Text

Evaluation of Diesel Generator Failure at Shoreham Unit 1, Failure Cause Evaluation, Final Technical Evaluation Rept
	ML20087P939
Person / Time
Site:	Shoreham File:Long Island Lighting Company icon.png
Issue date:	04/06/1984
From:	Ahmed S, Herrick R; FRANKLIN INSTITUTE
To:	Giardina R; NRC
Shared Package
ML20084L493	List: ML20087P939;
References
	CON-NRC-03-81-130, CON-NRC-3-81-130 TER-C5506-426, NUDOCS 8404100105
	Download: ML20087P939 (114)
	v • d • e

. .

l TECHNICAL EVALUATION REPORT !

EVALUATION OF DIESEL GENERATOR FAILURE AT SHOREHAM UNIT 1 FINAL REPORT, FAILURE CAUSE EVALUATION zm NRC DOCKETNO. 50-322 FRC PROJECT C5506 NRC TAC NO. -- FRC ASSIGNMENT 20 NRC CONTRACT NO. NRC-03-81-130 FRC TASK 426 Prepared by Franklin Research Center Author: R. C. Herrick i

20th and Race Streets

! Philadelphia, PA 19103 FRC Group Leader: S. Ahmed Pmpared for Nuclear Regulatory Commission Lead NRC Engineer: R. J. Giardina Washington, D.C. 20555 April 6, 1984 i

This rerort was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, or any of their employees, makes any warranty, expressed or implied, or assumes any legal liability or responsibility for any third party's use, or the results of such use, of any information, appa-ratus, product or process disclosed in this report, or represents that its use by such third .

party would not Ir. fringe privately owned rights.

l l

A l

00. Franklin Research Center

$1/g//Cd/06 DJ Wai n f The Franidin Institute The n=,mm Fruen rwk y. Ph6i... Pa. I9103 (21S) 4441000 m

dlOOFranklin Research Center

. A Division of The Franklin Institute April 4, 1994 U.S. Nuclear Regulatory Commission Washington, D.C. 20555 Attention: Mr. M. Carrington (MS-540)

Project Officer

Subject:

FRC Project C5506 NRC Contract NRC-03-81-130 FRC Assignment 20 FRC Task No. 426

Title:

Evaluation of Failure Cause, Diesel Generator Failure Shoreham Unit 1 Dear Mr. Carringtont The attached report presents FRC's technical review of the investigation of the diesel generator failure at Shoreham Unit 1.

Submittal of this report completes FRC's efforts on Task 426 of this Essignment. ,

Very truly youts, l

RJ S. Pandey Project Manger

. SP/SA/cm Enclosure cca R. J. Giardina (1 copy) -

-#R. ' Caruso (1 copy)

C. Berlinger (1 copy) 1 -

I

)

20th & Race Streets, Philadelphia, Pa.19103 (215) 448-1000 TWX-710 6701889

TECHNICAL EVALUATION REPORT EVALUATION OF DIESEL GENERATOR FAILURE AT SHOREHAM UNIT 1 FINAL REPORT, FAILURE CAUSE EVALUATION NRC OOCKET NO. 50-322 FRC PROJECT C5506 NRC TAC NO. -- FRC ASSIGNMENT 20 NRC CONTRACT NO. NRC43-81-130 FRC TASK 426 Prepared by Franklin Research Center Author: R. C. Herrick 20th and Race Streets Philadelphia, PA 19103 FRC Group Leader: S. Ahmed Prepared for Nuclear Regulatory Commission Lead NRC Engineer: R. J. Giardina Washington, D.C. 20555 h

April 6, 1984 This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, or any of their employe6s, makes any warranty, expressed or implied, or assumes any legal liability or t

responsibility for any third party's use, or the results of such use, of any information, appa-ratus, product or process disclosed in this report, or represents that its use by such third party would not infringe privately owned rights.

Prepared by: , Reviewed by: Approved by:

l$ f4d/ b 1.1lj h *;I * '/ A i ' >> $ ' /

Principal Author Project Manager Department Director (Acting)

Date: f3/f4 , Date: -+ /s 6 '9 Date: ' l ' I' "

si %

. 00. Franklin Research Center A Division of The Franklin institute The Benpamin Franksn Parkway. PNia.. Pa. 19103 (215)444 1000

a .

TER-C5506-426 CONTENTS SQction Title Page 1 INTRODUCTION . . . . . . . . . . . . . 1 2 ACCEPTANCE CRITERIA. . . . . . . . . . . . 3 i

3 PRELIMINARY INSPECTION AND REVIEW . . . . . . . . 4 3.1 Onsite Inspection . . . . . . . . . . . 4 3.2 Preliminary Technical Review-and Evaluation . . . . 6 4 TECHNICAL REVIEW AND EVALUATION. . . . . . . . . 20 4.1 Review of Crankshaft Metallurgical Examination. . . . 20 4.2 Review of Crankshaft Design . . . . . . . . 31 4.3 Review of Crankshaft Dynamic Testing . . . . . . 48 .

4.4 Review of FaAA Dynamic Model and Crankshaft Stress Analysis . . . . . . . . . . . 54 4.5 Review of Replacement Crankshaft Design . . . . . 59 5 CONCLUSIONS . . . . . . . . . . . . . 69 6 RECOMMENDATIONS. . . . . . . . . . . . . 71 7 REFERENCES . . . . . . . . . . . . . . 72 APPENDIX A - TORSIONAL CRITICAL SPEED ANALYSIS BY TRANSAMERICA DELAVAL, INC.

APPENDIX B - AMERICAN BUREAU OF f.3IPPING RULES FOR BUILDING AND

, CLASSING ST'.EL VESSELS APPENDIX C - INSPECTION COMMENTS CONCERNING DIESEL GENERATORS APPENDIX D RECOMMENDATIONS FOR MPCKANICAL AND ELECTRICAL COUPLING INVESTIGATION l

l APPENDIX E - COMMENTS BY H. W. HANNE3E ON THE

SUMMARY

OF SELECTED FAILURES AND EVENTS REPORTS OF TDI DIESEL GENERATORS l

l

d+a til db Frankhn Research Center A On.s.on af The Frannha insatute l-

s .

f TER-C5506-426 FIGURES Number Title Page 1 TDI Torsional Stress and Critical Spee6s . . . . . . 9 2 Typical Lumped-Parameter Torsional Mathematical edel . . . 35 3 Tangential Effort Diagram and Harmonic Components . . . . 38 4 Depth of Compressive Stress vs. Almen Intensity for Steel . . 66 5 Distribution of Stress in a Shotperned Bean with No External Load . . . . . . . . . . . 66 TABLES Number Title Pace 1 American Bureau of Shipping Tensile Property Requirements for Carbon Steel Machinery Forgings . . . . 22 2 Mill Certified Crankshaft Properties . . . . . . . 23 3 Chemical Analysis of Snoreham Crankshaft . . . . . . 24 4 Summary of Tensile Tests . . . . . . . . . . 24 5 Historic Values of Tn from Classical Soutces . . . . . 39 6 TDI Harmonic Coefficients . . . . . . . . . , 43 7 TDI Stresses for DEMA Rules (11-inch Crankpin) . . . . . 44 8 Comparison of FaAA's Tn Values with Ttzase of Lloyds and Ker Wilson . . . . . . . . . . . 56 l

9 Properties of Replacement 13 x 12 Crankshafts . . . . . 61 e

nidin Research Center A Cmesor of T A4 Fr.nWin inehade

a

TER-C5506-426 FOREWORD This Technical Evaluation Report was prepared by Franklin Research Center under a contract with the U.S. Nuclear Regulatory Commission (Of fice of Nuclear Reactor Regulation, Division of Operating Reactors) for technical assistance in support of NRC operating reactor licensing actions. This report

, constitutes the final report of the two-phase effort.

i l

I r nklin Research Center

~ ~ - . - .

)

i TER-C5506-426 i

1. INTRODUCTION In August 1983, a crankshaft of one of three emergency diesel generators

- (DG 101,102, and 103), manufactured by Transamerica Delaval, Inc. (TDI), and

. installed at the Shoreham Nuclear Power Station, owned by the Long Island Lighting Company -(LILCO), fractured during plant preoperational diesel generator tests. Inspection revealed severe cracking in the crankshafts of the other two diesel generators.

l During the failure investigation that followed, Failure Analysis Associates (FaAA) and Stone and Webster Engineering Corporation (SWEC) were

, engaged to carry out intensive analytical and experimental investigations.

Early inspection and evaluation indicated that.of the two remaining diesel generators (DG 101 and DG 103), sufficient crankshaf t operational life would be available from diesel generator DG 101 for an instrumented operational test

~

program if the cracks in the crankshaft were ground out. An operational test program was planned, and operational. tests were completed on September 28,

1983. In the meantime,' the two diesel generators that could not be operated ,

were disassembled for detailed inspection and rebuilding. Sections of the i

fractured crankshaf t from diesel generator DG 102 were taken to FaAA l laboratories for metallurgical examination of the fracture.

1 The Nuclear Regulatory Commission (NRC) requested that Franklin Research Center (FRC) provide an independent technical review of the failure investiga-tion performed by the Licensee and thereby provide a technical basis for the NRC's licensing actions regarding these failures.

Phase I included the following:

.a. attend an onsite inspection and review of the Shoreham diesel crankshaft failure, and review operation and maintenance history provided by the Licensee

b. analyze the data and-information obtained in the onsite visit and prepare an' interim report providing initial findings and conclusions

, regarding the events leading to crankshaf t failures

c. . provide in'the above any conclusions about other mechanical problems the Licensee has had with these diesel generators. .

-1_ _

II Frankhn Research Center

. A Opneson of Th, Franttn ingE0 MIS

TER-C5506-426 Phase II included:

a. review and evaluate submittals and data provided by the Licensee on the causes of these failures

b. provide technical assistance to the NRC lead engineer in evaluation of applicable data on diesel generators provided by the staff.

Accordingly, FRC participated in the onsite inspection, reviewed test procedures and diesel generator operating history, reviewed the crankshaf t ,

design analysis methods employed by the engine manufacturer, and participated as an active observer in the operational testing program prior to submitting an interim report (1]* covering Phase I. Subsequently, FRC reviewed the metallurgical examinations, diesel engine standards, specifications and design rules, and crankshaft design and analysis methods available to the industry, as well as performed a review of the testing methods, data output, and failure analysis conclusions of the operational testing program. In addition, the design and analysis of the replacement was reviewed.

This report includes the salient features of the interim report (Phase I) and elso reports on the subsequent events. In addition, it includes, as Appendix E, the commentary submitted by Mr. H. W. Hanners regarding selected problems experienced by the diesel engines manufactured by TDI. Mr. Hanners, an independent diesel engine consultant who was co-author of NUREG/CR-0660,

" Enhancement of On-Site Emergency Diesel Generator Availability," participated with FRC in the initial onsite inspection.

,

Numbers in brackets refer to references found in Section 7. -

l l

I s d '3, NFranklin Research Center A Dewesson af The Franklin insmde

TER-C5506-426

2. ACCEPTANCE CRITERIA Diesel generators are manufactured and purchased in accordance with various industry standards.

These standards include " Standard Practices for Low and Medium Speed Stationary Diesel and Gas Engines" by the Diesel Engine Manufacturers Association (DEMA) [2), " Rules for Building and Classing Steel Vessels" by the American Bureau of Shipping [3], and " Rules and Regulations for the Classification of Ships" [4], which includes " Guidance Notes on Torsional Vibration Characteristics of Main and Auxiliary Oil Engines" [5] by Lloyds Register of Shipping. Other rules and standards are available from European diesel manufacturer associations. In the absence of an ordered set

of acceptance criteria for this review Lod .' valuation, commentary regarding the applicable specifications and standards is included with the discussion rules, staM..: .ds, and methodology in Section 4.2.1 of this report.

l

'( ,

en as UN[ Franklin Research Center -

A Dmeen of The Franh>n Instaute

. i TER-C5506-426

3. PRELIMINARY INSPECTION AND REVIEW 3.1 ONSITE INSPECTION 3.1.1 Preliminary Briefing on September 1, 1983, a preliminary briefing about the current state of events and plan of action (6) at the Shoreham plant was held in the NRC Resident Inspector's office; a brief overview of the performance history of -

the three dierel generators was included. The briefing was conducted by the NRC Senior Resident Inspector and supplemented by the Director of LILCO's Office of Nuclear Power.

3.1.2 Inspection Tour of Diesel Generators A visual inspection was made of the three diesel generators. Although the observations are described in Appendix A, a brief summary follows:

o Diesel generator 101 was located in its operational room and was being prepared for a limited test program. LILCO reported that this unit had performed the initial qualification testing program for nuclear plant service and had been subjected to dynamic torsional testing as a part of that program. However, cracks were observed in the crank pin fillets of cranks 5 and 7.

o Diesel generator 102 was the unit with the fractured crankshaft. The

diesel engine and generator had already been moved to the main turbine deck where space and crane facilities were available to disassemble the unit, make a thorough inspection, and rebuild it with the 13 x 12 l

crankshaft now recommended by TDI. Considerable attention was paid to i

the crankshaft fracture in this inspection because of the imminence and magnitude of the failure. The entire engine was also studied to .

gain a perspective necessary for an adequate review of the many types l of failures experienced previously by the Shoreham diesels in order to determine if there may be a root cause not evidenced by the July 1983 ,

study (7].

o Diesel generator 103, reported to have crankshaft cracks developed to an extent that precludes further engine operation, was observed in its

! operational room. It was being prepared for movement to the main

! turbine deck for disassembly, inspection, and reassembly with the 13 x 12 crankshaft.

.. .2 Ob Franklin Research Center A Oms.oa of the F#annan Inattute

TER-C5506-426 3.1.3 Preparation of Requests for Information A meeting with the NRC representatives was attended at the Shoreham plant on September 1,1983, during.which the immediate and past problems experier .ed by the diesels at the Shoreham plant and their implication for similar diesels at other plants were discussed. Questions were prepared concerning the aspects of the diesels and their performance records that would be required for an adequate independent evaluation. These questions were submitted to LILCO by the NRC during the public meeting held at the Shoreham plant on

, September 2, 1983.

3.1.4 Public Meeting Onsite activities included attendance at the public meeting held at the Shoreham plant on September 2, 1983 for discussion of the diesel engine problems. Representatives of tia following organizations were in attendance:

Nuclear Regulatory Commission long Island Lighting Company Hunton and Williams, LILCO legal counsel Stone and Webster Failure Analysis Associates Counsel and Technical Consultant for Suffolk County Newsday Franklin Research Center.

During the meeting, the following points were established o There is no nuclear fuel at the Shoreham plant and consequently there is no demand on the safety systems.

o There is concern for similar diesels in other nuclear power plants.

o The problems with the Shoreham diesels are broader than the present crankshaft problem.

o The tests on the Shoreham diesels are significant for the whole nuclear power industry.

o V2e failure analysis team consists of LILCO, FsAA,'and Stone and Webster Corpcration.

o .TDI_is cooperating with the' failure analysis team to provide disassembly, inspection, alternate crankshafts and rework as

'necessary,- and reassembly of the engines.

vs Iankhn Research Center . >

, A Dnawan of The Frenehn ineeMe

TER-C5506-426 l

o TDI management is committed to the failure analysis and engine

- cebuilding program and will submit its own assessment and recommended actions; for objectivity, however, the program is under the direction

, of an independent investigator, FaAA.

o Concern of the community is high as represented by counsel for Suffolk j County and a technical consultant.

o A comprehensive failure analysis effort will be carried out to fully understand the failures so that the corrective action will be most effective.

3.2 PRELIMINARY TECHNICAL REVIEW AND EVALUATION l

3.2.1 Review of LILCO'S Master Plan A copy of LILCO's master plan [6] for the failure analysis and recovery 4

i of the diesel engines was received and reviewed. Comments (8) were submitted to the NRC, indicating where the reviewer's direct participation as an observer would be advisable. The plan was found to be acceptable.

t 3.2.2 Review of Test Procedures An early copy of LILCO's test procedure [9] for operational testing of 1

DG 101 was also received. The procedure was reviewed and a copy was forwarded I

to Mr. H. W. Hanners, an independent diesel engine consultant. Commentary and recommendations of this review were combined with those of Mr. Hanners and reported to the NRC in e?rly September 1983 before the start of operational tests. In the course of operational testing of DG 101, the original procedure and three revisions [10, 11, 12) were reviewed.

Recommendations for modifications and additions to the last revision [12]

to the procedure were made on September 24, 1983, by a meno [13] ' submitted via .

the NRC Resident Inspector at the Shoreham plant for expediency. Although the technical aspects of these considerations are discussed at greater length in

. Appendix D of this report, the recommendations provided means to assure that (1) all voltage phase references would be available.and known,-(2) transients associated with attachment of the generator to the electrical grid and to i

/ h -6 -

NIf Franklin Research Center m an.r,e m.

m-

a .

4-TER-C5506-426 major electrical loads would be recorded, and (3) testing at a significant synchronous loading would be recorded for power factors ranging from 0.8 to 1.0. These considerations were included in the tests carried out on September 28.

3.2.3 Preliminary Review of Diesel Dynamics Because the crankshaft failure and many of the earlier problems of the three diesels showed evidence of being associated with the dynamic response of the diesel engines, the torsional dynamics analysis summary prepared by TDI as

a part of the original design effort was reviewed. These analyses, which were stated in the September 1 and 2 conferences at the Shoreham plant to be verified as sufficiently accurate by FaAA defined the equivalent mass-elastic torsional dynamic model of the mechanical system, including the flywheel and generator rotor. They included the calculated natural frequencies and 5

critical speeds. This information was needed to form a basis of understanding by which the reviewers could evaluate the test data an'd recognize the response

of the various vibratory modes to the engine excitation orders in the course

of the test runs.

i i

3.2.3.1 Review of the TDI Mass-Elastic Model The mass-elastic model [14] employed by TDI to represent the dynamic natural frequencies and mode shapes of the engine is made up of 11 inertias 4

and 10 torsional springs. The inerti.as, identified in their order of position

. from the gear case end of the diesel to the generator, are:

4 o gear case and water pump inertia o eight equivalent inertias representing the piston, connecting rod, and rotating portion of the crankshaft for each cylinder assembly

-o flywheel t

o generator rotor.

J Torsional springs equivalent to that portion of the crankshaf t or generator rotor shaft were calculated by TDI for application between the inertias in the model.

g{Ikdi.3 Frankhn Research Center 4 won oe rh. nonen m.uw.

. , + - . - - - .- - - - ~ ~ ~ - - , - .

. s l

TER-C5506-426 TDI's analysis summary [14] indicated that the inertia for each crank assembly was the equivalent average inertia. This was an average of the real inertia comprised of a rotating crank, linear motion piston, and a connecting rod that combines both motions, all of which combine to form an inertia that varies with crank angle. Various methods are available to average these crank angle-dependent inertias to equivalent average inertias for use in the mass-elastic model yielding the natural frequencies. The detailed methods by which TDI calculated the equivalent inertia and torsional spring constant for each crank assembly were not evident from the analysis summary [9). Information to identify the methods used was requested from TDI through NRC. Methods developed by the industry over many years have been known to yield generally reliable results for most engine designs. This is not to imply that TDI did not use these methods correctly, only that its detailed methods were not evident in the analysis summary.

I For the interim report [1], it was assumed that the calculated inertias and springs, resulting natural frequencies, and critical speeds were suffi-ciently accurate. This premise was based upon the facts that qualificWiion testing was reported to verify these frequencies and that FaAA reported in briefings during the Shoreham inspection visit that it had obtained virtually the same values using more comprehensive computer methods.

Using its model for the resonant frequencies, TDI calculated the participation of the various orders of known engine excitation and plotted the amplification factors of the more dominant excitation orders as shown in Figure 1, which is a reproduction of the TDI chart from Reference 14. Note that the I-4 (4th order) curve of amplification factor remains the single largest participant, providing more than double the amplitude of the 4 1/2th order during operation at 450 rpm. The excitation frequency of the 4th order is that of 4 times per revolution. This is the firing rate of the cylinders which generates a sharp and dominant excitation. In short, using its analysis, TDI expected.to experience a large component of oscillatory torque at 30 Hz (engine firing rate) in the crankshaft. However, a large cyclic torque at 30 Hz is not necessarily bad, provided the associated stress levels d r:, ud Franklin Research Center nom ,an r,.wamau.

a .

TER-C5506-426 do m /Staal /wr#nxd $stMr .

h/Aens. Ad4RAtsst bLfg*2

. . . JJeaxaM4694#)GGM Aff CW . - . _ . . _ .

fMrat /d.* MM 7dasa/st

.. . . . . . . .... . .% c,ur . 73* di A rum 62miv7' 8975dW .

. _ . - - . . . . -. . -. . . Montw. Eceauancat .~ .. . . . . . . . _ _ . . . . .. . _ . _ . . . . _ .

M. . .tMe em .

.. . . . . % .. _ .. M45. .s _ . _ ._ .... . . . .._

s4 .64f5 ..

D MKPrw _ ..

% ewnu . . _

y- W' '

' ~

~ _

a ~ ,,, > - u .u f'to % de ew.

. /.

f.

. .. e. :

' ' ~

' s .z;st - ^f.g

... .1

'M .. . _ *

! / '.C 'q s

/ Q / V N ) Q~

Z th a

s aw .. ..i .. . Amur ,h.

= :oz

. . . . . ,, f, ,,,,,,,,g, t

l i

. ,4rs..E 2/M.

l t n* f 9.od 30 2n J.4 )$ ts, d,l,y d .# 94 Figure 1. TDI Torsional Stress and Critical Speeds

[from Reference 14]

/:b Ob_ Franklin Researen Center A Onim of The Frenham insek.de

. . .-= _ . .

TER-C5506-426

[

In combination with stresses from other sources in the crankshaft are adequately within the endurance limit of the material.

3.2.3.2 Investigation of Additional Constraints in the Torsional Mass-Elastic Model The influence of rotor-stator electrical coupling upon the purely mechanical torsional model employed by TDI was investigated in this review.

i Rotor-stator coupling of a synchronous generator may be approximated as an equivalent spring rate between the generator rotor and the inertia of the 5 electrical load. When the generator is connected to the electrical power i

grid, this equivalent inertia can be very large. In such cases, it is valid

] to approximate the ef fect by calculating the equivalent spring rate of the 1

rotor-stator system and inserting it into the torsional mass-elastic model between the generator rotor inertia and a new fixed rigid member (infinite inertia).

l

~

Review of TDI's mass-elastic system revealed that:

o The generator inertia was exceptionally large.

o The flywheel inertia was only between one-third and one-half that of the generator rotor.

o All other inertias representing crank assemblies, water pump, etc.,

i were very small by comparison.

Thus, although the introduction of the rotor-stator equivalent spring had an exceedingly small effect upon the natural frequencies of the rotor and I their dynamic response under the engine excitation, it did define a new mode of vibration not heretofore available from TDI's torsional model. This was essentially a rigid-rotor oscillation of the combined crankshaft, flywheel, ,

and generator rotor system with the rotor-stator spring connecting the j rigid-rotor system to the nearly infinite electrical load inertia mentioned 4 previously. The natural frequency (resonance) of this _ vibratory mode was independently _ calculated to be approximately 3.0 Hz. This vibration mode

! contributed to.the cyclic variation of power previously observed from the i control room to be at 3.75 Hs,'which is the frequency at which one complete i

ranklin Research Center

, - - . , _ . _. . .; - , . , _ . . . . . . . ;, . , . 4 - -- .

TER-C5506-426 set of eight cylinders fires, or the rate at which any one cylinder fires.

Portunately, the natural frequency of just under 3.0 Hz was sufficiently far from the 3.75 Hz excitation to prevent large amplitude vibration with resulting large swings in power. Also, the amortisseur windings of synchronous generators provide damping under oscillatory motions to limit amplitude buildup.

For diesel generators that may be coupled to the electrical power grid, TDI should have addressed this electrical-mechanical, rotor-stator coupling.

3.2.3.3 Investigation of other Mechanical-Electrical Dynamic Coupling When received, the torsional analysis report (14] indicated that the 30-Hz firing rate of the engine (4th order) would be sufficiently close to the first mode natural frequency, 35.5 Hz, to build moderately high amplitudes of oscillation at 30 Hz, which is one-half the electrical generation frequency.

