ML20042E656
| ML20042E656 | |
| Person / Time | |
|---|---|
| Site: | South Texas  |
| Issue date: | 04/18/1990 |
| From: | Rosen S HOUSTON LIGHTING & POWER CO. |
| To: | NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM) |
| References | |
| ST-HL-AE-3422, NUDOCS 9004260360 | |
| Download: ML20042E656 (176) | |
Text
_ .. . _ . . . . .
The Light eompany$wt Houston Lighting & Power esas Project Electric Generating $tation P. O. Bos 209 Wadsworth,Tenas 77483 April 18, 1990 ST-HL-AE-3422 File No.: G02 10CFR50.36(c)($)
U. S. Nuclear Regulatory Commission Attention: Document Control Desk Washington, DC 20555 South Texas Project Electric Generating Station Unit 2 Docket No. STN 50-499 Final Special Report Regarding A Diesel Generator Valid Failure on November 28, 1989 Pursuant to the South Texas Project Electric Generating Station Technical Specifications 4.8.1.1.3 and 6.9.2, Houston Lighting & Power submits the attached Final Special Report regarding a diesel generator valid failure which occurred on November 28, 1989.
If you should have any questions on this matter, please contact Mr. C. A. Ayala at (512) 972-8628 or myself at (512) 972-7138.
osen Vice President Nuclear Engineering BEM/n1
Attachment:
Final Special Report Regarding A Diesel Generator Valid Failure on November 28, 1989 9004260360 900416 FDR ADOCK 05000499 3 PDC h
NL.90.079.02
, A Subsidiary of Houston Industries incorporated k
f Houuon lighting & Power Company ST-HL AE 3422 South Texas Project Electric Generating Station le N . : G02 g
cc:
- Rtgional Administrator, Region IV Rufus S. Scott Nuclear Regulatory Commission Associate General Counsel 611 Ryan Plata Drive, Suite 1000 Houston Lighting & Power Gnapany Arlington, TX 76011 P. O. Box 61867 Houston, TX 77208 q%
, G :rge Dick, Project Manager '
U.S. Nuclear Regulatory Commission INPO W::hington, DC 20555 Records Center.
g 1100 Circle 75 Parkway J. 1. Tapia Atlanta, GA 30339 3064 S2nior Resident Inspector c/o U. S. Nuclear Regulatory Dr. Joseph M. Hendrie C:mmission 50 Be11 port Lane P. O. Box 910 Be11 port, NY 11713 Bay City, TX 77414 D. R. Lacker J. R. Newman, Esquire Bureau of Radiation Control N;wman & Holtzinger, P.C. Texas Department of Health 1615 L Street, N.W. 1100 West 49th Street W:thington, DC 20036 Austin, TX 78704 D. E. Ward /R. P. Verret C:ntral Power & Light Company P. O. Box 2121 Corpus Christi, TX 78403 J. C. Lanier Director of Ceneration City of Austin Electric Utility 721 Barton Springs Road Austin, TX 78704 R. J. Costello/M T. Hardt City Public Service Board P. O. Box 1771 Sin Antonio, TX 78296 NOTE: Only people designated by asterisk receive attachment.
Revised 12/15/89 L4/NRC/
i
~ j
l .' t I
c l
I ,
l l ~
I l
I !
' I SOUTH TEXAS PROJECT g ELECTRIC GENERATING STATION s .
g Final Special Report -
g Diesel Generator Valid Failure g
on November 28,1989 <
I l April 18,1990 I
I I
I J
. _m, B
I -
Attachment ST-HL-AE-3422 Page 1 of 9 I
g 3 I
I South Texas Project Electric Generating Station Unit 2 I Docket No. STN 50-499 Final Special Report Regarding A Diesel Generator Valid Failure on November.28, 1989 I
I I
I I
I I ~
g ,
I NL.90.079.02
lI Attachment ST-HL-AE-3422 Page 2 of 9 TABLE OF CONTENTS I
I. Description of Events II. Cause of Event III. Corrective Actions IV. Testing V. Conclusion Figure 1 SDG Connecting Rod-Figure 2 011 Port in Connecting Rod Figure 3 Drilling Configurations of 011 Passage s Appendix 1 HL&P Engineering Report-Summary Report on the Failure of Diesel Generator 22 Connecting Rod.
Dated March 14, 1990 Appendix 2 HL&P Material Technology Division-Investigation of Diesel Generator Engine Connecting Rod Failure South Texas Project Unit 2. Dated December'1.1, 1989 Appendix 3 Aptech Engineering Services-Signifi t;ince of Over-Drilled Oil Moles on Fatigue Life of the KSV-4-2A Connecting Avd in the Standby Diesel Engines at South Texas Project. Dated March, 1990 Apprndix 4 Cooper-Bessemer-Finite Element Analysis of the KSV-4-2A Master Connecting Rod.
Dated February 26, 1990 Appendix 5 Battelle Laboratorias-Failure Analysis of the KSV-4-2A Master Connecting Rod. Dated February 27, 1990 NL.90.0??.02
I .
Attachment ST-HL-AE-3422 Page 3 of 9
~
South Texas Project Electric Generating Station Unit 2 Docket No. STN 50-499 Final Special Report Regarding I A Diesel Generator Valid Failure on November 28 1989 DESCRIPTION OF EVENT:
On November 28, 1989 Unit 2 was in Moda 5 for a maintenance outage. The Technical Specification required twenty-four hour. load test was being performed on Standby Diesel Gen 6rator_(SDG) 22. At 0957 hours0.0111 days <br />0.266 hours <br />0.00158 weeks <br />3.641385e-4 months <br />, approximately-ten hours into the test, a loud knocking noise was heard by two maintenance technicians. The technicians immediately evacuated the area. The #4 master connecting rod subsequently failed and the engine tripped. No. indication of
(
I the trip was received by the control room. Immediate inspection of SDG 22 revealed that the #4 articulated rod, both #4 pistons and other debris had been ejected through the rigitt centerframe door. The #4 cranhcase overpressure relief valves were pushed off the centerframe doors on both sides of the engine and the right sida centerframe door was driven into the adjacent starting air dryer. The master rod, cap and counterweights were in the engine base.
Investigation and recovery teams were formed to evaluate the failure and implement corrective actions. A thorough inspection of the engine was performed. The following was observed:
The master connecting rod was broken into two pieces through the c't I passage between the crankpin bore and the articulated rod pin bore and through the upper end of the bails.
The left side connecting rod cap bolts were broken and the right side bolts bent.
All f our cap nuts had been si npped f rom the studs.
f' -
There was evidence of significant heating of the articulated rod
- pin, its bushing, and the rod on both the bail and-load sides of the articulated pin.
The centerframe sustained minor damage including several cracks and impact damage.
I -
The #4 crankpin was gouged at o veral locations.
The crankshaft runout was 0.003" '
5 I NL.90.079.02 I
i
t ll i Attachment ST-HL-AE-3422 -
, Page 4 of 9
- ll .
- - There were numerous superficial gouges on the crankwebs and crankpin thrust faces.
- g - The piston pins had slight impact damage.
>4 '"'
. i
- The inlet valves were slightly bent on both heads, tl.e rocker arms were cracked and the push tods were bent.
- The #4 cylirder liner skirts were broken.
] CAUSE OF OCCURRENCE:
The cause of this failure was the presence of an improperly drilled and repaired oil passage hole serving the. articulated rod pin bushing in the master connecting rod assembly (see attached Figures 1 and 2). The connecting rod oil passage hole was improperly drilled and repaired on March 5, 1979 by t Cooper-Bessemer Industries. This manufacturing defect acted as a stress riser. +
and an initiation site for high cycle fatigue crack propagation.
Primary fractures initiated on both sides of the improperly drilled and repaired oil hole. This hole was perpendicular to the main oil channel. The !
primary fracture surface displayed charactaristic signs of fatiguo pointing to a double origin on either side of the oil hole. The edge of the oil hole was sharp and acted as a stress riser. The edge also had crack like tears which were capable of propagation by fatigue. The fatigue crack had propagated over .,
-I a large percentage of the fracture surface before the final fed are1 occurred.
This confirmed that the cyclic stress levels were low and the critical flaw l
size for this material was high. Scanning electron microscope examination showed closely spaced fatigue striations confirming high cycle, Icw stress - .
failure.
Secondary fractures occurred in the strap sections of the connecting rod
- which hold the articulated rod pin. These fractures also-occurred as a result of fatigue, however, scanning electron microscope examination showed more I coarsely spaced fatigue striations indicating these fractures occurred as a result of the increased. stress placed on the straps by the primary oil hole fracture. The final fractures were seen in the articulated rod pin bolts and the bolts which connect the master connecting rod to the bearing cap. All these failures occurred by tensile overload. Scann'ing electron microscope examination revealed ductile failure of all parts characteristic of a tensile overload.
Laboratory analysis of the material composition and mechanical properties of the failed connecting rod indicated that the. material met j specification requirements (Appendix 2) and that it had good fatigue 9 properties. (Appendix 5).
NL.90.079.02 I
I Attachment l
ST-HL-AE-3422 s .Page 5 of 9 !
CAUSE OF OCCURRENCE Cont'd.:
- Other master connecting rods inspected at STP and other plants indicated no other cases of the type of manufacturing defect associated with"this failure. However, several cases were found with sub-surface holes drilled i beyond the nominal depth, similar to Figure 3, Case C. The overdrilled hole .
depth data was statistically analyzed and it was. determined that the probability of occurrence of a worst case sub-surface hole (with a thickness at the tip of about 0.125") was 2.4 percent with 95 percent confidence. The mean thickness at the tip was 0.48". . The worst case sub-surface hole ,
configuration was analyzed by Cooper-Bessemer, and separately by Aptech Engineering Services, using three dimensional finite element methods. It was.
shown that it would not result in a reduction of fatigue margin below the i design requirements.
A few cases were found with the configuration shown in Figure 3, Case B.
This is a benign mode since this design is.enployed for other through-holes in the same secticn of the connecting rods.
Analyses of the connecting rod failure are attached as follows:
Appendix Preparer Subj ect 1 HL&P Engineering -Summary Report on the Failure <
of Diesel Generator 22 ?
Conneeting Rod. .;
j Dated March 14, 1990 -
2 HL&P MTD Investigation of Diesel >
c Geners or. Engine Connecting I
Rod' Failure South Texas .
Project Unit 2 .
Dated December'13, 1989 l >
3 Aptech Significance of Over-Drilled ,
Oil Holes on Fatiguo Life of the KSV-4-2A Connecting Rod in I
the Standby Diesel Engines at the South Texas Project Dated March, 1990 4 Cooper-Bessemer Finite Element Analysis of the -
KSV-4-2A Master Connecting Rod Dated February 26, 1990- 1 5 Battelle Tailure Analysis of the
~'
KSV-4-2A Master Connecting Rod .
L Dated February 27, 1990 NL.90.079.02 >
r l
I Attachment ST-HL-AE-3422 Page 6 of 9 Cooper-Bessemer has ident11 tied that the repair of the master connecting rod which failed was documented and dispositioned in accordance with Cooper Energy Services procedures at the time of manuf acture. This was the only i master rod repaired in this manner. However, the existence of other connecting rods with similar, though less severe, manufacturing defects confirma that a less than adequate manufacturing process was in effect at Cooper-Bessemer at the time of fabrication of the #4 connecting rod.
As licensee, HL&P was responsible for ensuring that the vendor's quality program was properly implemented. An evaluation of the HL&F vendor control program in effect at the time of SDG manufacture has concluded that Cooper-I Bessemer failed to comply with purchase order requirements which specify that defects which are corrected by repair be documented and submitted to the purchaser for approval. Cooper-Bessemer documented that the repair toult place I on a Material Review Request however, they did not forward that disposition to HL&P's agent for approval. Since the repair disposition was reviewed and approved by Cooper-Bessemer engineering and did not necessarily represent a deviation from the purchase specification, it cannot be concluded that either HL&P or the architect-engineer would have determined that the repair was inadequate had they reviewed the Material Review Request. The HL&P vendor quality program at the time of the diesel generator manufacture assured compliance with 10CFR50, Appendix B through specification requirements, vendor audits and vendor inspections. A less than adequate repair disposition by Cooper-Bersemer was outside the control of HL&P and does not represer,t a breakdown or inadequacy in the HL&P quality program.
CORRECTI\p_ ACTION:
The fq1 lowing is a summary of corrective actions taken to repair SDG 22 and restote it to a like-new conditions Following this incident, the crank 2 haft was examined by-Reynolds-French, a specialty machine shop contracted to Cooper-Bessemer for eagine machining repairs. It was found to have a maximum runout of 0.003*TIR. Corrective action was taken to reduce the runout to lesc than 0.002"TIR. 1 i
Repairs were made to the e4 crankshaf t webs, the crankshaf t journal, counterweight landings, cracks to the centerframe, and access door bolt hole damage.
- As r precautionary measure following this incident, inspections of the
- 3L, #3R, #5L, anc. #5R cylinders and valve trains were conducted. All results were satisfactory.
1 I
NL.90.079.02 1
)
A Attechment ST-HL-AE-3422 Page 7 of 9 I The engine was flushed to remove contamination in the block as a result of this incident. Following this system flush, the internal lubrication system hoses were replaced. Oil samples were taken before I
and after the system flush and indicated that the cleanup was successful.
Due to concern that the #4L and #4R cylinder heads may have been I impacted during this incident, inspections were performed to verify the torque of the head holddown studs, elongation of the studs, and proper thread engagement. These inspections revealed no deficiencies.
Warm web deflection measurements were conducted following an 8-hour period. All results were satisfactory.
I -
Following a run.of the diesel at 100% of rated load, Cooper-Bessemer required a torque verification of the 44 cylinder ronnecting rod cap fasteners. The results were satisfactory.
As a precautionary measure, a torque verification of all SDG 22 anchor bolts was performed. The results were satisfactory.
The torque of the bolted coupling between the generator and the engine was checked. The results were satisfactory.
The non-emergency shutdown trip devices were re-calibrated.
Generator inupections were conducted. No damage from the engine failure was found.
I -
The lack of alarms at the time of the engine failure both locally and in,the control room was traced to a blown fuse in the engine control panel which supplies the annunciator. The diesel generator
_I surveillance procedures will be revised by July 1, 1990 to require an annunciator test prior to each diesel generator test.
I TESTING A preoperational testing program was developed and performed to satisfy Technical Specification requirements and vendor recommendations. The program consisted of the following tests: -
Jacket water system pressure test following the breach of the jacket water system during the incident.
NL.90.079.02
l l
l Attachment ST-HL-AE-3422 Page B of 9
~
Lube oil system functional test following.the lube oil system flush and component replacement.
Safety shutdown and control system test.
Starting air compressor functional test following damage to start!ng air system components.
Fuel oil system pressure test.
Initial runs for engine analysis.
8 hour9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br /> break-in run.
76 hour8.796296e-4 days <br />0.0211 hours <br />1.256614e-4 weeks <br />2.8918e-5 months <br /> endurance test.
Performance of 2 PSP 03-DG-0002, Standby Diesel 22 Operability Test.
I- - 23 Consecutive starta, g During the performance of the preoperational testing, additional failures g occurred which are outlined below:
On December 27, 1989 during post maintenance testing following the I recovery, several problems occurred with the #4L fuel injector pump.
An abnormal knocking noise was heard. During troubleshooting, it was discovered that the holddown studs for this pump were loose. They I were most likely loosened during the original incident. The pump, studs, and nuts were replaced.
during testing.
Subsequently, the new pump seized During the removal of the failed pump, a stud was broken and replaced. However, the other three studs were not replaced at this time. The second new pump broke free frca its pedestal.during further testing. Following this failure, inspections were performed by the pump vendor (Haynes Corporation) and it was determined that the most likely cause of the pump brea ting free from its pedestal was the weakened condition of the tr ee stids that were not replaced following the pump seizure. The-bolting tor,:ue was then verified to be correct on the other pumps on SDG 22. As a corrective action, the maintenance procedure to replace fuel injector pumps has been revised to require replacement of studs each time a pimp is removed.
NL 90.079.02 3
Att ac hrwent ST-HL-AE-3422
- Page 9 of 9 i
On January 7, 4tvu during testing followird the recovery, f requency I oscillations were observed while SDG 22 was in the emeraency mode.
The test was terminated and troubleshooting performed. It was determined ths.t resistor RD-1 in the electrical governor circuit was I -
improperly soldered. When the electrical governor failed the mechanical governor took over engine speed control. . A misadjustment of the fe,edback valve on the mechanical governor caused the speed I oscillations observed. The electrical governor resistor was soldered correctly and the mechanical governor. linkage was adjusted. The RD-1 resistor in the electrical governors on the remaining SDGs will be inspected during scheduled train outages by June 1 1990.
I CONCLUSION:
Successful completion of the retest program combined with a' rebuild of I SDG 22 has established a like-new condition as certified by Cooper-Bessemer.
As such, future valid tests and failures will be counted from the completion of the retest program. The current surveillance' interval for SDG_22 is once I per 31 days.
I i 1 1
i 1 i j
NL.90.079.02 I
w FIGURE 1 E CONNECTING ROD l
I O% ,,, ,4.
oli Passage I
g ,
! ArticG1e'ted conneettng Rod Haster g;nenin, I
l l Articulated Rod
'Sa %
enre,s.!e Which Ini iated I en Passages h \=
l lb
~"
YMA- Bearing Shell I
s/ Section A A gl _
Bearing Cap
^
) .
1
- FIGURE 2 SECTION A-A ~
OIL PORT IN CONNECTING ROD r
dlL PORT B (TYP)
O!L PORT + l - - I<a- PRESS. FITTED (T P) o [
p h e ,L r e ,T A (1 ,,
01L FLOW (1) NORMAL CON 0lTION OF SECTION THAT FAILE0 SHOVlNG OIL DISTRIBUTION HOLES.
I PARTIAL BREAKTHROUGH SHARP CORNERS i M VA VA F/M i * > I i e
/ "
(2) NON-CONFORMING CON 0lTION.
I HDLE TAPPED AND PLUGGEO M, VAFVA M r - -
/ "
o !
(3) CORRECTIVE ACTION, l
g FIGURE 3 DRILLING CONFIGURATIONS OF ll OIL PASSAGE
.g m.i -
PLUG l N n THAWA, i l l kv '//////////A Y/ *
=
yv vX /////////1 i .L l a l CASE A m
l li TffA WJA a _.
kv
'//////////A s v "v K /////////l i L /
CASE B g 7- . . .. ..
-y . .
E ky '//////////A' '//
g N v "v 2 h
/////////I ~
L --
= .'
g- I*
1 CASE C
_n
ll I
I I
( .-
! l l
l i
r 4
I l
i
< APPENDIX 1 l l i i HL&P ENGINEERING REPORT-
SUMMARY
REPORT ON THE l FAILURE OF DIESEL GENERATOR 22 CONNECTING ROD j MARCH 14, 1990 i
l 1
\
'I l
)
. I l
l 1
i M
D NL.90.079.02 t
I l
g I
HOUSTON LIGHTING & POWER SOUTH TEXAS PROJECT ELECTRIC GENERATING STATION ,
I StiMMARY REPORT ON THE FAILURE OF DIESEL GENERATOR (22 CONNECTING ROD REVISION 1 I CODES, STANDARDS & MATERIALS DIVISION SUPPORT ENGINEERING DEPARTMENT I
I I
Prepared by l
f h:s .-
S. Timnaraju Senior Consulting Engineer Date: 3/14/90 ST/WS100/CDS
HOUSTON LIGHTING & POWER A I#
EUMMARY REPORT ON THE FAILURE OF DIESEL GDfERATGR 822 CONNECTING ROD On November 28, 1989, a connecting rod of Diesel Engine (22, at the South Texas Project, Unit -2 failed. At this time, the engine had-operated for approximately 634 hours0.00734 days <br />0.176 hours <br />0.00105 weeks <br />2.41237e-4 months <br />. Examinations showed the connecting rod fractured at several places (Ref. 1).
A muts surgical analysis of the fracture surf aces by HL&P (Ref. 2) showed that the primary failure (i.e. chronologically the earliest initiated - all other failures following later) was fatigue failure at' the cross-section between the crank shaft bore and articulating rod pin bore. The fatigue crack extended almost across the entire width of the '
connecting rod and over half its thickness, the balance failing~ by ductile overload.
I There are a number of oil distributing holes in this section. One of the radially oriented holes at the center' of the section was over-drilled past the central longitudinal hole all the way to the articulating rod bore, which it penetrated partially. The hole was plugged by threading and screwing in a plug but the hole edge remained sharp and ragged. The fatigue crack originated at the reentrant corners of the tip of this hole and progressed along both directions by high' cycle fatigue (Ref. 2) The very fine striations confirm that this I process took many millions of cycles. Based on the known 634 hours0.00734 days <br />0.176 hours <br />0.00105 weeks <br />2.41237e-4 months <br /> of ST/WS100/CDS 1 ;
l I
I operation at the time of failure, it is estimated that the connecting rod experienced at least 12 million cycles of alternating stress.
A part of the section in which the fatigue occurred failed by. ductile overload; and other parts (balls) failed by fatigue with a lower number of cycles or by overload (Ref. 2) .
{
=
An inspection of connecting rods from diesel generator (22 and spare parts at STP, and some connecting rods at other plants with Cooper diesel engines showed that it was not unusual for the radial hole to be drilled past the transverse hole, and that the depth of the hole was variable. A statistical analysis of known hole depths by Aptech Engineering Services, consultants to HL&p, indicated that the mean hole depth was 1.1 inch. The worst case hole depth reported was 1.45 inch, F which is equivalent to a minimum thickness at the hole tip of 1/8 inch and a minimum ligament of about 1/16 inch. The minimum ligament occurs to a side of the drill tip due to the skewness of the hole and curvature of the bores, (see graphics in Ref. 3 and 4). If the ligament were to become thinner than 1/16 inch, the drill would tend to I break through and be detected by visual inspection.
A three dimensional finite element analysis of the ligaments with three different hole depths was conducted by Cooper Energy Services. The depth selected represented a 5/16 inch ligament (conriervative relative I to mean thickness), a 1/16 inch ligament.(comparable to the worst case subsurface hole) and a break-through case similar to the failed-ST/WS100/CDS 2
connecting rod. Aptech Engineering Services perforned, at HL&P's l request, an independent analysis of the 1/16 inch ligament case and reviewed the other cases performed by Cooper Energy Services. The results of Cooper Energy Services and Aptech were comparable, though different software, and slight differences in modelling were used.
Both companies used loads developed by cooper Energy Services.
The results of finite element analyses indicated that subsurface holes geometries have substantial safety margins to fatigue crack initiation, assuming a conservative endurance limit of 35 ksi. Subsurface hole geometries are, therefore, not a concern even with a worst case 1/16 inch ligament.
The results of Cooper Energy's analysis of the partial-break through hole (similar to the failure case), indicated that in this case the l
stress concentration factor at the reentrant corner is significantly higher, and it reduces the safety factor below the values used for design.
(The stress concentration factor is difficult to model exactly by finite. element methods, due to the small' local dimension at a feathered edge compared to the mesh sizes. The estimated values vary between 4.55, based on Cooper's' analysis, to somewhat higher values, based on a review of literature by Battelle Laboratories, consultant to Cooper. Regardless of the number, the conclusion is unchanged).
Accordingly, it is possible that the fatigue crack may have been initiated from the sharp undressed corners of the partial break through hole, with or without preexisting defects.
ST/WS100/CDS 3
II The possibility that the surface condition of the hole may have caused fatigue crack growth was investigated. The surface of the drilled holes typically contains minor surface imperfections which are tangentiallyorientedandareconsideredharmiess. A review by Battelle Memorial Laboratories, consultant to Cooper Energy Services, demonstrated that the break-through surface of the hole contained sharp discontinuities which are typically radially oriented (Ref. 5). The hole corners in the failed section were' damaged by the. failure and the features were not sufficiently clear to detect a crack starter; however, the typical condition around the rest of the tip indicates that potential crack starters existed at the location.
Fatigue crack growth analysis, assuming a very small pre-existing crack of the order of a few mils indicates that the crack would grow.
Estimates by Battelle using simplified models~resulted in a limited projected life, possibly of the order of 17 million cycles (Ref. 5) .
The limitations of linear elastic fracture mechanics in predicting crack growth rates from very small initial crack sizes are discussed in the Battelle Report. While life estimates may vary due to assumptions I and modelling, it is clear that life will be reduced to a finite value j by a small preexisting crack. This postulation'.is consistent with the observed failure at about 12 million cycles.
I -
ST/WS100/CDS I- 4 i
The materia) properties of the failed connecting rod were tested, in '
part by HL&P (Ref. 2) and further by Battelle Laboratories (Ref. 5).
