ML20217D756

Safety Evaluation Concluding That Testing of Ingersoll-Dresser Pump Model 8HN194 at Test Facility Demonstrates That Crystal River Decay Heat Pumps of Same Model Can Operate at Flows of 100 Gpm for 30 Days

ML20217D756
Person / Time
Site:	Crystal River
Issue date:	10/01/1997
From:	NRC (Affiliation Not Assigned)
To:
Shared Package
ML20217D754	List: ML20217D751 ML20217D756
References
TAC-M96314, NUDOCS 9710060098
Download: ML20217D756 (6)

Text

-

3 L SAFETY EVALUATION BY THE OFFICE OF NUCLEAR REACTOR REGULATION ON PUMP TESTING BY INGERSOLL'-DRESS 8ER PUMP COMPANY TO VERIFY CRYSTAL RIVER UNIT 3 DECAY HEAT PUMP CAPABILITY AT LOW-FLOW CONDITIONS FLORIDA POWER CORPORATION DOCKET NO. 50 302 TAC NO. M96314

1.0 INTRODUCTION

in July 1996, the Nuclear Regulatory Commission (NRC) Office of Nuclear Reactor Regulation -

completed the second phase of an Integrated Performance Assessment of Crystal River Unit 3 (CR 3).

The CR 3 Integrated Performance Assessment Process (IPAP) Final Assessment Report (NRC Inspection Report Number 50-302/96-201) dated August 23,1996, identified that a single failure, such as the loss of an emergency diesel generator, could result in the loss of electrical power to one decay heat pump and the inability to successfully open the drop line. This would result in a mission time of the remaining decay heat pump on recirculation flow to exceed the 10-hour maximum allowed time limit. To resolve this issue, the licensee performed an extended low-flow endurance test at the vendor's test facility to verify the low-flow performance of the decay heat pumps.

2.0 BACKGROUND

NRC Bulletin 88-04 requested all licensees investigate and correct as applicable two minimum flow design concerns: (1) strong / weak pump interaction in common minimum flow piping; anci (2) minimum -

flow capacity. Florida Power Corporation (FPC or the licensee) responded to the bullet!a in a submittal dated July 6,1988. The submittal stated that current minimum flow requirements were being requested - i from pump vendors. In a letter dated December 9,1988 FPC statea that as a result of changes in vendor recommendations, it identified a potential concern with decay heat pump's flow capacity in certain operating modes.

- CR 3 Licensee Event Report (LER)89-009, dated April 14,1989, stated that the decay heat removal pumps would be required to operate at flows less than the manufacturer's revised minimum flow capacity for certain small break loss of coolant accidents (SBLOCA). At the time of the submittal, the plant was shut down for a reactor coolant pump repair. The licensee committed to perform testing representative of accident conditions, and detail the test results and required actions in a supplement to the LER.- In situ flow testing of decay heat pump 3B wat wrformed on April 14,1989. Additional-information was provided in a submittal dated April 17,1s59, which stated that the vendor (then, Dresser) recommended the following operating limits: (1) minimum continuous stable flow (MCSF) of 2350 gpm; (2) operation between 2350 gpm and 1200 gpm for durations of 100 hours0.00116 days <br />0.0278 hours <br />1.653439e-4 weeks <br />3.805e-5 months <br /> per year

. maximum; and (3) operation at 80 gpm for 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> of cumulative operation after which the pump bearing and shaft must be replaced. The licensee stated that these recommendations were incorporated into their plant procedures. The licensee noted that the decay heat removal pumps have operational modes at flows below the MCSF.

7 The most severe operational mode was minimum recirculation which is calculated to be a minimum flow of 100 gpm. In addition, for operation in the low pressure injection mode for an SBLOCA, the flow g0Y 9710060098 971001 PDR ADOCK 05000302 P PDR  !

2 requirement was a500 gpm. Normal decay heat system operations required pump ibw rates of 21500 gpm. The licensee also stated that this was a significant change from the manufactured previous MCSF recommendation of 800 gpm.

