and flow meters 1
During the Phase 2 tests, the flow rate through the CS pump was again held constant at the maximum predicted flow rate equivalent to 4885 gpm, except for several tests in which CS flow was set to zero to simulate a HPSI flow only scenario. HPSl flow rate was varied ranging from the equivalent to 200 gpm to an equivalent maximum run-out flow of 13 10 gpm.
1 2.3.2 Scaling Considerations The same 1/61h geometric scaling used in Phase 1 was used for the Phase 2 experiments. Flow rates were scaled to maintain the same Froude number that would exist at Palo Verde. The Froude number relationship was maintained for both the total flow and the individual flow rates to the simulated HPSl and CS pumps.
Preliminary Safety Significance Determination REDACTED VERSION
REDACTED VERSION Page I 1 1
2.3.3 Phase 2 Results and Observations A series of twenty-eight tests were initially performed with varied flow rates, containment level, and containment pressure conditions. Additional tests were later performed to investigate the air transport process during potential LPSI pump start scenarios. Key observations from the tests were:
1 Flow Patterns Digital movie cameras were used to record the flow patterns in all the Phase 2 tests. Each test was initiated by simultancously opening the sump containment isolation valves. As the valves open, water is seen to enter the initially voided horizontal piping segment and induce mixing of the water and air. The air is swept out of the horizontal segment and into the vertical piping segment.
HPSI Air Ineestion Rates Preliminary Safety Significance Determination REDACTED VERSION
REDACTED VERSION Page 12 Preliminary Safely Significance Determination PmJPwlTMWN-REDACTED VERSION
REDACTED VERSION Page 73 These results show that the air flow ingestion rates increase to their maximum value within approximately [ ] seconds for the scaled experiments and then subsequently decay towards zero as the air inventory in the horizontal suction header becomes insufficient to enter the HPSI line.
Similar evaluations for scaled HPSI flow rates were also performed.
The air ingestion rate information was then scaled up to determine the air ingestion rates that could have been experienced at Palo Verde under postulated accident conditions. Since the Phase 2 tests were conducted on a 1/6'h linear scale, the mass flow rates were increased by six cubed (216) and the time interval increased by the square root of six (2.44) to develop the conditions
[
I Preliminary Safety Significance Determination REDACTED VERSION
Page 14 REDACTED VERSION that would have been experienced at Palo Verde and for use as input into the full scale pump performance tests. Using the results from the Phase 2 tests, these scale factors are applied and the results illustrated in Figure 2-4 for the case of a full HPSI flow rate of 13 IO gpm. As shown, the meaningful delivery period for the air flow is approximately [
I Since Reference 1, and other pump performance tests described in the literature, indicates that pump performance is typically assessed as a function of air volume fraction, the peak mass flow rate data obtained during the Phase 2 tests was converted to air volume fractions for use in the subsequent full-scale pump tests.
Preliminary Safety Significance Determination REDACTED VERSION
REDACTED VERSION Page 15 1
Preliminary Safety Significance Determination REDACTED VERSION
REDACTED VERSION 3.1 Description of Test Facility The pump performance tests were conducted at Wyle Labs in Huntsville AL. The test facility consisted of two closed pump loops each drawing suction from, and discharging to, a common 30,000 gallon pressure vessel. One loop was constructed to provide for testing of the spare HPSI pump. Suction and discharge pipe sizes were selected to correspond to the actual pipe sizes at Palo Verde. The specific suction piping configuration leading into the HPSI suction nozzle was explicitly reproduced. The second loop was provided for testing of the representative CS pump.
1 3.2 Test Conduct A series of tests were conducted at each base case flow rate. The base case flow rates were selected to produce the same dimensionless Froude number as the cases tested in the Phase 2 scale model tests, and were meant to span the range of flow rates that could be expected at the time of RAS during a postulated LOCA.
For each base case, tests were performed at incrementally increasing air injection mass flow rates.
