ML20206N321
| ML20206N321 | |
| Person / Time | |
|---|---|
| Site: | Duane Arnold  |
| Issue date: | 12/07/1998 |
| From: | Hiller J External (Affiliation Not Assigned) |
| To: | NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM) |
| References | |
| REF-PT21-98 NUDOCS 9812160243 | |
| Download: ML20206N321 (12) | |
Text
- -. .- -.. -- . . . - . . . . . . - . . ... -. - - - . - . - .
],
EL:It
. k=11 RALPH A. HILLER COMPANY 6005 ENTERPRISE DftwE EXPORT, PA 15632 TELE:724 32!E1200 FAX: 724 7331825
. December 7,1998 Document Control Desk United States Nuclear Regulatory Commission
- j. Washington, DC 20555
Reference:
Hiller Document 98-NRC-004
Dear Sir,
L As the responsible officer for the Ralph A. Hiller Company, I advised the Nuclear Regulatory Commission of a possible defect in a letter dated May 19,1998, and updated the status of our investigation in a letter dated July 17,1998. Copies of these documents
, are included with this mailing. This was per the applicable requirements of Title 10, Chapter 1, Part 21 of the Code of Federal Regulations y As required by 10CFR21, enclosed is our final report, which indicates our position that
- _ no reportable defect exists. Our recommendations are included in the report as applicable.
Should you have any questions, you can reach me at (724) 325-1200, or fax me at (724) l 733-1825.
i' Thank you for your attention in this matter.
i Yours truly,
/.)
olp Chief Executive Officer Enclosure (1) " Report on the Piston Failure of MSIV Actuator at Duane Amold Nuclear Station" cc: file 9812160243 981207 gon noocn ..
ogt]3 b'- a
l Report to the United States Nuclear Regulatory Commission Final report on the Piston Failure of MSIV Actuator At Duane Arnold Nuclear Station Manufacturer: The Ralph A. Hiller Company J. Randolph Ililler, Chief Executive Officer Phone (724) 325-1200 Fax (724) 733-1825 12/07/98 The Ralph A. Hiller Company Page 1 of11 6005 Enterprise Drive Export, PA 15632
Table of Contents Introduction 3
- Operation - 3
. Design Basis 4 Verifications 4 Material Testing 4 Finite Element Modeling 4 Further Testing 4 Analysis 5 Loading Above the Ultimate Strength 5 High Cycle Fatigue 5 Low Cycle Fatigue 5 Low Cycle Fatigue with a Pre-existing Crack 6 Impact 7 Summary 7 Appendices A - Preliminary notice of possible 10CFR21 reportable defect from J. Randolph Hiller to the U. S. Nuclear Regulatory Commission, Dated 05/18/98.
12/07/98 The Ralph A. Hiller Company Page 2 of 11 6005 Enterprise Drive Export, PA 15632 i
L .
l tr!1roduction l
l On May 18,1998 the Ralph A. Hiller Company reported a possible 10CFR21 reportable defect. A copy of the letter is included as appendix "A". What follows is the final report as indicated in that letter.
- IES Utilities Inc., Duane Arnold reported that the Inboard Model SA-A101 Main l Steam Isolation Valve "A" (MSIV) Actuator en the C loop of their Main Steam Line showed evidence ofleakage, in both directions past the pneumatic piston in excess of Duane Arnold's maximum. The actuator was returned to the Hiller Company's production facility for evaluation. Evaluation verified the leakage and discovered a fracture of the pneumatic piston (part number 629-10382) was the cause of the leakage. The crack consisted of a cross sectional (thr ough) crack in the web and reinforcement ribs near the center hub involving a 2180 of the hub area circumference.