The results of the pump test were documented in a letter dated May 30,1989. The pump testing was conducted over s. range of flows from 3000 gpm to approximately 130 gpm, including a 10-hour run at 400 gpm in order to bound the expected worst case pump operating condition (SBLOCA) of 500 gpm for 5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br />. An evaluation by the licensee's consultant concluded that the decay heat pumps would operate acceptably at or above 400 gpm for a period of 3 days based on: (1) the pump operating characteristics at 400 gpm; (2) no significant temperature increase in process or bearing cooling flows; (3) vibration measurements below the American Society of Mechanical Engineers (ASME) Code required action range with average values of 0.46 inches per second (ips) with a peak of 0.55 ips during the 10-hour test; and (4) acceptable bearing lubricating oil analysis results after the pump test. The pump was disassembled and subsequent inspection found no indications of unusual component wear or

- damage, in response to a request from the staff, the licensee submitted in a letter dated June 12,1989, additional justificatian for the acceptability of extrapolating the proposed pump mission time limit from the 10-hour test rur at 400 gpm to 3 days, in a letter to the licensee dated June 15,1989, the NRC stated that althoug's the likelihood is high that such a conclusion is correct, the information available does not fully justNy the conclusion that the decay heat removal pumps can operate at 400 gpm for 3 days. Restart was made contingent on a commitment from the licensee to a full resolution of the issue.

The licensee's commitment is documented in the revision to the April 14,1989, LER dated June 28,1989. The evaluation of the test results performed by the licensee's consultant, MPR Associates, was provided to the NRC in a letter dated July 11,1989.

On August 17,1989 a meeting was held to discuss the licensee's evaluation of the decay heat removal pumps and the extrapolation of the allowable run time from 10 hours1.157407e-4 days <br />0.00278 hours <br />1.653439e-5 weeks <br />3.805e-6 months <br /> to 3 days at 400 gpm. As a result of the meeting, the licensee agreed to establish a realistic operating time limit for the pump necessary to bring the plant to a safe condition and provide additional justification or perform additional pump testing if it was determined that the pump mission time would exceed 10 hours1.157407e-4 days <br />0.00278 hours <br />1.653439e-5 weeks <br />3.805e-6 months <br /> (see licensee letter dated-November 1,1989). In an intemal NRC memorandum dated September 27,1989, the NRC staff and its

contractor, Oak Ridge National Laboratory, concluded that there was not sufficient justification for extrapolating the test results because: (1) no margin between the nominal test flow rate in the April 14, 1989, test, and the accident flow rate; (2) test data uncertainties for flow in the April 14,1989 test; and (3) static fatigue evaluation of the pump shaft did not adequately represent the extremely advero: and complex dynamic operation conditions associated with low-flow pump operation.

A meeting was held on March 30,1992, documented in an NRC letter dated April 29,1992, to discuss the remaining issue concoming the mission time of the decay heat pump. During this . meeting, the licensee agreed to provide a detailed description of the simulator scenario to verify the pump mission time. The simulator evaluation, documented in a submittal dated August 26,1992, concluded that a SBLOCA would require decay heat pump operation in a low flow condition for approximately 5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br />.

_The licensee stated in this submittal that the Bulletin 88-04 issue for the decay heat pumps should be considered resolved. The NRC stated in a letter dated October 29,1992, that the licensee's response to Bulletin 88-04 was satisfactory.

~

,c, .

3-During the July 1996, Integrated Performance Assessment of CR 3, the staff identified that a single -

failure, such as the loss of an emergency diesel generator could result in the loss of electrical power to one decay heat pump and the inability to successfully open the drop line. This would result in a mission time of the remaining decay heat pump on recirculation flow to exceed the 10-hour time limit referenced above. This _ issue is documented in inspection report 50-302/96-201, 'The Crystal River Unit 3 -

Integrated Performance Assessment Process Final Assessment Report " dated August 23,1996 as unresolved issue URI 96-201-01. To resolve this issue, the licensee pedormed an extended low-flow endurance test at the vendor's test facility to verify the low-flow performance of the decay heat pumps.

In a letter dated May 13,1997, the licensee submitted Report TR 9640, Revision 00, " Hydraulic Evaluation by ingersoll-Dresser Pump Company for Florida Power Corporation, Crystal River 3 Plant, Decay Heat Pumps, IDP Model 8HN194,* to provide information necessary to close out URI 96 20101.

The pump used in the test was a spare decay heat pump from Three Mile Island (TMI) that was identical, including the model number, as both Crystal River decay heat pumps. The purpose of the testing was to perform low-flow testing at 100 gpm and to determine if the MCSF of 2350 gpm could be reduced. The testing consisted of a pre-test inspection of the pump, baseline performance testing, a 30-day low flow test at 100 gpm, an 8-hour " flow rangeability test" to determine the MCSF, and post test pump inspection.

