ML18066A900
| ML18066A900 | |
| Person / Time | |
|---|---|
| Site: | Palisades  |
| Issue date: | 02/28/1997 |
| From: | Bordine T CONSUMERS ENERGY CO. (FORMERLY CONSUMERS POWER CO.) |
| To: | NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM) |
| References | |
| NUDOCS 9703120023 | |
| Download: ML18066A900 (17) | |
Text
consumers Power POWERING llllClllliAN-S l'IUllillUS Palisades Nuclear Plant: 27780 Blue Star Memorial Highway, Covert, Ml 49043 February 28, 1997 U.S. Nuclear Regulatory Commission ATTN: Document Control Desk Washington, DC 20555 DOCKET 50-255 - LICENSE DPR PALISADES PLANT Thomas C. Berdine Manager, Licensing SUBMITTAL OF REVISED INSERVICE TESTING RELIEF REQUEST NUMBER 8 On August 30, 1996, the NRC issued a safety evaluation that, in part, gave interim relief to the requirement to test the Palisades charging pumps, P-55A/B/C, in full compliance with the code referenced by 10 CFR Part 50.55a(b ). This interim relief was granted in accordance with 10 CFR Part 50.55a(a)(3)(ii) based on the hardship or unusual difficulty without a compensating increase in the level of quality and safety that would ensue ifthe requirements were immediately imposed. The interim p~riod was granted for 6 months during which a long-term resolution needed to be determined.
Investigation for the resubmittal of Relief Request Number 8 was delayed due to the 1996 refueling outage and the multiple vibration problems associated with the generator. These activities prevented technical personnel from addressing the issues, including a survey of equipment manufacturers, on a more timely basis. We are therefore requesting prompt consideration of an extension of the interim approval of the Relief Request Number 8 that was granted on August 30, 1996. This extension should be for the period needed to consider the approval of this revised Relief Request Number 8. contains a revised Relief Request No. 8 for the Palisades charging pumps. This relief request provides additional information on the instrumentation issues and the proJ:>osed alternative testing and preventive maintenance in addition to 9703120023 970228 GI l
~,PD~. ~ADOCK 05000255 p1;,r **
- PDR I
A CMS' ENER6YCOMPANY
that contained in the original request. Guidance for this submittal was gained from the NRG safety evaluation dated August 30, 1996, and NUREG-1482 "Guidelines for lnservice Testing at Nuclear Power Plants". This relief request maintains that the code requirement for low frequency vibration testing of the Palisades charging pumps is impractical. Palisades is requesting review and approval of this relief request. contains schematic drawings of the charging pumps P-55A/B/C which may be useful in understanding the vibration monitoring problems with these pumps.
2 contains a report on the investigation of vendor capabilities with regard to low frequency vibration monitoring capabilities.
SUMMARY
OF COMMITMENTS This letter contains no new commitments and no revisions to existing commitments.
Thomas C. Berdine Manager, Licensing CC Administrator, Region Ill, USNRC Project Manager, NRR, USNRC NRG Resident Inspector - Palisades 3 Attachments
ATTACHMENT 1 CONSUMERS POWER COMPANY PALISADES PLANT DOCKET 50-255 SUBMITTAL OF REVISED INSERVICE TESTING RELIEF REQUEST NUMBER 8 6 Pages
- RELIEF REQUESTS RELIEF REQUEST BASIS NUMBERS SYSTEM:
Chemical and Volume Control - Charging PUMP:
P-55A, P-558, and P-55C (Charging Pumps)
CLASS:
2 FUNCTION:
These pumps return deionized and filtered reactor coolant to the Primary Coolant System (PCS) at a rate equal to the purification flow rate and the controlled bleed off rate. Upon receiving a safety injection signal, they are started and discharge concentrated boric acid into the PCS.
TEST REQUIREMENT:
OMa-1988 to OM-1987, Part 6 (OM-6), Paragraph 4.6.1.6 states, "The frequency response range of vibration measuring transducers and their readout system shall be from one-third minimum pump shaft rotational speed to at least 1000 Hz."
RELIEF REQUESTED:
Relief is requested on the basis that imposition of the OMa-1988 requirements for vibration frequency response range is impractical since the low speed design of the pumps would require unavailable field instrumentation. This specifically follows the responses to Comments 5.4-2 and 5.4-3 in NUREG-1482 that addressed similar problems. It may be possible to set up laboratory instrumentation with the necessary capabilities in the field. However, it is not considered feasible to expect this system to be adequately calibrated and sufficiently accurate to provide repeatable and trendable data. The proposed alternative testing and preventive maintenance, using available equipment, will provide an acceptable level of quality and safety based upon the design of Palisades' Charging Pumps.