Accordingly, recommendations were made (13] to ensure that any possible electrical-mechanical interaction would not be missed by the recording of data. These recommendations are included in Appendix C.

3.2.4 Review of Torsional Dynamics Testing 3.2.4.1 Initial Tests The torsional dynamics testing program, with operation of instrumented DG 101, was conducted to establish correlation with a detailed computer dynamic model of the diesel formulated by FaAA, as well as to investigate the dynamic interaction of the diesel with various loadings and operational conditions. The torsional testing program was primarily concerned with the catastrophic failure of the crankshaf t in DG 102 and near failure crack propagations in DGs 101 and 103. The testing was observed as a part of this review.

Li9 tings of measured parameters, sensors and transducers, and data i

recording equipment are provided in the test procedures (9, 10, 11, 12].

These include most engine operational temperature and pressure data, i.e.,

l

% b Franklin

[ Research Center m a n. r ma m.=w.

TER-C5506-426 lubrication temperature and pressure, combustion air pressure, each cylinder's exhaust pressure, etc.

Instrumentation for the measurement and recording of vital dynamic data included the following:

o Cranks 5 and 7 were instrumented such that crankpin fillet and web dynamic strain were measured by three element strain rosettes bonded to the fillet and by a single gage on the crank web.

o Dynamic torque in the crankshaft adjacent to the flywheel'was measured q by a strain gage torque bridge.

o Cylinder firing pressure of cylinders 5 and 7 was measured with high-i pressure piezoelectric transducers.

o Shaft dynamic displacement was measured by a torsional displacement transducer mounted on the gear case end of the diesel crankshaft.

o Linear acceleration of the engine base was measured by accelerometers mounted on the base at cylinders 5 and 7.

o vertical, horizontal, and axial acceleration (vibration) were measured

for the bearing housing next to the flywheel.

o Crankahaft position and revolution tachometer were referenced to top dead center of cylinder 7 provided by an optical sensor mounted on the generator shaft.

o Generator output voltages were recorded to measure the voltage difference between phases, (VA - V B) and (Vg - V )C *

, o Generator output current was measured for individual recording of each phase.

Instrumentation on the rotating crankshaf t was battery powered with -

signals transmitted by FM telemetry.

The initial tests were started on September 19, 1983, using the test "

procedure [11] dated September 15, 1983. Strain gage problems continued with gages dropping from service until five of the eight strain gagus in the fillets and webs of the crankshaf t were not operational. Testing was suspended at that point to repair the strain gage instrumentation. However, the test program had progressed through the initial checkouts, through the variable gm lib Franklin Research Center 4

4 % .e w,.naa w.w.

._.c . -- ~.

TER-C5506-426 speed torsiograph tests, and included the 1750-kW synchronous load test with the generator connected to the electric power grid. The full load tests, 3/4 load tests, and the TDI torsiograph tests remained to be accomplished.

3.2.4.2 Completion of Torsional Tests Following repair and improvement of the strain gage instrumentation in the crank fillets and on the crank web, testing resumed at 2:44 an on September 28, 1983. These tests included the test program in the test procedure dated September 23, 1983 [12). The test program, with instrumentation performing satisfactorily, continued to completion at approximately 7:30 am that same morning.

Testing began with Section 7.1 of the procedure [121, which involved measurements for verification of the analytic model at FaAA in Palo Alto, CA.

This was a correlation procedure in which the initial dynamic measurements were telephoned to the FaAA offices and checked against the analytic (computer) model both to verify the model and to permit the model to predict the available run time on the engine before crack propagation would preclude further testing.

Testing continued through the balance of the test procedere, including measurements recommended prior to the test and in the course of this review,-

and concluded with tests requested by TDI using its own.torsiograph and associated instrumentation.

Observations of data during the acquisition and recording of data on magnetic tape were somewhat limited, but these observations disclosed no instabilities with the electrical system or adverse transients in these tests upon connecting the diesel generator to various loads. As expected, the crankshaft torque signal and the crank fillet strain gages showed a signifi-cant 30-Hz component keyed to the pressure rise of each cylinder.

With the test data ~ recorded on magnetic tape, the review plan at the completion of testing was to permit LILCO and its contractors to review and verify calibration and zero settings of the various data channels in their ;

home facilities before conducting an independent review of the test data. l nWin Res

-n ea_rch Center

TER-C5506-426 3.2.5 Preliminary Review of Diesel Status Prior to Crankshaf t Failure 3.2.5.1 Review of Diesel Generator Test History Documentation of the test program at TDI prior to delivery of the diesel generators to the Shoreham plant was requested but was not received for review; however, statements by LILCO and TDI at the September 1-2 briefing indicated that a number of manufacturer's operational tests were performed on the engines in addition to the nuclear qualification program performed using DG 101.

It was also stated that the test data confirmed "to within 14" the critical speeds calculated during design. No statements were made concerniag whether these tests confirmed the amplification factors of each significant i order of vibration.

Reference 15 is a summary of Shoreham's test program for the emergency diesel generators received for review in advance of complete documentation.

This summary indicated that the test program was responsive to Regulatory Guides 1.108 and 1 9 and IEEE Std 387, in accordance with LILCO's commitments

~

in the Shoreham FSAR.

The test program was described as being of the " building block type"; it started with checkout and initial operation tests for individual components, and the components were then combined into subsystems and tested again. The checkout and initial operation were stated to consist of 138 tast packages in addition to 12 flush procedures, followed by 15 functional test procedures

[15].

Af ter the above tests, the diesel generators were operated for the first -

time as follows [15]:

OG 102 in October 1982

DG 103 in March 1982

, DG 101 in April 1982. .

Testing of eich diesel continued accceding to procedure, and the final test was performed to demonstrate the capability of_the diesel generators to i

' l complete successfully a total of 69 consecutive starts. According to 1 Reference 15, "By June 24, 1933, the emergency diesel generator preoperational

/, {sONI.nFranWin Rewarch Center !

A bmen d Dw FewWm hmNee

)

I TER-C5506-426 test program, including all mechanical, electrical and qualification tests, war completed for all three diesel generators."

In August 1983, all three diesels underwent a cylinder head stud replacement program, and one diesel generator completed the high load retest

[15]. LILCO's summary continues, stating that "one remaining demonstration of diesel capability was scheduled prior to fuel load; the integrated emergency core cooling system and emergency diesel generator operational demonstration."

LILCO reported (15] that as of the August 12, 1983 crankshaft fracture, the diesel generators had accumulated 2182 hours0.0253 days 0.606 hours 0.00361 weeks 8.30251e-4 months of operation as follows:

DG 101 -- 646 hours0.00748 days 0.179 hours 0.00107 weeks 2.45803e-4 months DG 102 -- 718 hours0.00831 days 0.199 hours 0.00119 weeks 2.73199e-4 months DG 103 -- 818 hours0.00947 days 0.227 hours 0.00135 weeks 3.11249e-4 months .

In response to a request for information by the NRC regarding the total number of operating hours on tach diesel generator and the total number of hours at 3900 kW or greater, LILCO responded [16] as follows:

Tbtal' Operating Hours for Each DG Unit at 2-hour Overload Rating (> 3850 kW)

Total Operating Hours (These hours included in total on Each DG Unit operating hours)

DG At At At At Unit TDI Shoreham Total TDI Shoreham Total' 101 128 518 646 3 16 19 102 30 688 718 3 19 22 103 40 778 818 3 20 23 3.2.5.2 Review of Conditions at the Time of Crankshaf t Failure In response to a request for information by the NRC about the test procedures in use at the time of the crankshaft failure, LILO3 rasponded with-the following description of the test [17):

r"1;.3 .bFranklin Research Center 4 %.e w re e.wou.

i.

TER-C5506-426

" Cylinder headc on DG 102 were replaced under R/RR R43-1001 with new design stress relieved heads. With all eight cylinders equipped with the new heads, the 102 DG was run for 12 hours1.388889e-4 days 0.00333 hours 1.984127e-5 weeks 4.566e-6 months to allow hot torquing of the exhaust header bolts and air start valve nuts. Following this run, a retest of the engine was begun under 8.7-R43-042. The specific scope of the retest under this 8.7 Form was tot

1. Verify proper diesel generator start to synchronous speed and rated voltage in less than 10 seconds.

2. Verify proper DG operation for four hours at the continous load ra ting .

3. Verify proper DG operation for 2 hours2.314815e-5 days 5.555556e-4 hours 3.306878e-6 weeks 7.61e-7 months at the two hour overload ra ting .

Refer to the response to NRC Request for Information II.2, pages 10.5 through 10.17, for a copy of the retest procedure 8.7-R43-042, as completed up until the time of the failure of DG 102."

On page 10.1 of Reference 18, LILCO provided the following detailed description of the events just prior to the failure:

"The diesel generator prior to the performance of 8.7-R43/42 was in its normal standby condition. An interim operating instruction was performed to ensure proper breaker positions, proper valve lineup and correct initial conditions. The diesel engine was started from its remote location, the main control room. Proper starting, acceleration to synchronous speed and rated voltage within 10 seconds was verified by the j test engineer and the OQA inspector. Plant Operator synchronized the diesel generator to BUS 102 by closing ACB 102-8 and then proceeded to increase the diesel generator load to 3500KW in less than 60 seconds.

Once at the 3500KW/300KVar load the operator was instructed to maintain this load for four hours. He was instructed that any deviations, caused by the LILCO grid, away from 3500KW/300Kvars should be corrected.

Another plant operator was stationed in the engine room.with verbal communications established between operators via headsets. During' the course of the four hout full load run, a LILCO technician was also stationed in the diesel engine room with the task of recording all pertinent test infopsation every 30 minutes. No abnormal readings were -

observed by either operator nor was the data written down by the technician'found to.be out of its normal operating range as specified by the engine manufacturer for this size load. .

Since this test was handled'similar to a Station Surveillance Procedure no special test equipment was utilized for data recording. All data written down was taken off of normal plant gauges either in the main control room or in the diesel engine room. The two exceptions were the generators bearing temperature and the generator stator temperature, both

/ W, NUf Frankhn Research Center A Onason af The Franneninet0We

. - - . . ~ -. -. . . . ..

l . .

TER-C5506-426 i of which were read off M6TE calibrated instruments. As stated in the 8.7 form high speed recorders were not used to record data on chart paper as a permanent record. Once, during the full load run the individual cylinder firing pressures were recorded and found balanced within i

manufacturer specified tolerances.

At the conclusion of the four hours the control room operator slowly increased the 102 generator load up to 3900KW/300 KVars. This load was t to be maintained at this level for the remaining duration of the test and

[ the operator was allowed to correct for any load deviations. During the increase in load, the lube oil low level alarm came in. The dipatick was checked and found to be below the shutdown level mark by 7-8". (This level is normal for high load operation of the DG units, and the alare i- has been an occasional occurrence on all three engines). Lube oil

, pressure and turbocharger pressure were normal and the test was allowed j '

to continue. Data readings were taken every 15 minutes. No abnormal noises were heard by the technician nor the local operator. Vibrations did not appear anything out of the ordinary; in fact the diesel engine seemed to be running fairly well.

The overload portion of the test was some one hour and 45 minutes into the two hour run when the diesel generator vibration was felt in the control room. The local operator reported no abnormal vibration.

l Generator load swings of 2.0 MW were c5perved in the control room meters, s

the operator reduced load to 1.0 MW tad the oscillations, subceeded. It was at this point that the generator 1;wd shot up to 4.0 MW where the operator tripped the output breaker ACB 102-8 and manually depressed the e 'stop' pushbutton. It was later observed that the engine overspeed trip had been activated and its alarm had been initiated. Other detailed descriptions of this failure are. attached, as well as a copy of the data sheets. Again no traces are available for analysis. Inspection of diesel crankcase internals s,howed the crankshaft web in the' area of No. 7 connection rod was cracked."

On page 10.3 of Reference 18, LILCO provided the following sequence of j, events when fracture is believed to have occurred:

[ "(Times are approximate and are intended .to illustrate sequence rather than exact time of occurrence)

Background:

EDG at 3.9 MW for 1 hour1.157407e-5 days 2.777778e-4 hours 1.653439e-6 weeks 3.805e-7 months and 45 minutes, 15 minutes from completion of scheduled 2 hour2.314815e-5 days 5.555556e-4 hours 3.306878e-6 weeks 7.61e-7 months run.

NASO - S. Livingston on headset at. EDG 102 panel in Main i Control Room, E. O. - M. O'Brien on headset 'in EDG 102 room.

5:15:00 Noticeable increase in vibration in Main' Control' Room - W.

Uhl, W. Nazzaro, W. Gunther approached Main Control room panel. Slight, but normal, fluctuation in load around 3.9 i

4 i

- wr# ef, NbFrankhn Research Center ao a w r e .mue

. - _ . .__ _ _ . -_ _ m - _ _ . _ _ _- - , _ _ _ _ _ _ . _ .

j i

TER-C5506-426 MW - no other indication of problem. Communication to EDG room for observation of any problem - only response was technician was in area taking readings.

5:15:45 Vibration continued and suddenly load swing of 1.5 to 2.0 MW commenced between 2 MK and 4 MW. Communication from CA to field - 'are yoit doing ar.ything'. Within 15 seconds, load was reduced by CR operator to 1 MW. Vibration ceased. This load was carried for about 15 seconds.

, 5:16:15 Load increased without cause to 4.0 MW. Vibration increased again. Again communication betwaen CR and EDG room regarding what was going on. W. Nazzaro, instructed Livingston to decrease load. Load would not come down.

5:16:30 W. Nazzaro instructed Livingston to trip the machine who immediately opened the output breaker. Speed was noted to reach 600 RPM before coasting down to rest.

Elapsed Time - 1 1/2 minutes" 3.2.5.3 Review of Previous Vibration Survey 1

There was ample evidence of concern over the vibratory amplitudes of the diesels. Review of the partial listing [19) of selected previous problems 1 with the diesel generators also provided evidence of high dynamic forces that i

had the potential of being associated, on preliminary evaluation, with large amplitude torsional vibration. Should subsequent thorough evaluation prove this to be true, then many of the various component failures would no longer be isolated independent events as previously reported but linked to a common cause.

Until more information is available, the following evidence of repeated

[ failures in components directly connected to the crankshaft remains circum-

stantial

l Date Failure Description I

l 3/30/C3 Holddown capscrews, rocker arm' assembly (EDG-103) 9/17/82 Jacket water pump shaft (EDG-102 and 103)-

10/05/81 Piston crown separated from skirt I

i r nklin Research Center

. A Dmeson of The Frenann Inseeuse '

-e .

l' i TER-C5506-426 Date Failura Description 10/05/81 hilure of attachment stud bolts 10/05/81 Grooving of crankshaft bearing and crank pin discolored 10/05/81. Wrist pin grooved and pitted, wrist pin discolored.

l Concern over vibration was sufficient to initiate a vibration testing program in the late spring and early summer of 1983 [7]. The conclusion of this study states:

1 "On the basis of comparisons of vibration data taken, the Shoreham diesels have only the expected and normal vibration and are not subjected to any excessive vibration and this normal, expected vibration does not i

prevent the diesels from reliably performing their functions."

, It is noted that the study was based entirely on linear vibration i

measurements without any measurement of torsional vibration. It is true tha,t rotating machinery can suffer from high torsional vibration with litcle evidence of linear vibration. However, the crank mechanisms of diesel engines provide coupling between the torsional and linear vibratory systems so that there is usually evidence of linear vibration associated with torsional <

i vibratioa.

i i l

i 1

1 f

4

-)

e.-A Frankhn Research Center .

A Onomen of The Fremen ensuue

- - - - __ . ~. - - . _ - . . . -- _- - . - _ - . _ - _ -

{

j TER-C5506-426 4

).

, 4. TECHNICAL REVIEW AND EVALUATICN i 4.1 REVIEW OF CRANKSHAFT METALLURGICAL EXAMINATION i

I 4.1.1 Material Specifications and Certifications 4

All documentation submitted by LILCO to the NRC regarding the crankshaft '

4 indicated that the only material specification for the diesel generators was

{ that provided in the diesel gencerator purchase specifications [20, 21]. Pages 9 [20] and 1-10 [21] of the purchase specifications cite " Standard Practices i for Low and Medium Speed Stationary Diesel and Gas Engines" [2], published by

, the Diesel Engine Manufacturer's Association (DEMA), under the heading of 1

applicable documents. No other document defining diesel engine material

!- specifications was noted. However, " Standard Practices for Low and Medium Speed Stationary Diesel and Gas Engines" [2] does not cover crankshaft materials other than to limit the cyclic stress level under torsional i vibration. Although DEMA's recommended practices are discussed at greater length in Section 4.2.1, it may be stated here that no documentation defining a required minimum quality level of the crankshaf t, or other engine componedt, was found.

i References 22, 23, and 24 indicate that American Bureau of Shipping (ABS) j Grade 3 steel was specified by the engine manufacturer, TDI, for the two d

crankshafts purchased from Ellwood City Forge, Ellwood City, PA. Reference 25 indicates that the third crankshaft was purchased from Mitsubishi Steel

Manufacturing Company, Ltd. in Japan. However, Reference 25 does not include an indication of the grade of steel specified by TDI, but does show that the

material conforms to ASTM A273, Gr. AISI C1042. Although the ABS rules covering steel machinery forgings are currrently written to be "in substantial '

f agreement" with classes of ASTM A668 steel, the ABS rules did reference ASTM A235 in the years of engine manufacture, 1973-1975. Study of ASTN A235 indicated 'that ASTM A273 could be specified where the forging mill desired to !

use semi-finished steel for the forgings. Thus, ASTM A273 is a specification j

for carbon steel blooms, billets, and slabs for forging rather than a specifi- l cation for carbon steel forgings (A235) . The conformance to ASTM A235 after

6 I ONdFranklin Raearch Center <

A Dwun d TM f teNM kemM

TER-C5506-426 l using ASM A273 fotging material would have depended upon the subsequent

forging processes. These processes were not indicated.

, The ABS. " Rules for Building and Classing Steel Vessels" (ships) [3]

I specifies steel forging material properties primarily by minimum material ,

properties, by certain limitations in steel processing, and by " substantial agreement" with designated ASTM specifications. The minimum properties of ABS Grade 3 steel, shown in Table 1, were taken from the 1973 and 1980 editions of i the ABS's rules. The 1973 edition indicates that ABS Grade 3 steel is to be

). in substantial agreement with ASTM A235-67 Class E and that the steel forgings i

j are to be annealed, normalized, or normalised and tempered. No restrictions on steel chemistry are noted in the 1973 edition, but the 1980 edition of the rules specifies that the chemical composition is to be reported and the carbon content is not to exceed 0.354 unless specially approved. Although this carbon content stipulation was introduced after 1973 and before 1980, the f interpretations of this rule appear to be such that if the steel meets the other chemistry and processing requirements, meets or exceeds the requirements of Table 1, and is in substantial agreement with the respective ASTM specification, the steel will usually meet with the approval of the ABS even though it may contain a greater carbon content. Thus, concerning the actual j crankshaf t steel, with properties as reported by the forging mills [26, 27]

and shown in Table 2, approval to qualify as an ABS grade can only be granted 4

by the American Bureau of Shipping following its submittal to them for' review.

As part of the early analysis of the failed crankshaft, FaAA analysed the steel's composition and tested its mechanical properties [28]. The results of these analyses and tests are shown in Tables 3 and 4 for comparison with

., Tables 1 and 2. The mechanical properties exceed the minimum requirements of t

Table 1 by a fair margin. With respect to chemical analysis, FaAA reported that "Except for the carbon and sulfur, which were determined by a combustion gas analysis method, the chemical analyses were obtained by an inductively-l coupled plasma technique." FaAA concluded that the steel met the ASTM 235-67 i

Class E requirements in accordance with the specification'of ABS Grade 3 steel. FaAA's chemical analysis indicated a chromium content of 0.34.-

Mnnya UUUU FrankNn Research Center A Onimen af The Fenden ineuhde

. . . . . ~ , , _ . ~ - , , . _ -. , _ - - ., , . _ . .

t c -.

"c 3 its 2"

-I IA:r -

2 la

'% Table 1. American Bureau of Shipping Tensile Property Requirements for Carbon Steel Machinery Forgings Longitudinal Transverse Size Reduction Reduction

-Under Yield Elongation in Area Elongation in Area g Over Tensile (1) (t) 8 Grade (in) (in) (psi) (psi) (t) (t) 12 75,000 37,500 22 35 18 28 3 8 75,000 37,000 20 32 18 28

> 3 12 20 li=

A b.

a e

e

.1

i

. =

,7...,

FLI.

g 13 l 2"

- ,3 D:r n

Table 2. Mill Certified Crankshaft Properties Mechanical - Proper t ies Crankshaft Material Chemical Analysis (%) Reduction-Eng ine Supplier Spec C Mn St P S , Tensile Yield Elongation in Area s

w DG 101 Mitsubishi ABS Gr. 3 0.42 0.64 0.21 0.014 0.011 87,100 47,200 24.2 44.0 4 74010-2604 Steel ASTM A273 87,600 47,700 24.2 42.0 AISI 1042 DG 102 Ellwood ABS Gr . 3 0.47 0.83 0.18 0.008 0.010 94,500 51,500 24.0 50.3 74011-2605 City Forge 97,750 52,000 25.0 51.7 DG'103-' Ellwood ABS Gr. 3 0.47 0.83 0.18 0.008 0.010 96,000 53,500 24.0 51.4 74012-2606 City Forge 98,000 55,000 24.0 52.2 N '

?

O e, ,

M

TER-C5506-426 Table 3. Chemical Analysis

of Shoreham Crankshaft

^

Ellwood City ASTM A235-Element 1st Piece 2nd Piece Mill Qnemistry 67 Class E C 0.47 --

0.47 0.4-0.47 Mn 0.6 0.65 0.83 0.9 max Si 0.12 0.12 0.18 --

S 0.014 --

0.010 0.05 nas P 0.01 0.01 0.006 0.05 max Cr 0.30 0.39 -- --

Ni 0.054 0.055 -- --

Mo 0.03 0.03 -- --

V 0.04 0.04 -- --

Cu 0.04 0.04 -- --

Al 0.004 0.004 -- --

Ti --

0.03 -- --

All elements are reported in weight percent.

Table 4. Summary of Tensile Tests l

Yield .

Stress Ultimate Reduction j Specimen (ksi) Strength Elongation in Area Number (ksi)

Upper Lower (4) (4) ,

R1 46.6 45.0 89.0 25.4 42.0 R2 45.3 44.9 89.4 30.0 45.1 Tl 47.1 45.9 87.6 37.1 49.1 l T2 46.9 46.9 88.2 39.0 47.6 L1 47.3 45.9 89.5 25.1 35.3 L2 47.4 44.8 89.1 23.0 30.6 l

i l

if a '

UdN Franklin Research Center 4 5 e n r, mea maw.

i l

T TER-C5506-426 It is noted that neither ABS Grade 3 nor the ASTM A235-67 Class E specifi-cation limits the chromium content so '.ong as it is a residual amount. A chromium content of 0.3% for this steel is considered to be residual. '

f FaAA's conclusion that the crankshaft steel, as analysed, meets the requirements of ABS Grade 3 specification designated by the engine manufac-turer is generally acceptable. However, it should be noted, as discussed 4

(bove, that only the American Luceau of Shipping can approve a material as conforming to an ABS grade if its chemical content is different than the range specified by A8S.

4

4.1.2 Metallurgical Examination of Fractured Crankshaft This section summarizes the metallurgical examination performed by FaAA 3 on the fractured crankshaft from DG 102 and reported by FaAA in Section 4.0 of j Reference 29.

j Appendix 3-1 to FaAA's report contains the agreement reached by LILCO, Stone and Webster, TDI, and FaAA regarding the. extent and proceduce for cutting and sectioning the featured portion of the crankshaft for meta 11ur-i t

gical analysis. The beginning of tnis process was described by FaAA [29) as I

follows:

l "The failed crankshaft had fractured into two pieces at the crank pin

), journal of cylinder No. 7. Fracture occurred mostly through the web

{ connecting the No. 7 crank pin journal to the adjacent No. 9 bearing

journal. The section examined was saw-cut from the crankshatta cuts were l made through ti.e No. 9 and No. 9 main-bearing' journals. This two-piece section containing the fracture was shipped to FaAA's laboratory in Palo

Alto, California for laboratory examination.-

Both pieces of the fractured section were examined visually; then the

}' metallurgical failure analysis was performed on the piece nearest to the ,

No. 9 main bearing. 'the other piece, with the mating fracture surface, has been preserved for any additional examination that may become

appropriate in the future."

