Text

Expresses Disagreement W/Conclusions Expressed in Util,Bnwl & NRC Staff Repts Re Reliability of Tdi Emergency Diesel Generators.Pattern of Deficiencies Exists Which Precludes Findings.Related Correspondence
	ML20096A651
Person / Time
Site:	Catawba
Issue date:	08/16/1984
From:	Guild R; CAROLINA ENVIRONMENTAL STUDY GROUP, GUILD, R., PALMETTO ALLIANCE
To:	Foster R, Kelley J, Purdom P; Atomic Safety and Licensing Board Panel
References
	OL, NUDOCS 8408310205
	Download: ML20096A651 (193)
	v • d • e

g o-e'

M ROBERT GUILD j anoa.c *r u. J POST OFFICE BOX 12097 CHARLESTON. SOUTH CAROLINA 29412 TELEPHONE 803/254-8132 35[

August 16, 1984 "0A 19022 g;;gg James L. Kelley, Chairman Dr. Paul W. Purdom Administrative Judge Administrative Judge Atomic Safety and Licensing Board 235 Columbia Drive U.S. Nuclear Regulatory Commission Decatur, Georgia 30030 Washington, D.C. 20555 Dr. Richard F. Foster Administrative Judge P.O. Box 4263 Sunriver, Oregon 97702 Re Docket Nos. 50-413 and 50-414 0 '

Catawba Nuclear Station, Units 1 and 2

Dear Administrative Judges:

In the limited time available to us, Palmetto' Alliance and Carolina Environmental Study Group have reviewed the rece.nt reports prepared by Duke Power Company, Battelle-Pacific Northwest Laboratory (PNL) and the NRC Staff regarding the reliability of the Transamerica Delaval, Inc. (TDI) emergency diesel rienerators at the Catawba Nuclear Station. We strongly disagree with the conclusions expressed in those reports that the TDI diesel genera-tors will provide a reliable source of back-up power as required by General Design Criterion 17 and that there is demonstrated reasonable assurance to support interim licensing of Catawba through the first refueling outage. We contend that a pattern of deficien-cies exists in the Catawba diesels, precluding such a finding.

These deficiencies, known and unknown," stem from inadequacies in design, manufacture and quality assurance / quality control by TDI."

SER on Catawba, Unit 1, TDI Diesel Generators, p. 1 (August 14, 1984).

The specific bases for our contentions regarding the Catawba diesels is amply reflected in the record and includes the long series of NRC Staff Board Notifications on the subject, beginning with B.N. No.83-160 of October 21, 1983, and the items referenced in PNL's Catawba Technical Evaluations Report (TER) at pp. 6-8.

To the extent these reports and documents are not already matters of record we ask the Board to consider them as such for purposes of further rulings on our contention.

8408310205 840816 DR ADOCK 05000413 PDR

\

-Judg:0 Kolloy, Footsr and Purdom

. - 'Auguot 16, 1984 Pag 1 Two In a series of rulings beginning with the Partial Initial Decision of June 22, 1984 (pp. 272-74, fn. 50) this Board has conditioned the conduct of hearings on the Board's former sua sponte diesel contention on further showings by intervenors of their ability to "make a significant technical contribution" through certifying either that a sufficiently qualified diesel expert will appear at the hearings for us or that such an expert

.w ill provide substantial assistance to us by preparing a " detailed statement of technical position" regarding the Duke, PNL and Staff-reports and prefiled testimony. This statement must also specify "the respects in which (intervenors) . disagree with these reports, and (describe) how (intervenors) propose to substantiate their-positions." Memorandum and Order (Concerning Hearing and Associated Dates and Expert Assistance for the Diesel Generator Contention) p. 4, July 20, 1984.

As we stated in our August 1, 1984 letter we were unable to certify that Dr. Anderson, our metallurgical expert, would be present at the Catawba hearings due.to his prior commitments to consult on with anddiesel the Shoreham prepare testimony generator for intervenor, contentions. Suffolk In its JulyCounty,4 20, 198 Order, p. 3, the Board observed that where Dr. Anderson's assist-ance to the Catawba intervenors was based upon his work ad; Shoreham such assistance would be inadequate because "the diesel engine models and the admitted contentions at Shoreham and Catawba are

different. Indeed, the Catawba contention is restricted to problems that have actually arisen in testing the Catawba diesels."

To the contrary, we maintain that the work of Dr. Anderson and the other experts retained by Suffolk County in addressing the TDI Owners' Group Program, the investigation by Failure Analysis Associates, the inadequacies in design, manufacture and quality assurance / quality control by TDI, as well as the adequacy of critical components of the Shoreham diesels bears direct relevance for the resolution of the Catawba TDI diesel generator contention.

As is reflected in the Shoreham emergency diesel generator contention, attached hereto as Exhibit 1, the intervenor Suffolk County and its experts focus their claims and technical analysis on four critical components:

1. the heavier replacement crankshafts;

2. the cylinder blocks; 3 the cylinder heads; 4 the Model AE piston skirts.

Suffolk County contends that the TDI diesels at Shoreham "will not operate reliably and adequately perform their required func-tions because (they) are over-rated and undersized, improperly designed, and not satisfactorily manufactured." Id.

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

Judg:s~Kolloy, Foot 0r and Purdom

) .- ' -, August :16,..1984 -

Page Three e n fWe maintain'that..the Shoreham experts' analysis of these

~fourJeritical components,as'well as of TDI design and manufactur-ing deficiencies -provides intervenors, here, with the ability to make a substantial contribution to a sound record for decision on .

~the adequacy of .these components and on other deficiencies in

'the Catawba TDI: diesels. -

~

.The NRC Staff and its consultant, PNL, themselves rely on Lthe results of Shoreham component analysessin reviewing the

? adequacy-of_the Catawba diesels,as' reflected in the August TER:

' Component PNL TER Page Cylinder' heads 16 Fuel Line Fittings 20 Fuel Oil-Injection Pump _ Valve Holder 22 1

Turbocharger Bedrings 24 -

' Turbocharger Lube Oil Drain Line 28 Turbocharger Prelube Oil Lines 12 9 Turbocharger Exhaust Gas Inlet Bolts 31 i

Lube Oil and Jacket Water Thermocouples 35 Rocker Arm (Subcover) Assemblies 37 Intermediate Rocker Arm Sockets 39 Exhaust Valve Tappet (Rocker Arm 3 Adjusting-Screw Swivel Pad) 41 Intake and Exhaust Valves 42

, Spring Retaining Nut and Roll Pin or Air Start Valves 44

Cylinder Blocks 51 Rocker Arm Capscrews 57 1

High Pressure Fuel Tubing 64 Jacket Water Pumps 65 i

i Clearly, no judgement can be reached about the significance of blown problems at Catawba or the ultimate safety of the Catawba

, diesels without reference to the knowledge gained through the l'

analysis performed on other similar TDI engines and components

by the-TDI Owners' Group and the expert consultants to other inter-

i. venors such'as at Shoreham. The Staff's consultants acknowledge as much:

'PNL's conclusions and comments are based on the available L

Duke Power Company documents, on on-site inspections of the Catawba engine components and examination of identi-cal.or at least similar components of TDI diesels in other nuclear facilities, reviews of the specific known-

, problem issue reports prepared by (or under the auspices i of) the TDI Owners' Group . . . .

PNL TER , p. 10.

7 L

I

~

. Judgaa Foot r and Purdem August K711cy$4 16, 19 l

.P4ge Four l l

I Palmetto and CESG specifically dispute the Duke and NRC Staff conclusions regarding the adequacy of the four critical com- '

ponents as analyzed in the pre-filed testimony of the Shoreham experts, attached hereto as Exhibit 2.

Crankshafts PNL acknowledges that three V-16 crankshaft failures.have been reported in non-nuclear applications, two of which were attributed to torsional stress. Indications, characterized as minor, have been detected in the Catawba 1A crankshaft. Since the TDI Owners' Group analyses of the RV-16 crankshafts "are not yet finalized to acceptable conclusions, in PNL's view, PNL cannot conclude in an unqualified manner that the Catawba crankshafts are unreservedly reliable." PNL TER pp. 46-48.- On the basis of the Shoreham intervenor experts' analysis we would seek to show that the Catawba crankshafts are inadequately designed and manu-factured. Exhibit 2, pp. 106-142.

Cylinder Blocks Numerous incidents of cylinder block cracking have been reported in TDI engines in both non-nuclear and nuclear applica-tions. .The Failure Analysis Associates study confirms that cracks will initiate in these blocks which they predict to be s " benign." While Duke's lA inspection has revealed.no cracks, PNL acknowledges that

~

(I)n light of the history of block cracks and the FaAA analysis, PNL and its diesel consultants remain concerned that even at Catawba there remains legitimate reason to maintain enhanced surveillance of the blocks at least through the first opportunity for heads-off reinspection and until a more definitive resolution of the problem is established by the Owners' Group and Duke.

PNL TER pp. 51-52, 84-85 On the basis of the Shoreham intervenor experts' analysis we would show that the Catawba TDI cylinder blocks are not properly designed and manufactured to withstand the stresses of service, Exhibit 2, pp. 143-183 and that far from being " benign" such block cracks could lead to catastrophic failure of the emergency diesel. Id., at pp. 151-156.

Cylinder Heads I Numerous reports of TDI cylinder head failures have been i identified in nuclear and non-nuclear applications, including a recent two-inch through-wall crack into the cylinder cavity in the DSRV-16 engine at Grand Gulf and small water jacket leaks in the Catawba lA and 1B engines.

~

.Ju d ges Kolley, Foster and Purdom August 16, 1984 Page Five PNL review of the FaAA report and the Shoreham analysis leads it to support interim licensing of Catawoa until first refueling "provided that the engine is barred-over . . . (periodically),

~

and thereafter prior to each planned start, to check for water leakage into the cylinders." PNL TER, pp. 16-19 On the basis of the Shoreham intervenor experts' analysis Palmetto and CESG would show that neither the pre-1978 nor the post-1978 designed and manufactured heads are adequate for the intended service; that unacceptable variations in dimensions induce stresses; that changes in manufacturing techniques have not solved the flaw and crack problems; and that the "barring-over" procedure will not identify all leaks into'the cylinder which could prevent starting, or cause catastrophic engine failure. Exhibit 2, pp.59-105 AE Piston Skirts After identifying four cracked piston skirts in the 1A engine

at Catawba, and upon the NRC Staff's insistence, Duke finally agreed to replace the Catawba " AN" Model piston skirts with TDI's improved design, Model "AE" skirts. FaAA analysis for the Owners' Group concludes that the "AE skirts may crack at 10% overload" but will not propogate. PNL and NRC Staff interim acceptance of the AE skirts is conditioned upon reduction in the generator loading from its 7000 KW nameplate rating to about 5750 KW such 2 that stresses are maintained below 185 psig BMEP -- equivalent to the only significant operating experience with the AE skirts in the Kodiak station application. PNL TER, pp. 11-13 The need for this is based on PNL and Staff concerns regarding the acceptability of crankshaft stresses, and the lack of substantial AE piston operational data at higher BMEP loadings.

1 SER p. 6.

On the basis of the Shoreham intervenor experts' analysis of the inadequacies of the TDI "AE" piston skirts, Palmetto and CESG would demonstrate that the Failure Analysis Associates conclusions underestimate projected crack propogation due to dimensional variations, imperfections, and actual operating environment, tem-peratures and pressures. Excessive piston side thrust and tin plating also exacerbate stress failure potential. The diesels ,

are over-rated for the "AE" piston design. Exhibit 2, pp. 25-59 The reduction in diesel loading to 5750 KW to accomodate the crankshaft and piston skirt uncertainties reduces the diesel's capacity to only marginally above their 5714 KW loss-of-offsite power emergency service loads. SER p. 6. Such reduction in design

= conservation is inappropriate to sustain licensing.

6

Fest:r and Purdom Judg:a Kolicy,4 August 16, 198 ' ,

, . . . Pag > Six l Palmetto and CESG continue to assert that the Board has ;

unfairly required intervenors to shoulder burdens not properly

-imposed upon us as the proponents of this diesel generator con-tention. We believe that Duke Power Company, as the Applicant, properly is charged with the burden'of proving the safety of its nuclear plant, including the TDI emergency diesel generators.

However, we submit that we have shouldered the burdens imposed upon us by the Board to demonstrate that our participation "may reasonably be expected this important to assist contention. 10 CFRin developing (a sound 2 714(a)(1) iii). record," on

- As observed by Judge Edles, concurring, in Washincton Public Power Supply System (WPPSS Nuclear Project No. 3), ALAB 747, 18 NRC 1167, 1182 (1983):

Our cases clearly recognize that. cross-examination can te an especially valuable tool in the development of a full record and that an intervenor may even establish its entire case through its use.

We submit the enclosed analyses of Dr. Robert N. Anderson, Professor Stanley G. Christensen, G. Dennis Eley, Aneesh Bakshi, Dale G. Bridenbaugh and Richard B. Hubbard, prefiled direct testimony on the Shoreham diesel generator contention, Exhibit 2, as an offer of proof pursuant to 10 CFR 2 743(e) and in support of our diesel generator contention as a statement of the substance of the evidence we seek to eatablish through cross-examination of Applicants' and NRC Staff'r tnesses and the documentary evi-dence referenced herein. -

We trust that this submission fully discharges the duties to be borne by intervenors and that the issues regarding the adequacy of the Catawba diesels will be resolved on the record of the scheduled public hearing.

I S ' .

Robert Guild cc: Service List (w/ enc 1. to parties)

( EXHIBIT 1 t

n. .

'j i

SHOREHAM EMFRGENCY DIESEL' GENERATOR CONTENTION contrary to the requirements of GDC 17, the emergency diesel generators at Shoreham ("EDGs") manufactured by Transamerica Delaval, Inc. ("TDI") will not operate reliably and adequately perform their required functions because the EDGs_are over-rated and undersized,_ improperly designed, and not satisfactorily manu-factured. There can be no reasonable assurance that the EDGs will perform satisfactorily in service and that such operation will not result in failures of other parts or components of the EDGs due to the over-rating or insufficient size of the EDGs or design or manufacturing deficiencies. The EDGs must therefore be replaced with engines of greater size and capacity, not designed or manu- ,

factured by TDI. (Suffolk County's Filing Concerning Litigation of Emergency Diesel Generator Contentions, June 11, 1984 (" June 11 Filing") at 2; Tr. 21,891}

BECAUSE:

1.(a) The replacement crankshafts at Shoreham are not ade-quately designed for operating at full load (3500 kW) or overload (3900 kW), as required by FSAR Section 8.3.1.1.5, because they do not meet the standards of the American Bureau of Shipping, Lloyd's Register of Shipping, or the International Association of Classi-ficatio1 Societies. In addition, the replacement crankshafts are

-not adequately designed for operating at overload, and their de-sign is marginal for operating at full load, under the German criteria used by F.E.V. (Tr. 21,878-79]

l t

i

.1 l_

-- . _ e:

i s

(b) The shot peening of the replacement crankshafts was not properly done as set forth by the Franklin Research Institute report, Evaluation of Diesel Generator Failure at Shoreham Unit 1, 1984, and the shot peening may have caused stress nuclea-4 April 6, tion sites. The presence of nucleation sites may not be ascer-tainable due to the second shot peening of the crankshafts. (Tr.

21,880]

(c) The crankshaft oil passage plugs on the replacement

~

crankshafts are inadequate, as evidenced by the failure of the same design plugs on a TDI DSR-48 engine owned by Rafha Electri-city Corp., which damaged the pistons of that engine. (June 11 Filing at 4; Tr. 21,881-82]

2. Cracks have occurred in the cylinder blocks of all EDGs, and a large-crack propagated through the front of EDG 103. The replacement cylinder block for EDG 103 is a new design which is

. unproven in DSR-48 diesels and has been inadequately tested. (Tr.

21882-3]

3. The replacement cylinder heads on the Shoreham EDGs are of inadequate design and manufacturing quality to withstand satis-factorily thermal and mechanical loads during EDG operation, in that:

(a) the techniques under which the replacement cylinder heads were produced have not solved the problems which caused the cracking of the original cylinder heads on the Shoreham EDGs;

~-_c w .r

b. '.

(b) the "barring over" surveillance procedure to which L1LCO has committed will not identify all cracks then existing in the replacement cylinder heads (due to symptomatic water leakage);

(c) the nature of the cracking problem and stresses exacerbating the cracks are such that there can be no assurance that no new cracks will be formed during cold shutdown of the EDGs; (d) there can be no assurance that cracks in the re-placement cylinder heads and concomitant water leakage occuring during cold shutdown of the EDGs (which would not be detected by the barring-over procedure) would not sufficiently impair rapid start-up and operation of the EDGs such that they would not per-form their required function; (e) there can be no assurance that cracks in the re-placement cylinder heads occurring during operation of the EDGs 1

would not prevent the EDGs from performing their required func-tion .

(f) variations'in the dimensions of the firedeck and water deck of the replacement cylinder heads create inadequate cooling, where too thick, and inadequate resistance to mechanical loads, where too thin, and create stress risers at their boundar-ies; )

(g) the design of the replacement cylinder head is such that stresses are induced due to non-uniform bolt spacing and the 1

different lengths of the bolts, l r - 4 - - - - , - e ,

- f (h) the replacement cylinder head design does not pro-vide,for adequate cooling of the exhaust valves; (i)- at least one replacement cylinder head at Shoreham has an indication:-

(j) the design of the replacement cylinder heads pro-vides inadequate cooling water for the exhaust side of the head; and (k) the replacement cylinder heads at Shoreham were inadequately inspected after operation, because: .

(1) a liquid penetrant ter.t was done on the ex-haust and intake valve seats and firedeck area between the exhaust valves on only 9 of the 24 cylinder heads, and,such tests were done after only 100 hours0.00116 days 0.0278 hours 1.653439e-4 weeks 3.805e-5 months of full power operation; (2) ultrasonic testing was done on the firedeck e

areas of only 12 cylinder heads;

~

(3) visua1 inspections were performed on the valve seat areas of only 32 of the 98 valves, and on only 7 firedecks of the 24 cylinder heads for indications of surface damage. (Suffolk County's Motion for Reconsideration of Portions of Board's July 5 EDG Order, at 1-3, as granted in part and modified (in sub-para-graph (j)) by order of the Board during a teleconference of the parties on July 11, 1984]

4. All AF piston skirts in the EDGs were replaced with TDI model AE piston skirts. The replacement AE pistons are of inade-

a. -

-s-quate design and manufacturing quality to satisfactorily withstand operating conditions, because:

(a) the FaAA report conclusion that cracks'may occur but will not propagate improperly depends on a fracture mechanics analysis of an ideal situation which is not valid for the actual conditions which may be experienced by the Shoreham diesels, (b) excessive side thrust load, which could lead to catastrophic failure, has not been considered adequately, and (c) the analysis does not adequately consider that the tin-plated design of the pistons could lead to scoring causing excessive gas blow-by, and thereby causing a failure of proper operation. (Tr. 21,88,6-88]

l e

4

EXHIBIT 2-SUFFOLK COUNTY, 7/31/84 1

-UNITED STATES OF AMERICA NUCLEAR REGULATORY COMMISSION Before the Atomic Safety and Licensing Board

)

In the Matter of )

)

LONG ISLAND LIGHTING COMPANY ) Docket No. 50-322-OL

)

(Shoreham Nuclear Power Plant, )

Unit 1). )

)

JOINT DIRECT TESTIMONY OF DR. ROBERT N.

ANDERSON, PROFESSOR STANLEY G. CHRISTENSEN, '

G. DENNIS ELEY, ANEESH BAKSHI, DALE G.

BRIDENBAUGH.AND RICHARD B. HUBBARD REGARDING SUFFOLK COUNTY'S EMERGENCY DIESEL GENERATOR CONTENTIONS INTRODUCTION ,

4 Q. Dr. Anderson, please state your name, address and oc-cupation.

A. My name is Robert N. Anderson, and my business address is Department of Materials Engineering, San Jose State Univer-sity, San Jose, California. I am a Professor of Materials En-gineering at San Jose State University.

Q. 'Please describe your qualifications and experience which are relevant to the matters you address in this testimo-ny.

i l'

I e

_y ._, ,_,_._-..._-_-.,-.y, -

, ,., . . - , - - , . . , , . - - - , . , , , . , - _ -, -y- , - -m, .,w-,--w--, w, ,w,.v. ,,,# - --+-. .---..--v.v. ,,,w--,

A. - I have a doctoral degree in metallurgy, a masters of science degree in- chemical engineering and a bachelor of sci-ence degree in chemistry. My duties as Professor of Materials Engineering include teaching courses in casting and nuclear ma-terials. I as a licensed metallurgical engineer and nuclear

~

engineer in the State of California, and I have qualified in court as an expert witness in metallurgy. I have actively con-sulted in the field of failure analyses for 10 years. During that time, I have served as consultant to a wide range of businesses, research facilities and local, State and Federal agencies and commissions, including the California Public Utilities Commission, Brookhaven National Laboratories, IBM, Memorex, Lawrence Livermore Laboratory, the California State

, Energy Resources and Development Commission, the Executive Of-fice of the President of the United States, Council on Environ-

]

mental Quality and Office of Science and Technology Policy, and the Office of Technology Assessment of the United States Con-grass. I have published over 50 articles and I have had numer-ous patents issued to me in the field of materials science, 3

including fuel cycle patents and a nuclear reactor patent. I

' am actively involved in professional activities, holding mem-bership in the American Nuclear Society, the American Institute of Chemical Engineers, the American Chemical Society, the

- 2- -

e

_r,m,.. . - - ,, , , , - - -~-,-~_rv,-w,-- -,,--,~rwww~--*-w~*-we~wsmm- - w-, --wv-+--~- w-w,e~ -w w e-= 3-enw=--~s--*~--w*~.nt

American Society of Metals and the National Society for Professional Engineers, among others. I am also a member and past Chairman of the Northern California Section of the Ameri- .

can Institute of Metallurgical Engineers. A further statement of my professional qualifications is attached to this testimony as Attachment 1.

Q. What parts of this joint testimony have you espe-cially sponsored?

A. I am particularly sponsoring all of the testimony pertaining to metallurgical science, including the properties i of materials, crack initiation, propagation and arrest, details of the casting process followed by Transamerica Delaval, Inc.("TDI"), and analyses of ,the various methodologies applied by Failure Analysis Associates to matters of crack initiation and growth. I have not provided testimony regarding the func-tions or NRC regulatory requirements for emergency diesel gen-erators.

Q. Professor Christensen, please state your name, address and occupation.

A. My name is Stanley G. Christensen. I am a Professor i

at the U.S. Merchant Marine Academy, Kings Point, New York.

I i

e 3-l

, +- , . , . - _,-.,,,_-.,m ,_-..y- ,c-- ,e-,.,.,---_.m. . , ,__,,,.y.. . .

..--.--.m ----,__v., ,.,-,..y..-. -. . _ ,,,m -, - . , ,

i

-Q. Please describe your qualifications.and experience

' which.are relevant to the matters you address in your portion of this testimony.

A. Since coming to Kings Point in 1978, I have had re-

- sponsibility for teaching various courses on diesel engines, including Internal Combustion Engines I.and II, Diesel Engine ,

L Maintenance, Marine Engineering I, II and III, Medium Speed Diesel' Engines, Diesel Propulsion Systems for Marine Engineers, Fundamentals of Marine Diesel Systems, and Diesel Ship

- Operation and Control for Masters and. Mates. From 1950 until 1978, I held various positions and was engaged in all aspects of diesel machinery for a variety of shipping companies and

served as an engineer surveyor at Lloyd's Register of Shipping.

I Prior to 1950, I served as a senior lecturer at Poplar Techni-4 cal College, London, England, and taught various subjects,

!- including Strength of Materials, Thermodynamics, Theory of Ma-chines, Mechanics Static and Dynamics, and Engineering Design.

I

! I also served at sea in merchant ships for 10 years, sailing 1

l finally as Chief Engineer. I have nearly 50 years experience with diesel engines. I am a long-standing member of the Insti-

! tute of Marine Engineers, having served on the Membership, Fi-nancial and General Purposes, and Education Group Committees and on the Special Committee on Engineering Institutions Joint l

I ;

i l

i

n. . - .ae - - _ - w, . . - - , - - , . ~. , , , - , , wv.,,w, . - , , _ _ ~ _ ..,--..e-,__.. , -.,,,_,,,_, w o-,,e,-w_.,-_...--,._.-

1 Council. I have also served as a member of the Institute's London Council and as a member and past Chairman of the Eastern l

United. States Council. I authored the latest edition of Lamb's Questions and Answers on the Marine Diesel Engine., I have read technical papers on subjects related to diesel engines and die-sel engine repairs at many Technical Conferences in, among other places, Singapore, Lisbon, New York and London. A fur-

'ther statement'of my professional qualifications is attached to this testimony as Attachment 2.

Q. Professor Christenson, what parts of the joint testi-g mony have your particulary sponsored?

A. I have addressed all of the matters regarding the de-sign, manuf acture and rating of diesel engines and their compo-nents. I'have not provided testimony on NRC' regulatory re- -

4 f quirements or on matters purely of metallurgical science.

l

! Q. Mr. Eley, please state your name, address and occupa-

2 tion.

1 i

A. My name is George Dennis Eley. I am a marine consul-l tant employed by Ocean Fleets Consultancy Service, Midatlantic l Corporate Center,1501 Grandview Avenue, Thorofare, New Jersey 4 08086. -

l i

).

1 e

, -- ,-.y ,...--,i,,y---. , . . - , .

Q. Please describe your qualifications and experience.

which are relevant to the matters you address in your testimo-ny.

A. I,am a licensed marine engineer currently employed in providing services to the marine industry, especially with re-spect to large diesel engines. I also act as a marine consul-tant on machinery damage investigations and system design for fuel consumption efficiency, and lecture on fuel technology at the U.S. Merchant Marine Academy and other educational institu-tions. Between 1969 and 1981, I was employed as a marine engi-neer responsible for operating and maintaining diesel engine power plants on ocean-going vessels. Between 1966 and 1969, I was employed by an engineering firm pcoviding consulting ser-vices on the machinery aspects'of shipbuilding projects. Be-tween 1959 and 1966, I was employed in the engine design de-partment of a British marine diesel engine manufacturer. A more complete statement of my professional qualifications is attached to this testimony as Attachment 3.

Q. Mr. Eley, what parts of the joint testimony are you sponsoring?

l l

A. I have addressed matters regarding diesel engine de-sign, manufacture and operation, especially with respect to the l

-v -

replacement crankshafts. I have not provided testimony on NRC regulatory requirements or matters purely of metallurgical sci- l i

ence.

Q. Mr. Bakshi, please, state your name, address and occu-pation.

My name is Aneesh Bakshi. I am a marine surveyor and A.

consultant employed by Ocean Fleets Consultancy Service, with my colleague, Mr. Eley.

Q. Please describe your qualifications and experience which are relevant to the matters you address in this testimo-ny.

. A. I am a licensed marine engineer. I hold a master of science degree in marine transportation management and a bache-lot of science degree in marine engineering. As a marine sur-veyor and consultant, I coordinate machinery (including diesel eng ines) repairs and undertake hull and cargo surveys on ocean-going vessels. Between 1978 and 1981, I was employed as a chief engineer / port engineer coordinating machinery (including diesel engines) repairs and maintenance on ocean-going vessels for a British shipping company. Between 1969 and 1978, I was employed in various engineering capacities associated with r

i l

l e

. . , , - , - . . , - - - - - -, ,,.,_.e- . . . , - . . _ , . ,._,,,,.,,_.._e

marine machinery (including diesel engines) for two shipyards

-and a shipping company. A further statement of my professional qualifications is attached to this testimony as Attachment 4.

Q. Mr. Bakshi, what parts of the joint. testimony have -

you particularly addressed?

.A. I have addressed similar matters as Mr. Eley, except that I have especially focused on issues concerning the re-placement crankshafts and cylinder blocks.

Q. Mr. Bridenbaugh, please state your name, address and occupation.

A. My name is Dale G. Bridenbaugh. I am president of MBB Technical Associates, a technical consulting firm on nucle-ar power plant safety and licensing matters located at 1723 .

Hamilton Avenue, Suite K, San Jose, California 95125.

Q. PleAse describe your qualifications and experience relevant to the matters you address in this testimony.

A. I hold a bachelor of science degree in mechanical en-gineering and am a licensed professional nuclear engineer.

Since 1976, I have acted as a consultant to a large number of domestic and foreign government agencies and other groups on I

l

i l

nuclear power plant safety and licensing matters. Between 1966 ,

and 1976, I was employed by the Nuclear Energy Division of Gen-eral Electric Company in various managerial capacities relating to the sale, service and product improvement of nuclear power reactors manufactured by that company. Between 1955 and 1966,'

I was employed in various engineering capacities working with gas and steam turbines for Gcneral Electric. I have written numerous technical papers and articles on the subject of nucle-ar power equipment and nuclear power pl, ant safety that have been published in technical journals and have given extensive testimony on those subjects. A further statement of my profes-sional qualifications is attached to this testimony as Attach-ment 5.

Q. Mr. Bridenbaugh, what parts of this

joint testimony i

are you sponsoring?

A. I have addressed matters regarding NRC regulatory re-quirements for emergency diesel generators and, generally, en-gineering concerns with respect to the diesels.

