ML20203E060
| ML20203E060 | |
| Person / Time | |
|---|---|
| Site: | Prairie Island  |
| Issue date: | 02/23/1998 |
| From: | Sorensen J NORTHERN STATES POWER CO. |
| To: | NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM) |
| References | |
| NUDOCS 9802260244 | |
| Download: ML20203E060 (67) | |
Text
_.
l Northetn States Power Company Prairie Island Nuclear Generating Plant 1717 Wakunade Dr. East Welch. Mennesota 55089 February 23,1998 U S Nuclear Regulatory Commission Attn: Document Control Desk Washington, DC 20555 PRAIRIE ISLAND NUCLEAR GENERATING PLANT Docket Nos. 50-282 License Nos. DPR-42 50-306 DPR-60 Response to Request for Additional Information Concerning Prairie Island Unit 2 Control Rod Drive Mechanism Lesk By letter dated February 18,1998, the NRC Staff requested additional information concerning the leak found on a Prairie Island Unit 2 part length control rod drive mechanism (CRDM). The attached information is provided in response to that Request for Additional Information (RAI).
During teleconferences subsequent to the issuance of the subject Request for Additional Informatiors, NRC staff requested Northern States Power (NSP) address plans for additional testing, inspection or other corrective action concerning Prairie Island Unit 1 part length CRDM housing welds. NSP understands that the issues Review Group (IRG) of the Westinghouse Owners Group (WOG) responded to verbal questions from the NRC staff and has activated the Regulatory Review Group (RRG) of the WOG to address the generic implications of the crack discovered on the Prairie Island Unit 2 part length CPDM housing weld. NSP mainta:ns membership on the IRG and the RRG and plans to support the actions recommended by the RRG. At a minimum NSP plans to perform additional inspections of the Prairie Island Unit 1 part length CRDM housing welds at the next scheduled refueling outage, which is currently scheduled to begin in April 1999.
The root cause evaluations and investigations relative to the Unit 2 part length CRDM I
weld crack are not yet complete and consequently NSP continues to evaluate r
information relative to the part length CRDM housing weld crack as it becomes j'M l
9802260244 990223 ff li,
s.
n USNRC NORTHERN STATES POWER COMPANY February 23,1998 Page 2 available, and may modify the above planned actions based on information received as part of the on-going investigation.
in this letter we have made the following new Nuclear Regulatory Commission commitments:
- 1. Northern States Power, at a minimum, will perform additionalinspections of the Prairie Island Unit 1 part length CRDM housing welds at the next scheduled refueling outsgo, which is currently scheduled to begin in April 1999.
- 2. Northern States Power will assemble the requested part length CRDM housing design description, with verification of the information by Westinghouse, and the iniormation will be provided to the NRC Staff by March 6,1998.
- 3. Information on the A-ferrite requirements specific for the weld and buttering metal used in the fabrication of the part length CRDM motor tube base is being retrieved by Westinghouse ard NSP, and will be provided to the NRC Staff by March 6,1998.
- 4. The formal root cause analysis report for the flaw in part length CRDM G-9 will be completed and submitted to the NRC by April 6,1998.
Please contact Gene Eckhol' (612-388-1121) if you have any questions related to tnis response.
mtMu -
Joel P Sorensen Plant Manager Prairie Island Nuclear Generating Plant c: Regional Administrator - Region lil, NRC Senior Resident Ir.spector, NRC NRR Project Manager, NRC J E Silberg
Attachment:
Response to Request for Additional Information Concerning Prairie island Unit 2 Control Rod Drive Mechanism Leak 4
t.
l t
i:
Response to Request for Additional Information Concerning Prairie Island Unit 2 Control Rod Drive Mechanism Leak i
PrairieJaland Unit 2 Corrective Actions j
Recuest 1:
Please, provide the NRC with a copy of your root cause analysis of the flaw Indications in the Unit 2 control rod drive mechanism (CRDM) partial length j
housing. The following are points that you may want to consider in your root cause analysis:
' A. A description of t!ie design of the partiallength CRDM housings including description of the materials of construction, fabrication rnethods, and design aspects for the housings and what the design code of record is for the part-length CRDM housings.
B. The A ferrite requirements, if any, spa.. Ic for the stainless steel used in the fabrication of the full length CRDM houtings in the Prairie Island reactor l
vessel designs and any records, if available, with respect the A-ferrite requirements.
C. A discussion of the impact of the corrosive environmental conditions and thermal-hydraulic loads on the detected flaw indications in the No. G9 penetration.
D. The failure mechanism for the flaw indications in the No. G9 partial length CRDM housing nozzle, and the basis for concluding that the mechanism is considered to be the predominant mode of failure for the housing.
Response
l Evaluation of the root cause of the flaw indications on the Prairie Island Unit 2 part length Control Rod Drive Mechanism (CRDM) motor tube base are still ongoing. The specific responses to items A through D below provide a discussion of the ir, formation F
currently available from the ongoing root cause evaluation. Where inadequate informatior is available to fully respond to your request, a schedule for when that information will be available is provided.
Item 1.A:
Preliminary information, in the form of design drawings, has been previously provided for the information of the NRC Staff. Those drawings included the following:
- TWestinghouse Drawing No. 883D194, NSP Revision 10
.n-.
,w.
RAI Respona 6
F bruary 23,1998 P:g) 2 Royal Industries Drawing No.121C142, Revision B Detail Nos. 2 & 3 of drawing number NRPN002 from Creusot-Loire Technical Manual, NSP No. X-HIAW-1001-1139 Northern States Power will assemble the requested part length CRDM housing design description, with verification of the information by Westinghouse, and the information will be provided to the NRC Staff by March 6,1998.
Item 1.B:
Per a telecon with the NRC Staff on Fobruary 2L,1998, it was clarified that instead of information on the full length CRDM housings, the NRC would like the A-ferrite requiremer,ts specific for the weld and buttering metal used in the fabrication of the part length CRDM motor tube base. That information is being retrieved by Westinghouse and NSP, and will be provided to the NRC Staff by March 6,1998.
l'em 1.C:
A discussion of the impact of the corrosive environmental conditions and thermal-hydraulic loads on the detected flaw indications in the No G-9 penetration is being developed by Westinghouse. A preliminary version is expected to be ready for release to the NRC Staff by March 6,1998. This information will be finalized in the formal root cause analysis report which is expected to be completed and ready for release to the NRC Staff by April 6,1998.
Item 1.D:
The failure mechanism for the flaw indications in the No. G-9 partial length CRDM housing nozzle, and the basis for concluding that the mechanism is considered to be the predominant mode of failure for the houcing is being developed by Westinghouse.
This information will be finalized in the formal root cause analysis report which is expected to be completed and ready for release to the NRC Staff by April 6,1998.
Reauest 2:
Summarize your Intended corrective actions for Prairie Island Unit 2 and the respective schedule for these actions.
Response
The corrective actions taken on Unit 2 include a design change to remove the part length CRDMs from the reactor vessel head and to replace them with Head Adapter Plugs similar to those used on existing spare CRDM penetrations. The Unit 2 part length CRDMs were removed 2 / cutting the lower canopy seal welds and by unthreading the assembly, a Head Adapter Plugs were screwed on and were seal
RAI R:sp:ns3
+
4 Fcbru:ry 23,1998 Peg) 3 i
welded using an inconel 625 fillet weld, These activities were completed on February 20,1998.
The design change also includes the installation of CRDM Cooling Baffle Dummy Can Assemblies on the threaded plugs at the part length motor can removal area. The dummy cans are the same as those currently used in the spare ChDM locations and are being fabricated in accordance with Westinghoum design requirements.
i Replacement seismic covers will be installed at the Part Length Rod Position Indicator removal area in the seismic restraint grid. Selsmic cover and dummy can fabrication and installation will be completed prior to Unit 2 startup.
Prairie island Unit 1 Assese ment 1
l Reauest 1:
Provide a safety evaluation of the flaw Indications in the CRDM housing and a justification for concluding the Prairie Island Unit 1 plant can be safely operated for the remainder of the current operating cycle. The following are items, inat 1
should be considered in this assessment:
A. A description of the NDE [ nondestructive examination) methods currently j
used to analyze the flaws in the CRDM housing in Unit 2, and a description of the techniques used to qualify the NDE techniques for use.
)
B. An evaluation of the orientation and size of the flaws in Unit 2 detected by the NDE personnel. Include in this assessment the following analysis elements:
(1) A description of the geometric orientation of the flaws and the orientation of the flaws with respect to each other.
(2) A conservative determination of the flaw sizes in the CRDM housing.
(3) A best-estimate determination of the relative proximity of the flaw Indications in the No. G9 CRDM housing to one another, and an assessment of whether or not the flaw indications need to be considered as one Indication in accordance with the ASME [American Society of Mechanical Engineers) Code proximity rules.
C. What are the worst-case sizes of flaws that are postulated to be present in partial length CRDM housings in the Prairie Island Unit 1 RPV [ reactor pressure vessel] head and what is your assessment of these flaws? Based on the results of the Unit 2 flaw evaluation, justify why any flaws postulated in the PraMe Island Unit 1 partial length CRDM housings would be considered acceptable to remain in service without necessitating a repair of the flaws over the remainder of the Unit 1 operating cycle, include in your flaw
,w----
.,..-.r
.w-.
e,-
RAI Response t
Fcbruary 23,1998 Peg) 4 evaluation allowances for crack growth of the flaws that are postulated to occur from the start of the Unit 1 operating cycle to the current time, and for the remainder of the current operation cycle. What are the margins against inatability for both normal operating and design-basis loads?
D. Provide a description of any enhanced leakage monitoring being conducted for the Prairie Island Unit i reactor coolant system. Provide a discussion of any additional compensatory measures taken on Prairie Island Unit 1.
E. Provide an analysis of the limiting postulated event (s) for the partial length CRDM housings, include in your assessment the following elements; (1) What is the limiting loss of-coolant accident (LOCA) associated with a failure of a partiallength CRDM housing?
(1) What are the estimated consequences of a postulated LOCA resulting from a complete failure of a partiallength CRDM housing? Please
{
address the following questions regarding consequences of the postulated event:
- a. What is the potential for missiles during the postulated event?
- b. What would the consequences be from hydraulic loads, jet impingement loads, and postulated missiles resulting from the postulated event? In particular, address what the consequences would be to the full length control rods, reactor core, and lines penetrating the Prairie Island reactor vessels during the postulated event as a result of postulated hydraulic loads, jet impingements, and missiles
- c. What are the dynamic effects as a result of the postulated full CRDM housing break at the indicated crack location (including a description of the dynamic analysir miodels used to assess the jet thrust force caused by jet flow from the postulated break, and a discussion stating whether or not the models are consistent with the Guidelines provided in NRC Standard Review Plan Section 3.6.2)?
(3). What is the minimum number of ECCS [ emergency core cooling system) pumps (e.g., low pressure safety injection (RHR) [ residential heat _
removal]) pumps, and high pressure safety injection (charging) pumps) that would be required to mitigate the consequen::es of LOCAs up to and inclusive of the size of the limiting LOCA associated with a complete failure of a partial length CRDM housing in the Prairie Island units?
(4) What is the minimum number of control rods required to bring the reactor core to a subcritical level fc postulated LOCAs up to and inclusive of the
RAI Resp:nn Fcbru ry 23,1998 i
P:g) $
size of the limiting LOCA associated with a compete failure of a partial length CRDM housing in the Prairie Island Units?