1 Initial examination of the spare actuators (2) at Duane Arnold found that one had no I cracks while one had partial cracks across two of the ribs. The actuator with the partial cracks on the piston, "B", had also been in service as the outboard actuator on same Main Steam loop (loop "C") as actuator "A" The Model SA-A101 MSIV Actuator is a Basic Component, and therefor safety related. It consists of a 20" diameter bore pneumatic cylinder in tandem with a 5" diameter bore hydraulic cylinder. The failure (safety) mode of the actuator is the rod exter.ded (valve closed) position. This position is achieved via an external spring pack and is assisted by the pneumatic portion. The hydraulic cylinder is designed for speed control.
The following are the findings of our investigation into this anomaly.
Operation The inboard actuator "A" had a crack, which had propagated through the pneumatic piston, causing leakage through the piston. Actuator "A" had been manufactured in1989 and put into service 1990. It had been refurbished per Iowa's planned maintenance schedule in 1996 at the Hiller Company facility. At that time it was postproduction tested which verified that the piston did not leak. It has been estimated that the piston was exposed to twenty-one (21) pressure fluctuations from 100 PSI to 0 and back to 100 PSI per fueling cycle for installation, set up and verification stroke testing. The piston was installed for four (4) fuel cycles. The piston was exposed to stress fluctuations in service due to compressor cycling. There is no existing mechanism to verify the number of compressor cycles for the inboard actuator, in a given period of time. Following the event Duane Arnold checked the compressor cycling and estimated it to be 3 times per hour. During that time, the compressor cycled between 95 PSI and 105 PSI.
i.
l 12/07/98 The Ralph A. Ililler Company Page 3 of 11 '
6005 Enterprise Drive Export, PA 15632
l The pneumatic piston of the actuator, which had been on the outboard valve of the
. same Main Steam Loop, actuator "B", had been found to have two partial cracks at Duane Arnold, which had propagated on the top of two ribs. Actuator "B" had been manufactured in 1989 and put into service 1990. It was later found through the use of I magnetic particle examination to have five cracks on ribs, none of them through I cracks. It was removed from service in 1998 and has not yet been refurbished. The actuator had been in service five fuel cycles. It has been estimated that the piston was also exposed to twenty-one (21) pressure fluctuations from 110 PSI to 0 and back to l
110 PSI per fueling cycle for installation, set up and verification stroke testing. The '
piston was exposed to stress fluctuations in service due to compressor cycles. There is no existing mechanism to verify the number of compressor cycles for the outboard actuator, in a given period of time. Following the event Duane Arnold checked the compressor cycling and estimated it to be one cycle per day. During that time, the j compressor cycled between 90 PSI and 130 PSI.
Design Basis l
The SA- A101 actuator is designed with a longer stroke than the Main Steara Valve that it will conud This added stroke is used during installation setup to insure that during operation of the valve and actuator the pneumatic and hydraulic pistons will l not impact the head and cap of the actuators.
The SA-A101 was designed per the applicable portions of ANSI B93.10 specification for the design of square head, fluid power cylinders, as required by the purchase specification, GE 23A6367. This required that "The group I (Main Actuator and Control Assemblies) shall be similar in design, construction and material to R. A.
Hiller Company Model SA-A068. The pistons used on the Model SA-A101 are the same material, construction and part number as the environmentally and seismically qualified SA-A068. The qualification test of the SA-A068 actuator included cyclic aging (5,000 full cycles). The qualification was to the environmental and seismic levels developed by General Electric for MSIV actuators specifically for this type of application. By definition, the piston is not a pressure boundary according to ANSI B93.10.
The piston was supplied to the Hiller Company by a subtier supplier. The following were the requirements to the subtier supplier. The material was specified as ASTM -
A48 Class 35 - B, Gray Cast Iron, which has a minimum tensile strength of 35,000-PSI.
Gray cast iron was selected as the material for the piston for a number of design considerations. In design of the piston the physical properties considered were the wear characteristics, corrosion resistance, machinability, the low shrink-rate of the casting, low melting point, and strength.