Q. Mr. Hubbard, please state your name, address and oc-cupation.

, i

. 9-e

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

r e - - - . . . , - - - , ,---,,..-n n., - , - - -n-- , - , , - - - - - - .,,

A. My name is Richard B. Hubbard. I'am vice president of MHB Technical Associates, which was identified by my col-league, Mr. Bridenbaugh.

Q. Please describe your qualifications and experience relevant to the matters you address in this testimony.

A. I hold a bachelor of science degree in electrical en-gineering and a masters degree in business administration. I am a licensed quality engineer. Since 1976 I have acted as a consultant to a large number of domestic and foreign government agencies and other groups on nuclear power plant safety and li-censing matters. Between 1971 and 1976, I was manager of qual-ity assurance for two departments of General Electric Company engaged in the manufacture of nuclear energy equipment., Be-tween 1964 and 1971 I was employed in various engineering capacities with the Nuclear Instrumentation Department of Gen-eral Electric. Between 1960 and 1964 I worked in various engi-neering capacities for non-nuclear elements of General Elec-tric. I have written numerous technical papers and articles on the subject of nuclear power plant safety and have given exten-sive testimony on that subject. A further statement of my pro-fessional qualifications is attached to this testimony as At-tachment 6.

~

_ 10 li a i s e i i

i i

Q. Mr. Bubbard, what parts of this joint testimony are-you sponsoring?

A. I have concentrated on the areas of emergency diesel generator fdnctions and regulatory requirements, the manufacturing quality of the pistons, cylinder heads and blocks, and the inspections of those components.

Q.. What is the purpose of your testimony?

A. The purpose of our testimony is to demonstrate the validity of the first paragraph of Suffolk County's Emergency Diesel Generator Contention by addressing the specific issues set forth in the numbered paragraphs of the Contention. The

! first paragraph states:

Contrary to the requirements of GDC 17, the emergency diesel generators at Shoreham

("EDGs") manufactured by Transamerica Delaval, Inc. ("TDI") will not operate I

reliably and adequately perform their required functions because the EDGs are over-rated and undersized, improperly desigred, and not satisfactorily manufactured. There can be no reasonable assurance that the EDGs will perform satis- *

' factorily in service and that such

j. operation will not result in failures of other parts or components of the EDGs due to the over-rating or insufficient size of the EDGs or design or manuf acturing l

deficiencies. The EDGs must therefore be

! replaced with engines of greater size and l

capacity, not designed or manufactured by l

TDI.

l 1 l

1

Q. How is.this testimony organized?

~

l

)

A. The testimony will first address'the capacity and ca- ,

pability _ of the EDGs to perform their safety functions, .and the regulatory standards for operating service and . safety functions which the EDGs and their ' components are required to meet.

Then, we will address the safety significance of deficiencies J

in four major EDG components: the AE model pistons, the cylin-der heads, the crankshafts, and the cylinder blocks.

Q. What are the-EDGa?

A. They are TDI model DSR-48 diesel engines with 8 cyl-inders in line, having a 17" stroke and a 21" bore. When com-bined with their generators, the EDGs are intended to provide reliable onsite emergency power to the Shoreham plant in con-formity with 10 C.F.R. Part 50, Appendix A, General Design Cri-i terion 17 ("GDC 17") . ,

~

1 j Q. Have the EDGs experienced problems?

f A. Yes. A broad pattern of deficiencies in critical TDI i

diesel engine components has become evident at Shoreham and at

! other nuclear and non-nuclear facilities. These deficiencies stem from inadequacies in design, manufacture and quality con-trol by TDI, and resulted in the NRC Staff losing confidence in, the reliability of TDI diesels including the EDGs.1/

1/ See Board Notification 84-020, February 13, 1984, " Report of Meeting of Representatives of the Transamerica Delaval, Inc. (TDI) Emergency Diesel Generators Owners' Group."

l

Q. How did LILCO address these problems?

A. LILCO and the other nuclear utilities issued the TDI Diesel Generators Owners' Group Program Plan (" Owners' Group Program"),l/ which embodied three major efforts:

1. Resolution of 16 Known Generic Problem Areas;

2. Design Review of Important Engine Compo-nents and Quality Revalidation ("DRQR") of Important Attributes for Selected Engine Components and;

3. Expanded Engine Testing and Inspection.

Q. Has the Owners' Group Program adequately resolved all th'e deficiencies in the design, manuf acture, and QA/QC of the EDGs?

A. No. We conclude, for the reasons set forth in this testimony, that this after-the-fact investigation of the EDGs conducted by the Owners' Group and its principal subcontrac-tors, Failure Analysis Associates ("FaAA") and Stone and Webster Engineering Corp. f ails to provide a sufficient level of assurance that the EDGs and their critical components, the 2/ Board Notification 34-051, March 12, 1984, " TDI Diesel Generators Owners' Group Program Plan," dated March 2, 1984.

\

13 -

i i

e

, - , - - - -- ,or - e - ~ . , - . - , , , - , , , - - --,n. .

mn... a- , ~--, .,-- - - -

pistons, the cylinder heads, the crankshafts and the cylinder blocks, will operate reliably and with appropriate capacity and capability to adequately perform their required functions and that additional parts and components of the EDGs will not fail.

Functions _and_ Requirements for EDGs i,

Q. What regulatory requirements must the EDGs meet?

A. The EDGs constitute the onsite electrical power sys-tem for.the Shoreham plant. They must meet the requirements of GDC 17, which stipulates that, assuming the absence of the offsite electric power system, they shall provide sufficient capacity and capability to assure that (1) specified acceptable fuel design limits and design conditions of the reactor coolsnt pressure boundary are not exceeded as a result of anticipated op-erational occurrences and (2) the core is .

cooled and containment integrity and other vital functions are maintained in the event of postulated accidents.

(Emphasis added). LILCO has described in the Shoreham Final Safety Analysis Report ("FSAR") how the requirements of GDC 17 are addressed.

Q. What are the major required safety functions for the i EDGs stated in the FSAR?

y. -

m

(

1 4

, s A.g .The FSAR provides that the EDGs must be fully capable performing two critical safety functions.

~

First, l

Lof reliabl L ,

it must b jemonstratedithatthe,EDGswillstartandreach .

rated frequency and voltage within 10 seconds.1/ Second, it L

further must be demonstrated'that the EDGs'have sufficient load

> carrying capability to satisfy the continuous and overload per-form nee rating.1/ .

Q. What is the rating for the EDGs?

A. Diesel engines'for generators, such as the EDGs, are !

rated by their manufacturers as to engine speed and horsepower, andasbeingcapableofNeetingcer'ainspecificperformance t criteria within allowable temperature and pressure limits. TDI rated the EDGs'at a speed of 450 RPM and at 4,890 horsepower

- (about 611' HP per cylinder), with a.perfor'mance rating as required by LILCO's contract specification 5/ and the FSAR for Shoreham.- Section 8.3.1.1.5 of the FSAR requires each EDG to be rated to operate continuously'(8,760 hours0.0088 days 0.211 hours 0.00126 weeks 2.8918e-4 months , or one year) at full load of 3,500 kW (with maintenance intervals required by 3/ FSAR, p. 8.3-14.

4/ FSAR, p. 8.3-5.

5/ Stone and Webster Specification for Diesel Genera' tor Sets, SHl-89, October 3, 1973 and addendum 1 to 5 thereto.

t 15 -

- A e- -w. .-m,y _ - - -- 3-.-, y , . - - e--g y w. ,p * - p ,,,,..w%.

- , _ . . . --*w-6 w- *~w We v 7 * -M-T we*- ie vve+=

TDI) and for 2 hours2.314815e-5 days 5.555556e-4 hours 3.306878e-6 weeks 7.61e-7 months per 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months at overload of 3,900 kW (without reducing.the maintenance interval established for the continuous rating).

Q. Isn' t this rating different from the actual loads the LEDGs are likely to see in service?

A. Yes. The maximum continuous load imposed on the EDGs is less than the continuous rating, and the maximum intermit-tent load is less than the 2 hour2.314815e-5 days 5.555556e-4 hours 3.306878e-6 weeks 7.61e-7 months rating.5/ However, the rat-ing requirement is to provide necessary confidence that the maximum actual power demands will reliably be met and that ac-cordingly the requirements of GDC 17 will be fulfilled. There-fore, the proper criterion for whether the EDGs and their com-ponents can, satisfactorily withstand operating conditions is ,

whether they can be expected to operate at the rated levcis without experiencing failures or incipient failures.

Q. Hasn' t LILCO applied to the NRC to reduce the perfor-mance rating for the EDG and revise'the FSAR accordingly?

A. It appears so. In a recent letter to the NRC,2/

6/

FSAR, Section 8.3.1.1.5, p. 8.3-8; see also FSAR Tables 8.3.1-1 and 8.3.1-2.

Leonard, Jr. (LILCb) to Harold R. Denton 2/ Letter from J. D.

(NRC), dated July 3, 1984, SNRC-1065.

e

,ew,- ,, , - - - - - -

e-w, - , , - - , - - , - , , - ------

_ _ w - - . _

i LILCO proposed to remove one service water pump from auto-start on an accident-signal, thereby reducing the maximum load on EDG 103 (prior to 10 minutes) from 3880.7 kW to 3442.7 kW. Section 8.3.1.1.5 of the FSAR would be revised to read:

l The rating of each diesel generator is 3500 kW. The required load on each diesel gen-erator is enveloped as follows: .

Continuous 3475 kW 2 hr per 34 hr period 3500 kW Q. What is your reaction to this proposal?

, A. , Insofar as it would reduce the performance rating for the EDGs, in contrast to reducing the actual loading on the EDGs, we believe the proposal would be detrimental for providing confidence that the EDGs* can operate reliably. The existing performance ratings already lack conservatism. For example, the maximum loads (af ter 10 minutes) on EDGs 101 and 102 are approximately 3400 kW each; the continuous rating is only 100 kW more, a margin of only 2.9%.8/ The EDG 103 load condition may be even less conservative. While the proposed change reduces the " prior to ten minutes" load to less than 3500 kW, LILCO's proposed change also includes the possibility 8/ FSAR Table 8.3.1-1, 4 of 4.

- 17 -

s

---ww-,<r-----e--- -

w-- + ------.--*-w, y -

---_..,-.----e q ,-- - , , %-y<

of manual start of the decond service water pump after ten minutes. Depending on what other equipment is available, it

. ay m not be possible to stop the second RHR pump coincident with this post-ten minutes load, and .the EDG 103 load would then be

'back up to-approximately 3900 kW. The usual practice for die-sel engines in non-nuclear electric generating plants and in marine applications is to operate them at only about 75-85% of their -ratings, in order to provide a conservative margin of safety.9/ To provide a similar safety margin for the EDGs, LILCO should have procured diesels with a continuous rating of at least 3910 kW. Clearly there should be no further reduction in the margin of confidence intended to be supplied by the cur-rent EDG rating.

Q. H'ow do manuf acturers of large diesel engines like the EDGs establish the rating of their engines? _

A. In our experience, manufacturers generally establish the rating of a new model engine by running thd engine on a test stand for thousands of hours at the load levels at which they. seek to rate the engine. The engine is inspected Mr. Museler of LILCO agrees. Deposition of William J.

9/ Museler (May 22, 1984) ("Museler Deposition") at 9. (Ex-hibit 1).

l l

l- ~~ - = - - . . - , . - . . - _ , _ __ _ _ , _ . , _ , , , .

periodically. If adverse effects appear, such as cracking or unusually high wear rates, changes may be made to attempt to remedy the problems and the engine testing will begin 'again.

In short, the rating process is empirical and involves many test hours.

Q. How did TDI establish the rating of the EDGs?

A. TDI establishes the rating for its engines by testing.10/ TDI'uses the DEMA standard for rating, which calls ,

for continuous operation at full load with a 10% overload for two hours in each 24 hour2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months period.11/ However, TDI tested the l

first DSR-48 engine (the model of the EDGs) rated at about 610 HP per cylinder for only 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months or less.12/ In our opinion, such a test is grossly inadequate to determine the proper rat-ing of the EDGs. Even the diesel expert for LILCO and the TDI Owners' Group, Dr. Chen, testified that to establish the proper rating for the first DSR-48 model engine it should have been tested in each 24 hour2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months period for 22 hours2.546296e-4 days 0.00611 hours 3.637566e-5 weeks 8.371e-6 months at 3,500 kW and for 2 hours2.314815e-5 days 5.555556e-4 hours 3.306878e-6 weeks 7.61e-7 months at 3,900 kW for at least 1,000 hours0 days 0 hours 0 weeks 0 months , be fitted with 10/ Deposition of Clinton Mathews (May 8, 1984) ("Mathews Dep-osition") at 27, 29. (Exhibit 2).

11/ Id. at 25-26, 30.

12/ Id. at 35.

i '

t l

strain gauges to estimate stresses, and then disassembled to look for wear rates and indications.13/

Q. Does TDI believe the EDGs were properly rated by the 24 hour2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months test of the first DSR-48?

A. Yes, because according to Mr. Mathews, vice president and general manager of the Engine Compressor Division of TDI, before the 24 hour2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months test, most components of the DSR-48 had been tested extensively in other model TDI engines at equivalent loads.li/ Those components common to the other TDI engines and tested in them included the cylinder heads and pistons, but not the crankshafts or cylinder blocks.15/

Q. Do you agree with this position?

~

A. No. We strongly disagree with TDI that testing of certain common components in other model TDI diesels adds suf-ficiently to a 24 hour2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months test to result in properly rating the DSR-48 engine. In testing an engine to establish its proper rating, it is imperative to adequately test the engine as a 13/ Deposition of Simon K. Chen (May 15, 1984) ("Chen Deposi-tion") at 55-58. (Exhibit 3).

14/ Mathews Deposition at 32-33. (Exhibit 2).

M / H . at 36-40.

____.___.________________m_ . . _ _ _ . _ _ _ _ _ . . _ _

whole to determine the operation and interaction of its many components. Further, we do not believe any engine can be prop-erly rated when its crankshaft and cylinder block have not been sufficiently tested. .

Q. Do you believe the EDGs were properly rated by TDI?

4 A.. No. We believe the EDGs are over-rated and under-sized.

Q. What do you mean by the term "over-rated"?

A. By the term "over-rated," we mean that the perfor-mance rating of the EDGs is higher than the EDGs are capable of meeting without suffering adverse consequences, such as cracking of components which may lead to catastrophic failure of the engines. In other words, the EDGs_do not and will not operate reliably at their rating of 3,500 kW continuously and 3,900 kW two hour overload; therefore, the rating given the EDGs by TDI was improperly high.

Q. What do you mean by the term " undersized"?

A. Simply that the EDGs are too small to reliably oper-ate at the levels required by the contract specification and

- Section'8.3.1.1.5 of the Shoreham FSAR, and thus the EDGs do not meet the requirements of GDC 17.

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

Q. - Does the operating history of'the EDGs confirm your position that they are over-rated and undersized? j

\

A. Yes. Operation of the EDGs has been confined to their testing by LILCO. . The EDGs are supposed to be capable of .

I meeting their performance ratings. Moreover, they are expected to last for the_ entire 40-year life of the Shoreham plant. Yet after only about 800 to 900 hours0.0104 days 0.25 hours 0.00149 weeks 3.4245e-4 months of testing all three EDGs had experienced extensive cracking of components.16/ These includ-ed: ,

1. Cracks in'23 of the 24 piston skirts;11/

2. Cracks in three cylinder heads resulting in water

~

leaking into the cylinders;18/

3. The severing of the crankshaft on EDG' 102 and cracks

~

on the crankshafts of.the other EDGs;19/ and 16/ See generally Board Notification 83-160 dated, October 21, TR3. (Exhibit 4).

12/ 10 C.F.R. 50.55(e) report dated November 16, 1983. (Ex-hibit 5).

i 18/ 10 C.F.R. 50.55(e) report dated April 15, 1983. (Exhibit 6).

I 19/ NRC Information Notice No. 83-58, August 30, 1983.

I -

22 -

- - - -,..w .me, , -

m- -r ,. g. -- , ,w - - , e

4. Cracks on the block tops and camshaft gallery areas of all three cylinder blocks.10/

- These f ailures of these four major components evidence that the ,

EDGs are over-rated and undersized. .

Q. Didn' t LILCO replace all of these four major compo-nents?

A. LILCO replaced-the crankshafts on the three EDGs with crankshafts having a larger (nominal 12" diameter) crankpin. 1 All piston skirts were replaced with TDI model AE skirts.

LILCO replaced the block on EDG 103 with a different design TDI block. Two of the cylinder blocks (on EDG 101 and 102) have :

extensive cracks on the block top, running from the stud holes radially and vertically to the cylinder bores, and cracks in the camshaft gallery areas. LILCO has not replaced these blocks and intends to use them at Shoreham during full power operation of the plant. All of the cylinder heads were replaced with TDI heads of the same design, but allegedly of better manufacte_e and quality.

i 20/ Design Review of TDI R-4 and RV-4 Series Emergency Diesel Generator Cylinder Blocks and Liners, June 1984 ("FaAA I Block Report"). (Exhibit 7).

l e

- - - - r,,-- .-r- ---y. --. . -- .,,- y ,--. . - -s.--._ . - - , . , , , , ,- . _ _ _- _ _ _ _ _ _ - _ _ _ - _ _ - - - -

u 1}. Did the replacement of these major components poten-j.

tially impact other safety functions of the EDGs? l

[.

A. Yes. In 1976 over 300 tests were conducted at the TDI factory to establish the capability of the prototype EDGs ,

to start and accept load. All these starts were performed on one EDG (EDG 101). No more than two failures in the 300 tests g were allowed.21/ Since these tests, there has been a wholesale replacement of critical' engine components including the four major components discussed in this testimony.

Q. Has LILCO adequately requalified the startup reliability of the modified engines?

A. No. We believe the 1976 prototype qualification tests are no longer valid and applicable, and additional quali-fication tests are required to demonstrate startup reliability.

' LILCO has proposed to perform only 100 starts on a single EDG, allowing no failures in 23 consecutive starts, and only one failure in the other 77 starts.22/ We believe that the origi-nal criteria for startup prototype qualification should be i

21/ FSAR, p. 8.3-14 and 8.3-15.

22/ Shoreham Diesel Generator Recovery Program Summary, atta-l ched to LILCO letter SNRC-1003, January 6, 1984.

f-7-~ _ . - -,- , - . , . , . - - .

implemented, which would require 300 starts on a single engine, or a minimum of 100 starts on each of the three engines.

Q. Has the replacement of the pistons, cylinder heads, crankshafts and EDG 103 cylinder block solved the problems ex-perienced by the EDGs in the past?

A. No. The EDGs are still over-rated and undersized, improperly designed and not satisfactorily manufactured. The reasons for this conclusion will be presented in detail in our testimony concerning each of the current four major components of the EDGs.

MODEL AE PISTONS Q. How does Suffolk County's Contention relate to.the TDI pistons in use at Shorehap?

A. The EDG Contention provides that its first paragraph is supported because:

All AF piston skirts in the EDGs were replaced with TDI model AE piston skirts. The replacement AE pistons are of inadequate design and manufacturing quality to satisfactorily withstand operating conditions because:

(a) The FaAA report conclusion that cracks may occur but will not propagate improperly depends on a fracture mechanics analysis of an 1

~

ideal situation which is not valid for the actual conditions which may be experienced by the Shoreham diesels,

.(b) excessive side thrust load, which could lead to catastrophic failure, has not been considered adequately, and 1

l (c) the analysis does not adequately con-sider that the tin-plated design of the pistons l could lead to scoring causing. excessive gas blow-by, and thereby causing a failure of proper

- operation.

Q. Why were the AE model piston skirts installed in the EDGs?

A. The AE piston skirts were installed after 23 TDI model AF-piston skirts in the EDGs were discovered to have lin-ear indications, that is, cracks, in the crown-to-skirt stud attachment bosses. Failure Analysis Associates ("FaAA"), an

-se..- '

._ organization retained?by LILCO (through'its att ineys) and the t

TDI Owners' Group, has published a report entitled "Investiga-tion of Types AF and AE Piston Stirts" dated May-23, 1984-(the "FaAA. Piston Report") ,.21/ which concluded that the cracks in the AF piston skirts were fatigue cracks.

Q. What are the bases for your conclusions that the AE l

pistons at Shoreham are inadequately designed and 23/ FaAA Report 94-2-14. (Exhibit B).

i.

-,y ...._,-.r - - , . - - _ , , ,y--.. _

w _.., .-

unsatisfactorily manufactured, as set forth in the EDG Contention?-

A. The bases for our conclusions are described in detail below.

Cracking of AE Piston Skirts Q. Did FaAA conclude that the AE piston skirts might crack?.

A. FaAA conducted a finite element stress analysis of 4

the AE piston skirt, which showed that cracks may initiate in the skirt.21/ FaAA also carried out experimental measurements of strain under static load in the AE piston skirt,25/ which

. predict that cracks will not initiate in the skirt under the .

cyclic stress levels obtained in the experiments.26/ The dis- -

agreement between the finite element analysis and the experi-i mental results is 28%,.which FaAA maintains is "quite good" agreement.21/

24/ FaAA Piston Report at 6-1.

25/ Id., Section 3.

26/ Id. at 6-1.

27/ Id. at 5-1. The disagreement between an earlier finite element analysis and the experimental results was 33%.

Initial FaAA Piston Report, February 27, 1984, at 5-7.

i y __ _ _

Q. Do you agree that the 28% disparity is "quite good"?

A. No. That disparity is the difference between two op-posite conclusions -- cracking or structural integrity -- which are critical to the results of FaAA's study.

Q. Which is more reliable -- the finite element analysis result or the experimental results? -

A. The usual methodology is to confirm the finite ele- l 1

f ment analysis by the stress experiments. The finite element analysis, when properly done, may be an excellent tool for evaluating a structure. It tends to be non-conservative (that i is, it would be expected to show less likelihood of cracking than experiments) because it averages the properties of the piston skirt material and ignores possible imperfections in the material. Because the experimental results differed signifi-cantly from the finite element analysis results, it would ap-The experi-pear to us that the experiments were inadequate.

! ments should have been carried out until crack initiation was shown, and then analyzed. Where, as in this case, the experi-ments do not confirm the analysis, additional work is required.

Instead of doing that additional analysis, FaAA concludes that the 28% disagreement of the results is acceptable and could be accounted for by incorrect assumptions in the finite element

{

- 28 -

f

,--.m ,.--- . , - - - - - - - - - . , , , . , . , . , , - ,

model,. omissions or approximations in the finite element technique, or inaccuracies in the experiments, or all of the above.11/ This is not a helpful conclusion, because the two results -- that cracks will initiate or will not occur -- are opposing. We believe that this conflict has not been adequate-

.ly investigated. We note, for example, that an unstated number of strain gauges in the stud boss area did not work.21/ Since no attempt has been reported to qualify the relative accuracies of the analytical and experimental techniques, and given the importance of the conclusion in terms of the safety require-ments for Shoreham, we believe the greater weight must be given to the results of the finite element analysis -- that cracks are predicted to initiate.

Q. Do you believe the FaAA Piston Repor't~ underestimates the probability that cracks will initiate in the AE skirt?

A. Yes. FaAA determined for purposes of its finite ele-ment analysis and experiments that "The maximum stresses in the piston skirt under peak firing pressure are of primary inter-est. This pressure is approximately 1670 psig as independently measured by FaAA and reported by TDI.=10/ To justify a peak 28/ Id.

2_9/' FaAA Piston Report at 3-6.

0 3,0f M.; see also M . at 4-1.

m _M-firing pressure of 1670 psig, FaAA cites only TDI reported val-ues'for a DSRV-16-4 engine at Grand Gulf Nuclear Station,21/

and FaAA pressure measurements of 2 cylinders at Shoreham which FaAA has acknowledged to be unreliable and too low.32/ In fact,,the peak firing pressure in cylinders of the EDGs at full load (3500 kW) is known to.tue as high as 1750 psig, and at overload (3900 kW) the peak firing pressure is at least 1800 psig. The stresses on the AE piston skirt used by FaAA in its analysis and experiments are thus understated. -

Q. What evidence do you have that the peak firing pres-sures in the EDGs are as high as 1750 to 1800 psig?

A. Test documents for the EDGs and for other DSR-48 die-sel engines establish these maximum peak firing pressures.

These documents are attached as Exhibit 46.33/ The test data show numerous peak firing pressure readings of greater than 1670 psig for the Shoreham engines at 100% load (the 1/24/76 run on EDG 102 shows 1750 psig, for example) and pressures as 31/ -Id. at 3-14 (Ref. 3-1) and at 4-7 (Ref. 4-2).

32/ Id. at 4-7 (Ref. 4-1); Emergency Diesel Generator Crank-sEaft Failure Investigation, Shoreham Nuclear Power Sta-

~~

l tion, FaAA, October 31, 1983 (FaAA 83-10-2) at 4-9.

33/ See Exhibit 46 at documents 5-9.

l I

.-y ,. - , -,-,e. ,-w -m. - - , . - , _ - - - , , , - - - - - , , , , . , - -

high' as 1800 psi are reported for the overload condition-(3/19/76 run on EDG 103. This evidence contrasts with the l readings on an EDG taken by FaAA. One must also remember that l firing pressures differ from cylinder to cylinder and engine to

. . engine. TDI gives no specific authoritative peak firing pres-sure for the DSR-48. Rather, its manual for operation of the EDGs' permits a variance in peak firing pressures of the cylin-ders in one engine of + 100 psi.li/ This means that any single peak firing pressure read in one cylinder may be exceeded in i another cylinder by 200 psi, so that firing pressures may be even greater than 1800 psi in the EDGs.

Q. What is the impact of the higher actual peak firing

, pressure on the FaAA Piston Report?

A. The higher actual peak firing pressures mean that cracks are more likely to initiate in the AE piston skirts in the'EDGs than FaAA predicts. FaAA underestimates the crack initiation in 3 respects concerning firing pressures. First, FaAA uses a too-low peak pressure of 1670 psig for its finite .

! element analysis and the reported strain gauge tests. FaAA tested the pistons to 2000 psig, but only reported the data at 31/ TDI Instruction Manual at 8-3 (Exhibit 9) 4 . .e. --~.--, .w,- ..,..---r -- _--- - . . - - - . .e- ..w,m.e. - - - - ,--.-..--,-.--v, .- m- - - - --- ,. , - - ,. -,s-

s the 1600 psig point. .Second, certain strain gauge measurements are limited'to a maximum of 1600 psig.11/ Third, FaAA made no analysis or strain gauge experiments at overload (3900 kW),

even though the EDGs have a 2 hour2.314815e-5 days 5.555556e-4 hours 3.306878e-6 weeks 7.61e-7 months per each 24 hour2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months overload rating and an actual maximum peak load of 3881 kW. TDI has testified that-the peak firing pressure of the EDGs at 3900.idi is about 1800 psi.}6/ These factors would, if taken into con-sideration by FaAA, result in a much greater likelihood of AE piston skirt crack. initiation than-predicted in the FaAA Piston Report. -

Q. Aside from the peak firing pressure, are any other issues of particular concern to'FaAA's conclusions concerning

[

j crack initiation? ,

A. The initial size of the gap between the outer ring of the AE skirt and the crown is, according to FaAA, important in predicting whether or not cracks.will initiate in the skirt.21/

- The FaAA Piston Report states

i 35/ FaAA Piston Report at 3-6 to 3-7, 3-16 and 3-17, 3-19.

36/ Deposition of Gerald Edgar Trussell (May 7, 1984)

("Trussell Deposition"), at 128-29. (Exhibit 10).

31/ FaAA Piston Report at 8-1; see, also Figure 3-2 for an 11-lustration of this gap.

e

The experimental results of Section 3 showed that the stresses due to pressure are dependent on the initial gap size, go, because this parameter influences the gap L

closure pressure and load transfer between F inner and outer load rings. As shown in Figure 3-2, the initial gap can vary from i

O.007 to 0.011 inch and still be within TDI !

- specified tolerance.38/

Neither FaAA nor the TDI Owners Group personnel has measured i the initial gaps prasent in the AE pistons in the EDGs.39/

Based upon foundry practices and the lack of effective quality assurance at TDI,_ discussed below, it is quite likely that the TDI tolerances'may be exceeded. Actual measurements of the gaps in the AE pistons at Shoreham would be useful in testing

' FaAA's assumption that all AE pistons have gaps within TDI's tolerances.

4 Another factor bearing upon the li,kelihood of crack

^

initiation is the tensile properties of the skirts.10/ We do not know the actual tensile properties of the AE skirts at Shoreham, but we note that the range of values reported for 38/ Id . at 6-4.

39/. " Design Review and Quality Revalidation Report, TDI Diesel f Generators For Shoreham," TDI Diesel Generator Owners Group, June 29, 1984 (the "DRQR Report"), Vol. 5, Pistons, l at B1 to B6. (Exhibit 11).

30/ FaAA Piston Report at 6-5.

r c

I l

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

typical material used at TDI shows ultimate tensile strengths as low as 85 Kai.$1/ If such a piston-were subjected to the higher firint, pressures possible (1750 psig or higher), the conclusions regarding crack , initiation would certainly be invalid.

Q. Do you agree with FaAA's conclusions that even if cracks do initiate in the AE piston skirt, they will not propa-gate?