The resulting air volume fraction, defined as the ratio of volumetric air flow rate to total volumetric air flow rate, was then determined. [
I Figure 3-1 illustrates the final test for the [
1 base case.
REDACTED VERSION
Page 17 REDACTED VERSION During every test, the duration of air injection was specified to assure that the total volume of air
[
data was taken during each test for subsequent assessment of the air ingestion on pump performance. Visual observations, and digital camera recordings, were made for all HPSI test cases.
] exceeded the total volume of air predicted by the scale model tests. Pump performance 3.3 Test Results Visual observations through the clear spool piece on the HPSI suction line confirmed Phase 2 tests. The visual observations confirmed the proper scale up from the Phase 2 tests and gives reasonable confidence that the Phase 2 and Phase 3 tests closely approximate the full-scale plant conditions. Pump performance data was taken using a data acquisition systcm that recorded each data point 10 times per second. The recorded data was then inserted into Excel spreadsheets to facilitate calculation of pump developed
] similar in nature to that observed during the scale model
[
1 a typical pump curve graph as shown in Figures 3-2 through 3-4. The data represents the calculated developed head (TDH) from the recorded pump inlet and outlet pressure data taken every 0.1 seconds, and the corresponding flow rates as measured on the pump discharge line. The data represents that obtained over a specific time period during which the air injection rate was at its maximum steady state valuc and the corresponding peak air volume fractions were obtained. The data points, as expected, fall along the test loop system curve. [
1 I
I Preliminary Safety Significance Determination REDACTED VERSION
REDACTED VERSION Page 18 Preliminary Safety Significance Determination REDACTED VERSION
REDACTED VERSION Page 19 As illustrated in the preccding three figures, and as would be expected, pump performance progressively degrades as inlet air volume fraction increases. This progressive degradation is consistent with data reported in NUREGKR 2792 (Reference 1). The following figure 3-5 is taken from Reference 32 as cited in the NUREG.
SULLCR R E S E A R C H NUMBER 1970 Figure 3-5 Degrading Pump Performance as a Function of Air Volume Fraction Preliminary Safety Significance Determination m
REDACTED VERSION
Page 20 REDACTED VERSION A maximum bounding degraded pump curve is then constructed as shown in Figure 3-6.
As illushated, the maximum degraded pump curve conservatively bounds all recorded data for the peak air volume fraction cases tested. The use of this maximum degraded pump curve results in additional conservatism since the Phase 3 tests conditions in some cases cxceedcd the specified air volume fraction from the Phase 2 scale model tests.
Preliminary Safety Significance Determination REDACTED VERSION
REDACTED VERSION Preliminary Safety Significance Determination Page 21 1
REDACTED VERSION
REDACTED VERSION 4.1 Thermal Hydraulic Analysis of Spectrum of LOCA Break sizes A series of thermal hydraulic analyses of the Palo Verde ECCS system were performed using
] These analyses 1 conditions that established the expected [
would exist at the time of RAS for a spectrum of LOCA break sizes. Operator actions as prescribed in the Palo Verde Emergency Operating Procedures (EOPs) to initiate a cool down and depressurization of the RCS upon diagnosis of a LOCA were explicitly considered in the analyses. In this way, best-estimate parameters [
at time of RAS were established.
[
1 The analyses provide two key results. The first is that break sizes of 2 diameter or smaller The second result is the RCS pressure that would exist at the timc of RAS for various size breaks. These results are provided in Figure 4-1, R C S P r e s s u r e a t R A S 1600 1400 1 2 0 0 1000 8 0 0 n
6 0 0 400 200 0
0 1
2 3
4 5
6 7
8 9
10 B r e a k Size Figure 4-1 RCS pressure at the time of RAS for various break sizes from CENTS analyses REDACTED VERSION
Page 23 F%QFwE-REDACTED VERSION This pararncter is uscd in thc idlowing section to [
I subs-qucnt usscssrncn[ on ECCS pcrfomiuncc (i.c. [IPS1 llow) undcr the niaxiniuiii prcdictcd air ingestion conditions.