Verifications 12/07/93 The Ralph A. Hiller Company Page 4 of 11 6005 Enterprisc Drive Export, PA 15632
i l
1 hiaterial Verification l Programmatically the original pistons were inspm ed for surface defects, l dimensionally verified, the material was verified to be cast iron via microstructure ;
and verified to be leak tight at post production testing. I After the leakage problem was identified, the through crack was discovered in the piston via a dye penetrant test. After discovering the through crack, the piston was Brinell tested for hardness, destructively tested to determine ultimate tensile stress and micro structure was verified by QC hietallurgical Labs for Alliant Utilities, Duane Arnold Plant. The hardness verifications were found to be 179 and 183 BHN, which indicates tensile strengths ofin excess of the 35,000 PSI minimum according to the AShi Metals Handbook Vol.1 Properties and ,
l Selection ofhietals 8* edition. The destructive test indicated the tensile strength to be 27,400 PSI. The metallurgic examination verified the material to be cast iron, graphite flakes were predominantly found to be type A. Some oxidation at the secondary cracks near the inception of the fracture was identified.
Gray Cast Iron is inherently non-homogeneous due to a varying cooling rate in the mold, along with other factors. This would account for the low tensile strength in a particular area of the casting.
j Finite Element hiodeling Dominion Engineering performed a fmite element model of the stresses Alliant
, Utilities, Duane Arnold Plant. The analysis concluded that a maximum radial i stress of 19,700 PSI would be created by a nominal operating pressure of the inboard actuator of 100 PSI.
I Further Testing Ultimate Load Testing The Hiller Company performed physical testing of an identical piston "C', which had been in service at a nuclear power plant since 1991 to verify the results of the finite element analysis. Physical testing demonstrated the tensile stresses near the I hub to be 16,000 PSI, at the force calculated to be equal to 100-PSI nitrogen pressure on the rod side of the piston (19,000 lbf). This demonstrates that the finite element analysis is within an acceptable range of the actual stresses demonstrated. Figure 1 shows the test set up.
12/07/98 The Ralph A. Hiller Company Page 5 of 11 6005 Enterprise Drive Export, PA 15632 l
l
1 s
, -3 e a
/ n_ \ )
\ %/ 1 Thernocouple Stroin run vie. . Gouge )
1 Piston Rod l Stroin Gouge
. m ,. ;-.
\ l E Plote Force Figure 1 i
Piston "C" was destructively tested to find the force required to create ultimate failure. The piston began to fail at 56,375 lbf. This corresponds to a maximum i
' tensile stress of approximately 61,600-PSI, which would indicate failure at a ;
pressure of 312 PSIG. Piston "C" was tested in three areas for Brinell hardness and found to be 180,185, and 200, which corresponds to tensile strengths in excess of the 35,000-PSI requirement.
Fatigue Testing The Duane Arnold piston with the two partially cracked ribs, " piston B", was tested. The purpose of the test was to determine if a cracked piston would fail when subjected to stress fluctuations like those, which the Duane Arnold piston would be subjected to in service due to fluctuations in supply pressure. The test set up was the same as figure 1.
12/07/98 The Ralph A. Hiller Company Page 6 of 11 6005 Enterprise Drive Export, PA 15632
i l
I l
1
)
Prior to cyclic fatigue testing the piston, was tested to verify cracks using wet 1
. luminescent magnetic particle examination. Five of the ribs were found to have some cracking. Wet luminescent magnetic particle examination shows both surface and subsurface cracking.
l The piston was then tested to verify cracks using liquid dye penetrant. l Examination showed that in addition to the two cracked ribs found at Duane 1 Arnold, two other ribs showed smaller cracks. A total of four ribs showed evidence of cracking prior to the beginning of the cyclic fatigue testing.
The piston was subjected to load variations from 18,000 lbf to 20,000 lbf.