The two pieces of the cutout fractured section of the crankshaft are

)

+

thown in Figures 3-1 and 3-2 of FaAA's report [29). Figures 3-3, 3-4, 3-5,

! cnd 3-6 of that report show the methodology of sectioning the half of the -

fractured segment used for metallurgical examination.

A2 [kFrankhn Research Center U

A Dunmen et The Frenen in.eMe a

l i

TER-C5506-426 4.1.2.1 Visual Analysis

PaAA (: D] reported the location, orientation, and charactecisation of the fracture surface as follows:

"The fracture surface exhibited an obvious, unmistakable fatigue crack l pattern. Concentric beach marks showed that the fatigue crack started at the surface of the machined fillet radius where the crank pin journal 4' blends into the web. The orientation of the fracture plane at the origin can be described using a visual analogy of a clock face: (1) the clock position is viewed from the output end of the crankshaft, (2) the clock face is centered at the crankpin axis, and (3) the 12 o' clock position is i

at the point on the pin journal furthest from the crankshaft rotation -

axis. The location of the fracture initiation is 0.055 inch in the radial direction from the journal surface and at a 4:30 clock position."

l FaAA supplemented its discussion of visual analysis with photographs that i clearly show a classic development of a fatigue crack. FaAA's Figure 3-8 shows the initial crack development area in the fillet between the crankpin and the web, wherein the early crack development is characterized by fine 1

beach marks indicative of slow progressive crack growth over a large number of

stress cycles.

The FaAA report characterized the crack growth and its orientation as being similar to that associated with pure torsion in a cylindrical member.

Because the developing crack plane deviated somewhat from the ideal torsional case, FaAA's discussion correctly noted the modifying influence of the tran-1 i sition in geometry from the cylind:ical crankpin journal to the crank web. '

, FaAA's basic findings of the visual analysis and their observation that j the crack's surface propagation is essentially normal to the maximum tension +

stress reaulting from the combined cyclic and steady output torque in the shaft were determined to be correct during this review. .

4.1.2.2 Scanning Electron Microscopy FaAA used scanning electron microscopy (SEM) to investigate and 1 j characterize the point of origin. This analysis was reported in considerable l detail, and 32 SEM photographs were included in the repc;t (29).

4 5

L I

%uu d Frankim Research Center A On=esen of f>e F#eneen inenswee

TER-C5506-426 In summary, the analysis confirmed that the fatigue crack began at one of a numbet of score marks on the crankpin fillet that was somewhat deeper than the adjacent machining marks and other score marks. The analysis also proved that the fatigue crack was not unduly influenced by the score mark or other small material imperfections present. No matter how ideal a material may be, small or even microscopic imperfections at which a crack will begin are always present. This means that, although the fatigue crack did initiate at a particular score mark, it would have initiated in that region a number of cycles later had the score mark not been present.

I 4.1.2.3 Metallography Af ter completion of the SEM examination, FaAA reported [29] that the sample containing the fracture origin was diamond saw-cut to expose a cross section through the point of origin. The cut surface was ground down in steps to permit viewing at several levels relative to the point of fracture origin.

The FaAA report included photomicrographs of the various cross sections.

One optical photomicrograph showed the depth of the score mark to be 0.002 1

inch in that plane. PaAA's description and characterization of the metal

, structure and the score mark follow:

"It is apparent, from the disturbed microstructure at the score mark, that local plastic deformation occurred when this score or anomaly was made on the machined surface; that is, the score mark was indented into

' the surface, not gouged out. This indicates that the score mark was made after machining. Local deformation resulting from the creation of the score mark may have left highly localized residual stresses that made this location more prone to be the point of origin of a fatigue crack than the surrounding machined surface.

i The microstructure was uniformily fine pearlite and ferrite. Figure 4-52

[9] shows this microstructure. The pattern of pearlite and ferrite reveals that the austenite grain size varied from ASTM 4 to ASTM 8. This duplex austenite grain size is consistent with the fact that the steel was not aluminum killed, and, therefore, it is not inherently fine grained. The microstructure is as to be expected for a steel of this type when'it has been given a normalizing annealing treatment at 1600'F.

r The steel was clean and relatively free of nonetallic inclusions. The

! inclusions observed were generally fine and were distributed uniformly i

L i

M '

likj Frankhn ao Research an,ee e .we C. enter

..-r., . .

TER-C5506-426 throughout. Figure 3-53 [9] shows the largest cluster of inclusions seen i on the metallographic cross section at Plane B. No unusual inclusions were found near the fatigue crack origin. It is concluded that the microstructure was normal and proper for a forged steel crankshaf t. The failure did not originate at an inclusion in the material."

These findings were reviewed and found to be satisfactory.

4 4.1.2.4 Macroetch Analysis 2

PaAA reported [29) that slab sections were escroetched to reveal flow 4 line patterns, segregation, general inclusion distributions, and any forging or ingot defects. Their report, supplemented with photographs of the etched sections, indicated that the sections had smooth forging flow line patterns with no metallurgical anomalies.

4 FaAA's report that the macroetch results (as shown therein) indicate that

! the crankshaft forging was metallurgically sound was reviewed and found to be 3

satisfactory.

4.1.2.5 Hardness Measurements i

j The following sumnary of hardness measurements made and reported by FaAA (29] was reviewed and found to be acceptable:

i j "Two conclusions can be drawn from these hardness test results. First, no systematic variation in bulk hardness values was observed. This supports the other metallurgical evidence that properties of the shaft

, are homogeneous and that any forging-induced or as-cast heterogenities 4

have been effectively removed by subsequent heat treatments. Second, the j hardness values are consistent with the mechanical properties measured on ,

samples removed from the failed crankshaft, and the measured hardness

values are appropriate for a properly heat-treated steel forging of this j type."

4.1.2.6 Residual Stress Measurements Residual stresses measurements were made and reported [29]. These.

l stresses were reported to be made using a position-sensitive scintillation 1

detector wherein a single exposure technique determined the residual stresses l l

1 t

in a surface layer about 0.0005 inch thick. The report, also indicated that l l

i

M Uhf Frankhn Research Center A Ommen of The Feerebn insenwee

O U-t 4

i-T 4

TER-C5506-426

'the area over which the stresses were averaged was 0.040'by 0.040 inch.

Reproducibility was reported to be poor due to errors in achieving sufficient

) accuracy in reposition to a previously measured area.

Residual stresses were reported to be measured first using a specimen 1

(designated Section H) of the failed crankpin and web. These stresses were 1

reported to be low in value, and were judged to have been influenced both by the relatively small sample size and by the fracture process and subsequent failure events: the engine did continue to operate briefly and to increase in i

. speed before being shut down following crankshaft fracture. 'the associated battering of the fractured surfaces could have instituted a stress relaxing

, process.

i Residual stress values for a subsequent larger specimen cut from the j No. 5 crankpin, where the post-fracture environment was much less severe, were reported to vary from approximately -20,000 to -55,000 pai and were reported (291 to be representative of the actual residual stresses in the machined '

i fillet radius of the original crankshaft.

The residual stresses for the second, larger specimen could also be in

[ error because the possibility of residual hoop stresses in the crankpin and j

i tillet was not considered. In obtaining the second specimen, an axial cut was 4

made through the crankpin that included only a small arc of the crankpin and fillet in the sample. If residual hoop stresses were present'in the original i

l crankshaft, they would have been largely relaxed by cutting out the material specimen. If these original' hoop stresses, when considered alone, produced

' tensile stresses on the surface, then their relaxation in conjunction with other compressive residual stresses woald have decreased the tensile component

[ and moved the state of surface stress farther into the compressive-region.

This would have been true for both the hoop and' axial stresses, which are 4

coupled.through Poisson's ratio. Reference to Table 3-9 of. Reference 29.

appears to give some evidence of hoop stress relaxation, but not enough toL h

' substantiate any conjecture that significant residual hoop stresses are being

' relaxed. It is noted in Table 3-9 of' Reference 29 that the axial stresses for the edges of the specimen (points B & C and H & I) where relaxation wouldt o

L anklin Research Center

. A Onisson of The honeen insanuse

o .

TER-C5506-426 4

be greatet are higher in cume.assive stress than the pointe in the center (points E & F). The pattern appears to be symmetrical as expected in the

, specimen. If this were true, it would indicate that the state of surface stresses could have been much less into the compressive domain than that shown by Table 3-9 of Reference 29.

In addition to the above discussion of residual fillet surface stress in the original crankshaft, it is believed that the presentation of the residual

, stresses in a 0.0005-inch-thick material layer at the fillet surf ace is misleading without a discussion of their ramifications. It is true that a

state of compressive residual stress is desirable in machine members subjected to cyclic (fatigue) loading. However, without the knowledge of how these stresses vary with depth, the stresses lose much of their meaning. They could represent a very shallow distribution sometimes induced by machining methods and may mask a subsurface state of stress conducive to earlier fatigue failure.

In summary, the conclusions in Reference 29 concerning residual stresses

refer to residual stresses induced by machining. These residual stresses induced by machining appear to be compressive, although possibly not to the magnitude reported and, as such, would not contribute to a state of surf ace stress conducive to crack formation.

4.1.2.7 Metallurgical Summary FaAA's conclusions about the metallurgical examination follow [29] .

l " Failure of the crankshaf t in the DG102 emergency diesel generator occurred by the formation of a fatigue crack. The fatigue crack started j in a surface score on the fillet surface where a crank pin journal blends I

into a web section. The identification of the fatigue character of the fracture is unequivocaule. Unmistakable beach marks are clearly visible

on the fracture surface.

Microstructure, composition, and mechanical properties of the forged steel crankshaft are proper as required by ABS Grade 3 or the equivalent ASTM A-235-67 Grade E, the pertinent steel specifications. The 4 crankshaft is metallurgically sound.

The location and planar orientation of the fatigue crack indicate that the crack was caused by cyclic torsion stresses superimposed on the constant torsion stress resulting from the engine output torque.

pg I 00b FranWin Research Center A Onas.on cd The Frankhn insetWe

TER-C5506 -426 Since the metallurgical characteristics of the steel are good, it must be concluded that the fatigue failure occurred because of excessively high cyclic stress as applied to the crankshaft during testing.

The fatigue crack was not caused by the score where it initiated, nor did the surface imperfections on the machined surface of the fillet near the crack origin influence the formation of a fatigue crack. Had the machined surface finish been better, the fatigue life of the crankshaft probably would have been longer. Such features are to be avoided in fatigue-prone parts. Had the crankshaf t not been subjecttd to excessively high service stresses, the fatigue crack would not have been initiated even at the deeper score mark."

The present review is in agreement with these conclusions.

4.2 REVIEW OF CRANKSHAFT DESIGN 4.2.1 Rules, Standards, and Methodology Applicable rules and methodology of design vary widely, depending upon the appli ition of the diesel engine. In general, design rules exhibit greater conservatism of design for applications where engine reliability is paramount. Rules governing tha design of engines for ship propulsion are usually conservative and reflect both the need for safety and the more limited repair facilities at sea. The following sections discuss two sets of standards made applicable by the purchase specification and by the-engine manufacturer's material specification. Crankshaf t design methodology. is also discussed.

4.2.1.1 Diesel Engine Manufacturer's Association (DEMA)

DEMA's " Standard Practico, for Low and Medium Speed Stationary Diesel and Gas Engines" [2] was cited under the listing of applicable documents in the purchase specification [20, 21] of the diesel engines for the ~Shoreham Nuclear Power Station. The 6th Edition is the latest edition, published in 1972 before the purchase of the diesel engines.

These DEMA standards constitute a set of nonmandatory guidelines for the purchase of diesel engines. The scope of the standards are best expressed by the foreword from the 6th Editions t((r nklin Research Center A Onomon af The Freneden Onethste

o .

TER-C5506-426 "This book has been published to serve as a reference for consulting engineers, government agencies, users, suppliers, power plant superintendents, and engine operators. It provides generally accepted standards for nomenclature, installation, application, operation, and maintenance of engines and accessory equipment in various types of stationary engine installations.

It is not the purpose of this book to attempt to set forth basic design criteria for engines because such approach would be impossible within this volume and yet do justice to the many types of engines on the market, notwithstanding the fact that many technical texts are available to the student who may be undertaking the design criteria aspects of engines in general.

The existence of, or the adoption of, a standard by DEMA does not in any respect preclude any member or nonmember from man'afacturing or selling products that differ from these standards."

With respect to the crankshaf ts, materials are neither specified not recommended. However, Chapter 7 is devoted to vibration, where design objectives and criteria are discussed under the topic of torsional vibrations. Three aspects discussed in the section apply to the Shoreham diesel engines. These are as follows:

o "In the case of the constant speed units, such as generator sets, the objective is to ensure that no harmful torsional vibratory stresses occur within five percent above and below rated speed."

o "For crankshafts, connecting shafts, flange or coupling components, etc., made of conventional materials, torsional vibratory conditions shall generally be considered safe when they induce a superimposed stress of less than 5,000 psi, created by a single order of vibration, or a superimposed stress of less than 7,000 psi, created by the summation of the major orders of vibration which might come into phase periodically." .

o "For the case of shaft elements variously know as ' quill shafts,'

' tuning shafts,' or ' torsionally resilient torque shafts,' and other .

elements which are specifically designed for the application, and manufactured from material of adequate physical properties, with careful attention to design and machining of keyways, fillets, etc.,

superimposed vibratory stresses at much higher levels may be acceptable. The design of such elements is always correlated in the torsional analysis."

With respect to the second item above, which recommends limits of 5000 I and 7000 pal for crankshafts, it is generally conceded that these stress

1._, ,

Udk40-w FranWin Re, search Center l a n r,. am.w, j

o e l

TER-C5506-426 i i

'~

limits apply to torsional stresses calculated for the crankpin diameter l without the app;ication of stress concentration factors. In fact, stress I

j concentration in the fillets, or the oil hole, is compensated by recomeniing i low cyclic stres.t limits for the crankshaft. However, such practices are not adequate if the stress concentrations are not limited by a parallel design !

j standard covering crankpin oil hole and/or fillet geometry where stress

concentrations ca1 be high. '

s By contrast, the third item from the DEMA recommendations above permits i, stresses higher than the 5000 and 7000 pai limit for quill shafts and other L elements specifically designed for the application and manufactured from

} material of adequate physical properties with careful attention to j stress-concentrating geometries. The contrast is this: the crankshaft limits 1

include an historical perspective of experience for similar crankshaft j geometry, whereas this is not generally possible for special equipments i therefore, full engineering analysis and judgment are required if these values ,

are to be exceeded. '

I j 4.2.1.2 American Bureau of Shipping (ABS)

The " Rules for Building and Classing Steel Vessels" [3] by ABS were not '

! invoked as standards for the design of the Snoreham diesel engines, other than the engine manufacturer specified ABS Grade 3 steel in the procurement of the crankshafts. However, the ABS rules are representative of the various rules i

used to classify ships, of which probably the world's best known rules are l'

Rules and Regulations for the Classification of Ships (4) and Guidance Notes

on Torsional Vibratien (5) by Lloyds Registet of Shipping. Lloyds rules, as l

i well as the Ass rules, are published each year to reflect constant updating.

f ship classification associations, such as the American Bureau of Shipping *

! and Lloyds Register of Shipping, represent possibly the oldest machinery !

l review and evaluation associations functioning today. Lloyds Register began l- operations in 1760 and published its first set of rules in 1834. As ships and !

i t ship propulsion systems became more sophisticated, the classification

i

' associations served as design review agents to evaluate functional adequacy l- and safety. Considerable experience in the review and evaluation of diesel

. :, ~33-hFraphiin rch C a wRese.a.seani enter 4 o, w. w.

t TER-C5506-426 engines was realised from the long-term use of diesel engines for propulsion l

and for electric power generation in ships. The ship classification rules ;

! probably represent the most extensive esper!ence in large diesel engines

available.

I l 4.2.2 'Ibraional Dynamic nossolas Anahsis i 4.2.2.1 Summary of Analysis Methods Reciprocating engines, especially diesel engines, have always constituted

! a challenge for the dynamicist. ne motions (kinematics) of the engine parts *

are complex diesel.s, in particular, are subjected to sharp cylinder pressure rises that serve to further excite vibratory natural frequencies in the complete power system as well as in the engine.

]

l The dynamic response analysis of an engine or complete power system to internal or external excitation begins with some mathematical model of the j system's dynamic properties. S e approach most used is based upon an appro-l_ priate consolidation of inerties (torsion vibration) and spring constants such j as that shown fn Figure 2.. With the mathematical model established, the l dynamicist has a choice of action l

o solve the dynamic equations directly to yield both the transient and j steady-state solutions

! o modify the equations and procedure of direct solution to drop the j transient response and retain the steady-state vibratory response i I

o make an electric-analog model and simulate the mechanical vibration i o resort to specialised numerical procedures that have been devised over '

l the years prior to the introduction of digital computers (e.g.,

Holser'a methods)..

In theory, the first course of actinn is the simplest: write equations equating the summation of forces (torques) on each mass (inertis) to the 4

acceleration of that mass (inertia) . Direct solution of these equations will provide the transient and steady-state solutions to yield a simulation of the

~

er.gine system's dynamic response conalstent with the accuracy and completeness of the model and equations employed. prior to the introduction of computers,

. [

a.

. #7% ~34-Nd o Frankh.n Re m.e,e.e.rch

.- Centet

=9=,

m y *g gu -

1 e '

ff l,s I

n i

I F-l I l I I i.

I I i 1 i i

_ _ __ _ . _ __ _ _ I g i I

in & la & 13 W la N& Is & Is & lr & le 9 fr '& la '4* *N IL I

., K K2 K3 Ka Ks Ks g n, Ks K, Ko I l y

i

- _ _ _J _ l I

+ a2 + 0s N es -

Fly-s wheel I l

< Crankshaft inertias and Stiffnesses 5 I I Generator I I Rotor I I L__J Electnce:

L:.act Figcre 2. Typical Lumped-Parameter 'rbraional Mathematical Model a

pt

?

a un St i

2:

TER-C5506-426 however, there was no feasible method of solving the equations. Even now, with the use of digital computers, a direct solution of the equations tends to consume more computer time than the modified direct solution methods discussed below.

The second course of action is a modification of the direct solution method wherein the solution procedure is modified to introduce certain types of expected steady-state harmonic response and the solution is made to determine the amplitudos and phace reisticnships of the harmonic responses.

This often called the mode superposition method (30), because the normal modes of vibration of the system are determined, each mode is subjected to the excitation, and the resulting responses of the separate modes are superimposed toyieldanoverallsolutiontothesteady-stateproyleA. In general, this is a little less comprehensive but also less computer-int ansive than the direct solution. Damping (energy loss) during vibration is not handled well by this method and approximate values for each vibratory mode must be substituted.

Ilowever, the method is highly recommended for application to diesel-generator i torsional vibration away from an actual resonant point. FaAA used this method I

l for its analytical (computer) model of the Shoreham diesel torsional dynamic

! response.

Electric-analog methods are powerful and ef fective but very time l consuming to set up and use. The method has largely been supplanted by the digital computer which of fers greater versatility.

l Since the turn of the century, various numerical tabulation techniques have been developed, of which flo1:er's method remains probably the best known today. These earlier methods are important because they form the foundation of most diesel engine torsional dynamien programs in use today for engine -

design. Further discussion of Holzer's method is presented in the next section.

4,1.2.2 llistorical 01esel Industry Analysis Method Tabular machods, such as Holzer's, have become the key analysis used by the diesel engine indurtry to determine the amplitude of torsional vibrations

., -e(Im aFrankhn Uh

n. .. w..w Res.ea.rth Cen.ter

i O' .

TER-C5506-426 excited by engine operation parameters and load variations (31). Although tabular techniques are very limited compared to computerised direct solutions, the industry had little else to use throughout most of the years of diesal engine development. Dusing the 1930s and 1940s, the industry sofined the tabular methods for the estimation of limited ranges of dynamic response. The trend was so well established that, when introduced, the digital computer was used to carry out, and to further develop, the familiar tabular techniques.

Perhaps this occurred because the computer was able to consolidate the basic analysis with certain of the classification rules (5) into one computer program. Althoulh engine manufacturers developed their own computer programs, et least one program (TORVAP by Structural Dynamics Research Corporation),

appearing to be based upon Lloyds Register rules (4l, was available between 1973 and 1978 on widely used computer time-sharing facilities.

In conjunction with the tabular methods that were developed for dynamic response analysis, Fourier series methods were developed to characterise the cylinder firing excitation and to provido a means by which it could be input meanim) fully to the dynamic response analysis. To do this, a curve of cylinder pressure versus crank amjle was employed with analysis of the crank positions to develop a diagram of tangential effort as shown in Figure 3. The tangential ef fort curve connects points of the instantaneous torque, normalised with respect to (divided by) crank radius and piston area. The effort curve is represented by the Fourier series as follows:

T = Tu + 1 (An sin ne + Bn cos no) or T = TM + 1 Tn sin (ne + $n) where Tm a constant mean tangential effort Tn a resultant coefficient of the nth order of torque excitation Tn a (An + Bn i !

o a crank angle n = order number, or the number 01 cycles of excitation of tha.

component per revolution of the engine.

o =37=

di3 awn rtenun r Re.w,erch

-. Center

> . . 1 i

o i ;

I Tangential Elfort Curve TER.C5506-426 j l i

- 4 l cl o

l i

x

, ~h hhh ,_

b- f cycle h! Tu ==2nb y ;

'N.J_ _ _._. . ._ .

9 180* ' 360' 540' 720' Crank angle .

, 3, 1 working cycle = 2 en0lne revolu'lons ,

3 , :

i i (a) ?,, -

e \,

l l

l 6 l

' I I !

-T,0. .

i n e= % .t /-

i i N /

- I O ..

l th,k -

l

'___.._ /\

2 i V V l I'

2W _ _._ b __. _ .-

I V ,

I l 3

3%

l 4 %%ew. ,W' 4% r ' ?AA l 5(N4 WM' '

S% eV l

! i 6 W ^'-y -

(b)

Flegure 3. Tangential effort Otagree and Normonic Components (from Reference 33l A

r *)S* '

[M$e i+ 4Reeeerth Ceneet FrenMan a teneum af the Poem tusense

i t

TER-C5504-426

( Table 5. Accepted Values of Tn from Classical Sources j' ;

! (Tn for mean effective pressure = 225 pett las pressure only) t i

i Lloyds i

l Porter [34) Nestorides (33) Net Wilson (32) Register (4) '

Ordei (1943) (1958) (1943) (1972) t 0.5 73.5 74.0 77.0 80.0 ;

1.0 79.8 90.0 79.0 88.0 t

1." 49.5 79.0 75.0 83.0 t 2.0 59.5 64.0 64.0 69.0 !

2.5 44.0 55.0 55.0 57.5 3.0 42.0 44.0 43.0 47.5 3.5 35.3 33.0 32.0 38.5 !

4.0 24.7 25.0 25.0 30.5 !

4.5 22.8 20.0 19.0, -

23.6 5.0 18 5 15.3 15.0 14.0 ;

5.5 15.1 12.0 11.0 13.4 6.0 12.2 f

. 9.5 8.9 10.5 i 6.5 9.4 7.3 7.3 8.5

7.0 7.3 4.0 4.0 4.0 I 7.5 4.4 4.9 4.4 5.5 0.0 4.7 4.1 4.0 4.4 8.5 3.5 3.1 3.4 3.7 '

j 9.0 2.6 2.7 2.6 3.2 1

i l 9.5 2.0 --

2.4 2.8

, 10.0 1.6 --

2.1 2.5 !

10.5 1.1 --

1.0 2.3 11.0 0.7 --

1.6 2.0 '

11.5 0.34 --

1.4 1.8

(

12.0 0.25 --

1.2 1.4 l

l r

4()jy!{i.a .Frenten.Reee.s.t.e.h

~.w - Center ,

j

\

i TRA-CSSO4-426 {

l Please refer to References 32 and 33 for full discussions of tangential l offort. Accepted values of Tn (4, 32, 33, 34, 35) are shown in Table S. For t order numbers through 9.0, for which all values are shown, the values cover a I span of 29 years and many diesel designs, including turbocharging (all are ,

4-stroke cycle). Se values are very consistent.