1
Response
The following discussion is a summary of the current Prairie Island corrective action program reviews on the extent and safety significance of the defect found in the Unit 2 part length CRDM at location G-9, serial number RA-70-090.
A safety evaluation per 10CFR50.59 was not prepared for the defect found in Unit 2 part length CRDM G-9, since a permanent change is not being made to the facility by leaving this defect in service. The defect on Unit 2 part length CRDM G-9 is being corrected. A safety evaluation was prepared to justify the corrective action, because the facility was changed to remove the part length CRDM housings to eliminate the defect. However, the defective housing is addressed under 10CFR50 Appendix B, part XVI, Corrective Actions. The root cause is being ans,1yzed, so the extent and any necessary corrective actions can be determined to preclude repetition of this event on Unit 1 or elsewhere in the nuclear industry.
On Jan 24th,1998 Unit 2 rcactor was shut down to rv.e leakage of 0.26 gpm from the Reactor Coolant System (RCS). The leak was from a w Oct located Ebove the part length CRDM intermediate canopy seal weld, in the motor tube base weld. Preliminary failure cause analysis results of February 16,1998 from the Westinghouse metallurgical examinations indicated the defect may be from original weld solidification (hot shrinkage) cracks. The crack appears to have traversed between the 403 center section base metal and the weld butteririg layer. The leak area, including 70 degrees of the circumference, was opened for examination. The crack was found to be covered with a dark, high temperature oxide layer out to the crack tip. The crack tip at the leak location was covered in the oxide layer and extended out to nearly the outside diameter. There was no evidence of any loads having propagated the crack. There is no indication of an stress corrosion cracking or fatigue growth mechanism. This indicates the crack has been stagnant and stable for a very long time (since fabrication). This also indicates the defect was due to a welding process control problem on this part length CRDM, with no growth mechanism, and is not a service induced flaw.
The part length CRDMs are manufactured by Royal Industries. The full and part length CRDMs are the only components at Prairie island manufactured by Royal Industries.
The part length CRDM geometry and fabrication is different than the full length CRDM houings. The full length CRDM housings do not contain any welds in the latch hota v or rod travel housings, other than the canopy seal welds.
Considering a potentially generic extent, the Unit 1 reactor coolant system (RCS) leak rates, containment sump run times, and containment radiation monitor trends were reviewed. Stable trends were found on all parameters. Unit 1 reactor head was recently inspected for leaks upon startup from the December 1997 refueling outage.
No leaks were noted. The inspection procedure specifically requires a sign-off for
1 l
RA! Responso Fcbruary 23,1998
{
Pag)6 satisfactory inspections at the intermediate canopy seal weld locations. Unit 1 Containment inspection was performed on February 11,1998, it include'. a visual inspection of the reactor head. No leakage was noted from this inspection. Enhanced plant management oversight and enhanced monitoring of the RCS leak rate and radiation monitors was established. Details of the additional monitoring of condition of the Unit i RCS are described in response to item 1.D below.
Original receipt inspection documentation for the part length CRDM motor tubes was reviewed. The finished weld was radiographed and liquid penetrant examined. The inspections were satisfactory. The original radiographs were not available at Prairie Island. The base material, weld wire and weld butter layer material heat information was reviewed. The part length CRDM motor tube base with the defect, (G-9) was welded with different heat batch end clad buttering and weld wire than was used on the Unit 1 CRDMs. Other heat numbers for base metals are similar between Unit 1 and others on Unit 2, except for Unit 2 CRDM G-5. Unit 2 location G-5 used a different heat batch for the motor tube base transition piece and end clad buttering weld wire. Unit 2 locations I-7 and E-7 used all the same mater;al as on Unit 1 (one part length CRDM on Unit i does use a different end clad buttering weld wire heat batch than all other CRDMs). Royal Industries welder number five fabricated G-9 assembly That welder did not weld on any other part length CRDMs for Prairie Island. The certified material test reports for the G-9 weld materials were reviewed and were determined to be within specification.
in review of documentation for the defective G-9 component, it was found that this part length CRDM motor tube base weld had repairs performed during fabrication. The Weld Record form for the weld between the motor tube base and the part length CRDM housing center section indicates a rework work order number. The documentation for the rework has not yet been located. The defect G-9 weld was etched to see if a repair weld was evident. A repair may be indicated at the root of the weld in part length CRDM G-9.
A second weld (upper extension weld) exists on each part length CRDM tube. This weld is above the motor tube center section, approximately two feet above the defect area. This weld was also reviewed to help determine if the defect in G-9 motor tube base was of a generic nature. This second v. eld on each part length CRDM used the same material specifications and processes as the motor tube base welds.
Unit 2 part length CRDM G-9 motor tube base and upper extension welds had some rework performed on the welds. The original Weld Record forms list a rework tag number in the comments section. The motor tube base weld Sr Unit 1 part length CRDM serial number RA-70-11, and the Unit i upper extension welds for serial numbers RA-70-11 and RA-70-53 also indicated rework tag numbers. The motor tube base and end clad buttering welds for the spare part length CRDM in the Prairie Island warehouse (serial number RA-70-51) also indicated rework tag numbers. Neither the rework work order or the nonconformances could be located in the Prairie Island documentation. No axtent for these reworks can be determined. The defect in Unit 2 part length CRDM G-9 motor tube base has not been correlated to repair welding.
l rat Respons]
Febru ry 23,1998 Pag) 7 l
Ultrasonic examinations of the spare part length assembly and the upper extension weld on G-9, which both indicated rework had no apparent discontinuities (NAD).
A preliminary crack Integrity Evaluation was prepared by Westinghouse discussing the margins which existed for the penetration based on preliminary ultrasonic crack depC) data. A copy of this evaluation was previously provided to the NRC Staff for their information. The motor tube base cross section was examined by ultrasonic testing (UT), and the crack depth was measured as a function of circumferential location.
From the preliminary UT results, the crack was continuous around the inside circumference, with the minimum crack depth being at approximately 18 percent of the wall thickness, and the average depth being about 48 percent of the thickness. The preliminary Integrity Evaluation based on the UT data showed the cracked piece would meet ASME c de allowables and be stable under emergency condition loads. All available ligaments were counted in this critical flaw size calculation. The flaw was stable with no growth mechanism. The flaw will not be left in service. The crack leaked at a rate that was detectable under these conditions, and the unit was shut down. The part length CRDM with the flaw has been removed form the reactor vessel head.
However, further verification of the crack depth is being obtained by actual sectioning in the Westinghouse Lab. Preliminary information from the lab indicates the UT data was not conservative in relation to crack depth. Therefore the conclusions of the Integrity Evaluation are not verified.
Another safety comparison can be made within the housing of the defective component. The cross sectional thickness of the motor tube base at the leak location was 0.325 inches. The cross sectional thickness of the upper extension piece above the motor tube base is only 0.15 inches thick. The wall thickness at the leak location is three times as thick in order to transition to the threaded end. This thickness is not required for structural stability.
The hydrostatic tects performed at the fabricators shop, and the Reactor Coolant System hydrostatic test performed at Prairie Island provide a reasonable proof test for the part length CRDM motor tube base weld joints. The RCS hydrostatic test was conducted at a pressure of 3107 psi. The fabrication shop performed a hydrostatic test on the finished part length CRDMs at 3450 psi (refer to ASME N-1 data report). Taking the fabricators hydrostatic test at 3450 psi, and information from the preliminary Westinghouse Integrity Evaluation, a flaw deptn can be determined that would yield failure in the part length CRDM due to the hydrostatic test pressure. From the preliminary Westinghouse integrity Evaluation this flaw depth is 72%. The hydrostatic test stresses can be compared to the emergency condition stresses by comparing flaw depths obtained between Figures 1 and 4 of the preliminary report. Component failure urider emergency conditions with a continuous circumferential flaw depth of 72% will occur, if a through wall flaw also exists for 16% of the circumference. This comparison shows that the hydrostatic test pressure loads exceed the emergency condition loads for the part length CRDMs at Prairie Island. This analysis has not yet tseen verified.
The preliminary lab findings of an oxide layer at the crack tip, no evidence of any loads having propagated the crack, no indication of a stress corrosion cracking or fatigue
RAI Respons3 February 23,1998 P ge 8 growth mechanism, indicates the crack has been stagnant and stable for a very long time. With no growth mechanism, the hydrostatic testing still proves the component structural integrity today.
A summary of the UT inspections (unqualified) on the part length CRDMs for Unit 2 and the warehouse spare assembly, shows only the G-9 motor tube base with linear indications. A total of ten areas were examined between the upper welds and motor tube bases for Unit 2 and warehouse spare part length CRDMs. The G-9 upper weld, the warehouse spare, and all other Unit 2 part length CRDMs were NAD. Part length CRDM G-9 upper weld and both welds on the warehouse spare assemb'; %d repairs performed during fabricatiori. These results potenilally indicate an indivioual piece problem on the G-9 part length CRDM motor tube base. Prairie Island will follow the root cause analysis, and the inspection results from any inspections performed at other plants utilizing the same part length CRDM assemblies, and will continue to evaluate the need to inspect or replace the Unit 1 part length CRDM housings prior to the next refueling outage.
The spec.ific responses to items A 'brough E below provide additional details on the current Prairie Island corrective action program reviews on the extent and safety significance of the defect found in the Unit 2 part length CRDM at location G 9, serial number RA-70-090. Where inadequate information is available to fully respond to your request, a schedule for when that information will be available is provided.
Item 1. A:
The following Tables provide a summary of the nondestructive examination (NDE) inspections that have been performed to analyze the flaws in the Unit 2 part length CRDM G-9:
NDE Inspection Performed on Prairie Island Unit 2 Part Length CRDM CRDM Lower Motor Tube Base Weld Area L c*t' a G9 l-7 G-5 E7 Spare CRDM RA70-090 RA70-023 RA70 014 RA70-031 RA70052 Sedal Number UT ABB/NSP ABB ABB/NSP ABBINSP ABBINSP Flaw identined NAD NAD NAD NAD RT uoSiNSP Flaw identifed Visual ID
- West, NSP NSP NSP Cract Identified No Flaws identified No Flaws identified No Flaws identified Visual OD NSP / West.
NSP / West.
NSP / West.
NSP / West.
NSP Crack identified No Flaws identined No Flaws identifed No Flaws identified No Flaws identified PT NSP No Cracks identified Dye)
RAI Response F:bruary 23,1998 Pig) 9 CRDM Upper Extension Weld Area Loc *'*a G9 l7 G5 E-7 Spare CRDM RA70@0 RA70423 RA70414 RA70031 RA70052 Serial Number UT ABB/NSP ABB ABB/NSP ABB/NSP ABB/NSP NAD NAD NAD NAD NAD Visual OD NSP / West.
14SP NSP NSP-NSP No Flaws identified No Finws identified No Flaws identifed No Flaws identified No Flaws identified RT MQS/NSP No Cracks identifed l
A visual examination was performed with the unaided eye directed onto the exposed OD and ID surfacas. Video examination of the ID was performed on the ID of part length CRDM l 7, G-5, and E 7 at Prairie Island using a Welch Allen fiber optic probe at a magnification of approximately 10x. A visual examination of the OD on G-9 upper and lower weld area, G-5 lower weld area and E-7 lower weld area has been performed at Westinghouse using a stereo microscope at approximately 6x magnification. After sectioning G-9 Westinghouse visually examir,ed the ID surface using a stereo microscope at approximately 6x magnification.