A. No, because that conclusion is based upon a highly theoretical fracture mechanics analysis which does not take all potential effects into account for predicting crack growth under the actual conditions that will be experienced at Shoreham. The FaAA analysis assumes:

(1) complete adherence to TDI drawing dimensions of the AE skirt (and crown);

(2) the AE piston material is isotropic, meaning it

,is free of any small inperfections such as sand inclusions or grinding marks, and with no subsurface defects such as hot tears or slag 41f M. at 2-7 -

e

inclusions, with the ultimate tensile strength

~

uniform in all directions; (3) a non-corrosive operating environment free of l

gases, water or vapor; (4) stresses resulting from a maximum peak firing pressure of 1670 psi; and (5) a uniform skirt temperature, both circumferen-tially and axially.42/

Each of these idealized assumptions is incorrect in terms of 1

the "real world."

and the im-Q., Explain why each assumption is incorrect, 4 pact of the error on FaAA's crack propagation analysis.

t A. (1)' The dimensions of each AE piston at Shoreham are

. not perfect. Only a very limited dimensional check on a sam '

pling basis was made on piston groove and ring height and pis-ton pin bore diameter and depths on the AE pistons at Shoreham.

No dimensional check was made of other parts of the piston 42/ FaAA did not independently measure the thermal gradience

~~

in the AE piston skirt. Harris Deposition at 41. (Exhib-it 12).

e

skirt, including the thickness of the boss greas or the gap between the piston skirt and crown.f3/ Even relatively small dimensional differences in the skirt and in the assembly of the skirt and crown would change the mathematics of FaAA's analy-sis, and could influence the results.

(2) The AE piston skirts in the EDGs are not free of defects. They are known to have some small defects, and it is highly likely that many more imperfections are present. At Shoreham, only 10 of the 24 AE piston skirts were subjected to liquid penetrant tests at the bosses for bolt attachment to the crown.di/ These tests did disclose some defects, but in any case were totally inadequate to determine whether there are small imperfections on the surface or subsurface of the AE skirts. Such small imperfections are likely to be present in _

the skirts in the EDGs. TDI does not use vacuum processes to ensure a dirt-free casting. Indeed, the foundry is poorly lighted 'and has a dirt floor, which increases the likelihood of sand or' slag inclusions. Control of scrap material for cast-ings is rather informal. Effective quality control is absent, 13/ DRQR Report, Vol. 5, Pistons, at B1-B6.

44/ Id. Eddy-current inspections See FaAA were conducted by FaAA on 12 Piston Report at 7-1 and skirts on the EDGs.

discussion below.

J l l

so that small imperfections are unlikely to be discovered. Mr.

William Foster of the NRC's Vender Inspection Program staff, who had participated in a number of NRC inspections at TDI, l

- stated recently that the nature and number of violations and non-conformances at TDI indicated to him that the TDI QA system was "inef fective."11/ The presence of'even a small imperfec-tion would permit a crack to initiate and propagate at. stress levels below those predicted.by FaAA as necessary for initiation and propogation. If a crack initiates in an area of -

the skirt where imperfections are present, its. growth may be entirely different than as calculated by FaAA, which assumed no flaws in the material. With the presence of some imperfec-tions, FaAA's fracture mechanics analysis is invalid.

(3) The environment of the piston dufing EDG' operation is not a vacuum. Combustion gases are present, and there may be small amounts of water or vapor. If a crack ini-tiates in the skirt, thede gases will tend to corrode the crack edges and hasten crack propagation. Corrosion products formed on the crack opening of a skirt during EDG operation will act I

as wedge when the crack closes (after EDG operation ceases),

15/ Deposition of William Foster (May 22, 1984) (" Foster Depo-sition"), at 16. (Exhibit 13).

i 1

I

b L producing additional crack growth. The FaAA fracture mechanics analysis does not consider these factors at all.

(4) FaAA's analysis postulates stresses resulting from a peak firing pressure of 1670 psi. The proper maximum peak pressure of 1800 psi, as discussed above, would result in greater stresses and a higher. likelihood of crack propagation.

(5) The temperature around the skirt is not uniform.

Actually, the side of the piston skirt taking the piston thrust on the firing downstroke becomes much hotter during EDG operation than the side taking the piston thrust on the com-pression upstroke. The temperature of these TDI pistons will-be even higher than 'is normally expected in other makes of en-gines where the initial side thrust is designed to be much lower, as discussed below.- FaAA assumes that the piston skirt is "nearly isothermal",16/ when in fact, one side of the skirt runs at a much higher temperature than the opposite side. Es-timates for the piston skirt temperatures were provided by TDI based on "templug" measurements taken on a non-Shoreham engine operating at only 213 BMEP.$1/ The EDGs operate at 225 BMEP, l

l 46/ "The Influence of Thermal Distortion in the Fatigue Per-

~~

formance of the AF and AE Piston Skirts", June 1984 (FaAA-84-5-18) (the "FaAA Piston Thermal Distortion Re-port"), at 2-7.

47/ Id. at 2-6, 2-7.

i.

and-would'therefore have higher piston skirt temperatures.

Q. Given all of these variations from FaAA's idealized assumptions, is it possible.to predict accurately how cracks in the AE skirt will propagate?.

A. No. It is not possible to make accurate predictions of crack propagation in the AE skirts, given all of the possi-ble variables. However, the FaAA analysis would have been far more useful if actual properties of the AE piston skirts in the EDGs had been recorded, to the extent possible, and sensitivity analyses performed to account for a range of potential vari-ables. Thus, the principal dimensions of each AE skirt at Shoreham could have been measured, especially in the boss area.

The gap between the outer ring of each skirt and the attached crown could.have been measured. Each AE piston skirt in the -

h EDGs could have been inspected for imperfections, especially in i

the boss area, by liquid penetrant tests, magnetic particle tests, eddy current examination and radiographic inspection.

The tensile properties of each skirt could have been sampled.

The analysis could then have been performed using a range of a

more realistic peak firing pressures (up to 1800 psi) and including the combined effects of maximum side thrust and its corresponding gas pressure, temperatures, and environmental

conditions.- The analysis could have included sensitivity tests to take into . consideration the potential for undiscovered di-mensional variations, defects in the skirt and differences in-

. tensile strength, and the possibility of multiple cracks. Such analyses would give a far better prediction of crack propaga-tion than the idealized study performed by FaAA.

Q. What else, besides the inspections and crack propaga-tion analyses you suggest, would be necessary to give adequate confidence that the AE piston skirts are adequate for operation 4

at Shoreham?

A. First, an adequate crack initiation analysis should be performed, using actual data as to dimensions, tensile

pro'erties, p imperfections, and gap sizes of the AE skirts at ,

Shoreham, and the appropriate peak firing pressures of up to 1800 psi. Experimental stress tests should confirm the results of finite element analyses, or a more refined finite element analyses or better experiments should be performed. The AE pistons could be instrumented and tested during EDG operation for additional experimental data. These analyses could predict multiple'eracks initiating with larger initial sizes, thereby affecting.the crack propagation analyses. The design deficiencies involving excessive pisten side thrust load and I

40 -

I .

l-

, - . . - - ,m.--- -

m ,-m-. -- .-,.n,. + m---- - -

+=---=--w-e -----e. ,-- - . - -

^

)

l tin plating of the skirt would have to be considered, as L .

discussed below. ' Finally, the AE piston skirts would have to

, be tested and inspected adequately in 'the EDGs.

i Q. Does FaAA'believe the AE piston skirts have been ade-quately tested and inspected?

A. Yes. FaAA has concluded that on the basis of the re--

sults. of its stress a'nalyses (which were contradictory as to crack initiation)' and "the results of inspections of engine-operated AEf skirts," the AE piston skirts "are adequate for un-limited life."18/ We strongly disagree that the AE skirts have been adequately tested or inspected to justify any conclusions about their expected life.

t Q. What inspections'was FaAA referring to?

A. FaAA was referring to inspections of 15 AE skirts, as follows:

(1) 12 AE' skirts of the 24 skirts were

- subjected to_ eddy-current inspections after over 300 hours0.00347 days 0.0833 hours 4.960317e-4 weeks 1.1415e-4 months of total operation each (including 100 hours0.00116 days 0.0278 hours 1.653439e-4 weeks 3.805e-5 months at full load), and no

" relevant indications", were found; (2) One skirtiin an RV-16-4 engine was inspected after over 6,000 hours0 days 0 hours 0 weeks 0 months of 48/ FaAA Piston Report at 8-1.

9

' _ 41 -

f v- ~ en- ,w-m 4 -- e + un ee ,- ee-, w- -.- - - - - - - , - - , - - - , - - , - -- , -- -----r - - - ~1 w--

t operation at a peak firing pressure of about-1200 psi, with no " relevant indica-tions" found; and (3) Two. skirts from a TDI R-5 development en- -

gine were inspected after operating at a peak pressure of 2000 psi or more after over 600 hours0.00694 days 0.167 hours 9.920635e-4 weeks 2.283e-4 months , with no finding of "rele-vant indications."49/

Q. Why don't you believe this experience and these in-spections are adequate to support FaAA's conclusions?

A. For several reasons. First, fifteen skirts is simply too small a number from which to reach any general conclusions, particularly without a valid statistical analysis.

Second, the inspection of only 50%, rather than 100%, of the AE skirts on the EDGs is inadequate. Mr. William Foster, the NRC Staff official with. responsibility for vendor inspec-tions of TDI, has testified that TDI has an ineffective quality l

control program, and consequently inspection on a sampling plan basis of TDI components "would not tell you anything."52/ In fact, Mr. Foster testified that even a 100% inspection of TDI components would not identify all defects.51/ We agree.

49/ Id. at 7-1.

50/ Foster Deposition at 14-16, 54-55, 82. (Exhibit 13).

l 51/ Id. at 55.

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

Third,'the numbar of hours and the amount of full loads and overloads run on each AE skirt at Shoreham are insufficient

'to\ reach conclusions about their expected life. To meet the rating specifications of'the EDGs, the AE skirts must be capa-ble of running many thousands'of hours, including significant hours at overload at 3900 kW. The AE piston is supposed to last the' lifetime of the Shoreham plant -- 40' years.M /

Testing them for only 300 hours0.00347 days 0.0833 hours 4.960317e-4 weeks 1.1415e-4 months without significant, if any, overload does not begin to be adequate. It is also important

- to' note that TDI did not test the AE piston before supplying it to customers in the field.M/

3 Fourth, the AE skirt in the RV-16-4 engine was operated at a peak firing pres 1ure of only 1200 psi, while the EDGs have a

. peak firing pressure of about 1700.to 1800 psi at full load and overload. Thus, the operation of that single skirt was at such low stress as to be useless for purposes of' reaching any con-clusions relevant to the AE skirts in the EDGs.

Fifth, the two piston skirts operated in the TDI R-5 en-gine are.of limited relevance. The R-5 engine is significantly 5_2 /

Trussell' Deposition at 111-13. (Exhibit 10).

53_/ Id.. at 107.

43 -

e m" :.m m ..

. different from the EDGs, including its operating speed (514 RPM). This would change the inertia effects which in turn low-ers the piston lateral loading. Therefore, before determining a

. the impact of the R-5 skirts on the Shoreham AE skirt report, .

study would have to be made analyzing the effects of the dif-forent parameters.

Sixth, the referenced inspections were incomplete and the standards for acceptance were unsatisfactory.

Q. Please be more specific about your last point.

A. FaAA stated that only eddy current examination was performed on the Shoreham piston skirts.Ei/ Further, only cer-tain portions of the skirt were subjected to the eddy current examination, namely, " machined areas on the boss where color contrast penetrant show (sic) -linear indications greater than 1/32 inch.=55/ This means that linear indications smaller than 1/32 inch, non-linear indications such as sand or slag inclu-sions, and areas of the boss which were not machined were omit-ced from consideration. As we noted earlier, even small 54/ FaAA Piston Report, at 7-1.

55/ FaAA NDE Procedure 11.5, November 2, 1983, para. 6.1.

(Exhibit 14).

e t ' a s s'

imperfections could significantly increase the possibility of crack initiation and propagation. Finally, the-only indications which were to b.e Yecorded were cracks " greater than 10% of the l -crack signal in the reference standard PAO-C-1."16/ Unfortu-nately, FaAA does not indicate, nor does the NDE procedure specify, the size of the flaw contained in the reference stan-dard, so there is no way to judge the sensitivity of this screening processing. In our opinion a crack eliminated from further consideration by these criteria could be relevant to issues of crack initiation and propagation. Accordingly, we have no way of knowing how many cracks or other imperfections there may actually be on the 12 AE skirts at Shoreham.

Q. What about the inspections of the skirts in the RV-16-4 and R-5 engines?

i A. On the RV-16-4 piston skirt, a liquid penetrant test showed an indication 3/4 inch long. This indication was i subjected to eddy-current examination and FaAA determined that i

there were "no crack-like indications."E1/ The two AE skirts from the TDI R-5 engine were not of the same design as the skirts at Shoreham.j8/ Three indications were found on one of l

16f Id. at para. 7.1.

Memorandum from D. Johnson (FaAA) to M. Milligan and B. l

}]/ '

Judge (LILCO), Feb. 17, 1984. (Exhibit 15).

j8/ Memorandum from D. Johnson (FaAA) to M. Milligan and N.

Irvine (LILCO), Feb. 3, 1984. (Exhibit 16).

- 45 -

- -~> -r ,---- . --- ,e ,m-w -

-am e , --w r -- -+,s, ,.mo ,,,w-,u >--o---,-r -n- - ,

i.

l-l ,

I .tte skirts, but FaAA decided these were "of no consequence to structural integrity of the skirt."19/ For the reasons given above, we believe the eddy current inspections do not support FaAA's conclusions that the AE skirts can be expected to have unlimited life. FaAA's standards for a " relevant indication" _

permit the presence of imperfections which could increase the likelihood of crack initiation and propagation; thus such defects should have been considered by FaAA in its analyses.

Q. What might happen if cracks in the boss area of the AE piston skirts do propagate?

A. Given the many variables and unknown factors, we can-not give any meaningful estimates of how cracks will propagate, oc how rapidly they will do so. We do know that the,tip of a crack is unstable. It is at higher energy than the surrounding material and will tend to corrode or link with impurities, inhomogeneities or imperfections in the metal to lower its en-ergy. Corrosion will increase crack propagation. At some point a crack, unless arrested by a sufficiently thick area or by physical movement of material allowed by the crack reducing the stress, will reach a critical point beyond which crack S9/ Memorandum from Wells and Johnson (FaAA) to Milligan and Irvine (LILCO), Feb. 9, 1984. (Exhibit 17).

e

growth will be very rapid. Circumferential crack propagation .

l could lead to crown separation from the skirt with disastrous results. Axial crack propagation, depending 'on location, could reduce piston clearance, adversely affect lubrication, and re-sult in piston seizure'or crankcase explosion or both.

Q. Please summarize your conclusions about the probabil-ity of AE piston skirt cracking.

A. FaAA's conclusion that the AE skirts are adequate for unlimited life is inadequately substantiated and invalid.

Cracks are even more likely to initiate in the AE skirts than FaAA's finite element analysis predicts, because the peak fir-ing-pressures in the EDGs are significantly higher than those used by FaAA. FaAA's experiments do not confirm the finite el- ,

ement analysis and should be reanalyzed to explain the signifi-cant 28% discrepancy. FaAA's conclusion that cracks initiate but will not propagate in the AE skirts is based on theoretical idealized assumptions which are unrealistic. Under actual

. operation cracks which initiate are likely to propagate due to such factors as variations in dimensions of the. skirts, the presence of imperfections in the skirt material, the operating environment in the cylinder, and actual firing pressures and temperatures. Finally, the tests and inspections of AE skirts ,

w-- -,e ,- - , - --._, ,_.y. . _

E .

cited by FaAA are insufficient to support conclusions that the skirts are adequate for nuclear service.

Excessive Piston Side Thrust i

=

Q. What is piston side thrust?

5 1

2 A. Piston side thrust occurs at all positions of the piston during operation except top dead center and bottom dead p center. In all of those other positions, the connecting rod is at an angle to the vertical line of the piston stroke. The side thrust on the piston is the resul,t of the force acting to

{

g the line of piston stroke.

s Q. Have you calculated the piston side thrust of the AE E piston in the EDGs? -

E

A. Yes. The calculations for piston side thrust of the AE piston are shown attached as Exhibit 18.' These calculations show'that at the first two midordinate positions the mean

unital thrust on the AE piston at Shoreham is over 123 psi and w

g 111 psi respectively.

=

Q. Is that unital side thrust excessive?

i A. Yes it is. An upper unital limit of 85 psi has been prescribed in a standard design text.60/ Another source states e E

[ 60/ Diesel Engine Design, T.D. Walshaw, Newnes, London, 1949, y at 140.

r

~

B;

that side thrust should not exceed 30 to 40 psi for slow speed diesel engines and 70 psi for high speed engines.ll/ Medium lim-

-epeed engines like the EDGs should fall within these two its.

In most engines with which wd are familiar built by other canuf acturers, the unital side thrust does not exceed 85 psi cnd we have reviewed the design of an engine comparable to the EDGs which has a unital. side thrust of 35 psi. Thus, the calcu-lated mean unital side thrust of the AE piston of 123 psi ex-ceeds the upper value by 44 percent. We believe that the actu-al maximum unital side loading of the AE piston will be more than the calculated figure, because the piston pin in the AE piston is located above the vertical center of the effective piston skirt height. The additional increase will depend upon th*e stiffness of the skirt. s Q. . What affect does this excessive side thrust load have on the EDGs?

A.

The excessive side thrust increases the temperature differences around the circumference of the piston skirt, by causing the side of the piston bearing the higher side thrust This to run hotter than if side thrust were normal.

61/ Internal Combustion Engines, V.L. Maleev, McGraw-Hill, 1945, at 501-02.

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

)

1 temperature non-uniformity will be exacer a eb t d by minor inbalances, minor gas leakage past the piston rings, or lesser lubrication availability after fitting new oil control rings.

As'the temperature differences in the circumference of the skirt increase, piston distortion begins. Distortion further reduces the-arc of contact between the piston skirt and the cylinder liner. As this contact is decreased,'the effective area of the skirt sustaining the side load is drastically re-duced, causing the unital thrust to increase. The increase of thrust increases the friction between the side of the skirt and the liner, further increasing the temperature differences.

Once the temperature differences increase above a certain crit-ical point, partial and complete piston seizure occurs very rapidly -- in just minutes or seconds -- and usually without warning. Piston seizure, if complete, will almost always cause catastrophic EDG failure.

Q. Why can piston seizure occur so quickly?

A. The breakdown can occur very rapidly because of the combined effect of distortion of.the piston in both the verti-

cal and horizontal plane caused by the differences in tempera-l ture in the circumference of the piston skirt. The vertical i distortion causes the piston to bend to the shape of a banana, I -

50 -

f s

w ,g,. ., - - , e,-

m-- . -- - - - -,

e- wm,-o---~~w- - , - - -

w,. -,,.--s-- - - - _ , - - - - - - - m-- - _ - - -- - , -__- - - - - ----

with the hot side-rubbing on the liner at-the outer part of the )

. curve-in the banana shape. As clearance between the skirt and l the liner further decreases, the top and bottom parts'of the l

l inner side of the curve on the cool side of the skirt rub the ,

liner, the. effective clearance approaches zero, a'nd the piston seizes.

Q. Are your calculations for piston side thrust in the EDGs at full load or overload?

A. Our calculations were based upon 4890 HP of the EDGs, the full load. At the rated overload of approximately 110%,

P

> the horsepower is 5379 and the maximum and mean gas pressure increases considerably. Under such conditions, the danger of piston seizure is even greater.

Q. Is the piston side thrust load affected by the fast start requirements of the EDGs?

A. Yes. During 6%e required acceleration of the EDGs to rated speed in 10 seconds the piston inertia forces go from zero to running " normal" while the firing pressures are high almost immediately. Since the inertial forces are subtractive from the side thrust imposed by the piston pressure, the later-al load on the piston is substantially increased during the

- - -m. . - - - ,.---.w-- - -- , ,e--- , .-m.,- - ,_- . . -

~

This load condition occurs

~

fast start portion of the cycle.

while the engine is still " cold" and before lubrication is l

fully' established.

L Q. Are you aware of any evidence of excessive AE piston side; thrust.in the EDGs or elsewhere?

A. According to the DRQR' Report for Shoreham, the TDI I Owners' Group inspections were supposed to verify " lack of l

' scuffing.at the piston skirt" in all'three EDGs.!2/ Scuffing l

was reported in the DRQR Report on a number of AE piston skirts,j3/ but we have not yet had an adequate opportunity to examine LILCO's deficiency and disposition reports cited in the DRQR Repor t to see how these conditions were evaluated.

These reports were only received a few days ago, so our review of ,

them has necessarily been preliminary and cursory. If our more complete review discloses significant information, we will file supplementary testimony. The DRQR Report concludes that "in-spections performed on AE skirts have not revealed excessive side load wear."fi/ Based upon our preliminary review of the 62/ DRQR Report, Vol. 5, Pistons, at B2. (Exhibit 11).

63/. Id. at B4-5, referencing TER Q-326; LDR 2275; TERs Q-41, Q-82, Q-83: LDR 2147; TER Q-159; LDR 2198.

. 64/. d I_d,. at 3.

52 -

y ._, _, ,, ,_.-- , . - _ ~ , , _ . . , _ _ e,- ,- _ , ,,y,y. , , , - , , _ - , . _ . , , , , , , , _ . . , . - . . - - -- - - -

inspection data and personal inspections of some AE skirts at Shoreham, we disagree.

Q. What inspections did you make?

A. During June of 1984, we inspected one AE skirt at Shoreham which showed a heavy wear pattern. The worn area of the skirt was completely devoid of any tin plating or sandwich layer plating. The appearance of the damaged area showed the light mottled patterning and surface roughness consistent with micro seizure. We believe this abrasion of the skirt most likely resulted from heavy side loading resulting in localized' distortion. The profile of the skirt indicated local distor-tion. During this same inspection, we examined seven other AE piston skirts. While these skirts did not show the same heavy wear pattern described above, they did show signs of distress in the tin-plated area (abraded surfaces and evidence of debris that had previously been embedded in the plating, but since removed).

Q. Are you certain the AE skirt you have described was damaged by excessive piston side thrust?

A. We cannot be absolutely certain, but that is the

, probable cause. Evidence of excessive side thrust is usually e

also evident on the cylinder liner against which the skirt has showed ev-rubbed. All of the liners we inspected at Shoreham idence of heavy deglazing, which obliterates any markings asso-We might ciated with high side thrust loading from the skirt.

Esurmise that side thrust markings made the heavy deglazing nec-essary. Deglazing is a maintenance operation in which the cyl-inder liner surface is honed in a criss-cross pattern leaving relatively deep " scratches" for the purpose of maintaining bet-and liner.

ter lubrication of the piston rings, skirt, Q.

Did the FaAA Piston Report address the issue of pis-ton side thrust loading?

A. FaAA has never addressed this issue, notwithstanding fac-that it is both a " functional attribute" and " evaluation" tor in the TDI Owners' Group Program Plan Component Design Re-view for Pistons, Part No.03-341 (DR-03-341-1). Under "Evalu-4 ations," item 9 states:

" Evaluate the effect of piston side loading on wear." We were surprised that FaAA' chose to ignore this matter, not only because of its importance to reliable EDG operation and the physical evidence of excessive side load de-scribed above, but also because of the impact of this issue on FaAA's crack initiation and propagation analyses.

e

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

- 2 -

Q.

What is the effect of excessive AE piston side thrust l

Cn FaAA's analyses?

I L

1~

A.-

' As explained above, excessive piston side thrust causes localized and later more widespread uneven overheating The resulting higher thermal stress will gener-of the skirt.

olly contribute to crack initiation and propagation, especially Jhere the higher surface temperature of the skirt is on the '

The hot other side of the section where the crack is located.

oide increases the tensile loads on the cold side, contributing FaAA supplemented the FaAA ,

to propagation of any crack there.

Piston Report with a second report documenting an investigation This report concluded of the thermal effect on the AE skirt.

that the influence of thermal distortion does not change the conclusions of the FaAA Piston Report as to the AE piston J

skirts.15/ The FaAA Piston Thermal Distortion Report, however, does not address the issue of piston side thrust at all and deals orincipally with effects of thermal distortion of the i piston crown.

Proper consideration by FaAA of the effects of excessive piston side thrust in the AE piston would likely change the analytical conclusions and probably would have shown 7

f crack initiation and propagation in the AE skirt to be more likely.

a 65/ FaAA Piston Thermal Distortion Report at 5-1.

\

1 1

V l

was evidence of excessive side thrust in-AE skirts (

-Q.- ,

'found in the TDI R-5 engine or the DSRV-16-4 engine referred to in the - FaAA Piston Report?

A. We don' t know. If the DSRV-16-4 ran at a peak pres-sure of only.1200 psi, excessive side load would be highly un-usual.

-Q. What do you conclude with regard to the piston skirt side thrust condition on the EDGs?

A. We conclude that the piston side thrust'is excessive and that the AE piston is inadequately designed to accommodate this' load. The FaAA reports have totally failed to address this concern.

There is, therefore, no assurance that the EDGs will not experience serious f ailures induced by this condition.

Accordingly, the EDGs have not been shown to be adequately designed to satisf actorily perform the service intended.

Tin Plating of AE Piston Skirt Q.

Did FaAA consider the potential effect of the tin plating of the AE skirt in the context of its design?

A. No. FaAA did not address this issue despite the fact that a functional attribute for the Task Description for pistons was

e i

5.- The piston skirt must provide a suit-able sliding surface against the cyl-inder liner.

Q.

What are your concerns about the tin plated design of the AE piston skirt?

A. During trips to Shoreham in 1983 and 1984, we observed relatively heavy vertical scoring in a sufficient num-The ber of cylinders to rule out a " case of one" phenomenon.

scores were vertical grooves located in line with the location Examination of pistons where maximum side thrust takes place.

during a visit in 1983 showed accumulations of detritus The scoring

. embedded in the tin plated surface of the skirt.

We believe was visible despite heavy deglazing of the liner.

this scoring results from detritus which tends to collect in The scoring in the the soft tin plated surface.of the skirt.

liner caused by detritus embedded in the tin plating of the skirt can result in gas blow-by. If the cylinder liner is scored, small grooves or deep scratches are made in the liner surface. The piston rings " bridge" the groove or deep scratch and high pressure gases blow down the groove on the outside of the piston ring.

This action in turn leads to piston ring distortion which will allow more gas " blow-by" . When this occurs, the piston e

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

This situation is potentially j skirts tend to overheat.

dangerous in the EDGs, where the piston design causes a high The high side thrust causes the AE Nidethrustontheskirt.

piston to run hotter leaving little reserve for a further tem-Small amounts of gas blow by perature rise from gas blow by.

may -therefore lead to an early piston seizure.

Q. Why_are the AE piston skirts tinned?

A. The piston skirts may be tinned to offset the bad This is yet another effects of very high,unital side thrust.

indication of over-rating of the EDGs. .

Q.

Aside from the liner scoring potential described above, does the tin plating present any other detrimental effects to reliable operation?

A. Yes. Tin and copper / tin plating of the AE skirts If the tin (or could initiate two types of failure mechanisms.

copper / tin) is electroplated on the piston skirt, catastrophic failure could occur through the mechanism of hydrogen embrittlement. The plating process liberates hydrogen at the This classical cathode which enters the metal structure.

embrittlement mechanism has been responsible for many dramatic failures of ferrous metals. It is difficult to detect and a t

hazard in all plated metal components. It-is difficult,

'therefore, to predict-if or when such a failure may occur. If 1

the tinning is applied by a " dipping" process, the resulting

. structure at the plating interface can contain an intermetallic .

compound that forms when the tin matter comes into contact with the iron. This compound is covalent so it acts as a ceramic.

This material, if present in significant quantities, can behave in an abrasive manner and thus contribute.to scoring of the cylinder liner and piston skirt. Such liner scoring could lead to the failures resulting from gas blow-by and piston seizure described in the side thrust. discussion above.

Q. What do you then conclude regarding the " tinned" AE piston skirts?

A. .We conclude that the EDG rating is.well in excess of the design limitation of the AE piston. Accordingly, there is no reasonable assurance that they will perform satisfactorily in service.

REPLACEMENT CYLINDER HEADS i

l Q. What is the purpose of this part of your testimony?

A. This part of our testimony addresses the County's i

concerns regarding cylinder heads; the relevant portion of the EDG Contention states

i l .

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

_ _ . . . ~

The replacement cylinder heads on the Shoreham EDGs are of inadequate design and manufacturing quality to withstand satis-factorily thermal and mechanical loads dur-ing ,EDG operation, in that:

the techniques under which the

( a-)

replact:aent cylinder heads were produced have not solved the problems which causedthe ,

on the Shoreham EDGs; (b) the "barring over" surveillance precedure to which LILCO has committed will not identify all cracks then existing in the replacement cylinder heads (due to symptomatic water leakage);

the nature of the cracking prob-(c) lem and stresses exacerbating the cracks are such that there can be no assurancethat no ne cold shutdown of the EDGs; (d) there can be no assurance that cracks in the replacement cylinder heads and concomitant water leakage occuring dur- -

' ing cold shutdown of the EDGs (which woul,d i not be detected by the barring-over proce-dure) would not sufficiently impair rapid start-up and operation of the EDGs suchthat the function; I

(e) there can be no assurance that cracks in the replacement cylinder heads occurring during operation of the EDGswould not their required function; (f) variations in the dimensions of the re-of the firedeck (and water deck]

placement cylinder heads create inadequate cooling, where too thick, and inadequateresista thin, and create stress risers at their i boundaries;

( (

4

, .. , _ _ . _ , , - . , _ . . . _ . , , _ , . . - _ , .,_,_.,,__,_,_...-y. . . . .pr-,*p.s p" . ..