It should be noted that the values of Tn in Table S represent cylinder pressure only. A full solution meet include the additions to Tn contributed by the following effects (32):

o inertia forces of reciprocating parts '

o dead weight of reciprocating parts o inertia of the connecting rod.

In the analysis for the amplitude of dynaalc response, the Fourier coefficients, Tn, are the root values of torque oscitation magnitude at a given frequency. Because of the amount of tabulation necessary, the analysis may be carried out for only those orders that appear to yield higher response amplitudes. In any event, superposition of evente, including the full phase relationship between orders of escitation and vibratory modes, is not generally attempted.

Where superposition is not attempted, the amplitude responses for each order are plotted as shown in Figure 1 (also see Tot critical speed analysis in Appendix A). Total amplitude of response at any particular operating speed is taken as the sus of the individual responses at that speed. Sese values are converted to shaft torque and to shaft stress by the usual procedures.

4.2.3 noview of Tot 'Ibraional critical aseed Anaiveis The Tot torsional critical speed analysis included in this report as .

Appendia A was reviewed and found to contain substantial deviations from the magnitude of the dynamic emottation employed, when compared with consistent accepted values, that is, the published values of Tn discussed above. This and other signifloant points are discussed in the following sections.

4

-f M)Dj% -

dr { g y g h { enter

o .

TER-C5506-426 4.2.3.1 TDI Mathematical Model TDI formulated a mathematical model, similar to that shown in Figure 2, for the calculation of natural frequencies and modes ** vibration. The calculated natural frequencies were First mode 2130 cpm Second mode 5455 cpm Third mode 6495 Ops These f requencies, with respect to the 4th, 4.Sth, and 5th orders, yielded critical speeds of 532, 473, and 425 rpm, respectively, as those nearest to the operaLinq speed of 450 rpm.

The TDI mathematical model did not include the load inertia or its coupling to the rest of the mathematical model through the rotor-stator electrical interaction. Although this has been proven to be of minor significance by the torsional testing of the engine, it prevented the recognition of an approximate 3.0-Ha natural frequency when the generator was connected to the electrical power grid as discussed in Section 3.2.3.2.

4.2.3.2 TDI Dynamic Response Analysis The results of TDI's dynamic response analysis are shown in Appendix A by the chart of response curves for each signficant order. TD1's response analysis summary is included in Appendix A and was performed in accordance with the historic analysis method discussed in Section 4.2.2.2 of this report.

, In the course of this review, TD1's use of the Tn values was studied with considerable interest. Firste TDI used a dif ferent set of Tn values for each of the three mode shapes calculated ( Appendix A), but, most significantly, tdt's values are about half the magnitude of the values shown in Table 5 which liste published Tn values. As a consequence, the stress in the crankshaf t.as calculated by TDI was about half the value calculated from dynsmic response analyses employing the published Tn values.

The TDI values of Tn should not, however, be compared directly to those of Table 5 because the values in Table 5 represent only cylinder pressure. Tn g.s bddatoFrankhn Renee.rth Center

.eNr. aaww

t l

4 i

t TER-CSS04-424 :

values for diesel engine dynamic response analysis should include att t

j contributions as discussed in Section 4.2.2.2. Because TOI neither described its source for Tn values nor defined the contributors included, it is not l possible to discern tdt's technical contene.. The source and technical i j justification of Tot's Tn values were requested but have not been provided as :

of this report. A review of the contributions from sources other than i cylinder pressure indicates that the added value is ses11 for the most part, '

i l encept for the lower orders. For esemple, piston reciprocating mass provides t 2'

the largest added value to Tn, =2.445 for the 4th order, 0.0 for the 4.5th order, and +0.404 for the 5th order. These values, added to or subtracted [

j from the historic values, still provide about double the values used by TOI i for the critical orders. "

i i TDI has recently made available its range of Tn values used for diesel l engine design over the past 10 years. These are shown in Table 4 for i 3

comparison with Tot's design values for the shoreham diesels and are taken directly from TDE's report (35)*. TOI, hwever, still did not disclose the source or content of these values.

i 4.2.3.3 Comparison of TOI Crankshatt Stresses to the DEMA Rules i

The discussion of the DW44 rules (2) in Section 4.2.1.1 of this report I shows that the stress criterion for a single order of torsional vibration is 1 *

, 5000 pai, with 7000 psi serving as the criterion for the summation of orders !>

j that can coincide periodically.

[

In Table 7, the 'old' stresses follow from TOI's original torsional '

critical speed analysis (Appendia A), whereas the "new' stresses reflect the t

use of To!'s latest set of Tn values in that same analysis (11-inch -

] crankptn). These stresses were calculated following Tot's analysis (Appendia l Al during this review.

, Comparing the 'old' stresses to the D9th rules, it is seen that the largest stress for a single dynamic order is 2142 pai, well under the DONA

single order limit of 5000 pai, and that the summation of the most significant orders remains well under 7000 pai.

t fa, UdW henhhn Research Cevset a en a.# me ressma manae

- - - . _ - ~ _ _ _ . - - - - - - - - _ - . ~ . ,

TER-C5506-426 t-Table 4. TDI Hermonto Coefficlents (from Aeterence 35) i Year 1974-1975 1975 1975-1977 1977-----

Listing From LILCO Cp6 L MP6 L Stride Narmonto .Grous 1 Grous 2 M Groun 4 0.5 11.00 90.88 97.00 155.45 1 20.62 89.78 94.34 94.21 1.5 19.00 94.84 100.70 129.21 ,

2 24.06 45.43 42.53 42.61 l

2.5 20.20 62.38 65.10 71.51 3 19.97 14.84 14.57 16.52 1 3.5 16.70 34.91 40.61 42.72 4 13.30 29.04 30.25 27.62 l 4.5 9.05 12.48 12.73 12.72 5 7.30 9.21 9.39 9.34 5.5 5.65 7.01 7.14 7.14

l 6 4.18 5.55 5.68 5.40 6.5 3.29 4.39 4.49 4.49 1 2.66 3.40 3.49 3.68 7.5 2.23 2.90 3.05 3.04 8 1.87 2.46 2.52 2.52 j 8.5 1.61 2.20 2.24 2.26 9 1.42 1.92 1.97 1.97 i

9.5 1.25 1.50 1.53 1.52 l.

10 1.11 1.25 1.27 1.27 10.5 1.00 1.13 1.14 1.14 11 0.91 1.01 1.02 1.01 11.5 0.82 0.88 0.49 0.89 12 0.74 0.78 0.79 0.79

/u ~43-(!dd reenWm a tm Reee.s.tch Conw w N r. wi kieme

i -

L t

TEA-C5504-426 f l

I I

r l

1 i

l l

Table 7. TDI Stresses for DE M Rules (ll-inch Crankpin)

Old New !

Oynamic Dynamic (

Old Strese New Stress Selected ElitL in ,,,[gg1L, Tn insil Orders.

I 1.5 19.00 320 129.2 2176 2176 l l

2.5 20.20 425 71.5 1504 3.5 16.70 241 42.7 718 4.0 13.30 2542* 27.62 5342* 5342 -

4.5 9.85 ,, lit 11.72 1020 3 4398 E 10,700 7534

' values used for cooperison with single order recommended limits.

l

. e d',- -44=

LMFrenhhn eh Ceret a w w me - -

I l

TRA-C5506-426 !

i For "new" stresses, the 5362-pet stress for the 4.0th order is beyond the '

single order critetton, and the sumsstion of only two orders exceeds the 7000-poi criterion. Carrying the summation further, Table 7 indicates that, i for orders 1.5, 2.5, 3.S, 4.0, and 4.5, which can coincide periodically but at a lower frequency, the sua of the stresses is double the DEMA criterion.

l A check of well-known literature would have indicated that the Tn values l should have been questioned. Also, the use of rules such as Lloyds Register (4) would have brought about an automatic check of the Tn values.

t l

4.2.3.4 Comparison of tdt's Crankshaft Design to the A38 Rules Seleu6ed paragraphs from the 1940 Autos for Building and Classing Steet Vessels (3) by the ABS are provided in Appendix 8. The TD1 13 M 11 crankshaft l 1s compared with applicable paragraphs, and commentary about the 1973 rules is provided where ditterences are apparent.

ANS Peraarash 14.3.4. Torsional Vibration Stresses ._

This Ana rule requires submittal of calculations including tables of l natural frequencies and vector summations for critical speeds of all orders up l to 120% of rated speed, and stress estimates for criticals whose severity f approaches or onceeds the linits in Ass paragraph 34.57 and Ass Table 34.3.

l An3 Table 34.3 timits the stress for a single harmonic (order) to the following values

, ANS Ocade 2: 1 2134 pet As8 Grade 3e e 2490 pet ANS Grade 41 1 2679 Pet Am3 paragraph 34.57.1 establishes that, for designe differing from previous installations, stresses at single herannic critical speeds are not to enceed the above Italting stresses. Ata Paragraph 34.57.1 also indicates that total vibratory stress in the interval of 904 to 1054 of rated speed is not to caceed 1904 of the above stresses. During this revloe, it was assumed that the method of totaling is not fined and that stresses due to all significant orders of vibration seay be reasonably totated secording to appropriate vector suma if the orders do not coincide periodleetly.

4,1._, =45-ljdd at=

Fren.Wm peeea,e.h e N. r.w. wwwe Cenw

l TSR-CSS04-424 '

l '

Note that the Aas limite for cyckte strees are approminately half the l l Denn [2) eecommended values of 5000 pei and 7000 pe1, respeuttwoly.

i Aaf Paraaraoh 34.17.1, Min)ppe Diappter of Cran 4 1ae and Journals f The formulas eenetituting the Ass analyste nothod shown in Appendia a l were used with the substition of the following values from the diesel -

generator design to calou14te the minimum crankpin journal diameter. - -

t O = 17.0 inches, diameter of cylinder bore .

P* = 1774 pet, estimated meatmum cylinder pressure at 3900 kW L = 14.5 inches, span between main 1, earings N = 3900 kW/0.744 W per hp = S320 hp ,

M = 450 rpm C

1.00, engine with more than 4 cylindere F = 2140, Grade 3 forgings (a constanti see Appendia 5)

M' = 0.131 70 3L = 1.100 x 100 (formula from Appendia B)

T = 43,000 N/A = 0.7315 a 10 (formula free Appendin B) d = 11.12 inches, sinteen oranspin or journal diameter !

(see formula, Appendia 0) i Note: 00 101, 00 102, and OG 103 had 11.00-in erankpine.

These calculations for minimum creakpin diameter under the Aas rulee show that, based on Aas Grade 3 (for which constant F = 3140), the required ninteum cranhpin diameter to d = 11.13 in. Per Asa Grade 3 meterial, the diesel generator crankshafte with orankpin diameters of 11.00 inches would be l

l underdeelyned. However, the actual erankshaft notestal has physteet .

properties more nearly equal to Asa Grade 4 and, eseept for one reduction of I

area measurement reported by FaAA (Table 4), could qualify as an Asa Grade 4 ,

meterla1. Thus, using the meterial constant, F = 3310, for Asa Grade 4 l

Cylinder preneures were not avellable. Newever, an infosmal telephone i commonleation with Dr. Johnston, PaAA, on February 7, 1994 indleated that peak eylinder pressere wee 1400 poi at the Slee-hW loed. Therefore, a asalmue pean pressure of 1774 pet at 3900 W wee computed as the equate root of the ratio of the power levels.

00

. Wee *. W

TER-C5506-426 l

material in the above calculation yields a minimum allowable crankpin diameter of d = 10.84 in. 'Ihis is less than the ll.00-in actual crankpin diameter. In summary, if full credit were taken for the actual crankshaft material properties, the ll.00-in crankpin diameter of the diesel generators just meets 4

the minimum ABS crankpin requirements as shown by calculations performed as a part of this review.

ABS Paragraph 34.17.4, Solid Crankshafc Web Dimensions In order to provide adequate bending stiffness in the web, ABS requires tnat the web dimensions satisfy the following inequality:

wt > 0.35 d d

where w = 21.0 inches, width of web t = 4.5 inches, web thickness j d = 11.00 inches, crankpin diameter..

Thus, (21) ( 4. 5) > 0.35 (11.0) 425 > 42.4 and the inequality appears to be satisfied. (The values for w and t were :.

acquired informally by a telephone conversation with Dr. Wells, FaAA, on.

February 9, 1984, and are assumed to be sufficiently accurate for this \'

, m .,

a s calculat. ion. ) -

' %. , - * ;,[ \

Summary of ABS Rules Application * ' '

I s

\- T ,s ,', :s Comparison of the TDI 13 x 11 crankshaf t design with the ABS rules s

indicated that the crankshaf t geometrical proportions were (thin the ABS ' 't( ,

3, . ,

rules, but the dynamic stresses in the crankpin were'not. Thus, the ABS rules '

are significantly note conservative with respect to' harmonically induc'ed ( -

-3 . .-;* .s .

' t, stress than the DDIA rules, or about half the DEMA recosenended limitsi'd Again,

4 . %

l.

this reflects the conservative design believed to beirequired for-safstymat '

l x- a Q, >

.b i sea, and probably is derived from the culmination of long-term d4perience'in '

}

^

i N 16 ;4 .s

. that industry.

l ,n M C G^

~

3 y. UN < g ~ Y '* .

, f, ^

, 'h;

q. .. 3 p m

.e :l ,.y3 ,

g

_47 .c - 4 0 0IfFranklin Research Center

~

- 9 ii' A Dmseca of The Franhan Inseawee i

. kA s '

. , ,_ N ~ W . -

?&

TEP.-C5506-426 4.2.4 Summary of Crankshaft Design Review The review of this section on crankshaft methodology and design may be summarized by the following statements:

J o The DEMA rules [2] a.te not design specifications and standards.

Supplementary specifications and standards are required.

o It is advisable to employ the more comprehensive direct or modified direct solution of the mathematical model equations for torsional dynamics. With the present development of computer methods and _

accessibility of computer systems, the direct solution methods are not more labor intensive than the present computerized tabular mbtnods and do provido more comprehensive design assistance.

o TDI used Tn values for torsional excitation that are very low compared to values recognized in the industry since at least 1942 (36].

o The TDI crankshaft (11 x 13) does not meet the DEMA or ABS rules for dynamic stress when the revised TDI values of Tn are employed.

4.3 REVIEW OF CRANKSHAFT DYNAMIC TESTING Dynamic testing of the crankshaf t is regarded in this review as the essential element of the failure investigation because it is only through carefully conducted measurements that t's actual engine dynamics and local component stresses are confirmed. Accordingly, great attention was paid to each aspect of the test program.

Dynamic testing of DG 101 using an instrumented crankshaft was performed i

on September 20 and 28, 1983 at the Shoreham Nuclear Power Station. Reviews

of preparations and procedures and an account of test observations were e, reported previously [1]. -

Instrumentation for the measurement and recording of vital dynamic data included that are shown in Section 3.2.4.1 Since the completion of testing, the recorded data were r?duced and l

reported [37] by Stone and Webster, and the implications f or the crankshaf t

-t' a n ~48-Franklin Research Center -

A Dras.on of The Freamhn Inswwe

s .

TER-C5506-426 l

i l

failure investigation were reviewed and reported [29] by FaAA. This section is primarily a review and evaluation of the reported test data [37] and the failure investigation conclusions [29] that were reached.

4.3.1 Instrumentation, Signal Conditioning, and Data Recording Reference 37 provides a description and statement of applicability of transducers employed, including those for strain, torque, torsional shaft displacement, cylinder pressure, generator voltage and current, linear vibration, and the combination of crankshaft position and rotational speed. A table listing their pertinent characteristics and applicable ranges is also shown. The instrumentation was evaluated and its installation observed by the reviewer at the time of dynamic testing.

Por the most part, data output from the transducers was good. Earlier problems of strain gages and data transmitters on the rotating crankshaft were largely corrected before completion of testing on September 28, 1983, although the reported data [37] do include noisy, but apparently functioning, strain gage signals, e.g., on the No. 7 crankpin fillet. Also, the transducers for cylinder pressure seemed to function satisfactorily but appeared to provide pressure data lower in value than the actual pressure. The application of instrumentation in these environments is difficult and the experienced experimental test engineer anticipates certain aberrations in these data-channels. Indeed, the essence of the test engineer's. work is-to plan and conduct the test to maximize the good data extracted. Data from the strain gages on the crankshaf t were telemetered to nonrotating receivers and were conditioned and recorded along with the other data on a 14-channel, FM mode tape recorder. With proper planning of. signal channels prior to a test run, this afforded an opportunity to record simultaneous events on parallel channels. The signal conditioning and recording equipment are described in Reference 37..

The application of transducers, signal conditioning, and data recorders was reviewed and found to be satisfactory.

-4 % UddhFranklin Research Center

A Deseon of The Frankhn insutute

C TER-C5506-426 4

4.3.2 calibration Procedures Measured values'are not necessarily more accurate than analytical estimates; experimental measurements are only as accurate as the accuracy of their calibration, and then only if the proper instrument was chosen for the task.

4.3.2.1 Strain Gage and Tbrque Bridge Calibration Fillet strain gages and the torque bridge (employing strain gage) were calibrated by the shunt resistance method, wherein a precision resistor of -

known value is shunted in succession across the available arms of the bridge circuit.

Shunt resistance of the strain gages provides calibration not only of the

strain gages, but also of the conditioning circuitry and recording equipment.

However, it calibrates the gage only for measurement of surface strain in the

, metal on which the gage is located. This is sufficient calibration for the 1

1 crankpin fillet gages which were for the measurement of surf ace strain.

Calibration of the torque bridge, which used strain gages, required

additional procedures because the measured quantity was that of shaft torque and not strain at a point. Consequently, the test engineers employed static torque tests and test operation of the engine at zero electrical output to confirm the calibration of the torque bridge.

! The static torque test yielded measured torque plotted against applied mechanical torque as shown in Figures A-10 and A-ll of Reference 37. ,

Considerable hysteresis'is noted in these figures due to the friction in the engine and possibly due, in part, to strain gages that are not fully exercised ,

following their installation. Industry experience has shown that the 4

relationship would be such more linear in actual operation, where the bearing

surfaces would be operating on developed oil films to greatly reduce hysteresis due to friction in the engine, and the strain gages would become

" exercised" for greater linearity.

\

l The zero-output tests of the instrumented engine are' discussed in Section

A.2.2 of Refe'rence 37, which includes a table of values measured at four -

i l' -l OkFrankhn Research Center

' A Owoon af The Frankhn ineesute _

TER-C5506-426 electrical loads. The normalized values of "kW/1000 lb-f t" showed a spread of

+4% and -6% about an arithmetic mean value. Using linear regression, the mean

ratio of the measured values of "kW/1000 lb-ft" was calculated by Stone and

, webster to be 1.21. Although Reference 37 explains this to be the stress-

concentration in the shaf t on which the strain gages are mounted, evaluation during this review indicated that the actual stress concentration is on the-order of 1.16 and that the balance of the factor is due to the experimental measurement spread of the "kW/1000 lb-f t" values previously discussed.

Shifts in zero reference of the data recordings were investigated as a j part of the data analysis as discussed in Section A.3 of Reference 37. The 4

) overall error due to static strain ranged from 1.0 to 4.24. Thus, the static offset does af fect the calculation of principal stresses by a small' percentage l because these are based upon both the static and instantaneous cyclic stress.

l It should be noted, however, that the stress range of the cyclic stresses is not affected by this offset.

j' 4.3.2.2 Calibration of the Torsional Vibration Displacement Transducer The torsional vibration transducer is the unit attached to the gear case' l end of the crankshaf t for the direct measurement of vibrational amplitude.

1 Sections 3.2.2, 4.3, 6.3, and A.4 of Reference 37 describe the application and; calibration of this unit, wherein calibration is performed easily by means of fixed limits on displacement built into the unit.

A problem that arose with the use of the transducer was corrected during j data reduction. As described in Reference 37, an-internal filter selection I

switch remained set to a 10-Hz cutoff frequency. This attenuated all signal:

t components above 10 Hz. Data reduction procedures were developed to amplify the attenuated signal components in an effort to correct the error. . The -

l procedure was reviewed, and the results of the error-correcting efforts shown f

~in Reference 37 were evaluated and found to be satisfactory.

i n

4.3.2.3 Calibration of Accelerometers 1

l' Sections 3.2.5 and 4.5 of Reference 37 cover the application and calibra-tion-of accelerometers for linear-vibration measurement. The accelerometers +

F anklin Research Center

( ~

A Onomon of The Free 6n inessume

(

. . , - - - , ~ . - . - - . :.--, ..,.c-. - , - - . , - . . - , , . . .

TER-C5506-426 were calibrated with the use of a B&K Hodel 4291 calibrator, which could serve as a transfer standard from the National Bureau of Standards.

This review showed that any use of this transfer standard capability was not stated in Reference 37 and that the calibration source was not known.

Although these data were not necessary in forming a conclusion regarding the cause of failure, calibration of the accelerometers and other instrumentation should nevertheless be traceable to the National Bureau of Standards.

4.3.2.4 Calibration of Cylinder Pressure Instrumentation Sections 3.1.4, 3.2.3, 4.4, and 7.3.5 of Reference 37 describe the measurement of cylinder pressure and its calibration. Time history pressure measurement was attempted by means of precalibrated piezoelectric transducers installed in the compression test cocks of engine cylinders 5 and 7. Calibra-tion of the data signal circuitry between the transducer and the tape recorder was performed using the B&K Model 4291 calibrator mentioned previously.

The cylinder pressure measurements were unacceptably low. Efforts by Stone and Webster and FaAA following these tests concluded that the gas flow path geometry (see Figure A.4, Reference 37) was responsible. Accurate cylinder pressure measurement was not necessary in this test for conclusions regarding the cause of failure.

4.3.3 Review of the Experimental Data Dynamic tests of engine operation were run at zero-output load (variable ,

speed tests) and at loads of 100 kW, 1695 kW, 1706 kW, 1750 kW, 2250 kW, 2550 kW, and 3500 kW, with constant speed (450 rpm) operation. Data for these tests were reduced by Stone and Webster and are presented as charts in References 29 and 37.

The test data as presented '[24, 37] are dominated by presentations of torque and crankpin fillet strain. Torque, as presented in Figure 4-21 of )

Reference 29, is characterized by a 30-Hz oscillation of varying amplitude superimposed upon a steady-state value. Torque oscillatory amplitudes for l

l iff"[g.3UdN Franklin Research Center

% a n. n.a amau.

j . .

J TER-C5506-426 3500-kW operation reach a value of 1175,000 f t-lb (350,000 ft-lb, peak-to-peak torque range) superimposed upon a steady torque of 57,000 ft-lb. Note that this cyclic torque is a little over 3 times (6 times for peak-to-peak range) i the steady torque required to produce an electrical output of 3500 kW from the

- generator. This single amplitude ratio of 3 stands in contrast to the ABS rules [3] where the single amplitude dynamic component is expected to be on the order of the value of the steady-state component (power transmitted).

This is explained as follows. Refer to Section 4.2.3.4 of this report and I

note that the allowable crankpin single-order torsional stress, using the example of ABS Grade 4 steel, is 12679 pai. For the 1004 load rating of the diesel generators (3500-kW output), the engine torque at the flywheel shaf t j (torque bridge location) is 57,040 ft-lb, which yields a crankpin torsional shear stress of 2619 psi. This is very close to the limiting torsional stress k level allowed by the ABS rules. This example was calculated for the 1004' load rating of 3500 kW, the maximum load in the torsional dynamic tests performed.

For the intermittent 3900-kW diesel generator loads projected for actual' service, the. steady state, and cyclic stresses would be proportionally higher.

i

)

The engine firing rate is 30 Hz.