According to the finite element analysis this created the same stress at the rib / hub l interface as the supply pressure fluctuations between 95 and 105 PSIG. The load l was cycled at a rate of 10 Hz. After 120,000 cycles no crack growth was noted. I At that point the cycle rate was increased to 15 Hz. After 160,000 cycles a new
{
surface crack was found. The cycle rate was then slowed to 10 Hz. No more new l surface cracks were developed throughout the remainder of the test. Through the test there was some crack growth. None of the cracks grew in excess of %" in 1.2 million cycles. !
l Non-destructive Testine !
A total of 7 pistons were tested for cracks using liquid florescent magnetic particle verification. No cracks were found on any of the pistons.
Analysis
- 1) L.oading above the ultimate strenuth The design of the piston was reviewed to determine if postulated loading exceeded the ultimate strength of the piston material. The designed tensile is 35,000 PSI. In the design of cast iron parts the tensile strength can vary throughout the part. The tested tensile strength of the material in the rib area of the failed piston is 27,400-PSI minimum. The finite element analysis indicates that the maximum radial stress of the piston is 19,700 PSI. A Brinell hardness of 180,185, and 200 was found on the piston destructively tested which indicates a tensile strength of 35,000-PSI minimum. The piston was tested to ultimate at 56,375 lbf. Although piston "C" had a higher Birnell Hardness than the failed piston "A", it can be extrapolated that since a piston with a tensile strength in excess of 35,000 PSI failed at 56,375 lbf, piston "A", with a tensile strength of 35,000 failed at approximately 52,500 lbf. As installed the piston should not be in contact with the head and cap of the pneumatic portion of the actuator. This is insured by the stroke of the actuator being longer than the stroke of the valve, and through the correct installation of the actuator on the valve.
12/07/98 The Ralph A. Hiller Company Page 7 of 11 6005 Enterprise Drive Export, PA 15632
Based on discussions with Duane Arnold no loading condition could be identified which would create a stress in excess of the 27,400-PSI. It is logical to conclude
' that the crack through the piston was not created due to a load in excess of the ultimate tensile strength of the piston.
- 2) Hinh cycle fatinue
' High cycle fatigue was considered as a possible cause of the crack. According to the ASM Handbook, Vol.19, tenth edition the fatigue endurance is 16,000 to 17,500 PSI for unnotched tpe 35 gray cast iron. This is specified for high cycle fatigue (cycles between 10 and 10 ).
' No loading condition.could be postulated which would have exposed either piston "A" or "B" to the high cycle fatigue range.
- 3) Low cycle fatigue Low cycle fatigue was considered as a possible cause of the crack. For low cycle -
fatigue ofcast iron the ASM Handbook says:
"For practical purposes, crack initiation in the low-cycle regime (102 and 10' cycles) can be assumed to occur on the first cycle ofloading. Thus, life prediction models for low cycle fatigue of cast irons should be based on the following chart."
I 102 !
A0/2(MPa)-140120100 90 85 g to f1 -
/
1 0.1 0'
104 ' ' ' ' '
1 10 102 103 toe ios gos N cydes Fig. 6 crack development in gray iron. Source: Ref 2
)
Figure 2 i
To analyze the probability oflow cycle fatigue, a cyclic stress needs to be l postulated over a number of cycles. The only cyclic load, which could be I postulated, is that of the pressure fluctuations of the compressor which supplies ;
12/07/98 The Ralph A. Ililler Company Page 8 of11 6005 Enterprise Drive Export, PA 15632
l ,
l l
nitrogen to the inboard actuator "A" According to the Finite Element modehng, !
the change in the maximum radial stress created in the fluctuations from 95 PSI to I 105 PSIis approximately 1,000 PSI (Ac=6.9 MPa). Following the chart and I calculating Ao/2 as 3.5 MPa. The stress created by this change in pressure should not create a failure, which would fall in the low cycle fatigue range.