4.2 Determination of Degraded HPSl Flow Preliminary Safety Significance Determination P
R REDACTED VERSION
Page 24 REDACTED VERSION The resulting HPSl system performance or operating points, given the degraded pump performance [
extended Bernoulli energy equation:
] can be determined from the The results can be depicted graphically as shown in Figure 4-3.
As indicated in Figure 4-3, the static head associated with the 1 diameter small break LOCA at the time of RAS is well above the developed head of the degraded HPSI pump under maximum air ingestion. It is assumed that the degraded HPSI pump would be unable to deliver flow to the RCS.
For break sizes 2 diameter and larger, Figure 4-3 indicates the degraded HPSI pump has sufficient developed head to continue delivering ECCS flow to the RCS for the short time until the volume of air originally resident in the voided piping is exhausted. After the total air Preliminary Safety Significance Determination p
R B
p R
(
v REDACTED VERSION
Page 25 REDACTED VERSION volume is ingested, the Phase 3 pump performance tests demonstrated the HPSI pump would recover and return to its normal non-degraded performance.
4.3 HPSI Pump Test Conclusion From the Phase 3 pump performance tests under air ingestion, a bounding degraded HPSI pump performance curve was developed. The bounding degraded performance curve envelopes the maximum predicted air volume fractions ingested by the HPSI pump, based on Phase 2 scale-model testing. This study then compared the resulting degraded pump performance with the calculated system resistance that would exist at the time of RAS, for the spectrum of break sizes. The comparison indicates the degraded HPSI pump would develop sufficient discharge head to maintain flow to the RCS for all break sizes except for the smallest breaks less than 2. The degraded flow rate delivered to the RCS would only exist for [
at which time pump performance can be assumed to return to normal.
, ] until the air inventory available to be ingested is exhausted, 4.4 Containment Spray Pump Test Conclusion Tests were conducted on the representative CS pump by injecting air at rates up to approximately [
amount of air predicted by scale model testing for all scenarios tested. The pump experienced a reduction in flow of approximately [
] during the period of air ingestion, then returned to normal baseline performance after air injection was suspended. It is concluded that the voided pipe condition does not have a significant impact on Containment Spray pump functionality.
] air volume fraction. This air volume fraction conservatively bounds the 4.5 Probabilistic Risk Assessment Conclusion From the CENTS thermal-hydraulics analyses and the Phase 3 pump performance tests, modifications to the Palo Verde Probabilistic Risk Assessment (PRA) model were made to assess the risk significance of the voided pipe condition. The Palo Verde model contains an event tree for small break LOCAs of 2.3 inch diameter and smaller. The model was revised by inserting a failure of the HPSI pumps at RAS (failing the high pressure recirculation function) for small-break LOCA due to air binding and modeling the subsequent plant cool down and depressurization and LPSI alignment for low pressure recirculation. Since the pump performance tests indicate that for breaks 2 inches in diameter and larger failure of the HPSI pump is not likely, medium and large LOCA events were unaffected by the voided condition. Thus the small LOCA event would be the dominant contributor to the risk increase due to the voided pipe condition. Some sensitivity studies related to this analysis are currently in progress and no analysis of the impact on Large Early Release Frequency has been performed at this time.
With the above described change made to the PFL4 model, the increase in CDF is about 3E-6Iyr.
Preliminary Safety Significance Determination
-w REDACTED VERSION
REDACTED VERSION Page 26
- 1. NUREG/CR-2792 "An Assessment of Resioual Heat Removal and Containment Spray Pump Performance Under Air and Debris Ingesting Conditions". Published Septemoer 1982.
- 2. Paranjape. S. S. et al. 2003. "Interfacial Structures in Downwaro Two-Phase Bubby Flow".
11'" International Conference on Nuclear Engineering (ICONEI 1). Tokyo, Japan.
- 3. Wallis, G.B., 1969. One Dimensional Two-Phase Flow, McGraw-Hi.,. New York.
Preliminary Safety Significance Determination REDACTED VERSION