ML20137D808

Safety Evaluation Supporting Use of Mechanical Stress Improvement Process in Primary Sys Stainless Steel Piping to Modify Residual Stress Pattern at Piping Butt Welds

ML20137D808
Person / Time
Site:	LaSalle
Issue date:	11/18/1985
From:	NRC
To:
Shared Package
ML20137D767	List: ML20137D763 ML20137D808
References
NUDOCS 8511270149
Download: ML20137D808 (7)

Text

._ - . . - . - - -. . . _ _ _ -

~, .

ENCLOSURE SAFETY EVALUATION OF THE MECHAHICAL STRESS IMPROVEMENT PROCESS Introduction By letter dsted July 19, 1985, Comonwealth Edison Company (the licensee) proposed the Mechanical Stress Improvement Process (MSIP) developed by O'Donnell & Associates as an alternative to Induction Heating Stress Improvc-ment (IHSI) that was previously committed for the LaSalle County Unit I during the first refueling outage. The purpose of these stress improvcment processes is to modify the residual stress pattern at piping butt welds. Normal welding practice induces a detrimental residual stress, putting the inside surface of ,

the weldment area in a state of residual tensile stress. This, in conjunction with service-induced stresses, causes IGSCC at the heat-affected-zone (HAZ) next to the weld fusion line. The intent of the stress improvement processes is to induce plastic strain in a manner that will leave the insice surface in the weld area in a state of residual cenpressive stress.

The beneficial effect of stress improvement effected by IHSI has been shcwn by_ analysis, residual stress measurements, and service experience in Japan, where it was developed and first applied. The NRC has reccuended the use of IHSI for BWR piping weldrents, and permits a significant reduction in augmented inspection schedules for piping so treated. The process has been applied to about one thousand welds in at least ten operating BURS.

Ongoing tests of IHSI treated welds in pipe test facilities at GE and Phl are sponsored by EPRI and the BWROG. It is expected that the results of these tests, in conjunction with actual field service will provide further verification on the amount of benefit in reducing the potential for inter-granularstresscorrosioncracking(IGSCC)overthelongterm.

l Description of the Process The HSIP process has only recently been developed as an alternative stress l improvement process. Instead of using a large tr.rporature gradient through the wall of the pipe to achieve the desired plastic strain pattern,F. SIP 0511270149 851110 PDH ADOCK 05000373 U PDR

f uses mechanical methods stated to be less expensive and time-consuming, and involves less radiation exposure to the technicians performing the operation.

t In the MSIP process, the pipe is plastically " squeezed" or contracted at a location about two inches to one side of the weld being treated. The  ;

force is provided hydraulically, working through split rings with flexible metallic pads between the rings and the pipe. A permanent reduction in diameter of about one to two percent is achieved by this process. After the equipment is renoved, elastic springback results l in residual tensile stresses in the squeezed area, balanced by compressive l stresses in the weld and HAZ area at the inside surface of the pipe.

O'Donnell&Associateshasperformedfiniteelementanalysestopredibthe resultant residual stress patterns produced in various weld configurations.

Some surface residual stresses were also experimentally evaluated, and confirmed the analytical results. The results of these evaluations show that the desired compressive residual stresses are developed in the weld and HAZ l area.

t Such analyses were also performed to determine the effects of general j and local out-of-roundness and mismatch conditions on the end result.

These showed that the resulting beneficial residual stresses were essentially not affected by these conditions.

The process itself is closely controlled by physical measurements before and after application. The amount of defomation desired is calculated, and mechanical spaces are used between the two split rings to positively  ;

control the amount of diametral reduction to that desired.

Magnesium Chloride Test

' A test that has been regularly applied to pipe samples to detect detrimental residual stresses induced by welding, and to evaluate process variables intended to reduce such detrimental stress is a boiling magnesium chloride test. This is routinely performed at the EPRI NDE Center (managed by J. A. Jones Applied Research Company).

A utility planning to use MSIP on replacement piping (Vermont Yankee) sponsored a standard magnesium chloride test on a piping sample containing two welds. One of the welds was intentionally given the HSIP treatment.

The adjacent weld was close enough that it also was subjected to some stress improvement. Although there originally was some confusion regarding details of the test results, these were clarified in a September 27, 1985 letter from the J. Jones Company, who perfonned the test. The test was run in confonnance with ASTM G36-73 (reapproved 1979), except that the exposure time was 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> instead of the 4 to 6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br /> usually considered adequate.

Neither the MSIP treated weld or the adjacent pipe butt weld cracked in this test, although the weld applied to hold the botten plate to the pipe test section did show cracking, both in the pipe and in the plate material (this bottom plate is put on to make the pipe test secticn serve as a container for the magnesium chloride). According to J. A. Jones, this cracking verified that the test was a valid one, even though the butt weld not intentionally processed by MSIP did not crack. A review of the actual sample geometry shows that the one-sided flSIP treatment was actually performed within two inches of the cptimum location for MSIP of the weld not intented to be treated. Therefore, it can be assumed that it was at least partially treated, and this partial treatment was sufficient to prevent cracking in the test.

