ML19308A222

Addl Interim Report of Significant Const Deficiency 9, Noncompliance to Welding Requirements for HVAC Seismic Support Hanger to Embedded Plates. in Addition to Immediate Corrective Actions,Hanger to Plate Welds Are Being Repaired

ML19308A222
Person / Time
Site:	Waterford
Issue date:	12/29/1978
From:	LOUISIANA POWER & LIGHT CO.
To:
Shared Package
ML19308A221	List: ML19308A220 ML19308A222
References
NUDOCS 7901040090
Download: ML19308A222 (17)

Text

A l

• e LOUISIANA POWER 6 LICitT COMPA'iY

n. WATERFORD SES - UNIT NO 3 ADDITIONAL INTERIM REPORT OF SIGNIFICANT CONSTRUCTION DEFICIENCY NO 9 KONCOMPLIANCE TO WELDING REQUIREMENTS FOR llVAC SEISMIC SUPPORT IIANGER TO EMEEDDED PLATES DECEMBER 27, 1978 790104co9o 4

s I .

ADDITIONAI. INTL: RIM RI: PORT OF SIGNIFICANT CONSTRUCTION DEFICIENCY NO 9 NONCO.'tPl.TANCE TO WELD 1NG RP.QU1REMENTS 1OR llVAC SI:ISMIC SUPPORT llANGER TO EMBEDDED Pl.ATES BACKCR0_liND_

As a result of a review of the American Welding Society - Structural Welding Code, it has beeil determined that 425 IIVAC seismic suppor' hangers at the Waterf ord 3 Steam Electric Station have been welded to embedded plates with-out applying the required preheat to the plates. The velding was performed by The Waldinger Corporation at various elevations in the Reactor Auxiliary and Fuel llandling linildings utilizing E-6011 series electrodes. As identi-fled in the Structural Welding Code - AWS Dl .1-75, metal:. with a thickness over 3/4" to 1-1/2" must be preheated to 150 degrees F when an E-60XX series electrode is used.

RESUI.TS OF INVESTIGATIO_NS

1. The single 1-1/2 inch thick embedded plate involved in this deficiency was selected as the apparent " worst case" for hydrogen embrittlement cracking. The hanger to embedded plate attachment welds were Investi-gated by magnetic particle tosting and liquid penetrant t.e s t in g . This testing was repeated as the weld was incrementally removed by grinding, including the heat affected zone. No indication os cracking was found.

2. A detailed weld evaluation was clso undertaken. This evaluation, des-cribed fully in the attached report utilized samples from plates avail-abic at Waterford 3 in addition to one specially procured plate repre-senting a range of carbon equivalencies.

This investigation concluded that the acceptability of the welds could not be demonstrated.

CORRECTIVE ACTION In addition to the immediate corrective action taken and reported in the Interim Report deted September 29, 1978, all of the hanger to embedded plate welds are now being repaired by:

a) Providing additional structural or mechanical attachments to carry the design loads of the af fected welds.

b) Removal of the unacceptable weld and replacement using low hydrogen elec-trodes.

This in described nore fully in the attached report.

A final report is now scheduled for December 1, 1979 at which time most, if not all, of this work will be completed.

4 MATERIALS APPL.ICATIO'J REPORT EVALUATION OF NONCOMPLI ANCE TO UP.I. DING REQUIREMENTS FOR llVAC SEISMIC SUPPORT llN;GER TO EMBEDDED PLATE UELD's i s.

APPROVED BY:

PREPARED BY:

Original signed by Original signed by J Brodsky L A Gunther J lirodsky L A Gunther

/

Materials Applica'tions Report

Subject:

LOUISIANA l'0WF.R AND LICllT COMPANY WATERIORD UNIT 3 EVALUATION OF NONCOMPLIANCE TO WELDISC REOUIREMESTS 10R llVAC SE1SMIC SUPPORT IIANGER TO EMBEDDED PLATE WELDS

_C_ONCI USIONS_

lt is the opinion of the Ebasco Materials Applications Department, an investi-based upon evaluation of information received from the site, gation of the literature and metallurgical testing performed on test weld-ments, that:

1. The velds made by the Waldinger Corporation to attach seismic support hangers to ASTM A-36 embedded plates 1 inch to lh inch in thickness using E 60 11 electrodes without a minimum 150 F preheat are i;naccept-able because of the possibility of cracks existing in ihe embedded steel under the fillet velds.