The results showed that the tensile strength (105.1 ksi) was well above the minimum required value of 85 ksi. Fatigue crack initiation and fatigue crack growth rate tests on notched and unnotched specimens from the failed rod by Battelle Laboratories (Ref. 5 and Ref. 6) indicated that the properties corresponded well with literature values for structural steels with tensile strengths over 100 xsi. The material therefore had good margins over the fatigue design assumptions. ;
CONCLUSIONS:
Three dimensional finite element analysis by Cooper Energy Services and by Aptech Engineering Services demonstrated conclusively that sub-surface holes retained high safety factors against fatigue failure using conservative fatigue strength values. Materials tests performed at HL&P, and fatigue property tests and studies performed at Battelle Laboratories, indicated that the material properties are superior to those assumed in the analysis, providing a substantial measure of conservatism.
The cause of the fatigue failure in the connecting rod that failed is 4
attributed to the fact that the lubricating oil hole was partially.
drilled throught and the edge was irregular and contained radially oriented crack-like discontinuities, which may have grown by fatigue .
crack growth. The stress concentration factors at a feathered reentrant hole, as in this case, are also estimated to be high enough i ST/WS100/CDS 5
i l
W a
to potentially cause fatigue crack initiation even without a preexisting defect. The growth. rate was relatively slow due to low alternating stresses in the region but the life was reduced to a few hundred hours, equivalent to a few million cycles.
I .
I I
g .
l 0
ST/WS100/CDS 6
REFERENCES:
- 1. " Interim-Special Report Regarding a'Diesal Generator Valid Failure on November 28, 1989"., ST-HL-AE+3326 dated December-27, 1989.
- 2. " Investigation of Diesel Generator-Engine Connecting Rod Failure,-
South Texas Project Unit 2" dated December 13, 1989,~HL&P Report i No. MT-2558.
- 3. " Finite Elenent Analysis of the KSV-4-2A. Master Connecting Rod",
by_ John M.' Morne, Cooper Energy Services, - Applied Mechanics Report AM-1852-C-1 dated February 26, 1990.
- 4. " Significance of Over-Drilled Oil Holes on Fatigue Life of the KSV-4-2A Connecting Rod in the Standby Diesel Engines At South Texas Project", Aptech Engineering Services,' AES 89121166-1Q-1,.
, March, 1990.
- 5. " Failure Analysis of the KSV-4-2A Master Connecting Rod":by M. J. Rosenfeld et al., Battelle Purchase Order No.-341B7139. i
- 6. Facsimile Transmittal from Cooper Bessemer to HL&P dated _l March 13, 1990, with revised page 10 to reference 5'above.
I j
l i
i ST/WS100/CDS
- ;i l 'l i-e I l~
7 i
-APPENDIX 2:
HL&P HATERIAL TECHNOLOGY DIVISION-INVESTIGATION OF 1 DIESEL GENERATOR ENGI!TE CONNECTING ROD FAILURE SOUTH TEXAS PROJECT UNIT.2 DATED DECEMBER 13, 1989:
l I i l
I ,
I ,
b i ;
I -
l i
NL.90.079.02 . ,
MMMW M N
w MATERIALS TECHNOLOGY REPORT
Subject:
Investigation of Diesel Generator Engine Connecting Rod: Failure South Texas-Project Unit 2 Date: December 13, 1989 Prepared by B. G. Steinberg s
.e***
Approved by T.'R. Drews
-e
.j Report Number MT-2558 3
l'
..y.
9 JNTRODUCTION The Materials Technology Division was requested to-investigate the cause of failure of a connecting rod from one of the emergency safety feature system diesel generators at South Texas Project Unit 2. The engine ha(dESF) failedat9:55a.m.onTuesdayforeember Nov 28, 1989. There were no indications of any problems be the failure occurred. The fractures-occurred in one of the connecting rods see Figures-1 and 2 . There were no signs of distress reported in the con (necting: rod bearing):shell or bearing cap.A Also, no fretting was re>orted on the a between the master connecting rod and t se bearing cap.gting surfaces CONCLUSIONS
- 1. The primary fracture oc~ curred by high cycle fatigue in the master-connecting rod through the oil passage to the articulated rod pin bushing, initiating at a manufacturer's defect. An extra hole inadvertently drilled through the wall at the oil chtnnel had been repaired by tapping and inserting a plug. The edge of the drilled hole had not been chamfered, leaving a sharp-edged.. geometric'ancmaly on the surface which acted as a stress riser and initiation site for '
fatigue. When the primary fatigue fracture.had advanced enough to significantly raise the stresses on other areas of the connecting rod,.
secondary fatigue fractures began in the-strap sections. The final fractures were due to tensile overload.
- 2. " Heat tinted". areas of the fracture had not undergone any microstructural or hardness changes. There were no signs of " heat' tint" or heat damage at any of the fracture origins. >
- 3. The connecting rod was made from AISI'1050 carbon steel. The-microstructure was uniform, fine grained,' partially spheroidized pearlite and grain boundary ferrite.
- 4. The mechanical properties (tensile strength elongation, reduction of area, and hardness yield strength, percent of the manufacturer's Specification No. C-5 all. formet the requirements forgings thicker than 7 inches.
RESULTS AND DISCUSSION Several 3 through pieces were received for the investigation ar.d are shown in Figures
- 10. Each > art received had been stamped by the manufacturer,.and these markings are slown on-the connecting rod schematic (Figure 2).
Macroscooic Evaluation A macroscopic evaluation indicated the primary fracture had initiated on see Figures 4, 12 and 13 This hole is ~
both sides of perpendicular a threaded to the main oil c hole (hannel and reportedly
.been uninten- had).
tionally drilled through the wall. To rectify this condition, the hole had been tapped and plugged. M're were also two half-moon shaped zones of heat tint on the primary fa cture surface place sometime during the failure process., indicating-heating The two straps which hadform taken the top of the articulated pin bore appeared to have failed as-a result of. the 1
MT-27Q
, , , , , , , , , , , , - . - - - - - = - " ' ' '
?
I; primary oil hole fracture, and therefore the strap fractures are considered to be secondary fractures bearing cap bolts and articu(see latedFigure.s pin bolts3were and 5).1The deformed fractures and displaof the a features characteristic of tensile overload (see Figures 8-and 10). yed 4 Primary Fracture '
~E '
E The Figures primary fracture 12 and 13 surface disp 1:iyed characteristic signs of fatigue (see mistakenly drilled), then pointing tapped to aand double origin plugged, hole. on either The edges side of of athe hole were sharp, acting as a stress riser. The macroscopic features of the
= fracture face indicated that the fatigue crack had propagated over a very 1arge percentage of the fracture surface before the final' overload failure
- took place. This indicates the cyclic stress levels were~ low and the critical flaw size for the material was quite high.
r A section containing the origin area of the primary fracture was examined in the scanning electron microscope Closely spaced fMi 2 striations were found, confirmin fatigue or low cyclic stresses see Figure 14>(g . The the(SEM).fracturehad.occurr closeness of the fatigue striations indicated the stress level)had .been low and the-part' had '
g undergone many cycles before failure.
! 6 Secondary Fractures t -
The strap sections of the master connecting rod which hold the articulated
- rod pin also appeared to have failed'by fatigue when they were examined macroscopically. The fracture origins were at the outside diameter of the straps see Figure 5 .
y been sec(ondary becaus)e when the fracture surfaces were exami m the fatigue striations were more coarsely s n.ced than those of the. primary fracture (see Figure 15
= the stresses were raised). This indicates tie straps began to fatigue when as a result of the primary. oil-hole fracture.and therefore experienced fewer cycles before the final failure.
One of the straps examined had two fracture origins (see Figure 5). Both origins e'xhibited signs of fatigue.
Final Fractures The final fractures, seen in the articulated rod pin bolts and the bolts
-- which connect tensile overload. All the master the bolts showed connecting macroscopicrod signs too theall occurred bearing b cap,f pre-fra plastic deformation and necking. The fracture surfaces examined in the SEM were The features
- 17) and fractureductile, dimple fracture (see Figures by 16 and tensile ovcricad modes seen are characteristic of failure .
-2 Failure by tensile overload was further confirmed when the threads in the
bearing cap bolt and nut were examined at higher magnification. Both the nut and bolt contained sheared threads from their mating parts. In addition, the thread crowns were plastically deformed (see Figures 18 through20).
=-
) 2 MT-2558
Metalloaraohv The microstructure was examined at the primary fracture origin,-at a " heat l tinted" region of the primary fracture in'a re fracture,andinathinregionoftheforging.gionwellawayfromanyThe microstructure was the I same wherever it was examined and consisted of fine grain JASTM grain size number 7-8) partially spheroidized pearlite surrounded by ferrite grain-boundary (see Figures 21 and 22).
There were no fracture origins. Thismicrestructural indicates he materialchankes sufficiently to alter the mechanical had not-been at the heated ' hest tinted" zones 1 or l< microhardness traverses taken across properties. This was verified by the " heat tinted" zone, the primary fracture or.igin and a region away- from the fracture. No changes in:
J Ei I
microhardness fatigue process.we,re seen except for those due to work hardening during the- '!
Mechanical Pronerties and Composition i The mechanical properties of the material- were checked,- and the results are ~
within the requirements of the manufacturer's Specification No. C-5B (see Table 1). The composition was also checked by optical emission spectrosco)y and was found to match both the material test re mill and tie requirements of. Specification C-5B (see Table 2) ..port from the i REFERENCE 1.
Conversations with Ted Fryar, S ecialist Consulting ngine'er, -
South Texas Project, Houston Lighting & Power, Decem er- 1,.1989. ;
I '
I 4 o
I
^
I a
I i
l 1 3
MT-255B
___,.p ,,..=i-"" ' - " ' - ' ' ' ' ' ' ' '
e i
Table 1. Results of HL&P Mechanical Property Tests Compared to the Requirements-Per Specification C-5B and Forging Supplier Report HT-60ll354 Ultimate 0.2% Offset Tensile Yield Percent ' Percent Strength Strength Elongation Reduction Hardness-fosi) _
(Dsi) - in 2" In Area (HB)
HL&P Test Results 105,100 -64,600 19.0 51.0 207~
i Material Supplier I Report, HT-6011354 108,800 66,300 26.6 58.1 ---
Specification C-5B to 4" thickness 90,000 min 55,000 min 20.0 pin 39.0 min 187-248:
4" to 7" thick 85,000 min 50,000 min 20.0 min 39.0 min' 174-248 over 7" thick 85,000 min 50,000 min 19.0 min 39.0 min -174 248 i
i I
MT-2558
Table 2. Results of HL&P Compositional Analysis Com SupplierReportHT-60ll354,SpecificationgaredtotheForging
-5B and AISI 1050 (WeightPercent)-
S 's HL&P Report Specification Element Analysis HT-60ll354 No. C-5B AISI 1050 0 0.53 0.55 see note- 0.48-0.55 Hn 0.78 0.75 0.90 max 0.60 0.90 P 0.010 0.010 0.050 max 0.040 max
$ 0.037 0.035 0.050 max 0.050 max Si 0.21 0.20 --- ---
Ni 0.12 0.10 --- ---
Cr 0.07 0.07 ---- ---
Ho 0.02 0.01 --- ---
Cu 0.02 0.02 --- --
A1 0.03 0.03 --- ---
l 1
Note: The use of carbon steels having-a carbon content of over 0 '5% requires approval by the manufacturer, i l
MT-2558
- 1 Y.+ 3,4N,,, *: - ~As 4C . .C ~ '~
l y .. , -y; , . sp% ._
- ., 7.,
t
- w .c" 2 . .- .
.9C:y
+
L ' .. 4t r - - '
r.,
m ,. L:
b.(J" '. [.
y.- -
,~: ;
' 4 ,.
l
{~f..a A s L . l- ; .
v.
r e
c' ,~,K, .; . . . ,
)l'yQ.Q.
p.r.:
4.p :; y .;
l = +, u: ,
- j 4 .,, i
&y . ; 5 [ ..
- 3 t.k;g Q:; ;%G~y~* ' 'i.
~
~
- 4 c y
^ % ' .
( > .
~, '
Q.f@
[ Q .l .
. , f ,; 5 g ?L s
m; g : $.m :F' ' "
- > .s
- , c.
y _
. c.e ,_ ,
.t / _
- .. ... 3 .,;
- < : ,xx ~a %
.a y .
(y. ;a .
q.- g%
. .g[i.[ M S} }Jf . ~. Ty;N %g-
=
j
- . , [ [ [ ; ._ ' ' _
- j
- l. ^
.p y[f y Qg
~
n;. -g, p .
- w. ;4 . s . a. .. + :
' ..v-
% "'y,_ ' l$; '*
r. '
d.
- $ g
.; gJQ7*:
-31 *.1 -
2 C: i.; ' % 4."'
4 w n g::~ $,^ e.&
49 r ..
pp
- ~~',, ,
i b
) ;
- s - .
.g y
nr> Vf .f i ,-
%kTN.Yq ;k j u. ': ik
- q e; ~: .J"J' ;e<. ~6 l r.. - ;.4 ;;,,
.y.s,7
. q i Q, *-
u.: ...
~ .L.
f*'!
J. ; * .; '
~~ ,
y#
0 '
s
- a. .- g -e p '
, = , ,.,
,. > v
, ' {i,
... .- .. . .. , q. ,
a; +
, ~; ;
i :'
~...[ .~.";@ h
, . . ,.;fls. <
J' #*i e .j ./ :
- .,j% .a%fy.Q: s . , 9; ' $ Q .'; -
l g , ,. . <
, p w. g, -
- i. k ,.- 1.n g:
.f.
.. = [
?
k '. -
.F '
l l j;. l&ff. - Qp% ^ ,
? ? -] [i ; W l '
L kNL' I Figure 1. Phttograph showing the failed connecting rod shortly after the failure. The arrow points to the primary fracture.
e MT-2558
~
E
- nO 1
i . !
Articula5sd '
Connecting Rod -
Mastw i e.connootin, noe q LUl10 i Secondary 8 B2229 Fractur 173 10 l 619.4s I Articulated PinW 8Hosos 1
I -
$ Prima l g
! Fracture '
'fML . l N
Bearing Cap.
t - T _ l Nuta 4g l 8A2006 and 7K3117 qse. -
4 1 and 4 I
4 8M0221 8M0218
-l .' i I
Figure 2.
Schematic of the failed connecting rod showing the primary fracture location and identification markings of various components of the connect!ng rod.
a MT-2558
~' - - -
^
f'\
'~
[ . [.3 c. ?
f.- f .
?; [i ,_
^
I i 1
,. u -
- 4
. . s.
.- y. ;.'
1 ;-
. n, ,
E }A \ 4 ~
h &
I
%.*. {, # .
, ,f .-
r p'
. 3
.-..-_e., . b.
_,, an . ,
, l* _ ,. -
k".'{",,_,; ~ "
s
~
j.e,,A ., , . ,
. 5-
. 7;
.< q, . ;%n -- vh ,
n 3 5- o ?ft I
+ -
s . ;p.
. g' .;
j
.J
. . , a- .
(( t
- 4
- i , +
,..g e; - "
Q. *#
, f ,,l. .: t ;
T .+
q--.~.7 -
g , , ,.9%-c .yg
- o c ."
I . .ug,wMamww amtsasemanMe wwww=aw may 2* kW? *?
i($ S J?. j[AM w %,J[ ' a . . k , ' , .-
~
h.Q$Y ' ' ' ,
- , T ['"? 4 .
t
-[ #
Figure 3.
Photogra rod siow nh of one side of the fractured master connecting both the prim and secondary fractures black arrows) arMa(whitearrow)0.1X) gnification: j I . % ,. **.
t
.... - - .. -MM-
-a a .
' tag n s.a.> ?
o M
l E ' '
., f "
{.
.g 4 ;> -t. .;
u:3 ::y QllCeQL' ;li. ..
),
J ., ^ ' tji P
e.; ..,y; 44 - J}g f. ,.
1
.y f-g
.p _ fi; y . ,; , ; . g.
.j{ - . __, ;
3 - y7 N ' *^ W .
g u g ; : : d s M e r ; M ..
l Figure 4.
Photographshowinkagnification:the to the otigin. ( 0.5X) primary fracture. The arrows po I
MT-255B
l L,. "b, . . . .
' i ?. s g .- '. '
) - ..} g a-t -
.', '11
...w .
y <
. , ,s . .
5 ggK$hL ,
, .r ,,., x
- e ,
% ~ r; , 9 <-
f -
? ,
r
,g
~
.. : ~
- q w.9- b g . .
.:-:.. f 4
' .).,k-'- , ..~ .
-?
[ ~ *
,Q
"' s r, s. .
~-
4 u g
+. A
+f'
- j. .. '
i
= ,.
- e y -Ws
~ -
,, -+4.j ,
4 ~[$ -,4.,
.: .. . c 9 .- . ."t .
. , a q , !,
y:. .-
..>+..?
. , ;7 .
Figure 5. Photograph showing the secondary fractures in the straps.
The arrows point to the origins. (Magnification:0.4X) 4
. s a '.Nu$, .
r.
'd J. ' . . . [, > g.@ e , q 4 g [i *' O y- . :.,u 4, ... - ,; - % -wN% &l ' q; 7.;.,n. , ./ . A ?
. ;q % g
'? l f;
g ..
.-r.
n: - -
n
%qr4 2:mg % g' '7. .
n .
, ,m.; -
7 .
." p.. ,
+ -
K:.n_<g;:n - .& R@ f %gt n
+ - p . . : u. ,..nn ,y y.
~
k
~
l
'l- Y.
- ?Q r" &(_ y,. %p.mypA .. 'n.
, . , '- - :_a_ .; -;;); fig.
k~w . p y x,w% .. + ya . ,
~ .c
,7. p
+
9 Q.' ~ ~ igdg f % ~
g
? ' , by Q Y
g.: ' - , z.'7 3,By. t / ?.' -
i.["i)(*t%jj$
.- /gf { r]
w ., ,
.y
.b.?yp
- t - '
. , *. a w
- m. a zy a ,,.
~
w: .L . .
.e ,( g eA g . g . ( , jx.,l %n-.:.3+ggh'
~...;A: --: . . , ,p
- c. ,.
- v. -
~.r 3 , :
' jygg }..
y; . ^ .
tj y
[ %
- e. w n
?N ,,.-a~
- k. , .-h[k[ $ f u u a e 'n'~ '
7i :n n.
~
b N $ '
Figure 6.
Photograph of the mating half of master connecting rod.
The ar ow points to the mating half of the primary fracture (see Figure 3). (Magnification: 0.3X) !
MT-2558
~
y '
{-
n
,{p *
' +
1
.c .
I #
$.1 h . .yt _- # l'
- g..
w g ,f, :
, . " ' .g ' " -
, , 4 g r 5%
s ,
..q; s
- se y7
.g 4.
?' : 1 N "D, _
{ ~. , . p , ...
- s. 9 .. ; &
s.,
g t .
~'
1:-~ ; hd' n.,4
~
g!' *
^
, ;, s.
[
% y?- Q
~
' ~ '
.y .,..4 x,
( f. g f.j;f '
, . - . : % ,5 ^~*
m3 -
.g. . . ; .
k' =~i * . , , * * '
p'
%Y
....e ~
j, -
. ._ g p % ~. ., 4 Qi '
.a
.? f.- -3d.3hyff:l.
- ..,. ,. y i v:/*,s . . _ .
[ -
a, 'l
. . . . . ?
l 1'}&?ylh'#t: .\>~
., *t Figure 7. Photograph of the master connecting rod showing the location of the primary fracture relative to the main oil hole in the crank case bore. The arrows point to origin.
(MagMficatior.: 0.3X) jgg ,
f .. .
Fl,
.. pl, l ... y -- Q gn[ ..
f
.- n.
~ .
.,,., s.
. . ;f -
[\ . :. ~ ,
r
,;^ @ I2*: ,h W .'bi'l. 22.' ;&% , Q
. + . .m .. .
-f.
\)*C . yR i'0 b.hpr 4? $y$'* c..% .
h ;sV%k ?
N
'??, % .
V (~. . l . E' O.
z ,.~
-i
.x
.?4 'D ~ . }? -
,f
? ,. x.. %.., x.
~
k .%% ;v , v. ~ .e %fR e
A- % , 7. -. %'T f ^ g ,,m p.Nyk,J .. ' d dg 4.++ (;3 <.; 4 2
.a t
m s..j "g p.n .. p2-.. e.
- x. g. ,, .
.. c r% -
wg<
h$ - I
d4 '. .
-. g, 4 m* .-
7? , Q .<:'975 4 3 W, . , . .
, , . .y, g m W.4.
W 7. 5,j.k;;.y{
t 5.in m q g. Ne
~
. g..h.- -
?3. - .,,.,- -: ;. .4,h-e , ;=g f s.t.X. n g t.y .
- . , . . 4
- %dg,. e
- w. g s v w e it ':
1 f -Y .
. ., y ft: 5; . ?* . ?
Figure 8.
Photograph of the failed bolts and nuts wh"h connect the master connecting rod to the bearing cap. Cagnification:
0.4X)
MT-2558
~ -
- ~
,\T :(( ifs.
,)e(3 x.> @
3 .
- 4g. ; s .s . .. <
q e J/
- y. :
f ./fpf. g d. y,r_;l' ._ [ \gy- .
^
E ,
s
% W >
n.:. n$ ..
, A ,
.i ., q g;. -
4 l
~
- Q *. y ?O" ; 'l . N l
- ?.
y , % ,j . s. ,
-gg,- ... . .
Wa. : ,
,. - , v. .
. ~ . ..
I 4 , .s=V r .k ' '. / ,
, . J p y ).> ' . . x.( -
X f(t',
. q r
I :; .. .s m
_y ff Aw..; . 4 p
1
( "9 , ,g I
.4. );. x u, 4 ,
, ,y . <
o
... . 7 - p
.y:' ('.l ,[h. .
, l J.We ,4 g
!k
.. . ,,p
- , m -
I . .i. .." gr ,,.
f,1
. '., ,f .' , %
4:1 d' f.
D. '-
g T u . ..
Q s.
a
..s
+ .
3-
.M~~m s 'i a
= i s-.ilf g< .
, , , , V -
Figure 9. Photograph of the articulated pin showing the location of the failed pin kits. (Magnif(ct. tion: 0.6X)
( .. .
w + . - Q r.
n - j. p
. r.,.
es yi, p
, p 4
,4-
< s
. _ y , ;11 - y y
.A k'
I# .J ','
- 'g y'nAsg
..,sy .~ h',f*'-
t .x
,.,7 .
~
y -a
.aw;
^ .~ m ; - .y
% p'~, ., . , .
~.4] ,7 +; ;
h.
f.af, -- %
- *y - + . . c .p -
'1- .__%
. p% ;, e 'y4.* *
'(y ;<Vg ' j 14
{A;j . . . f '; ,~#
.. . 4,
- c. J. '
\% 3 gs,. - 1 g
_g q:~, .
g, +
. w f(- ;h'- : 9 . .
j., TAR .
3
- k. Y,' % -l 4.g4.w:,. - 1 .
- 7. ; $
4;4 A[
- l,,3 q- -
L.}
p yf, ,M}.f. ..
mm-:_
. ;Re;ml ;, --lll*}fi,Gf;.c n
s-
, . yg . y. . . . m# v. ..-t'.. m ! , y* *5,.,; .
. ; .'u ; & .
%. . C -
.;, - # - ~ ? N s- .
NI i g M; f > ,f[44 v 4. ge
-v g y~ g . A(.h' [,. g.fg}. jekN} j] {.j.gc.-.
ev J-
, . v t . y. e 9
We hyg :
%.Q.-Y.'%d y.m., ., - ' j.$f:nW. z
,;c.x.
- ^
.Y. ? .? '
- &0w m
_ > ~-
' 'y
@.-a; ). L Q 4? is . . ,l [ .) P . l' ?+.. *
g%.f: . . j;;.
\
%".. .. . .. ;. y op, ~ , '~ l * *pW- . (. .h.* ' 'L. .W f' <C '
Figure 10. Photograph of the failed articulated pin bolts.
(Magnification: 0.5X)
MT-2558
- n .a 17 y w - jgj
I i
N.? .
f Q ~4
- [$ ,
- rgqs: % M .
' a ~ QD .- .,%
,:,.. . ve . n. .
h-I k
t E. .
Figure 11. Photo drillbaph of the
, then tappedprimary and fracture plugged showing hole and some the mistakenly of the heat I tint seen on the fracture surface articulated pin bore side is at th arrows)(.
0.5X) top.
The Magnification:
~- ~
.,,.,.p f www w - v. .
hip 9 . .
sy f ,. .-
gG~t;. u: :s._ .