Further, if the change in stress created by a change in pressure from 0 to 100 PSIG (20,000 PSI, o=137 MPa)is postulated to have occurred 1.5 x 105 times, the Ao/2 as 69 MPa. This is considered to be an overly conservative approach as the i number of postulated pressure fluctuation cycles is examined over a change in l pressure, which could not be postulated to have occurred. However, this would still indicate that the cast iron piston should not fail in the low cycle fatigue range based on the ASM chart. l Actuator"B" had partial cracks of two of the ribs of the piston. Following the same logic as above, it was postulated that it had been exposed to one (1) compressor cycle per day for the time in service since 1990, between the ,
3 compressor set points of 90 to 130 PSIG. Ifit had been exposed to 3.4 x 10 I cycles between 90 and 130 PSIG, following the same procedure as above the maximum radial stress created is approximately 4,000 PSI (Ac=27.6 MPa), and Ao/2 is 13.8 MPa. The stress created by this change in pressure should not create a failure, which would fall in the low cycle fatigue range.
- 4) Low cycle fatigue with a pre-existing crack Since the metallurgical examination indicated that some of the secondary cracks near the through crack had evidence of corrosion which could only have occurred at the time of casting shake out, a pre-existing crack was assumed. However, according to the ASM Metals Handbook 9* edition, Vol.1 Properties and Selection Irons and Steel, pg. 20:
" Fatigue Notch Sensitivity. In general, very little allowance need be made for reduction in fatigue strength caused by notches or abrupt changes of section in gray iron members. The low-strength irons exhibit only a slight reduction in strength in the presence of fillets and holes. That is, the notch-sensitivity index approaches zero; in other words, the effective stress-concentrationfactor for these notches approaches one. This characteristic can be explained by considering the graphite flakes in gray iron as internal notches. Thus, gray iron can be thought of as a material that is already full of notches, and that therefore has little or no sensitivity to the presence of additional notches resulting from design features" i This applies to gray iron material, which is subjected to high cycle fatigue.
However, assuming the piston had a pre-existing crack small enough to be undetected by the casting manufacturer and inspection at assembly, the conclusion may be drawn from the above information on fatigue that a pre-12/07/98 The Ralph A. Hiller Company Page 9 of 11 6005 Enterprise Drive Expon, PA 15632
1 existing crack would have little effect on the fatigue life of the piston in the low l
. cycle fatigue range.
l The logical conclusion is that the postulated and verified stresses could not have created the failure due to fatigue within the postulated number of cycles.
' I The cycle testing performed by the Hiller Company verified this conclusion. I 5.) Impacj l I
The other possible cause of the piston failure is an impact load. Testing indicates that a force of 56,375 lbfis required to begin an ultimate failure of the piston.
When this force was applied to a piston, a maximum stress of 61,600 PSI was measured. The finite element analysis of the piston indicates that the maximum l operating stress in the piston is 19.7 KSI, due to pressure from the nitrogen !
source. The actuator, when correctly installed and operated as designed, will not realize any impact loading on the piston. The stroke of the actuator is greater than the stroke of the Main Steam Isolation Valve. By installing and adjusting the actuator correctly, the piston will move back and forth between the front and rear l pneumatic heads. It will not impact them. l The Model SA-A068 actuator had been qualified for a life cycle of 5000 cycles.
During this testing a minimum of 250 cycles were at full pressure with no valve
{
attached to the actuator. For this test the piston was permitted to slam against the j head and the cap.
l Discussions at Duane Arnold Nuclear Station, considered a spike in pressure in the steam line to 1031 PSI, which occurred in 1997. Afler review it was concluded that this abnormal situation could have caused the crack due to impact in piston "A", but could not explain the crack in both pistons "A" and "B" Creep was not considered as a possible failure mode as the ambient temperature of the actuators was much too low for this to have occurred.
There were no indications that corrosion was the cause of the through cracks.