2. Since no method of nondestructive testing is available to examine the heat affected zone of the embedded plate base material under the fil-let welds for the presence of underhead cracks each of the seismic support han>;cr at tachments should be repaired.

3. The seismic support hanger attachments may be repaired by providing additional structural or mechanical attachments which will be designed or to carry the loads which the ouspect welds were designed to carry by removal of the suspect weld, verification of the integrity of the excavated base material, and rewelding using low hydrogen electrodes.

INTRODUCTION It was determined on June 16, 1978, that 425 llVAC duct seismic support hangers at the Katerford 1 Steam Elcetric Station Site had been welded to t

• -7 embedded plates without applying preheat to the plates as required by the American Welding Society DI.1-75 Structural Welding Code for the base nate-rials and weld filler metals involved. The 425 seismic support s and approxi-mately 1300 embedded plates involved are located at various elevations in the Reactor Auxiliary Building and the Fuel llandling Building.

The seismic supports and embedded plctes are fabricated from ASTM A36 material. The enbedded plate r'aterial is 1 inch to 1!$ Inch thick. The weld-t.he llVAC Contractor, using ing was performed by the Waldinger Corporation, E-6011 Class electrodes as identified in their Welding Procedure Specifica-tion No PQW-1. In accordance with the AUS Dl.1-75 Code, welds made to AST.i A36 metal with a base material thickness of over 3/4 inch to 115 inch using the shielded metal arc process and other than low hydrogen electrodes must be preheated to a minimum of 150 F prior to welding. This requirement applies to welds made using E6011 electrodes which have a cellulose type The fail-coating and are desigred to contain between 2% and 4% moisture.

ure to preheat the embedded plates was a violation of both the AUS D1.1-75 Structural Welding Code and the Waldinger Corporation's Weld Procedure Spe-cification No PQW-1.

The welding deficiency was originally reported in Nonconformance Re-port No W3-957.

The deficiency was classified as reportable because of the extensive.

engineering evaluation required to determine the significance of the de-ficiency under 10CFR50.55(e). The deficiency was reported to the U S 9 Nuclear Regulatory Commission as Interim Significant Deficiency Report No

9. Further welding using the deficient welding method was stopped and the use of 7018 low hydrogen electrodes in accordance with a revised welding procedure was implemented on hanger to enhedded plate weldments.

SURVEY OF hTTI:RATtfRE Hl: LATED TO !!ONCONFOR:!ANCE The existence of the underbead cracking problem in carbon manganese steels and the causes for the occurrence of the problem are well documented in the metallurgical and welding literature. The three elements required to cause underbead cracking are hydrogen, restraint-induced st.resses and the presence of mhrtensite in the heat-affected zone of the base material under the weld deposit. The primary cause of underbead cracking Is moisture.

Moisture causes atomic hydrogen to form in the welding arc and this hydrogen is absorbed by the weld and base metals during welding. The moisture con-tent for which E6011 electrodes are designed to operate, i.e., 2% to 4%, can produce sufficient hydrogen to cause a cracking problen. The restraint-induced stresses are present on the HVAC h:inger to embedded plat e welds due to the fillet weld. The third rclevant factor, the presence of martensite in the heat affected zone, can occur because of the cheuical composi. tion of the embedded plate material, and the material thicknesses which are suf fi-cient to induce high enough cooling rates to promote the formation of martensi. .

The hardness of a weld heat affected zone is usually considered as a good indication of the presence of martensite and thus potential cracking.

An extensive program of research on the weldability of carbon manganese steels was conducted by the Battelle Memorial Institute. This study in-cluded an investigation of underbead hardness as a method of evaluating the weldability of these steels. The generally accept.ed industry criteria for underbead hardness which evolved from this work and earlier work was "350 Vickers maximum to insure absence of cracking, and 250 Vickers maximum for satisfactory performance of the heat-affected zone under service loading, particularly instructures which are to serve in the as-welded condition".

A Vickers hardnest, number of 350 corresponds to a Rockwell hardness number of 35.5 on the C Scale.

. -4 Another aspect of the P,attelle research program was an evaluation of the effect of preheat on underbead crack sensitivity using bead on plate upecimens which were sectioned and examined for underbead cracking.

