NRC-10-0010, Response to Request for Additional Information Regarding Relief Requests PRR-006 and PRR-009 for the Inservice Testing Program Third Ten-Year Interval
| ML100540147 | |
| Person / Time | |
|---|---|
| Site: | Fermi  |
| Issue date: | 02/19/2010 |
| From: | Plona J DTE Energy |
| To: | Document Control Desk, Office of Nuclear Reactor Regulation |
| References | |
| NRC-10-0010 | |
| Download: ML100540147 (8) | |
Text
Joseph H. Plona Site Vice President 6400 N. Dixie Highway, Newport, MI 48166 Tel:
734.586.5910 Fax: 734.586.4172 DTE Energy 10 CFR 50.55a February 19, 2010 NRC-10-0010 U. S. Nuclear Regulatory Commission Attention: Document Control Desk Washington D C 20555-0001
References:
- 1) Fermi 2 NRC Docket No. 50-341 NRC License No. NPF-43
- 2) Detroit Edison's Letter to NRC, "Submittal of the Inservice Testing Program Relief Requests for Pumps and Valves - Third Ten-year Interval," NRC-09-0064, dated November 3, 2009
Subject:
Response to Request for Additional Information Regarding Relief Requests PRR-006 and PRR-009 for the Inservice Testing Program Third Ten-Year Interval In Reference 2, Detroit Edison submitted proposed relief requests for the third ten-year interval of the Inservice Testing Program for Pumps and Valves. In an e-mail from Mr.
Mahesh Chawla to Mr. Alan Hassoun dated January 7, 2010, the NRC requested additional information for Relief Requests PRR-006, Service Water Pumps Suction Pressure Accuracy for the Comprehensive Pump Tests and PRR-009, Relief from Fixed Reference Value Testing. This was discussed in a subsequent telephone conversation between NRC staff and Detroit Edison personnel on January 11, 2010. The additional information requested by the NRC staff is enclosed.
There are no new commitments included in this document.
Should you have any questions or require additional information, please contact Mr.
Rodney W. Johnson of my staff at (734) 586-5076.
Sincerely, Enclosure
USNRC NRC-10-0010 Page 2 cc: NRC Project Manager NRC Resident Office Reactor Projects Chief, Branch 4, Region III Regional Administrator, Region III Supervisor, Electric Operators, Michigan Public Service Commission
ENCLOSURE TO NRC-10-0010 FERMI 2 NRC DOCKET NO. 50-341 OPERATING LICENSE NO. NPF-43 RESPONSE TO REQUEST FOR ADDITIONAL INFORMATION (RAI)
REGARDING RELIEF REQUESTS PRR-006, SERVICE WATER PUMPS SUCTION PRESSURE ACCURACY FOR THE COMPREHENSIVE PUMP TESTS AND PRR-009, RELIEF FROM FIXED REFERENCE VALUE TESTING FOR THE INSERVICE TESTING PROGRAM THIRD TEN-YEAR INTERVAL
Enclosure to NRC-10-0010 Page 1 The following is Detroit Edison's response to each NRC request for additional information (RAI):
PRR-006-001 Under 'Reason for Request', Paragraph 2 states "[t]he calibration cardinal point of 24 feet bounds readings obtained for pump surveillance testing." Please elaborate on why this is true.
Response
The RHR Reservoir level instrument E11R802 is the source for determining the service water pump suction pressure. Quantifying the accuracy of the suction pressure values involves analyzing the accuracy of the E11R802 instrument. Per the calibration specification sheet for E11R802 (E11N559A-SS), the instrument calibration is verified at cardinal points of 6, 12, 18, 24 and 30 feet. The relationship of reservoir level in feet to service water pump suction pressure is such that at 24 feet the suction pressure is 6.7 psi and at 30 feet the suction pressure is 4.1 psi. A query of the IST database showed that over the past ten years the lowest recorded suction pressure for any of the service water pumps was 4.7 psi and the highest was 5.5 psi. This equates to a range of reservoir level of 26.8 feet to 28.6 feet. The E11N559A-SS specification establishes a maximum as found tolerance of 0.22 feet at each cardinal point, therefore the degree of error for this instrument is lowest at maximum reading (30 feet) and higher at lower readings. The cardinal point of 24 feet is bounding in the lower direction for all of our recorded surveillance data, therefore the accuracy specification at 24 feet is appropriate to use in quantifying suction pressure accuracy.
PRR-006-002 Under 'Reason for Request', Paragraph 2 states "[t]he As Left tolerance at 24 feet for this instrument is 0.79%." Please discuss the significance and purpose of this information.