A PT examination was performed by NSP on the ID of part length CRDM l-7 using a solvent removable fluorescent dye penetrant. The examination procedure ISI-PT-3, Rev. 7 was used, which incorporates the requirements of ASME Section XI,1989 Edition and vas qualified in accordance with ASME Section V, Article 6 of the 1989 Edition.
RT examinations were conducted on part length CRDM G-9 by vendor MQS. The RT photos were reviewed by NSP's and MQS's level lli RT specialists. See Attachment 1 for further details of the RT examination.
The UT examination technique was based on previous industry experience with detection of IGSCC. These examinations utilized a calibration blocP configuration similar to the CRDM housing with the exception that the calibration block was constructed of 304 material without any weld configuration. The calibration block has an ID notch in the area that the visual indication was found.- The sensitivity established by use of the notch was maximized at 80% screen height. Scanning sensitivity was increased to a 10 - 20% base line noise level. Transducers used were 2.25 and 5.0 MHz,0.25" diameter with a refracted 45' shear angle relative to the scanning surface.
Additional inforrnation on the UT examination techniques is provided in Attachment 3.
Item 1.B.(1h
- A UT inspection on the defective G-9 component was performed as part of the Westinghouse metallurgical examinations. The UT results indicated circumferencial
RAI Response February 23,1998 Peg') 10 cracking 360 degrees around the inside diameter surface. These UT inspection results are being verified as part of the remaining metallurgical examinations.
Item 1.B.(2):
The UT analysis of the part length CRDM G-9 remaining ligament was estimated to be a contiguous 360 degree indication with varying depth profiles (see Attachment 2). A detailed exrenation of the techniques used to detect and messure the ligament size is provided a
.achment 3.
Item 1.B.(3):
The information requested by item 1.B.(3) will be addressed with the completion of the metallurgical examinations and the root cause evaluation. The ligament thickness measurements are to be completed in the Westinghouse Lab by March 9,1998. The root cause report will be completed by April 6,1998.
Item 1.C:
No flaws are postulated in the Prairie Island Unit 1 part length CRDM housings. With clean ultrasonic testing results on all but the defective CRDM, potentially unique weld processes on Unit 2 G-9 part length CRDM due to different heat batches for the butter layer and groove weld wires, different weld personnel from Royal Industries performing the defect G-9 weld, a rework of the defect weld, and no welds in the full length CRDM rod travel housings, no flaws are considered to be present in Unit 1 CRDMs. As stated above, Prairie Island will follow the root cause analysis, and the inspection results from any inspections performed at other plants utilizing the same part length CRDM assemblies, and will continue to evaluate the need to inspect or replace the Unit 1 part length CRDM housings prior to the next refueling outage.
Based on the preliminary root cause analysis by Westinghouse, the Unit 2 defect showed no growth mechanisms. The Unit 2 crack should have been stable under normal operating and design basis loads.
With clean ultrasonic testing results on nine of ten samples, no leaks observed on Unit 1, increased monitoring of Unit 1, no growth mechanisms and no service induced flaw mechanisms identified in the defective piece, and having performed hydrostatic testing that potentially bounds the emergency loads placed on the equipment, the safety i
significance of the defect found on Unit 2 part length CRDM G-9 to Unit 1 is currently believed to be minimal.
Item 1.D:
Considering the potential generic implications of the Unit 2 part length CRDM leak, the Unit 1 RCS leak rates, containment sump run times, and containment radiation monitor trends were reviewed. Stable trends were found on all parameters. The Unit 1 reactor vessel head was recently inspected for leaks upon startup from the December 1997
(
RAI Respons)
F:bruary 23,1998 P ge 11 refueling outage. No Inaks were noted. The inspection procedure specifically requires a sign-off for satisfactuy inspections at the intermediate canopy seal weld locations.
When the UT of part length CRDM G-9 showed reduced margins to the critical flaw size calculations, additional monitoring of conditions of the Unit 1 RCS were put into place.
Enhanced monitoring of the RCS leak rate and radiation monitors was established on February 9,1998. The enhanced monitoring is summarized below. The extent of the enhanced monitoring will be periodically assessed against the results of the ongoing metallurgical examinations and the root cause evaluation and the monitoring may be increased or reduced as deemed appropriate.
The following is a summary of the enhanced monitoring and actions taken:
- 1. Operations personnel were instructed to perform the RCS leak rate calculation each shift. Previously this inspection was performed daily. This inspection also includes review of the leak rate trends, sump run times and containment radiation monitor trends.
- 2. Containment radiation monitors 1R11 and 1R12, and containment humidity were put on continuous trend display in the control room. Increasing containment activity would be indicative of RCS leakage. Should any of these trends increase, a RCS leak rate calculation or containment inspection could be conducted.
- 3. Surveillance was also increased on the Containment Fan Coil Units condensate collection tank levels. Increased condensate would be indicative of higher humidity and possibly of a leak inside containment.
- 4. Operations personnel were instructed to inforrn the Operations Genaral Superintendent if the RCS leak rate should increase above 0.1 gpm, or of any increasing trends on the monitored parametera. Plant management would then discuss the need for containment inspections to determine the source of the increased leakage.
- 5. A containment inspection was conducted on February 11,1998. A visual inspection of Unit 1 reactor head was performed. No leakage was noted from this inspection.
- 6. Do umentation was reviewed, verifying operations personnel were trained recently on Loss of Coolant Accidents (LOCA). Requalification training completed on June 27,1997 included review of the LOCA emergency response procedure and related simulator scenarios. Additionally, recent requalification examinations included LOCA simulator scenarios on 9 of 24 tests.
- 7. Awareness training was also conducted to refresh the operations department personnel on the LOCA procedures.
l RAI R sponse February 23,1998 P:g) 12 I
Item 1.EJjji The limiting loss-of-coolant-accident associated with a failure of a part length control rod drive mechanism (CRDM) is a small break loss-of-coolant-accident (SBLOCA) af up to 4 inches in diameter.
A review of the figure in Attachment 4 was done to qualitatively assess the potential magn:tude of the SBLOC A. The figure shows the lay out of the part length drive rotor assembly relative to the location of the motor tube base weld (the location of the crack found on Unit 2's CRDM housing G-9). From this figure it can be seen that most of the rotor assembly, and especially the thrust bearing retainer assembly, cannot be ejected past the inner taper of the motor tube housing. Thi3 will limit the effective break area to the annulus around the thrust bearing retainer assembly plus the area between the thrust bearing retainer assembly and the leadscrew. This effective area will be much
-less than the 4 inch inside diameter of the motor tube housing and will reduce the break flow rate and corresponding loss of RCS inventory. Westinghouse confirms that for breaks that could result from the failure of a part length CRDM housing, the design basis SBLOCA is bounding. The design basis SBLOCA for Prairie Island has over 1000 F margin to the limit of 2200 F.
Itsm 1.E.(2):
As part of the NRC approval of the Westinghouse SBLOCA methodology it was demonstrated that the limiting SBLOCA location is the bottom of the cold leg. A break at this location tends to prolong the time period where break flow is predominately liquid, which results in a higher vessel inventory depletion, lower core mixture levels and lower system energy removal. The combination of these affects yields higher peak cladding temperstures than would be experienced following a failure of a motor tube base weld. Also, for a break in tha upper head region credit can be taken for the safety injection and accumulator flow in both loops injecting into the cold legs as opposed to spilling out of the break at the bottom of the cold leg.
As discussed in the response to question 1.E.1, Westinghouse has confirmed that the design basis SBLOCA analysis bounds a postulated failure of part length drive housings for break areas of 4 and 8 inches. This was done by running sensitivity cases to quantitatively address the failure of one and up to four part length control housing failures. The cases assumed a 4 inch and equivalent 8 inch vessel herd break. From the top core node vapor temperature plots (Attachments 5 and 6) only a minor heatup occurs for both runs, but the temperature does not exceed the steady state temperatures. These results reconfirm that a SBLOCA in the vessel head is bounded by the standard SBLOCA limiting break location assumption of a break at the bottom of the cold 'eg.
~
RAI Resp;nse Fcbru:ry 23,1998 P:g)13 item 1.E.(2).a:
Assumirg a circumferential failure of the part length housing assembly at the motor tube base weld, there is the potential for portions of the rod travel housing and the rod position indication stack to become missiles.
Item 1.E.(2) b:
The consequences of a rod travel housing circumferential failure are addressed in the Prairie Island Updated Safety Analysis Report (USAR), section 3.5.4.1.3. The USAR states the broken-off section would be ejected vertically due to the vertical driving force and guided upwards by the positinn indicator coil stack assembly and the drive shaft.
The travei of the housing would be limited to less than two feet before it came in contact with the missile shield dissipating its kinetic energy. The USAR also states that the water jet forces would continue to push the broken-off piece against the missile shield. The conclusion in the USAR is that a circumferential failure will not cause damage to adjacent housings that would increase the severity of the initial accident.
The conclusion above is supported by the following:
The original design of the reactor missile shield included calculations of the missile's energy and penetration into the missile shield. The missile shield is a steel lined one foot thick concert slab located approximately one foot above the top of the rod drive housings. It is designed to withstand the impact of an ejected drive shaft and the drive shaft plus drive mechanism.
Assuming a complete failure of a part length drivo housing at the motor tube base weld of a part length drive housing, there would be insufficieni clearance betvieen the location of the break and the missile shield for the leadscrew extension to be completely sjected from the lower part of the drive housing. Therefore it is likely that the broken off section of housing would be ceptured.
The water jet from the break would be directed upwards and would not directly impinge on the adjacent drive housi gs nor any lines penetrating the reactor vessel.
As pointed out in the response to question E.1 above, the internals of the part length drive assembly aloag with the drive shaft would be captured within the adapter portion of the drive housing thus significantly reducing the break flow area and jet forces.
The physical lay out of the plant is such that the failure of a part length drive housing can not damage any engineered safeguards components.
Item 1.E.(2) c:
The dynamic Offects the jet forces would have on the resultant missiles was calculated per Section 5 of the Westinghouse Report " Protection Criteria Against Dynamic Effects i
0
- RAI Respor so February 23,1998 Pago 14 Resulting From Pipe Ruptures" dated January 6th,1970. A copy of Section 5 is contained in Attachment 7.
The following discusses NSP's understanding of the review procedures in Standard Review Plan Section 3.6.2:
Section 111.1 This section reviews the locations and configurations of breaks in high-energy piping. Since this Reque st for Additional Information is focused on a failure of a part length drive assembly, review of the location is not necessary.
Section 111.2 This section reviews the analyses of pipe wh.pping for impact on J protective barrier or component important to srfety. There are no bends or changes in direction of a part length housing and therefore the jet thrust of a complete failure would not result in pipe whipping. lhis section of the Standard Review Plan also states that:
i "An unrestrained pipe should be considered capable of causing circumferential and longitudinal breaks, individually, in impacted pipes of smaller nominal pipe size, and developing through-wall cracks in equal or larger nominal pipe sizes with thinner wall thickness..."
This statement implies that impact due to pipe whipping on pipes of the same size and wall thickness does not need to be considered. This would be the case for a failure of a part length drive impacting on a full length control rod drive mechanism.
Section III.3 This section addresses the jet impingement forces to show that they will not impair or preclude essential functions of nearby safety-related structures, systems, and components. As pointed out in responses to item 1.E.(2) u above, ;he jet forces resulting from a complete failure of a drive housing would be vertical and therefore would not directly impinge on any safety-related structure, system, and components. The affect the jet forces would have on the drive housing missile was addressed as part of the reactor missila shield design and discussed in the USAR section 3.5.4.1.3.