This engine firing rate is sufficiently- '

j. close to the first mode torsional natural frequency of 35.5 Hz to produce the l 'large dynamic response in the absence of significant damping. . '

Measured fillet strains on Crank No. 5 varied to a maximum peak-to-peak

~

range of.1800 microstrain (1800 x 10 inch / inch) as reported. for stain gage 5-1 in Figure 4-21 of Reference 29.' Table 6-2 of Reference.29 reports the major principal stress component of the measured strains to be 57,300 psi at

[ Crank No. 5, corresponding to a measured total peak positive torque of 230,000 l .

f t-lb (cyclic and steady-state) and negative torque of -153,000 ft-lb.

In the ~ absehce .,-

of direct access to the data and data reduction =instrumen-tation, observation of the diesel generator tests ~plus analytic investigations-I of the data reported in References 29 and 37 performed during this review provide basic. arrangement with the range and characteristics of torque and crankpin fillet stress' reported by References 29 and 37. Note that these are l

n 1:

measured values subjected to the measurement errors discussed previously.

L

[ . However, it does appear that these values are accurate to within~_+104.

I,c 1$$hdo . . -

P- 6 , Uu'JU Franklin Rmearch Center =

l}f j. A m.a .e the re.nemaineenues 3C 4- .. - . . , . _ ~. -,- -w...._. . , . . -- - . - _

i . .

1 l-

, TER-C5506-426 i

4 In addition to indicating high cyclic torques and stresses in the crankshaft, the test program yielded the following observations, with which l this review concurs:

o The rotor-stator electrical coupling within the generator which acted to couple the electrical load inertia to the engine dynamic model produced varying generator output current at 3.75 Hz when connected to the electric power grid, but did not contribute to the failure of the crankshafts.

, o Operation at 0.8, 0.9, and 1.0 power factors at the 2500-kW load range indicated that operation in this range of power factor did not

contribute to the crankshaft failure or dynamics of the system.

o The 30-Hz major dynamic response of the engine is not compounded by f j

any observed interaction with the electric loads, electric power grid, '

or plant loads. l I

J o The sudden initiation of plant loads was observed to cause a j smooth-orderly response of the engine and generator' and was not seen to cause cyclic fluctuations.

1 o connection of the diesel generator to the electric power grid was observed to be smooth and without significant transients, although it kl is realized that considerable care was taken at the time to make a proper connection. Connection of generators to the electric power grid without adequate synchronization can be damaging.

i 4.4 REVIEW OF FaAA DYNAMIC MODEL AND' CRANKSHAFT STRESS ANALYSIS 1

4.4.1 Dynamic Response Model In the course of the failure investigation, FaAA prepared and used a.

digital computer dynamic response model. From a discussion,* it was learned that the model is generally of the mode-superposition type discussed in Section 4.2.2.1 of this report.. Reference 29 indicates that the model-used -

the'same basic lumped-parameter (inertias'and spring constants) model as -

i formulated by TDI (Appendix A) with the addition of the rotor-stator equivalent spring constant and the electrical load inertia (see Figure 2 also) .

j.

Discussion with Dr. P. Johnston, FaAA, during test _ of DG 103 on January 7,

.1984, at the Shoreham Nuclear Power Station.

L7 anklin Research Center '

, A Onesson af The Fransen inessute

. ., i 4 TER-C5506-426

} FaAA's computer model output, as indicated by Figures 5-3 through 5-6 of Reference 29, has a remarkable similarity in character and amplitude to the values measured by the engine test. FaAA did not initially provide, in its report (29], the list of Tn values employed in its mathematical model. When it was suggested that the Tn values would be valuable for comparison to TDI's i

design values and to those from other published sources, FaAA made them available.* Table 8 includes FaAA's values with accepted values from Lloyds Register and Ker Wilson which were included here Isaa Table 5 to facilitate comparison. Comparison with values employed by TDI saa made using the TDI

} values of Table 6. In these comparisons, it was obverved that the FaAA values compared favorably with those of Lloyds Register and Ker Wilson. The FaAA values were more than twice TDI's design values (TDI 1974-1975 list in Table 6) in the critical range of orders 4.0, 4.5, and 5.0, and even greater for other orders. Thus, the Tn values for FaAA's mathematical model for which FaAA reported [29] excellent agreement.of computed dynamic response with that i exper'imentally measured further confirms the validity of published Tn. values over that employed by TDI for design.

t Even if FaAA's excitation had been prepared only to achieve the same dynamic response amplitudes as measured in the engine tests,-the model would have provided a highly useful interpolation function in portraying the dynamic action at points not available for measurement.

t I ~

As discussed in Section 4.2.2.1, computer models following f' rom the direct solution of the dynamic equations are very powerful in describing the full dynamics and interactions of a system. FaAA's computer model confirms this. The first task for the model was the prediction of the available cyclic-life of DG 101 remaining 'throughout the course of diesel generator testing on September 20 and 28, 1983. Here, initial dynamic response data measured 'at -

the beginning of each test session were ' introduced to the computer model. for comparison and prediction of the'available life cycles. remaining.

1 1 l

Telephone call from Dr. P. Johnston, FaAA, March.9, 1984. 1 m

'l

.c .

UbFranklin Research Center -

aon ona w re.n w =niw.

I l

l l

TER-C5506-426 r

Table 8. Comparison of FaAA's Tn Values with Those of Lloyds and Ker Wilson Vibration Lloyds Order FaAA* Register ** Ker Wilson ***

0.5 74.5 80.0 77.0 1.0 86.0 88.0 79.0 1.5 75.1 83.0 75.0 2.0 75.6 69.0 66.0 2.5 54.0 57.5 55.0 3.0 12.3 47.5 43.0 3.5 37.7 38.5 32.0 4.0 28.7 30.5 25.0 4.5 24.7 23.6 19.0 5.0 20.7 18.0 15.0 5.5 16.9 13.8 11.0 6.0 13.8 10.5 8.9 6.5 11.2 8.5 7.3 7.0 9.4 6.8 6.0

Calculated independently by FaAA. Includes effects of reciprocating masses.

- From Table 5. Not known what effects, such as reciprocating masses, are included.

- - Prom Table 5. Values for cylinder pressure only. -

f p . lj0bg Franklin Research Center A Onasica of The Fraren ineaewte

TER-C5506-426 4.4.2 Crankshaft Stress Analysis 4.4.2.1 Finite-Element Model FaAA formulated a finite-element structural mathematical model using three-dimensional, eight-node, isoparametric elements to represent one throw of the crankshaft. With application of torques from the dynamic response analysis, the model had the capability to indicate the highly stressed points in the complex crankshaf t geometry. Unless extremely fine element grids are employed, finite-element models generally underestimate the stress concentra-tion at local regions. Accordingly, FaAA used the same element distributions in an axisymmetric model of the same diameter and fillet radius so that the lack of stress concentration definition could be assessed by comparison to well-established values [39). The ratio of the established value and the finite-element stress concentration factor was used as a multiplier for the final stresses predicted in the fillet region by the crankshaft throw finite-element model. This was reviewed and found acceptable. The alternative method of using many more elements in the fillet would have been much more costly in both modeling and computer run time.

FaAA did not include a description of its method of torsional load application in its report [29]. However, when it was shown that the method of torsional load application employed by FaAA in the finite-element model was needed to complete the review of FaAA's crankshaft analysis, Dr. Wells (FaAA) provided a verbal description of the torsional loading method during a

, document review at the Shoreham Nuclear Power Station on March 8, 1984. The loading method was said to consist of a unit angular displacement applied to the journal end of the crank-throw finite-element model, plus a lateral displacement constraint applied to the side of the journal to represent the lateral constraint provided by the journal bearing. . The axial location of the lateral' constraint representing the journal-bearing reaction was said to have been varied to study its effect upon the computed stresses in the crankpin fillet. This effect was said to be relatively small. During the review of the crank-throw finite-element analysis and method of loading, it was noted that the unit angular displacement method of torsional load application along with the lateral displacement constraint to induce .the ' journal-bearing Y h b Frankhn Research Center

.. A Drnsson of The Frarden ineehde -

TER-C5506-426 reaction is a generally accepted method, which was deemed acceptable by this review.

4.4.2.2 Bending Stresses on pages 6-8 and 6-11 of Reference 29, FaAA discussed an investigation of bending stresses in the finite-element model due to an ef fective piston load at top-dead-center. When the associated bending stresses were indicated to be on the order of 4500 psi, as compared to approximately 40,000 psi for the torque load, the contribution of the connecting rod load in consideration of the fillet stresses was considered to be negligible, especially when the maximum fillet stresses occured when the crank was 130 degrees or so after top-dead-cente r .

Bending stresses, however, did appear to play a part in the stressing of the fillet as indicated by Figure B-100 in Reference 29. This bending action however, appeared to be local bending in the web and crankpin as part of the gross torsional loading. Consequently, it became a part of the stress concen-trating mechanism that caused the highly stressed region to develop at an angle of approximately 130 degrees from the 12 o' clock position on the crankpin.

4.4.2.3 Crankshaft Stress Analysis Summary The usefulness of a comprehensive stress model is readily apparent. The stresses predicted by the finite-element model appear to be in good agreement with experimentally measured values, even acknowledging the fact that the .

experimentally measured values contain an inherent error band of up to about

+10%.

Although the use of finite-element models for theoretical analysis, as well as for extending experimental investigations to regions not measurable, in to bc strongly encouraged, the validity of the failure investigation was considered during this review to be most relevant in the experimental measurement of crankshaft fillet stresses in actual engine operation.

Analytic techniques, such as the dynamic model and the finite-element crank-throw model, while quite powerful, were looked upon in this review as supplemental and confirming investigations.

t ha s"hd:FranWin d

m emv m .we Research Center 1

O O i

TER-C5506-426 4.5 REVIEW OF REPLACEMENT CRANKSHAFT DESIGN

[ Polloi #ing failure of the crankshaf t of the Shoreham diesel generators, the engine manufacturer, TDI recommended the use of an improved crankshaf t design, designated the 13 x 12 crankshaft. Whereas both the failed crankshaft (13 x 11) and the recommended replacements had 13-inch main journal diameters, the replacement crankshaft featured an increase in the crankpin diameter from 11 to 12 inches, as well as an increase in the crankpin fillet radius from 1

one-half to three-quarters of an inch. Analyses of the replacement crankshaft by FaAA [39] and TDI [40] are reviewed in this section of the report.

4.5.1 Review of Analysis by Transamerica Delaval, Inc (TDI)

I TDI used the same method of analysis as shown in Appendix A for the analysis of the original 13 x 11 crankshafts, with the exception that they substituted the Tn values shown in Group 4 of Table 6 of this report.- Here the Tn value for the 4th order is 27.62 as compared to the previous value of 13.30.

In summary, although the critical 4th order Tn excitation val'ue was doubled, the following considerations produced a reduction in the. calculated stress for comparison to the DEMA-recommended values:

o The larger crankpin permitted a 22% reduction in crankpin nominal torsional stress, o The increased natural frequency from 35.5 to 38.7 Hz reduced the dynamic magnifier for a 30-Hz excitation from.3.51 to 2.51.

This yielded a 4th order stress of 2990 psi as calculated by TDI for i

comparison to the DEMA recommendation of <5000 psi.

4.5.2 Review of FaAA Dynamic Response Analysis and Crankshaft Stress Analysis 4.5.2.1 Response Analysis FaAA employed its computer _ dynamic model using mode superposition to analyze the dynamic response with all significant modes considered. Inertia and spring constant elements for the model are shown in Table' 3-1 cf Reference

.39, and.the resulting natural frequencies are shown in Table.3-4 of that same

_nklin Rese_ arch

_ Center - .

,,, ..n n . . , ., - , .w-,

. o TER-C5506-426 reference. The dynamic response was computed by FaAA for " full load". The use of the term " full load" does not carry full definition since the design rating of the diesel generator is 3500 kW, but it is expected to operate at 3900 kW for short periods. For this review, 3500-kW generator output is inferred to be " full load".

Comparison of TDI and FaAA dynamic stress values to the DEMA recommenda-tions follows:

~

Average Torsional Average Torsional Stress (psi) Due Stress (psi) Due to Method of Analysis to 4th Order Summation of Orders TDI Analysis 2990 --

FaAA Modal Superposition 3300 5640 DEMA Recommendation <5000 <7000 Comparison of these stresses to those updated stresses for the 13 x 11 crankshaft, as shown in Table 7 of this report, indicates reductions in stress by a factor of 1.79.

Comparison of these stresses to the ABS rules, similar to'that shown in Section 4.2.3.4 of this report, indicates that the ABS rules may or may not be satisfied depending upon the interpretation that would be approved by the ABS following its review.

Assuming that an ABS Grade 4 steel was used for the crankshaft, the' ABS ,

allowable stress for a single harmonic is 2680 psi (see Section 4.2.3.4),

whereas the calculated stress (TDI) is 2990 psi. Thus, TDI's stress of 299r ,

psi and FaAA's stress of 3300 pai were both in' excess of the ABS allowable stress for a single harmonic using a nominal ABS Grade 4 material.

2The actual mechanical properties of the replacement crankshaft material, however, were shown by the quality control documents _at the Shoreham plant to be those provided in Table 9. Whereas Appendix B shows an ABS Grade 4 material' to have an ultimate' tensile of 83,000 pai, the minimum ultimate tensile

/Isa -

00$Frankhn Research Center A Omeson of The Franten measuae

)

l 4

l TER-C5506-426 Table 9. Properties of Replacement 13 x 12 Crankshafts Mechanical Properties J

Yield Ultimate Production Crankshaft Point Stress Elongation Area Brinnel Sample Number (psi) (ps i) (1) (t) Hardness Location 693 (DG 103) 58,310 100,360 25.0 54.1 205 --

59,470 106,460 24.0 58.9 212 --

694 (DG 102) 57,290 101,820 25.0 50.9 210 --

58,310 106,460 25.0 48.7 215 --

695 (DG 101) 52,650 100,800 24.0 50.9 205 Top 48,590 100,800 23.0 49.8 210 Botton Chemical Analysis Crankshaft C Si Mn P S Cr Al Number Heat (1) (1) (t) -(t) (t) (t) (t) 693 (DG 103) 821-487 0.50 0.05 0.70 0.006 0.010 0.63 0.003 694 (DG 102) 821-487 0.50 0.05 0.70 0.006 0.010 0.63 0.003 695 (DG-101) 811-167 0.46 0.12 0.65 0.010 0.008 0.69 --

123 JI bFranklin Research Center a w w n r,.,en m

TER-C5506-426 strength of the replacement crankshaf t materials as shown in Table 9 is 100,360 psi. To take full advantage of this material, an allowable value of 3090 psi for a single harmonic could be presented to ABS for approval in accordance with Note 4 of ABS Table 34.3 in Appendix B.

If full advantage of the material is to be taken, then it is also appropriate to use the full calculated dynamic response due to a single harmonic exciting factor. TDI's stress of 2990 psi was calculated using only the first mode response. Although TDI's analysis does show a small response for the second and third modes of torsional vibration, the second and third modes are seen to add very little to the first mode stress of 2990 pai. Thus, should the interpretation of the ABS rules discussed above be accepted by ABS, TDI's single harmonic stress would be within the ABS limits. However, FaAA's calculated stres of 3300 psi for a single harmonic excitation, based upon a somewhat higher value of Tn and upon greater modal participation, would not.

ABS also requires that the total vibratory stress from all harmonic excitation not exceed 150% of the allowable stress for a single harmonic exciting factor. For a nominal ABS Grade 4 material, this allowable stress is 4020 psi. For the interpretation of the ABS rules to use the full properties, the allowable stress is 4640 psi. TDI's total stresses cannot be compared to these ABS allowables because their analysis methods do not facilitate such summation of stresses. FaAA's calculated torsional strers for the summation of excitation orders is 5640 pai, which is well beyond even the interpreted ABS allowable stresses.

4.5.2.2 Crankshaft Stress Analysis FaAA used the finite-element method of analysis reviewed in Section 4.4.2 ~

of this report to compute the stress magnitude and distribution for the replacement crankshaft.

Stresses were reported to be reduced from the. previous cyclic principal stress range of 60,000 psi to a range of 37,000 psi. This constitutes a reduction by a factor of 1.78 to a cyclic range that is only 56% of the former cyclic range. The reduction was due to the larger crankpin and increased

).nNJ Franklin Research Center A Dnnsion of The Fraren insttute o

o .

TER-C5506-426 stiffneas with resulting increased natural frequencies as previously discussed, and was supplemented by the increase in the crankpin fillet radius from one-half to three-quarters of an inch. The analysis was considered to be acceptable.

4.5.3 Crankshaft Shotpeening FaAA reported [39] that shotpeening was introduced to the crankshaft processing in an effort to assure a " consistent, high level of compressive residual stress in the surface and to eliminate machining marks." The report continued by stating that the fillets "will be inspected by a high-resolution, eddy-current method after the break-in run."

Shotpeening has a long history of use in closing microscopic surface cracks and establishing a surface layer of the material in compressive stress.

Although the Jasic idea is good, it was noted during the review that while various levels of'shotpeening are available, no description'of the process was provided.

i Accordingly, the NRC arranged for a document review at the Shoreham Nuclear Power Station on March 8,1984, during which quality control documents pertaining to crankshaf t shotpeening were reviewed, and an informal discussion was held with Dr. Wells of FaAA. It was learned from Dr. Wells that two of the three replacement crankshaf ts, Nos. 693 (DG 103) and 694 (DG 102), arrived from TDI with the crankpin fillets already shotpeened.

The crankshafts were inspected and the results of the inspection are described by Stone and Webster Engineering Corporation's Coordination Report

, No. F-46109-G [41] as follows:

" Problem

Description:

Delaval has identified ' holidays' or lack of peen coverage in the fillet areas of new diesel crankshafts purchased in accordance with E&DCR F-46109-C. These ' holidays' have been dispositioned as functionally acceptable by TDI, however, recent analysis performed by Failure Analysis Associates indicates that 100% peening coverage is beneficial."

In conjunction with the review of documents on March 8,1984, photographs of the original shotpeening supplied by TDI were reviewed. Although the y.2 gSU Frankhn Research Center ~63- I U

am an r,-w-wam=

]

-. . . . _ . . - ~ - - - .- - ..- .. - . . ..- _ _~ _ _ _ .

4 TER-C5506-426 1

photographs did not provide the desired detail, the photographs gave an l 1

impression of surface texture more like grit blasting than shotpoening, i.e.,

the surface appeared to have been gauged by sharp particles instead of dimpled by round, smooth particles. Although the photographs provided only a limited

, view of the fillet surfaces, this evaluation of the initial shotpoening concurs with the results of the inspection [41) by Stone and Webster Engineering Corporation.

Stone and Webster's Coordination Report No. F-46109-G (41] provided a recommended solution as follows:

i " Problem Solution: Since the crankshafts are delivered to the site, Metal Improvement Company, a local firm with expertise in shotpoening will perform the rework. The fillet areas shall be repeoned in accordance with the requirements of MIL-S-131538 to assure 1006 coverage

of the fillet areas. Peening shall be performed by Metal Improvement Company on site and the crankshaf ts inspected by OQA for 1006 peening at l

) the fillet areas.' '

a Accordingly, LILCO Repair / Rework Request R/RR R43-1632 specified

{ shotpeening to include the following parameters:

} o Shot size; MI-550 l o Intensity, 0.008-0.010, Almen "C" test strips

! o MIL-S-131658, Amendment 2.

t j Quality control documents were reviewed .and indicated that the - Almen test strips for the repeening provided readings within the specified intensity of l.

! 0.008 to 0.010 inch (arc height) with the exception of one test strip which l was measured at 0.011 inch. -

4 ,

Photographs of the repeened surface wers reviewed and show an improvement

in'aurface texture, indicating an improvement in the quality of shotpoening of the crankpin fillets.

Crankshaft No. 695 for DG 101 was. received at the Shoreham plant direct from the supplier, Krupp-Stahl in Germany,.without shotpoening. ~ Crankshaft No. 695 was shotpoened at the Shoreham plant to the same specifications as those described for crankshaf t No. 693 and No. 694 above. Records reviewed at the Shoreham plant showed that the Almen test strips for crankshaf t No. 695 s

/% +: ~64-d$ Frankhn Research Center A Dnesson of The Frenteninsenute ,

o a 1

l l

TER-C5506-426 shotpeening indicated that the intensities remained within the specified range of 0.008 to 0.010 inch are height.

Shotpeening of this intensity is shown by Figure 4 to induce a compressive stress to a depth of from 0.027 to 0.034 inch, with the induced stress distributed as shown in Figure 5. Figures 4 and 5 are taken from Reference 42.

The purpose of shotpeening is to induce a compressive stress in the material at the surface of the crankpin fillets. Since the smooth surface la being disturbed by the particle impacts, it is necessary, once shotpeening is begun, to assure that the shotpeening coverage is uniform and of an intensity, with the right size of smooth shot, to achieve a suitable depth of material in compressive stress. Otherwise, improper shotpeening could serve as a source of added stress concentrations to make the crankshaf t more susceptible to fatigue.

The actual peened surface were not available for inspection in the course of this reviews therefore, this evaluation was made using the specified parameters, recorded Almen test strip measurements, and photographs of the peened surface. The shotpoening performed at the Shoreham plant is acceptable for the new cetrkshaf t (No. 695) not subjected to shotpoening in advance and will serve to increase the fatigue life of the crankpin fillets. Inspection of crankshaf t Nos. 693 and 694 revealed inadequate initial shotpoenings for these crankshafts, the rework shotpoening discussed above would be sufficient to counter the undesirable effects of the previous shotpeening, provided that the shotpeened surfaces that were photographed and made available for this review were representative of all crankpin fillet shotpeening. With this provision, the rework shotpoening is acceptable.

As an alternative to shotpoening, a surface layer under compressive Ctress can be induced into crankpin fillets by rolling techniques. This is accomplished by pressing a rolling element against the fillet surface with oufficient force to produce stresses in the fillet surface material that are just beyond the yield point. With the proper design of rolling element, the distribution of induced compressive stresses can be controlled to an ideal

m Ub Frankhn Research Center A Dews.on af The Frankha W.showie

1.0 TER-C5506-426 l

Rc31 0.035 /

2 e

f

=

0.030

{

0.75 t

5 g 0.025 / g l ' 5 0.020 > 0.5 e / 5 l M E 0.015 <

/ 3 0010 I 0.25

$ f f a

0.005

//

l 0

0 0 0.002 0.004 0.006 0.008 0.010 C Intensity Figure 4. Depth and Compressive Stress vs. Almen Intensity for Steel (from Reference 42]

l l

l PEENED SURFACE l

TENSION 0 COMPRESSION l

I t

C i

AESIDUAL STRESS C I'

C-T i

l Figure 5. Distribution of Stress in Shotpoened Beam with No External Ioad (from Reference 42)

/ ._, h0$ FranWin P.esearch Center A Cheem (A The Frenhan Wmte

e s z.

4 i- 1 l

TER-C5506-426

>. r profile of magnitude and depth in addition to providing'a smooth fillet

} !aca for optimum fatigue resistance. Fillet rolling provides many advantages: however, there are fillet geometeries for which it is dif ficult to design a coller, e.g., recessed fillets similar to those of the TDI crankshafts. In addition, the technique requires the proper machinery to .

1 i

hold, load, and rotate the crankshaft and roller. Where this technique is '

possible, benefits follow. Lacking the means,' shotpoening is rec amended.

4.5.4 _Surmary of Replacement Crankshaft Design j 1he increase in crankpin diameter from 11.00 to 12.00 inches provided a significant crankpin stress reduction by reducing the direct torsional stress :

1 i l

in t*e crankpin due to larger cross section and by stiffening the shaft to l produce a higher natural frequency and thus reduce the dynamic multiplication <

factor.

! Stresses calculated by TDI and FaAA were within the DDtA [2] recommenda-l tions for a single harmonic excitation. FaAA's summation of stresses for all' j excitation orders was also within DDIA's recommended values. TDI's analysis did not permit comparison of total stresses with those recommended by DEMA.

f

} TDI's crankpin stress for single harmonic excitation does not satisfy the

} ABS limiting values [3] for ABS Grade 4 steel, except through an interpreta-l tion of the rules in which full advantage of the crankshaft material properties, I l is taken. Such interpretation would require study-and approval by ABS. TDI's i .

analysis did not permit the comparison of total stress due to summation of

, ' orders with the ABS allowable values. Crankpin torsional stresses calculated' ,

I by FaAA for both single harmonic excitation and summation of orders were in j ;

j excess of ABS allowable values, including the higher allowable values

determined by an interpretation of the ABS rules that used the full material properties of.the crankshaft material.