5 (g) the~ design of the replacement cylinder head is such that stresses are

-induced due to non-uniform bolt spacing

[and the different lengths of the bolts];

the replacement.. cylinder head

[(h) design does not provide.for1 adequate cool-ing of'the exhaust valves]; .

(i) at.least'one replacement cylinder head at Shoreham has an indication;

[(j) the design of the replacement cylinder heads provides inadequatehead]; cooling .

. water for the exhaust side of the and (k) the replacement cylinder heads at Shoreham were inadequately inspected after operation, because:

(1)- a liquid penetrant test was done on the exhaust and intake valve l seats exhaust andvalves firedeck on trea only between the 9 of 24 cylin-der heads, and such tests were done d'

' after only 100 hours0.00116 days 0.0278 hours 1.653439e-4 weeks 3.805e-5 months of full power operation; (2) ultrasonic testing was done e on the firedeck areas of only 12 cyl-inder heads;'

(3) visual inspections were performed on the valve seat areas of only 32 of the 98 valves, and on only 7 firedecks of the 24 cylinder heads for indications of surface damage.

deleted The bracketed portions of the foregoing contentions are and not addressed in this testimony. .

~

l

h adequacy of What are your conclusions regarding t e Q. placement cylinder heads?

tha design and manuf acture of the re FaAA in its Contrary to the conclusions reached by A. 6(/ and by the DRQR Report ecport evaluating TDI cylinder heads en cylinder heads, we conclude that:

inadequate for The replacement cylinder heads are ini-( a) to the potential for cracks to their intended service due d leading to leaks into the tiate and to propagate in the hea s, cylinders.

heads of the The potential for flaws in replacement (b) i g techniques for cast-EDGs still exists, since the manufactur nt heads have not ing, inspecting, and. testing the ireplacemencieg which resulted in been demonstrated to resolve d the defic en the cracking of the original hea s.

k water into in the replacement heads could lea (c) .Crackt The "barring ld shutdown. '

the cylinders of the EDGs during co ust 5, 1983, proposed by over" surveillance procedure, dated Augf water in the cylinders.

LILCO will not preclude the presence o i

Delaval,

" Evaluation of Cylinder Heads of Transamer ca84-5-12, May, 19 66/

~~ Inc. Series R-4 Diesel Engines," FaAA(Exhibit 19).

( the "FaAA Head Repor t") .

sam, , ,., ,

e

%~-v=w<-

A. ~

- h:

id startup

. r.

Cator in the cylinders could impair or prevent rap I cnd : operation of the EDGs. ..

ible.

The-casting process at TDI is not reproduc (d) ill exhibit ,

Thus, there is no assurance that each casting w i istics.

.idcntical or even'aimilar. character l f

The inspections of the replacement heads as ter insuf-(e) i cperation were inadequate in that the operating t me wa Further, the sampling inspections ficient (only 100 hours0.00116 days 0.0278 hours 1.653439e-4 weeks 3.805e-5 months ). t demonstrated utilized were not appropriate since it was no that the population of heads was homogeneous.

d to dem-The stress analysis performed by FaAA faile (f) i of the replacement enstrate that $he pred,icted deformat on gress to .

heads due to thermal and mechanical loads will not pro heads.

the point of impacting acceptability of the oncur with

' In addition, as a result of our evaluation, we c t for Shoreham the Owners Group conclusion in the DRQR Repor that:

The Shoreham absence cylinder of detectable heads does not flaws in the preclude the eventual propagation of a crack from a suusurface defect or a defect in an inac-

' cessible location.fl/

(Exhibit 20).

i DRQR Repor t, Vol. 8, Cylinder Heads, at 3.

61/

i I

_ . . _ . , ____,_____._.~._.._____-...___._..___..._-...,_v.,,.. u m +r, . .

5

.3 Indeed, Dr. Wells of FaAA acknowledged at the June 22 the owners' Group and'PNL that:

coeting between-4 not knowing the distribution of flaws below

the' surface of these heads, that we would in fact acknowledge the ppssibility that cracks would grow and leaks would develop, and confidence in the -- or lack thereof,

- in the behavior of these heads really has to be established by inspection and . . by11/-.

i examining the causes of leaks..

that Based on the preceding conclusions, we do not believe l r ser-the replacement cylinder heads are adequate for nuc ea ill vice, and thus, there can be no assurance that the EDGs w parform satisfactorily in service.

t l I

o What prompted Suffolk County's concern with the cy -

Q.

I inder heads in the EDGa?

A.

Three of the original cylinder heads in the EDGs l

developed cracks in their firedecks which allowed cooling water t

1 Subsequently, the County filed a to leak into the cylinders. h l

contention in these proceedings, which was admitted ced. by t e Board, and discovery concerning the cylinder heads commen linder LILCO then committed to replace all of the original cy d f heads in the EDGs prior to fuel load of Shoreham with hea s o

/

allegedly superior manufacturing quality.fi at 124.

68/ Meeting Transcript (June 22, 1984) 22, 1983, para 3.

Affidavit of Edward J. Youngling, July 19/ .

(Exhibit 21).

l I

- rs.,-----r w -, --,---,w--- --,-paww w ,- ,,----s,- w.- ,-,,,m-,v- ,,rr,+-,,----.e- o- e v~ m s - v

t i der heads in the Did.LILCO replace all of the cyl n Q.

EDG37 FaAA states in one part It is unclear that they did.

A. h t "all but two (E71 of its cylinder head report of May 1984 t aeplaced with heads cnd F64)" of the original heads h had been reinafter called the

' cact by TDI after September, 1980 ( er that all elsewhere in the report FaAA says

'

Group III heads"); replaced with Group of the original heads 'in the EDGs have been heads have been The DRQR Report asserts that all 5

III heads.2S/ 1/

rcplaced with Group III heads.2 i inal cylinder Are the failures of three of the or g Q. III heads relevant to heads at Shoreham and of other pre-Group your conclusions?

TDI heads FaAA acknowledges that pre-Group III A. Yes.

but asserts that these were subject to numerous defects, facturing processes defects were caused only by inadequate 12/

manuBased upon infor l

and/or poor quality control at TDI. f

..turing techniques,22/ FaAA l

r given by TDI as to changes in manu ac FaAA Head Report, at 1-3 and ii.

70/ d at 3 (Exhibit 20).

DRQR Report, Vol. 8, Cylinder Hea s, 11/

FaAA Head Report at 11 1-2 to 1-4.

12/

at 1-5 to 1-6.

1

]3/ I_d,.

~~

p7~ --. ~

ds (including the rGpitcGm:nt ded ser-

}.

h30 c:ncluded that Group III heaare " adequate for their inte tly examined this hocoo st Shoreham) d vico.*li/ However, FaAA has not indepen en"has not been verified dato and in fact stresses its data lity assurance proce-nprColly required under FaAA's quais a potential for duroc."25/ FaAA also concluded that "therei ting flaws in the h h

ceccks to propagate from pre-ex st that "the potential for _ t e to leaks into cylinders," bu heads is significantly less pec-existing flaws in Group 25/

III thnn for" heads cast earlier. d above?

Do you agree with FaAA's conclusions state The Q.

not adequate.

A. No. The replacement heads arege part upon TDI's revi FCAA conclusions are based in larhanges in TDI's manufacturing ll prop-pre-Group III heads and ad hoc cwith FaAA that cracks may we While we agree heads, causing water leaks processes. h cgate from pre-existing flaws in t tehe likelihood of such into the cylinder , we do not agree thatds has been demonstrably flaws existing in the Group III hea 74/ Id. at 4-1.

25/ Id. at 1-5.

26/ Id. at 4-1.

we

,99

,ep

"**%F N*r S

~ ~

i e '.

Our

[

' manufacturing processes. :

y occur in educcd by changes in TDI s

,0ctic:ny will also address the fact that cracks h re-maother than th tha ecplacement heads for reasonsficiencies in the design of t e cacting flaws, including de picccrent heads.

heads in the EDGs of the Q.

Are the Group III cylinderham heads and others cast ceno design as the original Shore prior to September, 19807 heads are All of these TDI cylinder A.

Ba'sically yes. change to weld thicker for a ding to of the same design, exceptcore holes in the head, accor t to covoring plates over the d This change is not significan the TDI drawing of the hea . d designs.

any of our conclusions as to the hea testimony organized? -

. How is this portion of the Q. ficiencies exist First, we will show that various t de Shoreham which may A.

of the replacement heads a ceptable varia-in the design These deficiencies are unac bolt .

lead to failures. he firedeck andwill non-uniform demonstrate tions in dimensions of ts. Second, we i ues have spacing which induce stresseTDI in manuf acturing techn q blems which result in that changes introduced bysignificantly diminished pro not solved or 67 -

-m ~

, 4 ,~ , ,,

i Third, wa will' I l t heads. l,J lcws.cr-cracks in the replacemen EDGs have not been docus:nt that the replacement heads in'thet one flawed

~

l adequately inspected-and include at leasd effects

. FOur th , we will= examine theh nature anreplacement-heads, and 1 initiotion'and-propagation in t e adopted by LILCO will dure cuplain why the "barringk over" the proce use of the replacement not identify all leaks f and ma efor nuclear service.

hardo sufficiently sa e ,

Inadequate Design

. ues with cylinder heads Q. What are the major design issdiesel engine in large medium speed ost intricate and The cylinder head is one of the m It must be A. in the engine. ,

ss to which it is difficult-to-design otrong enough to componentswithstand th cubjected during engine operation,s water passages in the he cient cooling through numerou Thus, the two major handled.

f ficient strength and permit thermal stresses to beconcerns are to prov general design ,

adequats cooling.

i over that TDI did not We were therefore surprised to d sc increased the of the cylinder head when it .

change the design

. . . . i m e,am

MM NWWD#H4+P-

t,[

4 ., ,

i ificant p. . &

61 b rcepower of the R-4' series diesel, because a s gn d mechanical +

L increase in horsepower'also increases thermal an ..

N I leads. .

to in-fu What changes were made in .the TDI R-4 engines l' Q.

crcase horsepower? >

d In 1966-1967 the R-4 series diesel was develope . .

4 A. i eng ine , the R-4 Cenpared to its predecessor, the TDI R-3 ser es increased fuel l increased engine speed from 375 to 400 RPM, and ure l i

cnd air supply to raise its brake mean effective press At the same time, changes were from 165 to 185 psi.21/ linder (BMEP) ende in the design'of the' pistons, connecting heads.28/

rods, cy block, bed plate, cylinder liners and cylinder ine was

the horsepower of the R-4 series eng -

In 1970-71 h10 HP per cylinder) by

[

boosted to that of the EDGs (about 19/ The BMEP in- '

f increasing engine speed from 400 to 450 RPM.

Mro deal with the consequent higher creased from 185 to 225. h ged from a one-piece

! thermal loading, the piston design was c an re removed l

l iron cast to a two-piece steel cas' ting and flanges we ,

f (Exhibit 10).

Trussell Deposition at 81-82.

22/

11/ Id. at 74-81.

29/ Id. at 82.

f

._.m._,e.

-- g .

,,,,.-,n,- .

a.a. _ ~

i

< ! < x.

' h nges were ordo.

f-Joe c3nnecting rod ~

Thebearings, replacementbut headsnoatdesign Shoreham c aare , l sTt ha cylinder head.!E/ igned for an engine '

f ;

ihus iif the :same design as the . heads desPM and hor tith c speed of only 400 R linder in the as compared to over 610 BP per cy

nly.445 BP, EDG3. in the design of the What deficiencies have you noted Q. .

firedcek of replacement heads?

thickness of the A.

TDI permits' wide variations in the i kness across the firedeck. The acceptance standard for the th c' nstrument a i

28-29, 1983, fircdeck was 0.500 inch +ll/.005 inch on July

+ .010 inch per applicable drawing.d ck thickness in thirteen re-NRC inspectors measured the fire efound variations from 0.460 to inimum accept-piccoment heads at Shoreham andTDI takes the p 0.881 inch.81/ 81/ Apparently TDI designers cble firedeck thickness is 0.400.

11, 1983 10/ Id. at 85-8'7. /83-25, August NRC Inspection Report at 3. No. 50-322(Exhibit 22).

81/ ("IEE Report 83-25")

82/ Id. with attachments thereto.15, 1983, for I&E Report Transmittal letter dated August 13/ 83-25. (Exhibit 23).

s e

i i .

~ ^^ ~ ^

^ ~ ~~~ ~ ~

~

L

~

i 7

50notapplyanyacceptancestandardsformax thickness. wide variations in consequences of the l-Q. .What are the fircdcek' thickness? it is susceptible to '

A.- Where the firedeck is too thin,l stresses imposed o i

crtcking from the high mechan ca ticularly at higher horsepow-s which ex-1 fircdeck during EDG operation, parFiring ik pressure or cnd, loading. Where the firedeck is too th c th.

heat transfer 1 cocds the. material istrengt cooling; the diminished '

.thore may be insuffic en wide variations in the ,

Where there are increases stress. in the replacement heads, a thickness of the firedeck, as h boundary of a thick and thin .

stress gradient is created at t ere likely to occur.

portion, which makes cracking mo analyses to deter-Q. Has TDI conducted any studies oro thick or too th

' mine when the firedeck wall is to ign engineering A. No. Mr. Lowery, TDI's manager of deshas t i kness]

and research and development, determine what the firedeck ting the reduction tions have been done toThe only documentation suppor should be."li/

("Lowrey y (May 11, 1984) !

- l Deposition ofatMaurice 85. H. Lowre(Exhibit 24).

gi/ Deposition")

-~q, %_

m

, -- -, , , , , . . .-,,er, , ~ * ' . . . _ . -

s _,..,,---,,n;.---.-,,T,,-

j- ,

e i m firedeck thickness to l tha cceeptance criterion for min rt datedmu Sy 0.4001(from

!. 0.500) covering the f r is an inspection l (.400 repoi edeck jbru2'ry.21,1981, khdo cnd. bearing the notation s

ndum, pr? pared "s Functionally of

>in. Tolerance), "Il/.and an internal TDI memoraured ftor the NRC inspectors had "The smallest nominal meas tating 3 hocds at Shoreham in July,-sd as 1/2 inch. This dimension is 81/ yei_ I

'fircdcck thickness is specifieof 0.400 inches" (sic.)

a minimum engineering evaluation c11cwsd to vary to adequate is an thor of these documents h l of ainimum firedeck thickness.

i deck thickness standard, en whic ,

TDI ignores the maximum f re ce allowances, except betwe 1 r

chould be 0.515 inch with h toleran with allowed tolerances i valve ports, where .765_ inc i k ess requirements 7 the intake 10 required.32/

Both of these maximum th c decks n of asured areas of the fire ero exceeded on at-least 20 me t Shoreham.ll/

18rNplacementcylinderheadsa l

~

(Exhibit 25). i No. Q-0783. [

_85/_ TDI Inspection Report 1983, from G. King ( then TDI,(Exhibit 2 ,

(TDI).

'86/ Memorandum dated August 1,tospecial R. Boyer thicknessand E. Wilson

~~

now FaAA) fer to thevalve ports.

I&E Report 83-25 did notintake re AA Bead Report at 1-8.

17 / - specification between the ~

l gj/. IEE Report 83-25 (Exhibit 22); Fa 1 72 -

e

, ,- m n. :-

- -- . , , . - p :C'~~T? [Y_ __. -

your belief that tha -

Q. Please describe the bases for fn3n-unifork head, stud spacing induces str lccnt heads. are pretensioned the d

from the A. When the cylinder head stu sthe bending mome head are bal-

.hacd is stressed due to The bending moments in theThe .

tcncion in the studs. h block and the liner.

cnccd by the bending moments in t ee induced from the tensile is b0nding moments in the block arWhen thefer- head stud spac otross in the studs. nts set up around the circum non-uniform, the bending momeiform. The stud location is cnce of the head are also non-un h head from stud preten-cuch that the bending moment on t e e direction (90 degrees to h crankshaft cioning is greater in the transversthan in the' direction crankshaft polar axis) i moment means thati head '

exis. This non-uniformity of bend ngis greater in the transve deflection from pretensioning d by thermal distortion rection. The head is further deflectension of the i

resulting from the thicker d me of the head may lead to ex-valves . The deflection the intake attendant with valve leakage. ,

haust valve leakage and problemshe design of the replace-Did FaAA adequately review t .

Q.

ment cylinder heads?

73 -

x

h. ..

e ,+w M-

' We y .4

l A. No. FaAA did not address all of the functiona head as set forth in the Task De-

.cttributes of the cylinder 89/ Rather FaAA ceription for the cylinder head design review. h rmal. ,and pres-licited its design review to an evaluation of timplified e cure stresses on the firedeck, using an extremely ptions which s idcalized version of the firedeck and making assum We believe the FaAA invalidate the conclusions of the review. apparently cnalyses is unreliable, and the TDI ownershGroup t no reli-Mr. Coleman of the owners Group agreed ta cgrees. h FaAA Head cnce could be placed in the design analysis of t e 22, 1984 meeting R3 port in his statement to PNL at the June is that we "The idea that we' re trying to give you today that: report either from the didn' t depend on the (cylinder head] s some idea of otandpoint of the analysis, other than of toour give u recommenda-i what's going on there, but our conclus onsd not have enough infor-tions are based onWethe f act that we diwere unable to do the mation in our analysis.

analysis necessary to get that."

AA analytical Why do you believe the results of the Fa Q. are invalid?

evaluation of thermal and pressure stresses Appendix B.