1, ,
..' d i-i.j Q h e .,
I y i i l' I i fl
! ! Mll ILI lg [ l i L jME
~ "
M<& yja J ! 85 I -
m-+
mm ,
w j. g h lQg @Q - j i Figure 12. Photogra h of the primary fracture showing fracture origins large arrows) at the sharp edge of the mistakenly drilled ole and macroscopic beach marks (rrcall arrows) f f(
l
! j characteristic of fatigue.. (Magnification: 0.SX) 4 MT-2iSB y; y:.(;3 gg g yg g mggwsl
I .
7=-
g /\,y
_ ,p. y. + . . _ - sgun :% ?'"f l k _g .
.}. .
]
7' .u, M? * ; - .
? %* . , 2: % %:%?d ~1
".:;' :' % _ : SV .
-l:: . .
,j j;- p . . . -.
. .),
f.;; q l q }{- AhN$ h f ;
=-
N hk%g :
l Figare 13.
Photograph of the primary the double >
at the edges fracture of the dri1 showinked he' - 3 ori (Mains (arrows)6X) nification:
j 1
y .
~
pi 5
]
I' h
I Figure 14. Scanning eiertron microscope photograph of the primary I
fracture near the ori in showing the fatigue strintions.
'# [ .
The striations tre or ented p endicular to the direction I
of crack opagation. The or n is in the direction of ;m
{ the top 1 t corner of the photogra h, and the crack is propngating away from the origin. Hagnification. 4000X) t
\
hl p NT 251>B o
- ===
" ~"
5 p
%r g,. ~+ ~'? j,. ~. 7, gj
- g y% {y' p f?f p, . ?
f W:, '
yg y , "G
~
h' . h ., 1, .,
.,4 7 %.;- 4 5 R'.g. *J 59 L -
'l, (; , ? %
- 1Lg y i
- ; *
.f- e q.",9, - .
.3 .,
.- .t ) ,, w r :
- f[. . .
-* '.i, f + _-
- ,9 p;df
). .$ ! =Gp? as<' .;. , 7?: ,
iti ; . ..
- N) MW L%O ; f h S Figure 15. ph of one of the Scanningalectronmicroscop$chhoto secondary strap fractures s ing fatigue striations near the origin-, These striations are much coarser than those seen in Figure 14. The origin and crack propagation dirnction are oriented as in Figure 14. f, Magni ficat ion: jJ 4000X) i w
L i
I I
... i
%I
)
- x -
E ,,.___4 Figure 16. Scanning electron rnicroscope photograph of the fracture l
wtface in the t'e>1ts conntectinD the mt. ster connecting rod l I and the bening cap. The fri.cture mode is ductile, dimple i
f rupture. (&gnification:2000X) ;
A l
n Pff-2550 R 1
- ' f
[ ,_ RRW)RENREEROFDEM Mf# E % N+...."##
jygtRYj p4h (
.h I
. p. a # 4,g., -
W .; ... u .. < ...
'e ;;, '. *
- L.( m ; ;: q :
- f&
- '
s.. y : f,h gn . ; eip ; 3,
- {}l
^
l0*j .E*f! .
yg
,.. ' .w
,g 3'
?
$y#A g.
- ' ~
y ' T . ,., , y .
. _.+ww ' ,,
- a. ,
Figure 17. Scanning electron microscope photograph of the fracture s surface in the articulating pin bolts. The fracture mode is ductile, dimple rupture. (Magnification: 4000X) { '
\
f ;('.. _.^ __..( ..:'- _ _
l
.p$1Iv. : '.
1 w,,7;,
., j
, : . '. 1, : 0
- i-.
l ?:
4-
- 5-
'.- { ., . .j.:....~..:: i
'3
- I. :
f' .6
.3 4
.h ' A f . ,;. . f' t ,ta-l7
~
l h {.j .R NI;(h : ( ia.
k$[N '
} .
- = ; g
- , 3 4 ;: ;'); : y r '#., -
l
- 3. Qc.L.QQ.k[,..f Figure 18. Photograph of the end of one of the bolts which connect the master connecting rod and the bearing cap showing both sheisred threads white arrows from the matin plastically defor(med threads )(black arrows). g nut and (Magnifict^ ton: 1.BX) l MT-2558
\
l
, p .-
c mi n i
I 7 A e .L :lj
_N ' ?
l 3 . p ._a ~ 5%+,3 %n.g. Q-[ & s '.' )
( fjw.,;1}[ :p3.
, ~. I g q/ 3 ;* [ - ~ E. .. 1 ;,
.. ag%. ; ::;< [ 4..j. ,.4 7 -L~' '.'~ -
- j-
.. osg . .. . :n. ::. y i q .: f * '. } ,, . y ;_; ; ,.
< p p eg y p i.l ? l>. T T " $ 7 g g g g g ;
i g ..
Figure 19. Photograph of a section of the nut showins the flattened thread crowns. This nut contained shearec bolt threads which fell out when the nut was sectioned. The arrow points to the thread examined in Figure 20. "
(Magnification: 1.8X)
,G - -
- 2. . . .
x
. &'Q
$( ,
a ... vgema s, ... e-p . y .
. 't:* . . ' b 4 ='
h , ~~ w .% Qf 5.< .~ '
/ ... o. '?
..5*
w i'v? . ..
.?. . ,
isWLt><_.d Figure 20. Photomicrogra is of the last engaged thread on the nut showing the t ical mechanical damage seen. The top of t
the thread forces seen (d ) ing has been flattened and folded over by ths >
the final tensile overload. The small I triangular section of metal to the left is a portion of bolt thread. (Magnification: 100X, 3% Nital etch, MM1871)
MT-2558 .
I -_
4 I J s
v figure 21.
Photomicrographofthetk>estructureisfinegrain failed connecting rod. ical microstructure seen in the i ;
pearlite and grain boundary ) ferrite. (Magnification: i 100X, 3% Nital etch, MM-1866 w --
.y lg.j, 'f X. Jy 'V. '
.i '*
, ' A
;ff - . ..
- i. v :,
y
;_ y h ,;;' l
,_; <e
.'L _y ; t
[;. ; 14
!y + '
j
._ _ A. -
s ;.w,. , 4 L . I
-.x z d$.
Figure 22. Hi h magnification photomicrograph of Figure 18 showing th fine partially spheroidized pearlitic structure (dark phase phase . and free ferrite at the grain boundary (white ; (Magnification: 1000X, 3% Nital etch, MM-1866) ( MT-2558
l I DISTRIBUTION I .- Requester: 5. Timaraju j l : cc: R.'L. Ulrey Materials Technology Staff ' I. I I i i g 1
; j i
e E ' i i MT-2558 !
, -(:
1 i i
/l
- i i
APPENDIX-3. , AFTECH ENGINEERING SERVICES-SIGNIFICANCE OF OVER-DRILLED-OIL HOLES ON ON FATIGUE LIFE OF THE KSV-4-2A CONNECTING l ROD IN THE STANDBY DIESEL ENGINES'AT SOUTH TEXAS PROJECT -! DATED MARCH 14,.1990 i
.I I 1 l,
g i l.
~!
NL.90.079.02 7y . i 4
'i I. .
x
..._...~... . . .. . .. . . . . . . .
I I- . I
-l
.I. I . . i I 1
~
~;
APPENDIX 3' APTECH ENGINEERING SERVICES-SIGNIFICANCE OF OVER-DRILLED-OIL HOLES ON ON FATIGUE LIFE OF THE KSV-4-2A CONNECTING mm ROD IN THE STANDBY DIESEL ENGINES AT SOUTH TEXAS PR0 JECT, , ss DATED MARCH 14, 1990 in 4 U l w
=i L :
l g
- t- :
11 'l 3
=s m
6 NL.90.079.02 3
. _ _ , , , , _ , . _ _ _ . . .. .. . l
g AES~8912116610-1 CONTROLLED 1 SIGNIFICANCE OF OVER-DRILLED OIL HOLES ON FATIGUE LIFE OF THE KSV-4-2A CONNECTING ROD IN THE STANDBY : DIESEL ENGINES AT SOUTH TEXAS PROJECT H Prepared by i Russell C. Cipolla Jeffrey L. Grover i Philip M. Besuner - Aptech Enginee' ring Services,Inc. ' Post Office Box 3440. i Sunnyvale, California 94089 3440 .; { i l i Prepared for - South Texas Project ' Houston Lighting & Power Company . Post Office Box 308 - Bay City, Texas 77141 Attention: Mr. Siv T'mmaraju;. March 1990 , 1257 ELKO DRIVE O SUNNYVALE D CA 94089 0 (408)745 7000 t POST OFFICE BOX 3440 0 SUNNYVALE O CA 94068-3440- 4' OFFICES O ALEXANDRIA, VA O (703) 683-2062 O HOUSTON, TX O P3) 558 3200 , 1 A ' ' o '
, , .,; g ;
p ,9
, i
f l l VERIFICATION RECORD SHEET l REPORT NO:AES 89121166101 l. l, TITLE: SIGNIFICANCE OF 'OVER DRILLED OIL HOLES ON FATIGUE' LIFE.:OF THE . ; i KSV 4 2A CONNECTING ROD IN THE STANDBY DIESEL ENGINES AT SOUTH i '. TEXAS PROJECT Originated by 1% MEND:
" Project Date ihi se((
fY ' 8D m 3 - !] - ?e Project lifigineer Date - 3//J #0 Approved by - Project Manager 6 ate' Verified by A--- = 1 A 3 /3 70 4 Verifier - Date -( I Quality Assurance m /7 #/O Review by Quality Assuraince Engineer Date Quality Assurance Approval by'
/3/90 Ouality Assura6de Manager Date - '
l - I I I
-- - . .- . - - .- - - ~ . -- .
ii l
-I j
) TABLE OF CONTENTS i ! Section EASE j EXECUTIVE
SUMMARY
iii
- 1. INTRODUCTION 1 1' 2 CONNECTING ROD DESCRIPTION 21. )i Connecting Rod Geometry 21' Oil Holo Geometry 21 Hole Tip Geometry 21 .j 3 STATISTICAL ANALYSIS OF OIL HOLE DEPTH DATA 31.
Introduction 3-1 ' i Measurement Data 31 ,
;i Analysis Method .
34 Best (Point) Estimates of F(R) 34 Confidence Bounds F,(R) of F(R) 3 6- , input Data 38 ! Results 39 l 4 STRESS ANALYSIS 41 General Description 4 11 Rod Loadings Assembly Loads 41 1 Operating Loads 42 Global Rod Model 42' . Local Ligament Model 44 Summary of Results 46 Global Model 46 i Local Model 4'-14 l 5 FATIGUE EVALUATION 51 , 6 DISCUSSION OF RESULTS 6-1. 3 L 7 CONCLUSIONS . 7-1 , REFERENCES .R 1 1 I
1 iii ;
')
e
-i P
EXECUTIVE
SUMMARY
The Standby Diesel Engine 22 at South Texas Project, Unit 2, had' experienced a connecting ! rod failure on November 28,1989, during a 24 hour full load test. The diesel engine was l supplied by Cooper Energy Services and is a 20 cylinder vee type engine employing KSV-4-2A -
, rod assembly. Inspection of the failure region by Houston Lighting and Power Company (HL&P) revealed the failure was caused by f atigue cracking originating from a lubricating oil hole which ' r had been over drilled to the point where it penetrated the rod pin bearing surface.-_ This penetrating hole condition was repaired by tapping and plugging the hole, during manufacture. '
The fatigue crack origin was determined to be the sharp corners at the surface that remained . after drilling. Although it was determined by the vendor that this was the only rod assembly manufactured: and repaired in such a manner, visual exa'mination of other rods revealed over-drilled holes to I
- various depths but they did not penetrate the bore surf ace. Because of the concern over_ fatigue !
from over drilled nonpenetrating oil holes, a statistica! analysis of available h' ole depth data was performed to establishthe worst case remaining (minimum) ligament between the hole and the bore surface. A detailed three dimensionalfinite element stress analysis of the worst case oil. j hole depth indicated that the stresses in the remaining ligament are not large an' d that this location is not the highest stress location when compared _to other normal geometric- ( l' discontinuities in the connecting rod. Furthermore, the cyclic stress in this region is approximately 14.6 ksi which is below the fatigue endurance limit for the material with a. l computed safety margin of about 2.9. It was, therefore, concluded that the over-drilling of an. . oil hole that represents the worst case nonpenetrating condition does not pose a finite life problem for the connecting rod. I l 9 lI
11 ! [y . J i 1 Section 1
- L INTRODUCTION The standby diesel generator at South Texas Project (STP), Unit 2, experienced a failure of }
No. 22 dieselengine on November 28,1989 (1). The failure occurred during the performance < -l of a 24 hour fullload test. The Houston Lighting and Power Company (HL&P) determined that- ! I the failure was caused by a fracture of No. 4 connecting rod assembly. The diesel engine was ; supplied by Cooper Energy Services (CES) and is a 20-cylinder vee type turbo charged system y ' i with a rated output of 5500 kW for the purpose of providing' emergency backup power to STP 2. At the time of the failure, the dieselengine was reported to have 634 hours of operation at 600 rpm. i i i In their investigation, HL&P determined that the fracture of the connecting rod assembly was due to fatigue cracking originating at a lubricating oil hole (2). Inspection of the failure region
- revealed the oil hole was over drilled to a point where it penetrated the inside surface of the .
articulating rod pin bore. Although this condition was repaired during manufacture by tapping I and plugging the hole, the surface penetration was left in the as-drilled' condition without removal of sharp and irregular corners.
~
I lt was further determined by CES that this connecting rod was unique as it was the only rod . repaired in such a manner. However, visualinspection of other connecting rods available for 1 examination indicated that it was not unusual for this oil hole to be over drilled to various-l depths without penetrating the surface.' The worst observed condition is a remaining thickness - , of 1/8 inch between the hole tip and the articulating rod surface measured along the centerline ' of the hole. This condition corresponds to a minimum ligament between the hole and bore at the shoulder of approximately 1/16 inch ' Because of the concern over fatigue susceptibility of other connecting rods in service at STP, HL&P requested Aptech Engineering Services,Inc.
~
(APTECH), to perform an evaluation of the over drilled hole geometry including statistical i
a ~ 12 analysis of existing hole depth inspection data, a detailed stress analysis of the rod, and a _ fatigue evaluation.
?
r The primary objective of this evaluation was to determine the impact of having over drilled oil holes on the f atigue life of connecting rods currently in service. This evaluation was perf ormed l in parallel with the vendor's investigation with the objective of providing an independent verification of the stress analysis results. Loading information was obtained from the vendor and reviewed by APTECH for input to the stress model. This report summarizes the statistical analysis of inspection data, the finite element stress analysis of the KSV 4 2A connecting rod for the worst case oil hole geometry,'and a fatigue endurance limit evaluation to establish the margins of safety under worst case conditions. I I I I I I I
.~ . .. - - . . -
21-1 Section 2 CONNECTING ROD DESCRIPTION :i I CONNECTING ROD GEOMETRY The KSV 4 2A connecting rod assemb;yconnectstwo opposite pistons to a common crank pin .! bearing asillustrated in Figure 2-1. The master rod is a solid spar which transfers firing loads- ! directly to the crankpin. The opposite piston uses an articulating rod which can rock relative I i- to the master rod through the rotation allowed by the articulating rod pin. The area of interest - - is shown as Section A A in Figure 21 where the central oil holes are illustrated (1)l For modelling purposes, the dimensions, of the rod were taken from engineering drawings for r the master rod and bearing cap Q, d). The nominalinner diameters of the crankpin bore and [ articulating rod pin bore are 12.25 inches and 0.025 inches, respectively, as shown -in Figure 2 2. The angle between centerlines of the master and articulating rods is 53 10' and' l the distance between centers is 10.437 inches. n-I- OlL HOLE GEOMETRY The 3/8 inch lubncating oil hole is 12 off the centerline between the rod axis.- Section A A l in Figure 21 is shown in greater detail in Figure 2 3. .This illustration shows the section through the center of the oil hole and the 11/32 inch transverse oil hole. Also shown in the section are the 1 inch wide by 1/4 inch deep oil groove, and the two outer 3/8 inch through-holes for supplying oil to the periphery of the rod bearing surfaces. - i HOLE TIP GEOMETRY The geometry of the over drilled hole tip near the articulating rod. pin is complex in that'one L must account for the local tool tip geometry, the hole angle of 12 , and the curvature of rod I f i
h 2 WCPTSCHI
<~o;uaonavcis*
i 1 N Piston, Pin - - 7.1 1 t 1
.,, ,4.. .
# u I'Artic01a'ted c l Connecting Rod Master Connecting-Rod ,
I Articulated hod Pin % f dV . 011 Passe - l' I Oil Passages y Q Which In1Nated raiiore 3 ,y i Setring Shell 5 Section A=A g w Searing Cap > Figure 21 - Schematic of KSV 4 2A Connecting Rod Assembly, i
---..-.-w.ws
.. . . - . , ,,-...u.. .a.-+.e- ..-a .,, aw_. u . a, .- , .. . ..,a.
.Y 2-3 $
\ . i = t 6 i
$ (
- l. . ,,
,l - \ /
/
.l g
/
\
$k;xj/ y
{ ,, 4 'l s i o t g (\ _\ l 1 g M ( > n k \
\ ; .
x
\
F \
\
\
.\
\
\ -
r \ ' i . r 1t
\
1 I . . . - - .
'l l g 4 IE51AP78 CHI e c.*iam siaen * - C1900 APTECM l1: I 3/8-Inch 1-Inch Wide-Oil Hole Greove I Crankpin
; e- Surface 1 ,
- 11/32-I nc h' !
Outer '* L 'i > Transverse ! Oi) m \ 's \ Y s -\ 'i E/ 011 Hole I Holes
~.._
N,NNN
..w__=-
y,\s v y,
~
i lQ = : ko i - 1 I Rod Pin Bore Surface ! i Figure 2 Local Geometry of Intersecting Oil Holes. ; g
25 pin bore surf ace. A schematic of the oil hole geometry for an over-drilled depth of distance d, 'I , from the crankpin bore surface is shown in Figure 2 4. The drill tip angle,20, and tip land distance, b, were estimated from plastic mold taken of an over drilled hole in a spare rod. Closeup examination of the mold indicated a nom,inal tip half angle of 60o and a minimum land at the tip of = 0.070 inch @). The tolerance on the measurement for 6 is approximately 3.4 . 4 I I I I g . i I
f p ,.- 2 .k.. f, Articulating Rod (- Pin Surface I i di l te h f I H h-b y
"+
(
<, 4 1 l dp '
I ,, l- + -d' i
/
V <
\ ;
a i l 1 .
,, t i
Crankpin Bore Surface Figure 2-4 Local Hole Tip Geometry. I
- t ,
[. 31 I
; 1
,i
-l l
Section 3 STATISTICAL ANALYSIS OF OlL HOLE DEPTH DATA -[ INTRODUCTION h A statisticalanalysis was perf ormed on available oil hole depth measurement data. These data - 3 1 . were compiled by HL&P from connecting rod inspections performed at STP and by inspections performed by other utilities. Both spare rods and rods from actual service were used in' providing the sample data. The purpose of this analysis is to establish the upper population tolerance limit on the exter.t of over drilling so that a conservative hole depth (i.e., minimum'. > ,, remaining ligament) can be selected for stress analysis. A description of the statisticalmethod l and criteria for determining a suitable worst case over-drilled ' depth is provided later in this
-i section.
MEASUREMENT DATA i A total of 35 KSV-4 2A connecting rods were used in establishing the variability in hole depth:
.}
22 service rods,12 spare rods, and one rod which was rejected for reasons other than oil hole - problems. The data were compiled from information obtained from STP and seven other plants .
. (f). The information included both physical measurement data as well as information derived from review of plant records. The measurementdata were first reviewed for consistencywhich identified the raw depth data from three plants did not include the' depth of the drill'tip past the shoulder of the hole (2). These data were adjusted to increase the measured hole depth by the extended distance of a truncated conical tip (Figure 2-4). 'The nominal hole tip geometry discussed previously in Section 2 was used in the adjustment.
The over drilled hole depths,in terms of the remaining thickness,are shown as a histogram in Figure 31. The thickness, t,, is the remaining thickness between the hole tip and the -
~
B 4:mAPT8 chi B 32 pyggym - 1 I I ll 10 i i i i i i i . . , i i i i l e F-Z LJ 8 - l 1 La m D I M
<I LJ 6 -
1 l L 4 - O
! m t.a cri 2 - -I Z D
Z 0 3/4 11/16 5/8 9/16 1/2 7/16 3/8 5/16 1/4 3/16 1/8 1/16 0 REMAINING HDLE TIP THICKNESS, t c [ INCHES] Figure 3 Histogram of Remaining Thickness at Hole Tip.
f 1 I
, -33
- articulating rod surface along the hole centerline. The results shown in Figure 31 reflect the adjustment for hole tip discussed above.
The overall thickness across the section between crankpin and rod pin surfaces along the 12 i radial line is approximately 1.58 inches. Based on this overall dimension, the relationship between d,, and t, for the normal section geometry used by HL&P in correlating their measurement data is: P dn + t, - 1.58 inches (3 1)
- i Equation 31 was used to correlate hole depth measurements with the remaining thicknesses-given in Figure 31.
The histogram shows a decrease in the number of observations as the remaining thickness approaches zero. This trend is to be expected since a correctly drilled hole would have a remaining thickness of about 0.72 inch and any value of t, smaller than this would naturally have a decreasing likelihood of occurrence. Therefore, a value of t, = 0.72 (Figure 3-1) l corresponds to a correctly drilled hole which terminates at the base of the transverse hole. A value of t, = 0 implies a hole which has penetrated the surface. Actually, when t,is less than 1 l -1/16 inch, surface penetration at the minimum ligament will occur since t, -0. This condition will be detected by shop inspection. Three connecting rods with surface penetraticg holes are shown in Figure 31; one rod being the failure rod at STP 2, second the being a spare rod at the same plant, and a third red which is currently in operation at another utility (H). It should be noted that unlike the failure rod, the other two rods were completely drilled through and the exit hole properly radiused to remove I rough / sharp corners. The minimum value of t, for a rod with a partially drilled (nonpenetrating) , oil hole is approximately 1/8 inch for a spare rod at STP. This hole configuration corresponds I to a minimum ligament, t,, of approximately 1/16 inch. l G l
s i. 34 ANALYSIS METHOD For convenience in performing the statistical analysis, a random variable R is defined to ! represent the remaining thicknesst, at the drill tip. Similarly, through the relationship of Eq. 31 , a random variable P to represent hole depth do could also be used. A standard nonparametnc technique of order statisticsis employed to compute upper and lower confidencelimits of tie cumulative distribution, F(R) of the random variable R, the remaining thickness of material at the drill tip. The technique, being nonparametric, requires no assumed probability distribution model to compute limits and plot them as discrete points although three parameter Weibull distribution is used to fit the nonparametric data points. < r q .Best (Point) Estimates of F(R) Following the recommended graphical procedures of Gumbel (2), the mean rank is used to' estimato the plotting position (R,F(R)) in a cumulative failure probability plot. . This mean rank is given by: h F(R) - l(R)/(N + 1) (3 2) l where N is the sample size and iis the order number of the value of R. That is,i = 1 is used for the lowest value of R, i = 2 is for the next largest, etc. In other words, the data are ordered by the procedure, so that R,i R,1. . 1 R,. In addition, the procedure employed in the data analysis reasonably and accurately accounts for data at the extreme values of the variable R. The procedure takes advantage of a standard 4 method to handle what statisticians call suspended data. For example,in the case where no . penetration beyond the transverse oil hole occurs, a value of R of 0.72 is usually recorded. These values were all entered as R > 0.72 to avoid the physically unreal and mathematically confounding effect of a large number of data points entered as exactly equal to 0.72. W _ . - - - - - _ - _ - - _ - - _ . _ . _ _ _. ,_ _ ,r-
-- - -, , m A
'I ?