The total length of the cracks in any one specimen was expressed as a per-centage of weld bead length, preheat temperaturer. of room temperature, 150 F and 225 F vpre used on several heats of carbon-manganese steels having varying chemical compositions. Comparisons were made of the aver-age underhead crack sensitivity, expressed in percent, with the chemical composition as defined by the carbon equivalent expresulon C + h + 7h .

The results showed in all cases, that there was a definiu relation be-For commercially tween carbon equivalent and underbead cracking sensitivity.

room ten-rolled 1 inch thick steel plate welded with E6010 electrode at perature cracks were found for carbon equivalent values as low as 0.38 and values crack sensitivity increased dramatically for carbon equivalent above 0.42. When preheat was increased to 150"F cracking did not appear in material of carbon equivalents below 0.44. The average amount of crack-ing in carbon equivalents above 0.44 was also decreased. At 225 F the crack-Ing tendency was further decreased and cracking did not appaar in material below carbon equivalents of 0.57. The preheat requirements specified for A36 steel in the AWS Dl.1-75 Structural Welding Code may have been estab-lished based upon the results of the work performed by Battelle.

EBASCO TESTING pROGR/d!S The recommended disposition initially proposed by Ebasco Engineering Site Support personnel for Nonconformance Report W3-957 called for magnetic Because particle examination of a sample selected from the suspect welds.

of the underbead nature of the potential cracking problem the Materiale, Applications Departuent did not belleve that this nondestructive test

method would be sufficiently searching to warrant acceptance of the results by the U S Nucicar Regulatory Commission. For this reason a program was initiated to conduct crack sensitivity tests which would examine the poten-tial cracking problem for materials with chemistries and thicknesses simi-lar to the steel used for the embedded plates.

An evaluation of the certified chemical reports for the embedded plates involved in the nonconformance indicated that the compositions varied from carbon equivalents of 0.37 to approximately 0.57 and the plate thicknesses varied f rom 1 inch to l's irch. A piece of A36 plate material 1-15/16 inch thick was procured from the Lukens Steel Corporation for the initial test work. The composition of this heat of material is given in Table 1. The Carbon Equivalent of this material based on the C + f" + SI expression is 0.607. This material was selected for testing since it conservatively re-presented the upper levels of the chemistry and material thickness range of the embedded plates. The plate material was flame cut to yield 2" x 3" test pieces as shown in Figure 1. The 3 inch dimension of the plate was oriented along the rolling dir ection of the plate. This Loagitudinal-Bead-Weld Underbead-Cracking Test Samole was deveJ oped by Battelle Memorial In-stitute and is a widely accepted method for testing the sensitivity of steels to underbead cracking. The E6011 c1cetrode was provided by the Waterford 3 Site from material used for site installation. The welding parameters duplicated those used by the Waldinger Corporation at the site and are presented in Table 2. The tests were run on 1-15/16 inch thick material at 78 F and 50 F initial temperature and on material reduced to 3/4 inch thickness at 78 F. The reduction in thickness was proposed to evaluate the effect of thickness on the veld cooling rate and the crack sensitivity of the weld heat affected zone. Only 1/8 inch diameter elec-trodes were used on these tests. The resultu are presented in Table 3.

Followingthecompletionoftheakiovetest it was decided that material whose chemistry was representative of lower carbon equivalents should be tested. It was thought that these chemistries would provide a more complete the site. A representation of the range of embedded plate material used at search was made at the <:ite to find material corresponding to 0.40, 0.45, 0.50 and 0.55 carbon equivalent chemistries. The site provided 1 inch thick t i.

material from three (3) heats covering the chemistry range of 0.414 to 0.475 carbon equivalent as presented in Table 1. In order to nake the test more the <;1te on embedded representative of the restraint conditions experienced at run on plates, the longitudinal bead weld underhead cracking tests were larger sampics. Samples 4" x 5" x 1" were flame cut by the site from each heat of material. The 6 inch dimension of the plate was oriented along the rolling direction of the plate as shown in rigure'2. The tests were run at 65 F initial temperature.

In order to provide data for the tuo E6011 electrode diameters and parameters used by the Waldinger Corporat ion on the fillet velds, tests were run on 1/8 inch and 5/32 inch diameter electrodes.