Summary Any of the possible failure modes for a cast iron piston could have caused the through crack in the piston under the correct postulated circumstances. However, there is no definitive evidence, that any of these postulated circumstances occurred to cause the failure.
l l The generically identical piston under consideration has been in service in various l nuclear power plants for more than thirty years with no similar failures. The actuator design was seismically and environmentally qualified.
12/07/98 The Ralph A. Hiller Company Page 10 of 11 6005 Enterprise Drive Expon, PA 15632
. For this reason it is the current position of the Ralph A. Hiller Company that:
The actual cause of the piston cracking at lowa Electric, Duane Arnold plant can not be identified with exact certainty. l
. The material of the piston, which cracked "A", was veri 6ed and has been demonstrated to exceed the stresses required by the piston design during normal plant operation and postulated LOCA.
. There are no known plant conditions, which could logically be identified as l causing stresses on both the pneumatic pistons (Inboard actuator and Outboard actuator) at the levels required to create the cracking. l
. Inspection and metallurgical examination of the cracked piston did not I definitively indicate what caused the failure. Since cast iron's failure mode does l not create a substantial amount of deformation prior to failure, inspection of the fracture does not clearly demonstrate whether the failure is due to fatigue, tensile, bending, or impact stresses. ;
e It is the position of the Hiller Company that this is a random event in which the l root cause was a combination of:
- 1) The inherent nature (cracked) of gray cast iron
- 2) Cyclic stress fluctuations due to variations in supply pressure
- 3) Some type ofimpact load which can not be identified by either the Hiller Company or Alliant Utilities, Duane Arnold Plant.
For this reason, it is the conclusion of the Ralph A. Hiller Company at this time that no 10 CFR 21 reportable defect exists. The following are the actions and recommendations of the Hiller Company regarding the 20" Bore Pneumatic Piston on MSIV actuators.
. The Hiller Company will incorporate NDE of each new cast iron piston prior to supply as a replacement (spare) part or assembly.
. All rebuild / refurbishment of pneumatic actuators for safety related applications will include NDE of the pneumatic cast iron piston.
. Any cracking in excess of %" found as a result of either of the above NDE on the piston will be dispositioned as " scrap, replace with new"
. There is no immediate need to perform this verification with actuators currently in service. The plants do not need to shut down.
. The Hiller Company recommends that all Nuclear Power Plants, as a part of their standard MSIV actuator maintenance return their actuators to the Hilles Company for rebuild / refurbishment. As stated above the Hiller Company will verify the status of the piston and replace as required. In lieu of this, the Plant (or their agent) should perform NDE of the piston and replace with new if any cracks in excess of %" are discovered as a part of their normal maintenance.
. No change in material or design is required.
12/07/98 The Ralph A. Hiller Company Page11of11 6005 Enterprisc Drive Export, PA 15632
. _ _ . . . _ . _ ,...._._ _-__ ~. . ,__
,m m. _
l a ?
- j. '.
!' '. )
l l S
' l k=d l RALPH A. HILLER COMPANY 6005 ENT[RPR$E DRIVE EXPORT.PA 15632 EfNIEi$
- October 29,1998 -
l 1
Document Control Desk United States Nuclear Regulatory Commission L Washington, DC 20555 !
l 1
Reference:
Hiller Document 98-NRC-003.
Dear Sir,
Per the applicable requirements of Title 10, Chapter 1, Part 21 of the Code of Federal Regulations, the responsible officer for the Ralph A, Hiller Company I advised the Nuclear Regulatory Commission of a possible defect in a letter dated May 19,1998 (Attachment A).
As required by 10CFR21, I will be submitting a full, final report, which will indicate whether a reportable defect exists or not. However, at this time we are currently proceeding with a final examination of which we feel is required to accurately determine )
the root cause of the cracked piston. I expect the testing, analysis and the report to be I complete and the detennination to be completed within the next sixty- (60) days.
. Should you have any questions, you can reach me at (724) 325-1200, or fax me at (724) 733-1825.