Response
The calibration specification sheet (E11N559A-SS) for the RHR Reservoir level instrument shows calibration is verified at cardinal points of 6, 12, 18, 24 and 30 feet. The specification sheet contains as-found and as-left tolerance requirements. At the specific cardinal point of interest, 24 feet, the as-found acceptance criteria is +/- 0.22 feet and the as-left criteria is +/-
0.19 feet. The analysis of suction pressure accuracy described in the response to RAI PRR-006-001 above uses the as-found tolerance which is conservative. The relief request includes mention of the tighter as-left tolerance only to convey the additional conservatism that this provides. Actual pump suction pressure values can be expected to be more accurate than what is assumed in the relief request.
Enclosure to NRC-10-0010 Page 2 PRR-006-003 Under 'Proposed Alternative and Basis for Use, Paragraph 10, discuss why the low DISPR reading of 32 psi is bounding. (A high DISPR reading would yield a higher error.)
Response
The basis of this relief request is that the differential pressure (DP) parameter is the critical measurement. The DP is what is compared to Code-based criteria and is the parameter which is trended to analyze pump performance. DP is a calculated value equal to pump discharge pressure (DISPR) plus suction pressure (INLPR). These pressures are added to derive DP because these are vertical lift style pumps. The essential argument of the relief request is that the INLPR values are much smaller than the DISPR values and therefore a small additional error in the INLPR value has an inconsequential impact on the calculated DP. In evaluating the numerical impact on DP for the additional INLPR inaccuracy, it is more conservative to use the smallest possible value of DISPR. This amplifies the INLPR error contribution to show the most conservative overall impact on DP PRR-009-001 Please discuss how the RHRSW pump optimum flow tolerance of +/-100 gpm was determined and why the associated MOV cannot be manually throttled.
Response
The existing procedural guidance for the RHRSW pump surveillances contains a requirement that the pumps be throttled to the IST reference flow but allow for a tolerance of +/- 100 gpm about that reference flow. RHRSW flow is throttled using a 24 inch Powell 1551WE motor-operated globe valve. Observations of the performance of these surveillance procedures show that the operators have a fairly limited ability to make small adjustments in flow. The most rapid possible operations of the throttling MOV controls yield changes in flow of between 30-60 gpm. Analysis of the last 35-40 recorded flow values for each pump shows the following:
Pump Range (gpm)
Deviation from No. of points within reference (gpm)
+/- 30 of reference RHRSW Pump A 5310 - 5457
-90 to +57 25 out of 36 total RHRSW Pump B 5315 - 5498
-85 to +98 24 out of 41 total RHRSW Pump C 5330 - 5500
-70 to +100 27 out of 44 total RHRSW Pump D 5315 - 5470
-85 to +70 26 out of 39 total The procedural guidance directs the operators to make a reasonable effort to get as close to reference flow as possible. The number of instances where operators were able to achieve
Enclosure to NRC-10-0010 Page 3 within 30 gpm of reference flow shown above clearly demonstrates this reasonable effort.
The 100 gpm tolerance band represents a workable range which has been in place for many years and the data above shows that there have been some occasions in which the full range was needed. The relatively equal distribution above and below the reference flow is evidence that there is no bias to lower flow values which would subsequently yield less conservative DP readings. The MOVs being used to throttle flow (E1150F068A/B) are located in High Radiation Areas. Additionally, the specific locations of the valve handwheels and declutch levers are such that operators would have to climb down from permanent grating onto the pipe for manual operation. The aspects of dose and personnel safety do not support routine manual operations.
PRR-009-002 Please elaborate on why limited access to necessary transmitter pipe connections makes installation of M&TE for flow measurement of CCHVAC Chilled Water not practical.
Response
These particular flow instruments have test access taps behind the instrument racks, which have a solid face. The test connection taps are low near the floor and access behind the rack is difficult. The location of this equipment is on the 5th floor Auxiliary Building which is accessible only by several flights of stairs. This would make transport of the flow M&TE equipment setup, normally moved to locations using a test cart, much more difficult.
PRR-009-003 Detroit Edison proposes to utilize a more conservative differential pressure acceptance criteria range for testing of the CCHVAC Chilled Water Pumps. In accordance with NUREG 1482, revision 1, section 5.3, "...the licensee may make a corresponding adjustment to acceptance criteria to compensate for the uncertainty..." Please describe the effect of the proposed differential pressure acceptance criteria range adjustment on uncertainty.