Section III.4 c
This section assures that the analyses of pipe break dynamics include the effects of both internal reactor pressure vessel asymmetric pressurization loads and expand asymmetric compartments pressurization loads, ts appropriate. The
'a RAI Responso February 23.1998 POge 15 inclusion of these loads does not alter the discussion relative to the previous sections.
Item 1.E.(3):
The Prairie Island licensing basis for the SOLOCA assumes failure of one ECCS train.
Since the SBLOCA analysis of record meets all the design criteria, one train of ECCS pumps is sufficient to mitigate the consequences of SBLOCA up to and inclusive of the failure of a part length housing failure. Prairie Islands ECCS configuration consists of one safety injection pump (high head) and one residual heat removal purrp (low head) per ECCS train.
Item 1.E.(4):
Item 1.E.(4) requests the minimum number of control rods required to bring the reactor subcritical. NSP has interpreted this to be a request for the number of control rods near a part length drive that could fail to insert and stil; aring the reactor suberitical with the remaining rods. The NSP Nuclear Analysis and Design department has calculated the number of rods that could fail to insert at the end of hot full power for Unit 1 cycle
- 19. The order in which the failures were assumed to occur is shown in Attachment 8, and was based on maximizing the failed control rod worth while staying close to the failed part length location. The results of the calculations are shown in Attachment 9 as a plot of K.n vs. the number of failed rods. From this plot it can be seen that the best estimate number of rods that could fail to insert and still be able to bring the reactor subcritical is 9 rods. The remaining 20 rods are assumed to operate as designed.
4 Response to February 18,1998 Request for Additional Infonnation Radiographic Examination Details i
i L
1 4
4 I
s i
4 k
i i
i I
e.9 X Ray Info, r +$= * =
g==*s' h.
r a** ' =
/
/
d seen= wo.e,us
[ 'ML' 12" l
/
ci,,in i,YM."a%
'i P
,v"- d _.L _.,, Pg u %e O
T u
al.e u ->
L
(/ ~ \\, 5LQ.\\2
--fW (11 "i t wTF I
u t
s a ~z
_n,
\\
'd@d'c E 'l,'MS,"y M
- t. n,......u.,, m,;
pso
,u..u.
I'I'IMM O' M
- 6 l,Qo tiitco'.ittelt tEtcu J
'W Gource-We6d Film Arrangement End view 5(t View Kodak R Film v-sourc.
-pt.
Indication was Used for Ex).osure present on
\\
/
i film when using
\\
\\
this arrangement
/
/
g i yg and a 30" source
, y r
to film distance
-,t 7
--, x e--
/
I 5
\\
l' 1
I 5
\\
n
,'j ~,,;f,1 h a,- -f.-,'--
Egosure Arrange nt When source source Indication could was placed Locat$a be seen when on this side A
O
" ure a placed of weld no
-n.
n indication was grj --
fIm present on film
/,
g
\\
s I
/,
g E 3
/
/
(
g l V
/
d V
M ym Excesure Arrangement ATTACHMENT #1
..._~
.=
Renoonse to February 18,1998 Request for Additional Information G 9 Remaining Ligament Estimate
.~r
~
_A a
g 4
b ll
\\
\\
l T
y on i
ll l
i i
i 4
i:
i i
i f>
ll' M
f l
l
(
i i
ii l
1 l
I i
~ OSZ i
i l
l l
l
=
- ll:
b m
2 l
I e
3 033 I
I 8
i i
4 1
gg
{
e E
U l
4
'oei j i
,, f b
l l
b l
l l
I l
l i,
I l
g
' ot6 4 3
g l
l i
i
' 02:
i i
s WI l
l l
oot l
1 I
I I
I Tk i
i 09 l
i I
i
,i, oe i
i o,
4 f
i i
' of I
I i
i r
y e'
l 8
R R
a g
s s
a kWh MM MM i
O a
I Response to February 18,1998 Request for Additional Information Description of the NDE Techniques and Qualification performed by ABB 1
i 1
e W
d
_______._..___m.______
._m-8 WJut?722 ABB AMDATA F-973 T-e19 P-902 FEB 23 '98 19:46 I
l-1 Prair4tenandUnh1 Anecesment i.
L 1A. Descripdan of 6e NDE Technignes and Quahfwnas
)
The NDE melbods that were used to evaluees 6e ;adicadoes in the CRDM housings are derived Scan sessbods dessesessed for IO9CC detection. Two specise melbods beve been use.d; 45 doyee sheer wave =
inspection and d5 doyee leagiandbal wave dad element testing The former is designed to be de 10 career erup siganis and Ibe Laser is designed to denen tip ddirection eigents Decense of the angle of the taper (-13 doyees), tbs effectm been angle for dw sheer wave test at to weld ID was _ --
^ 'j $g degrees Par the tip ddaraction inspection, the angle of the esper was irreleved since this test was not used speciAcaDy to isearrogde the beer serboe. Akbough moeSed to needs the speclSc geomsey of the t
j ocaical taper la to npon of abs supect weld, both Ibane mesbods beve success 6d goes dumash abe EPRI 7""-i- : for IOSCC and Perfonnance Demoartuslos lakistive (pDI). In bot ' ;
t=, the
'~
scamung duentice was toward the weld 9 amine tiener (apper) side. A, ""- "= block was machiaad tem annonitic wakkaa steel to makh the taper geometry.1his block had a==ekhad poove 360 doyees aromid and 0.030" deep to esemblish seasidvity ibr the 45 dayee shear wave test. In adde-there were 8ve EDM noeches at various depths in the block to demonstrate the tip ddfreetica webalepas. From thw EDM acech resubs, a sizing accuracy and RMS error was eale=1*ad 1hsee eising results ashs6edthe l
ASME Cude Appendix $ avguiresneste. A mon detened desenption of these tsobnigt.as is given below.
PrairielanandUnk1 Assessment
[
- 18. Descriptica of the NDE Techniques and Quatharion IshintimePaekmamanas i-A drawag ofIbn overall ebape of the CRDM housing taper region was provided and a mockup (CRDS.I.
P!) was fabricated fmm ausesnitic seeinless sesel la order to periana en ' aformadanal test of the lealdag housing, as well as tlw olber thme beesangsc Thsee inspecticas were perfonned at the Prune Isled Plaut s8er reunoval ofIbo pressung housing asseenbly. The papoes of the initial tesebs was to dommaserste doesction of cucusaisestial cracks and nestber depsb nor length sizing was ar.senpeed with any precieics.
The drawings abowed a dies;adler metalwold joining 403 ferritic mahW= seest to 304 anseemtsc seminisse
- steel with a weld buster of 309 ce the 403 sectica and 30s weld metal. The weld cross section has a noeminal 13 degno taper on the OD wkb a range of thickasseen from 0.151" to 0.723". The initial assenption was that tbs flows would be in the 304 semica in the heat althceed sons, typical of!GSCC and otw failwee paviously observed in sissiler -_n lhe mockup, abown la Figue 1, rephcated the taper region and comnined a cacundseatial ID notch F,
0.030" deep.1here was also an OD notab and a side drill bole tbst was maintly oriented. Nether of these Isant two soflectors were need for the bifonnanna irspection The 0.030" aceCa was nonnal to to ID surfbce at the mid point of the taper, which had a mandaal dt===da= of 0.45", or slightly over 6.6%
through wall.
Based ce preview experienes with lasCC detecnon b stamiess sonst piping -kadaas including
., " - - " --work perforsnod boib for IOSCC Ins' ectica CertiScanor. and Performar.cc Demmestremos imidadve proyass. -hwead by EPRI, trudace sad websiques was selected for his specific t
- appliemlos Because of the conical shape of the OD surface, a anell diameter elemed was roquaed, so a 1/4" esmeest Kraulkramur crysed was used on a miniseurs waise for a 45 degree sheer wave exantimados.
' Ibis techmque has base used for the vanous EPRI 7"" ^, and was =adhd fbr the specific geomsey. (Note that te EPRIIOSCC -; "" =M use pipe wolds that range from 0.7" to 1.2" well thickness and lesser duineter cylindrical shapes.) It is important to recognias that ces fundammental assenption of these inspectica *=chal = used for IGSCC is the existence of a conur trap signal to g
produce a reDectme. For a welding process type fisw, this may not be a valid assenpdos. However, the
+
me v
a an,.
- -. - ~ -
~r.-~.-
.,..n.<-
.,n-
2c N 722
.ABB AMDATA F-973 T-019 P-003_,,FEB {'98,191, O
l l
teclunques are very robust k the deseedon of heded flowe, which essentially act as a disiribuesd group of corner refiscears through es -:
A thid ness.
no enlibration was pesfonned by senhas tlw noech nepoess toe the 304 side,ibst is scanneng Ace 4 the thick put of the taper towud the tin part, to s9% Adl semen height (FSH). Due to Ibe 13 degree taper, this provided an _offective 32 degne angle niedve to the ID surface, which is not optimma, but the beam spread provides omf5cient energy in the 35-45 degne range to provide easedve seasidvity, in e.
L.,
4
. coming aca the th iowed ibe thick esotion, ibe etsective angle is se dessus, which is aceseny the lowest senskivity ar.gle fbr cornw renectors. Again, the been spread premdes sufBeiset energy, but tlw 4
oburved =suli wee sat se sesidvky in sie seen dirootion wm -s da missive to se oem ds.ction.
Prior to inspecting the 09 housing, a spese, unused housing was hispected in the warebeese at Prarle Island Plant. Tbs scemains was performed in both axial dinctions whb the appropdsee sensitivity seeunge and no reportable indications (>20% FSH) was noemd.
l In the first inspection at the plant on CRDM G9, the plan was to scan from the tidek side. However, the provmus weld repak on ilw canopy seal roemoted the scan ares suf5cient'y abat this was not a meenh.gfW inspection. De *=aning was performed with +612dB of gain to or==g====s* for $e reduced senestivity.
The results of this inspecean are shown in Figas 2. There wee one very prong renector contered at 180 degrees (pbat sout) that was approximately 5" long ad a soccod, lower amplitude roSoctor three inches long 9an approximasely 270 degnes (cw) to 30 degrees Dere won lower ampikude signals noted for the entire 360 degrees in addition to tbc two areas with stronger responses.
During ilw inspecdon, wee baportant ibeture of the signal response was ibe effect of skewhis etw treanducer relative to the ladication onentation. la dissimilar metal wPis, it is o9se possible to obtain a falm signal at the weld fusion line dee to d:Serences in acoueuc impedances between ilw two metals ne modud inethod of overcoming this effbet is to skew ttw transducer. False signals tend to diminieb rapidly with skewing due to the nuner type aspect of tim reflective ines. Trim cracks tend to give responses over 1
large skew angles (up to 20 degress) due to the generally faceted nature of crack surfaces.
Uains the ASMB Sectica XI flew evaluation rules (IWA 3300) for this type weld (essescry BO,IWB-1 3523), the two larger reflectons would be considered linked since the separanon was less the twice the maximuen depth. However, tide assessment point was moot, since tim indicmion was that the actual Dew was 360 degrees anyway, ne rejectable fisw depth for volcmetric examination resuhe is 10% ofwall nickness, which was clearly exceeded Het Ceu M. " - Detoedom and Steing
]
L A seceed lampar+== campaign was perfonned at the Wessagbouse Science and Technology Casser k both detectwa and stahg. De C h inspe: con technique was the scene as that used at the plant, but corrected for the respones diffenece aseodated whb the scanang dinction, as described estiler. This time, the response to the Haw, nonnalized to the response to the 0 030" noech, clearly showed a 360 degree
==i-asiaa in several locations, two crack liin neponse were otoved at diffbnat metal pass, ind6eseng
- either multiple flaws of brensking of a crack, ne asiplitude roep wee profl3e for this inspection is shown in Figure 3, and is evident that the respon* txceed the calibration notch response fbr almost the amtire 360 degnes.