I j- 30 i Similarly, for the situation in which the drill has penetrated the articulating rod bearing inner [ surface, the data are treated as suspended to avoid an unrealistic bias with several points i coincident at O. Thus, the method accounts for data points in which (1) the hole drilling does not go beyond the center hole,(2) completely penetrates the back surface, or (3) stops at any point between these extremes. ; in addition, the procedure uses a rigorous nonparametric confidence bound estimation method to handle small sample sizes. This avoids the errors of asymptotically normal distribution confidence levels which should only be used for large samples. For suspended data samples, C the best-estimate equations for F(R) are: 1 F(R, . 3) - F(Ri ) + 1/(N, + 1); I - 0,n, (3 3) , t' where, l F(R,) denotes the plotting position of the ith of n, ordered data values for which R is known precisely (i.e., unsuspended values of R)
- F(R ) - 0 (3 4) and N, - Effective number of units with R>R, l
l N' N, - N, + { (R, - R)/(R i. 3 - R) j-1 I ~ I-L e-,,-r er
k
-t 36- i
-I where, , N, = Number of units for which R is known to be > R,,, j N =. Number of units for which R is known to be > R,, where R,1 R,& R,., ! Use of the above algorithm is equivalent to assuming a piecewiselinear cumulative probability function for observed values of R. Confidence Bounds F,(R) of F(R)- l For complete samplesin which the value of R of one unit is independent of all other values of , R, the exact confidence bounds for the ith order statistic in N are given by the' cumulative' ; binomial distribution. Figure 3 2 reproduced from Whittaker and Besuner (1Q) illustrates the I relationship between F for the order statistics and the parent distribution (using N rather than - 8 i to denote the ith value of R). The specific equation used is given below: i l-1 Ni (3 4) y {k-O k!(N - k)1 p,k (3 , p,)w.k I I where y is the specified confidence level and F,, defined as F, - F,(R,, i, N)- l is the desired confidence bound estimate of cumulative R probability. This means that y is the probability that the true cumulative value F(R) lies in the interval between 0 and F,. For all but , j the simplest situations, the above equation must be solved implicitly through an iterative numerical scheme. t For the case of suspended data, the previous set of equations are used with N, used to denote . the effective size of the sample rather than the complete sample value N. The parameter N, t is completed from the relationship N - (1/F(RJ) - 1 (3-5) m ; '~. j
, , . - - - . . . ,. ~ . . . . . . . ,
t - J I .' I . 3-7 M APTECH;.. newe mau * - tteso 4.PTECM t
~
i t
?
I lI l l ..
+
, i M(tlABILifY OF A RANDOMLY CHOSE *i F LEE T Mf M8E A. R
- 1 + F r, z, Ocoa99 - 0 'J999 0 "90 0 "9 0M sw 10negismi _ _ _'_ _ i C : E _.
f,c " 0 2C NM,N .* ;gi og R ll' Al - O **0 -l . Z s 1/10 ; i M* t. Ne t.000 O'so , ~ g Me?,hel00 .o 0M M e ?, N.1,000 . 3 o
.-.M.1,N.10
.p
) e 0 02 ' M
- l. N
- 100-
- 4 3
d1 / -
~ M
- 1. N e 4d . 0 99 - C ~
2 i:100
/
p.J l [ - Me!,Ne20
. 9 7 _ Aq m. ). -
l 8 r'
.}.,,,,3/ * ' "' / -
. u. i. n. i ,,, 's o .
- M 3. N 30 .
g O g - \ _ Me?, N.4 . t ! l E 1/10' i t i i i/ t i e i , E o 1/106 t/10 6 1/10.000 1/1,000 1/100 1/10 f 1.0 $ FAtLURL PROSASILITY OF A RANDOMLY CHOSEN FLitT MLMcLR,F d a ( 1
-4 I
b 5 F t Figure 3 2 - Distribution of Order Statistics. L. L 1 . . . - _ __ . . _ . . . . . . .
lfJ ( 38
. for each [R,, F(R,)] point plotted.
This procedure accounts for the fact that the fewer the values of R, the less the accuracy in making estimates of R. In general, N, is not an integer. A linear interpolation is used to estimate the confidence bounds, F, for noninteger values. The specific equation used is given by: F,(R,, N.) - F,(R,, NB) + (N - NB)(F,(R, NA) - F,(R,,NB)) (3-6) l where N, lies in the closed interval between the two integers NB and NA = NB + 1. The above procedure, while complex in nature, was benchmarked against en independent I analysis method with excellent agreement between the two methods. I I r. INPUT DATA I The measurement data for t, were used to define the random sample of R. All data were transformed by the relationship R' - 50 + 100R (3 7) to simply shif t the data where R' is the new transformed variable. This transformation is for > convenience and has no impact on final results, i 3 For the sample population, one rod was the failure rod which did not go completely through the surface but indeed ruptured the surface enough to cause complete breakthrough at the hole centerline. This data point, therefore, was treated as unsuspended (precise) data with R = 0.
- For the spare HL&P rod with a completely drilled through hole, the data point was treated as suspended data with R <0.063, since at this value (t, = 0.063 inch) t,= 0 and rupture of the rod pin bore would be detected.
s ll t1 -39 In addition to the measurement data, advantage was taken of the 330 rods in nuclear service ll (11). It was assumed that 308 rods (i.e.,330 rods minus 22 service rods already in the data - base of Figure 31) have hole depths somewhere in the range of 0.86.5. d, < 1.58 inch. For these 308 rods, a value of R> 0.063 was assumed in the statistical analysis based up'on physical arguments presented earlier. This assumption was based on the assurance that all service rods have nonpenetrating holes Q2) except for one rod which is currently in operation (1). l; RESULTS The results of the statisticalanalysis are shown in Figure 3 3 along with confidence bounds i (tolerance limits) at various values of R. The bounds include an " upper" value of 95% and a
" lower" value of 5%. The probability of having an over drilled nonpenetrating oil hole with a l1 remaining thickness less than the worse observed condition (i.e., t, = 1/8 inch) is approximately 1.3% for the best estimate analysis. This probability is approximately 2.4%
- ) when 95% confidence is used. The mean value for t, at 95% confidence from Figure 3 3 is approximately 0.48 inch which indicates the expected value for the remaining thickness will
!i be much greater than the minimum measured value. Based on these results, an assumed value of t, of 1/8 inch at the hole centerline (or t, = 1/16 inch at th'e shoulder) will represent a very q pessimisticsituation f or an over drilled nonpenetrating hole given the very low likelihood of such a condition. I ~ I I
=- ='
TN M M=
- -- ~
I M ^- ~ M~ --~ '
-N 'i-
- rnir *-
- ssis '-
Ewassa sensua oc - t eene nw
,ggg // /
.99 UPPER CONF 10ENO: 89UND - 0 GrA00 g
.8 '
sxy
/$/
.5
+
+ -
+
jo /:
] 15
;;; X d '7
*i2 g ,g / J s/ '
o / /' / p
.05 f
f /, *
?
o
/c
$ +
E * / 8 .015
.012
,/ ,+ , / y /
, s. ,
A /
.005
/
o
+
.002 _
.0015 - '
.0012
.001 ^
4 5 6 7 8 g - 2 TRANSFORMED REMAINI'iG THICKNESS, R*'= 50 + 100 t c Figure 3 Cumulative Probability of Remaining Thickness R *. rom,,- n -y --
,y, y-n- -
g gm ,_, --_ _ _ _ ___.__ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ , , _ _
l 41 1 Section 4 STRESS ANALYSIS
. GENERAL DESCRIPTION The KSV 4 2A connecting rod was modelled in three dimensions by the finite element method.
Two finite element models were developed: a global model of the complete connecting rod, and a local (refined) model of the region containing the oil lubricating holes in the region-between the crankpin and the articulating rod bore surfaces. Nominal dimensions from the engineering drawings were used to define the geometry (3, A).: The results from the global' l model were applied as boundary conditions on the local model to obtain accurate stress values in the region of the oil holes. Additional details of the global and local models will follow later in this section. l A general purpose finite element analysis program called COSMOS /M (12) was ^used in the-development and solution of the stress models. - The COSMOS /M program is a PC based analysis system which runs on a PC/386 microcomputer. The finite element models were set up for linear elastic small strain analysis with a modulus cf elasticity of'30'x 10' psi and Poisson's ratio of 0.3. I i ROD LOADINGS The loads for assembly and operating conditions were orovided to APTECH by CES (14,15). i A brief discussion of the loading conditions is given below. .
; Assembly Loads The assembly of the connecting rod and bearing cap produces an assembly stress due to the interference fit between the crankpin bearing shells. .The assembly stress is modelled.by a f uniform radial pressure at the crankpin bore surf ace. An upper bound estimate for this pressure
r [ 42 is 960 psi as determined by CES to encompassthe high limit of the interference fit between the crankpin bearing shell and the rod. Ooeratino loads -
- Per the discussion of the engine operation provided by CES, the total operating' load en either connecting rod fallinto two categories: firing loads and inertia loads. Firing loads act in the -
downward direction and are transferred directly through the bearing to the crankpin in the - region directly_under the load application. These loads are compressive in nature and do not significantly contribute to the stressesin the region of the lubricating oil holes, inertia -loads act in either upward or downward directions, although downward inertia loads are reacted through the crankpin bearing in a similar manner as firing loads. . The maximum operating load occurs when the upward inertia loads are highest at near top dead center during the exhaust stroke. The upward inertia loads are reacted through the bearing shellin the lower.- half of the crankpin so that tangentialstressesare developedin the articulating rod bail area and 1 in the section between the crankpin and articulating rod bore surfaces where the oil. holes are located. I The maximum upward inertia load from either cylinder is 44.6 kips at a normal engine speed of 600 rpm. Since the two cylinders fire on alternate revolutions of the crankshaft,there is no j significant interaction between peak upward loads of the two cylinders. I GLOBAL ROD MODEL i The full model of the KSV 4 2A connecting rod is shown in Figure 4-1 Q6). The global model ; includes the bearing cap, the articulating rod bail, and a portion of the master piston rod. Also included in the model is the 7/8 inch central oil hole between the crankpin and articulating rod l bearing surfaces;however the details of the 3/8 inch oil supply hole,11/32 inchlateral oil hole.- ' and the 1/4-inch deep by 1-inch wide central groove at the crankpin surface are represented in the local model only. The global coordinate system used had the origin at the center of the i crankpin bore with the x axis oriented along the midline of the master rod and the z axis oriented axially with the crankpin. The master piston rod was truncated a short distance away ;
4 - APTSCHf!
, egsg oguv.cu c I
Articulating Rod Bail j-%s g y Master-Piston-f
\ / /"'% .
s < /
. \ , \' '. ' -
, , , p
))W" -~ '~
/. /j ,
I i
/r
-y NN N /3:
/
// a s
--- - ~
N R //g/
~ I. / ~ ~
L f y,7 - /s:n =NQ // p/ 7/8-Inch Eswx i s :_.
/,/ N '
'y' .
011 Hole % J{ I h ~ h . -lyi xx NN
,1
/- L l N- NN " '
% s -
jl' x N s\ sN , l -# Hh' kN
//
pM t l 'N Nw - gx N g 1 g ag 6% l %Q:-% f w eeer4eo car i 5 Figure 41 cieai Finite se-t of the xsv-4-m-ecting noe.
\-
1 i f- 44 from the crankpin bore and a plane'of symmetry along the midline (z = 0) of the rod was used to simplify the modelling of the complete rod.
. The global connecting rod model shown in Figure 4-1 contains 1.182 eight-noded isoparametric s,olid elements with a total number of 1908 nodes. Three loading conditions were applied to this mcdel: (1) assembly load case,(2) peak upward inertie load on the master bank rod, and (3) peak upward inertia load on the articulating rod bail. The assembly load was modelled by ,
a uniform pressure of 960 psi acting on the crankpin bore surface. The inertia load on the i master bank rod was introduced by applying nodal forces on the crankpin bore directly opposite i the master rod. The master rod end was pinned in the x direction, and the nodal forces were ' adjusted to give a reaction load at the master rod boundary of 22.3 kips (i.e., half the totalload , of 44.6 kips due 'to symmetry). In a similar manner, the inertia load (Case 3) acting on the - articulating rod bail was modelled through nodal forces acting on the crankpin bore opposite to the rod centerline. Spring elements were used to react the inertia load at the inside surface of the articulating rod bail. Again, the applied nodal forces were adjusted to give a total spring j g reaction force of 22.3 kips. i-LOCAL LIGAMENT MODEL The section between the articulating rod bore and the crankpin bore surf aces was modelled as - a local region for more refined stress definition. The local model is defined by'the section j between a 90 arc of the articulating rod bore and a 450 arc of the crankpin bore about the 7/8-inch central oil hole. The finite element model of this local region is shown in Figure 4 2 l (ll). This local model contains the geometry of the 7/8 inch central oil hole, the 3/8 inch central lubricating oil hole, and the 11/32 inch transverse oil hole. The outer 3/8 inch l lubricating through-hole, whichis 3 5/8 inchesaway from the central (midplane) location of the rod, was not modelled being far away from the area of interest. The local finite element model contains 4204 elements with 5348 nodes. The 3/8 inch . lubricating oil hole was assumed to be drilled past the transverse hole to a depth within 1/16 inch (i.e., t, = 1/16 inch) of the articulating bore surface at the minimum section (see again the - r description of the over drilled hole geometries in Section 2). A conservative estimate of hole i 1 tip geometry (0 = 56 and b = 70 mils in Figure 2-4) was used to define the depth and shape of the tip. This assumption givss a remaining ligament thickness at the centerline of the hole, L F l_____-__-____. - .
.q E: $5F 'PT80HI E 4-5
. og ime ,c -
=
)
'I ^ !I :
,6 Articulating j Rod Bore E Surface -
l! I -
/~
i L ,I 4 7/8-Inch ! -011~ Hole I a 5 4 3/8-Inch 011 Hole L s Figure 4 2 - Local Model of the Section Containing Lubricating Oil Holes. E F"~ . i
.e
() 46-t,, a value of 0.11 inch in the finite element model. This condition is assumed to be the worst I case hole geometry for connecting rods in service short of penetrating the surf ace and is slightly - more conservative than the worst case measured nonpenetrating hole. This hole configuration has less than a 2.4% probability of occurrence with 95% confidence in the population of.a service rod based upon the statisticalanalysis in Section 3. The displacement solution from the global model was used as bounda; y conditions to the local j model at the common boundaries for each of the load cases. A uniform pressure of 960 psi-was applied to the crankpin bore surface to account for the assemblyloading conditions acting on that boundary of the local model. I
SUMMARY
OF RESULTS Global Model The displaced shapes and principal stress contours of the global model for the assembly, master rod inertia, and articulating rod inertia load cases are shown in Figures _4 3 through 4 8, j respectively. The displacement plots show the exaggerated shapes of the connecting rod under I the imposed loadings. The stress contour plots give the distribution of maximum principal stress (a,) along the exterior surface of the rod. These plots are oriented to show the stress 1 gradient at the 7/8 inch central oil hole and the articulating rod bail rin0 l The maximum principal stressesand the von Mises stresses M) at the 7/8 inch oil hole, rod pin ' bore surface in the vicinity of the 3/8 inch oil hole, and the bail ring surface are summarized in Table 41. The assembly load case produces a 20.1 ksi stress at the intersection of the 7/8-inch hole with the rod pin bore. This is the highest stress condition predicte.1 by the global model. The action of the upward inertia load acting on the master rod causes a 9.0 ksi principal f stress at the same location. These stresses are considered local nominal stresses since they - ! do not contain the effect of the 3/8 inch and 11/32 inch oil holes which are absent in this . model. The results for the upward inertia load on the articulating rod produces relatively low stresses in the area of the lubricating oil holes as shown in Table 41. This is also evident in the stress contour plot shown in Figure 4 8. The action of the upward inertia load through the bail
y - -- g y- ; y. HMAPf808" 1 c m mw. Table 4-1
SUMMARY
OF GLOBAL MODEL STRESS RESULTS Calculated Stresses (ksi) Articulating Rod Master Rod Upward Upward Assembly Load inertia Load Inertia Load Location . Node g, f g, f 2. .f 7 7/8-inch Oil Hole at 369 20.12 15.22 8.99 8.28 -0.28 2.51' Rod Pin Bore Surface 7/8-Inch Oil Hole at 359 6.60 -6.47 8.42 7.91 -1.19 ' 4.75 - Crankpin Bore Surface
' Rod Pin Surface ~ 12' Off ~ 398 1.86 ~ 2.70 7.38 5.73 0.21 0.80 Centerline Bail Ring ' . 2033 .O.27 1.11 0.06 1.46 16.04 15.11 e
~
, , . m-m -
- - - - - - - - - - a ^^ -"" * - -"--"-" ' ' ^ - - -
~
's l v : i ~-: F.
l I *,<, H
\\ %
x x xxx wm em mm_
, r ,
/
\
///// 1 i i Ii l )\ f J)ll/N#
f\\\'\\\\ !
w, , l, ,l,lull \\\:\\\\\\\ i fII' n I lli I n
' \ g d gl g- a g c.-,
m . , - a 5 Rh i 4 ,
- lll)lll i NNxxxxx m sssss. /p/// .i
. " %' ////// Q d r_ -
I t t ; Dl f IIl' t a i4 ;i , olli r e u i i _______j
;; I 4 - ,;' 7 ;; -
4-9 - r ; .
- - 4 h ~., '
g ._ l
- w ll l \\\\
/\\\\\
\\NXN I x xxx
\ x\x l %K $
I Es -m 1 I - my sD N i ji I /s // ///// X/ ij f ljly i QQ l G T il { {lill [ i o,, muw '\\T\\\\\\\ ! i uuw zi,71 2 g\\'n jJ ia
; \\\\\\\
A x N N xxx P 4 fI llI 1 l ~ nxm / / /l/ j /// .g
- - xx "y///'/jhfffff ; f s
j l }
+ f '
b f l E u 4
h 4-w j i v = T :.: 1 l ls ! .. . ,l .. j [*~
~
l I
! \\%
sNw i l NN% VA i l ! i l l i I
--- l l 6 i
///// } i
' ' 'N ;
' ////////
' %\\\\
/////g I \\
l/// ///// 7 k\\\\ls [i>
]:
\\
4 Q7 f / 1; f 'i % l mtt
. 1
,,,n
++
~ jjyr)r.><sun &
E
't i i mx s \ \ \\ \\\
\\\\\\\
/////j
///N ]//// !!
i
//////
W/ .
'~ a :
- / / ms ?
'm ( j ( i ,
n T 1 I ll'( j s Il ll 4 [ ] ) ) f ))i [
?
E Y [. '
-i c >
,/
////// /////////////// ///////
vh sL
//
s / .'
~
N g s s s
\
s
+.
+
s
,'///,
/ / ,// ,/ .- '
h -3 3, >
- i. $
', s
\'
%u*
,e ,s' / , . - .' / p' 4
\
\
% O j
p
% T.
S o
') \ &
x g g
\ \.
\ i
\ $,
\ \
\ e
,s
\
x t k g x 1
\ k
3 . 1 . .
- I
,I B _B B _B R I B I l
. e i
I s a I i i I I i
! h i e a
i ! I
! i e
i I 2 X i r
4
; 4 14 !
transfers the load through the rim region around the circumference of the crankpin and very I little load is carried at the mid plane of the rod where the central oil holes are located. The cycle due to this load case can be neglected for the purpose of the fatigue endurance limit ! evaluation of the over drill hole geometry since it does not produce a significant alternating l stress. On the basis of these results,it is concluded that only the master piston inertia load I case contributes to the peak stress in the cycle which will occur once every two engine i revolutions. The stress in the bail ring is highest for the upward inertia load acting on the articulating rod as expected. This loading condition causes a 16.0 ksi maximum principal stress at the inside ! surface of the boil as shown in Figure 4 8. 1 Local Model
'[
The local model containing the lubricating oil holes was solved for two load cases: the i assemblyload only, and the comt,inedloadings of assembly plus master rod inertia load. These two cases represent the minimum and maximum stress levels in the load cycle and the stress levels at four locations are summarized in Table 4 2. The maximum principal stress for the combined assembly and master rod inertia loads are shown in Figures 4 9 and 4-10. The highest principal stress occurs at the intersection of the 3/8 inch oil hole with the transverse hole with a stress value of 23.1 ksi. The second highest l stress location is the 7/8 inch central oil hole at the intersection of the rod pin bore surf ace as r predicted by the global model with a value of 21.5 ksi. It should be noted that the coarseness i of the mesh in these areas and the representation of the surface intersection as a sharp corner will cause the calculated stressesin these locations to be less accurate. (Note: This location is not the area of the failure). I A closeup of the stress contours at the 3/8 inch hole tip is shown in Figures 411 and 412 for I the assembly plus inertia and assembly only load cases, respectively, At the tip of the hole subsurf ace to the rod pin bore, the maximum principal stress is 19.0 ksi for the combined load I. case. The assembly stress is 4.4 ksi at the same point. The stresses are lower at the rod pin bore surface where the stress anges between 11.3 ksi and 1.1 ksiin the cycle.
,h 4-15 mane <w
;y g g ,7 y m is ac l
u E I I Table 4 2
SUMMARY
OF COMBINED LOAD CASES FROM LOCAL MODEL 1 Calculated Stress (ksi) Assembly Plus Master Rod Assembly Load Inertia . Location Mada Ei 2. g Ei 4 7/8 Inch Oil Hole 1979 14.17 13.60 21.52 20.08 I Intersection of 3/8 and 1528 14.76 11.14 23.14 18.01 11/32 Inch Oil Holes I- Bottom of 3/8 inch Oil Hole 75 4.39 5.18 18.96 15.00 Rod Bore Surf ace at 3/8 inch 4 1,10 2.09 11.27 9.18 Hole i
p---~
+
1
- t. - %
{ g . -- ,, .,_ -- " ' __ 1-7 i 6 3 a 1 1 i s
.\
b i 5 k o 5 5
\
1 i 1 3
, 9.
c.
m M,6 24,5 u,-e.m.Aa.m.MJaus.-.aA.mme m,s a rm --- .p.-w -aa. s. --
~
l ! ^ l -
.e
.2 a
4-27 E , wm ' ' ' ' ' ~ - - .._. . , , [ E E l E 5 i i u. li a i E f,
.2 E !
21 8a , 8' _ j M,I m ! e 58 _g l
.If t-I .
kE l i Ithb E at . O E
- f i '
i? l e m .u t a
--, - - -- - . . ----. --- - - . . + , . - - - . - .a
c-.,, .,,-:---- ,
- w. ,;_;a-:u.;;= =;.:,---: ,n_;;; -~ m -
l' 1 g, + k ,, ' o, I k'. - t * lf - - 4 18 j
. 1 i
I I >
)
1 i I i i b I S ; O i
- a.A
,;L . . .+, v. sl r" ?" '
j r
- Maf!JO%g+e. ~. .* .; T;~ .Pf"-
y p Z,; ,
-opogg y f , .
' TQ': ;%. ,:n] w ,
e m i
.J
.9 I
L&A ic1
- .6 i
t"#:.tJ;&.#w
. . . ., g .". w g
i NE4 4.A s ., gE
-a
- r. s -.;..; yg ~
I E :
.E M I w .
O> ' a3 ' n I
. )
u<
. ),
P 7- \
- l
- i 5
I m
" )
l I i l l o: ,i
]'t g /
i! i J; y, :'-
.: : :, *n' -
4-19
~
If; .. __ _ .. .' . . . ~ 6 . l l :: -: :
- y. r k ,
I g e. j
- u. :
ll l 3 ,
., .h g . .;,. F lg ^
l y
~
g 5 .
. s 4 .
m
* ~
f; * '. '. E U e r
,si
.J M
f - y 3
- 3. .
i I 1 4
.. 't -
. gp b
.e o u.
' ~ '
4 l l . ___ _ _ _ _ , . _ _ _ _ _ _ , . _ - _ _ _ _ - - _ = - - - ' -
i' 'l 51 ! 1 [ Section 5 i FATIGUE EVALUATION l t A fatigue evaluation was performed to predict life behavior where both mean and cyclic stress components are present (.1.8). Following the vendor's design evaluation, a modified Goodman Diagram is used to determine the safety margins for infinite fatigue life for an assumed . I endurance limit of 35 ksi and a specified minimum tensile strength of 85 ksi(.L(). The actual tensile strength wss determined to be 105.1 ksi(2) which indicates the f atigue analysis, based i upon the above minimum properties, will contain additional conservatisms. A summary of the cyclic and mean stresses, as derived from the' results from Section 4, is given in Table 51 for five rod locations. Although the highest stress locations were not l associated with the minimum ligament at the over drilled hole tip, locations of the greatest stress range per load cycle are the bail ring and the bottom of the hole tip with a computed ! stress range of 16.0 ksi and 14.6 ksi, respectively. The stress range and mean stress at the f rod pin surface is lese than the hole tip region. i Following CES's definition for fatigue safety margins, the safety factor (SF) for preventing fatigue failure ls computed from
}
SF - S" (51) e,+(S/Sjo,,,,,n ' where S,is the endurance limit, S,is the specified minimum tensile strength and a, is the ! alternating stress (4o/2). The computed safety f actors are also given in Table 51. The safety I l factor for the hole tip is 2.89 which is a significant margin given the assumed worst case ! analysis for hole geometry. All other locations have safety factors greater than 2.9. Based on this analysis, the connecting rod is not life limited by the condition of over drilling the 3/8 inch hole to the thinnest plausible remaining ligament, r l
I ' 5-2 REiAP78CHf gpg.,.gi-cise j i ' l
- l ,
Table 51 i 4
SUMMARY
OF CYCLIC PRINCIPAL STRESS COMPONENTS
\
i Cyclic Stress . Components (ksi) Fatigue Location A2 P. Safetv Factor I . 1. 7/8 inch Oil Hole 7.35 17.85 3.17
- 2. Intersection of 3/8- 8.38 18.95 2.92 I! Inch and Transverse .