Prior to the test, the pump was inspected and critical components measured and compared to vendor

' drawings. The wear ring bores and the stuffing box pilot fit to the casing were determined to be 0.005" undersized. The vendor stated that these discrepancies would not have any significant effect on the  !

test results and stated this configuration was a more rigid test because the rotating element of the pump would have contacted the wear ring mere easily. In addition the TMl pump impeller is 18.50 inches where the Crystal River decay heat pumps each have impeller diameters of 18.75 inches. The vendor stated that this would not affect the low-flow performance of the pump because this is a function of the Impeller inlet configuration.

Performance data, including tank level, suction temperature, discharge pressure, differential pressure, bearing housing temperature, and approximate pump mechanical seal leakage, was collected at two-hour intervals during the 30-day endurance test. - Pre and post test performance curves plotted total '

. developed head, efficiency, and brake horsepower as a function of pump flow rate. Vibration measurements were performed using the licensee's test equipment and included unfiltered and filtered vibration in ips_ and pump spectral data. Flow rangeability testing was performed at flow rates from 500 -

gpm to 4680 gpm to determine the MCSF. The vendor concluded that the pump could operate from

- 100 gpm to 1200 gpm for up to one year after which the pump rotor must be refurbished. MCSF for this pump was determined by the vendor to be 1200 gpm. Post-test inspection of the pump, including non-destructive examination of the pump shaft and impeller, and bearing oil analysis, did not provide any indication of abnormal pump degradation.

-3.0 Evaluation The baseline pump performance curve (flow vs. total developed head) for the spare TMI pump taken on October 7,1996, before the start of the 30-day endurance test, is plotted in the graph below. Also plotted on this graph is data taken from a minimum flow test for the CR-3 decay heat removal pump conducted on April 14,1989, Performance Test Procedure PT-321 (PT-321). These curves are essentially parallel with approximately a 5% difference in total developed head (TDH) between the

4 l spare TMI pump and Crystal River pump 3B at 500 gpm and approximately a 9% difference in TDH between the two pumps at 3000 gpm. This offset is consistent with the Crystal River decay heat pump 3B impeller being slightly larger than the TMl spare decay heat pump impeller Therefore, the use of the TMI pump provides a valid comparison of performance at low flow conditions for the Crystal River decay heat pumps.

t Crystal River and TMI DH Pumps Head vs Flow 450 '

t i i l i t 400 -

ii '

l 350 A 4

3 300

$ 250

, -w-Thal AnarnDH Piimn W7/QA I

,, 200 150 C 51A MUALFSAR Fgure 6-9 E 1 Cnstal Rver DH Pumo 3B 4/14/89 50 o I O 500 1000 1500 2000 2500 3000 3500 Flow Rate gpm The vendor report did not include any calibration data for instrumentation used during the performance testing. However, a letter from Ingersoll-Dresser to the licensee states that the total test error for flow may be "slightly" above 2%. For the purposes of this test, the actual accuracy is not required to be in accordance with the ASME Operations and Maintenance (OM) Code provided the instrument accuracy of the test equipment is considered in the test results.

The test procedure states that bearing housing temperatures are to be monitored and recorded, it appears that less stringent criteria were used because numerical bearing temperature data was not included in the endurance test data sheets. However, bearing temperatures were taken during the flow rangeability study showing acceptable bearing housing temperatures in the range of 69 to 71 degrees Fahrenheit. Additionally, significant vibration data was taken during the spare decay heat pump endurance test, which provide acceptable indication of pump mechanical performance, The TMI spare pump performance curve is virtually identical to the baseline performance curve generated after the 30-day endurance run on November 7,1996 (not included in the graph), which provides evidence that the pump hydraulic performance did not degrade during the low-flow endurance test. Vibration data taken by the licensee during the low-flow endurance test did not reveal any significant vibration trends that were attributed with the pump. A post test pump inspection and non-destructive examination of critical components did not reveal any significant wear in any pump I

-_m_J

5-

' components. Therefore, the pump clearly is capable of operating at 100 gpm (recirculation flow) for a period of 30 da The vendor determined that the MCSF for this pump could be reduced from 2350 gpm to 1200 gpm.