In addition, a sizing see wee perfbrand manually using *=d=37== abet had been quahfied at EPN for 10S00 through both tlw ICSCC ed PDI progreau. This technupw was a 45 dagne dual element lang*=h..awave probe ihr detection of tip difDactica signals Again, dua to the conical geometry, the minimum footprint possible was necessary in orJer so maintam good cot ting for this 6-*6=
To quahfy the sizing tschelqies, a series of Sve EDM notches won fabricased in the opposite 180 section of the mockup block, CRDM-PI A depth cah'bration cave was established using a stainleen steel block with
_ -. ~._
2036882722 ABB AMDATA F *J73 T-019 P-004
' FEB 23 '96 19:47 r
+ +
l/16 diameest side drIDed bones at 0.1" dep6 lectements through the ness of bearest. The $w EDM notches were then sneasund ad ttw remaining desment was ami= mad toe to response. The naaeh=hl and UT messenments an given in Table 1, below. From tido dua, te RMS error value was cabl=a d and the results west compared to its PDI requasasnes in Appendix 8 of the ACME Code for piping inspection. (Desennitar metal welds are not yet covered in his document, so the amannide pip,
.J were esed.). The RMs error is 42 mins, vs a requirement of <!25 mils.
The rage of applkebibry of the tip difbachen maassement was determined fresa the notch responses. At the 1D_surlben,6e H=im= is the ability to discrimind between the reistively large corner trap signal and the==allar tip signal By -i s =. it was found that the manneen separasion requhat was 0.090".
1 1hassfore, any time a corner trap signal is escoceed, it meet be assunned that the flew is at Ismet 0.09" deep frain the inner surfmee j
At the near surthee, component OD, the lunitation is the reduced sensitivity some at depes shallawer than the habac focal depth of the deal element transducer. A roof angle between the two elements establishes a feast mone that provides maannum sensitMey. Using the noeches, it was observed abat the sensitivity was i
+
greatly reduced at depts less than 0.1" froom the outer swface. Accordagly, any reedhg that is less tbse J
0.1" reunaining must be assumed to bd ranainine ligameet.
Table 1 *==ining Ligament Measurenunts Using Tip Diftaction Dual Elanent Tr===dua=r f
Idaa a h: moneurun et UT Error mus mils mils i
97 104 7
63 112 49 24 24 2
209 2sg 79
+
341 350 9
RMS error 41.9 la se hat cell, a series of manual tip diffraction measurements were made nosainaDy every 10 degrees I
around the circunfuence. The row dets for r===minqr ligament as well as dans corrected for RMS error l
are plotted h Figure 4. The row dets indicased an average r==.inmg hgement of 211 mus vs a corrected l
value of 172. Wadi en uh " Wy measured wall thickness of 0.393" at this point 's the taper,
(
approximasely 44% of tbo section is a=d==ead to reunam The acasal randnee f the tip diffractwa inspechen was difBouk to analyze due to Ibe inniemiaan of manual o
inspecnon techniques. Multiple tip signals were observed ttumughout the cross escdon as endt angular position at different depebs sad et slightly different transducer adal locations. This gave the appeanece of multiple arock surfaces, but a dighal imaging system would be neces% to provide any doenus on these
. complex responses Imaging Syeeen Technigee In parallel whb the infonashonal testing, an summased seemote scanner and imaging system was being The equipinent is bened on ibe intraSpect Ukresenic Imagog System and preparedihrfases i. ^'
AMAPS Scanner. This equipment has been quaktied for pipe inspections at EPRI through both the IGSCC i
and PDI proyees. The syment '.s satable for both the 45 degree sheer wave (ooener trap)laspecuan and the 45 degne longmubnal wave dual eleinent (tip diffracnon) technique. A custom track and delivery syssum were fabncated for the 9ecific geotnetry and access for the part length CRDMs. A rarnfind j.
calibration standard was thbricated with EDM notches on the ID and OD. as shown b Figue 5.
I
+,
m.
3036882733 ABB cMDATA F-973 T-019 P-005 FEB 33 '90 19:48 l
i
& m4v advantage to the muerrsated digital system is the ability to digitiae and display the entire -
ultrasonic response with a variety of views, includag the most important cross section view abowing 1~wk of the various tip signals. Also, th' transient nature of the signals from manual exams are no long er a problem with pumanent recordkig of the signal and location in a precise grid of des points.
Based on the assne depth cabbratics) tectex[ue used for the manual inspection, the RMS sizing error was 15 mils. It tmist be nMed that tehse measures.:at s were made with a celibrWwn black fabricosad Dom 304SS while the actualinspecnon was performed from % 403 SS side of the weld. The sound velocity was measured on the maani component and my differeact will be factored into the analysis of the data, ne real capabihty, however, will be demonstrated on a section of the CRDM housing from 09 that has been maintained for this purpose.
The knagbig system was used on a total of six welds or partial welds on February 20,21,1998 at the W Sciece and Technology Cor aer, includag thw lower and upper dieniniliar metal welds of CRDM housings E7 and G5. Also, the 120 degree segment freen G9 was inspecsed as was the upper weld in this boasang.
For the the lower welds, its 45S and 45L Dual were both used. For the upper wolds, only the 45L method was used due to geometdc restrictions from the taper in that regiou. The reouks of these inspections showed abat no other indreiam were desectable ir any of the welds other than those noted in the 09 lower weld desenbod above. For ime 09 lower weld section, a corner trap (inner surface connected) indication was observed by either the 45 degree 6&h! wave transduceer or the 45 degree sher wave transducer for the entire sector.
The depth sizing dets for the 120 degree sector of the lower weld of G9 mwill be provided to the NRC when it is available (equipment is still in the control zone as of this writing)
2036882722 ABB AMDATA F-943 T-007 D-006 IEB '33 '98 15:37
~
~
~~
~'
FIGURE 1 CRDM HOUSI M UT MOCKDP
. - =
4.221
. =.
!=
3.919 '
='
0.150 %_
~
.Pz a
O O
4 0
o T
n cu g
3 1.000 -
ll
_t u
u u
/
/
5.365
=
=-
7 Hole dia.
=
.062 Hole depth
=
1.0 I. D.. groove = 0.030 deep, flat bottog 0.062 wide ABB Combustion Engine 6 ring Nuclear Servic'es EBCD *d 6Ptc ES4 EZP E3J10 TBS TBld WED 880 6P:91 a661-6T-E3d
b$9-1pge 1?,t 49 ABB CDC FIEl.D SEE81CES 423 752 2449 P.034r3 p
.(
Q t
o.
fqI l
II I
I e
o+ e-i e
D b"
- 1
- h
(
a' d
2
.g 1
14
^
d4%
1 i
$17 8
l
\\ "S l
t
\\
ns M
d=
d N
./
q w
Wg 4
e e-wJ gf y
!~
J M4 U
Um g"
W N
a w
k I
h 6o k
N w5.
el e
2 TOTAL P.03 4E:I1 86. EE E33 400-d 400-1 EP6-3 el m 830 CNDE
t l
i i
1 i
)
p HEilt G fn:V5
..Mi@Yy&gjh,jfk!j]ijlj{W[a y{j pig
, d m g
m w F,gi l
c ww wew; 1]s.! inh a l
r
'h.'
j 3
am t-g i
.;-..J#id,s2.1rC!:
e s
.2'l:p'iqp$.$..it@4ca..
t:m z
- +. u,,.ik psis:
unw= w :.
i
$g E..i,7.'
e L,$$t f.,?
ihr+
ifSi c
=
p.L.:.p Tg f
l@yi,.,l"'ICK g
gb E o
9n. _
l 5 j.
r
...,,,,,,,, 1g..,
- f *.
.m l
W W
2E*ad.
.M,id y........mc, 6 5
. h'ygS S
.'t.6 y n, i
3,. r. #. +..
~g' *
- a.
z ' ;': L l
',4
- as
'7 j
Q.
. ;i..,. e
, jpg..,:
y.*.i,N.;Q(i,j.h j f k E
. w,.-
.. r g
- 1 s..
- 4.,. m
. A q.
g.
x'.
' pg.e]
. :.~
+
~
, i'N "
up,. ~-
k 7
g,
}4.:
w 6
i q
l 3
g g
f, c. t.-
g f*
.Y x
i r
r I
w I
a l
~
~.
,~-
i o
4 oc, 4
.s M
1 t
R R
R R
8 8
8 epreadwy posts 800 d 4,00 1 Et'6 elt s 880 W 889 N
.~-.
-4 artil 06.. EC 833
- - - ~
!i ll
<(_
l l
l
' *c e
ogg i,
e o
li.p
' or.
l osz ng
,e i
' h(e t
E l
l l
ore l
l me.
o-l l
l l
l l
l e
i nos yg i
5 on
.i l
l on
- La i
8 l
l l
l l
' ozi
-u.
s, e
l oot 0
i l.
8 i
n l
i i
,i s
i 4
I
'N i
I E
R 6
g g
g O
i I
l l
t l
co:11 86. Ec H33 sco-d 4,00-1 ct6-3 eitGJO Bad ETR Be9C03
FIGURE 5 Call 8 RAT 10N BLOCK FOR AUTOMATED UT OF CROM PRESSURE 1100 SING a
i 4
e s
ALL 6*0 i
'I. D.
.25 TO E 75 TYP -*
yyp >
l d
N_
1 250 l
e
.V L
.087
.175 i
.262
.350 DEEP DEEP -
DEEP DEEP SECTION "B-B "(o-s)
ROTATED UNLESS OTHERVISE SPECIFIED
- 1. DIMENSIONS & TOLERANCES To AstE Y14.5M-1994,
- 2. ALL DllEN510NS ARE IN INCHES.
- 3. TOLERANCE ON:.X=
1,.XX= 1.06 XXX=.010. ANGLES = 0*30'.
- 4. lECHINED SLRFACES TO E 250AA OR BETTER.
- 5. BREAK ALL INTERNAL AND EXTERNAL CORNERS WIT l
Response to February 18,1998 Request for Additional Information Cross Section o/ Part Length Control Rod Drive Mechanism
/
pyh'r lj l4 q l'
/
/'x A@
c[4 v
u.
s se e.
aj Co p
- 3
l
[
7, '
EE x' Qp y
h$
I mn n
5 gp t;, !
mH l
' gf Q
l e i LL
=
\\
4 y' q
4 k
h.
ll 1 U;
j) t i
\\
s LOCATION l
i o
OF WELD o
l m.
8
\\w I
I i
i s
e 2
(
i m
O 1
o 9
U
=
a g
i 8
u y
e M
1
?
I MOTOR k
/
TUBE HOUSING N
b
INNER TAPER
! i) i h5 b
6 i
e v
\\
s
!f "
hw n s
j p
/
zo N
a go l}ys) kh:
8~5 i
!7 1,
>y
/
$ 5
\\
\\
E u$
h@!-4
]l I h 8
/-
s l
THRUST BEARING Q
WJL gTh
/
RETAINER ASSEMBLY l
m::N
\\
i L
$i5$!
{
I 2
3
/
S E
r sx e
/
o E
a
=
- es=pN m
)
g
.- M 2
m v
s w
s c
E 5
i y
E ATTACHMENT #4
$e a
W m
h
I:
ay:
t.1: '
, -n Q::.