Oil Holes
- 3. 3/8 Inch Hole Tip 14.57 11.68 2.89 l c
, 4. Rod Bore Surface at 10.17 6.19 4.58 3/8 inch Hole 3
- 5. Bail Ring 16.04 8.29 3.06 l
r s u . c . i f r
i 61 I l B Section 6 DISCUSSION OF RESULTS i I Despite the worst case condition for minimum ligament of an assumed over drilled hole, the stresses in this region are relatively low. From a maximum stress viewpoint, other normal regions of the connecting rod, namely the 7/8 inch oil hole and the intersection of the 3/8 inch oil hole with the transverse oil hole are locations of higher stress. These results indicate that ' the condition of having oil holes driiled past the transverse hole does not pose an over stress g problem in the remaining ligament when compared to other geometric discontinuities typical of 5, a normal rod. i Of the five rod locations evalua?ed for fatigue, the 3/8 inch oil hole tip the intersection of the 3/8 inch hole with the transverae hole, and the ball ring inner bore surface are the areas of 4 highest fatigue usage. However,' e computed safety factors basedupon a Modified Goodman Diagram exceed 2.9 in these regions for the minimum design mechanical properties reported by CES. Therefore, some other unique attribute of the failure rod must exist, such as having a penetrating hole with rough surface edges to ceuse the failure, since the calculated cyclic stress by itself is insufficient to initiate a fatigue crack. l The stress evaluation was performed for a normal engine speed of 600 rpm; however the engine could achieve a maximum speed of 660 rpm during engine testing or load trip (overspeed) conditions. Such a condition would cause about 21% increase in cyclic stress which will be encompassedby present margins. Becauseof the infrequent nature of overspeed loads, the fatigue safety margins will not be significantly reduced by these events. l h
.. _____w
l.' II . 71 I Section 7 CONCLUSIONS j' The following major conclusions can be crawn from this eveiuauon: I 1. A detailed finite element stress analysis of a worst case over-drilled oil hole depth indicated that the remaining ligament at the tip of the hole is not the highest stress location in the KSV 4 2A connecting rod assembly.
- 2. The cyclic stress range at the tip of the hole is 14.6 ksi which occurs once every two engine revolutions.
- 3. The safety factor against fatigue failure originating at the minimum ligament is-approximately 2.9.
- 4. From a cyclic stress viewpoint, the over drilling of the 3/8 inch oil hole to various depths short of penetrating the rod pin bore surface does not create a finite life problem.
- 5. Extending this evaluation to the observed fatigue failure of the No. 4 rod assembly,it appears that the f ailure cannot be explained by the local stresses resulting from over-drilling.
1 I e l l l
l I R1 l REFERENCES 1
- 1. Draf t Justification for Continued Operation, " Thrown Rod Standby DieselGenerator 22",
Houston Lighting & Power Company (December 6,1989) (ECD 7).
'] 2. Steinberg, B. G., " investigation of Diesel Generator Engine Connecting Rod Failure, South Texas Project, Unit 2", Houston Lighting & Power Company Report MT 2558 (December 13,1989) (ECD 7).
- 3. Drawing KSV 4 2A, " Rod Power Piston", Cooper Bessemer Corporation, Revision 12 (June 21,1989) (ECD 1).
- 4. Drawing KSV 41 A1," Cap ConnectingRod", Cooper BessemerCorporation, Revision 7 (January 24,1984) (ECD 2).
E 5. Document No. 1166101, " Estimation of Oil Hole Tip Geometry", APTECH Project AES 8912116610 (December 21,1989) (ICD 2). A 6. Document No. 116610 2, " Review of Hole Depth Messurement Data". APTECH Project AES 8912116610 (January 27,1990) (ICD 3), p 7. Telecopy From R. Koogle (HL&P), " Connecting Rod inspection Data - Background Description" (January 11,1990) (ECD 6). I 8. Meeting of Cooper Bessemer Owner Group Steering Committee, Atlanta, Georgia (February 15,1990).
- 9. Gumbel, E. J., Statistics of Extremes, Columbia University Press, New York (1958).
- 10. Whittaker, l. C., and P. M. Besuner, "A Reliability Analysis Approach to Fatigue Life .
1 Variability of Aircraft Structures", Wright Patterson Air Force Base AFML TR 69 65 (April 1969).
- 11. Telecopy From R. Koogle (HL&P), " Total Number of Connecting Rods" (January 11, 1990)(ECD 5).
- 12. Telephone Conversation Between J. Horne (CES) and R. Cipolla (APTECH). '
- 13. Lashkari, M., " Stress Vibration, Buckling, Dynamics, and Heat Transfer Analysis,"
COSMOS /M Users Guide, Release 1.52A, Structural Analysis Resebrch & Analysis
.l Corporation (May 1989).
1 1 I R2
- 14. Horne, J. M., " Stress Analysis of the KSV 4 2A Master Connecting Rod". Cooper.
Bessemer Report AM 1852 C (February 4,1987). I 15. Engine Loads Information and Stress Summaries Obtained From J. Horne (CES) (December 6,1989). I 16. Document No.116610 3
- Global Finite Element Model." APTECH Project AES 8912116610 (March 5,1990) (ICD 4) (In Course of Preparation).
I l
- 17. Document No. 116610 5
- Local Stress Analysis of Oil Holes,* APTECH Project AES 8912116610 (March 9,1990) (ICD 6) (in Course of Preparation).
I 18. Juvinall, R. C., Enoineerino Considerations of Stress Strain. and Stranath, McGraw Hill (1967). I I 5 k il 1 I I
m 1 I I I I APPENDIX 4 I COOPER-BESSEMER-TINITE ELEMENT ANALYSIS OF T!fE KSV-4-2A MASTER CONNECTING ROD DATED TEBRUARY 26, 1990 I I I NL.90.079.02
.5 i
I !
)
i l l 1 I f i :
+
t i
- I t !
I i l ! APPENDIX 5 ! I BATTELLE LABORATORIES-FAILURE ANALYSIS OF THE KSV-4-2A MASTER CONNECTING ROD DATED FEBRUARY 27, 1990 f i i i I 1 ! l
~
L I { k 8 NL.90.079.02 \__-,_._._.._. . - . . . . _ _ _ . _ _ . . _ _ . . . . _ , . - . . . - - . .
F - Cooper-Besemer Reciprocating Products, Division l Cooper Industries, Inc. I I Applied Mechanics Report AM-1852-C-1A g Final Report Finite Element Analysis of the KSV-4 2A Master Connecting Rod i I I I Prepared by: John M. Horne, Manager Analytical and Compressor Engineering Grove City, Pennsylvania February 26, 1990 l
w History A I On November 28,1989 the No. 22 Diesel Generator (KSV-20-T, SN-7193) at Houston Lighting and Power, South Texas Project, experienced failure of the master I connecting rod on throw #4 Examination of the failed e m ponents indicated the rod failed by fatigue, originating at a point where an over drilled lubricating oil hole partially penetrated the surface of the articulated pin bore. The I hole had been tapped and plugged, but the reenterant corner had not been removed, nor had the edges of the hole been radiused to removed damaged material. Examination of quality control records revealed no other rods had been salvaged in this manner, but two others had been drilled completely through, with the edge of the hole radiused. Visual checks on spare rods and on others which had been temporarily removed I from various engines revealed some rods which had the center oil hole extending past the transurse nole, but not into the articulated pin bore. he deepest of these found to date is about 1/8 inch from the surface. A three dimensional E finite element analysis was done to determine the effect of these partially E over drilled holes on the stresses in the connecting rod. This report covers the results of that analysis. Loading The total loading on this master rod can be divided into three catagories; assembly, firing, and inertia. The assembly load is due to the interference fit of the crankpin bearing shells in the rod, producing a n minal radial pressure of 833 psi in the crankpin bore of the rod. Firing loads always act in a downward direction, and thus are transferred directly to the crankpin in the area closest to the load application. These loads produce relatively low [ empressive stresses in the lower portion of the master rod, in a direction ' normal to the bearing surface. Inertie loads, en the other hand, act in both upward and downward directions. I The peak downm 1 inertia loads occur near the bottom center position of the crank throw, and produce stresses similar to those from the firing load, but lower in magnitude. The peak upward inertia forces occur near the top center position of the crank. Weir reactions occur in the lower half of the crankpin I i bearing, so that tangential stresses are produced in the articulated pin bales and around the crankpin bore of the master rod as the loads are transferred to the bearing reaction area. Since the KSV is a "4 Stroke Cycle" engine, the peak upward inertia forces occur twice in the complete power cycle of each c/linder, at the end of the exhaust stroke and again at the end of the compression stroke. On the exhaust stroke the pressure on the piston is very low, so the net load is upward and very nearly equal to the inertia load. On the empression stroke, the force frm empression pressure at top center is equal to or greater than the inertia force, so the net load is zero or downward. , The two cylinders on any crank throw fire on alternate revolutions of the crankshaft, so there is no significant interaction between the peak upward loads of the two cylinders. The maximum upward inertia load fr m either cylinder will be 44.5 kips at 600 rpm. When the master rod is at the top center on the exhaust stroke, there is a downward load of 65 kips at the articulated pin.
-- a
Paga Two { February 26, 1990 L Procedure A three dimensional solid model of the lower end of the connecting rod was [ developed with the aid of the Conceptstation sof tware developed by Aries Technology, Inc. Only one-half of the rod was modeled, since it is symetrical about its axial center plane. This model is shown in Figure 1. Small details, such as the lubricating oil holes, could not be included in this model, because of problem size constraints. A finite element mesh was generated in this model using second order tetrahedral elements (Figure 2). Two load cases were applied to this model
- 1. Assembly load - A unifom radial pressure of 960 psi was applied to the crankpin bore. This value was chosen based on the high limit of the interference fit between the crankpin bearing shell and the rod.
- 2. Master rod upward inertia force plus the assembly load - The m:> del was held at the shank, and a unifom pressure was applied to a portion of the crankpin bore directly opposite the shank. This pressure was selected to produce a total force of 22.5 kip on the model (45 kip on the total rod).
The 960 psi assembly pressure was also applied to the entire crankpin bore, in both cases the model was restrained at the center cutting plane, in the direction perpendicular to the plane to account for symmetry. Both load cases were analysed using the linear static subset of the ANSYS finite element analysis program which is furnished with the Aries package. Previous work on a two dimensional model of this rod showed that the articulated rod upward inertia force case did not produce significant stresses in the area of the rod near the lubricating oil holes. Three small section models were constructed of the portion of the rod between the crankpin and articulated pin bores. These included the lubricating oil hole drilling in that area. One model had the central 3/8 inch hole drilled to within 5/16 inch of the articulated pin bore, the second was drilled to within 1/16 inch of the bore, and the third had the drill tip just breaking through the surface (Figure 3, 4, and 4A). The finite element mesh for the "1/16" model is shown in Figure 5, and the mesh for the " break-through" model is in Figure SA. Because of model size limitations, the outer one-third of the model width was removed. A check run was made on the full width model for one case, using a larger computer system, to insure this assumption did not compromise-the validity of the results. The deflection results from the full rod model at the two section cutting planes were used as input to the small section models. The assembly load pressure was also applied to the crankpin bore side of the models. An . additional load case was run for the "1/16" and " break-through" models, using an added pressure of 1625 psi on both the crankpin and articulated pin bores in conjunction with the master rod inertia deflections, to represent the added effect of the downward load at the articulated pin, l
Pag 2 Thrca _ Fcbruary 26, 1990 [ A Modified Goodman Diagram procedure was used to calculate the Cyclic Failure Factor (safety factor against fatigue failure) in the area near the over drilled oil hole. In all cases, both the maximum principal and the Von Mises i stresses were examined, and the larger of the two used for analysis. The average of the assembly load only and the master rod inertia plus assembly load I cases was used as the mean stress value,and the difference between the maximum and mean stresses was used as the stress amplitude. A further review of the " break through" model in figures 4A and SA in I comparison with the failed rod indicated this model did not truly represent the failure case, in that the actual drill point penetration was slightly less than this model, and the finite element mesh at the drill point was not fine enough I to represent the reenterant corner geometry. For this reason a further atudy was done, using a smaller section model with arbitrary loading, to characterize the stress concentration effects of the "1/16 inch ligament" case and the failed case. The undrilled, "1/16 inch ligament", and " break-through" models I. are shown in figures 29 through 32. An arbitrary pressure of 10,000 psi was applied to th. right-hand face of each of the models, and the left-hand face was held. The dimensions of all three nodels were identical, except for the I drilled hole. Results Figures 6, 7, and 8 show the stress results for the full rod model with the master rod inertia plus assembly load case. 'Ihe three nodes selected are on the articulated pin bore and rod center cutting plane close to the point where _I an over drilled oil hole would intersect the articulated pin bore. Thus these represent the "ncninal" stress in the rod at this point. These values range from 6.04 to 9 44 ksi. The maximum stress value in the full rod model is 15.14 I ksi which occurs in the articulated pin bore at the narrowest point, as shown in Figure 9 I Figure 10 and 11 shown a cernparison of stresses in the full width section model and the model with the outer 1/3 removed. Note that in the narrower model the maximum principal stress is 12% higher, while the Von Mises stress is 7% lower. All further work was based on the narrower model, j Figures 12 and 13 show the stress results on the section model with the small central oil hole drilled to within 5/16 inch of the surface. The maximum I stress occurs at the edge of the 7/8 inch central oil. hole, in the articulated pin bore, with Von Mises values of 18.4 ksi for the. assembly case and 23 9 ksi for the master rod inertia plus assernbly case. Figures 14 and 15 show an i enlarged view of the smaller oil holes in this model, with the stress contouring range reduced to provide better resolution in that area. The , maximum stresses occur at the intersection of the small central hole and the transverse hole, with values slightly less than 12.0 for both load cases. .'
I Pcg2 Four Tcbruary 26, 1990 3 Figures 16, 17, and 18 ksi show the results on the section model with the small 3 central oil hole drilled to within 1/16 inch of the surface. The maximum i values occur at the larger oil hole, and are similar to the previous case. The third run, with the articulated rod force added, shows slightly lower maximum i r values. Figures 19, 20, and 21 show an enlarged view of the smaller oil holes-in this model, with the reduced stress contouring, ranges. At the intersection of the radial and transverse oil holes the maximum principal stresses in this " model are 17 3 ksi for the assembly load case, and 20.0 kai for the master rod ; inertia plus assembly load. Addition of the articulated rod force reduces the stress at this point. The addition of the articulated rod load increases the maximum stress near the drill tip slightly- , The three nodes near the drill tip with the largest stress values, with ! articulated rod force included, were selected for failure analysis, with the , following results (See Figures 21A, 21B, and 21C for node locations): ; I. Maximum Principal Stresses - ksi. ; Node Assembly Maximum Mean Amplitude CFF 641 3 17 16 30 9.74 6.57 3 31 ; j 278 3 82 18.70 11.26 7 74 2.83 ! 277 3 85 17 30 10.58 6.73 3 16 II. Von Mises Stresses - ksi. ' Node Assembly Maximum Mean Amplitude CFF 641 3 26 16 90 10.08 6.82 3 19 278 3 28 17.00 10.14 6.86 3 17 L 277 3 72 17.20 10.46 6.76 3 17 The 18.70 ksi maximum stress value, when compared to_ the average nominal value { of 7.74 ksi frm the full rod model, indicates a stress concentration factor of 2.42. r- Figures 22, 23, and 24 show the results on the section model with the drill tip just breaking through the surface of the articulated pin bore. Figures 25, 26, L' i and 27 show the results for the most highly stressed node at the " break-through" point. The minimum cyclic failure factor at this node is 3 03 based on the maximum principal stress, with articulated rod force included. [ , Figure 28 shows the lowest CFF values plotted on a Modified Goodman diagram. { Figures 29, 30, and 31 show the results of the small section submodel study used to determine effective stress concentration factors. Figure 33 shows the nminal stress in the undrilled model at the node closest to the maximum stress point on the two drilled models, while figures 34 and 35 show those maximum ' stress nodes in the :: rilled models. In all cases the minimum principal stress-is the significant value, since the model is loaded in compression.
- a. _
Paga Fiva
, Tcbruary 26, 1990 I The stress concentration factors predicted by these subnodels are:
Submodel Principal Stress SCF
-psi Without hole I 1/16 inch ligament Hole broken through
-35200
-84700
-160000 2.41 4.55 Discussion of Results The stress results from the full rod model (Figures 6 through 9) are very I similar to those obtained from s two dimenaional finite element model of this same rod as shown in Report AM-1852-C, February 4, 1987. ne differences in maximum stress level between the full width section model in Figure 12 and the I,
i narrower model in Figure 13 are minor, and the stresses in the articulated pin bore of the narrow section model, Figure 13, away from the oil hole, are. similar to those in the full rod model for the same loading condition. All of these factors tend to validate the assumptions made in arriving at the final section models. The maximum stresses in all of the section models occur at the junction of the 7/8 inch central oil hole and the articulated pin bore. R ese stresses are I. somewhat conservative, since the model has a sharp corner at this point, while the actual part is radiused. The cyclic failure factor at this point is 2.71 based on maximum principal stresses. When the small central oil hole is 5/16 inch from the surface, the highest l stresses in that region are at the intersection of this hole with the transverse oil hole, rather than at the drill tip. The stresses at the hole intersection are quite moderate, with cyclic failure factors greater than 4.0. 1 With the hole tip 1/16 inch from the surface the stresses in the thin ligament J at the drill tip increase, but they are still less than the values at the 7/8 5 inch oil hole. The results with the articulated rod downward load added to the master rod inertia load are conservative, since the load was applied to the entire surface of the model, while the actual part has a slot in the i articulated pin bushing at the oil holes. W e 1/16 inch value was chosen for the analysis since this was the closest the drill point could credibly e m e to the surface without being noticed by the machinist and the inspector.- If the I drill point were less than 1/16 inch fr a the surface it would tend to create a
" blister" in the articulated pin bore.
The original results with the drill tip breaking through the surface (Figure 4A model) indicated lower stresses, and thus a higher cyclic failure factor, than the "1/16 ligament case", assuming the " break-through" is smooth with no torn , or cracked edges. This case does not truly represent the failed rod however, - since the mesh is not fine enough to model the reenterant corner. The case with the hole drilled completely through and radiused was not run, but the above results indicate the stresses would be lower than the "1/16 ligament case". This hole then becomes identical to the two holes near the outside of the rod, and experience over the entire design life of the KSV engine indicates the stresses at that point are acceptable. II II
Page Six February 26, 1990 m The stress concentration factor for the "1/16 ligament case", calculated L from the small section submodel results, is in good agreement with the data from the larger section model. The results from the small section submodel representing the " break-through" case indicate the stresses are 1.89 times [ those of the "1/16 ligament case". This would imply a cyclic failure factor of I 1 50 for the " break-through" case. Bis low CFF value, in cambination with the observed machining cracks at the break through point, would certainly explain the failure of the connecting rod at the South Texas Project. A finer finite element mesh might shown an even higher geometric stress concentration factor at this point. All of the stress results are based on normal engine operation at 600 rpm. On I rare occasions during engine testing, the speed could, approach the overspeed shutdown setting of 660 rpm. We engine load would be nearly zero under this condition, but inertia forces would increase in proportion to the square of the speed. For the "1/16" ligament model at node 278, this would tend to produce a I mean stress of 12.44 ksi and an amplitude of 9 37 kai, for a CFF of 2.42. This is conservative, since the downward load ecstponent from the articulated rod pin would actually be much lower under this condition. All of the calculated cyclic failure factors for this rod are well above the normal design valve of 2.0, except for the rod which failed at the South Texas Project. Conclusion E
' Any of the KSV-4-2A connecting rods which have the central 3/8 inch lubricating oil hole drilled past the transverse oil hole, but not through the surface of I the articelated pin bore, will have a minimum cyclic failure factor of 2.83 near the tip cf the hole, under normal operating conditions. The CFF will.be greater than 2.42 under the worst credible operating condition (overspeed trip setting). Rods which have the hole drilled through and radiused will have a I higher CFF. Thus all of these rods presently installed in KSV engines are safe for continued operation.
John H. Horne, Manager I Analytical and Compressor Engineering JMH/sas Attach. l
I
- I I
I E I . . ~ ... .-.. . _ . . . . . . . I - I I : I .! 1 KSV-4-2A Paster Connecting Rod Full Model - Figure 1 l
t
, ' ' ~ ~ ~ ~ ~ ~ - ' --- .__ _ _ , _ _
k
/
r i l ( l , L [ XSV-4-2A POWER PISTON ROD f l Finite Element Mesh Figure 2 t
I l
\
{ Q+- ., ; j ,,. .. - . j
~;. .
\...ba. . , , .., ,
s
, y g$ f . ^i. , '
- l. &
-t .- l y.p'%, + l 4
..x .-
i
' ; .,,~~- . -
+
~
j
},t
'~% ~-
s _[ . E 3
.g f '.
}
g
;. r- ' -
; )( '
l g , KSU-4-2A PISTON ROD COOPER _ HOLES SOLID SECTION Blind hole is 5/16 inch from bore l Figure 3 j i
i l \1 .I t ( T . . .
.ff .,; .
s <
'- ' o
* \y. ...*Q v .' $. g 4 . ~ . , + u ~ ~ . , - - ..
;)' .
i
%Mff$7k.;y'*a[q,,. .% . "~.-. .,'
I :[ .,',
.% ffL:n mo
.- m% u .'
o u
-e s.,
g.r, .. I
> e . .. . . , . , . . .. .
, ,. s l ~
~.
~
, ,_ ., . 1'\
- r* ; , .
- * }1 .
I
\
" 1 i, , -
t - s -
. . , . :.1.TA . *
+ ,,
, -.~ . .
; , \..
' ? _.s..- ' - .__
e.. .,
. , ' ..- .;g ..f '
l 1 #l i t
+
w l . {
/- .
v ( l . . _ ._. -- " . ~. F KSU-4-2A PISTON ROD ROD SECTION_1/16 SOLID SECTION m Blind hole is within 1/16 in. of bore m-m Figure 4 ]'
\
_j
f- ! :-i 5 I I - r I e. , ag u , _ _ . , .. s
; j s y.-
.t .. ,. . \. ,
..c. . . . . .. . , , . -
, . . .:y '
c_ . 6 y ~. .n a, # ': <
' :)
4 U , y . I: * 'I
$. 4 N
; C+ g_ f
,y)g:y*
.#* Q
, ; .-1 , e .
.g
' 7 21.4 a ;g ,.. w s^
-e
,4
. p't
,e 3
. 9%,
. 4, c, l .
y: 4 q*y . \ , .
- <, t 1 .y p .
, s ,.
j ' f), jd%' sap . - s .,
- e. d.pw 3
.s
. 7" . . . - t.
q *
. . y l
. ' . f.'c)
. 'T ,; '
t
,,j i? h;
\ c,y . ,
+
[
- ~ >
i ti y. : . ,
- : un - ,
r r -
.?
' .i.~ r 1 > . ., . - _\,m s =
3., q 1 .c "1 '. . / , - .: 4 .' - w :- - - . n. s a- if
.c-s >. .
- . cr, ;,. ,
.. yy.* .;?,
s+, 3 '-7.
..L...
I , 1.N. k
. gg .
%;_ _ %__ .f 7p XSU-4-2A POWER PISTON ROD Blind hole broken into bore -- acomed in l
i l I Figure 4A I
I I I 1 l 1 l J
.. . . . - - - - - . - . ~ - - - - - - - - - - ~~
4 k i I - lI I KSU-4-2A POWER PISTON ROD Bltnd hole is within 1/16 of bore I u, Finite Element Mesh Figure 5 I I
I k I I '
~ _ - - _ __ ._. ~ . . _ _..
I I I I
diy g*, ~ : i.i; 't i.f.d v r .. t 4*" . > , c >y j rs4 j *I . ,\ .,
L4 P.
, ,. e g .- s $ 4 ..., t 5, 2,. ,s< ..s . .. -' t r3 , . .:
1JM e<*. m .c.( pi tr. 1 p.c. Ja 4r 4 s t 4, t. vu i e. ?' :t. *. 1 il . r,..... &-
- MSU-4-2A POWER PISTON ROD Blind hole broken thru into bore Finite Element Mesh i
Figure SA '
l I I l I 1 I I l OPEN I SETUP STRESSES RT NDDE 557 l DISPLRY I Rt.UULT_INFU l 55 FEM _ INFO MRNREE l 0 .00 6 E] RIBE ~ [IT n I Y2 SHERR =
-2.61 L pf/inas T?