' Crankpin fillet shotpoening of the replacement orankshafts was evaluated through the review of documentation and photographs at the Shoreham plant.

1 i

l

/ .;f libf Franklin Research Center

.. . A Dnme.en of The henten Instause

'l

. . , . _,. _ . - . - - , .-. - . . , - - - - - . - . _ . _ , - _ _ _ _ . - ,--,w

O TER-C5506-426 1

i Crankshaf ts No. 693 (DG 103) and No. 694 (DG 102) were found to have been previously shotpeened by TDI. When inspection at the Shoreham plant indicated that the initial shotpeening was unsatisfactory, the crankpin fillets were repeened at the Shoreham plant. Crankshaf t No. 695 (DG 101), received direct from the supplier in Germany, was not initially shotpeened by TDI and was shotpeened only at the Shoreham plant. The crankshafts could not be inspected directly, and the shotpeening was evaluated only through the review of documentation and inspection of photographs of local regions. The shotpeening 4 and rework shotpeening performed at the Shoreham plant were found to be acceptable insofar as the photographs inspected are representative of all shotpeened surfaces.

It must be noted that all of the TDI and FaAA stresses reviewed herein

! pertain to the 3500-kW electrical output loading (100% design load) and not to the short-term 3900-kW load required by the Shoreham plant.

f.

i i I i

_ _a Obb Frankhn Research Center A Dews.on af The Feaanha inature

, _ . ~ _ _

. . . . . - ..- . ._ -. . . . - . - . - .-. ------ .~ . .. . . - ~_ .-- - -

, -s TER-C5506-426

+

i 5. CONCLUSIONS I

Based on the findings of the failure investigation reviewed herein, it is l 1

i concluded that !

1 I

o The crankshaf t of diesel generator 102 failed in high cycle fatigue. !

! o Suf ficient cause for the high cycle fatigue failure was crankshaf t

] design based upon exceptionally low values of cyclic torque excitation

(Tn) coupled with a natural frequency fairly close to the dominant excitation frequency.

, o The specified design standardr were not definitive and contributed to the failure by not providing esign review material by which the i design would have been evaluated and found to be in question prior to i p the diesel generator's application as safe shutdown equipment.

i With respect to the replacement crankshaf t design, it is concluded thatt o The combined static and dynamic effects of a 1.00-inch increase in

crankpin diameter from 11.00 to 12.00 inches serve to reduce the

crankshaft stresses calculaced by TDI and FaAA to within the DBMA l recommended values for single order excitation and for summation of

order excitation. -

?

I' o Although stresses from TDI's analysis for the replacement crankshaf t do not satisfy the ABS cules for a single harmonic using a nominal Grade 4 material, they would just meet an interpretation of the ABS rules for a single harmonic wherein the actual properties of the

] crankshaft material are used. However, such interpretation of the ABS i tules is subject to review and approval by ABS.

o TDI did not present an analysis by which their summation of stresses i from all orders can be compared to the ABS limiting value for that ;

condition.

I

o FaAA's crankshaf t analysis predicts higher dynamic stresses due to (1)

. the use of slightly larger amplitudes of excitation (Tn values) than

-those used by TDI and (2) the superposition of modes resulting from j the direct solution of the equations of crankshaf t dynamics.

Vibratory stresses computed by FaAA do not satisfy the ABS requirements for a single vibratory order or for the summation of i --'ers, even considering an interpretation of the ABS rules to fully i l use the mechanical properties of the crankshaf t steel..

i IWanun Research Center 5 A Onneson of The Frenten insaeues J

y -y T e- - g 1p-t- r 9 ~ v-*- i eyn- er- *- sy-pyw-wry-- a4.wr-W---i--=-r---*M C P--

4 TER-C5506-426 I

o All analysis of stresses performed by TDI and FaAA pertained to the 3500-kW full load condition and not to the 3900-kW short-term overload l required by the Shoreham plant.

o Crankpin shotpeening ot one crankshaft and rework shotpoening of two crankshafts performed at the Shoreham plant were found to be acceptable only insofar as the evaluation from documents and photographs of localized shotpeened areas is representative of all crankpin fillet areas.

From the broad evaluations performed in the course of this review, it is summarily concluded that a set of standards more definitive than DEMA's

" Standard Practices for Low and Medium Speed Stationary Diesel and Gas Engines" is required for diesel engines essential for sa'e shutdown of the Shoreham plant; that " Rules for Building and Classing Steel Vessels" by the American Bureau of Shipping is representative of definitive standards for safety at sea; and that, with the possible exception of TDI's stress for a single harmonic, the stresses evaluated in this review do not meet the requirements of the ABS standard.

/{a 00bIFranklin Research Center a %,an a no r-w w.

TER-C5506-426

6. RECOMMENDATIONS The following recommendations are oftered:

l o The application of a torsional vibration damper on the Shoreham diesel generators to reduce the present high amplitude of torsional vibration and the associated high amplitudes of cyclic crankshaf t stressa should be investigated. The higher torsiona? amplitude of the crankshaft is the face, or gear case, end which is available for damper attachment.

o Specifications and standards employed by diesel engine manufacturing groups and user groups in the United States and Europe shoiald be

^

evaluated & ta for the purpose.of compil standards and specifications for the procuremen%dn t of diesel appropriate generators set of for nuclear power stations so that these standards and specifications can also serve as acceptance criteria for design and performance review. Although this recommendation is made for the review of the crankshafts, the recommendation is also applicable to other engine components. !

I l

e g1 ~71' Ub Franklin Research Center i 4 % a m r,m mww.

. . i l

TER-C5506-426 }

7. REFERENCES 1

1. Franklin Research Center, Evaluation of Diesel Generator Failure at Shoreham Unit 1, Interim Report on Phase 1, Failure Cause Evaluation FRC Project 5506, Task 20426, November 18, 1983

2. Diesel Engine Manufacturers Association, Standard Practices for Low an'd Medium Speed Stationary Diesel and Gas Engines, Sixth Edition, New York, New York, 1972

3. American Bureau of Shipping, Rules for duilding and Classing Steel Vessels, New tork, New York, 1973 and 1980 Editions

4. Lloyds Register of Shipping, Rules and Regulations for the Classification of Shipst Part 5, Main and Auxiliary Machinery, Chapter 2, Oil Engines

5. Lloyds Register of Shipping, Guidance Notes on Torsional Vibration Characteristics of Main and Auxiliary Oil Engines, 1972

6. Master Plan, Crankshaft Failure Analysis / Recovery, Beergency Diesel Generator 102 Shoreham Nuclear Power Station August 23, 1983

7. P. J. Holden, Project Engineer Diesel-Generator Operational Review Program, Prepared for LILCO, Submitted to E. J. Youngling, LILCO July 14, 1983

8. S. Ahmed, FRC Informal Technical Communication to J. T. Beard, NRC

Subject:

Regarding LILCO Master Plan for Recovery from EDG 102 Crankshaft Failure September 9, 1983

9. LILCO Procedure for Crankshaf t Testing, Emergency Diesel Generator No.101 September 2, 1983

10. Revision, LILCO Procedure for Crankshaf t *asting, Emergency Diesel ,

Generator No. 101 September 9, 1983

11. Revision, LILCO Procedure for Crankshaf t Testing, Emergency Diesel Generator No. 101 September 15, 1983

12. Revision, LILCO Procedure for Crankshaf t Testing, Emergency Diesel Generator No. 101 September 23, 1983 gn hJO Frankl.n Research Center a % .e te r,.,wa =

. +

f i

TER-C5506-426 1 l

13. R. C. Herrick, Principal Engineer, FRC ;

Memo to C. Petrone, NRC Resident Inspector, Shoreham Nuclear Plant

Subject:

On-site for Torsion Test Monitoring '

September 24, 1983 l

l t

14. Delaval Engine and Compression Division Torsional and Lateral Critical Speed Analysis, Prepared for Long Island

[ Lighting Co.

August 1974 Revised March 18, 1975

15. LI140 Preliminary Response to NRC Request for Information I.3 1

Transmitted September 20, 1983

16. LILCO Response to NRC Request for Information 5 l

) 17. LILCO Response to NRC Request for Information II.1 {

Transmitted September 20, 1983 l

t 18 LILCO Response to NRC Request for Information II.2 Transmitted September 20, 1983

~! 19. J. T. Beard, NRC j

Memo to Gary M. Nolahan, Chief, Operating Reactors Assessment Branch, NRC i Subjects Recent Failure Experience of Transamerica Delaval, Inc.,

! Emergency Diesel Generators

] August 22, 1983 1

20. Stone and Webster Engineering Corporation, Specification for Diesel <

] Generator sets, moston, MA, October 3, 1973, page 9 4

I 1 21. Stone and Webster Engineering Corporation, Specification for Diesel j Generator Sets, Revision 2, Boston, MA, January 26, 1983, page 1-10 !

1 1 22. tang Island Lighting Company, Conference / Trip Report--Trip to Ellwood [

}-

City Forge Co., Ellwood City, PA, August 31 - September 1,1983 i

23. Stone and Webster Engineering Corporation, Crankshaft Quality Document l j File, control No. 140, Engine No. 2605 ;

24. Stone and Webster Engineering Corporation, Crankshaft Quality Document ;

I File, Control No. 141, Engine No. 2606

25. Stone and Webster Engineering Corporation, Crankshaft Quality Document File, control No. 258, Engine No. 2604

26. Mitsubishi Steel Manufacturing Company, Ltd., Japan, hesults of Testing and Inspection Crankshaft for Delaval Turbine, Inc., June 29, 1972 1

1 g ,, ,

lib Frankhn Reeeerth Center A Daseien ei the reensen t=sumuse

TER-C5506-42G

27. Ellwood City Forge Company, Crankshaft Test Report for Delaval Turbine / Enterprise Division, Elwood City Forge Shop Order No. 0-957, Ellwood City, PA, June 10, 1970

28. Failure Analysis Associates, Preliminary Metallurgical Evaluation of Failed Crankshaft from DG102, Palo Alto, CA, October 1983

29. Failure Analysis Associates, Emergency Diesel Generator Crankshaf t Failure Investigation, Shoreham Nuclear Power Station, Palo Alto, CA, October 31, 1983

30. S. Timoshenko, D. H. Young, and W. H. Weaver, Jr., Vibration Problems in Engineering, 4th Edition, John Wiley and Sons, New York, 1974

31. Ker Wilson, Practical Solution of Torsional Vibration Problems, Volume 1, Frequency Calculations, 3rd Edition, Revised, John Wiley and Sons, Inc.,

New York. 1956

32. Ker Wilson, Practical Solution of Torsional Vibration Problems, Volume 2, Amplitude Calculations, 3rd Edition, John Wiley and Sons, Inc., New York, 1963

33. E. J. Nestorides, Handbook of Torsional Vibration, Cambridge University Press, 1958

34. P. F. Porter, Harmonic Coef ficients of Engine Tocque Curves, Journal of Applied Mechanics, March 1943

35. Transamerica Delaval, Inc., Letter from C. S. Matthews, Vice President (TDI) to T. M. Novak, Assistant Director of Licensing, NRC, Transmitting TDI Response to Users Group Questions, December 16, 1983

36. R. S. Stansfleid, The Measurement of Torsional Vibrations, Proceedings of the Institution of Mechanical Engineers, London, England, Volume 147, No.

5 (1942)

37. E. Bercel and J. R. Hall, Field Test of Emergency Diesel Generator 101, Stone and Webster Engineering Corporation Report B1-1160-041-1, October 1983

30. R. E. Peterson, Stress Concentration Factors, John Wiley and Sons, New York, 1974

39. Failure Analysis Associates, Analysis of Replacement Crankshaf t for Emergency Diesel Generators, Shoreham Nuclear Power Station, Report FaAA-83-10-2, PA07396, Palo Alto, California, October 31, 1983

40. Transamerica Delaval, Inc., Proposed Torsional and Lateral Critical Speed Analysis for Engine Nos. 74010/12, Oakland, California, August 22, 1983

.., -g[l al Frankhn Research Center m m e ,.e.=

.,.g, _

A

y. 4 - . . - .-

, s ' '

x

, \

j ,

, TER-C5506 $2d ,

s s

1

41. Stone and Webster Engineering Corporation, Coohdinating Report No.

F-46109-G, dated September 16, 1983 i

42. Metal Improvement Company, Inc., Shotpeening Applications, 6th Edition, s 1980 -

, 1; f [

. w

, .In

. s .

1 i

n*

i

\

S t

f

.. ~2 0 j,n}h=

! Franklin Research Center A Dnaman of The Feardde N -

-.c . -- - .- .- - .

APPENDIX A TORSIONAL CRITICAL SPEED ANALYSIS BY TRANSAMERICA DELAVAL, INC.

J 8

l Franklin Research Center A Division of The Franklin Institute The Bengan Franklin Parh*ay Phala Pa 19103 (21Si 448 1u00

- .. .. .. _. _. , - - ~ . .

+ e 1

l I

J TOR $10NAL Arc LATERAL CRITICAL $ PEED ANALYSl$

ENGINE Nur'.8ER$ 74010/12 DELAVAL-Et.TERPRl5E ENGINE MODEL DSR le8 3500 Kv/4289 eMP At 450 APM FOR i STONE & WERSTER ENGil;EERING CORP.

LONG ISLAND LIGHTING COMPANY 1

a 1~

i, .

T i / * *W . ;#-&?. -

7/r-e .: ./;- . , .'..-

I ** t . [-hd - -

%, , . . . s,*;

- *p; i DELAVAL ENGINE'f, C07 PRES $0A OlVISION k rN-b',- '

, 550 - 85th AVENut N^ . ':

oAKL'APO ,'CAllFORNIA 94621 A/

. .(~

[

$l .l

qa*)

.p.z ?t% . - - -

2,v,s s.i.in} en Rot m ANa -

. Au sst 1

%\.

/-

I<ma hr.(c..

2NP

rest.

.p.,.;a, u T

IlT:'- c l'

l f

I t,

e i

ifmc _

tjuuu Franklin Research Center

. 31 ,

A Onamen of Tite Freden eneseuse

- _ _ _ = ' - - - _ = _' _ _ _ _ _ - __ . _ _ . -

4

.e s e

1

/

7s5-74 Jew /:s as ban wes G annay' h-ins aa .n ru:ss! 2:2 */. S -

?

y' ' JScoca (~fFe 4-?

5 dic AA'? g ;, s

. q

^

=0 a && b) n k

}

( ( C dmed du.-s >= 7+?cto/st N. *:

g 6- '

ip T,

' 9 4 A g ',

% % 9 i I I I i k t <

4 i I I

sim, .fe nt> c bnre.

OweC,Mnsv GU<' ff 44 Av'A me 515 Zeerva' 63.t o dt.; y /x ii 9 7o gro.orr 9.63 1/8 i4 nt/7' In 6 sos sn.rr toc'

&sec Mo. /

1 DAM /sT92

SE, i 4'iMr7 4/. P /

/6J Y 33 A.A /r

I e de 7 95

Chc de J - 7

% unnar 44.e 7

P'" M8 *)

1 $ Wo.ts . He Cw TJ 7. do Ocon WP. 4,9 qq OMr. Wr. 437pg M9 .st, a. s~r" Zo A/ 9 497

O'*w Me 8 03*HK /J92 3Z,,

<?wer 7s. e 3 .

166 4 t4 AC. /t' fr St.723 .

CAacarm" iCox.: 245.e8

~ ftrastra. 72e4l' 4 At 76Al. -

Gawweed'm Rs.e . .9J.co 18493 e8 4A f?' fo pp oo .as .

6sxersrec }%wt 8427C a.M' Z* 24S0 43b .

70 % (Jet's /JJ740 AA #7'

#., A-2 N Frankhn Research Center A Onee.on of the Frannen insenwee

e .

, . . e J .# ,

d. / * ]

,f:, .., ,u ru r /r ~--a h./ '

Ec~r Ganc s CrowAan Ha. / .cotyW de 44SM .oo

srst./ne a3rr nza Cyra~ aces cos3:J' Me Y/. 244 so' fran ska>

G'uwacc M 8 % A rv 8442

c ei26.C*go 70. 880 = o o ' FYea,/tes.

A raam. s sw-m i d <

w. sa. sa, af. _,,- Jr dr n' ) '$

pt

_ f. . '

l . , ,

I .} .I *,

'7 .__ . __. g.4._.g _a '9 _ e 4q i

, m

(

-u -

I ./

MOJ6 eI 4 a f *l r F I_oi" sl

.n.c si; * -

g, a..e s.s . ,i .v , .. ss . ,a - e . . . .c. , , , , , , _ _ ,, , . , , , , , , , , ,,,,,,,,,,,g,:

.** <*' <a- ,u,, ,

.. .su . .... a ea.. ..sist... . sea. ....a m - g,;u,,,,,,ee rr a,;

5%~ h.,n., a fo.n cio> < e. Cro-m M.. / 6-2(f Atrw Ca'vue.e4 // w.

(Yaw.%x avi.8 % /Ir w // **

A runar % Guam la oa.

[en'WW4EL We rna,r e 6Nf aA Gurewec A1 e A4ure i7sio ad.

St!,evar A **.ta:.

/ fateuswerti Aar dose .t/se 19 2 Estoss Moss .folts VA.4 17 pes Mm gws vs.,

l c=>vo .v.-e % % e / w ou, A4 wr Amtm. J / 9/ vres.

l H #$?, ./','l} *,, * / ;*s em ^ Ror.o -Q. i.;S4 .

'. A-3 d! ,!b. ;d~ Franklin Research Center 4o .e n r,

, .~ _

O i

{

&OJe !

omega squ-3rao in iraosans/secondl==2 = 0.04;74543 r atura l f r**cuency in v.p.m. = 2124.n4 no. Inertia theta tom 2t stoma m shaft k dthata l c.b I.00000 0.33V 0.339 54.6 0.00020 2 30.6 0.993o0 2.518 2.H50 71.2 0.04000 J 4V.5 0.95380 2.346 5.198 71.2 0.072Vc 4 .49.5 0.83064 2.169 7.'367 II.2 0.10340 5 49.5 0.77743 1.914 9.261 71.2 0.13027 o 49.5 0.04716 1.593 10.875 71.2 0.152c4 7 49.5 0.49452 1.210 12.092 11.2 0.169/S e 40.5 0.324eo 0.800 12.892 71.2 0.18073 9 o f. I 0.14334 0.370 13.202 70.9 0.18711 10 18v3.1 -0.04327 -2.3$3 10.895 270.8 0.03930 Il 26o0.4 -0.08263 -10.894 0.000 moue i omaca squarad 0.04974543 natura l f raquency /129.e4 s t ;n.a tatnetaesi 2438.6705 stJma i=tnata**2 2755.3442 '

f thi 1062e.20 .

T cAf 2c37.24 strPs$aJ d\dmPtaf of PXtpTO81 shaft 16.Q) coullibrium amelituce 0.00VVG$26 f in 74e7.33

f Int 10.59 f eut 2.o3 fe 152G2003.64 .

to 0.

f I ct O. fp f cs 0.

tc o ro e r. rpm O.

tn

./**vec tstint tstaxt phi tmaxi 'tmava ;

U.5 42'aV 11.03 0.719 83.1. /2.4' -0.223' 2J12. -633..

1.0 /12/ 20.02 0.149 32.0 S.7- 0.057 V24 247.

l.5 141V 19.00 1.4J2 2e9.2 10.9 0.166 6802. 2179 2.0 1004 24.03 0.3e4 97.6 26.8 0.261 '2760 739, 2.5 c51 20.20 1.432 308.4 -91.8 0.blo bo76. 2387 3.0 109 19.07 0.14 V - 31.6 d.4 0.0d4 894 .239

3. 5 e0H I6.70 0.789 127.1 '33.9 0.330 3301 961.

4.0 332 13.30 5.216 734.0 196.1 1.956 20307 5555.

4.5 413 9.e5 0.719 75.0 20.0 0.200 4124 507.

5.0 425 7.30 0.149 11.5 - 3.1 0.031' 327 07.

d.5 387 5.65 1.432 85.7 22.9 0.226 2427. 646.

o.0 354 4.18 0.304 17.0 4.5 0.045 481 128 o.5 J21' 3.2V l.432 49.9 13.3' O.IJ3 1413. 377.

1. 0 -304 2.66 0.149 4.2 1.1 0.011 *110 32.

7.5 283 2.23 0.719 17.0 4.5 0.045 : 4 51. - 128.

6.0 26$ l.87 -5.216 103.3 27.6 0.275 2026. 781. .

u. d ' 250 1.61 0.7.19 12. 3 3. 3 0.033 347., - 9 3.

0.6 0.006 9.0 234 1.42 0.149 2.2 64 17 9.5 224 1.25 1.432 19.0 5.1 - 0.051 537 143.

10.0 212 1. 11 0.Jod 4.5 1.2 0.012 129 . 34 10.5 202 1.00 1.432 15.2 4.0 0.040 429.- 115.

16.0 - 143 0.91 0.144 1.4 0.4 0.004 41 II. 4 16.5 IMS- C.d2 0.719 6.2 ..l.7 0.017 177.. 47.

12.0 #7/ L.74 t.2le 4 t. . v 10.4 0.109 1859.. 300.-

.x A-4

[Mllfl Frerddin Research Ca.m A Dn.euen af The Feennha Inseeuse

i 4 mocc 2 omena scuarea in (r,-J tans /seconcl **2 = 0.32c372V7 natural frecuency a n v . p . n.. = 5459.41 no. inertia theta tom 2t. slyma m Shaft k dthsta i 6.8 1.00000 2.221 2.221 54.6 0.04070 2 50.8 0.V5V30 . 15.904 18.125 71.2 0.25440 3 49.5 0.70440 11.3B7 29.512 71.2 0.41423 4 49.5 0.29007 4.096 34.206 71.2 0.48014 5 49.5 -0.86947 -3.001 31.I47 78.2 0.43716 o 49.5 -0.626c5 -10.123 21.024 71.2 0.29609 7 49.5 -0.V2114 -14.890 6.134 71.2 0.08609 8 49.5 -1.00763 -16.281 -10.147 71.2 -0.14242 9- 31.7 -0.60541 -14.609 -24.756 70.9 .-0.34027 10 1800.1 -0.58614 -185.30o -210.004 276.8 -0.75e08 11 2650.4 0.242b4 210.000 -0.004 I moae 2 oniega sou3 rad 0.32031297 nattsral f requancy 5455.41 519mn 1*tnata**2 2133.d4e4 s a gr.a tatnata**2 8207.9240 4 1 t i.: 27414.36 1 txi 34704.17 stretsea siameter of external shaft 10.00 coullibrium 9molitude 0.000J5398 f in 7427.JJ f int 1.40 .

f ext 2.7v f e l l 21816 7v. 00 t 3 0.

f cr O.

f es 0.

f o 0.

orcer rem ti vec tstint 'tstext phi teuxi teaxe 0.5 10v10 11.00 1.389 21.3 42.0 0.05d 6o02. JIV7 8.0 5455 22.51 U.43e- 13.6 27.5 0.03b 1034 2063.

l.5 363o 19.00 J.738 99.1 197.7 0.271 7433. 14R39, 2.0 2/27' 7.81 1.23e 13.5 27.0 0.031 1015. 2025.

2.5 2162 20.20 3.731 105.3 210.2 0.268 790c. 15776.

3.0 lol8 14.27 0.436 8.7 17.4 0.024 c55. 1309.

3.5 lo$e 1o.70 1.399 32.4 64.7 0. ors 9 2432. 4654 4.0 IJo3 13.3J I.050 30.6 61.4 0.084 2311. 4611.

4.5 1212 9.65 1.389 19.8 38.1 0.052 1435. 2b63.

5.0 109t' 7.30 0.43d 4.5 8.9 0.012 335. 669 o.5 v91 5.65 3.731 29.5 58.8- 0.0HI 2211. 4412.

6.0 vo? 4.18 1.238 -7.2 14.4 0.020 543. 1053.

o.5 039 3.29 3.731 17.2 34.2 0.047 1289. 2509.