Response to February 18,1998 Request for Additional Information 6.
Fl; seM:
LOCA Analysis Results for 4 Inch Vessel Head Break B?'
=.
4 e
ATTACHMENT #5 I
NSP 4-INCH VESSEL HEAD BREAK TVFN 6
0 0
TOP CORE VAPOR m
LA.
~
\\
4
~- -
O g
i-1 Time (s) mu
i 1
Response to February 18,1998 Request for Additional Information LOCA Analysis Results for 8 inch Vessel Head Break i
.__e
._.mm
-+a-aa-4 ATTACHMENT #6 NSP EQUIVALENT 8-INCH VESSEL HEAD BREAK TVFN 6
0 0
TOP CORE VAPOR m
k
~
w N
/
e
)I n.
o.
3 S
\\
W L-Time (s) t
,9
Response to February 18,1998 Request for Additional Information Excerpt From Westinghouse Report
" Protection Criteria Against Dynamic Effects Resulting From Pipe Ruptures" Dated January 6,1970 Section 5," Missiles Associated With a Loss of Coolant Accident"
.q..sy.3.mayny,q.v;gy ym. pef wy7. n.
9..
HISS!!.r.S ASSOCIATP.D WITil A 1,05S OF RP.ACT0lt COOL.AllT ACCIDENT 4
P~
5.
5.1 Proli et t on of Contritnnent Tunct inn The micailes, that might be generated in coincidence with a loss of reactor coolant shall not cause loss of function of the engineered safeguards or loss of " Containment" integrity.
.2
\\
.W was
- 7 The systems located inside the reactor containment have been examined to identify and classify potential missiles. The
)
basic approach'is to assure design adequacy against generation of missiles, rather than allow missile formation and try to contain their effects.
hl Types of missiles that could be postulated with a loss of 4
N reactor coolant accident are as listed belows
- '?
Valvo stems and bonnets (except as noted in Section 5.3.3 under paragraph - Ifisciles Associated
. ; 'i4 With Reactor Coolan't Piping);
4h Instrument vc11s and thimbles; n.
Huto and bolts;
'7 Control rod drive nhaftc and/or houcings.
t 9
.h
' 'l e
.W u.y=?
..s d.
G :
i Catautrophic f ailure of the reactor vecocl, steam generatorn, pre muriver, reactor coolant pump caninga, at d piping leading
- fhe reanen for to generation of minniles tu nut pnntulate<l.
not providing protection for thene types of rfucilen to that massive and rapid f ailure of thos,c components is incredibic I
because of the material characteristics, inspections, quality l
U control during f abrication, crection, and operation, conserva-V tivo design and prudent operation an applied to the particular l
component.
It is the F ES pocition that valvos and the reactor coolant pump flywhcci are not sources of micciles for the reasons
' outlined in Section 5.3.
AEC agreement on the position that j
~
F the reactor coolant pump flywhcci is not a source of misslic has been obtained. Ito feedback on the D ES position on valves has been roccived yet from the AEC. For this reason and to
,f further decrease the risk associated with a loss of reactor coolant, it is strongly recommended that, whencycr practical, "d>
valve bonnet ejcetion be postulated and the general criteria 4'i given in Section 2.1 for the Reactor Coolant System branch k4 pipe rupture he followed. This can be accomplished by adopting one or more of the following approachest 1
i)
Valves can be oriented and located so that no vital equipment or piping or structure is in
- j D
their potential ejection trajectory.
,2 si t
T
,gy
<j w,. g,p 7.g,.
.n y
- a..
s,.,,
_-__:__-_____x-
m wwmp nap.,-ru.m.y,,,.,,.,p.v.y r
G. ).
$1 11) narriero (cor 'rct o nial"8 or nteel plati": or
..,tr
,.. v,
"'I""" " 3 blant m.1t:,) can be locat '"1 hetWC""
vital coulpent t'r litplon or ut rucW.
licdundant econn to tasten bonnetn to V M C iii) body can be provided.
,9 t
The nuts and bolts are of no concern, because of the small i
i amount of clastic energy that can be stored in the bolt ma t e ri al.
Sections of piping are not credibic free missiles, however, consideration is to be given to whipping of pressurized pipinC as discussed elsewhere in this document.
The gener 1 critorion to be followed is that the consequential,
?
ef f ects of a reactor coolant system pipe break, i.e., missiles, forecs, shall not propogate outside pipo reactions, and thrust u
the reactor compartment.
i The principal barrior protecting the containment structure Additional structures, from missiles,is the accendary shicid.
J i
no required, chall be provided to block the path of minnlics which might othervinc cucape f rom the reactor compartmento I
h through openings around equipment, venting holco, or any ot er opening in the minnile barricro.
~
l o
.W.
Jami.iry 6,1970
$~4-The of f ecus of the:m minnlieu on the reactor compartment valls The depth of penetration into and the effect t
shall be evaluated.
of onurr.y tranufer to the concreto structuren shall be analyr.cd.
Thenc structuroc shall be capabic of stoppiny, the potential mioniles, and still perform their functi6n.
Other than for the CCCS which must circulate cooling voter to the vecsci, the engineered safeguards are located outside the reactor compartment and are protected by the same barricts which The ECCS lines which penetrate the protect the containnent.
secondary chield shall be routed around and outside the secondary shield to penetrate the cceendary shicid in the vicinity of the loop to which they are attached.
The steam generator shc11 is considered ample to resist penetra-For the tion by postulated missiles listed in this section.
lower steam generator shell connecting lines, routing of these lines shall be such that they are not in the direct path of postulated minciles.
~
u
^ri.
- t.
.1.- ;2, p
~
M 5 " m "! q : r e;rs ? i n g,,,,,y m;,,,n,
Janutit y. 6. I')70 l'
5 G
$,2
}tlello l', f f e c tj Imp'Let of0 1)
Specifie 9tructur.61 cffcets as the result of missilu mist'llo penetration and 2) structural responne to dynamic impact load, lloth cficcts should be considered in the design of concrete and steci structures to resist missiles.
5.2.1 Pcactration The ponctration formula recommended for general use with concrete in the This formula has roccived general industry,
!!odified Petry formuis 9-acceptance to date and in the velocity range of interest below 'u00 f t sec.
/
The ponctration formula presented it gives rear.onably conservative results.
ir. Section 6.6.2 of Omit-::51C-5(2) is recommended for use with a
structures.
5.2.2 impact Load
.i tither of two methods are reccanended for use in calculating dynamic load The first method is empirical in nature and is based effects of missile.
of the Army on the fittit C of twcudo-clastic strcsc indices to Department The second method in baced on connervation of momentum x
test data m
4
- i$
6 3 principleu(^) and own generally accepted ultimate ntrength dtwir.n utrenn
~ * *
(D)(5) for added contetvnt1:m the u'ic of valuco and ductility f actor s (6) code ba ;ed ultimata utrength deniga utt'em iuy be conuidnd.
$.2.2.1 pseudo-Strean Method Roference 3 presents a method for impact design utilizing clastic stress
. y, The design method prescribed
,7 multipliern matched to various damsgo indices.
d with while having no analytical basis has the advantacc of being correlate Since the actual tests woro conducted a number of experimontal results.
f using ham-or impact on target plates, orily clasto-pinstic deformation o b
the beam was availabic te determine the equivalent static load.
M
,4 The method as described in the reference suggests the use of a dynamic load
- ar and deoth of penetration to determine the equivalent static load.
it Sinco dynamic load f actors accociated with rapidly applied sustained loads for structures which allow maximum on an clastic system are largely lost deficction of two or more clactic deficctionn (v > 2), the use of a dynamic The equivalent static lead to be used for load factor is not recommended.
design is determined from:
rd 1/2 mv
=
wherci n = mass of.the taissilo j
?
v = impact veloetty of the miunilo F = Equivalcut static load
- 7.,'
d = total dinplacement of minnlic af ter irmaet
[
te y."
' y,1 4.g 1, g.,;.4,.g,,.,,, j.g
.' y '.:',.,al. 2 % p't,q p
- -,., ~
+ -.
, r W..q. ',%,a7. Vg. gggf.
g.
a t
m m.
y.
l W " M *7 (5 1 for nondernrmable misolics (2) d=p+6 whero C
p = total depth of missile penetration 6 = total deflection of the impacted structuro For relatively low mass hir,h velocity missiles d is essentially a function
,.,- G 4$
m' of penetration while for relatively largo mass low velocity missile d is For a function of deformation or deficction of the target structure.
potential missiles which are capabic of istge plastic deformation a third term may be added to displacetont which represents the crushing of the
~
missile.
As an upper limit 6 should bo limited f or concreto structures to a u 's 3
}
and u 1 10 for low carbon structural stool structures 6
" '" I (3) e y
and 23 1
6 = deficction of the structure at yield of reinf'orcing s"tec1.
y 1
t The value of r thus determined may be uced as the equivalent static load I
for deutr.n'purponen ac outljned in Reference 3.
It should be underntood that the determination of the final F in to come del;rce on interativo
, s.fr
?M&W,.,, * * - i,a a
..,,, o s
^
.:a
.p~
5.
.I procenn nince a valuu of r munt be deterinined in order to defind 4
......,,g, which in turn determinen F.
5.2.2.2 Connervation of t!c-entum liethod
...r The Pseudo-Stress }icthod described in Section 5.2.2.1 has been critized BI
,3 for its largely empirical naturo with no clear extrapolation to structures
' "'G other than the test beams considered. To overcome this concern a somewha more analytical method is proposed which assumes conservation of momentum principics.
. m v +m v
- I"1 * "2) #3 j
i 3g 22 vhere
= mass of missile m3
]
= cf fcctivo mass of target structure m
y
{
= initial velocity of missile v3
..i:
= initial velocity of target structure (usually equal zero) vy Y
= final velocity of target structuro (plastic impact) v 3 dsy In many The valuen of m. v and g are generally known quantitics.
g 1
' ['
instances the mass of the target structuro is orders of magnitudo larr,or the than the tarr.ct misclica and therefore it is quito unlikely that Study of numeroup q
misnile impact ic fuit by the entire target structure.
...;Q{
. M :-
4 fl,.*f
lt% mr m gp.2, g.,
3,.
,5,
.fanunry fe. 1970 6-9 ioen.
minnlie offeet photop,raphn indicaten a circular damano area approximately 10 timeu tlu' least imp.ict diameter of the minulle( }.
It 1u therefore recommended for ta gut structurca having impact surf ace diameter in execun of ten times the ler nt missile impact diameter that the ef f ective mass of the targct structure be limited to that mass associated with ten times the With m. m ' "I and v thus defined it i
least dimension impact diameter.
g 2
3
'5 from Eq. 4.
is possibic to calculate final velocity v3 Assuming the initial velocity of tho target structure is zero at time of impact the energy absorbed by the target structure may be expresned as
' E = 0.5 (my + m ) "3 s *. ( 5) 2 To determine the equivalent static load for design purpose F=f (6) where F and d are as defined in Section 5.2.2.1.
It should be emphasized that neither the Pscudo-St,rcus nor the Conservatio's of Homentum mo'thods are rigorous or exact solutions of the clanto-plastic renpouco of structures to the dynamic impact of missileo, llowever, the methodo presented uhould give some simpic practical guides to missile impact d.,a
.y X
i denin requirements.
- )
R
- ?
,4
.h
,k,N.
- , Ns sf.es ' 6% '..'