89/ FaAA Head Repert, po maw ,

we w, wsm

-_.______ _------ - __-__ __ ^^'

~~~

"these are approximate A. Dr. Wells of FaAA stated thatthe general levels of ther-calculations only intended toWe show agree for the following rea-fa cal and pressure stress. RE/ ional model o cons. First, FaAA used an idealized one-dimens fore assumed a uniform flat plate for the firedeck, /

and thereSolutions in the thermal an thickness of the plate.21 iform plate thicknesses.21/

wore obtained for 3 different un and has many thickness vari-The actual firedeck is non-uniformThe temperature distribution in ations in a single firedeck. ffected by these thickness vari-the firedeck is significantly aSecond, FaAA assumed a peak firing t

ations, as explained above.The actual peak pressure is abou pressure of 1600 psi.21/ h assumed and the This large difference between t e sults of 1800 psi. would substantially alter the re actual firing pressure stress evaluations by FaAA.

both the ther' mal and pressure lysis idealizes the firedeck Third, FaAA's pressure stress ana lamped at its outer ,,

as if it were a plate uniformly c the head are In reality, the bolts holding down boundary.21/ ,

ting between PNL and the 22, 1984 mee '

90/ Transcript of JuneTDI Owners Group, at 136.

FaAA Head Report at 3-1.

91/

Id. at 3-3. (" Wells 92/ (May 14, 1984)

Deposition of Clifford H. Wells (Exhibit 27).

93/

~~ Deposition") at 130-31.

FaAA Head Report at 3-5.

94/

- - _ _ . __ ~--+e--

tresses in ,

/

FaAA admits that of the"ths local o complexity Cot uniformly spaced. defy analysis because in support of ltho' critical areas ... Fourth, the underlying datareport's conclu-fof'thageometry.ll/ ovided and thusconclusion the that " pro-tho calculations is not prFor instance, the d Oicno are inscrutable. not exceed twice the yiel even if vided the range of stress doesdimensionally stable Fi-tions.

ctross, the fire deck should be" is not supported d ma- by calcula yielding occurs . . . . sume a perfect cylinder hea The strength of nelly, the FaAA evaluations asts or imperfections. will af-terial free of any defec presence of imperfections tho actual casting and the withstand mechanical and ability of the firedeck to feet the thermal stresses.

, hniques Changes in Manufacturing Tec .

original cracks in the three 1980 Q. Do you believe that theheads cast prior to Septemb suggested in the FaAA Head Shoreham heads and in other result of casting flaws, as are the Repor t?ll/ I 95/ Id. at 3-6 96/

Id,. at 1-2 to 1-4.

i w=

"dee-

,usegq_

Neh Nh

.- D Y '_ .. * *-w _

f the three From the two TDI failure analyses o ,

A. Yes. is evidence of shoreham head f ailures,91/ we believe thereBut there is no bas costing defects in those heads. of failures of these and including oliminating other contributory causesFaAA Bead Repor't, other TDI heads re"ferred to in the the design defects described above.

l engine cylinder head The causes of cracks in anyi 4-cyc eof stresses from cylin-are generally related to a combinat on oling strains (set up -

and der pressures, thermal stresses from coling of the castings),

during the solidification and the cooheads onto the engine frame.

stresses arising from bolting the occurred at Shoreham can Failures such as the ones hthat facthave that stresses affect come about from fatigue and from t eFailures can also occur i gs.

the en' durance limit of the' castcess, n even if the '

if there is thinning in the casting proorosity or hot tears in the thinning is insufficient to cause p n then overstress the thin casting, since gas pressure loads ca is a thickening of ,

The areas. Failures can also occur d if theredue to core shift. .

critical areas of the cylinder hea s h thickened material does reduction in working stresses from t e 0150 and 0151, March 28,

~

TDI Failure Analysis Reports No.(Exhibit 28).

97/ 1983, signed by R. A. Pratt.

.e,...,.

- e hpm ar. g y ,

^ -

~* - -

l' rease in thermal otrece, j ,

St gcnorally.

~

compensate'ithfor cracks thedevelop inc tg usually' occurs, starting w ing outward. ..

nd fcilure cooling water space and mov s pithin'the e is porosity or hot tear Will cracks develop if ther Q. ,

slag in-linfthe casting process?

i kage and sand or n result A. Porosity, hot tears, shr nof casting defects which c ,

clucions are all examples For this reason, it is ap- '

in the-in cracking of the cylinderlyheads. the cracks that were found f propriate to discuss not on heads, but also the kinds o original Shoreham cylinder cracking.

three cause which can hm cacting defects of the cracks in the three Shore a cause Hasn' t the Q. j established? i cylinder heads been lyses performed by TDI, Based upon the failure ana original cylinder A.

t the cracks in the threees acting upon latent of cyl-

.LILCO asserts thacaused by operating stress l-i heads were of cylinder

i[

casting defects - hot teard inclusions in thei case s unjustified inder head S/N E94 and sanThis assertion, however,d o heads S/N E27 and E31 analyses areossible inadequate an b because the failure analyses do not rule out p t The completeness. t 78 -

4 4 f pp.genem '"

"W M a

'?"

' '"AW ,,; *%., .

Complete failure anclynoo [ ,

c. ,

$ributory causes of the failures.- h ical analysis, Id~Eavoincluded_metallography,bulkcemhaps localized chemic

nning electron microscopy, and perLMetallography would have ,

31y510 of:the fractured surfaces. - itably heat treated.

Jclocad-whether the heads had been suled informati It would-tollography would also have revea failure site.

icin otructures of the casting at the t rial, which can be jvo.dotected the presence of coring ma eting, and could have in-z Clotarious to the integrity of'the cas Scanning electron Jicoted the presence of residual stresses. site of crack initiation

~

h Jicroccopy would have identified t e of failure.

would have helped reveal the mode firmed the type of

?nd therefore TDI Socalized chemical analysis would have conWithout th lcac' ting defect present. h cause or causes of . __

l could not have accurately ascertained t e erally be For example, while a hot tear can gen the cracks. ting surface, the cracks ex-recognized on a clean and fresh casback to TDI for analysis, were parianced at Shoreham, when sent l ned and Thus, the cracks should have been c ea corroded. i te means such as scanning electron I,

.oxtmined by appropr a f the corrosion - '

' aicroscopy. Among other things, examination otion of how lon i

products would have given an ind ca With respect to However, none of this was done.

had existed.

79 -

~

l 5

\L <-

. -. - = ~T ^ ; ' .-

~-

1.

c -m

44. -

f the' filings would ) -

l tna sand-inclusions, chemical analys s o i

While TDI frf hsve.

differentiated sand from slag inclusions.

i s, they failed to f performed a microscopic assessment of fil nglish its true origin. s.

chemically analyze the material to estab if " sand inclusions" ,

Moreover, metallography would have shown porosity. Ag ain ,

ware concealing other defects, such as gas however, these tests were not performed.

ry? J Why is a complete failure analysis necessa E Q.

) of the A. Because until the actual cause (or causes h adequacy of solu-cracks is determined, judgments regarding cracking t e to occur h

tions to the problems which permitted t e .

cannot be made with any degree of assurance.

fidence Is it possible to determine with any con Q. turing techniques adopted by TDI .

f

,whether . the changes in manu ac linder heads?

have solved the casting problems with cy hich No.

Since 1976 there have been over 74 changes w A. hniques and foundry TDI has reported it made in its casting tec spe-Not all of the changes were in response to l procedures.28/ in Indeed, many of the changes were made cific problems. f 1

However, the multiplicity of response to production costs.21/

03-360-03-OF" (undated) (Exhibit 98/ "4 Valve Steel Head -- 3, 1983), at 29).

See Crjosition of Edward S. Dobrec (August 99/

52-62. (Exhibit 30).

or - -- ---y +- w e S V TY'I -

m

.n~

5 k5s an as-l chcnges and the interrelationship between changes, Coccrent of the total effects of do not reflect a clear Thus, the changes claimed by TDI blo. but rather an ad hoc cvolution in techniques and procedures,In addition, TDI h Chit or miss" approach. h original cylinder foundry practice procedures at hthe time t eriginal and the re-h0cds at Shoreham were produced, so t e ot be accurately compare piecement head casting processes canno i ular The FaAA Head Report refers to part c al-Q. port its conclusion that, Etnufacturing procedures to sup ht crack due to pre-existing though the replacement heads mig the pre-September flaws, they are less likely to do so than Do you agree?

1980 heads. of inad-s FaAA states that TDI's casting problems A. No. d poor gaging procedures equate mold quality, core shifting, anly addressed" by:

in machining the firedeck "were apparent in mold and core design, and

" Improvements" d (1)

" Chang e s" in materials used for mold anodi-(2) core fabrication, especially use dofand a suse um silicate ester sand for the mol of core shells.lSS/

_100/ FaAA Head Report at 1-5 to 1-6.

..-e -,

p _

e, 6**=

' v -

, - - te T= wf- T Y

t ry flow cnd Cupply of '

S^cond, FaAA says that unsatisfac o lidification

'iquid setal to the mold and the inadequate so fottorn in the sold "were apparently d addressed" b

" Modification" of gates and risers, an

.(3)

(4) Use of chills.lEl/

h ges have solved FaAA' does not conclude that these c anAA could not sup TDI'o casting problems, because Fa f the changes in conclusion. FaAA has undertaken no analysis oIt does not d

'tho techniques which it mentions.h t techniques were used by TDI tdmits it has not verified, w aWithout a careful analysis of both prior-to the current ones. i g procedures, one simply tha current and the previous cast n " impr ovements . "

cEnnot conclude that the changes are

h particular mat- ~

Please explain your views as to eac Q. evious answer.

i ter referred to by FaAA in the pr d core design are A.

(1) The " improvements" in mold an t with no This is a purely conclusory statemen ,

not described. t changes FaAA believes supporting analysis; we don' t know whaso we can' t evaluate w ,

ware made in mold and core design, golf pi. at 1-6.

i k.

~~-~~' a

~ .

s ts.

Our ,own investigation diccioacd

~

not they were improvemen ings of molds, so accurate com-could not ih3t TDI never used design draw 3cricons of current and older t mold andmold ccre de- and core des As TDI-experimented with differen . lts could be Ordo. - may have resulted, but poorer resu l lcigno, improvements occurred with equal probability.

cico have t r sand for molding (2) The change to sodium silicate es eporosity in the cas ocnd can increase the chances for gas-d is more resistant to ga

-cinco-sodium silicate estercasting. san cnd the gas can be trapped in the isers were modified is (3) The manner in which gates and ris no information to j, ddressed not described by FaAA, and thereh ther the changes effec ,

i in a mold *

.a basis to deter'mine w eFor example, placing a gate the casting problems. that the flow is too slow, the nay reduce the rate of l metal flow soady in the mold to cool b thereby permitting metal a re too small can contribute A riser which 19. .

casting is complete.

to shrinkage defects. ihood of ho t

f of chills may reduce the likel lihood (4) The use

, Chills also increase the like tears in some instances.

of gas porosity in the casting.

1 1

4

= =s e ee

- - - . . _ . '%w -**hm._ ~*'em < _ . . . .

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

. .~ .. .. , .. . . . . .

L 1 ting methods and tech-Tho claim that these changes in casd TDI's casting problems is quac have successfully addresse blicd by the extremelyrehigh rate of defects in cylinder p

made. TDI docu-3oduccd since all of these changes wei d r heads show that nearly Ontclg2/ for foundry rework on cyl n ehad defects and required re-1982-83 311 of the heads cast in Jorking.

d manufacturing What other changes in cylinder hea Q.

A?

processes were referred to by FaA to post weld heat FaAA refers to a 1978 TDI procedure A. f the Stellite valve seat treat the heads after deposition o Information Memo changing overlay and to a 1980 TDI Service l92/ The changes are not an-valve seat weld repair procedures. . tion, or effects. We '

clyzed as to their adequacy, implementa h ges have solved the have no basis to believe that these c anits seen in TDI he crccking of the Stellite weld depos Grand Gulf.

hining Problems with RV-4

-102/ TDI Documents, " Casting and Mac(Exhibit 31).

Cylinder Heads."

103/ FaAA Head Report at 1-5.

{

,,n.-.," w "'* ' ^ * '

s heads cast after '

FCAA also states that'TDI claims that all f treatment. = Stress tob5r 1978 received a second stress relie Bicving , or normalizing ,- can' reduce -stresses in the fcasting I:=.

I c ~.

kldsco not eliminate or affect geometrically-induced h of which stress, l

)c. porosity,f inclusions, shrinkage or hot tears, eaci :

- 47 Ms ccuse cracking of the he' ads and . increase crack propagat '

during -

FEAA raises the problem of poor gaging procedures d that

. f d

ircdeck machining ,-but does not state how TDI. ad resse i in firedeck thickness in

)cobica.

.Given the large variat ons '

believe TDI has not t (ha roplacement heads at Shoreham, we Moreover, in our opinion it is not a (ddreseedthisproblem.

a firedeck thickness is rather a design, pcgingproblem; '

utnufacturing, and quality control problem.

d Does the operating history of the Group III hea s '

Q. ) support FaAA's (including the replacement heads at Shoreham te for nuclear conclusion that the replacement heads are adequa isting corvice and are significantly less likely to have pre-ex e ,

flawa than earlier heads? I No.

The only operating history of Group III heads ;

A.-

l ent cylinder verified 'Inr TDI is that pertaining to 16 rep acem ys-it has hondo.at Shoreham, out of 311 Group III heads TDI saf pecduced.101/ FaAA states that Messrs. Trussell and Pratt o ..

)

(Exhib-

_304/ FaAA Head Report at 1-2; Wells Deposition at 103. !

it 27).

~

L

. I.~

> ; m.

---y ,

- $=*._g

l .- ~

i TDI said in December, 1983, that "there have been only five in-stances _of water leaks in Group II and Group III cylinder heads that-have resulted in water in the cylinders.... 105/ However, Mr. Mathews, vice president and general manager of TDI,

~

L 1

testified in May 1984 that.TDI had never in the past two years conducted any review of its files to ascertain failure rates of cylinder heads.106/ There is simply. insufficient evidence from TDI's operating history to conclude that Group III cylinder heads will not crack or will have any less likelihood of I

cracking than pre,-group III heads.

1 Inspections of Replacement Heads Q. Have the inspections of the replacement cylinder heads at Shoreham ensured that they are adequate for nuclear service?

A. No. The inspections performed on the Shoreham re-placement heads have been inadequate in a number of respects.

First, there can be no confidence in inspections carried out by TDI before the heads were delivered. Second, the inspections 105/ FaAA Head Report at 1-4.

106/ Deposition of Clinton S. Mathews (May 8, 1984) ("Mathews Deposition") at 79-82. (Exhibit 32).

- 86 -

l r

I.

._,__w--- , , . . - _ . , - . ., , ,.,,w-~ -, . . . . , - ,r- ., . - - _ - - , . - - . _._m__.

L i

l of the replacement heads performed at Shoreham after 100 hours0.00116 days 0.0278 hours 1.653439e-4 weeks 3.805e-5 months of. operation were insufficient because inspections were done on a sampling basis, not all of the inspection techniques neces-sary to detect flaws were used, and the inspection standards ,

were inadequate.

Q. . What inspections is TDI supposed to have performed on the replacement cylinder heads?

A. .TDI has written procedures for carrying out visual inspections, magnetic particle inspections, liquid dye penetrant testing and hydrostatic testing.107/ However, these

. procedures are seriously deficient. The magnetic particle in-spection procedure does not 'specify which areas of the cylinder heads are to be inspected. Ind,ed, e the procedure is only a general procedure "for.the testing of ferromagnetic parts and 1 ,

assemblies," and is not specifically written for cylinder head inspection. In f act, TDI did not inspect the replacement heads by magnetic particle techniques, because it only started to use this procedure on cylinder heads in April 1984.108/ The hydro-static test procedure, which is written for use in " welded i

107/ Respectively, TDI QC Procedures 600-10, 600-30, 600-20 and 600-70.

108/ FaAA Head Report at 11.

I L

assemblies and cast products," is not expressly ~ applicable to cylinder heads. TDI interprets its inspection procedures to

~

l permit acceptance of a cylinder head which fails a visual in-spection (by having a visible indication), so long as it does i

not leak - during the hydrostatic test.109/ A further example of TDI's inadequate. test and inspection procedures is the in-process inspection procedure (I.P.-300), which directs the QA inspector to use the same gauge blocks as the machinist and 1

sets forth no measures for ensuring that the gauges are proper-ly controlled, calibrated and adjusted so as to maintain accu-

, racy. Mr. Mathews of TDI testified that TDI may well deliver cylinder heads to nuclear plants that have cracks or sand in-clusions.ll.E/

Q. Are TDI's inspections and testing techniques, if tIhey are properly performed, capable of detecting all casting defects and cracks in the replacement cylinder heads?

A. No. It is unlikely that any of the techniques used by TDI will detect cracks or other casting defects more than 1/4 inch beneath the surface of the casting. Visual I

109/ IEE Report 83-25 at 4. (Exhibit 22).

110/ Mathews Deposition at 86-87. (Exhibit 32).

4

(

l i

l i nspections and dye penetrant testing, if done correctly by trained personnel, will only reveal surface cracks. Hydrostat-ic testing only. discloses through-wall cracks in or around the cylinder head passageways that are tested, and will not detect subsurface cracks. Magnetic particle inspections can reveal subsurface cracks or other casting defects, but only to an ap-proximate depth of 1/4 inch.

r Q. What inspections were subsequently carried out on the L

replacement heads at Shoreham?

A. A liqui'd penetrant test was done on the exhaust and intake valve seats and firedeck area between the exhaust valves on 9 of the 24 cylinder heads, after 100 hours0.00116 days 0.0278 hours 1.653439e-4 weeks 3.805e-5 months of full power operation. Ultrasonic measurements were taken of the firedeck areas of 12 cylinder heads. Finally, visual inspections were .

performed on the valve seat areas of 32 of the 98 valves, and 4

cus 7 firedecks of the 24 cylinder heads for indications of sur-face damage.lll/

Q. Were these inspections adequate to conclude that the replacement cylinder heads at Shoreham are qualified for l

" unlimited operation"112/ in nuclear service, as FaAA 111/ DRQR Report, Vol. 8, Cylinder Heads, at B3-B4. (Exhibit 20).

112/ FaAA ' Head Report at 111.

- 89 -

-y 3 ,# r . . - .

. -e_ - - - - _. - - - , - . - _ -

9

' concludes?

'A. Absolutely not.- In fact, the DRQR Report for Shoreham states The absence of detectable flaws in the Shoreham cylinder heads does not preclude

.the eventual propagation of a crack from a subsurface defect or a defect in an inac-cessible location.113/

We agree with this statement, but we also believe that the in-spections were not sufficient to detect even all relevant flaws and defects in accessible areas of the replacement heads. Ac-cordingly, the probability of cracking of the replacement heads may be much higher than indicated in the DRQR Report.

Q. What.are your reasons for concluding that these in-spections did not suf ficiently disclose' even surf ace defects in the replacement heads?

A. First, only a limited number of samples of the re-placement heads were inspected. As described above in our dis-cussion of the AE piston skirt inspections, a sampling inspec-tion is particularly inappropriate because of TDI's ineffective QA/QC program. Region IV of the NRC informed TDI that results of NRC Vendor inspections of TDI show 113/ DRQR Report, Vol. 8, Cylinder Heads, at 3. (Exhibit 20).

[s]erious deficiencies have existed in the implementation of your committed quality assurance program for manufacture of emer-gency diesel generators. What concerns us greatly is that certain of these findings are of a nature which brings into question

' both the adequacy of existing manufacturing process controls and the level of compli-ance by manufacturing and quality control personnel.ll4/

We agree with Mr. Foster of NRC Region IV that TDI's ineffec-tive QA/QC program makes a samplying inspection next to useless and means that even a 100% inspection is unlikely to reveal all defects.115/ However, given the importance of the heads, a 100% inspection should have been performed.

Second, of the sample heads, only selected portions were examined. For example, the liquid penetrant test was performed on .he firedeck only in the area between the exhaust valves. -

Other areas of the firedeck are as likely to have indications or inclusions.

Third, inspections were restricted to visual and liquid penetrant. The ultrasonic measurement was done only to measure firedeck thickness. It is likely that more defects would have 114/ Letter dated January 17, 1984, from V. Potapovs (NRC) to C. Mathews (TDI). (Exhibit 33).

115/ Foster Deposition at 54-55. (Exhibit 13).

i been detected if magnetic particle examination, eddy current examination, and radiograph testing had been employed. The vi-sual examination is unfortunately of limited value.

~

.Q. . Why is the visual examination of limited use?

+

A. Apart from the obvious fact that it is limited to what the naked eye can see, the results of the visual inspec-

' tion have apparently been ignored.. The NRC Staff discovered an indication about '3/8 inch long on the machined bottom part area of replacement head S/N H-34 at Shoreham.ll6/ TDI advised the staff that this crack was within TDI's acceptance criterion be-cause the head had not leaked under hydrostatic test. LILCO and FaAA have not replaced the cylinder head with this indica-

-tion, apparently accepting TDI's cr-iterion.

Q. Do you believe that the LILCO response was appropri-ate?

A. No. A 3/8 inch indication such as on head H-34 may grow under operating stresses and with the effects of corro-sion. Yet LILCO, TDI and FaAA would permit one or more small cracks or inclusions in the replacement cylinder heads. This 116/ IEE Report 83-25 at 4. (Exhibit 22).

I l

l t

-~~ , - - - -

' 'is also shown by the acceptance criteria used by LILCO and l- 'FaAA.

Q. What is the basis for the FaAA/TDI Owners' Group in-spection criteria for cylinder head inspections?

A. No bases are provided for the liquid penetrant in-spection or the ultrasonic measurement criteria cited in Appen-dix A of the FaAA Head Report. For the magnetic particle in-spection, no basis is provided to demonstrate that the ASTM cc,iteria are appropriate for the intended service. For the firedeck UT measurement, the thickness is only required to be recorded. No maximum thickness is specified and the technical basis for the minimum thickness is not cited. The bases for

.all the acceptance criteria should have been provided by TDI, and assessed by FaAA. The acceptance cr'iteria bases must be demonstrated because without knowing the distribution of flaws below the surface, any crack or void can be assumed to grow.

? ,

Cracks in Replacement Heads Q. If cracks si'milar to those in the three original 7

heads occurred in the replacement cylinder heads at Shoreham, is it true that only a very small amount of water could leak i into the cylinders after shutdown of the EDGs?

f e

r,-,, n , - ,. --

A. No. This proposition was asserted by LILCO based upon TDI's inadequate and incomplete failure analyses of the !

l original failed heads, which determined that the cracks were caused'only by operating stresses acting upon pre-existing casting defects in the cylinder heads.ll7/ TDI contends that since these operating stresses are caused by the cylinder fir-ing pressure, once the EDG is shut down and operating _ stresses are substantially reduced, any cracks would close.ll8/ In ad-dition, TDI asserts that the stresses are further reduced when the cylinder heads cool to a steady-state temperature. Thus, it was concluded that the cracks were self-relieving and would not have propagated.

While the evidence suggests that a cause of the failed cylinder heads was casting defects, there is no support for ,

4

' TDI's assertion that only the operating stresses were acting upon the casting defects and that the cracks were therefore

' self-relieving -and would not have propagated. In fact, cracks i

117/ Af fidavit of Edward J. Youngling, dated July 8, 1983 (Ex-I hibit 21).

118/ Contrary to the preceding assertion, PNL consultant Mr.

Louzecky stated at the June 22 meeting "

(Tr. at 129) be-in the tween PNL and the Owners' Group that . . .

cooling-off period, that's usually "

when your (cylinder head) crack opens up . . . .

9

- 94 -

e

~ -- r ,- ,, v, - - -w

l s

buchLas those[found in the three failed cylinder. heads at Shoreham will always propagate and grow, unless arrested by heavy material or a void.

Q. .' What f actors other than operating stresses would cause cracks to propagate and grow?

A. Cracks propagate (i.e. , deepen and/or travel) and grow (i.e. , lengthen and/or widen) due to operating stresses, residual stresses (i.e., manufactured-in stresses, such as from the casting and welding processes), geometrical stresses (e.g.,

stresses arising from design, such as stresses which exist at What must be kept in mind sharply-angled edges) and corrosion.

is that cracks are stress raisers, and that stresses other than operating stresses will propagate a crack. Residual and geo-metrical stresses commonly accelerate crack propagation and

growth, and corrosion occurs preferentially at cracks. All of these mechanisms (residual stress, geometrical stress and cor-rosion) will act to pro'pagate a crack even when a diesel is not in operation. Further, the environment may increase the growth of the crack at a higher rate than one would calculate by merely summing the cyclic loads.1197 Indeed, cracks in the 119/ " Introduction to Fracture Mechanics," Kare Hellan, McGraw-Hill, 1984, at 135. .

6

,-....v . . . . . . _ __

, . . . . , _ _ _ -.. ,,,,-..m_.. _ - , _ _ , , ,m . _, ,

i l

. cylinder head by their.very nature propagate and grow until L they hit a massive part or a void, such as an exhaust valve.

When a crack enlarges, the fldw of water through the crack will j increase.- Furthermore, cracks are seldom self-relieving, ex-cept perhaps when they split open, and cracks never decrease in dimension, especially when the crack surfaces are covered with corrosion products. Therefore, water can continue to leak into the combustion chamber after shutdown and at any time thereaf-ter. ,

Q. Could cracks in the replacement cylindes heads first begin to. leak during cold shutdown of the EDGs?

A. Yes. For example, a crack which initially occurred

. from operating stresses may not leak during operation. , That same crack may not leak for some time after the EDG is-shut down. However, stresses other than operating s' tresses, such as stresses from corrosion products acting to force the crack j

apart, may cause the crack to propagate or grow after shut- -

down.120/ Cracks may grow very slowly for some time, but once a crack reaches its critical size it will grow very rapidly and 120/ " Analysis of Oxide Wedging During Environment Assisted Crack Growth," S.J. Hudak and R.A. Page, NACE, Vol. 39, No. 7, July, 1983, at 285 to 290.

\ l

~l

i L

i rupture. Thereafter, the flow of water through the crack could I

I.

be significant. The amount of such water leakage would depend upon the number and the size of the cracks and their location.

The existence and interaction of these factors cannot be predicted. However, depending upon the circumstances, signifi-cant leakage could occur in a matter of days or even hours.

Q. Would undetected leakage from a cracked head into the cylindar affect the rapid restart capability of the . diesel gen-erators?

A. Yes. If liquid is contained in the cylinders, there will likely be damage to the engines. Quantities of. liquid can cause dangerous pressure rise within the cylinders. If liquid is contained in the cylinders, the compression pressure in-creases and will continue to increase until it equals the fir- -

ing pressure; the volume of liquid contained in the cylinders If the liquid in then becomes known as the " critical volume."

the cylinders is greater than the critical volume but less than the clearance volume, the liquid may not show up during the barring procedure proposed by LILCO, and dangerone praasures may build up during the start period. This very high and dan-gerous pressure buildup can cause studs holding the head in place to stretch, thereby " blowing" the head gasket. When this

. l l _

l

O-occurs, the EDG cannot be operated because of flames blowing out.between the head and liner faces.

Water leakage into the cylinder head could also lead to a catastrophic failure should the cylinder head go " solid with water." .The Shoreham piston crowns have'a dished configura-tion, and should there be leakage the dish area could fill up and the water overflow down past the piston rings into the lube ~

oil sump.- This could cause water contamination of the lube oil. Leakage, even in very small amounts, could also impair lubrication of the cylinder. Scoring of the cylinder liner bores can occur, followed by rapid seizure'of the piston and consequential damage.

Catastrophic consequences can also result from cracks in the cylinder head firedeck, even when there is no water leak-age. Higher pressure combustion gases can leak into the cool-ing water space. kntheshortterm,thecombustiongasesenter the cooling water and may " air lock" the heads. Alternatively,

~

the' heat exchangers may not be able to handle the heat input to the cooling water and a rise in temperature could cause a shut-down. A further problem could arise when the cooling water pumps " gas up, " causing the cooling water temperature to rise and the engine to shut down.

4

s Q. Could the corrosion inhibitors in the cooling water 3

of the EDGs affect rapid restart if leakage occurred?

These corrosion inhibitors can alter the cylin-A. Yes.

der liner diameter by building up salts and other corrosion l

products, if cooling water leaks into the combustion chamber and cylinder space. This, in turn, prevents adequate lubrica-tion and causes a number of dry strokes during the starting of l

the engine. The dry strokes would result in localized heating, with probable additional failure of lubricant and seizure of the pistons.

m Q. But won' t the corrosion inhibitors prevent corrosion in the cylinder,'should leakage occur?

No, the corrosion inhibitors act to passivate a sur-A. s f

f ace by providing a stable film to act as an oxygen barrier.

f Corrosion would preferentially occur in the space between the 7

cylinder walls and the piston. Thus, it is possible that cor-h rosion products could form that would act as a barrier and pre-I vent the passage of water betweeen the piston and the walls and into the lube oil sump. In other cases, auck, carbonaceous ma-

?

- terial and detritus from the piston ring grooves can act as a r

sealant and prevent leakage down the side of the piston. Then, r water would collect in the cylinder, causing the cylinder head E

k I to go " solid with water."

i E

E E

E P

f

I Q. How fast could this corrosion occur?

A. The passivation occurs immediately on contact with the metal. However, the speed at which subsequent corrosion

' , , processes occur is dependent upon a variety of factors and their interaction, including-temperature, surface area and driving force. Hence, the speed of corrosion development for this case is inherently unpredictable. What must be kept in mind is that the concern lies not only with corrosion in the cylinder, but also with the effect of corrosion on cracks in the cylinder head. As previously mentioned, corrosion products put a stress on cracks. Thus, a crack may grow slowly until it reaches a critical size. Thereafter, however, it grows much more rapidly. Indeed, cracks can change significantly in a

~

matter of days or even hours.

Q. Will water flow through a crack during cold shutdown even though there is essentially no water pressure to drive the water through the crack?

A. Yes, because the water has substantial driving force through the crack without the water pressure of the cooling system. The cooling water flows into the crack in an effort to dilute the corrosion products and creates an osmotic pressure.

In addition, the driving force from the capillary action causes

- 100 -

' flow through the crack. As the crack grows, the flow of water increases proportionally.

L Q. But isn' t a steel cylinder head strong enough to re-

'sist cracking caused by corrosion?

A. No, it is not. The stresses generated by corrosion products are extremely high. Moreover, the tip of a crack acts ~

as a stress riser and can synergetically exceed the tensile strength of the metal without any additional stresses. In ad-dition, TDI has changed the steel in its cylinder heads to a lower . strength alloy (TDI's No. 7 steel) with less carbon con-tent. This seduction in carbon can cause cracks to initiate, to propagate, and to grow.

Q. Could leakage from cracked replacement heads also have an adverse impact on EDG's performance'during operation?

A. .Yes. Operating stresses could cause the cylinder head to crack or could exacerbate existing cracks' growth.

LILCO and TDI contend that there would be no adverse impact on the EDG's performance, since any water leaking into the cylin-der during operation would be expelled along with combustion by-products. However, depending upon the location and size of

' the leak, water in the cylinder could be sufficient to impair

- 101 -

l 1

. - - - . - . , . _ , _ _ - ,, -g - -- . - . - -

.v.

lubrication in the cylinder and cause coizure of the piston and ~

fracture of the piston skirt, leading ' to engine shutdown. In l-other cases,-only partial seizure may occur; however, this can

-lead to . heavy bearing wear and misalignment.

Q. But isn' t water sometimes injected into the combus-

- tion chambers of diesel engines to improve performance?

A. Yes. Sometimes distilled water in very small amounts is homogenized with fuel and injected into the combustion cham-bers. This is done to reduce the emissions of nitrous oxides with the exhaust gases. However, this process requires strict control of the quantity of water that is homogenized'with the fuel prior to injection. In addition, the cooling water in the EDGs contains corrosion inhibitors. If the cooling water leaks inte the combustion chamber and cylinder, the salt residues ,

from these corrosion inhibitors can cause abrasive wear on the cylinder liner bore, thereby reducing piston ring life.

Q. Could cracks in the cylinder heads also cause prob-lems in the long term?

A. Yes. While it is true that water leakage into the cylinders generally flashes to steam and passes out with the exhaust gases, if any water remains it is sprayed out with the

- 102 -

i

.. . - - ,. ... . . . . .. .. = . , . . - - .. ,. ,

e exhaust gases and erosion of the turbocharger blading will occur in a manner similar to steam turbine blading erosion. In that event, the turbo blower speed falls and the engine overheats due to a reduction in air flow. Moreover, cracks in the cylinder head firedeck may cause a reduction in cooling water pH value, leading to the formation of acids which attack various parts of the engine cooling system and cause corrosion of the engine. Water leakage may also damage or score the cyl-inder liner, damage the piston rings, reduce power and allow gases into the cooling water system. The scored liners allow hot combustion gases to blow down between the cylinder and the piston skirt. This causes distortion of the piston, further scoring of the cylinder liner and serious overheating, which may eventually lead to a crankcase explosion.

Q. Has LILCO committed to perform a "barring over" pro-cedure at certain intervals after EDG shut-down to detect water which might have leaked into EDG cylinders due to cracks in the replacement heads?

A. Yes. LILCO inter.ds to use the procedure referenced in SP27.307.02.121/

121/ DRQR Report, Vol. 8, Cylinder Heads, at 3. See LILCO Pro-

- cedure SP 27.307.02, Emergency Diesel Generator Cylinder Head Leak Detection Test. (Exhibit 34).

- 103 -

4

Do you believe that the propocad barring over procc-

~

. Q. ,

. dure, if implemented, will ensure that leakage, if it occurs L

during testing or operation, will be detected?

)

i

-A. No. LILCO's proposed procedure will not ensure the In fact, detecti'n o of leakage of water into the cylinders.

given the nature of cracks in cylinder heads, no barring over procedure can ensure that leakage will be detected prior'to an emergency' rapid startup of the diesels. Cracks which occur during operation may not leak during operation or even within the first 12 hours1.388889e-4 days 0.00333 hours 1.984127e-5 weeks 4.566e-6 months .after shutdown, the time under LILCO's pro-For posed procedure when the EDGs would last be barred over.

example, cracks formed during operation could be focal points for corrosion, which would make it difficult for the cracks to close. ' Water could therefore leak into the combustion chamber of the EDG after shutdown,' including more than 12 hours1.388889e-4 days 0.00333 hours 1.984127e-5 weeks 4.566e-6 months there-after, in amounts sufficient to impair an emergency start.

Such a leak would not be detected by LILCO's proposed barring over procedure. Even swinging over the engines 'with starting air might not detect small amounts of water symptomatic of a leak.

Q. Would your concerns with LILCO's barring over proce-dure be alleviated if the barring over were performed more frequently than proposed by LILCo?

l

- 104 -

- - - - - , - - - s- - ,n,, -m. , - - - - - , - - - - - - - - . - --

e

A. Even if barring over were done more frequently, there would be no assurance that leaks which could impair emergency It is not possible to startup of the EDGs would be detected. ,

predict when emergency startup would be needed, and it is therefore impossible to bar the engine over immediately before startup is required. Unless the barring over is done immedi-ately prior to emergency startup, there can be no assurance that water from one or more cracks would not leak into the cyl-inder of one or more EDGs and impair startup.

- 105 -

^ - ~ - " - - - - - - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

REPLACEMENT CRANKSHAFTS '

Q. How does Suffolk County's dontention relate to the crankshafts in use at Shoreham?

A. The EDG Contention provides that its first paragraph is supported because:

( a) The replacement crankshafts at Shoreham are not adequately designed for operating at full load (3500 kW) or over-load (3900 kW), as required by FSAR Section 8.3.1.1,5, because they do not meet the stardards of the American Bureau of Ship-ping, Lloyd's Register of Shipping, or the International Association of Classification Societies. In addition, the replacement crankshafts are not adequately designed for operating at overload, and their design is marginal for operating at tull load, under the German criteria used by F.E.V.

(b) The shotpeening of the replacement crankshafts was not properly done as set ~

forth by the Franklin Research Institute report, Evaluation of Diesel Generator at Shoreham Unit 1, April 6, 1984, and the shotpeening may have caused stress nuclea-tion sites. The presence of nucleation sites may not be ascertainable due to the second shotpeening of the crankshafts.

Q. What is the type of crankshaft now in the EDGs?

A. The EDGs now have replacement crankshafts with 13-inch diameter main bearing journals, 12-inch (nominal)

- 106 -

e d

diamator crcnk pins (or connecting rod journale), and 3/4-inch crank pin fillet radii. The original 'erankshaf ts had 11-inch e

(nominal) diameter crank pins'and 1/2-inch crank pin fillet radii.- The replacement crankshafts were installed after the original crankshaf t on EDG 102 fractured into two pieces during an-engine test run following the replacement of cylinder heads.

The fracture occurred mostly through the web connecting the number 7 crank pin adjacent to the number 9 main bearing jour-nal. Subsequently, inspections identified cracks in the number 5 and 7 crank pins of EDG 101 at the generator end and cracks in the number 6 crank-pin of EDG 103 at the governor end. FaAA has published a report entitled " Emergency Diesel Generator Crankshaft Failure Investigation, Shoreham Nuclear Power Sta-tion" dated October 31, 1983, that concluded that the original crankshafts were inadequately designed and had failed due to high cycle torsional fatigue.

Q. Have you examined any materials concerning the origi-nal and replacement crankshafts?

A. Yes. We have examined the drawings for the original and replacement crankshafts and associated parts, and have reviewed numerous documents concerning the crankshafts, including the various reports by FaAA. LILCO also allowed us

- 107 -

4 6

e

to toko~c'briof vicual_inspaction of the replaccmsnt crankshaft as-installed in EDG 103.

Q. Have you reached any conclusions concerning the ade-t quacy of-the replacement crankshafts in' the EDGs?' :

'A.- Yes. We have concluded that the replacement crank-shafts in the EDGs are inadequately designed for operating at the FSAR-specified full load (3500 kW) or overload (3900 kW).

The replacement crankshafts do not meet the published standards

~

of the ~ American Bureau of Shipping, Lloyd's Register of Ship-ping, the-International Association of Classification Societies and'other major international classification societies for operation at the full load or overload operating conditions of L

the EDGs. The replacement crankshafts also are not adequately designed for operating at ov'erload, and their design is ma,rgin,-

al for operating at full load, under the standards of the ,

German design criteria used by the TDI Owners Grou' p 's diesel engine consultant, FEV. -

In a ddition, we have concluded that the shotpeening of two of the replacement crankshafts was improperly performed ini-

. tially and may have caused nucleation sites which may not be ascertainable due to the second shotpeening of those crank-shafts.

108 -

_ _ _ _ _ _ . _ _ _____ _ m -.__-.-.,_...-__,-.--_.,,._...,.m,. , . , em, .,w__.m..._..,_ , _ . _ , _ . ,.y. ..y.,9-f-.y_. - - -. .

Ste:ndnrdn for Crnnkchaf t Deaign Q. Are there any standards governing the design of L crankshafts in_ diesel engines?

A. There is no single set of engineering standards governing the design of crankshafts in diesel engines. Howev-er, the various' ship classification societies have adopted standards for evaluating the adequacy of the design of crank-shafts in diesel engines in marine applications. We believe that these standards provide minimum guidance for applications where reliability is a significant evaluation factor. The ship classification societies include Lloyd's Register of Shipping

("Lloyd's"), the American Bureau of Shipping (" ABS"), Nippon Kaiji Kyokai ("NKK") , Det Norske Veritas, and Germanischer Lloyd.

Q. What are ship classification societies?

A. To assure the safety of their vessels, shipowners re-quire shipyards to build and equip their vessels in compliance with the rules of classification societies. Those rules in-clude limitations on propulsion equipment such as diesel en-

! gines. Engine builders use these rules as design criteria when designing new engines and major engine components, when l

- 109 -

t i

increasing the rating of an engine, and whon changing the.

design of major engine components. Prudent engine builders en-sure that their engines comply with these rules.

. As reported by the NRC's Consultant, Franklin Research

. Center ~("FRC"):

" Ship classification associations such as the American Bureau of Shipping and Lloyd's Register of Shipping, represent possibly the oldest machinery review and evaluation associations functioning today. Lloyd's Register began operations in 1760 and published its first set of rules in 1834.

As ships and ship propulsion systems became more sophisticated, the classification as-sociations served as design review agents to evaluate functional adequacy and safety.

Considerable experience in the review and evaluation of diesel engines was realized from the long-term use of diesel engines for propulsion and electric power genera-tion in ships. The ship classification *

- rules probably represent the most extensivt experience in large diesel engines avail- (

able.al22/

Q. Why do you believe that the standards set by ship classification societies should be applied to determine the ad-equacy and reliability of the replacement crankshafts at Shoreham?

122/ Evaluation of Diesel Generator Failure at Shoreham Unit 1, Final Report, Failure Cause Evaluation, April 6, 1984, by Franklin Research Center ("FRC Report") at 33-34. (Exhib-it 35).

- 110 -

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

A.. Beccuco thoco.ctendardo cabody tha only ccmprehensive

! - collections of meaningful guidelines controlling crankshaft de-sign in diese1' engines to be used in applications where reliability is a -controlling factor. There are no other ade-

- quate standards.

! Q. The purchase specifications for the EDGs required

' that the crankshafts conform to the guidelines of the Diesel l

. Engine Maufacturers Association ("DEMA"). Aren't those guidelines a reasonable alternative set of design standards by which adequacy of the design of the replacement crankshafts can be measured? .

A. No. Those guidelines are not a design code. As the foreward to the DEMA guidelines explicitly states, "(I]t is not the purpos'e 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 en-gines on the market, notwithstanding the fact that many techni-cal texts are available to the student who may be undertaking the design criteria aspects of engines in general."123/

l 123/ Standard Practices for Low and Medium Speed Diesel and Gas Engines, 6th ed., 1972 at 111. i l

l I

111 -

l 5

4

. .. . g u

=

Q. Generally speaking, what factors do the classifica-tion societies take i.nto consideration in evaluating the ade-quacy of crankshafts on diesel engines?

A. The various classification societies evaluate the ad-equacy of the design of diesel engines in different ways and in

- varying degrees of detail. For instance, Lloyd's rules evalu-ate the adequacy of the design by calculating the maximum power rating for engines. This calculation takes into consideration 26 inputs, including the manufacturing or forging process of the crankshaft, the strength of the crankshaf t material i .d the

?

existence of fillet radii. Lloyd's rules also calculate the In addition, maximum allowable torsional vibration stresses.

unlike most other rules, Lloyd's rules require that auxiliary oil engines that are coupled to electrical generators must be capable under service conditions of developing the power to

} '

drive the generators for 15 minutes at an overload power of not

[

less than 10 percent. Lloyd's rules also consider misfiring in the cylinders.

f The ABS rules evaluate the adequacy of crankshaf t design R by calculating the minimum allowable dimensions of the crank-shaft pins and journals, and crankshaft webs. These calcula-tions take fewer inputs into consideration than Lloyd's rules.

=

m

- 112 -

y L .

[

For example, the ABS rules do consider the strength of the crankshaf t material, but do not consider the forging process

- nor do they directly consider the existence of fillet radii.

The ABS rules also calculate the maximum allowable torsional vibration stresses. The ABS rules, however, make no provision r

L for operating an engine at an overload condition.

The draft rules of the International Association of Clas-sification Societies ("IACS"), which are used by some of the classification societies, are somewhat unique in that they con-sider the adequacy of the crankshafts on the assumption that the most highly stressed areas are the fillet transitions be-tween the crankpin and crankshaft web as well as between the mz E journal and the web. Rather than calculating the adequacy of crankshaf t dimensions or torsional vibrations, the IACs rules calculate a factor of safety based upon torsional and bending stresses and stress concentration factors.

E Q. Do you believe that the rules of any particular clas-

=

sification society should be adopted to evaluate the adequacy of the replacement crankshafts?

E A. No. We do not believe that any particular classifi-2 cation society has the " ideal" standard. However, it is pe ti-nent that Lloyd's generally is considered to be the most

- 113 -

m M

i .

I E LE = ; WH'

conservative of the major classification sociatioc, hence .

. providing - the greatest margin of safety. In view of the poten-f tially catastrophic consequences resulting from the failure of l

the EDGs at Shoreham, we believe that, at a minimum, the crank-shafts should be compatible with the rules of all of the major classification societies.

Q. Professor Christensen, have you-performed any calcu-lations under Lloyd's rules to determine the adequacy of the design of the replacement crankshafts at Shoreham?

A. Yes. I have performed calculations under Lloyd's rules for maximum allowable' horsepower for the replacement crankshafts at Shoreham. Those calculations show that for 1680 psi, the highest peak firing precsure assumed by FaAA in its

$ studies at full load (350.0 kW), the allowable' horsepower per-mitted under Lloyd's rules is just under 4621 HP. Using the actual measured peak firing pressure of 1750 psi, the allowable maximum horsepower under Lloyd's rules is 4422 HP. In addi-tion, my calculations also show that for 1800 psi, the peak firing pressures at overload (3900 kW), the allowable horsepow-er under Lloyd's rules is just under 4252 HP. Shoreham's horse-power rating of 4890 HP at full load and 5379 HP at overload

{ exceeds the allowables for horsepower under Lloyd's rules. A copy of my calculations is attached as Exhibit 36.

- 114 -

l-I

~

Q. Messrs. Eley and takchi, hcyo you clco~ performed cal-culations-under Lloyd's rules to determine the adequacy of the design of the replacement crankshafts at Shoreham?

A. Yes. Our calculations confirm that the replacement crankshafts fail to comply with Lloyd's rules for maximum al-lowable horsepower. Our calculated figures are only slightly different from those of Professor Christensen. Our calcula-tions show that, for 1680 psi firing pressure, the allowable horsepower under Lloyd's rules is just under 4636 HP; for 1800 psi firing pressure, the allowable horsepower rating-under Lloyd's rules is just under 4269 HP. (Exhibit 37). Shoreham's horsepower rating of 4890 HP exceeds the allowable horsepower under Lloyd's rules.

Q. What accounts for the differences in your calcula-tions?

A. The minor differences result from conversions and roending of decimals -- such differences are normally encoun-tered when different individuals make computations of this na-ture.

Q. What is the significance of your findings?

- 115 -

6

_ . - . - _ . - ----..-.-.-._---.-.--_.---.._--._--_n