7.0 779 2. 60 0.436 1.6 3.2 0.004' 122. 244 4 - 7.5 127 2.23 1.380 4.3 '8.6 0.012 325. 648.

o.0 081 f.87 1.056 4.3 8.6 -0.012 325. 648.

e.5 64I- 1.61 1.389 3.I c.2 0.009 235.- 468.

9.0 006 1.42 0.438 C.9 1.7 0.002 65. 130.

V.5 574 1.25- 3.731 c.5 13.0 0.01c dec. 976.

10.0 o45 . l. Il I.23u- 1.9- 3.8 0.005' 144 285 10.6 919 .1.00 3.731 5.2 10.4 0. 014 391._ 783.

81. 0 495 C.vl 0.43e 0.o 1.1 0.002 42. E3.

11.5 4 /4 0.b2' 'l.3c9 1. o - 3. 2 . ' O . '00 4 114 234 a2.u 454 c.vo I.c5c 2.2 4.5~ J.OJ . 1o7. 334

/ .' A L M Frenidin Research Center 4 w oe n. r . m m ,

+ s m0Je 3 omeues squarea in I ranians /secon3) **2 = C.4c2o4 30e natur*! frequency in v.p.n.. = e493.21 no. Inertia theta tom 2t s193a m shaft k dtheta i o.6 1.00003 ~3.146 3.148 54.6 0.05769 2 90.8 0.94231 22.143 25.293 7I.2 0.35501 3 49.5 0.56729 13.449 38.742 71.2 0.54378 4 49.5 0.04352 0.996 39.739 71.2 0.5s777 5- 49.5 -0.51425 -11.776 27.963 71.2 0.39246 o 49.5 -0.90o73 -20.764 1.1 99 71.2 0.10604 7 49.5 -l.03777 -23.077 -15.878 71.2 -0.2228I B 49.5 -0.764VO -17.974 -33.852 71.2 -0.47516 9 . 58.7 -0.30976 -7.412 -41.265 70.9 -0.5821e 10 1800.1 0.27242 138.642 97.378 276.8 0.35163 l 11 2030.4 -0.07941 -97.371 0.000 moae 3 omana sauerec 0.40264306 natural frecaency 0492.21 sauma t= theta *=2 2421.4 eld s : 2:a letneta==2 Joe 3.3370 I I t.I 33uov,H7 i nAl 2535t.65 stretteJ Jiamater of ex ternal shaf t Ic . 00 coullinriue'arrlitude 0.e0000013 f t r. 74e7.33 f int 2.65 t *xt 2.03

-t e 14C7/b74s.00 tu 0.

f er O.

f cs 0.

fa D.

oroer rrm tn vec tstint tstext . phi tmaxi tmere ,

0.d 12vod . I I . 03 0.bol 25.1 19.2 U.029 955. 732.. '

l.0 64V3 32.e5 0.94e 62.5 03.2 0.095 .3430. 2405.

1.5 4333 19.00 2.V4e 14e.3 113.7 0.171 5o40. 4325.

2.0 3247 e2.42 2.b20 615.9 4 72. 3 0.706 23423. 17461.

2.5 .23V5 20.20 2.946 197.7 123.9 0.181- 5997. 4505.

3.0 2107 17.28 0.948 43.2 33.I 0.050 1o43. 1200.

3.5 toS5 10.70 0.861 38.1 29.2.-U.044 1449. IlII.

4.0 1623 13.33. 1,000 - o8. 7 . '52.7 0.077 2014 2004 4.5 1443 v.65 0.801 22.5- 17.2. 0.026 855. 655. '

5.0 1299 7.30 C.94d - 16.3 ;4.1 0.021 697. 534 3.5 ~ 1880 5.oS 2.940' 44.1 33.8 0.051 1677. 1236.

o.O 1082 '4.13 2.820 31.2- 24.0 0.036. 1868. 911.

o.5 V99 3.29- 2.946 25.7 19.7 0.030 977. 749..

7.0 V27 2.co. 0.948 6.7 5.1 .0.008 254 195.

7.5' 'ooo 2.23 0.861 5.1 3. 9 0.000 1V4 140.-

6.0 d ll - 1.67 l.950 9.7-. ~ 7. 4 0.0 l l . Jo8. 2E2. i 6.5 764 .l.ol 0. col 3. 7 ' 2.8 0.034 140. 107.

9.0 12 8 1.42 .0.046- 3.6 2.7 0.004 136. 104 19.5 683 1.23 2.946' 9.8 7.5. 0.011 Jil, c65.

10.0 c49 1.11- 2.820 8.3 6.4- 0.010 ' 315. 242.

10.5 . old 1.00 2.946 7.d .o.0 0.0D9 - 297. 227 16.0. 500 0.46, 0.940 2.3 1. 6 0.003- bl. 67.

11.5 364 0.t 2 C.eei 1. v. l.4 -0.002- 7 1.- 55.

12.0 -9 41 0.s4 :l.9ow 3.5 2.4 s.3G4 145. 112.

ra A-6

-[N[:Franklin Research Center A Dmeson of The Frankhn ineamme -

+

& e

% 7 ts

.. . . -., ... . u-~-- %. N N. t

'~~*~

4 g

,.; r .

.. .. - ...... . . ~ .. .yf.g ._ -

J.

N N9 4

_ .. w .~. _.

YJfC//A O dt9E $

e T .\,i a .

s e r.

-~ .. .

} '.i:'.

p. ~ -

e -

. .,,qq -

.. .. __ e. ._a _. ' 4 l~, * . - -. . - - . -

n...

n Q .~.

.4 e. . . . . . . . . ~ . - . .

L

..y x .. > t W D, N . g$ !

.. _. . . a_._....-...- ..'

\yN k **

. . 3 .

g .

h 4, **4!

.e g g . ,.

s.

.%o .-

.. yd _ ., - _ . .

Q ks ,4

...-4.,.

\ . . . - . . ~

(. . - . .. -

. .3 N S.

4 9% %

i M % tm 4 g T $ 69

_1 QN N .{k

4) v y g . . . . . . . . . . . .

.. . .. g ..

. .. .k ls .q ~

.f .5 _9

..i .J .*. .. -..

,.4N. . . . . - .

-Aq ;; ,

.a -f . . .

4 e .

...._..g . .t .. t g __ _ . . _

Y -

.e t

. 'E.

Q...n

. ...-- .._ ...- . \. . . . .

_ y .

,. g .k .

- 4

-. .. p

.L,- . _ . . . ,-...a.- -._. . - -

p. . . . ..;...

- . , - : .:. . . .. 2:1 a

.-" " ' ' * * ~ * '

_a... '

.. : c.

y... , .

.~ '

.; Ei:+i: .1 : :i

-c . -

i ::ren .

-~~- -

..~

L. . .a.....

~

... .' . .:,Td R

. y .;

t . ..._:. , .

(lg, .*W 4 o

^

- .. -...n.ai4 .$ .pe ..

E

-- .. . =-

s:.:4 .

. nw]

t

-}.

, 12.; - A-7

' b. Franklin Research Center -

%on a m re.ama mau.

. o

, b i W " k' e

ivrw ll:31v0T 07/18/74 V . 3 74 900 lli+ 0 i left reaction * -16858.

ric% reaction = -12694 nelta dia weight inertia shear moment slope deft

a. O. O. O. 1685tl .1 O. 0.000127 O.

.312 13.000 143.2 1402.0 16714.9 $ 3 900. 0.000124 0. 0004 d 3 a.z5J 13.000 47.0 1402.0 locos.O e4d54 u.003122 0.000633 2 .c87 14.VJJ 247.d 1965.7 10420.2 178941 0.093108 0.001292

.5UJ 24.0Jo 516.1 lo2to.C 15644.1 221535. 0.003 00 0.0017,o J. u. 0435.0 16250.0 PVOO.I 22 535. 0.03JIOo 0.0017/o

.00) 2 4. O'Is 57o.1 1o280.0 6333.0 290330. 0.03J1GJ U.002244

.00J I o . 0dv 512.1 J217.0 7623.9 363022. 0.000073 0.033053

> .00J Ic.00J 432.1 5153.0 73be.e 4J6651 0.000059 0.033453

..JSu I e . uJe 522.1 o153.0 ob63.7 45J327 O . 0030 3 9 O . JU 3 L i

n. Ibo l o . o la 534.0 3703.5 6 332.7 0183/4 '

-). 003303 0. 0(; 39'to

s. O. 17156.C 370J.5 -10317.3 Sic 074 -0.00dOOJ 0. 00 3 '4 'i O e.25v- 1c.510 060.4 3647.2 -113/7./ 415422. -3.000)40 0.0J3770 1.09v i c . 4 've 60.5 Jo33.6 -11433.2 4J4014 -0.003044 0.03373:.

s . oJJ 16.4o5 602.n 3610.2 -12048.0 2 ecol 8. -0.001,73 0.OGJIJJ

. 50) 16.40e 093.2 3610.2 -12734.2 1441tu. -V.03J3h O.002115 i .25s 6.Os> 100.0 201.1 -12HV4.2 -0. -J.OvJi33-0.ucsJss lateral f recJency t)y raleigh's methoc = 3191.

HUHAw SiUF Ai l o bs

,tb 4I U..I f 5 i t.

J av .15 CHU 000U.05 TCH 0032.21 AC

r AT 11: 33 % 01/1e//4 a

/ 3 A-8 Nib Franklin Research Center A Dewson af the Frankhnlasutwe

APPENDIX B AMERICAN BUREAU OF SHIPPING, RULES FOR BUILDING AND CLASSING STEEL VESSELS 1

1 i

&m Franklin Research Center A Division of The Frankhn Institute The Bengan Franklin Park.ay. Ph.la Pa 19103(21 % 448 1000

I-APPENDIX B Selected Sections Pertaining to Crankshaft Design Rules for Building and Classing Steels Vessels American Bureau of Shipping,1980 The following are selected paragraphs frota ABS Section 34, Internal Combustion Engines, and ABS Section 44, Materials for Machinery, Boilers, Pressure Vessels, and Piping.

Although these selections are from the 1980 rules, the technical content differs very little from the 1973,1974, and 1975 editions of the rules, t

-a B-1 Ob Frankhn Research Center A [hmwm of The Fsankhn lesWe

1 i

I 34.1 Construction and Imtallation 34.1.1 General Comtruction and installation of all internal-combustion engines and reduction gears intended for propubion in claned vewch and auxil-iary engines and reduction gears of 135 horsepower (hp) and overare to be carried out in accordance with the following requirements, to the satisfaction of the Surveyor. Srnaller auxiliaries are to be of approved constniction and are to be equipped in accordance with good commercial practice, but need not be impected at the plant of the manufacturer, whose guarantee of the engine will be accepted subject to satisfactory performance witneaed by the Surveyor after .

! intallation. For engines driving generators see aho 35.21.

34.1.1 Comtruction-survey Notdication i

' Before proceeding with the manufacture of materiah subject to test and inspection, the Bureau is to be notified in writing that survey is desired during construction, such notice to contain all the necenary information for the identification of the machinery to be surveyed.

34.L3 Certification on Basis of an Approved Quality Control

' Program Upon application, comideration will be given to the acceptance of stand,adhed, m.us prodacd en;;ir.es ar.d r:doction gears without

] test and inspection of individual units subject to approval of the j

manufacturer's quality control program.

j 34.3 Plans and Particulars to Be Submitted 34.3.1 All Engines The particulars to be submitted for all engines are to include the type of eisgine, maximum continuous brake horsepower and revolu.

tiom per minute, maximum firing pressure, mean indicated presure, critical-speed data, and weights of reciprocating parts, weight and diameter of flywheel or flywheel effect for the engine. Material specifications are also to be submitted for approval.

34.3.2 Main Engines la addition to the plans showing the general arrangement of machin-t ery in the vessel, shafting, stern. bearing details, the siies and types i of various auxiliaries and the sizes and purposes of suction and dis. -

charge connections of the' pumps. as required in other sections of

the Holes to be submitted for approv61, the following plans are to be submittest in quadniplicate for approvah Sectional assembly, bedplate' or crankcane including details of

. breather arrangement, sump ventilation and explosion relief valves, 1

cylinder incheding jacket and liner, cylinder head, piston and con.

necting rods, shaft ng, couplings. clutches, vihration dam rs, tie rmh if fitted, pressure piping. air containers and details of t following when driven by the main engine: air compressors, scavence pumps or blowers. turbochargers or superchargers; for indirect drive, plans of the gears, clutches, couplings, generators and motors nre to be sid>mitted in accordance with Sections 33 and 35.

J

^

Franklin Research Center A Dmeeun af the reanesn Insteuse

, I , - , , . . , . , _ . . . . _ _ . ~ , , . . _ . _ , , _ . . _ . . _ . .

. , . - , . _ ~, - , _ _ _ - _ _ . _ -. . , - , , . , _ . . - . ..

i l

1 34.3 3 Auxiliary Engines Plans for auxiliary engines are to include a sectional assembly and i crankshaft, piston rods, connecting rods, couplings, clutches, vibra.

tion dampers, together with pressure piping and air containers and, where fitted, supercharger or turbocharger in sufficient detail for

, design analysis. The plans are to show details of the breather arrange.

ment, sump ventilation and explosion relief valves when they are required.

343.4 Torsional Vibration Stresses The design equations do not take into consideration the possibility of dangerous torsional vibration stresses, and where propuhion criti-cal-yed arrangements are such that dangerous torsional vibration may occur within the operating range, calculations are to be submit.

ted including tables of natural frequencies, vector summations for critical speeds of all significant orders up to 120% of rated speed,

< and stress estimates for criticals whose severity approaches or exceeth the limits indicated in 34.57 and Table 34.3.

i 34.7 Material Tests and Inspection 34.7.1 Speci6 cations and Purchase Orders

! Except as indicated in 34.1.3 and 34.73, the following material

, intended for engines which are required to be comtrscted imder survey is to be tested and inspected in accordance with Table 34.1.

The material tests so indicated are to be witnessed by the Surveyor in accordance with the requirements of Section 44, ami copies in

duplicate of pi.rchase orders and specifications for material are to 4

be submitted to the Bureau for the information of the Surveyor. The Surveyor will inspect and test material manufactured to'other speci.

l ficatiom than those given in Section 44, provided that such sgweilica-tiom are a, proved in connection with the designs and that they are i clearly indicated on purchase orders which are forwarded for the 1

Surveyor's information. All other tests in Table 34.1 are to be carried

. out by the manufacturer whose affidavit of tests may be accepted

by the Bureau.

34.7.2 Steel-bar Stock ia llot. rolled steel bars up to 229 mm (9 in.) in diameter rnay le used when approved for use in place of any of the items indicated in e Table 34.1. See Section 44.

34.73 Alternative Test Requirements Material for engines and reduction gear units of 500 hp or less, including shafting, gears, pinions, couplings, and coupling bolts will be accepted on the basis of the manufacturer's certified mill test reports ami a satisfactory surface inspection ami hardness check witnemed by the Surveyor.

! # .+ . B-3 Uh Frankhn 4 o, a rn. r Resea.rch Center

31.13 Cylimien and Covers, IJners, and Pistons ,

Patts such as cylmders, liners, cylinder covers, and pistinn whid .iie suliect to high temperatures or pressures are to be made of ni.hoal suitable for the strewes and temperature to which they are exposed.

< When the cylinder diameter is over 230 mm 19 in.), a relief valve, set to relieve at not more than 40% in emeess of the maximnm firing pressure is to be fitted on each cylinder of reversible engines and engines using air for starting. For auxiliary enigmes other effettive iiieans for determining the maximum cylinder pressnre, uh as a maxinium-presure indicator, will be specially cormidered. ,

L 1

34.17 Craninhalts 34 17.1 Diameter of Pins and journals The aliameter of the crankshaft pitis and journals. In mm or in.. is not to he less than d as' determined by the following esl nation.

d = c , 31 + (3tJ + 41'J)a< a f

N Alettse t' nits inels l'>nnel t' nits 4

St = l Rftt'Dal, St = 0.131\'ly1, T = 1.02 x 1011/11 T = fu.tuloll.'It n = liameter of cylinder lune in mm or in.

. I' = maximum firing pressure, in kit /cm2 or psi 1 I, = span between bearings, measured over the web, in mm or in.

I 11 = hp at rated speed y, sy% fjrf4fe#

H = rpm at rated speed i

c = 1.lfi for one. cylinder engines

= 1.13 for two.c'ylinder engines

= 1.10 for three. cylinder engines

= 1 n7 for four cylinder engmes

= 1.(H for five. cylinder engines -

= 1.n2 for sis. cylinder engines

= 1 fMI for engines with_ more than six cylinden -

f = 1.!Mut for Grade 2 forgings

= 2,140 for Grade 3 forgmus -

< = 2,310 for Grade 4 forgings Values of f for other snuterials are stihject to special coenideration.

I Mwe The alwive erpret6nn wdi maally apply to enumes w here a heannt ailenne ca h 1

s6.le af euh trank an.1 where imcle ungmhn m rnr at cereal intersalt it mas 4 apply to other enenies if U li nunhheil to restest the appropriate henilme

nunnents. Intte.pnlihmemann mav he re rored a here a reekal-. pre,I areance.

mente tw stren own entraepwn are not fav.wahle h here e rankshaft elememann

' are penpned wheth are ten than thme ilrternnned hv the alane cereat>=.

ownplete mppirtu.c data,6mlinimt decaded itens analons, are in he whnnt.

. fed for sprual ownplcrate,wi.

l

.., a-4 IM 4Frankhn % a rke r, Resear.ch C. enter

.in na..

-- e p- ,a .i -,q

i

-. .?,,

{

~ 34.17.2 Manimum Firing Pressure and BilP The Surveyor is to verify the masimum firing prnsure P and brake t honcimwer thering the full power trial of the engine. When the engine Innkler has dennuistratetl to the Surveyor by means of tests on a pilot engine that the design value of f is rmt eseceded within est iblished limits of production tolerances and settings which wonki

affect it, venfication of P will not he required for an engine Imilt on a prmluction line, provided the engine delivers its rated power within the established limits.

4 34.17.3 Highee Ratings

,. Sule.cepient adjustments for the purpose of obtaining higher powers or higher maximum presmres will he subject to special cimsideration.

A 34.17.4 Solid Crankshaft Webs The proportions of the crankshaft webs are to be such that the cifective resisting moment of the wels in lumdmg is not lew than fWr% of the resisting moment of the minimum respiired dumeter of 3 pins and journals in bending: that is,

wet h n.35d' w = ellective width of weh in mm or in.

i t = thickmw of web in nun or in.

Where the prog artions are such tlut pins and journals overlap. t may be taken tt, ue the minimum diagonal tintance through the wch.

4 34.57 Torsional Critical speed Arrangements and htrens Limits i 34.57,1 Allowable Stresnes f' Where torsional critical speed arrangements daller sigmlitantly from

. previous imtallations. the torsion.al vibration strew in propeller slialts

, and progmhiose crigme cranksluf ts, due to a snegle lunnonic caestmg

,[ factor at the rtNmant peak, is not to neced the limit indicated in' 4

Table 34.3. Total vibratory strew in the interval from mrN to 105'% q

'of rated speed due to remnant harmonin and the dynamically

,magmfied parts of siumlicant nonrewnant harau--- a suit to ew.eit 13r% of the . allowable stresh for a single harletonic esenting lattoE 34.57.2 Barred Hanges .

4 When tonional vileratory strewes esteed llw foregoing limih, .at an rpm withm. the operating range lmt leu than Mr% of rated speed.

a lurreil ravige is to be provided. Tlie tat-Immeter is in le nurked, 1

asul a wariaing notice htted to the engme and at tlw operating cou.

trols, to the effect that contimium ogwratum within the lurred range is to lie avonfed. The width of sneh lurred range is to talie mio

~

tomaleratum tiie breadth and severity of the critical, Imt is to estend a

/ ., B-5

' NLfFr'anklin Research Center

, A Oswisan d the Feeseen enessuse " C

, , . . . . - ,, .w

8 0 at least 5% almvc anel 5'% f riow the speed at the remnan) peal A harted rar:ge is not acceptable in the interval from 'MP% to yNP%

v4 rated speed.

34.57.3 Other Effects pecame torsional vibration has deleteriom effects other than shaftinq fatique, the limits in Table 34.3 are not intenited for direct appli-cation as design factors, and it is desirable that the service range almvc 'Mr% of rated speed be kept clear of torsional criticals inwfar as prai-ticable.

34.57.4 Torsiograph Tests When the calculations indicate that criticals occur within the operat.

ing range whose severity approaches or exceeds the limits in Table 34 3, torsiograph tests may be required to verify the calculations and to assist in determining ranges of restricted operation.

34.57.5 Vibration Dampers When torsional vibratory stresses exceed the limits in Table 34.3 and a barreil range is not acceptable, the propulsion system is to be redesigned, or vibration dampers are to be fitted to reduce the stresses.

34.57.6 Gears When the propeller is driven through reductmn gears, or when annliary eqmpment, such as a blower, is driven through gears. a barred range is to he provided at the critical speed if gear.trmth chatter occurs during continuous operation at the critical.

.. B-6

[$lF'O! rNnklin Research Center A(Jeweenwsof the trankhnlesqueuse u

J O TABLE 34.1 (continued)

v. .r. .,r .

te ' u,,u, ; r...r. ,,i, rnnu

%esel gese uhrels f.or sam 4iaft ilmes alwwe 4twi mni alanc 4:uiemn t117 in.l luirc t 117 ina l=>re

%.giresharget, lorl=wharcer shaft snel rotor alim e ifwi m m al=n* 48timm (ll.M ina lwire il17 m.i lmwc

('ast strel cIrments, imimling their wcl.leil

.. nne. tunis (ne licilplates ic g mam beanne h.anm.c ' all all lie.lplate of wel.Inl omstnet um. piaies anil tr.imirne in armes rmlen maile ni forgeti ne t ast steel all -

Framc aml eranLe aw nf welitrei ronst nit a lim all -

Ent.il.latures of wehint tomtniction all Thnnt sh.ifts. lineshafts.cnnplings. emipling -

twilts, progeller shafts. alvi generat,3r shafts anil inwitor atul gear shaf ts Int inclirect altive 'all t.re, i un e hic .i. re n..I n.ver parts .,1 the enene s .s h as p pes .in.1.. c, ..ncs ..I the .earimg met wssem a.nl edhre perwire system, 2 u.oiene p.rti. ir I p..I practrant ne c.pi.Ils e#n m e tests 4.c t., hr arew.t ...e Iwre n.. Icsin.ctive te emt 65 in h..tril.

1 t h,4 .. . tr e.nc n t.hnanait, mr.ornl.

I l'ivrew.nor testing is adihtenneliv re.purcel where the lore eu-ecik list nmi 4117 no .

1 N.nwintrmteve inung mov he reepute.1 liv the Nervevne f.it test trel imnionients. '+ c il lt %.

st..icre el enes. to he intnew,1 hv the % rvesne.

E-8

$. Frankim Re.se.a.rch C. enter W}N a t-.e tii e, . an . e

_ _ _ _ . = __.