A
- pq'
. * ' d *N,. -
f f
\\Y$
",,.9 s.
y w w w.w + g.N. m s e.v.un.,q p g y w., w ei.v wr. w
- 1 5.
~~
1.1ST or nl:Pr.nl:lici:S r
1.
Amirikian, A. Desir,n of Protective Structurcs, Bureau of Yards I
L' and Dochu Pui 11 cation 14o.11AVDOCKr. P-51, Department of the llavy Washington, D.
C., 1950.
- j
.u 2.
Cottell, W.
B., Savolainen A. W., U. S. Reactor Containment
.j,g
.r Technology, OPJ1L-!! SIC-$, August 1965.
d
'1 4
a 3.
Department of the Army, Protective Desien. Fundneentals of Protective i
1 20icn (flon-t!ucient). T!! 5-685-1.
July 1965.
}
4.
flo r ris, C. 11., lions en R. J., e t. al. Structural Desinn for Dynamic 3
Loads _ tt: Craw - 11111 Book Co., !!cu York,1959.
.6 ll 5.
ASCE Committee on Structural Dynamics, Deninn of structurcs to Resist 1
Nucient Ucanonn Cf fects, ASCE tlanuals, of Standard Practico No. 42,
=.;
p American Society of Civil Engincors, 1964.
.g 2
'l 6.
ACL Committcc 318-63, ACI Standard Buildinr. Code nequire Aentnfor.
j' '
Reinforced Cnnereto American Concrete Instituto, June, 1963.
f;
)
.j<l 7.
C. A. Trexci, Tents and Denirn of itemhproof Structures of Itoinforeett
- 4 3.E(!
.y Concrete, Navy Department, U. S. Covernment Printing Of fice.
.y f
Wanhinr.hounu, 1961.
- ?l4,g
,. '.* N
.,s.k.
. ;.w:
..s.v.,,...p,..,,,.~.'..
'l..c p,g, g,_ p, s'$?....'d' W ll?lY*
rg %gj;;gy a
i,.
~,,7;
~;
,,,. f, s Jtnuta ry (,, l')70 S,
r.g.
a 5.3 Pm: t u t a t i,dgi_9 t en i
- 5. 3.1 Reactor coolant Pump Flywhcol The following precautionary measures, taken to precludo missile formation from the Reactor Coolant Pump flywheel, assure that the flywhcci vill not produce missiles under any anticipated accident condittons.
The flywhcci is'f abricated f rom rolled, vacuum-degassed, ASTil A-533.
a) t b)
Flywhcci blanks are flamo-cut f rom the plate, with allowancq for exclusion of flamo-affected metal.
c)
A minimum of thrco charpy tests arit mado from each plate parallel and normal to the rolling direction to determine that each blank satisfico design requirements.
d)
An HDTT less than 10'F is specified.
c)
The finis,ned flywhccl is subjected to 100% volumetric ultraconic innpection.
4
- I/US+Y
- v.... m. v:w. a -
2v
.+m.,... a,,..n u...... ;.,.,,g. :.,.,,c.n... s a.a.~ w. y.
+.
(
s U.
.lanuary 6, 1970 5 '
O rs-f)
The fininhed masnined bores are alno subjected to magnetic particle or liquid penetrant examin.ition.
Thuuc decian fuhrication techniquen yield flywhecla with a primary stress at operating speed lens than 50% of the minimum npecified material yicid strength at room temperature (100 to 150*F).
a fi4 1:
The reactor coolant pump is driven by an induction motor. Thus its rotatienal speed is controlled t,y supply frequency. I;ormal operating speed of the pump is 1189 rpm with a synchrorious speed of 1200 rpm; however, in acecrdance with tig!M standards, it is designed for an overspeed of 125%
of synchronous speed or 1500 rpm. The most adverse operating condition
~
of the purnp motor flywhcci is visualized to be the loss of outside load situation. During a loss of outside load condition at the blant, the d
et
- trbinc, generator will overspeed and the reactor coolant pump will tend G
t to follow. Analyses of the turbine control system with the utilization
'I; of redundant inlet valves (turbino stop and governor valves) and redundant
,}-
re-heat valvos (re-heat stop and intercept valves), show that the maximum f
%('
turbino overspeed under such conditions will not execed 120% and would
- y not persist for more than 30 seconds. Therefore, with an auxiliary power f.
system arranged with each reactor coolant purnp supplied from a separate
- p dintribution bus an utlpulated in the 11alance of plant criteria and the
,k B
\\
cxpected turbine-generator protection ryntem arrangement for tripping the
.[
4!
auxiliary power syntem from the cencrator on a turbie u trip, even if a bun failn to trip uff the entrenpunding unit it will reach a maximum npced less than 120% of nynchronoua upced for the abovu turbine overnpced condition.
+
= h.k.a.J w.tv..+
.:,:2
. ~ * ' " *
. ls
\\
y g
- h gy2
- 4 ek
(
p e
' WNNGt7.tawf gg.ngy,gg3,gg i
January (e, 1970
'ls
,q 5 -Q r
BurntIng speed of the flywhenin han been calculated un the baulu of Crlf flth-!nelu'u recults
, to bc 3900 rpm, more than three timen tho
- 4' operatlig: n pe.:d.
A fracture mechanien evaluation van made on the reactor coolant pump flywhccl. This evaluation considered the following assumptions:
Maximum tangential stress at an assumed overspeed of a.
, ).f
.dtT 125%.
.g b.
A through crack through the thickness of the flywhcci at the bore.
400 cycles of start-up operation in 40 years.
c.
Using critical stress intensity factors and crack growth data attained
'ff on flyuhcci material, the critical crack size for failure was greater
_.g a
than 17 inches radially and the crack growth data was 0.030" to 0.060" S
1000 cycles.
.$u f
i An ultraconic inspection capabic of detecting at least 1/2" deep cracks from the ends of the flywhcci and a dye penetrant or magnetic parcicle
)
.i. !}
test of the bore both at the end of 10 years will be more than adequate Y]
as part of a plant surveillance program.
x TJ Mi The Reactor Coulant Pump motor hearings arc of conventional design - the y
O radjal hearingu are the segmented pad type, and the thrust bearings arc j
d tilting pad Kingnhurg hearings. 411 are oil lubricated - the lower radial hearing and the thrunt bnaringa are nuhmerged in e,il, and the
'k
.8
.q
'%b i
. 'E 1 rnunt 1.. liohtunnn, "llurnting Tests of Stent
- Turbine Dink Whecin,"
- N Jmiunty 6,1970 1
?
{.
j.
upi.cr radial hearing la vil fed f rorn an trapeller integral with the T'
t h r u.; t runner.
1,uw oil levelu would ninnal an alarm in the control rouin and require r; hutting down of the puinp. 1:at.h r.:otor hearlon contains embedded temperature detector::, and no initiation of f ailure, separate irc,m lohn of oil, would be indicated and alarm in the control room as i
a high bearing temperature. This, nanin, would requiro pump shut down.
t gven if these indicat.ons were ignored, and the bearing proceeded to i
- failuro, su molting point nabbitt metal on the pad surfaces would ensure that no sudden soiture of the bearing would occur.
In this event the motor would continue to drive, as it has aufficient reserve capacity I
i to operato, even under such conditions, llowever, it would demand excessive currento and at some stage would be shut down because of high current demand.
It may be hypothesized that the purep impeller might severely rub on a stationary member and then soir.e. Analysic has shown that under such conditionn, assuming instantaneous seizure of thu irnpc11c t, the pump chaf t would f ail in torsion juct below the coupling to the motor. This would constitute a loss of coolant flow in the one loop, the offect of which is analyr.cd in tho TSAR. Following the seizure, the motor would continue to run without any overspoed, and the flywhcci would maintain its integrity..as it would still be supported on a nhaf t with two bearings.
There are no other credible coureca of chaf t scir.ure other than impeller ruba. Any acir.ure of thn pump hearing would t..' precluded by chcaring of G
tini nraphitar in the hearinn. Any neinure in the nealn would renult in
=
NM v p m e ryc. w,p.,
,y,
,.1 a m u r y (., l'HO (I 9 a chcaring of the anti-rotation pin in tim neal rJun.
The motor hau adequato power to enntinue punp operation even after the above occurrencer..
Indicationu of pump malfunction in the::e conditionn would be initially by high temperature signalu from the hearing water temperatura dctcetor, and excessive No. 1 seal icakoff indicatione respectively.
To11owing those signals, pump vibration icvels would be checked. These would show execesive g,
Icvels, indica:ing some mechanical troubic, and pump would be shut down for investigation.
The design specifications for the reactor coolant pumps includo as a design 4
condition the sttesses concrated by a maximum hypothetical carthquaho ground acceleration of 0.2g.
Desides examining the externally produced loads from the nozzles and support lugs, an analysis is made of the effett of gyrnscopic reaction on the flywheel and bearings and in the shaft, due to rotational movements of the pump about a horizontal axis, during the maximum scismic disturbance.
The pump would continue to run unaffected by such conditions.
In no case does any beating stress in the pump or motor execed or even approach a value which the bearing could not carry.
The design requirements of :the bearings are primar.11y aimed at ensuring a long life with nonligibic wear, so as to give accurate alignment and smooth operation over long perioda of time. To thlu end, the surfaco 9
J l
j' 7,..
.s r
.innua ry 6, l')70 4
7 5.
bearing streficen are held at a very low value, and even under the most severu seismic trannient:: or other accidentn, do not henin to approat:h loads which cannrat be adequately carried for short periodo of timo.
5.3.2 Control Rod Hechanism A failure of a control rod mechanism housing sufficient to allow a control 4
rod to be rapidly ejected from the cora is not considered credibic for the following reasonst a)
Each control rod drive nochanism is completely asccebled and shop-t'ested at 3450 psig, ej
?
b)
The mechanism hous,,s are individually hydrotested to 3107 psig 3
as they arc installed on the reactor vessel head to the head adaptois, and checked during the hydrotest of the completed i'u Reactor Coolant System.
?g A
c)
Stress icvels in the mechanism are not af fected by system A
transients at power, or by thermal movement of the coolant loops.
.1 1
d)
The mechaninm bouninco are made of type 30I. statnicos steci. Thin 1
materl.il exhtuita exec 11 erit, notch toughncsc at all temperatures that will be encountered.
(,
O 6
m g.4
.. ym,.,,.,,,,,
V e
... n....
l f G v
4' L.
In addition, a minnile chield Structure is provided over the control rod whf ch will block any mir.ulles which might be generated i
drive tuechantnm9 in the event of a f racture of the prensure housinn of any mechanism.
y Control 1od t)rlyo itechaninm tfinstle Shield l
a, Th3 analysis performed to identify the potential missiles associated d
With the rupture of a control rod drive mechanism housing is presente 4
in this section.
e 4
The criterion to be followed is that these missilns shall not jeopardize 4
~
It is recommended to locate a concrete slab the containment integrity.
I, With steci f acing on the top of the CRDH housing, as close as possible to i
the housing to limit the velocity of the ejected missiles, to minimire' 1
the protability of missiles missing the sh'icid and striking the containment liner (and in the case of the ice containment, to prevent penceration of the compartment barrier), and to minimize the probability of missiles i
ljj l
ricocheting and damaging other CRDH housings.
II i
s o
It is recommended to design Jthe shield in such a mannor that it will i
behave structurally as a single unit formed by interconnection of the l
'this will prevent lif ting of a singic block and removabic blocks.
i J.,
subsequent drop above the reactor vessel head.
s
..(
3r!