Above the MCSF, the decay heat pump can operate without any stringent time constraints. Pump

. standards, such as the American Petroleum Institute Standard 610, state that vendors should determine the MC5F but do not give criteria as to how this value should be established. The pump vendor stated that this value was established by performing " flow rangeability testing," which, it believes, is standard engineering practice. Criteria to establish the MCSF could include specific vibration limits, complete

elimination of cavitation, or some other hydraulic performance criteria, in addition, vibration values of the TMl spare pump are between 0.1 ips and 0.2 ips for pump flow rates below 3500 gpm. Examination of the test data for the TMI pump reveals that the 1200 gpm value was not a part of the licensee's flow

' rangeability testing. However, in examining the baseline performance data, a significant change in the slope of the head curve does occur above 1000 ppm. At 1200 gpm, the pump is no longer operating on the flat portion of the pump / head curve which is typically regarded as a region of unstable pump operation, The vendor states that, based on the 30-day low flow endurance test, the decay heat removal pump can operate at flows between 100 and 1200 gpm for a period of one year. This appears to be a function of the projected life of the mechanical seal (reference March 12,1997, letter, from John Crane

- Company to ingersoll Dresser Pump Company included in Vendor Report TR-9640, Revision 0).

Therefore, the MCSF and the extrapolated duration of the pump under low-flow operations appear to be based on vendor experience as opposed to strict criteria. Given the fact that the original concem was operating the decay heat pump up to 3 days at flows of 100 gpm, the vendor's test satisfactorily bounds this requirement.

4.0- Qhservations

.Two observations were noted during the staff's review, which are described below:

The first observation involves the margin between the actual and expected performance of the decay

- heat removal pump. The above graph shows the performance curve for the Crystal River 3B decay heat removal pump, based on test data from PT-321 and the decay heat pump's performance curve as o described in the Crystal River FSAR (Figure 6-g). FSAR Section 6.1.3.2 states that one decay heat removal pump can deliver approximately 3000 gpm to the reactor vessel.- According to the graph, at 3000 gpm there appears to be litts or no margin between the actual performance of the 3B decay heat pump _and the expected pump performance documented in the FSAR. If ASME Code hydraulic required 3 action ranges are applied fully to this pump, there is potential that the pump could be allowed to . '

. degrade below the FSAR specified hydraulic performance limits. This should be investigated and

" addressed by the licensee.

The second observation involves the vibration of the Crystal River 3B decay heat pump. Vibration measurements of the TMI spare decay heat pump indicated that, with the exception of loosening of the pump baseplate during the endurance test, no significant increases in vibration values were recorded.

i Spectral data indicated that the main vibration contributor was from the impeller blade pass frequency.

- The absolute vibration values and the spectral plots demonstrate that the TMI spare pump did not sustain any signl5 cant degradation of mechanical components. However, these results cannot be

. . . . . r . .

__.-_s---- -

6-transferred directly to the current performance of the Crystal River decay heat pump. In examining the performance data from completed Performance Test Procedure PT 321 dated April 14,1989, pages 14 and 17 clearly show that the motor side pump beanng (vibration point C) exhibits vibration values above the ASME OM Code alert range criteria of 0.325 ips at flow rates below 800 gpm. This is not indicative of the TMl spare decay heat pump mechanical performance, Although these values are below the required action criteria of 0.700 ips, licensees are expected to initiate corrective action for pumps determined to be operating in the alert range during their inservice test. The licensee should ensure that the vibration levels are in accordance with ASME Code requirements or take appropriate corrective action.

5.0 Conclusions On the basis of its review of the licensee's test report (TR 9640), the staff finds that testing of Ingersoll-Dresser pump model number 8HN194 at the vendor's test facility demonstrates that the Crystal River decay heat pumps of the same model and manufacturer can operate at flow conditions of 100 gpm for 30 days. An assessment of the manufacturer's claim that the pump can operate for 1 year at flow rates of 100 gpm is difficult to make because the manufacturer appears to have used a combination of qualitative criteria and supplier recommendat!ons on component life to determine the expected life of the pump at this operating point as opposed to quantitative criteria such as trend data of component wear rates or degradation of hydraulic performance (note: there was no apparent degradation of hydraulic performance during any of the testing of the TMI pump). Considering that the test bounds the '

expected mission .ime for operation of the decay heat removal pump during a SBLOCA by an order of magnitude, an assessment of the manufacturer's extrapolation of pump performance to one year is unnecessary.

Principalcontributor J.Colaccino Date: October 1, 1997 d

_ _ _ . _ _ _