ML20141N887
| ML20141N887 | |
| Person / Time | |
|---|---|
| Site: | Grand Gulf  |
| Issue date: | 02/13/1986 |
| From: | Feil J, Tommey M MISSISSIPPI POWER & LIGHT CO. |
| To: | |
| Shared Package | |
| ML20141N885 | List: |
| References | |
| 85-M-017, 85-M-017-01, 85-M-17, 85-M-17-1, TAC-60165, NUDOCS 8603180246 | |
| Download: ML20141N887 (19) | |
Text
--.
4%
- e METALLURGICAL EVALUATION of RECIRCULATION PIPING CENTER CROSS SURFACE CRACKING FOLLOWING IHSI GRAND GULF NUCLEAR STATION - UNIT 1 FINAL REPORT Mechanical Section Nuclear Plant Engineering Report No.
- 85-M-017 Prepared By: ? :.t c ) / n71t.t u s4f / / MF6 pf
' Responsible Engineer Date' Reviewed By: ren // / ( 7 7 , 6 a S ' 9 e/ // 'v!4 S Engineering Supervisor Date Reviewed By: / n Principal Engineer Date Approved By: a / /
Director of Nuclear Plant Engineering Date D 6 e
( %
NOTICE This report supplements and is to be used in conjunction with Mechanical Section Interim Report No. 85-M-017, dated November 21, 1985 4
L .
I:
TiBLE OF CONTENTS PACE
1.0 BACKGROUND
............................................................. 1 2.0 SAMPLE MATERIAL........................................................ 3 2.1 CHEMIC AL COMPOSITION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 2.2 MICROSTRUCTURE.................................................... 4 3.0 METALLURGICAL EVALUATION............................................... 5 l 4.0 DEPOSIT ANALYSIS....................................................... 6 5.0 TENSILE TESTS.......................................................... 7
! 5.1 SPECIMEN DESCRIPTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 5.2 TEST DESCRIPTION AND RESULTS...................................... 8 5.3
SUMMARY
.......................................................... 10
' 6.0
SUMMARY
AND CONCLUSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1 l ' 7.0 DETERMINATION OF GENERIC APPLICABILITY AT GGNS................. .... 14
8.0 REFERENCES
............................................................ 16 l
t
] '.
PAGE 1 0F 16 1.0 RACKGROUND ,
' Induction Heating Stress Improvement (IHSI) was performed on selected welds attaching piping and fittings in the 304 stainless steel reactor recirculation system of the Grand Gulf Nuclear Station, Unit 1. The IHSI procedure was performed by Nutech Engineers, Inc., in accordance with the Electric Power Research Institute (EPRI) guidelines with all process parameters within the EPRI perscribed range. The piping connection welds of the twenty-four (24) by sixteen (16) inch cross which connects the twenty-four (24) inch recirculation pump discharw s piping to the sixteen (16) inch manifold header and the cap attached to the center cross were treated as part of the IHSI program. The induction heating procedure calls for liquid penetrant (LP) examination of the monitoring thermocouple locations following the IHSI treatment. During the LP examination several small irregular indications were noted in a region approximately six (6) inches away from where the heating coil had been locetad. Further surface preparation revealed indications over large areas of the outer surface of the twenty four (24) inch regions of the cross. The sixteen (16) inch arms, the cross to cap weld, twenty four inch pump discharge-to-cross weld, and cross-to-header welds did not reveal LP indications. The recirculation system crosses are SA-403 WP304 with a base material to SA-182 and fabricated by the Taylor Forge Division of Gulf and Western Manufacturing Company.
In order to determine the cause for the indications, a metallurgical evaluation was undertaken. The samples for the evaluation were obtained from the two unused SA403 WP304 crosses from GGNS-Unic 2 that had been fabricated to the same design and from the same material as the GG.:S-Unit I crosses.