MPE"'I ' 15 - IMUM SHE AR -- = - 3.15e- 83 '.
t ' l 7
_\ , li 1 N
/N N
I I Prompt: 5 elect a node . " - ~ " ' ' - - ' ' " QUTUNbOlBRCKUPlHARDCDPYlGRIDIRXISIUIEWPORTISETUPICRLC IMERSURElINFO IWINDOW KSU-4-2R POWER PISTON ROD FULL ROD - NO SMALL HOLES Articulated Pin Bore j Master Rod Inertia Force Figure 6
\
I I i I I 5 I [ OPEN l PICK l I SETUP STRESSES RT NODE B33 l DISPLRY I NE5 ULT _InrU ! 1 -7.94R Ygn_ 55.31E INFO I FMNRE E
-8.99 6 NH3 l
1 l UPEN ! SETUP t )ISPLRl' I Rt.5 ULT _INFU 1 ggn_ INFO RRNME : 1,,_ 13 5 - M REE M NORMAL = 6.68e+83 'i 7 nas "i7 4.55e+82 I
. = nas
.. = 3.33e+82 I /i it I< /
..e FWI . .
~I = . l'eT8T"' in
=
M d~ nas MLNIliUM PRINCIPAL = 1.7!ie +82~ 7 N'MISES = 6 .7,te+03 I j MAXIMUM SHERR = W he+03 I
~"i
.i 1,
I '
- ] .
E N i I f
^ ' '
Prom.pl: 5 elect a node. ' Command: H QUIT! UNDO BACKUPIHARDCDPY(GRID lRXISIUIEWPORTISETUPlCRLC lMERSURElINFO l WINDOW KSU-4-2R POWER PISTON ROD FULL ROD - NO SMALL HOLES Articulated Pin Bore Master Rod Inertia Force Figure 7 l 1
1 I ; I [ OPEN I BETUF l 1I5FLRY iS4 I RE5 ULT _Inrv I TEri INFO DM t C- C E M I,_ j l { TWDBlK. . . . . . . . . . . . _ T~NUNE . YVIREEt 7 '55ii T- i7 n==
=
, - . . _ . , = . -5.21e+83.'.i.q a .8Be+ >2 i7 l T'i,9 M ]
I 'UDITTRTREI MISES
' =
9.10e+83 m f/~n 7n n y i y ' f-
N !
I g x,s i i I " Prompt :~5eTeil 'a node . Command: QUIT IUNDD !BRCKUP lHRRDCOPY IGRID lRXIS IUIEWFORTisETUF lCRLC IMERSURE IINFO IWINDOW MSU-4-2A POWER PISTON ROD FULL ROD - NO SMALL HOLES I Articulated Pin Bore Master Rod Inertia Force Figure 8 1
1 m i < . 7 .
' ~
~
[ l '
.Q:~
_ .- jf.Q.aW:'~;" jig'w if",, e l
~ " ' :P .
. {gy -
s g -l .
. ; .- , p. .. ~ *
] ; .c .p.< - -
- , : , .. . . - va t a ,. ,. ua : ..
I .. . . KSU-4-2R POWER PISTON ROD FULL ROD 200M i Articulated Pin Bom Master Rod Inertia Force ' Figure 9 l
l J i l l I l l 1 mimm44ttimp4MJMMImmi11'19mthhB .. 1Mittl1MK41-: lim EdMMminil4444Snalf'f4RMMBd3 M4RAMRTR1 naics . ~ SXMXSimR4mmI.3fEEEDM-I FEDLGS).mMtm )E11FAKEDM 64MIEl1FEEEDMi .
=.
s a ! j E333:141mK5m)RI1fAttDM I riai:141mKim*EI1FEEEDMl
- -a N:mi;ittammmmmmmesatmtmI1fEttDamm
.Eli13ti!I1;mMtH4IL1mKittt1MI1fEttrim
- 4 -
'.Eilt004SalH49MMEIM*EI1TEEEDM '
I Rill 3tii31;EMlH41 allee 3R4M1B1Xf1EEDM . E'blM;ltf46Et@Mt3fAEtDM ' i . t
> E;i136!IliE.difttm. 4Mt.MSMitfAEEDM. . .. .
i
,,7.. .
...- s -=
n
., : 1.
~.1 m .
S v wr !.'-
- s l
, u
~
M ..
. L, .
- , n .,
+
.^
' ]'-.
l m .
\
c ~ lllO [; ff:
- ,l ,
l ..,, . p.u '
. . . h- .
. -w
?
- 4[nnm.n. .
n, ~'. lF
, , ..
- w . +. .- .
?-
, , e og v. --n u SJ ..^
I ~ (
...(.".
- t. .
.: %We %3- - ~*-tP~. M t-
.n n .a u3 .
.? -
= ,
....# ....e t . ,,
.:. . . . :.1 a
K4:n stil:Idet t143121 IM 4444 attHiMn3ktDJ MMil3 dE4W1214n te ti) Qd4J49titelbindsam KSU-4-2A POWER PISTON ROD (ARIES RAN) 1 "5/16" Full Width Model Run at Aries Technology, Inc. Master Rod Inertia Force 1 Figure 10 l 1 l
I 1 i l , I I > I ~ mlJ Miaminumm141simmunin tJumasa mMemminM*issAaflifen.-;h nHM1544 nfM sueat MMmeMktM6minmut:4tMm t!$ : EElillZ}mWimIElifEIZDMi I
^
FEl.1,Gil m Dem*E11FJnae M
~.
^
t a 1.L G LI N E I M 4 p l 1 FE I T D M , RIR3 MAN 1lrkFTI$E11FETIDM .. ETE311ti EC33M1625 - R1M*WgEtID M;< - EhDGl:1;tJ3Clell;1mEffCIMITFAIIDM - * ' 1'E100sl3J3CI4ti;1pMfM4E11FEITDM I & 'T
- G CIG!IEE J3014 GLIMEf tTHEl3 FJt2DLM
- E8MIEill44mWimtElYfACIIIM. :
EhDGIIRE4MimR4M4ElifATIDM
~ ' ' ~ '
I
~
Y
?lD$'.:?-'-
~-
sy ,
\
'. , a p
-x
.y n . , + . .
I 4 N t,' k $-Q , .
~
~
4% -
.N,,.gn. y.y '%b .,,.Q;. '
mi . ,bm , an., ..o 4 ,. , m .- 2
'l. .
I flh e:" . ' . ' I 8 '$ ii 1
. 7 i. . , ,
+i .
...-- . - ,.-,wwMr .. .. .:
, . . , y E4:1 tat:1&lil : - 'I 1:libl4slWaldIslaMt'ileJunit4tijnl4i14elMMihMBCfietM4Niil!M XSU-4-2A POWER PISTON ROD (C-B)
"5/16" Model With Outer 1/3 Of Width Removeo Run at C-B Reciprocating Master Rod Inertia Force Figure 11
i i ~ - i i
"n ..e +
(s _ s 4 g b E I l . s . - 1
$2
-? . . .
l. t l KSU-4-2A F0WER PISTON ROD LOAD CASE 1 (assW I , l Assembly toaa Figure 12 1 1
I ' I 1 I l I ' I I I a m n '..(pru
- mw 4 4 i nes . . .. , . r sp. , .
rt tuu , o . o -.w r.g,m, a.- , r: *. 1. + , . < . *,f. *- a to; . , -i ;4 ( .. . . . . . . - 4 .. 1 kkJ[h4WA =+ b i% 5t' eM( .$ hht[ Sk. hh$ b4 vs is it tu , l it s'W lyr 6 i . Ix
- 4 x.,5 e t* ..w I
KSU-4-2A POWER PISTON ROD &c40 C AS 6 2 (mesrcg <co ,m e r, A) Master Rod Inertia Force Figure 13 i t A lll . . .
- 1
] ] m I , I I l
-M
' i' !
I N v
^.
9 l I ... ...... I HSU-4-2A POWER PISTON ROD LOAD CASE (assy) I Assembly Load Figure 14 I
I I i I . . . _ _ . _ _ _ _ _ _ . _ . - _. . . _ . . __ _. . l
- - ~ -
s e
. .x . ' ;. s _ . -
., 4 i .
x <
.g.,- , .
..>:. 4
.- 4 x
;. s "e;, * '
/' %. '
t
-t, , *
!.t v
. I t .
'f U 4
I 4 .
~ . .- .
n .. N
' . q. ' '
% .. , s
.q: ? . ,
I ;,; <
, y
\
, ,m. .
s .
,..t .,
s,: - > c. o .
. .....m .
us sr 6- ' ,. .
.p 4
- + m . n 3,in *
. ., , j ; c I :.-u. is a . c... .o. ..n 1 ai.4 .
..< . g -
;t l.
f um Fev ex ts.,W u kh tag
~
%,>. p .e . . . . i .. .e, , ,, ,
- - . . _ ~ . - - . . . . . . - - . . . . . . . + ;. - n
- d 4 lh ,,[
KSV-4-2A POWER PISTON ROD Load case 2 (master rod inertia) Master Rod Inertia Force Figure 15 1
- - . - - ~ .
l ~. I , l 1 g esiUn i p W I I I
,, , ,, ,g . , . .,
t as.a ,o , s su s a; . c:.n ,. , . - . . . = , . fl. }NOIN 4 I.t% 64et I tfi e 3Y IAW ' td '- %s
- I ; !%IK Ahv v, 1 7+ .I d iif;(1 f t ,, li-v ,
1a 14... i ! 5:4 t4 :w hr f bt r ) i si - T . 4: . e = v . . e- . I KSU-4-2A POWER PISTON ROD Blind hole is within 1/16 in, of bore Assembly Load i 1 Figure 16 i {
l I I
'm*
4 4 e a
._ j Y
a t.
"*-_ w,,
K I 1 , . .:. 1 'r
,.-.4 KSU-4-2A POWER PISTON ROD LOAD CASE 2 (master rod ineritia)
Blind hole is within 1/16 in. of bore f Master Rod Inertia Force Figure 17 l
/1
l .I II I I ;
~
I u
.T.
I 1 k.
.t
'x
- 5. .
1 4 1 ew. t < i . ,. . u v t,. n E' i4 [l b 1' Tk,, t , 't
.t
.y o ,
.;.t b.6 a ,' i '' ~' ' '
_ ,. a.. ;.
~
tyypo . .a. m, t . p -'- S 4.~ <t4. , n ..',.,...; ..g..,' '4{'j} c-I , res W q. y .u t#4 a , . 4
,o a u.a.,
a ;te g r.,..
,,r. ; ,, . ,. g.,._ .
KSU-4-2A POWER PISTON ROD LOAD CRSE 968+1625 Blind hole is within 1/16 in, of bore Master Rod Inertia Force Plus Articulated Rod Force Figure 18 I
. . . . . , . . . . . . . . . . . . . . . . . . . . . ~ - - " - -- - " " ' -
. ... . . . . . .. . 2M W
L I
<' a
+
' - ~ ' ' ' "
- a. '
~
4> l
=,Y -,.
4- k . - . . .I
, + ; .s e y
8 *'
~
W - 2 v "
?.
it 1 4 ,-
# 4 T '. .
1
'? s
^*
3
.. g .. . - . . . ,
~
6
; .. .'l- . .
s .- f *j 4
- ',) ' , s < ,k. f'- i s 1
- b- ,
< p[.. .# ' * '
I / [
. , 4
'W. '
N.; ' 4 a 8o8 .h # g g a l
*' f ..i ,.3,- @ .
,, -f
- i s. .
- b W Ji6- *. ..
+
-} , k
- I
.A*- * '. ,,f
f \ *' * '
} g ::*].$~'- ,
p.. . . ,
. w
- n e. . ,..: - 4 .
.k.5 Nj;h ' .a j .. - ~
, - : i' c. +. .-
, y . .
.~ ,
f
,l *'[ -
k' l J'Ag. e e * -' g g 4 e'
,, , ,;. t h .?' ' **' ~ ' * ' k '
.s ,
< _m _
9 / __ ,
~ -""5 ' ' " -
l k A-I '> '-'PI ' ' -
!'" - i ,, -$ . I l j 'v. 4 I KSU-4-2A POWER PISTON ROD Blind hole is within 1/16 in, of bore I asee -
Figure 19 f
1 i 1 l
.. ... ... . .. . . 4- -
. 7 j> .c.
,' f 4 , ,
t .
- s. ,
3~' ., . <. -, 4,. , -
$g f' ,
4 p w:
;e Vy Q,? r
w- *t . . . . .,
N {
.adW a g.- . .
h "'d:.A_ . ..
. Kh .: ..
s... p _. w r.. g., n a > r, , - - -
.,,. it w ,
x , , , , , . 1 n. .
-. .c , -
[$l1 6%+s 4P . L A# 1TM Na l4 i 4 i" * * * - i +-
. . . -" + & . ' v . i '
a t 'E * ' < * ' ' ' KSU-4-2A POWER PISTON ROD LOAD CASE 2 (master rod inertia) Blind hole is within 1/16 in. of bore Master Rod Inertia Force Figure 20 I I _ _ - - - - - - - - - - - - - - - - - - - - - J
~ r L
), '
__- . .t [ ,
.* y:
- s" . m , '
T-
'? '
3 .
'g
,- f 3, .
4 5 L . s
}
!s..r ~
i n
- L' ' '
hy~*L
- g '
A. , i
. i +
e . [
.' 4
(
.' ' h ,,, 4%
$h . +
. . 9%
L[ : vy M f. [g..
; ;,y .
.5 5 - L,m41. ... LN;4.;
l s g . , ,, , ,,.
~ m.
f f'.
; tw ti Wgu
. t -. 3 .s y i n ji a, w
.+
,t , g. , ,
.l.
ing A, :.
. . , . 3 . -
. .,. g,; : , , .
- pt. iu: +i**.. te a s. a- : . :.e
<n.. w < _ , . , + . . ,, .mn, . , - =- .
+2 ,-. .. .
[i f-i+-. KSU-4-2R POWER PISTON ROD LORD CASE 960+1625 [ Blind hole is within 1/16 in. of bore Master Rod inertia Plus - Articulated Rod Force { Figure 21 1 I
- I il l
I I : .jg e . ' SETUP Mill J4aFMnNfiefsiggjgigegggg;pgggggig BMelmMitJ4M453Bil461MT9
~
l s .'
' M M M M )hi} M ?R:PR afMitikS 3 M14111mm)MJ41FM M:ftfj94 Cr#E MidWE11NK1L$f1 M I
J
~ M N:Bi m itJ4M413E:1448 G m M W @ M $R u m e%E11 WII s '
5:EH1R$mmtm:Isl1FJEEDM ~ FEHmimmettitaMFAEEDM -
- i I
IWhil;TJmRt5ntal4FEEEDM .
%: ErmiM;t6km4Bl1(EEmm
'stadul-EEmtal1FEttnamm ~
i t{E@ttm&1MtBI1FAECHM ; I 1 '
.R I2ti;M;EAMM4ti3ME*ft11El1FJEEDMW'
- E;ll*)482:HM4ti3M4kITita11FEEDMd '"
W. r
^
l EMBENiB AMH4tJ3MRlrm*EITTEEEDM A l.: EMel! elm 4mRtmIE11FJEEDM ' :44-r ' EMIGNjEG41mtRimtR11FJEEDM
'.?.h . -
w - .
- s < .
..+
'e
_f* !( ,,
.i,l;o _
j , i. yl -
% e- , ..;a
.s
~
'*; $.?' 'i Mt [:) K ,
.ls. 9- ,5
~
;.y; tr 6,
u n ...a. w ~ ..a w , ,n n m .t a . - # . ,
, o *
$ "rin m , a s et h n . v . . u , ,. e . . .'c, .o
' .4 - -
s.
-..:~v-~~-.. .~
( l a ps'agilMin:IntailattinM4&atetHela:Haut#Wunhetismenselinn.181MentMat9 m91?am NODE 641 - STRESS RESULTS j l t Blind Hole is 1/16 inch from bore Master Rod Inertia Force Plus Articula ted Rod Force Figure 21A d.
l l I I '
' mandatmansumus44gummM J8hramm mmWinjuic pE ntsjutopmpimili m
' M2i 4Mg4dttt*W3S 1449hMMnM MKUlM BWmy fikl!5 .
I pm 4J;'AmEirma11FEtttim. ', B'aMJir3mR4m1B11fEEEDM
'95H4r3 m mrk m 1BlifEttD M - 7 , '
l - 53'atithmS4fn1ElifEttDM 4 --
. , i: :
E;' scil;Ed.HM44236 MimGilfEEEDM ..--
~
'4
;E34444MM:144736 34 "EtE11FEEEDM I
1 E=;H:H;4;Ei:lCl4tl3MEL 'Y1*Bl4fEEIDM V . ' .- E'bl:R;IM4mRD c - M13f i M
\} j $
ed ErJ3tD1;EtfGm!RRE4EllfE i;. W ?% . - I I . , 9.c
- ;~. , .
l a . x . .. . f -
?
OfdQf.g i j;. ' ,, ,
)#pgyyw}yg I
$)' k nM . g., y,w % %, ' i' -
,dV H e .3 t e s -3 1 p. ,p
. . . *^
l $ . I 1 !X -ab e 4
.- -- y ..
J 'M .3..p b
.- ' < . - ..c,..-
.g
<.* I 1
Es14tnuM41i13uual:133444tal11t4I33M i . W iuMAIOsletetrinjastabeitseminem NODE 278 - STRESS RESULTS Blind Hole is 1/16" from bore Master Rod Inertia Force Plus Articulated Rod Force Figure 21B
/I
I il l
- I I '
- j. 3 l .,E ml. - ,m m !m.. m. . m. _ _
\
JIstM m hjet1413n:1444tk m $ BM M 4WM M ?M6R$ {tN;11i i { . EEleklD1mmtMEEl1FEETIPM ' " I
- 3EIldElmE*mtElffAE1Ilm E4DLlf;1 6 &Mfal4filID M
. E3'B@t1;mRtTD*El4FEITDM .-
i E15dM;1; m R$ m tal1FEETIt M t
/ :
, ~; .
' Ef;13G91;.J;tel41;T:1MMrhmM14FEIHFM i -
i 530444SJ;tHI4IJilm:MWFDialTFAMM "
. .l H
' l 6
(EHCHZilMME14Mi1EMIStrtital1FEETIPMh*;.z~ : my ; . m; . 1
.L313GMEdMim41pm4BITFA i - M %,
I ,
, ~
J :
..% e s -
e a
. -a s c p
ng" '%u.+a-tp 5_ I l , I . A %$.@5 ' i /
~
-vgi '
o r p,. '
.'y!
- / o n . , * .
53,,
,. m i v.
~
'h 'nnt
#9- . ~ .
- t s .s -. .r t- 11, - . . ,
- i'
,;;' . -.. :'= - -
+ ,s E4tleditilld.II:I 1E0121:f;)?bl44te141bEli13CH8it4N43n@illinI4 ins.listkiih44B1:IsMI*ft:iihlag NODE 277 - STRESS RESULTS Blind Hold is 1/16" from bore Master Rod Inertia Force Plus Articulated Rod Force Figure 21C
,,,\\ , m , , . . ' ~
m ' u r~ l 1 l '( I I I I , t i F Ats t < e .n c: kche t M y et il ' 4 iii.Ii{ [{ }$4 laj A4f;u 1 ' ' ' k .$.h.T4 l $ (kk&1* f 1 I 4 % '* k I ra 0 ** s
.) ~#t .t g* 1 lts r, . '?
. OS:t
- O' 'II< I I# t . . * . I b- E- ** .
itw we cur
- eg4xuy tr S u a eq4 w 1+,c m, o m ne n tma ; .,,. , , e i . ,,
1 i KSU-4-2A POWER PISTON ROD Blind hole broken thru into bore l l Asse% h W d Figure 22 ____-____.--.M
i a l J l t 4 l [ [ [ {, " 4 l ii A6' + 1
- l li l A .i tit 4A 4 t - .. ! 4 cua , , .; y . ,. . . . -
(..+,.. .. t . , < Ii 1' i1 1 50,54 st ws it* ,
- f. 4 Irx. " is 4 -
3'- - * .. - l + s hIik kh*I t 1
' wit il is .hh ks Jd /A 14 a ' -[.Ev. 4+45 f *. a.-. [ *,.1(33 i t 4- , ,. - .+ h MSU-4-2A POWER PISTON ROD
( Blind hole broken thru into bore [ Master Rod Inertia Force Figure 23 l l l
W c
- - - - - - - - - -- I 1
[ i i t
; . ., . rm sit i,i ,, r
- s *tJ . , 4; .i t u 4 ring '
c v-m t- 2 , /, u , . "
' u e -e . < ', . ./ 6 su i > > ,i- ~ is t* o:a ' * < - - - -
I k 4hk r ay { fkB ti II , lh.4 )%. /- > 1 A_E+2.9 lvV 4 C u ;i
- 1eset I' s '
4 .:vs & h-KSU-4-2A POWER PISTON ROD { Blind hole broken thru in to bore
-( Master Rod Inertia Force Plus Articulated Rod Force Figure 24
_ . . . . . . _ . . _ . . . . _ _ _ ~ . _ _ _ _ _ _ . _ _ - _ _ ._. i t I h I N'W T UTEN I EETUF ' - E d h u M Lia'E M 1 6 4 0 1 6 r MRhiMLE%M*KTE% nMWS
- ~
l
~. ..-. ..
,.J X NDRMAL
~
1.55e+03 lbf /in.* M -
+
i drB'3:1. e ,4M1 1FEEZ D M
.. '. n.
4 ,
.' e + -: 3tTFEI222M i
7' , i
. . n o s .m c
, .e.. . - f
*e
- a . ,- .- i n e . M' . l ;
. 4.,...; .
i n. . m , -
- nA m .
' .. . . u e. . in,. . .
- s. - ' -
e
^
. . . ; o . gl1gEIEx}M , ~: < ,
MINIMUM PRINCIPA * -4.M,Ilsl311FEI211M ' +
/
Mm13tif1122}ama
,i.' -
....B i m *t s tYrJt2 n m 4 .
l . ,
** . e 4 ..
't.'... s, ; 4 4 '
5
~
g%,; ' - , e . l e%4.
' .j-:ti % j g
+
~
i
- 4
.h u , ': .
??_ . .
, ,p; ".
L. ,# M.;$. y,
%Ae ,
+
w ,, . . A_ . . . . t 't A, i e
~ i i ee ry n. . i - t
[ ' nr4 m.e . ; l a tu( 1. y.i n ( . , > i - .. 1 I { J$ l . ' romp : e oc a no.a .
! Command : i iQUITIUNDO!EACKUP!HARDCDPYIGRID1RX15lUIEWPORT15ETUPICRLC lREREURElINFO (WINDOW i KSV 4 2A POWER PISTON ROD Blind hole broken into bore l
Assembly Load Figure 25 1 l l e y ,-r--,- -m-- yf
u eq;gy. :' ..
- jqgyntap3;mangggggggggg
> pit 4MMin.14414 SAM 2144R Eb MN[M%ELiMnRS *N11 G 3
EEM.1; Lim.MMT11RlY fAI2DM j l'EI4XhimtG TtI-FRIT 7AIXDM
'i item &1 N Im1E11EE2D M
; agtinammmmmmmmeR3m1B11FJIMIIM ~
L ..Kf5tlNR4f111511f21XEM; ~ 7 3 1M5 tith m RII11FElTfAlHIl m tih3GlIKEJ3D LIME:imIpilfEIXD . EinDUIILBJ3 E;) 4*)81Gi 31(@I(1 LIMEMT11511fAEEDM i ilMNim FElifEIID : , 1 - Et ?ElH4mEimLElifAEGIM
. I i
E;s13fil4EtiLI.m. . R, IF.t.14. E. l'1FJtXD. M- . t
; y .
j .
~
L..... I " +
~
(s/ p \ . I,,>,
, 93
~ .. .
l g/' \ ,,
\
y ' t...h q.f~.y . 5.N:4- \, .
.....s,.,.. $ -
. .rw * .'.u>a .-
i ~. l p ..
- ~
., . , , , . n. 4 i
' ' ~
- 4. d t i >' +g th ** :
4, 1
- in
' a t ,h e co ! , .. g + v <>y y
- i ., . ,; r: .: . -
j t .e.. ix ., ' n. 3 %. } g; him k.. - .i .. .ei. .. [ .. . . . . l '.. I
. .. % 7 R4thstellliPEI4i14Xejat;ht;444Mti.H4fil3Ph8Jt4%4.6AM4i8L it)4elliM18hl491 llettaCli *
- KSU-4-2A POWER PISTON ROD Blind hole broken thru into bore 1
Master Rod Inertia Force Figure 26 I ! l
a l L 1 g M-E n. .~.n. .... - M AMimmisA4M45.lE141sRGmM E t m L K4 ) WH rR S AR n t ,.