ALTERNATIVE TESTING:
As an alternative to the code requirement, overall (IPS-RMS) vibration readings will be obtained over a frequency response range of approximately 6 Hz to 16 kHz. Additional
RELIEF REQUESTS RELIEF REQUEST BASIS NUMBERS 2
readings are taken and used in a spectral analysis to determine trends in the vibration information. Additionally, oil analysis will be performed on a monthly basis with detailed sample analysis being performed when the samples are not cont~minated and may be free released for shipment off-site. In the event that the oil samples cannot be analyzed by an off-site facility, the vendor recommended oil change intervals will be observed and an inspection of the crankcase internals performed at that time.
BASIS FOR RELIEF:
The following information is applicable for the Palisades Charging Pumps:
Pump ID P-55A P-558 P-55C Motor Speed 1786 rpm 1790 rpm 1790 rpm Pinion Speed 620 rpm 1100 rpm 1100 rpm Crankshaft Speed 115.5 rpm (1.9 Hz) 203.2 rpm (3.4 Hz) 203.2 rpm (3.4 Hz)
To satisfy OMa-1988 requirement, the frequency response range for a vibration monitoring system would need to be from 0.64 Hz to 1000 Hz for Pump P-55A and from 1.13 Hz to 1000 Hz for Pumps P-558/C due to the low speed design of the pumps.
Palisades currently uses Predict/DU hardware and software. The system is well established and has been in place for six years although the low frequency response is 6 Hz. Predict/DU does not offer any equipment with a better frequency response than is currently being used.
Three other manufacturers, CSI, Entek IRD, and SKF, have also been contacted to determine if their equipment is capable of operating at the required calibrated frequencies. The result of the investigation into the vendor capabilities is that no data collector manufacturer has successfully completed any instrument calibration to frequencies below 2 Hz that is NIST certified to within 5% accurate. Therefore, no standard data acquisition systems were found that could satisfy OM-6 requirements down to 0.64 Hz or even down to 1.13 Hz. Palisades has therefore determined that the procurement of standard, commercially available equipment, capable of accurately measuring the extremely low frequency vibrations required by OM-6, is not currently possible. Without the equipment capable of measuring these low frequencies, it is
RELIEF REQUESTS RELIEF REQUEST BASIS NUMBERS
.3 impractical to test the charging pumps in accordance with OM-6 due to their low speed design. This revised Relief Request Number 8 is therefore based on the basis that meeting the existing code requirement is impractical. This specifically follows the responses to Comments 5.4-2 and 5.4-3 in NUREG-1482 that addressed similar problems. These responses stated that "a relief request must be submitted when it is impractical to meet the frequency response range," and "Apparently, the code committee did not establish the frequency response range for all types of pumps.
When the requirements appear to be impractical, the licensee may request relief."
Pump Description Palisades' Charging Pumps are reciprocation pumps and have components both above and below 600 rpm. Pump P-55A (P-558/C) input shaft and pinion rotate at 620 rpm (1050 rpm) and have two rotating element bearings. The bullgear and crankshaft rotate at 115.5 rpm (203.2 rpm) and are supported by two sleeve or journal bearings. There are no seals or couplings within the pump crankcase, nor is component rub a vibration concern based upon the pump design.
Vibration Monitoring The OM-6 Code committee has recommended that velocity measurement be used for rotational shaft speeds above 600 rpm and that displacement measurements be used for speeds below 600 rpm. This recommendation was made because vibration readings in velocity units may not be indicative of certain problem severity in slow speed machines, while vibration readings in displacement units may not be indicative of problem severity in higher speed machines (Reference NUREG-1482 Section 5.4). Velocity units are used for vibration monitoring of the Charging Pumps at Palisades due to their design.
The high speed pinion, supported by rolling element bearings and the reduction gears, is the largest concern when examining vibration and overall performance.
Problems in these components will appear at frequencies well above their running speeds and any degradation will be best detected using units of velocity.
Shaft alignment problems, intermediate gearbox problems, or associated with one times the pinion speed will be adequately detected using current equipment and the available frequency response range. The contribution of the low frequency vibrations associated with the crankshaft and bull gear will not be accurately proportioned due to roll off of the frequency response. However,
RELIEF REQUESTS RELIEF REQUEST BASIS NUMBERS 4
the contribution of vibration at these frequencies (in velocity units) to the overall vibration readings would not be significant due to the extremely low operating speed. As assurance that problems with low speed components will be identified, problems associated with the crankshaft/bullgear assembly will result in higher vibration amplitudes at gear mesh frequencies which are well within the frequency response range of current vibration equipment being used. By using velocity in lieu of displacement, we can more accurately monitor the more failure prone components within the pump crankcase.