Thank you for your attention in this matter.
- Yours truly,g
-y.
W,
/. h +&&
< .J. Randolph Hiller
. Chief Executive Officer
. Attachments: (1) " Notice of Possible Defect per 10CFR21" cc: file qg03-oeB5 Ip
L ,- -
l l _C.E k=11
' RALPH A. HILLER COMPANY 6005 ZNTERPRtSE drive EAPORT. PA 15632 -
L'I' U$'$Il$$$
July 17,1998 Document Control Derk United States Nucler ibgulatory Commission ,
Washington, DC 20555 '
Reference:
Hiller Document 98-NRC-002
Dear Sir:
Per the applicable requirements of Title 10, Chapter 1, Part 21 of the Code of Federal Regulations, as the responsible officer for the Ralph A. Hiller Company I advised the Nuclear Regulatory Commission of a possible defect in a letter dated May 19,1998 (Attachment A).
As required by 10CFR21, I will be submitting a full, final report, which will indicate whether
- a reportable defect exists or not. However, at this time we do not have sufficient data to make that determination. We are currently proceeding with physical testing which we feel is required to accurately assess the situation. I expect the testing to be complete and the determination to be completed within the next sixty (60) days.
Should you have any questions, you can reach me at (724) 325-1200, or fax me at (724) 733-1825.
Thank you for your attention to this matter.
Yours tmly, RALPH A. HILLER COMPANY 7
. & A A y l; J. Randolph Hiller Chief Executive Officer
! Attachments: (A) Hiller Document 98-NRC-001 JRH/klj L cc: File Qt3o 09N I P ._
, ~. .. ~ ~ - . . + . - + - - - 1w > -es e ' ' ' '
. 3 + m.mns
.. _ _. i
- > < * ~ ~ ~ ~ . - ~ . . ,
.. o h.*
b l W :
RALPH A. HILLER COMPANY.
6005 ENTEnenest Dnivt tapont.PA 15632 f
TELE: 412 32" 1200 Fa a: 412 733-1825 I
f May'18,1998 '
r Document Control Desk United States Nuclear Regulatory Commission l Washington, DC 20555 l l
Reference:
Hiller Document 98-NRC-001
Dear Sir,
Per the applicable requirements of Title 10, Chapter 1, Part 21 of the Code of Federal -
Regulations, it is my responsibility as the responsible officer for the Ralph A. Hiller l Company to advise the Nuclear Regulatory Commission of the attached "Possible Defect".
As required by 10CFR21, I will be submitting a full, final rep. ort, which will indicate if a reportable defect exist or not. ]
Should you have any questions, you can reach me at (724) 325-1200, or fax me at (724)
-733-1825.
Thank you for your attention in this matter.
Yours trulyU
[f,..q.!h ff '
w
.[K Randolph Hiller Chief Executive Officer .
Attachments: (1)" Notice of Possible Defect per 10CFR21" ,
i cc: file
,. j Q W .2?bO M SPP
<w...
8 k=,i RALPH A. HILLER COMPANY 6005 ENTERPaist Onivt EsponT. PA 15632 T t t.E. 412 3258200 ra x. 412 7338825 Attachment 1 " Notice of Possibic Defect per 10CFR21" Dackground The Ralph A. Hiller Company of Expon PA is a supplier of safety related components to Nuclear Power Plants with a Quality Assurance Program per the requirements of Title 10, Chapter 1, Pan 21 of the Code of Federal Regulations. A significant portion of these components has been valve actuators.
Description of Event IES Utilities Inc., Guare Amold has reponed that a Model S A-A101 Main Steam isolation Valve "A" (MSIV) Actuer showed evidence ofintemal leakage,in both directions past the pneumatic piston. The actuator was returned to the Hiller Company's production facility for evaluation. Evaluation verified the excessive leakage and discovered a fracture of the pneumatic piston. The crack consisted of a cross sectional (through) crack in the web and reinforcement ribs at the center hub involving a 2180* on the hub area circumference.