Response
In this relief request Detroit Edison proposes to use [for CCHVAC Chilled Water Pumps] a value of +/- 3.5% for the total flow tolerance, which is the accuracy of the flow measurement added to the throttling range allowed by the test procedure. This is higher than the total flow tolerance standard of 2% as described in NUREG-1482. To validate this request as an acceptable alternative, Detroit Edison also proposes toreduce the allowable band for differential pressure from +/- 10% to a band of -8% to +6%. In terms of equivalency, the 2% change in Minimum DP acceptance criteria would revise the actual test criteria for DP from 22.1 psid to 22.6 psid, an increase of 0.5 psid. In the range of the actual test flow on the pump curves, the relationship between flow and discharge pressure is approximately 2.2 psi per 40 gpm. Using this relationship, the 3.5% test flow tolerance is
Enclosure to NRC-10-0010 Page 4 equivalent to +/- 8 gpm, or 0.44 psi. This is a very conservative equivalency analysis because the true deviation in test flow tolerance that is requested is only 1.5% (3.5% - 2%).
Therefore, the proposed use of the more restrictive DP criteria is a valid offset for the requested flow tolerance deviation.
PRR-009-004 NUREG 1482, revision 1, section 5.3 states "[f]or a tolerance greater than the allowed percent (which may be necessary depending on the precision of the instrument), the licensee...may perform and document an evaluation to justify a greater tolerance." For each of the three pumps proposed (RHRSW, EECW MU, CCHVAC Chilled Water) the discussion states that the proposed tolerances as the best achievable. Please provide a justification for the greater tolerance for each group of pumps, i.e. why does the increased tolerance provide an acceptable level of quality and safety or why would compliance with the requirements not provide a compensating increase in the level of quality and safety?
Response
The basis of Relief Request PRR-009 is that pump performance can be accurately monitored even with the slight increase in total flow tolerance bands. The impact of the larger flow test band is increased scatter in the differential pressure data. Detroit Edison routinely uses a flow normalization method to analyze short term trends in pump performance. Fermi 2 IST Program Technical Position TP-12 describes this method.
Excerpts of TP-12 are as follows:
DATA NORMALIZATION:
For pump testing where a reference band is allowed, there will be some scatter associated with the measured variable data. This guidance will allow the evaluator a method to normalize that variable data in order to properly assess trends. An example is a centrifugal pump where flow is set to an established reference value but allowance is given within the procedure for test flow to be plus or minus a given range about the reference value. Discharge and suction pressures are then measured at some measured flow value within the allowed test flow range.
The evaluator needs to establish a known linear pressure-flow relationship, either from baseline data or from an assessment of a large quantity of pressure to flow data accumulated over several years. Using MS Excel, this linear relationship will need to be expressed as a table with at least 4 rows of data relating pressure to flow. One set of values should be at or below the minimum end of the allowable test flow band, another set at or above the maximum end and the remaining data at or very close to the middle (reference flow). The evaluator shall then use the Excel Forecast function to input an actual measured test flow and derive what the DP should be. The numerical difference between the derived DP and the DP at the center of the flow range shall then be
Enclosure to NRC-10-00-10 Page 5 subtracted from or added to (as applicable) the actual measured DP for that test flow.
This should be done for all measured data to be used in the trend analysis.
DP analysis performed in this manner is accurate regardless of the degree of scatter caused by variation in test flow. In addition, long term pump performance analysis is done using a moving average technique that eliminates the effects of scatter in the DP data. The response to RAI PRR-009-001 above demonstrates that there is no discernible bias in the historic flow variance for these pumps. The number and amplitude of points with test flow below reference flow is roughly equal to those above reference flow. Averaging of the DP within the data scatter produces an accurate plot for pump performance.
The ability to accurately trend pump performance despite increased data scatter due to higher flow tolerance bands is the basis for considering this relief request to be an acceptable alternative.
PRR-009-005 Please describe how this testing was performed during the second IST interval.
Response
The testing during the second ten year interval was done in the same manner as described in relief request PRR-009. The industry only recently became aware of the incorrect interpretation of NUREG-1482 section 5.3. For twenty years Detroit Edison had established proceduralized test flow tolerance bands of up to 2% exclusive of flow instrument accuracy.
Based on the recent NRC inspection findings at other plants, Detroit Edison now limits our total test flow tolerance bands to 2% inclusive of instrument accuracy with the exception of the pumps cited in this relief request. While the ASME OM ISTB committee is working on a resolution to this issue which may preclude the need for similar relief requests in the future, this relief is needed and an acceptable alternative has been presented.