The parameters are presented in Table 2. The samples were sectioned longi-The rr ' ' ; are presented in tudinally and examined for underbead cracking.

Table 4.

In order to provide additional information on the condition of the base material heat af f ected zone under the weld deposit , transverse cross sec-The samples tions were removed for hardness testing as shown in Figure 3.

were run using 1/8 inch and 5/32 inch diameter E6011 electrodes and a 65 F initial welding temperature.

The cross sections were prepared metallographi- .

cally and microhardness survey was performed on the weld, heat affected zone and base metal using a Knoop llardness Indentor and a 500 gram load. Since the maximum underbead hardness of the heat affceted none is the data of con-cern to t his evaluat ion these results are presented in Table 5. The results are listed as Knoop liardness Numbers and their equivalent Re liardness Values.

Magnetic particle and liquid penetrant tests were run at the site on a fillet veld on a #P-53 embedded plate which is 1 inches thick. The paint was removed from the weld and adjacent areas and the wald surface was examined by magnetic particle (!Tf) and liquid penetrant (LP) test methods.

Additional MT and LP examinations were performed after the weld had been ground 1/3 and 2/3 of its original size and flush with the base material.

il An attempt was made to examine the weld heat affected zone by grinding 1/16 The re-inch into the base material and performing MT and LP examinations.

sults of these tests are reported on Peabody Test.ing Certifled Reports of Only minor Nondestructive Examination Numbers PBT-MT-285 and PBT-PT-633.

undercut and rounded indications were reported.

Discussion of Test Program Results The Longi t udinal-Bead-Weld Underbead-Cracking Test results exhibited unc'erhead cracking in samples of t.he 1-15/16 inch thick 0.607 carbon equi-valent material, liowever, no cracking was observed in camples of this material at 3/4 inch thickness or in the other sample .taterials of lower carbon equivalent. chemistry tested at 1 inch thichness. The results show less cracking sensitivity for steels havine lower carbon equivalents, but do not agree with the results of the more extensive testing program con-In addition the test is limited, as ducted by Battelle Memorial Institute.

are all available underbead cracking tests, in their ability to duplicate all conditisas experienced at the site. For example, the Longitudinal-Bead-Weld Underbead-Cracking Test does not provide any information on the effects of restraints other than that developed in a bead-on-plate on the*

cracking tendency. Therefore, the results of these tests are not considered as conclusive evidence that underbead cracking is not present in the em-hedded plate materials.

g_

The underhead crack: observed and those der,cribed in the literature are nornally concentrated in the hardest region of the heat affected zone adjacent to the weld fusion line. This ret; ion may be less than 1/32 inch wide depending upon the welding energy input and the cooling rate experi-enced by the material. There is a high probability that attempts to iso-late this area by grinding through the fillet weld into the base material f l.

prior to magnetic particle or liquid penetrant testing would miss the most critical area. Therefore, the resul ts of the exploratory grinding and non-destructive testing are not considered as conclusive proof that underhead cracks do not exist in the embedded plates.

The results of the underbead hardness tests indicate that weld heads deposited using 1/8" and 5/32" diameter electrodes on 1 inch thick material even at the lever carbon equivalent chemistry level are capabic of producing hardness levels above the Rockwell C 35.5 value in the heat affected zone adjacent to the weld fusion line. The values observed on the test ;amples indicate that a hardened area containing martensite in formed. The high hardness values observed are considered unacceptable for the restrained fil-let welds nade using E6011 electrodes which can provide adequate hydrogen to cause underbead cracking. These hardness values would be very unfavorabic for any structure welded using these electrodes and left in service in the as-welded condition.

Sensitivity and Table 1 - Corposi t ion of A16 Plat e Material lised for Crat i liardnes . Tes ts MATEl!I Al, PROVIDED liY WATERl'ORD 'l SITE

!! EAT DESIGNATIO'l PLATE Pl!OCl!nED FROM 1.UK!;S STEEh

_ 45204 491W1221 56927 11.