Effect of Varying Strength Across the Weld Area Soce concern has been expressed regarding the efficacy of the process censidering the variation of strength levels of the base material and the weld area. It should be pointed out that the residual stress on the HAZ depends primarily on the plastic strain in the compressed area outside the weld. The yield strength in this area of the criginal pipe will be at least 30,000 psi, and after: compression to 2 22, plastic strain would be expected to be in the range of 40-50,000 psi - the same range as the weld and cold worked region adjacent to the weld would have before treatment. Because the weakest material will yield first, and the deforma-tion is compressive (no reduction in area during plastic flow), the differing yield strengths will tend to become more uniform. That is, the weakest areas will defonn most, and therefore undergo the most strain hardening.

The resulting residual elastic stress pattern will therefore be dependent primarily on the strain hardened elastic properties of the originally lower strength base material, and not the original properties.

Possible Deleterious Effect of Cold Work It has been well known that severe cold work, such as that produced by abusive grinding, will enhance crack initiation in 8WR primary coolant.

General Electric has recently submitted proprietary data regarding the effect of lesser amounts of cold work on the susceptibility of austenitic stainless steels to IGSCC in the absence of classical sensitization.

Although these data cannot be disclosed, the results are not inconsistent with other test results and service experience that are not proprietary.

t .  ;

(

i 1

f 1 I

t The Standard Review Plan for Control Rod Drive Structural itaterials, I

4.5.1, permits the use of cold worked stainless steel with a maximum i of 90,000 psi yield strength. This usually corresponds to something j in the range of 10 to 12% plastic strain. This criterion was based on ,

early published infonnation and unpublished work. Although we might now question the resistance of 90,000 yield strength material under high ,

! load controlled stress in BWR primary coolant, the information available  !

i does not indicate any expected problem with material with 5 to 8% cold work j at about 65 to 80 ksi yield strength.

The MSIP process has been stated to produce up to 2% plastic strain in the base material of the pipe at the location where results in residual  !

I tensile stress. The process is well controlled, so it is unlikely that l j the amount of plastic strain would exceed twice that amount. There are j no data that we know of that would indicate that even 4% plastically strained

! material would be subject to IGSCC initiation in BWR primary coolant.

t i

Staff Evaluation -

i i

l Although the basic concept of MSIP appears theoretically sound, and the

! process is inherently controllable, there has not been sufficient time j to develop confirmatory data to provide the same degree of confidence

! as is generally felt in the IHSI. The analytical work performed by l O'Donnell & Associates is ccmparable to that performed by vendors of l the'IHS1 process, but no independent analysis have yet been performed by j the staff.

i The successful service experience obtained in Japan on !HSI treated welds l l (4000hoursuntilpipeswerereplaced)providedadditionalevidencethatstress improvement, and IHS! in particular, reduced the probability of IGSCC.

The ongoing EPRI pipe tests on litSI treated weldments is expected to i

i

j , .

4 4 l 1

, provide even more quantifiable data, because stress levels are known and f controlled. These are expected to confirm the staff's conclusions regarding the degree of recuction in IGSCC severity and probability that l

i IHSI can be relied upon to produce. .

[

i  !

, There are three main actions that should be taken to further confirm that }

l NSIP is also a viable stress improvement process, and to provide a quantifiable assessment: l j (1) Pipe samples containing MSIP treated weldcents should be included in the

ongoing EPRI test program. l I

I l (2) independent clasto-plastic analyses should be perfonred to confirm those ;

j already done by O'Donnell & Associates; and [

l (3) through-the-wall experimental residual stress analyses should l be performed to confim the theoretical analysis results.  !

)

r

} L 4

Conclusions  ;

I It is the staff's conclusion, on the basis of information now available, that j the MSIP process will be proven to be effective. It appears likely that it f can be as effective as IHS! proved to be. The staff's interim positions on (

its use are as follows:  !

(1) MSIP appears to have no deleterious effects, so can be used on replacement  !

piping to further reduce the probability of IGSCC. [

i (2) IISIP can be used to reduce the extent and frequency of augmented inspec-I tions of piping susceptible to IGSCC. This reduction is expected to be i comparable to that afforded IHS! treated wolds if and when action items  :

listed above are satisfactorily completed.  :

l >

(3) The long range benefits of MSIP can only be proven out by long term pipe ,

tests and actual service experience. All types of stress improvement will be continuously evaluated as more test data and service experience are accumulated. As a result, augmented inspections to confirm the performance of MSIP treated welds are recomended.

1 i