Engineering Background De Model SA-A101 MSIV Actuator is a Basic Component, which has been classified as safety related. It consists of a 20" diameter bore pneumatic cylinder in tandem w .h a 5" diameter bore hydraulic cylinder.
The failure (safety) mode is the rod extended (valve closed) position. This position is achieved via an external spring pack and is assisted by the pneumatic portion. ne hydraulic cylinder is designed for speed control. This family of actuator is used as the MSIV Actuator at many other plants.
De actuators designed by the Hiller Company for safety related applications are designed per the guidelines of ANSI B93.10. De piston materialis designed to be ASTM A48 cast iron with a tensile yield strength of 35,000 PSI. This material was chosen for its excellent compressive strengt'i, good tensile strength, good machinability, and excellent resistance to gauling characteristics and is extensively used for pneumatic industry applications. However, the material does not have good bending strength.
When the Model SA-A101 MSIV Actuator s were purchased in 1989/90 by Duane Arnold, the production verification of the pistons for safety related service were programmatically required to be dimensionally checked, assembled, the correct assembly was verified and finally a leak test was required. Our records indicate each of these verifications was successfully performed.
Investigation The failed piston was tested to verify the material and failure mode. The failure was found to be caused by bending stresses due to impact. Tne material was found to be cast iron with a tensile yield strength of 27,500 PSI. No inclusions were found which would initiate this type of fracture. No other indications of damage were found in the remainder of the actuator that could cause leakage.
Inspection of other MSIV Model SA-A101 MSIV Actuators, which have been in service at Duane Amold, found another with a similar, less severe crack, actuator "B" Dis actuator,"B", had been in service on the same Main Steam Line at the plant. The piston web was cracked in the same area, however the crack did not propagate through the part and no leakage was detected. As pan ofIES's maintenance plan, both
___._.m_m_ _
4
, +
4 Actuators "A" & "B" had been refurbished and postproduction tested a second time approximately two years ago with no leakage.
Review of the exact nature of these failures indicated that the safety function of the actuator would not have been prevented, even if the fracture had been catastrophic and the piston had become loose from the piston rod. The piston failed safely and would not have prevented the safety mode from occurring, produced by compressed air and springs. Further testing on actuator "B" indicated that normal cycling of the actuator did not continue to propagate the failure.
Discussions with the owner, IES Utilities Inc and the plant designer, General Electric, indicate that i operationally there is a standard test, which could create a pressure anomaly through the line. Discussions with the valve manufacturer, Edwards Valves, concluded this type of event couto cause the valve to force i the Actuator's piston rod into the Actuator (retract). The initiating event is still under investigation. The hydraulic control syvem is not designed to mitigate shock waves that produce excessive internal hydraulic pressures. De excessive energy could cause damage of the pneumatic piston and/or the hydraulic control assembly. He original specification did not require that the Actuator to be designed for this type of event 4
and there has not been a similar rep'orted failure of a piston on any actuator designed by the Hiller Company.
\
Plan
- The Ralph A. Hiller Company is currently working to detennine the cause and extent of the problem. This I work requires a more thorough engineering review of the function of the MSIV Actuator in the plant and the stresses this could create in the piston. Per the applicable portions of the Ralph A. Hiller Company's Quality Assurance Program and 10CFR21 the findings of this investigation along with any I recommendations will be reported to the Nuclear Regulatory Commission.
Summary There has been a crack found in two pistons, of two of the MSIV Actuators, in service at a single Nuclear Power Plant. A loss of function of the actuator did not occur. These actuators were on the same Main Steam Line and it has been determined that the leaking piston would not have prevented the safe function of the basic component. At this point in time, the situation is being reviewed to determine if this failure could pose a " substantial safety hazard" to the particular, or any other Nuclear Power Plant.