0.18 0.19 0.22 O.26 C

0.90 0.97 0.98 Mn 1.15 0.011 0.005 0.013 P 0.009 0.027 0.071 0.029 S 0.028 0.036 0.010 0.041 S1 0.24

"" 0.414 0.4 35 0.475 C+ -+ 0.607 4 4 Table 2 - Teyt Welding Parameters Emp, loved With E6011 Electrodes Measured Approxinate Measiired Ayy_,roximig Voltnyjn Travel Speed Enerqv_ inytg Dianeter Current Volts Inches / Minute Kilo .loules/ Inch Ann 25-27 4.5-5.3 27-40 1/8"95-110 25-27 4.5-5.3 31-50 5/32" 110-140

T on Table 3 -_ Resul t s of I,ongittvlinal-Cead-Weld Underbead CrackinLyst:.

pgterial Procured From I,ul: ens Steel Preheat Temperature Results No. of Sanples Teste_d Sample Size All samples 2"x3"xl 15/16" 78 F 5

exhibited underhead cracks.

50 F 3 of 5 5 2"x3"xl 15/16" sampics exhibited underbead cracking.

78"F No cracks 5 2"x3"x3/4" (*)

observed.

testing.

(*) Plate material was reduced to 3/4 inch thicknens before Table 4 - Results of 4x6"xl" Lonr;itudinal-Bead-Weld Underbead Cracking Teuts on } tat erlal Obtained f rom Waterford 3 Site No. of Samples Results_

lle a t_ Carbon Electrodes Equivalent Deameter Tested No.

No Cracks 45204 0.414 1/8" 1 45204 0.414 5/32" 1 491W1221 0.435 1/8" 1 491W1221 0.435 5/32" 1 56927 0.475 1/8" 1 56927 0.475 5/32" 1

• Initial Welding Temperature 65 F

Table 5 - Maxtruim lleat Af fected Zone (IIAZ) liardrn";s Electrode Saraple Maximum llAZ liardness Diarac te r Size Kiln KC llent No. Carbon Equivalent 2"x 3"x l" 378 38 45204* 0.414 1/8" 31 2"x 3"x 1" 321

" " 5/32" 2"x3" x l" 395 39 491W1221* 0.435 1/8" 42.5

" n. " 5/32" 2"x3"xl" 432 2"x3"xt" 382 42 56927* 0.475 1/8" 42 2"x3"x1" 423

" 5/32" Lukens 15" 46 2"x3"x1g 490 Material *** 0.607 1/8" 2"x3" x 3 / 4" 510 48 1/ 8"

• Initial. welding temperature 65"r 0

• Initial welding temperature 78 p

t REFERENCES

1. Metals llandbook Volume 6, " Welding and Brazing", American Society for Metals, Metals Park, Ohio, 1971, page 7.

p l.

Voldrich, C.B. and liarder, 0.E.  : " Review on the Weldabillty of Carbon-2.

Manganese Steels", The Welding Journal, Vol. 28 No. 7, Research Supplement, July 1949, page 326s-336s.

and Voldrich, C.B.  : "The

3. Williams, R.D., Roach, D.B., Martin, D.C.,

7, Weldability of Carbon-Managnese Steels", The Welding Journal, Vol, 28, No.

Research Supplemert, July 1949, pages 311s-325s.

4. Stout, R.D., Daty,W.D., Weldability of Steels, ed. Epstein, Sand Somers, R.E., 2nd Edition, New York: Welding Research Council, 1971, pages 253-254.

e

.

~/g /i WEAP l-

< OEAD

- 1 ,,

3,,  : .

i, aLJ

- 43 k

.) l1!'o? lb; e Q k

,,. ')

C4t N 1

y- x .

O I, ..

-+ a- A PCiL

./6 -

Y Y

mw A

(UTZ 77ticmtgSs

'I

~

"W A >

.S.p_ec imen Prenarat ion : _

l. Ueld Specimen

2. Age for 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> at room temperature to allow sufficient time for cracking to occur.

3. Saw cut specimen to the right of weld center line and grind to wcld center line.

4. Examine using 20x magnification for cracking.

Figure 1 - Longitudinal llead - Weld Cracking Specimen Developed at liat te lle steraorial Institute.

n

l.

._p, /A

.a i

6.,

, W420 0SID li 1' r

/

f' $.

k Q. ',

i.., ,,n

MP. IG).

i

> k I

D Y {!

u m  %~()['[', 3l, th

~

lo S cd a

I te Y _

\l

,u

-g- yu

<j,'. -

0'

+.

y_

i Figure 2 - Modified Dimensions of Longitudin:i1 !>ad-1l eld Cracking Snecimen u

.