. 0 TABLE 34.2 Test Pressures for Parts of Internal combustion Engines P i, the nminimni no,Lmg peewnre m the gun (n.ncened tr.we I st twuute Cslunter unct tontmg spate 'TLctm J ilini gno Csluuler Imer, over the whole length of 7 kg '(md iltui pso umhng spas e Osimder gas Let, unehng spate 4 kg. t md i57 ino Imt not ten than i TP 1:sliaint salse. umhng pa c 4 kg tmJ iS7 gno 1 nl not Icw ti..m i IP Pnton t ri.w n. untmg spate ? Lg!sm dilies pse 6 taller ancinlily woh poton enth fiert inlrtlieHe % stung l'miip luuly, perunee ude 1.5PorP + %ElLg/cmJ 01170 gno uhnbrier a lew T als e 13P n, P + isl Lg, und i ll?tl pso whnheser a leu Pipe 1.5P or P + .1u0 Lg, undi 4270i no whnheter n Irw

s. .enee pump t,iemic, 4 sg/und .s? pso Turbobl..wcr, unihng spare 4 kg'sm' t57 pso Imt init Icw than i TP lShains pipe, um.hng spat e 4 kg 'un J 457i no I.ut ned icw than 13P 1:ngme-ilenen air unnprewor. tylmden.

unen. interometen aml alterumlen ait ude 1SP waire sale 4L4amd :57 pso lait not Irw Ilun 13r 4 '. .len, rai h ide i krun i57 d

gno init n d lew than i TP l.ngme ihnen pmnps .uil. w ater, furt. Iniget 4kgtm3 .5* ino Imt not 1.w tlian I it' TABLE 34.3 Allowable Stress Values for Crankshafts and Tail Shafts Due to a Single Harmonic (Grade 2 Steel)

I neme %g -rel 0 IN ..e ten el %N 0 *M . l.c ott I eM lh neirf re NWimm 14tlO Lg/t ml 1291 kg/ tent :l':lSillg /t tn J 17dl L g.'r mJ

. Il 't m I or Icw (5.f #1 ps0 (1591 ps0 t2,114 gnH 11XiHgno t h inicere es o mm .120 k g, un' 1:00 ku/un' :100 kg i m' 1:<ui kg m' ell (6 m i ut more (4 M i ps0 ;2.M5lnH f 1.7117 p*O ( 2.'sl1 lnd v,.ee, I seern I min f..e spent = mecemeikaie letween th.ne sh.n.n in Table it 1 emt foe ih.ifu l .eween u o an.1 eaai mm i11 it m an.lilit in i m ihameter, maw lie .J e 4med in uiterp444am In the T4 lc, il n epm at t4ted speed. whn h H ihe speed at mainmem omtienmu e4tmg in, re g .lae npre.ition in went e %iecise, see noenmal vol.ies heed .. ih4mvect of g e nbpm..

.w ... the mmunmn peopeller.. haft diametet twtween Ihe big en.i ni the tapee and Ihe Inew se.1 itren et. ml. dnerver.img steen omi enstati.m f4.t.ws j M here the 9rttite H tmit that the vettel will e.perate Int a ugitifit Ant putum ni tot tervle e life 4e =preds b .l..w is e% of este l oper.l. the iteen limit, ne the me -ev41 o mit. I mit .41411, le i er t. he me.1 m im h pent rence.

1 llInem.441 reitu al speed see,ingeenents see umilar to pervemn 6mtallatomi pemen h, ersu e cyn ew m c. 4.m**ireatum will f.e gn en to h.che, iterne. upm u.bm.tt44..t f..il dre.i i.

4 sorn 'mne, so, s e mnsliarn m.nle ni t;es.le t .. s iv seppemed allav. steel f..egenei mn I.e nu er oril by t u .. tineils e4 the gwerentage nutece m ultimate teemle sterndh n*ce t j lir e im i toisupn i

d..

hd Frankhn Researt h Center B-9 s u w.n .a si e r . une m,nu,

a .

44.19 Steel %Iachinerv I'orgings

\..ee I., mine.intui ae arment with \sT\1 einn:n.aemaw

\te-4,t4w H toe glW (.r.n!c ?

\eahr.l.m l} for \In (.r.wir .)

Afd.ht.t.sw E tor \lt% (itaile 8 44.10.1 Process of %lanufacture a General The followmq reeinirenwnts sover eatinm. steel l~

mgs intenik d to be med in machinery constnntion. Thn does n. a pret hnle the me of alloy steels as pernntted by 41.1.1. The ocel n in be fully blicd anel n to be maile hv one or more of the tolig

'protewet open hearth, basic osygen. elettric inen. ace, os mth other powen as may be approved. The uow.wotional area of the m.iin im edy ol the unmattune l. hnbhed forging is ne to exceed one thml of the area of the ingot; palnn. flanges ami simdar enlargemenl> on the forging are not to esceed two. thirds of the area of the ingot. A mihuent thward n to be made from c.nh ingot to seture facedom from pipmg and undue wgregation.

b Chemical Compositiori The themical cominniinm is to Iw reimrted and the earlam umtent n not to eu ccil o l% unlew nie:

_ually anmused Speually approved grades havmg umre than u..nt tartum are to have S marked after the grade numlwr.

44.19.2 hlasking and lietests a .%fotAing in aikhtson tu .ippenpriate klontdic.ition ma: Lings of the mannfacturer, the ihncau ma:Lmgs,ludicating satolactory unm-ph. ante with the linle retiturements. ami .n Im nkhed by tlw 'm ve,s on, a to be stamgwd ani all forgings in such hicatmn .n in be ihwern-ible . alter m.nhimng and untallation in adihtion Grade 2. (.sade .1.

aml Grade 4 forgings are to be stamped . . .aml respec-Inciv.

b Hetesta ll the remits of the physical tests for any forgings m

.iny lot of forgmgs do m>t conform to the ret lidsements spet shed.

the mannlactmer may re. treat ther torgmgs, but amt more thais linee i

adihtion.il timim. lietests of an adihtumal speumen or speentwin are l

to le made ami are to conform to the trijmrements spet theti.

44.1921 lleat Treatment

.i Genetul Unlew a ilepartiire from the followmg pnnetimes b l spenhtally approved. Grade 2 and 1 fmemes are to le annealed normabied or normahied ami temiwred. Grade I imumg* .nc to lie normaheed and tempered or double.nonnakevil and tempen d.

Tiw Inrn.ne n to be of ample prognertioin to lirisig tlie longings to i motorm temiwrature.

h Cmiling Vrior to lleut Treatment Alter fosumg aml beloon seheating for heat treatment, the forgings are to be allowed to unel in .i m.imwr to present inlnt) and to asunnplish teasoformation.

c Anneuling The forgmgs are to lie rehealeil to and hel i at the proper ainteentielog tempe atuse for a milnient time to elles t Ilie deuseil tr.imlonnat uni .ind then be allowed to uml slow lv ami esenls m the furn.ne untd the temperatuse h.a fallen to alumt IMC (N?>0l's or lower.

d Nonnalizing The fosgmqs ase to be scheated to .nni Iwbl at lhe peoper temjwr.itute alnave the tramlosmation sange for a mih-t trut imw to elletI the tiestred trainluem.itimi unil then w allahann lonn the hirnate and allowed to umi m air.

e few;=Wng Tlw hugnigs are to las rehealeil to .nni brhi at ;he pmiwr tempesature, whhh will be lwlow the tramtonnahon s.ing".

ami are then to he unled undes untable unuhannn.

g, . . 11-10 nter.

N.ll ar f r.inkhn ma1.r,.,*.or.Researt h Ce4.

, 44.19.47TImile Pr rtiep a Car &IUl&f orgings ne carinmateel forginc are tu rnnform, to the. reeluirements of Table 44.11 as to temite propertiet 7 Large Forgings in the ca.se of a large forgmg reyintme two temion tests, the range of temde strength is not to eweed 7 kg mm:

(ItHMMI psi).

c Application Subject to the approval of the appropriate mate-rial for each design application. Grade 2 is approved for all purimses; Grades .1 and 4 are approved for all purgmses exceptmg pmpeller shafts .

d Alloy or Special Carlma Steels When alloy steel or carbon 1 stech ehfiering from the almve requirements are progmseil for any purpme, the purch. der's specification is to be mlumtred for approsal in connection with the approval of the dnien intimling meh appli-cation. Specifications mch as AST\l A117 or A47fl or other steels smtalde for the intended service will le eenmdered. ,

44.19.5 Test Specimens a locotion of Specimens The phnical propertin are to be deter-mmni Innn int specimem taken from prolongatunn havim: a sec-tmnal area not lew than the luxty of the forgmg. Spetimem may he taken in a direction parallel to the atis of the forgings in the shrection in which the metal n most drawn out or mat he taken j) tratoversely. The axis of longitudinal specimem is to he' h>cateil at any pmnt midway between the center and the wrface of solid forg-ings and at any point midway between the mner and outer mrfaces of the wall rf hollow forging <. The amis of tramverse spechnem may 4

he becated time to the wrface of the forgings h flollmcJrilled Specimens in lieu of prolongatmm. the test specimem may he taken from forgine mlumiteil for eac h test lot, or if uthfactory to the Surveyori test specimem, may be taken imm forumo with a hollow drdi, e Smoff Forgings in the cases of small forumo weighing few than l I 8 Le olvilb) cat.h. w here the foregoing pmtedorn are unpractica-ble, a special forging may he made for the imrgmw of obtamme test sperimem, provided the Surveyor inathhed that the e int specimem are reprewntatise of the forgine mhnnited for int. In mch cases the special forgings shonhl be whjecteti to approximately the same amount of working and redm16on as the foreme reprewnted and should lie heat treated with ihme forcino.

d identification of Specimens The int specimem are not to le i

detas hed from the forgine until the final heat treatment of the forgings has been emnpleteil not until the tot specimem have leen stamped by the Surveyor for identihratmn.

44.19 fl Number of Tests a large Forgings in the case of large forgine with rough ma chined weights over 4080 hg (9000 lb) each, one temion test h to he made fmm each end of the forgine h Smaller Forgings in the case of forgioc with rough mailuned weights lew than 40M) kg (90(M) Ihl each. ewept as holed m the following paragt.tph one temion test h to I e inade Innn cath forg.

ing 1., B-11 l *h ,Franki n Research Center a o, na n., r .mia inaw.

. o e Small Forgings In the case of unalt forgings with rough ma-

, chined weights less than 227 kg (500 lb) each, one temion test mas be taken from one forging as representaine of a Int of 90.4 kg i2tx)0 the or lew, provided the forgings in each such lot are of sinnlar site. are of one grade and kmd only, are made from the same heat Jud are heat treated in the same harnace charge. For lots over 908 Le (2txn>

lb), only one tenuon test need be taken from one small forging as i

representative of .a lot proviiled 20'% of the other forgings in each such lot, not sulajetted to temile tests, are ushjetted to llemcll haid.

new tests and meet the following reiguirements.

l limall liardusu Tot .

.tluumurn Grade 10 mm ball. Jul La I. uni s 2 120 3 150 4 170 d Specsal Situarium In the case of a numl=r of pii.ces ont hum a dngle heat treated forging, individual tests need not necewarily lie made for each piece. but such forging may be tested in accord.une with whicheser of the foregoing procedures is applicalite to the primary heat. treated forging involved.

44.19.7 Impection All forgings are to be inspetted by the $ntveyor alter Imal he.a

treatment and they are to be Ionnd free from elefech.

l TABLE 44.11 Tensile Property Requirernents for Carbon-steel Machinery Forgings Ismestml6nn! Tenme mr ynmens % prune.ee M I. 1.1, mere.

f. nnte nnnot tem ne Merlur. Ilame.o. It. l , .

sr u %rarne.h. tirlal 96 mne to.n of rem m re,m o.(

mm 9,rneek se m . .t,,,.. vs m .g,rn.

e s e, Aor . r, a.

c

Ac, ,,,n.3 . ,,n ,, 2 n. i nm, i :,.nl. . me s se r mm un o opu, opn> g,,m n.o gnm ent an n eur ,,, ,ra no 2 3nl a,,5 34 (121 42 21 to $1 Ital enrunun ($unun . 24 all

! ill) l 1 2m 21 ao l sn i

i 2ful $nt

~

thi ,125 M 2fi 5 . 22 1*

~ "

j sisuun rtT* sun y fit Sul Iful till .Jn 2n 12 W ~ "

r2n1 l's an I .W5 31.5 2n .11 17 27 (R1:n un (nruun l % . .r u 1.cn eang.n. ut sp2 emcm are taken fr nn a heels. nngs. remt ihse s. en . In a hnb :he inap.o i

final hnt wenk6ng is ni .he e.mu neial ihrcten.n. the trimic t.%e es%Its a,e to mere the er,,...n .ri. , i.., t.meii..hn..I .i.ri-unem

.., B-12

-*YiNIFrankien f ri ni,e n., tr ch C n Resear.enin erWer e

I j

i

{

APPENDIX C INSPECTION COMMENTS CONCERNING DIESEL GENERATORS i

I i

w%

Franklin Research Center A Division of The Franklin Institute The BenAmin Frank).n Parkway Phila Pa 19103 (2 t Si 448 10r4 l

I t -. .-_ .- -.. , -

...-_-- _. . ~ . - - - - . - - - . . - . - _ . - - - - - - - - - - - . - - .

e e APPENDIX C -

4 The comments herein are based on an inspection made as part of an initial visit to the Shoreham Nuclear Power Station on September 1, 1983.

C.1 Diesel Generator 101 Diesel generator (DG) 101 was located in its operational room and was being prepared for a limited test program. The crankshaft was exposed for inspection, cracks in cranks 5 and 7 were being ground out to reduce the associated stress concentration, and preparations were being made to install instrumentation. Instrumentation specialists at DG 101 indicated that the planned instrumentation included a strain gage torque bridge on the crankshaft adjacent to the flywheel, strain gage rosettes in the crankpin fillets of cranks 5 and 7, vibration transducers (accelerometers) on various bearing journals, pressure transducers in two combustion cylinders, an angular displacement transducer (torsiograph) on the free end of the cranksnaft (opposite the flywheel and generator), and a sensor on the generator shaf t to indicate shaf t position relative to top dead center of crank 7. The instrumentation was in various stages of preparation and installation.

The cracks on the crankshaf t appeared to have been nearly ground out in accordance with the torsional test procedure.* Cracks in the crankshaf t of DG 101 were reported to have been approximately 1 inch deep prior to grinding.

Crack locations included cranks 5 and 7, and the cracks made an approximate angle of 45' relative to the crankshaf t (or crankpin) longitudinal axis. The crack on crank 7 was located in a 5 o' clock position relative to top dead center of crankpin 7 and on the fillet toward crank 8. The crack on crank 5 was located in a similar manner, but in a 7 o' clock position on the crank pin fillet toward crank 6.

l C.2 Diesel Generator 102

DG 102 was the unit with the fractured crankshaft. DG 102 had already been moved to the main turbine deck where space and crane facilities were

Revision, LILCO Procedure for Crankshaft Testing, Emergency Diesel Generator No. 101, September 23, 1983.

y C-1 Obj F nklin Research Center A Dmeun of The Frankhn Inseeute

O O available to disassemble the unit, make a thorough inspection, and rebuild it with the 13 x 12 crankshaft. Disassembly and inspection of the whole engine was progressing part by parts although it had not progressed to the point of removing the fractured crankshaf t, the fracture was clearly visible and open to close inspection through the sides of the engine block where the cover plates had been removed. Inspection of crank 7 revealed a fracture through the crank web and partially through the crankpin, with the fracture passing through the crankpin fillet at approximately the 5 o' clock and 7 o' clock positions with respect to top dead center of that crank. The tip of the V-shaped crack propagating out into the crankpin reached approximately to the midpoint of the crankpin bearing surface.

Further inspection of the fractured crankshaf t (still assembled in DG 102) revealed that one edge of the web at the fracture had a large discolored area characteristic of heating to a temperature range of 400*F to 600*F. This discoloration was attributed to the considerable energy dissipated in sliding contact at the point of fracture and against the connecting rod during the short time (approximately 1 1/2 minutes) that the diesel was believed to be under power (see page 3 of Reference 18) following the fracture.

Inspection of the sump revealed considerable debris under crank 7 as compared to other crank positions. Although accumulated dirt in the engine sump was heaviest toward the flywheel and generator end of the sump, the excessive accumulated debris of crank 7 proved to be mainly bearing material scraped out of the connecting rod bearing by the displaced fractured segment of the crankpin that acted as a sharp cutting tool during those moments of operation following crankshaft fracture.

Further inspection of the crankshaf t failure was not conducted because the crankshaft was to be removed over the holiday (Labor Day) weekend and transported to the facilities of Failure Analysis Associates, Inc., in Palo Alto, CA, for immediate extensive examination.

C.3 Diesel Generator 103 DG 103 was observed to be under disassembly in its operational room in preparation for moving it to the turbine deck.

/ ., C-2 d} Frankhn Research Center A Dunswm of The Frerden inetouse

o e l

In the initial briefing, the crankshaf t of DG 103 had been reported to i 4 contain cracks of suf ficient depth and magnitude to preclude further f operation. Testing was to be performed using DG 101 only. The action plan I

for DG 103 reportedly called for complete disassembly, inapection, rework as

necessary, and reassembly with a 13 x 12 crankshaft (13-in journal bearing

, diameter and 12-in crankpin diameter) now recommended and supplied by Transamerica Delaval, Inc. Analytical studies of the engine with this crankshaft were to be carried out concurrently, with updates to the analysis i

being made as test data on DG 101 became available.

I

.3 C-3 dlIk Frankhn Research Centet

% a n r, e .ww,

APPENDIX D mi RECOMMENDATIONS FOR MECHANICAL AND ET.ECTHtCAL COUPLI!.0 INVESTIGAT10tl 1

i l

I I

l l

l

(

l e, f. ,_ w_ ,

.U. Franklin Research Center A Division of The Frankhn Institute nen.a,.. r,.nonp.,...,pu. p. isioirais 44n i. a l

l l

. e l i

i APPENDIX 0 The following three recommendations were made* to ensure the recording of possible electrical-mechanical dynamic interactions in the course of diesel generator testing.

1. The generator output voltages for the three phases were not all stated for recording -- only the voltage difference between phases, (Va - Vs) and (Vg - V C ) . Hence, if electrical interactions occurred during the tests, a positive voltage phase reference would not be assured, but would be dependent upon the 3-phase generator voltage output remaining balanced. That is, only two measurements were being made, and that fact required the third voltage to be calculated from the 3-phase electrical vector relationship. This is possible only with the assumption that the voltage remains balanced on all three phases. When it was reported that the third voltage for recording could only be obtained with considerable difficulty, it was recommended that the voltage on each of the three phases be read and recorded separately (from the control room) so that each voltage would be known, should it be required for vector calculations.

Although it was believed that the electrical power grid would I

certainly remain balanced during the recording of data during the synchronous load tests, the reading of the voltages would remove all doubt. For loadings derived from plant equipment (coce spray, etc.,

versus the electrical power grid), the measurement of voltage was more meaningful.

2. Even though the generator rotor inertia was large, it was believed prudent to provide for the investigation of generator instabilities, especially since there could be significant cyclic torque at 30 Hz.

Accordingly, it was recommended that all vibrational data be recorded at power factors between 0.8 and 1.0 in synchronous load tests under as high a load as feasible. A load of 2550 kW was chosen from the tent procedure to minimise engine run time at or near full load.

The background of this recommendation is that synchronous generators tend to be more unstable with low excitation (1.0 power factor) than with higher excitation (0.8 power factor). Also, page 1 of f,ILCO's response to the NRC Hequest for Information !!.2 (September 20, 1983) indicated that a significant amount of the testing on DG 102 was performed at a power factor of 1.0 and that DG 102 was operating at

H. C. Her r ick, FRC Memo to C. Petrone, NRC Hesident Inspector, Shoreham Nuclear Plant Subject s On-site for Torsional Test Monitoring Deptember 24, 1983

/.., D-1 LM FranWen Rmerth Center a o a re e. nw w.

e ~

power factor of 1.0 at the time of failure. Further, means were not available immediately prior to the test to determine if the 30-Hz cyclic torque could aggravate generator instability. Hence, this consideration appeared to be a prudent co 2rse.of action for thorough coverage of possible instabilities or me'.tanical-electrical interac-tions.

3. It was also recommended that assurance be provided for the recording of any transients, mechanical and/or electrical, associated with synchronization and attachment of the diesel generator to the electrical power grid, and to any of the isochronous loads.

(

4 4

t F

A l

' t L

i 4 ,

? I. d i

r

,- 4 - -,

q,, ,

,I l :

\ '

-' -} , .

,. j i

! < L. . H '

l f* .3 ,

"Y4 i i

!t.

11

. \ I

'=

M. g '. \ : , ' , -fl

! o. qv s l- , ._ cy. .,.

g-l -

,.j.

t . $ ... ..

i. O!.7dErankhn Research Center +

1

'/ 3 ' ' , - 2I

<s a o- a w r,.,*a in.eius. . , - 7's..',' , ' " 'N{ ,,i,, , 7 - < ' c

is_ l . e ,% ,

. e l

l l

APPENDIX E COMMENTS BY H. W. HANNERS ON THE

SUMMARY

OF SELECTED FAILURES AND EVENT REPORTS OF TDI DIESEL GENERATORS i

I

{

l l

J 5 l __ _

Franklin Research Center l A Division of The Franklin Institute The Oenjarmn frankjen Parkway. Ph:la Pa 19:03 (21Si 448 LY;0

a *

, 6 4

! APPENDIX E Comments on the items as dated:

08/12/83 The broken crankshaft is bellefed to have been the result of excessive

?

stress due to torsional vibration.

03/30/83 These screws may have failed as the result of inertial forces from engine operation at or near a torsionally critical speed as well as possible low quality in material, design or manufacture.

03/08/83 The cracked cylinder heads could have been the result of design, but the new design apparently needs to be tested by actual use and acceptance tests.

03/03/83 The high pressure fuel line failures could surely be reduced by design improvements of the shroud (usually called sleeve) . Again, sufficient proof remains to be seen through in-service experience and acceptance tests.

12/13/82 Better quality obviously needed.

09/.17/82 The omission or removal of the keyway in the water pump shafts could be I

avoided by eliminating the " stress raiser" effect of the keyway in torsional vibration. An impeller design change to reduce the rotary moment of' inertia l could also help.

l l

i

07/22/82 i l

j Probably fixed by the change of design.

)

! I I

~E-1

/lI[

O!ffun,nklin

.- Rewarch Center l a am.n s n. rr=w.mau. l

06/23/82 Change to neoprene is certainly an improvement over isoprene.

05/13/83 This probably should have been 05/13/82 from the " backtracking" review scheme being used.

05/13/82 The shorter capscrews may be satisfactory, but were the original screws actually too long or the threaded tapped holes too shallow?

03/19/82 Neither the problem nor the solution (or fix) are clearly explained. The 53-minute bleed down time is too long as a practical fix. Even 53 seconds is rather a long time to consider acceptable. Successive starts should certainly be allowed more often than 53 minutes apart. Seismic qualification of the sensing line is recommended.

03/15/82 This implies that the rear crankcase cover is a stress bearing part.

Neither the strongest bolts nor the reasons for the basic failure give faith in the TDI remedy or explanation of the failure. More proof and further explanation are needed.

12/09/81 The TDI remedy of a lower oil cooler mourting secos reasonable, but the complete system should be reviewed.

11/05/81 The use of Belleville washers in the two-piece piston design may or may not be satisfactory, depending on whether the heat from the hot piston crown anneals the Belleville washers. Heat barrier design may also be required for success.

l

/ .a E-2 l

$$ Frankhn Research Center i A Chwes.on cd The Franh6n insutuee

4 .

l l

1 The cylinder liner grooving and the bearing grooving may very well be

, caused by " built in" dirt and chips in the original factory assembly. All 1

three engines should certainly have the crankcases or bedplates thoroughly l 1

cleaned and all bearings examined and replaced as needed. A tedious, careful, i and expensive job is indicated. This means not only the bearings, but also the surface of the mating parts such as the crankshaft crankpin and main jcurnals and other parts would be damaged.

07/14/83 (or 07/14/81)?

Another indication of possible excessive vibration and or lack of proper clamping to prevent cracking of oil lines. Danger- of fire from oil line fracture should certainly be given more attention.

03/23/81 Motors should certainly be qualified and not merely be stated to be equivalent.

12/16/80 A redesign of the lube oil system is indicated as necessary so that the turbocharger bearings get oil immediately af ter a start. This may mean a change to an intermediate drain back sump in addition to the main oil sump of the turbocharger. Acceptance of occasional " fast starts" is not sufficient as this is tantamount to saying that dry bearings are tolerable.

All of the remarks and critique of the items regarding the TDI engines

, are intended to be constructive and helpful. However, practically all of the suggestions are subject to testing in actual service and qualification under NRC regulations.

During the nuclear power plant survey and inspections done in 1978 and 1979, performed at the University of Dayton (Dayton, Ohio), a grand total of 288 items were investigated. Most of these subject items were found in'every-power plant of this survey.

/ . . .. E-3 UkdUranklin Research Center A Denmon of The Franhba insature

- , , - - - - ,, -y - - . -

ML20087P939

Contents

Text

Subject:

Title:

SUMMARY

Background:

Description:

Subject:

Subject:

SUMMARY

Navigation menu

ML20087P939

Text

Subject:

Title:

SUMMARY

Background:

Description:

Subject:

Subject:

SUMMARY

Navigation menu

Search