A. Lloyd's rule on allowable horsepower calculates the maximum. power that can be developed safely and reliably in an engine when taking into consideration its various parameters.

The Shoreham EDGs are required to operate at a higher horsepow-er rating -- 4,890 -- than would be considered acceptable for

' them under the Lloyd's' Rules. The failure of the Shoreham EDGs to comply with the allowabl'e horsepower. limitations under i

Lloyd's rules signifies that the EDGs cannc't be operated reliably at their rated power.

Q. What is the IACS?

A. The IACS is an organization consisting of three minor l

i and nine major ship classification societies, including the ABS and Lloyd's.

Q. Are you familiar with the draft rules of the IACS en- ,

titled " Rules for the Calculation of Crankshafts for Diesel En-gines"?

A. Yes. Those rules are based upon a proposal by an in-ternational group of engineers, CIMAC, entitled " Rules on Cal-l culation of Crankshafts for Diesel Engines (4. Draft)" which is i still under discussion among IACS members and between CIMAC and IACS. Portions of these rules are being used by the various

- 116 -

)

The ruloa ctate,that they ' ara to ba classi(ication societies.

applied for checking the suf ficient dimensioning of crankshaf ts for diesel engines for main propulsion and auxiliary purpos-l es.=124/ .

Q. What does a calculation under the IACS rules involve?

A. In order to determine the adequacy of the design of a particular crankshaf t under the IACS rules, you must first de-termine the' nominal alternating bending-and torsional stresses.

Those stresses, when multiplied by the appropriate stress con-centration f actors .using the deformation hypothesis (von Mises' Criterion), give a conparative alternating stress. The IACs rules state that adequate dimensioning of the crankshaft is en-sured where the ratio of the fatigue strength to the compara-

' tive alternating stress is greater than or equal to a f actor of safety of 1.15.

Q. Ha[e you performed any calculations to determine the sufficiency of the dimensions'of the replacement crankshafts

. under the IACS rules? .

4 124/ Exhibit 38, at 1.

l l

I

- 117 -

e

, + - - - -,-,4,. -

_.-_,m_,___..---e,

A. No, not directly. However, we have reviewed TDI's calculations under the IACS rules, a copy of which is attached as Exhibit 39 and we agree that they-are correctly computed.

Those calculations show that the replacement crankshafts at Shoreham do not comply with the IACS rules. Significantly, those calculations were performed by TDI using 1650 psi as the maximum firing pressure. As previously indicated, the actual maximum firing pressure in the Shoreham engines is higher (by as much as 100 psi at full load). When the correct maximum firing pressure of 1750 psi is taken into consideration, the replacement crankshafts fail to conform to the IACS rules by an even greater margin.

Q. Professor Christensen, have you performed any calcu-

- lations to determine the adequacy of the design of the replace *-

ment crankshafts under the ABS rules?

A. Yes. I have perforr,ed calculations under the ABS rules to determine the adequacy of the design of the webs on the Shoreham replacement crankshafts. Those calculations dem-onstrate that the replacement crankshafts do not meet the ABS rules. (Exhibit 40).

Q. What are the ABS requirements concerning crankshaft webs?

- 118 -

A. In order to provide for adequate bending stiffness in the crankshaft webs, the ABS rules dictate that the crankshaft webs should be in a specific proportion. Section 34.17.4 of the ABS rules for crankshafts with solid webs provides that The proportions of the crankshaft webs are to be such that the effective resisting mo-ment of the web in bending is not less than 60% of the resisting moment of the minimum required diameter of pins and journals in benling.

Normally, this rule is expressed in the formula ud't.35d , where w equals the effective width of the web, t equals the thickness of the web, and d equals the minimum required diameter of the pins and journals.

Q. Was it possible for you to use this formula?

A. No, because the Shoreham replacement crankshafts have a reentrant or crankpin fillet radius of 3/4 inches. The exis-tence of this fillet in the replacement cranksh' aft precludes the use of the formula because the effective resisting moment cannot be obtained from the rectangle created by w, the effec-tive width of the web, and t, the thickness of the web. This is so because the replacement crankshafts have a re-entrant fillet cut into the crankwebs. Thus, the fillet section, in-stead of being positive in value, is negative in value. If

- 119 -

B 4

EE E E I mi g

" " " "

b W k - ' l 11 , nl

1

.these. negative valuac cro not concidered in the calculation of the moment of inertia, the value of the resisting moment will be too high.

L Q. What are the-specific results of your calculations?

~

A. My calculations show that.the web strength in bending is equivalent to a crankpin or journal diameter of 10.9337 inches.- Using this value, I then calculated the maximum allow-able firing pressure for the replacement crankshafts. My cal-culations show that the' maximum allowable firing pressure under the ABS rules is 1746 psi at full load and 1651 psi at over-load. Thus, when the actual firing pressures of the EDGs are considered, the replacement crankshaf ts do not comply with the

! ABS rules at overload and are marginal at full load.

Q. Do you know of any other design standards that bear on the adequacy of the webs on the replacement crankshafts?

A. Yes. The rules of Nippon Kaiji Kyokai ("NKK") and the standards of the German register reflecting the experience of German engi'ne manuf acturers.

Q. Have you performed any calculations under the stan-dards of the German register?

- 120 -

i 4

e

. - , . - - - , .--~-.7, , . ---,-.,,~--,.y._ , , - - , . . . - - - ,-.m.,y- , , . . - .,--e.-v m_r,-,%.-- - ..~1..,w--

. l

. A. ; No,lbut wa havo rcvicwhd tho dopocition of Dr.

Pischinger, FaAA's diesel engine consultant, who has performed calculations using the German register standards. According to ,

Dr. Pischinger, this register is used for designing and assessing the adequacy of the design of diesel engines and is accepted by most of the European diesel engine companies.125/

According to Dr. Pischinger's calculations, which he described as " preliminary", the dimensions of the crankshaft webs are in-4 adequate under the German register.126/ Dr. Pischinger ex-pressly stated that he would have designed the webs thicker.127/ Dr. Pischinger also regards the cyclic stresses in the crankshaft as excessive under the register.128/ Dt.

Pischinger's preliminary conclusion was that the replacement crankshafts did not meet the standards of the German register at overload and were marginal.at full load.129/

e 125/ Deposition of Franz Pischinger (June 21, 1984)

- ("Pischinger Deposition") at 94, 97. (Exhibit 41).

i j 126/ Id. at 108.

127/ Id. at 98.

128/ Id. at 185-187.

129/ Id. at 100-101.

4 l - 121 -

1 l

I l I

Have you ncdo cny calculatienc concorning the cdsquc-O.

cy of the crankshaft webs under the NKK Rules?

A. Yes. Our calculations show that-the webs on the re-placement crankshaf ts do not comply with the NKK rules for full load or overload conditions. The NKK ' rules provide for two different ways to determine whether the webs are adequate. One method is based upon the relationship of the ratio of the breadth of the web and the actual diameter of the pin to the

- ratio of the thickness of the web and the actual diameter of the pin. The other method requires a calculation for the diam-eter of the pins and journals which takes into account various factors such as the maximum firing pressure. We made our cal-I culation using 1680 psi and 1800 psi as the maximum firing pressures. A copy of these calculations is attached as Exhibit

42. . _ _

Q. Have you undertaken any other calculations under the ABS rules regarding the adequacy of the design of the replace-l ment crankshafts?

A. Yes. We have evaluated the design from the stand-point of torsional vibration stress and found that the replace-ment crankshafts exceed the limits for torsional vibration stress set forth in Section 34.47 of the ABS rules. The total

- 122 -

l e

-~>--,m- g -n , -- <,---w ,w--m-,- ,,--,- e. w.-- 1, - ,ev- w ..m ---,+..e -- m

E torsional vibration stress imposed on the replacement crankshaft was calculated by FaAA to be 5,640 psi at the member

~

between Pistons 6 and 7. 130/ By contrast, the maximum stress

allowable for all harmonics is under the ABS rules for a crank-shaf t of the same design is 5,069 psi according to our calcula-tions. The calculations of torsional vibration stress by the

[

ABS yielded a slightly lower limit of 5,035 psi.131/ Thus, the total torsional vibration stress imposed upon the replacement crankshaft exceeds the maximum permissible under ABS rules for l

the design and materials in question by a f actor of more than 10 percent.

_ Q. Did TDI obtain ABS approval of the replacement crank-shafts?

l A. Yes. In effect, ABS has approved the Branksha'f ts in a letter dated May 3, 1984 from the ABS to TDI. ABS stated that it has "no objection to the submitted torsional critical f speed arrangement for use on diesel generator sets on an ocean 130/ Analysis of the Replacement Crankshaft , dated October 31, 1983, at 1-2. Dr. Pischinger, FaAA's viesel expert, s

believes that the Tn values used by FaAA in this calcula-tion are very reasonable. Pischinger Deposition at 110.

131/ Exhibit III to the Depositions of Richard Woytowich, Boward Blanding and Robert Giuffra (" ABS Deposition")

(July 18, 1984). (Exhibit 43).

E=

r

- 123 -

=

I _ - - ______

. going vessel, insofar as our classification requirements for marine service are concerned." A copy of this letter is atta-ched as Exhibit 44. However, the ABS letter was issued on the basis of special consideration of supplemental information sub-mitted by TDI which we believe is inaccurate and incomplete. .

Q. What information did TDI submit _to the ABS in seeking approval of the crankshafts?

A. TDI's submittal consisted of a " Report on Crankshaft Torsional Stresses, Transamerica Delaval Model DSR-48, Serial No. 74010/12 for Long Island Lighting Company," dated April 4, 1984. A copy of TDI's submittal is attached as Exhibit 45.

Q. Did the ABS issue its May 3 letter in reliance on all of the information submitted by TDI?

A. No. Because the predictions of torsional vibratory stress submitted by TDI exceeded the allowable limits under the published ABS formula, the ABS relied on supplemental informa-tion submitted by TDI -- the alleged effect of shotpeening the crankshafts, strain gage test measurements, and certain operating experience of the EDGs.132/ The ABS also performed 132/ I_d,. at 163, 165.

- 124 -

~

- - - - - - - - - - - - - - - - - - - 1

'its calculations using the value given to it by TDI for the maximum firing pressure'in the EDGs, 1700 psi.133/

Q. Did the ABS independently verify the accuracy of any of the supplemental information submitted by TDI?

A. No.134/-

-Q. What is the basis for your belief that the supplemen-tal information submitted by TDI and relied on by the ABS was incomplete and inaccurate?

A. We have reviewed testimony and documents obtained from TDI and LILCO showing that (i) contrary co the representa-tions in its submission to the ABS, TDI did not believe that sho.tpeening would substantially improve the fatigue endurance of the crankshaf ts, nor did *TDI disclose to the ABS that the original shotpeening of two of the replacement crankshaf ts was performed improperly; (ii) the actual maximum firing pressure in the EDGs is higher than the value that TDI submitted to the ABS; '(iii) the strain gage measurements are based on tests sub-ject to significant inaccuracies that aff ect the accuracy of i

133/ Id. at Exhibit III; Id. at 112.

134/ Id. at 167.

- 125 -

l l

l

4 the measurements, but TDI.did not inform the ABS of those i inaccuracies; (iv) the EDG operating experience data submitted j by TDI to the ABS did not include any of the many significant problems experienced by the EDGs.

Q. Has Suffolk County notified the ABS of your beliefs that the information submitted by TDI was incomplete and inac-curate?

A. Yes. By letter dated July 25, 1984, the County's counsel notified the ABS of our beliefs and identified specific data which we believe is more accurate and complete than the information submitted by TDI. A copy of that letter is atta-ched as Exhibit 46. There has been no response.

~

Q Did the ABS perform any calculations on the replace-i ment crankshafts?

i A. The ABS performed six calculations of combined safety factors for the replacement crankshafts under two methods and compared those calculated values against its desired minimum value for safety factor.

Four of those calculations showed that the replacement crankshaf ts did not meet ABS's desired minimum safety f actor value. None of those calculations considered the effects of

- 126 -

l l

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

shotpeening. Only when the ABS took into consideration the full 20 percent increase in the fatigue limit from shotpeening as submitted by TDI, did the ABS calculations show that the re-l placement crankshafts exceeded its desired minimum safety fac- ,

tor. A copy of these calculations is attached as Exhibit 47.

Q. In making these calculations, did the ABS ascertain whether the shotpeening of' the crankshafts had been performed properly or whether the shotpeening would in fact increase the f atigue limit of the crankshaf ts by 20 percent?

A. No. The ABS performed its calculations and reached its conclusions on the assumption that the crankshafts were shotpeened properly and that the shotpeening would in fact in-crease the fatigue limit of the crankshaf ts by 20 percent. The i ABS made no inquiry as to whether the shotpeening was performed properly, even though it believes that improperly performed shotpeening could increase the stresses in a crankshaft.135/

Q. Is 20 percent a conservar.ive minimal value of the in- .

crease in the fatigue limits from shotpeening the fillet re-gions of the replacement crankshafts as asserted by TDI in its submission to the ABS?

135/ ABS Deposition at 93, 98, 99. (Exhibit 43).

- 127 -

e - - - . -- - - - - - , . - ...,=...c ,.e -.- . ,-.. . . . - , . . .-,--....,r----..-. ,,s - - - -,,-,. ., v.+1,_,.,...,,,, , --..,--r. .,.-.---..,,,...w.- , -

A. No. In fact, TDI had recommended tgainst shotpeening the crankshafts based upon its experience and upon the opinion of its metallurgical consultant, Professor Wallace, that ,

J shotpeening would not provide more than a 5 percent increase in the f atigue strength of the crankshaf ts.136/ In addition, TDI was informed by Kobe Steel, Ltd., a Japanese manufacturer of l

' crankshafts for TDI, that shotpeening crankshafts of this size is "a waste of time" because the hardened depth by shotpeening was estimated to be quite shallow compared with the depth of the highly stressed area at the fillets.137/ TDI never informed the ABS of any of this information.

Significantly, in its April crankshaf t report, FaAA stated that "the effect of shotpeening may produce widely differing increases in f atigue endurance limi't; however, a conservative range of values of this increase is 5% to 20%.a138/ However, i

FaAA has withdrawn from that position in its May version of this report, which suggests no range of increases in fatigue The May endurance limit to be expected from shotpeening.

f 136/ Trussell Deposition at 45-48. (Exhibit 10).

137/ Letter dated February 17, 1984, from Shinpei Dench to f Gregory M. Beshouri. (Exhibit 48).

l 138/ FaAA Crankshaft Report, April 19, 1984, at 3-11.

- 128 -

+ -, m-+ ~ - - ,,,,e<< ,-<n- mswe, e c--,rn--w n- w,-- --mm---r.,--,-----ww.w-,-- v-vr+,- ---~~-wen,n-- - - - + - - ---,-ee-,e-,w- .,,w-~

report .merely states generally that shotpeening "will produce increase in fatigue endurance limit [ sic]."139/ According to

~

FaAA, the references to a numerical value were deleted when the !

' final version of the report was reviewed by FaAA's quality as-surance program.140/ We-are unaware of whether TDI has informed the ABS that FaAA no longer attributes any numerica_

value to the increase in the fatigue endurance limit from shotpeening.

i O. If the full 5 percent v..lue ? were used for the in-crease in the fatigue limit from the effects of shotpeening, would the replacement crankshafts meet the ABS's desired mini-mum safety factor?

Assuming that the shotpeening was performed properly, A.

and if _the full 5 percent value were taken into consideration

~

i using the f atigue limit values derived from the information submitted by TDI to ABS, the replacement crankshafts would not 1

meet the ABS's desired minimum safety factor (1.34) under one of the ABS's methods (1.2852),and would only marginally meet the ABS's desired minimum value under the other method (1.3713).

1 139/ FaAA Crankshaft Report, May 22, 1984, at 3-11.

140/ Transcript of TDI owners Group Meeting, May 24, 1984 at 114.

- 129 -

i e

,,+wma-,,---,,-v----n- - . - ,,,,am-n-,, -,,-,.-----a -

e-,,--- - , , --, - ~ - ~, - , ,-,

i Q. Were the crankshafts in fact shotpeened properly?

A. No. As described in greater detail below, two of the crankshafts were not shotpeened properly. As a result, the fa-j ,

tigue limits of the two erankshaf ts may actually be less than if they had not been shotpeened at all.

Q. Did the ABS safety factor calculations consider the actual maximum firing pressure of the Shoreham EDGs?

A. No, and it is pertinent that the maximum firing pres-sure is a significant factor in those calculations. The ABS calculated the safety factors based on a maximum firing pres-sure of 1700 psi,.the value supplied to it by TDI.141/ Howev-er, as we have explained above, the firing pressure in the Shoreham EDGs has been measured as high as 1750 psi at full .

' load and is conceded by TDI to be as high as 1800 psi at over-load of 3900 kW.

Q. Were the strain gage measurements submitted by TDI and relied on by the ABS accurate? .

141/ ABS Deposition at 112. Id,. at Exhibit III.

1 130 -

e

+w -, --e.~ , - - ~ ,v ,--- ~-,-e-~~ - , . - - - . .- ,-,.,,--,-------,,w, . - - . , r-m-

-A. No. TDI submitted the results cf strain gage mea-surements from tests on EDG 103 with a replacement crankshaf t and EDG 101 with the original failed crankshaf t. TDI, however, did not submit the actual test reports. Those reports explic-itly state that the strain gage measurements could be as much as 5 percent higher.142/

Q. Was the service experience of the.DSR-48 engines sub-mitted by TDI and relied on by the ABS comylete?

A. No. TDI submitted data on the service experience of the Shoreham EDGs, but that information consisted only of the total numbers of hours that the EDGs had operated, and the hours they had operated at 3500 kW and above (EDG 101 -- 114 hours0.00132 days 0.0317 hours 1.884921e-4 weeks 4.3377e-5 months ; EDG 102 -- 116 hours0.00134 days 0.0322 hours 1.917989e-4 weeks 4.4138e-5 months ; EDG 103 -- 110' hours).143/

TDI did not specify how many of these hours were at full load or how many hours were above full load. The ABS incor-rectly assumed from this information that the Shoreham EDGs had operated the entire 114, 116 and 110 hours0.00127 days 0.0306 hours 1.818783e-4 weeks 4.1855e-5 months above full load, and relied on this operating experience in issuing its May 3 let-ter.144/ It therefore appears that ABS was relying on a 142/ Exhibit 49 at 7-3, Exhibit 50 at 7-2.

143/ Exhibit 45 at 28.

144/ ABS Deposition at 81.

i l

- 131 -

\

. misunderstanding of TDI's operating experience data.

l Q. Did the ABS in fact know whether the EDGs had oper-

[

ated at loads higher than 3500 kW?

A. No.145/ In fact,Las of April 30, 1984, none of the Shoreham engines had accumulated as many as 100 hours0.00116 days 0.0278 hours 1.653439e-4 weeks 3.805e-5 months of operation above full rated load since the replacement of the crankshafts.146/

Q. Did TDI inform the ABS about the other abnormalities that have arisen during actual operating experience of the EDGs, such as the cracking in the blocks?

A. No. TDI only submitted information to the ABS con-cerning the number of hours that the EDGs had operatgd.

Q. Based upon'your review of the information submitted by TDI to the ABS, your knowledge as to what factors the ABS relied upon in issuing its May 3 letter, and your belief that significant information submitted by TDI and relied on by the I ABS is inaccurate and incomplete, do you believe that the ABS will reconsider the conclusions stated in its May 3 letter?

i 145/ Id. at 80.

146/ Design Review of TDI R-4 and RV-4 Series Emergency Diesel Generator Cylinder Blocks and Liners, June 1984 ("FaAA Block Report") at 1-8, 1-9, 1-10. (Exhibit 7).

i

- 132 -

- , . , --..----,-.,---v----.- ,.,.--.-+--.v.--~-w,..--, .vm, y ---,y,--*--y-.~,-y--.-ww.,-...- - , - , - - - . ----.-~-w-,-,-

A. Yes. In fact, Mr. Blanding of the ABS testified that the ABS would have to reconsider its conclusions if any of the information submitted to it by TDI were incomplete and inaccu-rate.147/ .

Crankshaft Shotpeening Q. What is shotpeening?

A. Shotpeening is a cold working process that produces a shallow layer of residual compressive stress on the surface of the metal being treated. The process consists of the bombard-ment of the metal surface with small beads of metal propelled by air pressure at high velocity.

Q. What is the purpose of shotpeening the crank pin fil- ~

let areas of crankshafts?

3 A. Shotpeening is intended to increase the f atigue re-sistance of the crank pin fillets, an area which is subjected to cyclic loading and which is the most critical area for fa-tigue initiation in a crankshaft. Shotpeening, however, cannot l

ine'rease the ultimate tensile strength or the yield stress of i the fillet material.

147/ ABS Deposition at 167-168. l l

- 133 -

Q. Would properly performed shotpeening of the crank pin

- fillets of the Shoreham replacement crankshafts significantly improve their fatigue resistance?

A. No. .Indeed, as previously stated, a major manufac-turer of crankshafts informed TDI that shotpeening crankshafts of this size is "a waste of time" because the hardened depth of the shotpeening is quite shallow compared with the depth of the highly stressed area at the crank pin fillets.148/ In addi-

- tion, the effectiveness of any shotpeening will be further re-duced if the material is subject to appreci.able heat as the crankshafts are.

Q. Did TDI recommend that the replacement crankshafts be shotpeened?

2 A. No. In f act, TDI recommended against shotpeening the

$ replacement crankshaf ts based upon its experience and the opin-ion of its metallurgical consultant that shotpeening would not increase the fatigue strength of the crankshafts more than 5 ,

percent. As Mr. Trussell of TDI explained, " Shot peening a h

148/ Letter dated February 17, 1984 from Shinpei Dench (Exhibit of Kobe 48).

Steel, Ltd., to Gregory M. Beshouri of TDI.

i l

- 134 -

I s

O

.nn, . . --,e,.._m- . . - . - . . , n,. . . ~ - n,.,. .__ , . . . . . -.n. , -- ,,... n. ,.- ,--,,-n.v,. , , , . - . , - - - ,m,- .-

thin piece of steel of the same specifications of the crank shaf t would substantially improve its f atigue strength, while applying the same surface improvement to a thick section, like a crankshaf t, would not provide d substantial improvement in the f atigue strength of the piece."149/ ,

Q. Generally speaking, are there any adverse side effects to shotpeening?

If not performed properly, shotpeening "could 2

A. Yes.

serve as a source of added stress concentrations which would make the crankshafts more susceptible to fatigue.=150/

e

, Q. Can even properly performed shotpeening cause any ad-i verse side effects? .

i t

A. Yes. Even if shotpeening is performed properly, shotpeening raises the stresses below the compressed surface.

When shotpeening introduces compressive residual stress on the

, surface layer, the adjacent underlying layers are put under tensile stress. This shotpeen-induced tensile stress is 149/ Trussel Deposition at 48. (Exhibit 10). See also Stone &

Webster Engineering and Design Coordination Report

("E&DCR") No. F-46109-G, at 4 of 4. (Exhibit 51).

150/ FRC Report at 65 (Exhibit 35); Pischinger Deposition at 168. (Exhibit 41).

l 135 -

._,,c.,-..- . , . , _ . _ . ,...__ w - - . , , - . - .--w,-4,-.- .m_.,mm_-m.-, w_,,,.-y,- - - - - - - , - , . ...--.-,o -- -#,-, ,-w ,- _.%.,. ,n .,

A fatigue additive to the already present calculated stresses.

failure does not necessarily have to begin on the surface of The criti-the fillet; it may begin in a sub-surf ace area.151/

cal area in this regard is the transition stage between the surface layer (which is under the residual compressive stress from shotpeening) and the layer of material further below in which tensile stresses is first experienced.

Q.

Were the crankpin fillet regions of all of the re-placement crankshafts at Shoreham shotpeened?

A. Yes, the crankshaft for EDG 101 was shotpeened once, while the by Metal Improvements Company at the Shoreham plant, crankshafts for EDGs 102 and 103 were shotpeened twice, once by TDI in Oakland and -once again at Shoreham by Metal Improvements Company.152/

Q. Was the shotpeening of EDGs 102 and 103 performed properly?

151/ Dr. Johnston of FaAA agrees. Deposition of Paul Johnston (May 9, 1984) ("Johnston Deposition") at 39-40. (Exhibit 52).

152/ FRC Report at 64-65. (Exhibit 35).

' - 136 -

i A. No. Although TDI's shotpeening procedure required

" full and complete intensity and coverage,"153/ some of the fillet areas of the crankshafts lacked shotpeen coverage. This was the first time TDI had ever shotpeened a crankshaft for a As reported by Stone & Webster " holidays,"

DSR-48 eng ine .154 /

or lack of shotpeening coverage, existed in the crankpin fillet areas. TDI reported that holidays occurred in two areas of the EDG 103 crankshaft: on the top of the number one crankpin di-rectly adjacent to the crankpin and at the outer edge of the crank radius.155/ Although TDI dispositioned the holidays as functionally acceptable, Stone & Webster recommended that the crankshafts should be shotpeened again.156/ According to Stone

& Webster, both EDG 102 and 103 were inadequately shotpeened in the lower third of the.re-entrant fillet of the crankshaft pin junction.157/

4 1

1 153/ E&DCR No. F-46109-G at 3 of 4. (Exhibit 51).

154/ Lowrey Deposition at 62. (Exhibit 24).

155/ E&DCR No. F-46109-G at 2 of 4. (Exhibit 51).

156/ M . at 1 of 4.

157/ Interoffice memorandum dated September 20, 1983, from Gary V. Luther to D. E. Ellis. (Exhibit 53).

- 137 -

e l

- . _ - - . - --.,---...-.-----,.,-.----,---._--,,,-,.,.,,,,,_,_,,_.,..-,.-n,-

, , - - , . , _ , , , , ,..n- ,- . - ,-- - . ., - .

t Q. Is the location of the inadequate shotpeening in the lower third of the re-entrant crankpin fillet area important?

A. Yes, the lower third of the re-entrant fillet of the

~ )

crankshaft pin generally is the most critical area with respect i to crankshaft failure. Furthermore, FaAA specified three rea-sons why'this area was most critical on the replacement crank-shafts. First, FaAA concluded that this area is highly stressed at loading. Second, FaAA discovered through x-ray diffraction that a residual stress existed at the fillets.

Third, FaAA found roughness in the surface finish at the fil-lets. Scanning electron microscope photographs in this area showed that cracks in the initial stages of propagation were initiating at one of the radially machined " valleys" of the fillet.158/ ,

Q. Have you inspected the original shotpeening on EDGs 102 and 103?

A. No. However, LILCO made available to us some, but not all, of the photographs taken of the original shotpeening.

From the photographs, it appeared that the depth of the 158/ Interoffice memorandum dated September 20, 1983, from Gary V. Luther to D.E. Ellis. (Exhibit 53).

- 138 -

undercut areas for machined tool runout was excessively deep in some areas, although it was difficult to tell how deep because of the effect of light and shadows from the photographs.

Reshotpeening would exacerbate the problem of stress raisers arising from the deep runout and may mask the critical point in way of the tool runout so that the residual compresive stress in these areas would be insignificant. In' addition, it appears that damage to some of the journal fillets may have occurred from deep single shot impacts which may act as stress raisers because the areas around the deep impacts go into tension --

the very thing to be avoided. Other photographs showed what appeared to be cracks in the shotpeened surfaces. In some cases it was possible to determine whether these deficiencies occurred in critical areas of the crankshaft, other photographs were insufficiently identified by their captions to be sure whether they showed pins or journals.

In addition, according to FRC's inspections of photographs of the original shotpeening , the surf ace texture of the shotpeened areas looked "more like grit blasting than shotpeening, i.e., the surface appeared to have been gauged

[ sic] by sharp particles instead of by round, smooth parti-cles.=159/ FRC believes that such improper shotpeening could 159] FRC Report at 64. (Exhibit 35).

J l

- 139 -

i

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

serve as a source of added stress concentrations to make the crankshafts more susceptible to fatigue.160/ We agree.

s Q. Have you come to any conclusions based on your review of these photographs and the documents identifying deficiencies in the original shotpeening?

A.- Yes. We have concluded that the original shotpeening of EDGs 102 and 103 was improperly performed and may have cre-ated nucleation sites on the fillet radii, the most critical area with respect to crankshaft failure. Shotpeening of pre-existing cracks in this area could cause the cracks to propa-gate further.

Q. Does the repeening of EDGs 102 and 103 by Metal Im-provements Co. alleviate your concerns about the original shotpeening?

A. No. Instead of adequately correcting the improperly performed shotpeening, the repeening of EDGs 102 and 103 serves l

to mask deficiencies already present on the fillet radii caused by the first shotpeening. The presence of nucleation sites due to the improperly performed shotpeening may not be 160/ Id. at 65.

- 140 -

i s

. - ---r- _ . . , . - . . - - _ . , , - , _ - . . - - - _ , .__._,_%-._____-.__._c,____m -.. __-_-. .,__ . _ _ _--__, _ _.w_,y

ascertainable due to the second shotpeening of EDGs 102 and 103.

Q.

ProfessorAnderson,doyouhaveanyconc.er$sabout shotpeening in general and about the procedures used for shotpeening the crankshaf ts?

A. Although it is generally true that shotpeening pro-duces compressive forces in the surface of the metal which en-hance its physical properties, shotpeening adversely affects .

the chemical properties of the crankshafts. .

The shotpeening procedure used for the Shoreham crank-shafts will produce some real reliability problems. Prior to shotpeening, the areas adjacent to the fillet radii are masked off. This result's in stressed (shotpeened) areas located di-rectly next to unstressed (un-shotpeened) areas. This differ-ence in surf ace energy is the driving force for corrosion and environmental attack of the fillet and stress cracking. Fur-thermore, since the un-shotpeened area is larger, the rate of corrosion is increased because of the cathode-anode area law.161/ 1 161/ Fontana and Greene, Corrosion Engineering, McGrah Bill 2d ed. 1978.

- 141 -

-, . . . . , - .-....-...-.---.-.,.---..,.~-n, ---..__e.,-,,n- , - . , , , , - - - - . . - . . . - , . . . , - - - - -

Q. Please summarize your conclusions about the replace-ment crankshafts.

A. - The replacement crankshafts should be required to comply with the Lloyd's rule on allowable maximum horsepower, the IACS rule for allowable safety factor, the ABS rules on ,

crankshaft webs and allowable maximum cylinder pressure, the NKK rules on crankshaft webs, the ABS rule on allowable tor-sional vibration stresses and the standards of the German reg-inter for crankshaft webs and cyclic stresses in order to en-sure their reliability for nuclear service. The failure of the replacement crankshafts to comply with these standards shows that the crankshafts are not adequately designed for operation at full load and overload and does not give adequate assurance that they can operate reliably.

Q. What are your conclusions regarding the shotpeening

of the crankshafts?

A. Any credit for increasing the f atigue strengths as a i

result of the shotpeening performed on the replacement crank-shaf ts is negligible. Indeed, the shotpeening may introduce detrimental effects. Furthermore, the shotpeening on EDGs 102 and 103 was not properly performed and may have caused nuclea-1

! tion sites which may not be ascertainable due to the second i

shotpeening of these crankshafts.

- 142 -

l l

CYLINDER BLOCKS Q. What is the purpose of this testimony?

I A. The purpose of this testimony is to set forth the re- ,

sults of our evaluation of that portion of the County's conten-tion which addresses the cylinder block problems of the EDGs.

That portion states: ,

" Cracks have occurred in the cylinder blocks of all EDGs, and a large crack prop-agated through'the front of EDG 103.

Cracks have also been observed in the cam-shaft gallery area of the blocks. The re-placement cylinder block for EDG 103 is a new design which is unprovan in DSR-48 die-sels and has been inadequately tested."

Q. What are your conclusions regarding the adequacy of the design and manufacture of the cylinder blocks?

A. We believe the block cracks are evidence that the EDGs are over-rated and undersized. The EDG cylinder blocks are not properly designed and manufactured to withstand the stresses to which they are sub'jected. We are concerned that LILCO' proposes to use the cracked blocks of EDGs 101 and 102 for EDGs in nuclear service during the operation of the Shoreham plant. Those blocks are unreliable and are likely to experience crack propagation which can lead to catastrophic

- 143 -

e

failure of the EDGs. The newly designed block for EDG 103 is unproven and inadequately tested.

Contrary to the conclusions reached by FaAA in the cylin-

- der block repore162/ and by the Owners' Group DRQR Report on cylinder blocks, we conclude that:

1. The cracks in the ligament between stud holes and liner counterbores of the blocks of the EDGs are not 4

benign and may be lead to catastrophic failure of the eng ine . ,Further, the cracks may not be fully 2

contained between the liner and the region of the block top outside the stud hole circle.

i

2. Field experience in non-nuclear service has not been systematically documented or reviewed in order to demonstrate the extent of ligament cracking or the

immediate consequences of such cracking.

3. The deepest crack (5-1/2 inc'h depth) between stud holes was measured after the immediate shutdown of EDG 103 following crack propagation during overload

~

i 162/ " Design Review of TDI R-4 and RV-4 Series Emergency Diesel Generator Cylinder Blocks and Liners," FaAA-84-5-4, Fail-ure Analysis Associates, June, 1984 (the "FaAA Block Ra-po r t" ) . (Exhibit 7).

- 144 -

1 i

p testing of EDG 103, and contributed to the decision to replace the block. The replacement block has not l

been adequately tested.

4. Blocks with ligament cracks (those of EDGs 101 and l

102) have not been demonstrated to be capable of withstanding a LOOP /LOCA event. While we agree with l FaAA's conclusion that cracks between stud holes are likely to occur and propagate in blocks with ligament cracks, we disagree that FaAA can predict with any accuracy when such cracks will initiate or the rate at which they will propagate.

5. The preliminary material evaluation by FaAA of the i microstructure of a small, region of each block top of ,

the EDGs is not representative of the properties of the entire block and does not demonstrate that the block EDG 103 is significantly weaker than the other two blocks. To reach conclusions regarding the suf-ficiency of the material strength of the blocks of

' EDGs 101 and 102 in comparison to that of EDG 103, the material of all three blocks must be adequately evaluated.

I 145 -

e

_ . , . . ~ , . . , , . . . - . , . , _ _ - , - _ . , .. ._...,.._..-,...,_.,,_.-_.,__,_..e--, --.,..w,.- . - .

6. The cracks in 'the can gallery support region of the EDG blocks may be detrimental to the operation of the engine. Further, the assessment of these. cracks has y

failed to demonstrate that the cracks will grow very ]

slowly at full load and not at all at'75 percent load, or that the cracks can be attributed sorely to the casting process.

Based on the foregoing, we conclude that it has not been demon-strated that the cylinder blocks of the EDGs will reliably per-form their required functions, and thus, there can be no assur-ance that the EDGs will perform satisfactorily in service.

Q. Please describe the cracks which have occurred in the cylinder blocks of the EDGs. , ,

A. There is no disagreement that numerous cracks exist on the block tops of EDGs 101 and 102, running in the radi-

.al/ vertical plane between stud holes and the cylinder bores.

These cracks are shown in drawings, and some of them are de-scribed, in the FaAA Block Report.163/ Similar cracks were l-found in the top of the block of EDG 103, which also had cracks between stud holes for adjacent cylinders 4 and 5.164/ On 163/ FaAA. Block Report at 1-2 to 1-3 and Figures 1-2 and 1-3.

( 164/ Id. at 1-2 and Figure 1-4.

- 146 -

I

- - = - - + , - . . , _ _ . . . , , _ _ _ , , _ _ _ , . , . , , _ _ , , , _ _ .

April 14, 1984, during qualification testing at 3900 kW, o crack was noticed starting under the no. 1 cylinder head and extending across the front of the EDG 103 block and about 5 inches down the front of the engine.165/ Subsequent inspection of the EDG 103 block showed that many. existing cracks had prop-agated, and that additional between-stud hole cracks had developed at four other locations.166/ In addition, there are cracks in the camshaft gallery areas of all three EDG blocks.167/ These cracks have been observed to grow in the EDG 103 block.168/

Q. Does the F5AA Block Report provide a satisf actory de-sign review of the cylinder blocks?

A. No. Rather than a design review of the blocks, it is a summary of FaAA's " investigation of the structural adequacy" of the blocks.169/ FaAA fails to address most of the 4

165/ Letter dated April 17, 1984, to Administrative Judges from E.J. Reis (NRC Staf f) . (Exhibit 54).

166/ FaAA Block Report at 1-2 to 1-3 and Figures 1-5 to 1-8.

167/ Id. at 4-6.

168/ Morning Report, NRC Region I, March 20, 1984. (Exhibit 55).

169/ FaAA Block Report at i and 11.

- 147 -

l 1

e--- , . . , - , - - . - - - - - - - , , , - - - - , .,e .. - - -, .-

l 1

functional. attributes of the cylinder blocks set forth in the Task Description for the Component. Design Review.170/ we believe it is significant that FaAA does not conclude that-the cylinder-blocks are adequate for nuclear service ar.d capable of

' unlimited operation. However, based solely upon the FaAA Block Report and its supporting packages, the TDI Owners Group con-

- cluded that the cracked blocks of EDGs 101 and 102 and the re-

)

i. placement block for EDG 103 (pending final material study re-sul s for the original and replacement EDG 103 blocks) are acceptable for intended function with implementation of routine inspections in accordance with E&DCR F-46505.171/

Q. What does the TDI Owners Group mean by the phrase

" acceptable for intended function"?

a *

! A. The DRQR Report does not expressly define this 4

phrase, but indications are that it refers to the ability of the cylinder block "to withstand with sufficient margin a LOOP /LOCA event."172/

There is no suggestion of what a "suffi-cient margin" might be. Mr. William Museler, a vice president 3

j 170/ Id., Appendix. l 171/ DRQR Report, Vol. 4, Cylinder Block, at 3. (Exhibit 56).

l 172/ Id.. at 2; see also M . at C1 and C2.

1

- 148 -

l l

l m,~ - -- -g-,

. . - - , . - . , - - , - , , . ,-,-,.m.a. --m.-., -

e.g..n_, n., - - , .-m--.. -,w, n..,+.n,w___-

of LILCO and former technical manager of the TDI Owners Group program, testified that the ad hoc acceptance criterion applied l by the Owners' Group program for adequacy of the EDGs was not the performance rating of the EDG established by the FSAR and the contract specification.173/ Rather, the TDI owners Group i

critIerion was reliable operation during the testing required to be performed plus one LOOP /LOCA event for seven days.174/

Q. Is the TDI Owners Group acceptance criterion intended to be applied to qualify the EDGs only for operation during the 4

approximately 18 month period until the first refueling outage I at Shoreham, when the newly purchased Colt EDGs are scheduled to be installed?

A. Nog according to Mr. Museler. He testified that al-

. though LILCO intends to replace the EDGs with Colt diesels by the first refueling outage, the owners Group criterion was in-tended to qualify the EDGs for a period "far beyond the interim i

period."175/

173/ Deposition of William J. Museler (May 22, 1984) ("Museler De po sition" ) at 7-8. (Exhibit 57).

174/ Id. at 14-17.

175/ Id. at 43-46.

4 .

- 149 -

Q. - - - - ,- - , ,-c-- - , - - . .

Q. Is the criterion used by the TDI Owners' Group cppro- ,

priate to ensure that the EDGs, and specifically their cylinder blocks, are adequate and reliable enough to meet the require- i ments of GDC 177 .

A. No. The Owners Group criterion is extremely limited, subjective and does not meet the technical requirements of GDC

17. As discussed above, the proper technical standard for GDC 17 is the performance rating for the EDGs set forth in the FSAR. That rating -- 3500 KW continuously for one year and 3900 kW for 2 hours2.314815e-5 days 5.555556e-4 hours 3.306878e-6 weeks 7.61e-7 months per 24 hour2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months period -- was established by LILCO and approved by the NRC Staff on the basis of the required service for the EDGs. There is no rational or regula-l tory basis to eliminate that performance standard.

~

Q. Did th,e FaAA Block Report use the same improper ac-captance criterion as the TDI Owner's Group for determining the adequacy of cylinder blocks?

4 L

FaAA issued an interim report on the cylinder block j A.

and liner, which concluded preliminarily that the DSR-48 cylin-der blocks may be adequate "for interim use" depending on fur- l i

ther analysis.176/ Mr. Robert Taylor of FaAA, who prepared the i

i 126/ Exhibit 1 to Taylor Deposition. (Exhibit 58).

i

- 150 -

. , - , , * ,.-w ,~ - ----------n .,c-----------en-.. - . - - . , . --.--- , - - - . , - .- - , - -, .-nm---,

s

[

i use,' he testified that in determining " inter m ;

intoria report, f about 260 .

cred an " intended load profile" for two years o d

'h::rs of EDG operation, including 80 hours9.259259e-4 days 0.0222 hours 1.322751e-4 weeks 3.044e-5 months at full load an In the final FaAA Block Re-1c00 than one hour at 3900 kW.121/ ot the cylinder is made as to whether or n use, so no accep- ,

port no statement h blocks are adequate for interim or any ot erHowever, FaAA appears t t nce criterion is expressly applied. er criterion of h ve further reduced the inadequate and impropbecause the FaAA tho two year " intended load profile," h an engine block with +

R port only specifically addresses whet er (so called cracks between the stud holes and cylinder borehole to stud ho aligament cracks'") , but with no stud ll8/ This crite-ccn be predicted to survive a LOOP /LOCA event.tandards required b rion is totally inadequate to satisfy the s .

s i

GDC 17.

of conclu-The FaAA Block Report sets forth a number Do Q. licable to the EDGs.

sions and recommendations which are happ racks in the liga -

you agree with the FaAA conclusionrbore thatare tec ment between the stud holes and liner counte

" benign."lll/

(Exhibit 59) .

_122/ Taylor Deposition at 69-70.

128/ FaAA Block Report at 4-3 to 4-5.

119/ Id,. at 5-1.

- 151 -

. . . _11M_E"l_ ._ _

ht We strongly disagree with FaAA's ccn31usicn t a A. First, FaAA states, and we these ligament cracks are " benign." ks might be agree, that one consequence of the ligament cracsuch

) llE/

leakage of coolant (although not into the cylinder trophic. ,

1cakage is far from " benign," and could lead to catas failure of the EDG. .

Now could the leaking of coolant lead to a cata-Q.

strophic failure?

ra-A.

The leaking of the coolant could result in tempe i d r liner and ture increases of the upper part of the cyl n e d block, The consequent thermal stresses on the cylin er head. ts increase cylinder heads, pistons, and other engine componen For example, the overheating of the likelihood of cracking. , d/or cause a partial the cylinder liner could crack the liner an A partial piston seizure makes combustion gas piston seizure. k se explo-blow-by highly probable, which may lead to a cran ca Lack of sufficient coolant sion and complete piston seizure. head, which could could also lead to distortion of the cylinder l Distor-cause the exhaust valves to fall to seatf gases complete y.

from the tion of the cylinder head and the leakage o

( at 11 to 111.

119/ id.

- 152 - l G

l- +

exhaustvalvescouldleadtooherspeedingoftheturbocharger and damage to the blades and rotor, which would stop the ;

I turbocharger. This 'would ' result in an insuf ficient quantity of air supply t,o the engine, further increased temperatur,es of the l

-s i-operating pa'rts, and ultimately to a complete piston seizure.

Complete piston seizure would cause bent or broken connecting rods, serious overloading and possible cracking of the main bearing shells,'dracking in the engine base and stretching of the main bearing hold down studs. A complete piston seizure will almost always_stop.the EDG.

Q. Can you predict how quickly the coolant would leak from the ligament cracks?

A. Coolant water could l'e^ak rapidly f rom ligament The coolant water is under pressure of 40 psi. The cracks.

rate of leakage would depend on the number of cracks and their widths. The leakage becomes crit { cal when the expansion tank (coolant reservoir) either cannot replace the loss of coolant water fast enough or is depleted. A dangerous overheating con-dition occurs when the temperature is high and the water low so' that the circulating coolant mixture consists of liquid and vapor. 1 i

\ 4 153 -

~ ,

e b

6

,-,.,,.w,- h+--,,- ,m-- . , - - - -. . - - --w- -v- -e-- w ,.w--,- - - ,-wwwv-,y-y

Q. Do you agree with FaAA's conclusion that the ligament cracks are benige because the cracked section is fully contained between the liner and the region of the block top outside the stud hole cir-cle.181/

A. It is not clear what FaAA means by this description.

FaAA describes the ligament cracks accurately as running be-tween the stud holes and the liner counterbore, so the cracks do-run to the stud hole itself. We believe that FaAA is refer-ring to the " apparent arrest" of the ligament cracks at the liner landing ledge.182/ This conclusion as to the " apparent ,

arrest" of ligament cracks is based upon observation of liga-ment crack depth on the EDG blocks, and unconfirmea183/ and in-complete inform.ation regarding selected blocks of TDI engine's in non-nuclear service. _

Q. Were ligament cracks " fully contained" during the testing of the EDGs?

l l

181/ Id. at 5-1.

182/ Id. at 1-2 and 1-3. .

183/ Id. at 1-1.

t 154 -

O

- - - , , w , - - + - , - - , - . , --n- a- - . ~, , , ,- ,

No.

The history of the ligament cracks on the EDG A.

blocks does not support the conclusion that they are " fully on the contrary, the large contained" and therefore " benign." d I 5" crack which occurred on the EDG 103 block during overloa i i h already testing ran from a stud hole at cylinder No. 1 wh c Compare Figures 1-4 and 1-8, FaAA Block had a ligament crack. d Report.

That comparison also discloses that after the overloa k had test was aborted, nine new stud hole to stud hole crac s Thus, even if the ligament cracks on the EDGs had initiated. t be not propagated downward past the liner landing, they canno If the ligament crack is in fact arrested described as benign. i at the liner landing ledge, it would appear that continu ng sufficient operating stress causes cracks to initiate and lig prop-a-

agate radially and vertically from the stud hole with the .

ment to adjacent stud holes or to the outer wall of'the Finally, Figure 1-8 contradicts FaAA's assertion block._184/

that ligament crack,s will not grow beyond the 1-1/2" depth of ks the liner landing ledge, because it shows six ligament crac with a depth of 2 to 2-1/2."

184/ Note that Figure 1-8 of the FaAA Block Report shows that most of the ligament cracks had reached a depth of at least 1.5", the reported depth to the liner landing.

- 155 -

-m -

, ,wm-o -r- --w -'~w ----~ =--=~ - - - - ~ - - - , - ---m-- n--~ -- --,~w-~vv= - - - - - - - - - -

n.

in non-nuclear Q. Doesn' t FaAA's data on cracked blocks

" benign" and carvice demonstrate that the ligament "?lll/ cracks are etnnot have adverse "immediate consequences FaAA The unconfirmed information given in the

. A. No.

l ion at all.

Dlock Reportl8{/ does not support FaAA's conc in initiation usthe FaAA concludes that the mechanism of crack during startup to cylinder block tops are low cycle fatigue firing pressure c.

high load levels, high frequency fatigue fromng at loads ab i 0 tresses, and overload rupture occurr p These factors, which also affect crack pro -

power levels.181/ i h an engine is cgation, are all related to the loads at whthe c stress and run, that is, the higher the load, the dgreater rapid propagation.

the more likely is crack initiation an have run, but does '

FaAA states the hours which theduring '

non-nuclear those hours.

f not disclose the loads at which they ran lied at all on the We believe it inappropriate that FaAA has re ked why FaAA had .

marine non-nuclear cases they cite.When ash at Shoreham, I decided not to examine cracks in blocks other t an Mr. Taylor of'FaAA responded:

ly / FaAA Block Report at 5-1.

18_6,/

6 Id. at 1-3 to 1-4.

181/ Id. at ii.

- 156 -

e i

~'-"*7-y99 " ' " " ** %,,__ 'New%., _

l

. j Well,-the engines in the Marine service see '

a different service Their loadthan profiles shore-based are different. en-gines.

t They' re operated dif ferently, and justlooking at the b I I

out knowing the size of the liners, how much the liners protruded, exact load his- ;

tory, even if I were to go look at that block, I wo*21d -- there's a wealth of other data that would be pertinent that I don' t '

have yet and orobably would not be able to reconstruct.lB8/

would ,

Mr. Taylor also testified that data such as load factors FaAA concedes ,

i mak'e examination of other cracked blocks useful. loads and that.non-nuclear engines generally operate at lower f

/

with fewer starts than nuclear diesels.189 . i Do you have additional comments on the specific cases Q. A?

of non-nuclear engine block cracks relied upon by FaA !

M.V.' Gott does not dis-A. Yes. The information on the by ;

hd r close load levels for this DMRV-16-4 engine, the met o s sult

^ h which crack depth was measured, or the fact that as t,e re ified.120/

of the cracks the engihe blocks were repaired and mod for its During the telephone conversation on which FaAA relies 4

188/ Taylor Deposition at 40- 41. (Exhibit 59).

189/ FaAA Block Report at 4-3. to 1983 from Lowrey (TDI)