.E:E !!shl2.l M R E M I E )1 N t3 D M .,
I B'E'l') 17.1631iMdB11DIZDM-1Elitamim1511FR3BM , ' nTat 4,1. Kim>E11st:tum s-
. watact.). 4RE11MEIDM '
I 3.. . MRtitmW4ft1FM11HfEDM ' EAElGISEJ.lll(414ilMMMRIERTNIZDM , EllG4115J lCl4thlM&MT11El'INEEIMI ~ 1 ElClGIERJ. LEI 4GL18EEarMMY11511FETIDM"' ' . gigggggggggygg , i EMElutEstiti mir11FEli st3D e N * ' -
;u . -v
' 7g;.y_;y _ .. : .
., 7,c : . v ,,
J. - } [fjf., ' Na[
; e .- -
.i A ', E
' '"?{ij.o .p'.i }, -
;~ I ? . r. . , ,
1 .r q, , j,> ". .- - _ [
; \{__ j. ;:.s 3, ,
.:) .c.
y
+-
.,,m. 3, y g b. .% ' r.
, : p ., ,.1';.' 1 1
.. m . . . a m .. .
- e. .> u i n. m .,
e a tt t.
- u1p** '): ;
.\
l - t m, , . . i y . -
. 2. . -
' rom- : -e oc a no o.
w tu u ,. 4 .. ) a ., 9 . j E*1tleAlliM*1334[214Milm:1.tb]qelgainM4I;1't6119l'(aj Misthnij alq;1sleJtpit1433 gupe1as a:iilelm KSU 4-2A POWER PISTON ROD Blind hole broken into boro Master Rod Inertia Force Plus Articulated Rod Force Figure 27
; , ! ;;4 'i ,
4 e ;< *' 4 e
- f s A 4 4 4 s y.srs .
/
4 I t 4 . o 9 o o o ! 0 *
/ ,x ,
9 N s . 5 xN 2 - C
)
l o se wN . x .- N - 2 . 0 x ~ (e s A C KsOlP - SBvD s _ x 2,ut r T - / - 2 A o . s
~v t 1f4 f/
? - .
N ,+ Do4 /rf A 2 .L - AL RAD
/
r Sur . _ fo rt A t R . f/ E r s ooo aP M 5sA ?! A b S
/ Aw R_J 2 L -
KR d / ) z0 - W r 0 v PP Tw a p e S tr . C cS f' x r/
/ 1 t
ci k pP
/
AE A t so N r oA 5 Cid .
& C cL o -
5 s L r - B N t 3 C w ]) T C t . K M cSc r c M S& i K
,G S 8~s- w / 4 tc .
S . s A 1 f R o T o 7 0 n v - 0 . D s E t
/ s E
; s 7 -
9 - 3 7 - . g o 1 o.
, ! ,; t!, -ii,.: --' . i t >- !
' 1 i I I i t 1 ; .I l
, ,; ?
C y
,. k
' ~
l I M-..-. ) ..d
- . ~ . . . -
i KSU 4-2R POWER PISTON ROD Without holes Small Section Submodel without holes 10000 psi pressure on right face I Figure 29
;I T" w -rye-
m i E I I i :,
~.
I I I B I w h /6o bore Small Section Submodel 1/16 inch ligament 10000 psi on right face Figure 30
e m t I I . l . [- .. l !. ! --
.[;f E
i l le I - i.
't 1 ~,'
li i 1 :.:
. . . _ . . . . . . . - - . . . . . _ . _ _ . - 1 I ** : .
I KSU-4-2A POWER PISTON ROD Blind hole broken thru into bore I Small Section Submodel I hole broken through 1 10000 psi pressure on right face I Fioure 31 I
ii -~ j I l
)
i t I l I F t I . . . . . . . - . . _ . - - . - . . . . . . . _ . . - . . . _ . . . . . . Tflf,
.tf i
I . rg& 4 %- %nn % %. 5"!,. 3
. .. l.
f f95 p- u I ?yEh*
, nr%ig,M \ r.
14i.W 3,,b1MQIc
.o y.,.2p;i:ve?'n
$ .' J, S.R
, .i. ,< f '
. N ,4 c,;
l2.' - WR Nl I 3 t
%@i . .. , ;
C.M'4.%.Ny:;,$e:.t*.
,..* 4 4 s ~ p% p'+.a+1,;
'.t w't., +
*;' "d * <: y i 5&k.E ;Q;}p ., f
. y;,,. u e;if x
.//-
4
- r. : ay.
e he, a
.M,:
.r,% ;
. l. .
- m. , ,.
R I i i KSV-4 2A POWER PISTON ROD ' Blind hole broken thru into bore i 1 Small Section Submodel - hole broken through Figure 32 I 1 e e
_ . - _ _ . . - _ _ _ _ . _ . - _ _ _ _ - - - - _ _ _ . - _ . . _ _ _ . _ _ _ - - . . - . . . . - - - - - - - . . . - - . - . . . - , ~ . . . - . . . . . ...~._ . __,. I l i I i < \ UrEN l EETilF 1 D 3rLRY I REBULT_INFC7
~~
r&R_ INFO VWNRF.E l i FICK 1 STRE55E:i RT NUDE 511 + 4 .5M ,g35 5 MI3PM N NORMR; * - 3.45e+94 r'/ n M
'" 'S:#~* .'
I . . T7 " n PRi M'AR E'r7 n 1 TTER 'AR
= ' J. -64.9 ^.[ y7 n MJ )D ' E PR' NCIF = -
I M':14II1UM Pl:INCI e - I a l ( l MRxl_pPLSHQR = 1.73 v'l
; r
?
[1 _4 Prompt: Select a node. Command: E ' QUIT (UNDOIBRCKUPIHRRDCDPYIGRIDIRMI5lUIEWrDRT15ETUFlCRLC lRER5UREiINFO KSV-4 2A POWER PISTON ROD Without holes Small Section Submodel without hole ! 10000 psi on right face j Figure 33
?
L
l I C lI I j . I . : I
~
1 .
- 4. . .
I R f_% ~ . a-Q tyg,'.l ' I I Prompl: Belect a node, l'omma nd : E QUIT JUNDOTBRCKUP jHRRDCDPV j GkID IRX15 lUTEWPORTl5ETUP lCRLC 'j'HER5URE ]INTO TuTwvow --- KSU-4 2R POWER PISTON ROD Blind hole within 1/16 of bore Small Section Submodel 1/16 inch ligament 10000 psi on right face figure 34 l
] ] l I I - y :: g .. NIMUM PRI
- I
. .' *KH h
h i Prompl telect a node. U UNb0lBRCKUPlHRRDCDPY{ GRID {RMI5lUIEWPORTj5EtUPjCRJ.ClMER5URJlINTO{ WIN KSU-4 2A POWER PISTON ROD Blind hole broken thru into bore Small Section Submodel hole broken through 10000 psi on right face Figure 35 I ll l ll l ll Il g i lilkil I E l t-. i
l 1 i l l l l i l k i l I l I :
?
p ) 1 I ! e APPENDIX 5 l l t ' l BATTELLE LABORATORIES-TAILURE /6NALYSIS OF THE KSV-4-2A MASTER CONNECTING ROD DATED TEBRUARY 27, 1990 , i i b l l i l i + n ,I NL 90.079.02 i
. _ . , . . . ~ . . . . . _ , _ . . . _ , _ _ _ _ . . . . . ~ _ _ . . _ . . _ . . _ . . _ . . . , _ . . _ . . . _ . . _ _ . . _ . _ . , . _ _ _ _ _ _ . _ _ . . . , . . . _ _ _ . . _ . . _ . _ . . , _ . . -
I 1 m)n n ! l a V a I ... MNAL REPORT I l PURCHASE ORDER No. 341B7239 ,I FAILURE ANALYSIS 1
'I OF THE i
l l KSV 4-2A MASTER : I CONNECTING ROD !I I n : COOPER. BESSEMER RECIPROCATING , COOPER INDUSTRIES ; 1g February 27,1990 , I I I I OBattelleh.!froc +u.... 1). I
w FINAL REPORT { PURCHASE ORDER NO. 3(187139 b on [ FAILURE ANALYSIS OF THE KSV-4 2A
- MASTER CONNECTING ROD I
to COOPER BESSEMER RECIPROCATING l COOPER INDUSTRIES i February 27, 1990 I by I l M.J. Rosenfeld, R.B. Francini, P and R.C. Rice l I BATTELLE 505 Kin Avenue : Coiombos. Oh o 43201-2693 l
, TABLE OF CONTENTS e Pace L SUKMARY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ii INTRODUCTION .............................. 1 FAILURE ANALYSIS ............................ 3 [ FRACT0 GRAPHIC ASSESSMENT . . . . . . . . . . . . . . . . . . . . . . 3 RAT 10NAllZAT10N OF THE FAILURE . . . . . . . . . . . . . . . . . . 4 Stress Concentration . . . . . . . . . . . . . . . . . . . . . - 4 ( Nominal Stresses ................... 6 Life Prediction Model . . . . . . . . . . . . . . . . . .... .... 6 MATERIAL CHARACTER 12AT10N . . . . . . . . . . . . . . . . . . . . . . . . 9-FATIGUE CRACK INITIATION BEHAVIOR ................. 10 Literature Laboratory Results Data . . . . . . . . . . . . . . . . . . . . . . . . 10
...................... 10 FATIGUE CRACK GROWTH BEHAV10R ................... 10 Literature Laboratory Results Data . . . . . . . . . . . . . . . . . . . . . . . . 10
...................... 11 REVIEW OF COOPER BESSEMER ANALYSES ................... 12
[ CONCLUSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 REC 0KMENDATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 REFERENCES
............................... 17 LIST OF FIGURES FIGURE 1. FAILURE SPECIMEN (lX) ..................... 19
[ FIGURE 2. DETAIL OF OVEP. LENGTH HOLE AND SURFACE PENETRATION (5X)20 .... FIGURE 3. DETAIL OF ORIGINS (20X) ................... 21 FIGURE 5. DETAIL OF TYPICAL RADIAL DEFECT AT SURFACE PENETRATION, VIEWED FROM OUTSIDE (a) AND INSIDE (b) DRILLED HOLE (1000X) . . . . . 23 FIGURE 6. INSIDE SURFACE OF DRILLED HOLE,1/4 INCH FROM TIP SHOWING SURFACE TEARS (50K) . . . . . . . . . . . . . 25 FIGURE 7. DRILLED HOLE BORE SURFACE DEFECTS (UNDER NICKEL COATING USED TO STABILIZE SECTION) (50X) . . . . . . . . . . . 26 FIGURE 8. SURFACE DEFECT SHOWN IN FIGURE 7b (250X AND 500X) .................. 27 FIGURE 9. THEORETICAL STRESS CONCENTRATION FACTORS . FOR PARTIAL HOLE INTERSECTION ....... ........ 28 FIGURE 10. CRACK GROWTH RATE DATA FOR FAILED ROD MATERI AL . . . . . . . 29 i
)
i
~ SUKu.ARY 1his report summarites the results of Battelle's investigation of a master cylinder connecting rod failure in a Cooper Bessemer peciprocating i Model KSV 20 stationary 4 cycle diesel engine. The failure occurred during l routine periodic testing after only approximately 600 total operating hours. The failed component was forged from 1050 steel, quenched and tempered. The failure occurred in the ligament between'the articulated rod bushing and the l' crankshaf t throw journal. Battelle's failure analysis of the connecting rod consisted of a fractographic examination and some basic fracture mechanics and fatigue g calculations to attempt to rationalize the cracking behavior. Some additional ? effort was also devoted to identifying the crack initiation and growth behavior of this material, through a search of the literature and limited laboratory testing. It is apparent that the failure originated at the location where the l articulated pin bore surface was penetrated with the drill tip. The drill tip penetration created a region of locally high stress concentration associated l with the re-entrant feathered edge. Radia1 crack-like defects on the order of 2 to 4 mils were found in this feathered edge (in orientations other than the failure plane). It is quite possible that they existed in the failure plane as well, serving as initial defects, however, this could not be established ! because the fracture origin area was obliterated by mechanical damage. However, we believe that the presence of the defects was not a necessary condition for' f ailure. In other words, the presence of an initial crack like ! defect would not have been necessary to cause the failure, because a fatigue I crack could be expected to develop in the local region on the basis of cyclic l strain levels alone. Once the crack had formed and grown outside the local region, there would have been sufficient crack driving force to propagate the failure, l if the hole had been drilled clean through the pin bore surface and the corners properly dressed, it is reasonable to assume that the f ailure J l would not have occurred. This assumption is made on the grounds that the stress concentration f actor for a through hole would be significantly lower ii
]
~
than that for the partial penetration hole ar.d that through holes exist on some other units that have survived much greater usage without failure. We are aware that there are a number of other rods still in service that have been overdrilled to varying depths, without piercing the pin bore E surface. It was not feasible within the scope of this study to validate the structural integrity of these configurations. From our brief examination of l similar cases studied in the literature it appears likely that remaining thicknesses, on the order of the hole diameter, and possibly less, represent i no more severe a configuration in tems of fatigue resistance than the many other acceptable hole configurations found in other units presently in I service, which have caused no service problems in general. However, this should be verified by detailed stress analysis, photoelastic analysis or some other means. If finite element analysis is used, it should be recognized that I the stress concentration will probably be underpredicted unless the mesh is j extremely refined. A fatigue analysis considering local stress or strain concentration factors should be performed in conjunction with any additienal j stress analysis. Cooper-Bessenier has perfomed extensive finite element analyses of the connecting rod with various hole penetration configurations. The results of these analyses were used as input to this investigation, and have been supported by similar results obtained independently by Cooper Bessemer's customer. Battelle reviewed the basic methods, assumptions, and procedures embodied in the analyses and found them to be appropriate and consistent with the standards of good analytical practices. The analyses were not revieaed in ' detail with regard to correct numerical implementation of those concepts, l however. l I
E. ^ FINAL REPORT j PURCHASE ORDER NO. 34187139 { on FAILURE ANALYSIS OF THE K5Y 4-2A MASTER CONNECTING ROD to I COOPER BESSEMER REC!PROCATING COOPER INDUSTRIES February 27, 1990 by M.J. Rosenfeld, R.B. Francini, and R.C. Rice I. 1 INTRODUCTION { A Cooper-Bessemer Reciprocating Model KSV 20 stationary 4-cycle diesel experienced a service failure in a master cylinder rod during routine periodic testing on November 28, 1989. The engine operated normally at 600 i RpH, producing 5000 kW of power. The unit was constructed about 10 years ago, but sat on site for seven years awaiting completion of plant construction. in the three years of service, the unit had experienced 200 to 300 starts, and > 600 to 700 total operating hours. The failed component was forged from 1050 steel, quenched and tempered, with specification minimum tensile yield and ultimate strengths of 64 and 85 ksi, respectively. (Actual ultimate strength for the material, as determined by Houston Light and Power (I) , was approximately 105_ksi.) The failure occurred in the ligament between the articulated rod bushing and the crankshaf t throw journal. This ligament is under maximum tensile stress when the master piston is at top dead center on the exhaust
- stroke. At this point inertial loads are the greatest, equaling approximately 44 kips. The ligament also experiences a compressive bearing load of about 65 kips during the articulated pisten power stroke, which occurs in phase with j the tensile inertial forces. !
a 2 7 On January 17, 1989 Hessrs. Richard C. Rice and Michael Rosenfeld of L Battelle met with Mr. John M. Horne of Cooper Bessemer Reciprocatilig Products . Division to review the failure and their ongoing stress analyses of the failed connecting rod. As a result of that meeting Battelle was tasked with the l perfomance of a failure analysis of the KSV-20 connecting rod. This report includes the results of, and conclusions from our failure investigation. I I I I l l l l l q . 1 I
e s ~ y e 3 , FAILURE ANALYS!5 L Battelle's failure analysis of the connecting rod consisted of a [ fractographic examination and some limited fracture mechanics and fatigue calculations to attempt to rationalize the cracking behavior. These two ( elements of the failure analysis are reviewed in the following paragraphs. FRACT0 GRAPHIC ASSESSMENT [ One half of the failure specimen was made available to Battelle for { examination by Mr. Siv Timaraju of Houston Lighting and Power Company. The specimen was subjected to unaided visual and microscopic inspection under the scanning electron microscope (SEM). The f ailure cccurred in the ligament between the articulated rod pin [ bore, 6.125 inches in diameter, and the crankshaft throw journal bore,12.25 inches in diameter. The failure plane, transverse to the ligament, contained { several intersecting oil passages (refer to Figure 1). The thickness of the ligament at the failure was about 1.5 inches, although this was not the ( thinnest part of the ligament. The width of the ligament was about 9 inches. The long lateral oil passage was located correctly according to the q design. Transverse passages near the outboard ends of the section were likewise correctly made. The short transverse passage at the section centerline was specified to be drilled from the crankshaft thron journal deep enough to intersect the lateral oil passage. However, it was drilled through and beyond the lateral passage, such that the drill tip broke the surface of the articulated rod pin bore. The passage intersected the pin bore surface at an oblique angle of approximately 60 degrees. This oblique partial penetration created a sharp feathered edge around the surface penetration in the failure plane. The hole at the pin bore surfan was somewhat elongated, approximately 0.10 inch in diameter in the plane of the failure. The overlength portion of the oil passage was tapped and plugged with a threaded ' insert, however, the surface penetration of the pin bore was not treated l further. Viewed at relatively low magnification, it is apparent that the failure originated at the feathered edge surrounding the pin bore surface
w 3 - 4 - penetration, on both sides of the hole (see figure 2 and figures 3a and 3b). However, subsequent mechanical damage obliterated the actual origin sites. The crack grew by fatigue on both sides of the hole simultaneously, first as I diametric corner cracks, then as diametric through cracks (through in the sense that it had reached the lateral oil passage), inspection of the hole at the pin bore revealed crack like defects 2 to 4 mils long radiating from it (Figures 4, 5a and $b). The cracks were in the feathered edge surrounding the hole, oriented radially, with overlapping edges. It seems probable that the cracks were due to tearing caused by the cutting edges of the drill bit piercing the pin bore surface. Further inspection of the hole below the pin bore surface revealed 1 that the entire hole surface is affected by tears caused by the drilling process, figure 6. A cross section of the overlength hole between the I threaded portion and the surface penetration was polished and etched. Several tears were observed in this section, as well as severely sheared and deformed l wear surfaces, figures 7a, 7b, 84, and 8b. The tears had cracks emanating from their bases, generally oriented tangentially to the hole. The radial l depth component of the tears was 3 to 8 mils. Although the othar drilled surfaces in the fracture plane were not inspected microscopically, they appeared to be of comparable surface quality. RAT 10NAllZAT10N OF THE FAILURE in an attempt to rationalize the f ailure we have examined the local I stress concentration factors and predicted nominal stresses. With this information, we have perfonned some basic fatigue crack growth and crack initiation analyses. l Stress Concentration Effect t l The local geometry associated with the partial drill bit penetration # of the pin bore surface, comprising a feathered re-entrant edge surrounding an elongated hole, represented a signif'icant stress concentrator. This l I configuration was considered comparable to an oblique or skew hole penetrating a plate at an acute angle. The theoretical stress concentration f actor for ;
l 5 such a structure has been developed (2). Evaluated for an edge angle of 30 denrees, which corresponds to the local geometry in the plane of the failure, the stress concentration factor, K ,t was estimated at about 14. The theory tends to overpredict Kg compared with experimental results, particularly for l skew angles in excess of 45 degrees, and ratios of plate thickness to hole diameter greater than 0.5. Photoelastic and strain gage tests indicate a K l t of around 6.5 to 7.0 for this geometry (3)(4)(5), Further support for high levels of stress concentration was found in g anothersource(6) , which used three dimensional photoelasticity to estimate the stress concentration factors associated with various configurations of intersecting and closely approaching coplanar holes, including ti,e case of a partial intersection, as would be caused by the tip of a standard drill bit penetrating an adjacent hole surface, figure 9. An important difference I- between the failure configuration and the photoelastic models is that the approached surface does not possess any inherent stress concentration, whereas the photoelastic models do. Thus, where these models suggest a Kt on the order of 13 for d/03 equal to 0.25, a Kt on the order of 10 would be more l appropriate for the case in question. The'results in Reference 4 also indicate that the K treduces to that of the approaching hole bore, Kt *3' l outside the re-entrant geometry. A 7/8-in diameter oil hole is located near the failure plane,12 degrees away around the crank throw bore or 1.283 inches. The distance from the edge of the 3/8-inch hole and the center of the 7/8-inch hole is then 1.096 inches, or 2.5 times the radius of the 7/8-inch hole. Theoretically, this results in an elevation of local stresses around the 3/8-inch hole by about 12 percent of the nominal stresses. In light of the above physical analogies, which are similar, but not identical, to the failed configuration, a theoretical Kt on the order of 8 would appear to be realistic for the site where the fatigue crack initiated. Such a severe stress concentration would promote fatigue cracking, even under conditions of relatively low nominal alternating stress. # f J
i , 6 - Nominal Stresses The maximum nominal stresses in the vicinity of the hole have been I calculated to be approximately 9 ksi during the maximum upward inertial l. l loadO). Compressive loads on the rod do not apply compressive stresses to the ligament area. The ligament is under a constant tension stress of about 1.8 ksi due to the bushing assembly lead. Thus, the nominal stress range, not l including any local effects of holes, is about 7.2 ksi, with a stress ratio (minimum divided by maximum stress), R, of about 0.2. Life Prediction Model It appears that the combination of one or more small initial crack-like defects, located within a geometry which itself was associated with a high stress concentration factor, was responsible for the relatively short l fatigue life experienced in service. Once the initial defect grew outside the local re-entrant area, a somewhat lower stress concentration, a Kg of about 3 l associated with the nominal hole, was operative on the crack. This situation was sufficient to allow the crack to grow relatively large, such that failure was inevitable. It would be desireable to perform rational analyses to show exactly how this occurred. One way would be to perform crack growth calculations, using appropriate linear elastic stress intensity solutions for the various geometries involved combined with appropriate da/dN - AK data for the I. material. In fact, this was done, albeit in a somewhat simplistic fashion. A stress intensity solution for diametrical corner cracks originating at a hole was used, with an additional stress factor of 2.67 (8.0/3.0), within the i re-entrant geometry. Nominal stresses were then assumed to control the crack l growth along, and outward from the 3/8 inch overlength hole, since the stress concentration associated with that hole is implicit in the stress intensity solution. The stress intensity solution used was based on a quarter elliptic # l crack solution developed by Neman and Raju(8) . The crack was assumed to grow with a circular shape until it reached the lateral 3/8-inch oil passage. At that point, a stress intensity solution for a diametrical pair of through-wall cracks was used. When the crack had 1
=
i
, 7
" grown" a further distance in excess of the original hole diameter, a stress intensity solution for a through crack in a finite width plate was used.
r The crack shape and growth was forced in increments, da, and the L alternating stress intensity, AK, estimated at each increment. The number of cycles, dN, required to achieve that increment in crack growth was then [, estimated using the average AK between steps. The number of cycles between l steps was calculated from ( log (da/dN) = 4.12 log (AK) - 11.01 , k which was obtained from crack growth tests performed on specimens machined from material adjacent to the failure. The total number of cycles was estimated using the Paris law, and using modifications of that by Walker and by forman, which account for the effects of positive stress ratios. [ These calculations resulted in predicted lives of 26 and 17 million cycles, for assumed stress ratios of 0.0 and 0.2, respectively. The estimate { for a stress ratio of 0.2 corresponds reasonably closely with the observed life of 12.5 million cycles. Furthennore, the calculations indicated that 90 ( percent of the life of the rod had been spent by the time the crack was only 0.5 inch long, on either side of the overlength hole. The life prediction calculations are very sensitive to the coefficients in the da/dN-AK relationship, the stress levels assumed to be operating on the initial crack-like defect and on the growing large crack, and on the calculated stress intensity. A variance of 20 percent in the stress or stress intensity results in a factor of two difference in total estimated cycles. Thus, it can be argued that a result that is within 50 or 100 percent of the *right" answer is sufficient in this case to rationalize the observed failure. It must be recognized, however, that linear elastic fracture mechanics is probably not valid for the combination of defect size and stress field that existed initially. There are a number of reasons for this. LEFM requires an assumption of constrained plasticity in the vicinity of the crack-tip, which occurs only if the plastic zone is small relative to the track size and the material thickness. The plastic zone size is estimated as rp =[(K/S)2)/2r y ,
, e d 8 " or around 0.5 mil. This is somewhat less than an order of magnitude smaller than the assumed initial defect size (one of the nomal validity criteria), - ( and is somewhat larger than the radius of curvature of the sharp edge around the hole at the pin bore. [ Secondly, LEFM makes certain assumptions about the material behavior at the crack front which are probably violated in a number ways in the area of origin. For one, the grain size was estimated to be around 1 mil, which is on
, the same size order as the initial defect, thus the initial crack may have been affected by metallurgical features in ways that a larger crack would not.