Typical rotating element bearing problems are indicated by vibrations at non-integer multiples of running speed. With the exception of cage defects, these faults appear at frequencies well above running speed. Additionally, rotating element bearing defects are generally not detectable by any overall vibration reading until the degradation has progressed into latter stages. By that point, harmonics of the fault frequencies and the general noise floor (both well above running speed) have elevated and begun to increase the overall vibration levels.
Therefore, vibrations at subsynchronous frequencies caused by rotating element bearing defects (cage fault frequencies) offer no measurable amplitude increase in overall vibration readings until accompanied by fault frequency harmonics and noise floor elevation, both at higher frequencies.
Journal bearings are susceptible to some problems detectable by subsynchronous vibrations such as oil whirl and looseness. However, these problems are nearly always accompanied with higher order harmonics of shaft running speed and the fault frequencies. This is especially true for conditions that are truly detrimental to the bearings and overall pump performance.
Additionally, the contribution of subsynchronous amplitudes to an overall velocity vibration level, especially from a slow speed rotor, is likely negligible in comparison to the vibrations at and above the dominating running speed(s). As previously discussed, any such journal bearing problem would also result in elevated gear mesh frequencies.
In addition to overall vibration readings, Palisades obtains vibration readings for spectral analysis as a requirement of each inservice test procedure. No matter what the shaft running speed is, vibration problems associated with any bearing can be identified by spectral analysis long before they are detectable by overall vibration readings. Although roll off is present at lower frequencies in spectral data as well, the analysis is not necessarily based on peak vibration readings at
RELIEF REQUESTS RELIEF REQUEST BASIS NUMBERS low frequencies (or at any frequency); it is based on trends and the mere presence of signals at particular frequencies. Spectral analysis is a significant part of Palisades' vibration monitoring program because we initiate action based on spectral analysis before overall vibrations fall into the required action or alert ranges.
Periodic Maintenance Palisades follows vendor recommendations for charging pump oil change frequencies unless detailed oil sample analysis provides information to extend this interval. The vendor recommends quarterly changes for continuously running pumps and semi-annual oil changes for idle, or intermittently operated pumps. During oil changes, the crankcase cover is removed to examine for particles and to wipe out the sump. Gear and bearing wear particles would be visible during this evolution. A general inspection of internal components is also performed during oil changes.
Oil samples are drawn on a monthly basis for all running and standby charging pumps. Occasionally, the oil is radiologically contaminated and can not be released off-site for detailed analysis. When the sample is radiologically contaminated, it is tested on-site for acidity, water content, and viscosity. When the samples are not radiologically contaminated, they are sent off-site and analyzed for acidity, water, viscosity, particle contents, and a wide array of other contaminai:its and additives. Extension of the oil change interval can only be made following an acceptable off-site analysis.
Eguipment Monitoring 5
During pump operation, auxiliary operators regularly monitor all equipment for abnormal noise. Additionally, the system engineer is required to perform weekly walkdowns of equipment which also provides opportunity for monitoring noise emanating from the pumps. Occasionally, recordings of audible noise are made in the vicinity of the operating pumps. This provides a means to identify and verify changes in audible noise. Although there is no acceptance criteria, this noise monitoring has been responsible for detection of several indirect pump problems in the charging pump cubicles such as failed accumulator bladders and motor winding problems. In particular, suction bladder failures have resulted in audible cavitation at low frequencies
RELIEF REQUESTS RELIEF REQUEST BASIS NUMBERS 6
of occurrence. Similar audible banging or ticking would be expected if problems were to be present associated with the pump crankshaft or pistons.
Preventative Maintenance Charging Pump P-55A which is continuously run has been thoroughly inspected and/or overhauled four times in the last ten years. A qualified representative of a pump manufacturer has been present to oversee and direct each evolution.
With the exception of pitting on the bull gear teeth, no significant findings have been recorded. No unusual bearing wear or degradation has been noted. All major maintenance evolutions are controlled per a permanent maintenance procedure (CVC-M-1) which incorporates vendor recommendations and original equipment manufacturer specifications for acceptance criteria. Additionally, Palisades regularly participates in NMAC and CE Owners Group charging pump workshops which provides information as to modes of failure and possible generic issues.