~~~/

190 Letter dated November 30,Blanding (American (Exhibit 60). Bureau of Ship 4

- 157 -

i I

i

- -- . , , . , . . _ . . , ,, .n... ,.. , , , , _ , ,_ . , , ..,_.,..n.na..,,..,,,-, . . . , , , , - . n ..,,_.,.-,,,,___,._n.__ , ~ . , ,E

l-the owners also told FaAA'that I

'information on the M.V. Gott, d to reduce (i)Lthe' blocks onethe Gott were being machine i l l the- engines on the Gott had been. so extens ve y ctresses, (ii) [.

idered "stoc k" TDI die- p todified they could no. longer be cons i s a maintenance / inspection program for the eng ne cols,.(iii) ram was __

cuch~more comprehensive than the recommended ith a cyl- TDI prog i l f

baing used, and (iv) 'the design of the TDI blocks, design which w inder liner placed in a counterbore, "is an oldmal prob-h nobody uses anymore because of the resulting t er infor- ,

1 The FaAA Block Report fails to disclose this !

lems."_91/ :

V nation. '

disclose load i

The statement on the M.V. Columbia fails to k d block levels or that the State of Alaska replaced the crac et ly .43%.1 l-I' and derated the TDI DMRV-16-4 engines by approxima j' r 35 HP less e - _ _t, '

Further, these engines were originally rated at ove '

Information on the St. Cloud, Cop- '

per cylinder than the EDGs. do not disclose per Valley, Homestead and Bhiel engine blocks l

Block Report Ref.1-3, Memo of Juneand 7,1984 te e-Liberty 191/ FaAA

--~ phone conversation between Spiegel (FaAA)

(U.S. Steel). (Exhibit 61). History.of, 192/ Evaluation of the Operational and Maintenance in the M.V.

i Sys-and Recent Modifications to, the Main EnginesC at 2-1. (Exhibit 62).

tems, Inc., April 1983, l l

- 158 -

l l

--c-, ., . * + - - M

e '

in lead levels or other pertinent operating information, such as !'

The engine at Homestead is rated at paak firing pressures. Three of the ;r i

8800 kW, but'is operated at only about 6000 kW.

by 20%.

TDI engines owned by Copper Valley have been derated from Maintenance history documents obtained by LILCO ort FaAA of Copper Valley disclose many problems, including replacemen to a block on engine S/N 75011, but do not specifically refer '

Finally, FaAA has supplied ligament' cracks in the blocks.193/

h i al

' no information on the block material properties or c em c Yet i

composition of the cylinder blocks in non-nuclear serv k ce. i FaAA believes these factors are very important to crac In summary, FaAA's information initiation and propagation.lli/ l ion that on non-nuclear service does not demonstrate its conc us the ligament cracks on the EDGs are " benign." .

Q.

Do you agree with FaAA's conclusions that ligament ~

d to occur cracks and stud hole to stud hole cracks are predicte high after operation at high loads and/or engine starts to load?lll/

i

_.193/ Maintenance History (Exhibit on TDI S/N 63). 75011 and 75012, Copper Valley Electric Ass'n.

iv.

194/ FaAA Block Report at 4-5 to 4-6, 195/ Id.. at 5-1.

- 159 -

7 o

e

p

~~ .

h~the

'A; Yes. But FaAA understates thetimates strescas the to whic i

biceks of the EDG are subjected, and thus underes t cracks .

likolihood and rapidity of the initiation of ligamen d of propagation lcnd stud hole to stud hole cracks, and the spee Thus, PaAA has failed to demonstrate that .

.Of those cracks. i bly with- ,

biccks with ligament. cracks are capable of rel a Otanding a LOOP /LOCA event.

re Please explain why you believe these stresses a Q. .

underestimated by FaAA.

h block First, FaAA understates pressure loads on t e A 00 p'sil91/ rather by assuming a peak firing pressure of only 16 t r at 1004 load..

than the actual value of 1700 psi or grea e loading Second, FaAA has not properly determined ithe pre col.lar i otress or how much of the preload isis borne borne by bythe the l ner onto the liner landing ledge and how much FaAA states that "much" of the preload is block.191/ i pon several vari-transmitted to the liner collar, depend ng u ables. But it does not address these variables in terms o The liner collar their importance or give any calculations.

196/ Id. at 2-3.

191/ Jd. at 2-1.