Also, the material at the surf ace of the drilled holes underwent severe shear deformations, resulting in an entirely dif ferent material structure froin the base material, and therefore, different mechanical properties in the vicinity of the crack. The drilling process also introduced a residual plastic and 9 elastic stress field, probably on the same size order as the crack, which was l not characterized. The service loading acting through the geometric stress ; concentration also imposed a local stress field which would have influenced j early crack growth. Finally, inspection of the radial defects around the drill penetration showed that they have a significant edge lap, which implies that initial crack growth occurred in a mixed mode. These factors can combine " to produce initial crack growth rates that are entirely different from what would be predicted using LEFM(9) . This can result in large differences between actual and predicted lives using LEFM, and suggests that the close correlation between the predicted life and actual life in this case may be " somewhat fortuitous. This also helps explain why the crack apparently grew in spite of the f act that the calculated AK was initially somewhat less than either the measured long-crack threshold AKth of 6.2 ksi/in, or the literature-reported short-crack threshold AKth of 4.4 ksi/in (for a stress ratio of 0.0). 7 In f act, the postulated initial defect probably was shorter than the transition crack size which defines the valid range for LEFM. Service life for such defects is better predicted using initiation models that essentially disregard the presence of a specific defect 00) To test this, the fatigue l . r notch factor Kf = 6.8 was calculated from the assumed theoretical concentration f atter ft = 8. The resulting cyclic strain and strain life was _
e 9 _ then estimated using cyclic strain hardening properties for a 1045 quenched u and tempered r. teel, with somewhat higher strength than the connecting rod material, yield and ultimate strengths reported as 92 and 105 ksi. [ respectivelyOI) . This material would be expected to have somewhat better fatigue resistance than the rod material. A strain life of 25 million cycles [ was predicted. If cyclic fatigue strength and ductility properties for the rod material had been available, this value would undoubtedly have been lower, probably by a factor between 2 to 10. { Thus, the finite life can reasonably be explained purely on the basis of a high local geometric stress concentration associated with the b partial penetration of the pin bore surface, located within a moderate stress concentration region associated with the 3/8-inch drilled hole. The presence of an initial crack-like defect would not have been necessary, because a ' f atigue crack could be expected to develop in the local region on the basis of ~[ cyclic strain levels alone. Once the crack had fortned and grown outside the local region (into where LEFM is valid), there was sufficient crack driving .[ force to propagate the failure, as was shown by the incremental crack growth calculations. [ KATERIAL CHARACTER 12AT10N As a part of the failure investigation on the KSV-4-2A master connecting rod, Battelle agreed to examine the literature and perform a limited number of crack initiation and crack growth tests to document the oehavior of the connecting rod material. A search of the literature for fatigue and fatigue crack growth data ( of mediuz carbon steel produced three sources of data. Copies of pertinent pages from these sources are contained in the Appendix to this report. [ The material used for the articulated connecting rod was forged AISI 1050 carbon steel. Mechanical testing reported by Houston Lighting and Power [ shons the tensile strength of the failed rod to be 105 ksi and itt yield strength to be 65 ksi. No exact match for this material was found during the 2 literature search and it was necessary to broaden the search to include ' { wrought carbon steels displaying similar material properties. This approach was justified by the overlap in the chemistry and properties of these steels.
w 10 FAT 1Glit CRACK INITIATION BEHAV10R Literature Data The Structural Alloys Handbook reports the endurance limit for wrought 1040 steel with a tensile strength of 100 ksi to be between $5 and 60 ksi. Thenotchedendurancelimit(apparentlyaKgofabout3)isreportedto l bebetween25and30ksi(12). The same source reports an endurance limit of ' 53 ksi for unnotched and 24 ksi for notched (Kg=2.45)specimensofhet l rolled 1045 steel with a tensile strength of 106 ksi. l Reference 11 provides cyclic stress. strain curves and strain' based l fatigue curves for AISI 1045 steel at a srinnell hardness of 225. Laboratory Results The fatigue crack initiation ttsts were performed in accordance with ASTM [466. The firt.t sample was notched with a Kg of 3.0. It was run with a g maximum stress of 20 ksi and a stress ratu of 0 0. The test was terminated
,a at 12.8 million cycles without a failure. The same sample was retested at the-same stress ratio and a maximum stress of 30 ksi. The retested specimen l failed after 650,000 stress cycles. Both of these values are in line with what was found in the literature for this m6terial, as described above.
l Two ur. notched fatigue samples were also tested. The first was testedatamaximumstressof60ksi(stressratioa0.0). It survived 357.260 cycles before failure. The second was tested at a maximum stress of 50ksi(stressratio=0.0). It did not fail after 11.5 million cycles. The g stress level was subsequently increased to $5 ksi (stress ratio a 0.0). and the specimen failed after 230,220 more cycles at the' elevated stress level. These results provide reassuring evidence that the connecting rod was not made ; from material with abnormally low crack initiation fatigue properties. I O l d IfI .. .
---_A
s J 11
- threshold stress intensity factors for this steel are shown in the table
[, below. I- Lon Short Cracks R Ratio AK (g Cracksksivin.) AK (ksivin.) 0.0 6.5 4.4 0.5 4.2 3.4 I The report also gives the following Paris Equation values for this steel: [ R = 0.0 da/dN = 2.3 X 10*II( AK)3.5 [ R = 0.5 da/dN = 1.9 X 10-10(3g)3.0, The values for the Paris equation that are typically used for design with [ ferrite-pearlite steels is given as(14): da/dN = 3.6 X 10-10(3g)3.,0, Laboratory Results ( Two fatigue crack growth tests were perfomed. Both tests were conducted at an R value of 0.1. The first test was conducted under constant [ amplitude loading conditions. The second test was performed using a stepped, decreasing load amplitude to determine the threshold AK for . crack propagation. Both tests were performed in accordance with ASTM E647. The resulting log. (- log curve (with both test results combined) is shown in Figure 10. The staga two growth region fits a Paris law equation of: da/dN = 9.77 X 10-12(3g)4.12 and the estimate of long-crack threshold AK is 6.2 ksi root inc.A. It should be mentioned that the crack growth-became uneven during the threshold test. The test was run under automatic control over the weekend. On the following Monday it was discovered that one of the pins used
12 to hold the sample had broken, possibly resulting in an uneven loading of the l sample. Attempts were being made to straighten out the crack and check the data obtained during the test. The reason that this data has been included is that 1) it does fit well with the data obtained during the crack growth test and 2) the threshold value is in line with the data obtained from the literature. The constraints of time and the lack of availability of more samples precludes a repeat of the test. REVIEW OF COOPER-BESSEMER ANALYSES l In order to better understand the cause of the failure, and the implications of that failure to other units in service with overlength holes, Cooper-Bessemer performed extensive finite element analyses of the connecting rod. The analyses assumed linear elastic material behavior and static load . conditions. The model geometry was generated using ARIES, a commercial CAD , system, while the analyses were performed using ANSYS, a reputable, public domain, general-purpose finite element code. A global analysis was performed of one-half of the complete connecting rod assembly. Only half of the rod geometry was modeled due to symet ry. The ligament containing the hole geometry of interest was then modeled separately in greater detail. The displacements and free-surface pressures from the global model were then applied as boundary and load conditions to the detailed models. The detailed models considered the cases of a blind hole with remaining edge distances of 5/16 and 1/16 inch, and the partial penetration geometry associated with the failure. Important geometric features that were included in the models were the 1-inch-wide oil groove in the crank pin bore, the 7/8-inch-diameter central oil hole, the intersecting 11/32-inch transverse oil passage, and the conical tip of the blind hole. Loads that were accounted for were the tensile inertial load from the master piston, the compressive load from the articulated rod, and the assembly load associated with the articulated rod pin bushing. .- Battelle reviewed some of the analysis output at CRoer-Bessemer's f acilities, and reviewed their stress analysis report (7) More importantly, . critical issues regarding boundary conditions, mesh size, load application, general modeling approach (global model followed by detailed models), and
~ 13 results were discussed with Cooper-Bessemer annlysis personnel. On these } bases, the analyses were judged to be appropriate and consistent with the standards of good analytical practices. 7 Battelle did not review the analyses in sufficient detail to verify " { correct numerical implementation of the cove concepts. In other words, the " models appeared to be correctly sized and proportioned, but nodal coordinates were not checked; likewise the load applications were discussed, but the actual element pressures input to the models were not specifically examined.. -l b However, the results appear credible, and are in essential agreement with those obtained independently by Cooper-Bessemer's customer. ! [ Cooper-Bessemer's models were used to obtain' local stresses at l various sites of geometric discontinuity. The local stresses were ' [ subsequently used to calculate " cyclic fatigue factors" ("CFFs," essentially factors of safety on fatigue life) on Goodman diagrams. The accuracy of calculated local stresses is highly dependent on mesh refinement relative to { notch accuity. The accuracy of the local stress calculations in turn affect the usefulness of estimated CFFs. ; [ The ratio of local stress to nominal stress outside the area of i notch influence (or the nominal stress that would exist'if the discontinuity i k did not) is the stress concentration factor (SCF), Cooper-Bessemer's models i indicated apparent SCFs of around 2.5 for the 7/8-inch diameter central. oil hole and 2.25 for the overlength 3/8-inch diameter hole. The theoretical _SCF i for a hole in a solid is 3.0 in uniaxial tension, and somewhat less in biaxial tension. The biaxial stress state could occur as a result of material constraint in a thick section like the connecting rod ligament. Thus, there { is a possibility that local stresses in the overbored hole geometries have been underestimated, by perhaps 20 to 33 percent. This would result in a degradation of CFFs, reported in Reference (7), from around 3 to around 2, at a stress ratio, R, equal to 0.2. Likewise, the effective SCF at the i intersection of the 3/8- and 11/32-inch holes was estimated by the finite : element analyses to be between 3.5 and 4.0, which is somewhat less than the ,- benchmark SCF of 5.2 obtained by photoelastic studies (5) The CFF at this location, with R equal to 0.85, would also reduce from around 3 to around 2. This is not necessarily a cause for alarm, since many of the geometries so
.. --,"44 -
4 4 14 represented, such as through holes and hole intersections,- are design details ~ that are known to be nomally reliable in this application. The SCF in effect at the tips of the overbored holes was estimated in Cooper Bessemer's analyses to be on the same order as the 3/8-inch hole I L shank. This is probably reasonable, based on the photoelastic studies in (5). Thus, it appears that the hole configurations which have been deployed in the b field, other than the failed case, have been adequately represented in Cooper-Bessemer's analyses. Accounting for local effects beyond the finite element ~ [ mesh resolution still indicates adequate fatigue margin. The finite element analysis of the failed configuration indicated a [ SCF of around 4.5 in effect for the reentrant geometry associated with the articulated rod pin bore surface penetration. As explained in the failure analysis portion of this report, the SCF was thought to be between 6 and 10, i or 1.3 to 2.2 times greater than the SCF estimated from the finite element analysis. The greater error is probably a result of greater notch accuity relative to the mesh. However, since Cooper-Bessemer is not attempting to qualify this specific geometry for field acceptance, this is not a concern. I 1 4 [ [ [ - [ l 4 { y 3
15 CONCLUSIONS. , The failure did initiate at the location where the articulated pia bore surface was penetrated with the drill tip. The drill tip penetration created a region of locally high stress concentration associated with the re-entrant feathered edge. Radial crack-like defects on the order of 2 to 4 mils were found in this feathered edge (in orientations other than the failure plane). It is quite possible that they existed in the failure plane as well, serving as initial defects, however, this could not be established because the fracture origin area was obliterated by mechanical damage. However, we believe that the presence of the defects was not a necessary condition for failure. . If the hole been drilled clean through the pin bore surface and the corners properly rounded, it is reasonable to assume that the failure would not have occurred. The stress concentration associated with a complete , through hole is significantly lower (Kt = 3), and through holes of the same or - longer diameter exist ir <imilar units without any fatigue problems. e l
16 ~ RECOMMENDATIONS L The practice of partially penetrating a free surface with a drill bit without subsequent treatment is one that should be. avoided. Similar i i configurations should be corrected by drilling the holes through completely and preparing the corners. Cooper-Bessemer has also indicated that a number of rods have been l l overdrilled to varying depths, without piercing the pin bore surface. The amount of material between the bottom of the holes and the free surface varies l from 1/16 inch to nearly 3/4 inch. It is possible.that the surface quality of the hole bores is similar to that which was observed in the 3/8-inch hole near g the failure, which contained numerous surf ace tears up to 8 mils deep -with j significant tangential orientation. l I, It was net feasible within the scope-of this study to validate the structural integrity of these configurations. However, photoelastic studies of closely approaching coplanar holes indicates that the stress concentration ! associated with remaining ligaments less than 1/10 of the diameter of the a approaching hole in thickness can be high. However, stress concentrations on l the order of the nominal hole stress concentration occur with remaining-ligament thicknesses equal to or greater than the hole diameter. It appears j that remaining thicknesses on the order of the hole diameter, and possibly I less, represent no more severe a configuration than the many other hole configurations found in other units presently in service, including coplanar intersections, which have caused no service problems in general. This should be verified by detailed stress analysis, photoelastic I i analysis, or some other means. If finite element analysis is used, it should be recognized that the stress concentration will probably be underpredicted I unless the mesh is extremely refined. This might be checked either by observing the change in stress with an increasingly fine mesh, or by comparing calculated stress concentration factors with known values from test or theory. ' A fatigue analysis considering local stress or strain concentration factors . L should then be performed. ' l l 1 m
~ 4 -: 17
.f REFERENCES (1)- Steinberg, B. G., " Investigation of Diesel Generator Engine
[~ Connecting Rod f ailure, South Texas Project Unit- 2", Materials n Technology Report, Report Number MT-2558, Houston Light and l Power,(December 13,1989) (2) Ellyin, F., and Sherbourne, A. N., "Effect of Skew Penetration on-Stress Concentration", Proceedings of the ASCE (December, 1968). I See also, Ellyin, F. , Lind, N. C.. -and Sherbourne, A. N. . -
" Elastic Stress Field in a Plate with a Skew Hole", Proceedings of the ASCE (February,1966).
, (3) Peterson, R. E., Stress Concentration Factors, (1974).
l (4) McKenzie, H. W., and White, D. J., " Stress Concentration Caused by a Oblique Round Hole in a flat Plate Under Uniaxial Tension", Journal of Strain Analysis, 3 (2), (1968). l- - (5) Johnson, L. R., and Leven, M. M., " Stress Concentration Factors at Intersecting and Closely Approaching Orthogonal Coplanar Holes", Exp. Mech., 17 (1), (1977). (6) Daniel, I. M., "Photoelastic Analysis of Stress Around Oblique Holes", Exp. Mech., 17(1),(1977). ! (7) Horne, J. M., " Finite Element Analysis of the KSV-4-2A Master Connecting Rod", Applied Mechanics Report AM-1852-C-1, Cooper < i Bessemer Reciprocating Products Division (January 31,1990). (8) Newman and Raju, from Gallaher, J. P. , et. al., USAF Damage 1 Tolerant Design Handbook, Air Force Wright Aeromautical Laboratories, AFWAL-TR-82-3073 (1984) . , (9) Leis, B. H. , Hopper, A. T. , Ahmad, J. , - Broek, D. , and :Kanninen, M. F., " Critical Review of the Fatigue Growth of Short Cracks", l Engng. Fract, Mech., 23 (5), (1986). See also Leis,-B. N., et i al., " Mechanics Aspects of Microcrack Growth in inconel 718-- I Implications for Engine Retirement for Cause Analysis", AFWAL- , TR-84-4041 (1985). l (10) Fatigue Design Handbook, Rice, R. C., ed., AE-10, 2nd Ed., Society of Automotive Engineers (1988). (11) Landgraf, R. W., Mitchell, M. R., and Lapointe, N. R., Fatigue ' l Properties of Structural Steels, Ford Motor Co., (Unpublished), June, 1972. (12) Structural Alloys Handbook, Mechanical Property Data Cer.ter, Battelle, 1989 Edition.
_ 18 - (13) Werner, G. K., and Stephens, R. l.. " Fatigue Crack Growth " Properties of 1045 Hot-Rolled Steel", from Multiaxial Fatigue, Aiialysis and Experiments, Leese and Socie, eds., AE-14, Society of Automotive Engineers, Inc. , pp. 149-156,(1989). (14) Rolfe, S. and' Barsom, J., Fracture and Fatigue Control in-Structures, Prentice-Hall, 1977. i i l 1 l o k, , 1
0
~
} . 4 i 4 ni g 1 i
+
.i.
i
& , L.
8 k TE) , i \
.U 'a 1
1 h b' I
)
( 5['h!I E t fL . i .. L :.. t 1 ( :
, A a
f a . ( ,.
\
q , . < p .* M '
}
\
20 [ go. ., I i ~ -
.A1
( - 1 -- .ng {u -
- ;A1
-rf
%'% a '.
[> . .';,,,..., - %,.l
.>s
.J.. . . v. , .
44./.'9'y ?-
='
y, : < . ~ _' .;. F L s [ i 'e:-, . r, ,g I . i .,1 r*(g , 3 . _ . . . . ... e f
.g s t g, FIGURE 2. DETAIL OF OVERLENGTH H0LE AND SURFACE PENETRATION (5X)
m m ,444g.ag-ye,.,. m 4. .-..- sm-- --ma-p .m. ww- -m. - > - - - - - - - - - - - - - . . - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -----------------------------Ti-r7! 21 g, .. . . .. . . _ . . . . . . . . - 1 I. . {@ %D } n p r'53;;,. , w *'"
$3).l @hl M n+f f
$$Q} )
- l jp., . .-'< . fr's. +.J'
s
. ; pp , .y;3 $ .lT 3 ;w;,. .
c:. . :, ' :. i 3 37: ._ .
~
A I .
. p:m-gdg%..;gy -
Lr .v 'p/g me " .. a. _. W
'E .J ,' _. ,}m. ?'e e - " . .' . . *.
_l '<mQ'l*.
}L [ & _, yll .}Q4 &f E
~
\f, 2.c kp , .
4 '
+ z,. .
,1., . 3m w. /' " * " ' ' ' .t *7 *
,. ,i . . A : . . - ~
ps.u :4 ;.s L,w x -s
~4_
}c. ' . ,. ? q
! . 4 _k
. .. . . . . ' _ . . . . . . . . . . . . . . . - . . . ~ . . . . . . -
~
g 2 8
$ t 0 }i R E T . . . _
822 nEit. 58M l $,_
*l ' l s .
q
.n O
I . N W 5
*# W' r.r. <* <..f.?.RQ g * /
6 -
\
\ %
tw I ) ' vv: : l , . . *V. . -
. . 5 l . .
<4 ,
\ ., .
* ~ ' ' ' " " '
I I , I-l
8 i l 1 l 1 E l . l l l ,. l I mus e I
- r. ..,
l ,.).$'d'\
)-
1 '. i\' ,e g lg
. y,' .;.,
p y {
; ,e p ; .
i '\ 'O i, *. _ L', *
.1.
. ,\
l .*
- e o
'h, ,
. y s.f.l v.
, .' . ' /
;: ', ' . .'.e-y ? '., h
& :' -~Ha Z ,
l. 1
. -i.i ,,
FIGURE 4. DETAIL OF PARTIAL SURFACE PENETRATION, WITH RADIAL DEFECTS (75X) l I l
_ 23 6 ~
, , , , , . , , , .. ,s - - ' " ' " " '
'S' 9', s. A
-:...."."'4 ' ,
9 O'AW - 3 .
,g .
g
~
.,. 4 h
~ Y ' h, fhh-I
.. 7, .
g I , , .
._' y
. \
~
. ps;;
9.a, .
- ,i
. - m , g I
% - 2 w
3h
+
w. 3 l
* - ) S. g,,,
r . b ,. . f-gg
-A
~
wh j . I, - w 4 me h i 1 - s-..
. ee I ,
gg f_ . I, .- O 1 . . . . . . . . . . . .. 4O CE E 4
, A 4 at , =.
g e 7 .h *% . . t: ,. e pm
- g 4 o==
. t M p
1 Q ~. '_ , ; - ;- % {0 >*- --,.-h-v ; > , gh
. . , .,..;....,,,~~.. f
.?
e j' -
- s. q. ,
s .
.m g
.h'.s, y7^ ~
;g l - . '. . .
t i > ,
. ' a g , - *
$g m 'k-4
, s
. x,
-i ~ <
8
"... m% -
4ss . ., ,
..).- '% * '
m
. W . .. . * '
.c n y c. , -
w
*-7 ~.
g
. 4 . h -
h k.. -
.' A ' ., ..,p Pt 6,
4*.,e}. '
. ' ^. :
v nh?'.4 ,. w
~_ .
e
-. , .-_, *. w m. ,
sc 4 .$ [_
.l
_3 g"7, .h.
, b
. ~'
- g. y +
& ?N. $%*; r , * ' ' '
_? .' > l
- , u-4 ,
Q ( ;.
. u.*
2
. :N
- a. ~
f . *f .. l' .
a e o l- 24 : E ! l2 , l 1. 1 , i _f 7 . . . . . .. l 4 we > eg . .
.M I
8 l . , ;
?- t y . ..
(w i 7% :
=c i
<% .4 1 4;
es *s8=*=, . ky . , x '" # e .f , . ; 4.M *
. ,, g is.* .. .-
m .. -' ,
- [, . .Y4 , y[li > li 1
y ^, h' __$j$ f *@' : ,
'sp
- f y . L.- r. ,. g.j.Qha.& ,.1 Rf ~ - & - ..;- ({ %%
, h .. .
.... g .....~ ,g
.~
l *
*yf ,%P --,,
<x ,; .gf, y ;g ,, 3. ,y .Ry
.4 %s,* .,s.g;,
y* i ,, . e s
, ,o 4 , '*Q ~ *?Y**'l,
- y. .-
1 ,i
- FIGURE 6. INSIDE SURFACE OF DRILLED HOLE, 1/4 INCH I
- FROM TIP SHOWING SURFACE TEARS (50K) 2
N
~
a - 25-
- M. . . . A p .
\
-,i
[ . 2 (a)
- - - . ~ _ ..,
=
%f .
. a.
l -e e g 9
/
4 4 8 f 5 l e+ 4 (b) FIGURE 7. DRILLED HOLE BORE SURFACE DEFECTS (UNDER NICKEL C0ATING USED TO STABILIZE SECTION) (50X)
Y l 26 J J B _ = - _ J L l
,rr .
g, ..,. e ,
,#W b.'t%,4 ,gt%.
'. ,.{. .. 3 l.y '.; 77 ~..
%, . ; <...ec, ' %. .
s . . . ( , 4 . ., . . t ' Ag , W ., 4 ,g .. % i e . b' .
;. p-4.siq;y;.; L e.x.-5.v.! __
,ra -
1 [.
' ' ; *. E:%;$, s .,* .,, ' ,.- Q g a, f m
- m. .
I ( l
-=
\ -. --
#M ..
M l N[ >::..-
.. ~.
\ ..
*;f .. ...
,Y '.. d,.,
','m.. .
1
- 7. . 5 kW .
se kat:.%af/f i (b) FIGURE 8. SURFACE DEFECT SHOWN IN FIGURE 7b (250X AND 500X) l ~ - - ~ _
~ 27^ ~ L I I I u , , , . li ' j i < : 5 g 1 Ca w io . ' '2 " - U -
. 11 ~
151 1 ~ l fvli
-C 3
/C ' '
recast.m 2 ' ' "t l E. g . Can 9 g, gg g, , g,3 .. I 1 l Caw ! - - -
~
p p C' s 4 -
* ' . Qwt:WDe05 3 i i 1 Di 'c = 0 * - .
0 02 0. 4 04 0. 8 10 3 0 0 01 C C2 44) - .0N-8/0 1 1 }/0 2. I FIGURE 9. t I THEORETICAL STRESS CONCENTRATION FACTORS FOR PARTIAL HOLE INTERSECTION I ! I i
'l ,
I
. . . = . -. - . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . . _ _ _
iq ll. p
'E Cll i
o. u I 7 l o it f
, ~e v
r I u c
- D O
D g .L A I 5' 1
.I
.R E
T [p 2 t .M A 4 - s e .D t . O R I 'h t D E N o w r L I A d g F
'/, R O
o I 3 o F 1 ) .A d K . A T
- (
d D s g t E o se T v I l t
.A R
d H K l o T W h O d s e .G R - r - h K O I t C
/
I .
. 1 A
R C 4 0 so 1 h I 1
.E
+ R t U -
G
+ s e I F
+ + t m.
h t m. 9 w m I O o r g 4 5 7 ., O
- - e. 9 ,
}~ - .O3U? _
I}}