ATTACHMENT 2 CONSUMERS POWER COMPANY PALISADES PLANT DOCKET 50-255 SUBMITTAL OF REVISED INSERVICE TESTING RELIEF REQUEST NUMBER 8 SCHEMATIC DRAWINGS CHARGING PUMPS P-55A/B/C 4 Pages
- CHARGING PUMP P-55A/B/C TOP VIEW
/
PUMP BEARING MOTOR BEARING GEAR BOX BEAR.ING VIBRATION LOCATIONS VIBRATION LOCATIONS VIBRATION LOCATIONS P-55A PUMP BEARING GEAR BOX BEARING VIBRATION LOCATIONS VIBRATION LOCATIONS MOTOR BEARING VIBRATION LOCATIONS P-558/C
3 134 48 CHARGING PUMP P-55A/B/C EXPLODED VIEW 128 Ci~
\\
131.
II 46
>4----- !!:
129 130 16
- CHARGING PUMP P-55A/8/C
- MODEL P AND MCP OIL LUBRICATION SYSTEM FOR EXTERNAL OIL PUMP RELIEF VALVE IS LOCATED OH ECCENTRIC BEARING CAP OIL PUMP A & B CONNECT A.
B EXTERNAL OIL PUMP (OPERATION 60 RPM OR LESS'
OIL FILTER
/OIL PRESSURE
- "':/
GAUGE WATER DRAIN LIME WATER DRAIN COHHECTION.
WELL AREA
73
\\
72
~..
9~ \\,
64 CHARGING PUMP P-SSA/B/C EXPLODED VIEW 55 __
I 87
ATTACHMENT 3 CONSUMERS POWER COMPANY PALISADES PLANT DOCKET 50-255 SUBMITTAL OF REVISED INSERVICE TESTING RELIEF REQUEST NUMBER 8 REPORT ON VENDOR REVIEWS TO DETERMINE LOW FREQUENCY VIBRATION MONITORING CAPABILITIES 2 Pages
REPORT ON VENDOR REVIEWS TO DETERMINE LOW FREQUENCY VIBRATION MONITORING CAPABILITIES Vibration Data Collector Manufacturer -- State of the Art Models Examined The following information was developed through direct contact with the technical personnel of the listed manufacturers of vibration data collectors.
CSI - Manufacturers of the CSI Model 2120 Data Collector Data collectors can detect vibration at frequencies as low as 0.5 Hz with standard accelerometers and lower with specialized accelerometers. However, no accuracy figure is applied to that 0.5 Hz value and calibration below 2.0 Hz is extremely difficult and has not been performed. Any such calibration would not be NIST certified through their calibration laboratory..
Predict/DU - Manufacturers of the Watchman 8603 and DC-78 Data Collectors Data collectors are 5% accurate to approximately 6.0 Hz. Below that frequency, significant response roll-off begins to occur. Predict/DU is reportedly in the process of lowering this range, but would not be able to provide 5% accuracy to 0.64 Hz.
Entek - IRD - Manufacturers of the Entek dataLINE and IRD dataPAC 1500 The IRD 1500 released last year is considered the unit with the best potential to meet 5% accuracy at 0.64 Hz. This model has a frequency response of 0.18 Hz to 73.3 KHz, however, no unit has been calibrated below 2 Hz due to procedural difficulty and lack of NIST certification below that frequency.
SKF - Manufacturers of the CMVA 10 and CMVA55 Data Collectors Data collectors can detect vibration at frequencies as low as 0.5 Hz but no accuracy figure is applied to that 0.5 Hz value. Hardware designers reported that modification of the CMVA55 would be necessary to meet a 5% accuracy at 0.64 Hz. Additionally, calibration below 2.0 Hz is extremely difficult and has not been performed. Any such calibration would not be NIST certified through their calibration laboratory.
Summary of Findings Data collectors are available that can obtain data at frequencies within the vicinity of 0.64 Hz. However, no data collector manufacturer has been requested to or successfully completed any instrument calibration to frequencies below 2 Hz that is
NIST certified to within 5% accurate. In many cases, data collector manufacturers would need to modify their existing hardware and develop an elaborate calibration process. Since there is no industry demand for such an accurate system at low frequencies over such a wide range, there are no off-the-shelf instruments available.
Transducers are available that may satisfy a 5% accuracy from 0.64 Hz through 1000 Hz, but when combined with an instrument loop, it is not practical to achieve a 5%
accuracy over this entire range. Even if such a system was engineered, calibration remains a challenge. Shaker tables are limited in low end frequency and transducer voltages become essentially DC. Calibration standards that will provide a reasonable ratio are not very common.
2