- 160 -

'O

~-,

l on the EDGo is protrusion, or " proudness,' above thedblock wouldtopresult in gecater than current TDI specifications, an FaAA measured the liner landing ledge.121/ i -

i grcater preload on G 103; the measure- ;

i tha liner proudness for the cylinders of ED csnts varied from 1 to 9 mils.121/ !

f thermal load Third, FaAA has not calculated the amount o FaAA 3EE/

cn the block due to thermal expansion oft the essliner.

of the liner correctly points out that thermal expansion sr depending upon the f ,

will not all be transferred to the block, I lEl/ But there are no ;

clearance between the liner and block. ounts of stress ;

calculations of the optimum clearance or the amFurther, there not transferred under those optimum conditions.h blocks of are no calculations of the actual tement clearancesthat "in- in t e t

the EDGs, so there is no basis for FaAA's staall as possible.=22 terference stresses in'the block are as sm t 198/ Id. at 1-5 s of DG 103, Project No.

d (Exhibit 64).

~~~/

199 Calculation " Liner Prou nesLau, dated 6/10/84.

' 03315A", by John H.

I 200/ FaAA Block Report at 2-2.

201/ Id. at 2-3.

- 20,2 / J,d_.

- 161 -

l l

chow -

element analysis accurately.

Q. Does FaAA's finite the block? ~

lb offects:of' stresses.on the top ccurately of reflect-g No.

The FaAA analysis does not ait incorrectly assumes [

A. First, ly .

stress effect. 1600 psi, thereby significant Octual probable ,

Second, it assumes a peck firing pressure of only stresses due-to pressures. y to .

understating the n the liner andIfblock necessar the optimum clearance betwee l expansion.202/ the actual close the clearance by therma the assumed optimum, '

l each cylinder is less thanThird, FaAA assumes therma clearance for ,

stress effect will be greater. This would only the cylinders. re the otresses are symmetric betweenand load in all cylinders we occur if the firing pressure differ significantly from i perating manu- i stmo.

Actually, firing pressuresof the same E cylinder to cylinder of + 100 psi.

j i

i f the variance al permits a dially in the plane of the top sesoin the thermal stresses act ra are also longitudinal stres is Actually, there thermal stress pattern block.

surfaces of the block, so the i upper I I

an oval shape.

l i, . ,

j

/ g. at 3-3.

- 162 -

O Wd*-

---w, _ __ =V't,e%. - - - - . - - - - - . - , , , ;, . . ,

j- ,. -

. 4- _

. rrect &nd/or j L..

-Q. Please explain how FaAA's incoffect its conclusions f' dicted to nan-conservative assumptions ato stud hole cracks are pre ..

CGnt' cracks and stud hole blocks? I

~ and propagate in the cylinder .

Tinitiate .

k could occur in fewer FaAA predicts that these crac s teady '

A. 0 to 90% power or above and/or s

.than 100 starts from 90% or higher power, with a oparation for over 100 hours0.00116 days 0.0278 hours 1.653439e-4 weeks 3.805e-5 months aterties.201/ The incorrect

~

d block having minimum material i prop s of FaAA and its understate m

i and/or non-conservative assumpt on33 kai (asmean compared to th pack total stress figure of f 32 ksi for a 2-1/2 section) d con-ultimate tensile strength ol initiate under FaAA's predicte

~

i that the cracks might wel h n minimum material I ditions in blocks having higher t a40 gray cast iron, or at belowi.

s, properti'es for ASTM A48-64 Classtion for fewer than 100 hour0.00116 days 0.0278 hours 1.653439e-4 weeks 3.805e-5 months pa .

90% of. power or at steady opera It is not possible to factors.

i or any combination of these conclusion is in error be-l state by what percentage the FaAAh as actual firing pressur .

cause the many variables, suc and " proudness" of the liner for inder block and liner clearance,ithout further experimental dat are impossible to predict w a specific engine.

h04/16,.at3-6 i

- 163 -

l r

- 'O

a .: 9

r. *l initiation to occur at stecdy I[ ',

FaAA predicts crack or above.lEl/

Q.

than 100 hours0.00116 days 0.0278 hours 1.653439e-4 weeks 3.805e-5 months at 90% power 90% cracks can initiate

running'for more FaAA's W3uldn' t one expect that at loads abovef operation, even ta Ct'fGwer than 100 hours0.00116 days 0.0278 hours 1.653439e-4 weeks 3.805e-5 months o ?

incorrect assumptions as correct hours .

operating load, the fewer The higher the FaAA does not ad-A. Yes.

ks initiate. A 90%

would ime required before crac ignificant omission.

required actu-dreas this issue.221/ This is a sW, well below the a lord on the EDGs is only 3150 k G is required to carry two during ;

P/LOCA event, al naximum load After of103881 minutes kWinto anaED LOO LOOP /LOCA event. i m coincident demand of about EDGs must each produce a maxWhen mu2gl/ this factor is com-i 3400 kW, or 97% of rated load. from past start-ups and {

block bined with accumulated damage,h t cracks can initiate in a '

operation, it is apparent t a than 100 hours0.00116 days 0.0278 hours 1.653439e-4 weeks 3.805e-5 months .

during a LOOP /LOCA is much less <

a 1

i i

r 205/ Id, state that 110% load "is clea -

00% load than 100% load is 206/ The- FaAA Block (at 4-1).Report doesly more damagin relative to 90% load" 2,02/

0 FSAR Table 8-3.1-1 at 4.

- 164 -

1 I

T gww. [ l 4

'~*%-.. ' ___ "Wu_ _ , _ 4e

FaAA suggests'that stud holo to stud hole cracks bl it .

Q. I

[

night not be dangerous, because "the deepest measured crack in -

did not degrade' engine operation

- this region (5 1/2-inch depth) [

or result in stud loosening."208/ Do you agree?

r A. No.

FaAA fails to state, indeed if it knows, when I

this crac,k grew to a 5 1/2 inch depth or how long EDG 103 oper- l Even if we assume that this crack grew oted with this crack. I during the " abnormal load excursion" affecting EDG 103 on April .

2 hours2.314815e-5 days 5.555556e-4 hours 3.306878e-6 weeks 7.61e-7 months before it 14, the engine could only have run less than The very deep was shut down and the crack was discovered.2091 stud hole to stud hole crack contributed to the decision to re-Such cracks could cause the loosening and place the block.

breaking of the cylinder head studs, with consequent loss of .

This condi- ,

power and overloading of the remaining cylinders.

tion would probably lead to engine failure. ;

Q. FaAA concludes that the cracked blocks on EDGs 101 and 102 can survive a LOOP /LOCA event if they have no cracks between stud holes and if the block material of the original 208/ FaAA Block Report at 5-1.

209/ Ioad 1-2. EDG Id. atexcursion," ran for 10 minutes after the " abnormal then was run for 100 minutes before being 1

i shut down when the 5" crack running from cylinder no.

was noticed.

- 165 -

L._. _ ._

l leec racistant to EDG 103 block *is shown to be sufficient y 40."110/ Do you fatigue than typical gray _ cast iron, class !

cgrse? orted The FaAA's conclusion is based upon a purp

A. No. d growth in cbility to accurately predict crack initiation anknown lysis of the EDGs 101 and 102 by " cumulative damage 3/11/84 ana and cxperience during operation of DG 103 between ises FaAA's analysis is based upon faulty prem 4/14/84.=211/ FaAA cannot accurately predict whether and insuf ficient data. 01 and 102 may cause cnd when the cracks in the blocks of EDGs 1 a failure during a LOOP /LOCA event.

What are FaAA's faulty premises? '

Q.

l tive dam-FaAA bases its analysis on a " linear cumu a to obtain the total fa-A.

i 4.1)

Ege approach (presented /

in Sect onThe use of the linear fatigue tigue damage" of a block.ll2 that is, it is assumed ap-damage index is not limited by FaAA, d duration. Ex-plicable for all ranges of stress, iload anare known to result in tremely high loads for a short durat on is not reflected failures or excessive cracking;211/ this fact 210/ Id. at 5-1.

211/ Id,. at 4-3.

212/ Id. crack running from

' 213/ Indeed, FaAA emphasizes that the largef the EDG 103 bl the no. 1 cylinder down the front o(Footnote cont'd nex t

- 166 -

l

I

,y :

^

Further, FaAA assumos thot the- X

'by FaAA's linear. damage index. /11/84 and 4/14/84 damage index recorded for EDG 103 between h vior of3 other '

10 an appropriate . benchmark to predict the be a j'

! On this basis, FaAA concluded that:

I biccks.

A block with no existing stud-to-stud l cracks and material properties sufficient y better than those'of to complete the LOOP /LOCA requiremen DG 103 t

/2 should s be able without any cracks as deep as the 5-1 inch crack in DG 103, while continuing to run normally.211/ .

i is erroneous.

However, the assumption for this conclus on .

What are the errors in the assumption?

Q.

hich First, it completely ignores the large crack w A. d testing and ran from appeared in the EDG block during overloa block front, resulting ,

cylinder no. 1 about 5 inches down theshutting down .

in aborting the test, the block. The damage .

contributing to the decision to replace bility of an EDG h

caused by that crack and its impact on t e da by FaAA. Second,

" continuing to run normally" is not assesse (Footnote cont'd from previous page) load. FaAA occurred after a 23 second unusually high Block Report at 1-2.

214/ FaAA Block Report at 4-5.

- 167 -

--- ,-- -_.. -r. . , . , . - . . - , . _ - _ _ _ . - . . . _ . . - - ,, . . - . , , a w-e9**i' --

~~ .-. __

._.-._____ ~

c, - ;s 1 and 102 in comparison i f'f} [a cyplying . FaAA s damage index to EDGs 10d period does no .

to the EDG 103 index for the statetra on the three en- .

t l ccccunt the effects of differing load' spec of loads as- c '

l Crack dynamics are affected by sequence FaAA provides insufficient evidence ginas.

wall as their duration. tated period is a worst

_ that' the EDG 103 block damage in the s possible case.

. ith the validity of

-Do you have other concerns w Q.

FaAA's analysis?. ity to re-Although we have not had an opportun A. Yes. lations which were only ob-view some of FaAA's underlying calcuerned with FaAA's conclus toined a few days ago, we are conc required to' initiate cracks that an amount of additional damage initiate must at Least between studs after ligament cracks d to initiate the ligament ~

equal _ the cumulative damage require take into ac-This conclusion does not appear to cracks.311/ l' ent analysis, which count the results of FaAA's finite e emformed, the transverse shows that af ter ligament cracks have This increase in stress i- 211/

stress between stud holes doubles.

2y / Jd_. at 4-1. ,

211/ Id. at 3-4.

- 168 -

e

-w . . . _ _ 9- ~ ,

'"~'--oww,_

f

, g

L ~.w .

~..

more

' 7' ,

~

ge level to accumulato required '^~

d:cppear to-cause the dama additional damage assumed iders, and the would be less than idly than FaAA consstuds to initiate

6:crccks between - h ease f FcAA.. he cast iron determines t e This is presented as Sacond, the quality of t a given damage index. unvarying con-normally an con-sf initiation for exponent) which is However, FaAA has On" (Paris law of " n" and g iv es a a given material condition. by Otent for i the best valuewould be determined d

cidorcble trouble in find The ngThe proper.valueconser ac-valua of 5.37 to 9.62 relation to the no testing the metal of the blocks.rt have The values are "n"

estimates of in the FaAA repo101,of 102, the and 103 bloc signif-tual values for fferent EDG for each block, becausend its poor qual- '

expacted to be di be I casting procedures all three blocks .

icant variance As discussedin.thebelow, TD a l properties, ity control. determine their materia may not be which may or properly evaluated to assumptions relying upon rather than empirically Correct. nalysis purports to be of signifi-ation Third, while the FaAA ahavior, it lacks inform initiate racks first based on EDG block beWhen did the ligament c cant importance.

169 -

1

, n n-m . !

. = . - . . ._. ._ -

~.____._

. i

, (*

c s...M -

i i.: ' ^;q blocks, and what was the cumulat ve hg-

- .w lin Ccch-of the three EDG oint? When did the original '-;J l dantgo index' of each at that p .

103 block first initi- l? '"

l

.crtek between'the stud holes in the EDGl damage index acc m Gto, and!what'was the additiona t cracks in the same block

. f itions did the 'origi-I twaOn the initiation of ' the ligamen s l

cnd that point? )Rien and under what cond in the EDG 103 block grow to I

When nal-crack between the stud holes its rate of growth?

' 5-1/2 inches in depth, and what was i der no. 1 down the front did the large crack running from cyl nte did it propagate?

of EDG 103 first inititate and at what rald provide some mean The answers to these questions wou smpirical data. techniques to predict Q. Did FaAA use hfracture mechanicscracked block to .

i the rate of crack growth of't e and 102? . fracture The FaAA Block Report does not ,use a A, No. th of ligament cracks or But

' nechanics analysis to predict the growt d hole to stud hole

^ the initiation or growth of s u predict the propagation of FaAA does use fracture mechanics to and of of the blocks cracks in the camshaft gallery areas We believe l

E piston skirts. '

cracks which may initiate in the Ay in the approach FaAA ha this is a significant inconsistenc used to predict crack growth.

- 170 -

<w+,e m

- ~s

~~

^^t

_ . _ m _-a . 2 ._

Y iginal block of xlA Can the. excessive cracking in the or W, Q. ly weaker material than

- EDG 103 be attributed to significant [(

thscs of EDGs 101 and 1027 - of any actual No. There is insufficient evidence small area of A.-

FaAA examined only a t iron

'bleck material properties.But within the same block the cas GCch block top.212/ h presence of trace ele-ial properties may vary widely due to t eA meaningful analys nGnts in certain areas. ld require metallurgical ex-properties of a cylinder block wou of the block.

amination of numerous sample areas l

d block is dependent on The performance of the EDG cylin er FaAA's exami-l of construction. was l

the properties of its materia sof the block tops" of the EDGs ~

nation of a "small region ials of each of the blocks. .

/ ~

inadequate to characterize the materis homogencus, but in' act f segregation which FaAA has assumed that because theofblock the the casting is not uniform Therefore, more ing process.

naturally occurs during the castbe evaluated to determine whethe l

than a single small area must tire blocks of EDGs 101, l

or not there are differences in the en" Specific materi ;

FaAA states, )

102 and 103.

i -

g / g. at 4-4.

f

- 171 -

i l l

! i a e s,

.- ,- . ,..y-- - . , -

--+--ym__-._m_,,-_.w.,py y .,m-%__ ,my._-_.,, , ,_,, m mw,-.. ,,.-.,-----yev, 94 ,.3,, *,w-.w., ,,ys.y--..-w,.y-s-,--j,.-,y'

e vn or'frccture P~ 9.Eb7 ,

de' gradation in fatigua= 2_1_8,/ We agree. ;', .~ i i o.s 4" Oguircd to quantify anyi k section' block l of the cast ng.

or iginal f

arcperties of the-th coses that only the materia If that block mate rwevGr, FaAA prop l tely evaluated. a nt to fatigue bleck for EDG 103 be comp esufficiently less ks of EDGs resista ric1.is "shown to be iron, Class 40,"111/ the blocof surviving .

then typical gray cast d by FaAA as capable of those ials 101 and 102t would be predicteThisial. assumesThere is that the mat

.a LOOP /LOCA even . strong as " typical" materTo reach conclu at least as assumption. 02 blocks are i no.edequate basis for th s blocks of EDGs l101 cks and 1 material strength of thematerial of all three b o about the 103, the compared to that of EDG l ck be cust be properly evaluated. cracking onof ,the EDG 1 Can the ' excessive Q. l load excursion" at Shoreham .

attributed to the "abnorma outage April 14? FaAA notes that the power ds, and that FaAA did not do so.

A.

excess load for 23 secon G effected crack EDGfrom 103the with an linder down the front of ED no. 1 cy the large 218/ Id. at 4-5 2y / Id_. at 5-1

- 172 -

l

{

1 t

49 e.

e "T

-> '-m., N,, ,, ,

~

71,

. ' n. C c' nt. But FaAA refrains U*

b.

!103 cccurred after the excesseffects loadofevebetween the ; th ,

frca caking.-any causal connection ibing the ;

h 23 Naithst FaAA-nor'LILCO. documents descramount of the load I

/ disclose the d for 100 powar outage 120 DG 103 ran at test _overloa CGconds., We do'know that E rack down the blo'ck front l to deter- .I cinutes thereaf ter before' the large cavailable Two.

facts we l With the d on the block.

was noticed. r ains what,- if any, effect the.23 secondsif haFirst, they the " a observations are in order.that accidents happen, even h uld ,

Dion" demonstrates again The EDG's and their blocks s o are thought to be unlikely. h an accident, which might have Second, EDG 103 l

ba strong enough to survive suction of a LOOP unloaded

/LOCA.

. occurred during the incep 23 second episode in an That fact, coupled ran for ten minutes afterater.321/ the overload condition and without cooling wdamage resulting from the i with the subsequent block of EDG 103 may have been test, suggests that other componentsthe entire start-up f

LILCO has committed to repeat of its replacement d amaged .

fter installation engine.112/ This test program with EDG 103 a d inspect the f block,.and then disassemble an '

984, from J.A. Notaro to W.E.

_220/ Letter dated April (Exhibit 65)24,1 Steiger.

221/ Jd. at 2. lk County's Filing ConcerningContentions 222/ LILCO's Response to Suffoof Emergency Diesel Generator Litigation 21, 1984, at 55.

P

- 173 -

'- - - ~ . ,

__. \

subject , ,

The inspaction abould be ,,

_[ /

t .

amaita:nt' is very importan .i s in this proceeding. ,

o the cerutiny of all part e h t the cracks '

Do you agree with FaAA's conclusion will not grow t a to -

Q. of the blocks in tha camshaft gallery areas Ony cignificant degree? applying its formula ' for one example FaAA gave A. No. d~the assumed crack the to fatigue crack growth, which predicteIn its analysis, ac- FaAA uses 313 /

grow, but at a slow rate. which does not takeoninto fa-simple Paris empirical relation,uch as mean stress effects d the count important parameters sin addition, FaAA iron evaluate tigue crack propagation. luation based on gray cast 103 block.

parameters in the Paris eva tly present in the EDG with-without the , defects apparen ck growth are meaningless -

k size to f atigue .

The conclusions out presenting the presented on crasensitivity of i of the he limi-and the physical propert esh t our general comments on d above with life We should also point out t amechanics analysis AA predic- discusse totions of a f t:a.:ture i ts also apply to the Fa cracks.

regard to the Af piston sk r camshaf t gallery area f the tions for the growth o

_223/ FaAA Block Report at 4-6 to 4-7. i 5 - 174 -

i 1

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

y _. _. . _

3 I

inconsistencies in the :

Q. Did you also discover otherft gallery cracks? '

AA Gveluation of the camsha values to n ,

First, FaAA assigns different d x (n = 9.6)

Yes.

-A. in their cumulative damage in eSince .

l is (n = 5.37). expo-6Paric Law exponent) camshaft gallery crack ana ys cnd in the o f " n" orme material is used in both ca.Second, value the ltha confuses the results. material used failed in thetoEDG

.nsnt value specific FaAA evaluated for the chould be h uld be recalculated. t rial.

103 block and Table 4-1ting s oof specific block ma e i of "n"

obtain the value from teside the basis ery for its select on sensitive

! Further , FaAA f ailed to provThird, crack' growth d in the rate is v values.

ganaric "n" of " n." For example, if n = 9.6 is useon page 4 ,

to the value ample given 10,000. ,

gallery crack growth rate exthe rate is increased by Block Report, camshaft gallery area of the l

Have the cracks in the Q. sured for propagation?

EDG blocks been mapped and mea h se cracks and some ap-ot A. Apparently 24/

LILCO did map t eThe FaAA Block I

peared to have grown.2_ , rt, i

(Exhibit 57); Morning Repo at 97-99 4. (Exhibit 55) .

12,4/

2 Museler DepositionNRC Region 1, March 20, 198

- 175 -

i

,p,+w yg PM

~ ^ ' + - - - ..m_, "*v'w-*rm-, , - , _ we,w,,,. ----~--..w..__. __ 3;

l

~

pagation of theso 9 jport cny empirical data concerning pro frccko. of the ft gallery area Q. How could cracks inof the i the EDGs? camsha Sylindor block affect the operat on-(and there is no If the known cracks propagatethat the first will not) they.

A.-

The flexing rcported metallurgical evidenced flexing of theldcamshaft.

Gff0ct will be increase adjacent bearings, which. cou will then increase the load ates on of cracks load at these on the loca-further increase the propagation rcamshaft takes place, the As flexing of the i will be reduced.

tions.

cylinder where camsh' aft h flexing is occurr ngother cylinders ~

As there appears Consequently, the loads on t e will be lost. ,

bility to creased, and cylinder balanceof power in the EDGs, In the a unbalance. .

to be almost no reserve seriously affected by the camshaft ;

ken take full load will becracks the worst case, the of the could result in a bro leading to irreparable damage eng ine . luated by FaAA? I Q. How is the load imbalance eva from changing loads due to i

A. The interaction resultingand increased loading in tion crack propagation in one loca i

- 176 -

l l

seampe ,p,..

~~- "

2 ack growth forecasts made i

Ghar :lecations is not part of the cr -l i

fyFCAA. am . gallery support .

Q. The DRQR authors conclude that c t full load and no Grccko "are predicted to grow very slowly aWhat is th /

ct all at 75 percent load."331

' conclusion? is provided in either the No basis for the conclusion Further , the information A. k Report.

DRQR Report or the FaAA Bloc and in fact contradicts, a t support, ,

-provided by FaAA does no ' t grow "at all."

conclusion that cracks will noh t the cracked blocks on by Q. Will FaAA's recommendation t ad for EDGs 101 and 102 be examine ion ensure the safe and reliable sddy current"after 'each operat op3 ration of the EDGs7315/ cks between stud l As discussed previously, cra ent and lead to A. No.

rapidly during a LOOP /LOCA evperiodic '

i holes can initiate Inspection of'the block after l

catastrophic failure. operation in an ,

ensure reliable f

' testing does not therefore l at 3 (Exhibit 56) .

115/ DRQR Report Vol. 4, Cylinder Block, I 22/ FaAA Block Report at 5-2.

- 177 -

l

-l

~ ~ ' ' - - .- __ W'e,- 9 ~ ~ . ~ . _ , .

+

- emargcncy. Morsover, ce diccussGd above, ligamsnt cracks can

- cause leakage of coolant which itself can result in catastroph-ic failure. The propagation of the large crack down the front of EDG 103 running-from a stud hold in cylinder no. 1 (which had a ligament crack) demonstrates that unanticipated and dan-gerous crack propagation, other than of cracks between stud holes, may occur rapidly during a LOOP /LOCA event. Ligament l

cracks similar to that on the stud hole for cylinder no. 1 cur-rently exist at two stud holes for cylinder no. 8 of EDG 101 and at one stud hole for cylinder no. 8 and another for cylin-der no. 1 of EDG 102.227/

Q: Aside from the radial / vertical ligament cracks, the cracks between stud holes, and the cracks in the camshaft gal-1ery area, have other types of cracks been found to occur in the R-4 and RV-4 series TDI cylinder blocks? .

A: Yes. The FaAA Block Report refers to cracks in the blocks of TDI DSRV-16-4 engines at Comanche Peak Steam Electric Station. These cracks appear to extend down the counterbore and through the counterbore landing.228/FaAA also refers to 227/ Id. at Figures 1-2 and 1-3.

228/ Id , at 1-3.

- 178 -

I t -

"circumferential cracka in tho linor counterbore et the linor landing ledge."229/

l Q: Has FaAA determined the causes of these cracks and addressed whether they could occur in the EDG blocks at Shoreham?

A: No. FaAA states that the cracks at Comanche Peak have been " metallurgically examined and were identified as interdendritic shrinkage or porosity resulting from the cassing process.=230/ However, FaAA does not state who performed this examination, give any results in detail, or address whether

~

similar cracks might occur at Shoreham. If the conclusion .

stated by FaAA is correct -- that these cracks are due to cast-ing defects -- it supports our view that castings by TDI, including the blocks, piston skirts, and cylinder heads, are not reliable. FaAA does not discuss the circumferential block cracks at all. When questioned about the circumferential block cracks, Mr. Robert Taylor of FaAA, who headed the block study, testified that the FaAA report would not address the circumfer-ential cracks:

f 229/ Id. at 1-1.

230/ Id. at 1-3.

- 179 -

1 .

[B]ecause I am rccoiving pioccuro fec3 Ocn- :

agement and LILCO to put a report out so p that they can start a dialogue with the NRC.. It's my understanding there have been a block

- promises made to NLCA (sic -- NRC) report willigo out . And jLnI the j ustvery can' new t -- (sic it just -- ;

near) won't future. be a complete2 analycis, but it will l

start things moving._31/

e Q: Are' you, concerned abo'u't circumferential cra :ks developing in the EDG blocks?

A: Yes.

Such cracks could be very dangerous and lead to '

There is no reason to believe they will not de-EDG failure.

velop in the EDGs.,The causes of the circumferential cracks have not been determined.

Q:

Did FaAA determine the causes of the ligament cracks and stud hole to stud hole cracks in the block tops of the EDGs?

A: Not precisely. FaAA only concluded that these cracks were service-induced and identified "three possible mechanisms in of crack initiation (act'ing separately or in combination)

. low cycle f atigue . . , , high frequency i

the block top, . .

f

- f atigue .

. . , [andi-overload rupture.=232/ These same 231/ Taylor Deposition at 67. (Exhibit 59) . <

232/ FaAA Block Report at 11.

l

- 180 -

o

,..- . y i

4 i mferentiall 1 cschanisms could,couco the intitiationiof the c rcu ,;

i t F cr ccks .'

of the ,

Q:

Do you agree that the cracks in the block tops i h

EDGs were stevice-induced?: .

I i ly supports A: All of the evidence available to us certa n h t the theory. We believe these cracks are indications t a .

that They cannot operate at -

EDGs are over-rated and undersized.i f the blocks and' ~

rated and required loads without the_ crack ng o < d Dr. Chen, the diesel consultant to LILCO an 1

other components. k high firing pressure the TDI Owners Group, testified that the d recommended of the EDGs contributes'to the block cracking, an1,'500'to l that peak firing pressure be reduced to would reduce the l

l Of course, such a reduction in firing pressure t for service ;

horsepower of the EDGs to below the required amoun 6

s h

at Shoreham.

h re-What is the basis for your assertion that t e O. design and has Pl acement block for EDG 103 is of an unproven l

not been adequately tested?-

s ,

s 15, 1984) ("Chen Deposi-ill/DepositionofSimonK.Chen(May

, tion") at 129. (Exhibit 66).

4

- 181 -

~+

h I

_4.

1 A. Mr. Lowroy of TDI testifisd that the design of the replacement block was only developed in - the. last two months of 1983,-in an attempt to solve the block cracking problems of the R-4 series engines. 234/ The newly designed replacement block was never tested by TDI, according to Mr. Mathews, the general manager.235/ - Rather , TDI relied on the f act that the top por-tion and boss section of the replacement block design' was the same design as similar portions of the block of the TDI RV-5 engine, and the RV-5 block had been tested.236/ A block is a 4

single casting. We do not believe that a new design of an en-gine block is adequately tested simply because a portion of the casting is the same as a portion of an entirely differently i

designed block.

Q. Do you believe that the replacement block for EDG 103 is likely to crack?

A. Even if the design were adequate, and we believe such has not been demonstrated, the material properties used in all 234/ Lowrey Deposition at 15-16. (Exhibit 24).

235/ Mathews Deposition at 106-107. (Exhibit 32).

236/ Id. In 1981 TDI decided to use the RV-5 blocks in current production for RV-4 engines, to address the block cracking problems. See Memo dated 4/1/81 from Lowrey to Pratt (TDI). (Exhibit 67).

- 182 -

-l

of FcAA cnalyDG2 cre dSpendsnt on the esating procacs. The casting process can introduce defects such as porosity, tears, inclusions, and degenerate phases which critically effect the results'of analysis. From the results of our inspection of the TDI casting processes and review of pertinent documents relating to changes made in those processes, we are not satisfied that TDI can produce a defect-free block. Therefore, any new replacement block must be completely inspected and tested.

Q. Have you recently received documents cited in the

" Component Review",section of the DRQR Report on cylinder blocks?

A. Yes. A number of the underlying documents were re-centl-y received by the County. We have only had time to preliminarily review these documents. Many are illegible or have missing pages.

4 Q. What do you conclude based on your initial review of some of these documents?

A. Contrary to the conclusion in the DRQR Report that i

the " Owners Group has completed its review of the TDI diesel generators installed at SNPS" (p. 4-1) and that the Report

- 183 -

i

i providoo tha romulta which provida the basic for the conclusion that the EDGs " presently installed are fully capable of l reliably performing their intended safety function" (Executive Summary, p. iii), we have discovered that final resolution of a number of unsatisfactory conditions documented on LDRs had not occurred when the Report was issued. Further, our review has disclosed that objective standards were not applied to re-solve identified deficiencies. Thus, rather than documenting the completion of the DRQR assessments, the Report in fact pro-vides only a status of the ongoing investigation. Should fur-ther review reveal additional information relevant to our tes-timony, the testimony will be supplemented.

4 i

- 184 -

1 l

_ _ . _ - . _ .. .-_ ..r ..

ML20096A651

Text

Dear Administrative Judges:

Navigation menu

Search