.'y
- r... re..,+.,;;
' *
- s ' d
'f r
.,tw
.; -..,.. c, o 4..
_1
. 3 5 *
- n...
For the ico containment tiv compartinent barrier m:iy alno be used to act au a minalic shield provided the barrier is deninned to resint potential missilo irrpact without significant crackinr, (localir.ed crackinr, limited to 1/4 inch in width).
The assumptions, method of analysis, and results of the calculation performed to identify potential missiles should a CRD>t housing break are summarized heroin. This analysis can be applied to any Westinghouse PWR, using the same CRDt! design and with a reactor coolant design pressure of 2500 psia.
1)
General Three types of missiles are analyzed:
a) Plug on top of the CRD)! housing b) Drive shaft c) Drive shaft and drive mechanism latched together.
.The worst case, assumed for design, is the,following:
The top plug on the CRMt housing is assumed to becomo loose to be accolcrated by the water jet, until it reaches the underside of
- .l u
1
.the minuile chicid and partially perforates it.
.g
l
. MFPNF)w..c'MRrW..KJW.wpe pyr.9 January 6,~ 1970 ~~ * " Y R,
, l-
..3,
-,a p.
t
/4 f.
g '
j f
i, h
'h v, -
4 In the meantime, an opon an the top plun cicarn the break, the drive chaf t and cc.ntrol rod cluuter are punhed out of the core t
j P
by the differential pressure of 2500 psi acrosa the drive chaft.
I, b
The drive shaft and control rod cluster, latched together, are t
il i
j, assumed fully inserted when the acciden'. starts.
.ifk 1
1
.G 1
'y After approximately 12 feet of travel, the RCC spider hits the a
undc ro i ',: of -he upper support plate.
Upon impact the ficxure 1
I arms in the coupling joining the drive shaf t and control cluster
~g vill fracture, completely frecing the drive shaft f rom the contror
.i 4
rod cluster. It was assumed that the control cluster would bc 6
3
'l completely stopped by the upper support piste.
fhe drive shaft,
].}
however, continues to be accolorated until its top hits the missile
'ejj
- t
,4*
shield.
~1
- i hl At this time t'.c shaf t pushes the plug through the non-perforated
.n T
'Y M
layer of steel and into the concreto slab.
g h
.h 2) guyinn Plur F.jcetion f
The reactor coolant diccharge flow rate from the break has been calculated using the nurncil equation. The coolant pressure has f.j I.
4 58
..lNa h 3, a w u.u
- u'Cb, %Sb "
5
.!..4 4,,
"!2*.2 h ' :'
r
~ - -
..y m
January 6,1970
[
5 m been nsnumed constant, at the initial' value. A fluid exit velocity of 534 ft/nce has been calculated. The velocity of the plug was calculatud equating the iiicresce of the plug momentum to the decrease of the vator jet momentum. No spreading of the water jet was assumed. The char.ictoristics of the plug and the velocity reached by the plug as a function of the trasil ate summarized in the following tabic.
TABLE 5.3.2-1 D
Plug Weight: 11 lb.
t.
P' lug 0.b.:
2.75 inches Travel, x Velocity, v Kinctic Energy (ft)
(ft/sec)
(ft/lb) 1 240 9,750
.2 335 19,000 3
370 23,300
]
4 415 29,200 5
440 35,000 Q
sl}
, ;c,-
- [I**T
....i..
,i '
ih 5
(
k
{p D[ '_
k#
e
.lanuary fe, l'170
.g 7
.: w Q
yw E
The depth or pen tration lit the minn11e chic!d nteel plate han been calculated an 111uutrated in O!!Nf.-NSic-5 page 6.150, using j
a value of 60,000 pai for the tarcet plate ultimate tenuile strength and 14 inches for the side icngth of the "squarc window."
..,j a
the depth of penetration is 0.5 inches and 0.6 inches, for the missile shield located 3 ft. and 5 f t. above the top of the CRDM
~
housings, respectively.
3)
Drive Shaft Ejection The drive. shaft and rod control cluster (RCC) have been assumed to -
t.
g be acccierated by the dif ferential pressure of 2500 psi across the A
h f:
drive shaf t.
After a rod travel of 12 ft., the velocity of the s
drive'shaf t and cluster (W = 270 lb.) is 130 f t/sec. The drive e
j shaf t has been assumed to become loose af ter the impact of the RCC spider on the upper support plate. No credit was taken for s
thc energy absorbed in breaking the flexure arms.
.f I
e t
{
Upon impact, the RCC (with the spider) is assumed to be completely f
stopped by the upper support plate. The drive shaft (with the
)
disconnect rod) (U = 120 lb.) is ansumed to be further accelerated by the dif Cerential prenourc of 2500 pai.
A cicarance of 1 foot is accumed betwecu the top of the drive shaf t when fully withdrawn
\\
ned a,e top or die h e inn. Therouwinnuweaqathe J
s.
I'
'}
...,.. > :. tv Wi.,
=
1 r d dida MSL h
M
.f mns iry (i, l') /0 g '
(
!.O h- -
characteristics of thic minnile and its velocity as a function of m
travel out of the huuntnn. Tabic 5.3.2-2 alno gives the micotic shield uteci pinte and concrete niab-Lhicknen::ca required to stop the drive shaf t as a function of the distance between the missile shield bottom and the housing top. The thickness of a
the steci plate is oscumed cont, tant and equal to 1 inch.
.n The critical Kinctic energy required for penetration is calculated as recommended in ORNL-NSIC-5 page 6.158, using a value of 60,000 psi for the target plate ultimate tensile strength and 14 inches for the side length of the " square window." A value b
of 48,000 f t.lb. has been found for the perforating energy for i
the 1 inch thickness of stocl. This value has been deducted from the drive shaf t Kinctic onec ;y at the time of impact and the new s '.
reduced drive shaf t velocity has been determined. The depth of penetratic. ir. the concrete slab was :alculated according to Nav.
Docks p-51, April 1951 and a slab thickness of 3 times the depth of ponctration has been chosen as design value.
h, 4) llousinn plun and Drive Shaf t Impact on the Same Itisnile Shield Spot For thin caso, whlch in the design case it han been ancumed that the plug perforates partially the ntec1 plate as indicated in 2).
Then the drive shaft hits the p1un and pushes it through the
<tf non-perforated nteci pinte layer and into the concrete. Two solutionn M
{
can be adopted. The firnt is to une the concretc tilab thickneun j
+
M va
)
'.j4 6..
-ikhY
- n. "... - 4..#
h k jhli'5* h $br Y Y $W Yl$$5 h b Y$NE
+.. -
C$
9 TABLE 5.3.2.
T DRIVE SHAFT EJECTION Dianei:er = 1.75 lii.
~
o i.ength = 300 in.
e j
k*eight = 120 lb.
Missile Shield **
Missile Shield **
e' '
Drive Shaft Travel Drive Shaft Drive Shaft Steci Plate Additicnal Concrete d
k Outside llousing*
Velocity Kinetic Energy Thickness Slab Thickness Y'
(ft)
(ft/sec)
(ftxib)
(in)
(la) si.
}[
151 42,900 1
u 1
1 2.5
+;.
2 162 49,000 1
10 3
171 55,000 t.'-
f C...
4 179 60,2000 1
16 5
189 66,500 1
25
[
269 134,7000 10 i
G=
t.
Distance from top'of rod travel housing to bottom of missile shield.
h..)
,5j.
These thicknesses are indicative only, and shield design can optimize between 16 steel and concrete thickness.
- 1' io at~;-
- c l
T lM:.
t.
k
- 1:!-:
y.
.. +
a.
-. ~ -.
6 ;
found in 3) and to increanc the steci plate thicknens by the plug perforation depth. This will overer.timate the concrete thickness because the drive shaft punben the plun inutead nf penc-trating directly (plut 0.D. = 2.75 inchen, drivu chaf t 0.D. =
1.75 inches).
The second solution is to keep a 1" stcci plate thickness, and g
increase the concrete slab thickness by 2 inches and 7 inches for the missile shield located 3 feet and 5 feet above the housing top, respectively.
i 5)
Ejection of Drive Shaft Latched to Drive Mcchanism n
s Y
The velocity of this missile has been calculated as.in 3). The s
n missile characteristics for this case are summarized in Table 5.3.2-3.'
,]
32 TABLE 5.3.2-3 T
.I r.
i,9 Missile Weight: 1500 lb.
51y Impact 0.D.:
3.75 inches
'I M<
Travel -
Velocity Kinetic Energy (ft)
(ft/coc)
(ftxib)
I 1
14.3 4,600
.3 2
- 10.2 9,200
. [-
3
.24.8 13,800
'N 4
28.6 18,400.
5
.32.0 23,000 u
. i. y,
~
- E M -'
- ' =.
a
-g
, 'e.-+.
...,j
. -....... w.w
_ x 7my.w;. gm.-
_ _ _ -__- -____ __-___ '.9 nn u a ryVD
,( e 100,000 ftxth.
The critient Kinutic energy for perforation in
. ',. - ~
,.,y a.
Therefore, no perforatton in expected.
The pocolbility of misnile shictd dJcplacement undur impact has Should the miccile shield be located 3 feet above been analyr.ed.
the housing top, a maximum vertical missile shicid displacement This of 0.3 inches has been calculated assuming an clastic collision.
e-displacement should not present any probica.
6)
Jet Thrust on the Missile Shield The jet thrust is 6000 lb., the weight of the missile shield is
...-6 more than 50,000 lb. if located at a distance of 3 f t. or more above the housing top, therefore no overturning vill occur.
3 ',
9
'e m-S E,
- , se
?%
e n
T s
,e S.
?
.s s 6.-
Response to February 18,1998 Request for Additional Information Core Map Showing Assumed Control Rod Failure Locations Followino Separation of Part Length CRDM G 9 a
4
'l 1
4 l
l
.w.
ATTACHMENT #8 PRAIRIE ISLAND UNIT 1 Cycle 19 1
2 3
4 5
6 7
8 9
10 11 12 13 A
A A
19 16 SD D
SD 21 17 13 c
c D
A B
B A
F D
PL c
PL D
Fall G
25 3
6 A
B B
A 29 5
c c
24 12 SD D
SD 22 18 14 A
A 20 16 Bank M -
order This attachment shows control rod locations and the order in which they are assumed to fall in this analysis. Failed rods are rods that cannot be inserted into the core due to damage caused by the separation of a partial length control rod drive mechanism from the reactor head, in this analysis, the mechanism in location G-9 is assumed to separate.
i
O *gD Response to February 18,1998 Request for Additional information i
K.n vs. Number of Failed Rods Following Separation of Part Length CRDM G 9 4
1 1
5 i
4
v f
li
[
[1>!
lttt L
[; '
[tI
.: 2
.?
9
+
2 82 72 6
O-2 5
+
2 4
2 3
2 2
2 1
+
P 2
FH 20 E
t 9
a 1
t 9
re 8
)
s t
1 r
n es T
i 7 n o
+
i N
1 t
o t
l E
a i
l 6
ia 1
F F
M
(
t 5 s a
+
d H
h o
1 t
R C
4 d s
+
1 e
d A
o ia l
R
+
3 F i
1 f
o I
s r
A v
2 e e
1 b
f m
fe u
K O
1 1
N 9
1 0
1 o
1 P
+
9
+
8
^
7 6
+
5 9
4 e
3 o
2 1
O 4
3 2
1 0
9 8
7 6
5 0
0 0
0 9_
9 9
9 9
1 1
1 1
O 0
0 0
O r E%x c
~
1 llll