One of the unused GGNS-Unit 2 crosses had been used to develop the IHSI
PAGE 2 0F 16 technique and coil for the IsSI treatment of the cross cap (X-cap) (the 24" pipe cap welded to the upper 24" inlet of the cross) . This cross was designated as unused cross No. I and was examined using the same LP technique and procedure. The examination revealed indications similar to those in the GGNS-Unit I crosses. Samples for metallurgical evaluation were taken from an area with LP indications and a non-IHSI affected area (ie., no LP indications). The second cross was designated as unused cross No. 2 and was LP examined. No surface indications were detected. Metallurgical samples were taken from the sixteen (16) inch arm of the cross and the twenty four (24) inch region corresponding to locations sampled in the unused cross No.
- 1. Locations'of metallurgical sampling for both crosses are shown in Figure 2 of the Mechanical Section Interim Report No. 85-M-017 (attachment to letter AECM-85/0372 dated November 20, 1985).
The original billet material for the CGNS Unit I recirculation crosses came from the same material heat produced by Colt Crucible and shipped to the Cicero Ill. plant of Taylor Forge. At that plant the billet was forged into four (4) seamless tubes at a specified working temperature of 2220 to 2250*F. The final dimension of the cylindrical preforms following machining of the inner surface was 26 inch 0.D. x 3.0 inch wall thickness x 38 inchen long. This forged preform was shipped to Taylor Forge, Woodstock, Tennessee division for hot working to the final 24"x24"x16"x16" cross. The final fabrication involved a plug pulling technique at 1900*F maximum in a series of up to 10 steps to form the 16 inch 0.D. nozzles. The final inspeccion involved liquid penetrant (LP) testing at the Taylor Forge Woodstock shop prior to shipment to CGNS. Two of the crosses were installed on CGNS-Unit 1 and the two remaining crosses were retained in storage after the CGNS-Unit 2 design was changed to use 316K stainless steel piping material.
PAGE 3 0F 16 2.0 SAMPLE MATERIAL The sample materials for this metallurgical evaluation were obtained from two center cross reducers salvaged from GGNS Unit 2 that had been fabricated to the same design and from the same material heat as the Unit I center cross reducers.
The sample materials were divided into two groups. Group 1 specimens were sectioned from the unused cross No. I which had been subjected to IHSI treatment during the programmatic adaptation of IHSI to the sensitized Unit 1
- center cross caps. Group 2 materials were sectioned from the unused cross No. 2 and have not been subject to the IHSI treatment. All sample materials had been solution annealed during fabrication and all specimens were sectioned from their respective center crosses to provide both I.D and 0.D.
surfaces for examinations.
Material Characterization 2.1 Chemical Composition Specimens from each group of samples were analyzed to determine the bulk i
- chemical compositions. The results obtained are tabulated below.
Chemical Composition C Mn Si Cr Ni Mo P S Sample A 0.06 1.61 0.35 18.29 9.22 0.21 0.027 0.007 (Group 1)
Sample G 0.06 1.60 0.35 18.30 8.99 0.23 0.019 0.006 (Group 2)
SA 403 0.08 2.00 1.00 18.0- 8.00 -- 0.045 0.030 WP304 max. max. max. 20.00 11.00 max. max.
l
! The compositions for both specimens conform to the requirements for SA j 403 WP304 material in every respect.
l l
k
PAGE 4 0F 16 2.2 Microstructure Representative specimens were taken from the 24-inch leg of unused cross No. 1 (IHSI treated) and from both the 16-inch and 24-inch legs of unused cross No. 2 (no IHSI) and examined metallographically to characterize the microstructure of the cross material. All of the specimens from the 24-inch legs of both crosses exhibited a very large grain size. In each case, the grain size was generally uniform through the wall thickness.
The grain size for the 24-inch leg was generally too large to be assigned an ASTM micro-grain size number according to the relationships given in ASTM G-112. For Sample C. (from unused cross No.1), grain sizes were in the range ASTM M10 to M9 (macro-grain size number), which means that the average grain "dianeter" was between 1.1 and 1.6 mm. In a few locations, grains with an average size of ASTM 00 (micro-grain size number) were observed. These grains have an average "dianeter" of 0.5 mm. The grain size measured for Sample C (from unused cross No. 2) was very similar to that of Sample C, although a greater spread in size was observed. Sample G had a grain size in the range ASTM M11.5 to M8.5, which means that the average grain diameter was between 0.65 and 1.9 mm. In some locations the grains had a grain size of ASTM 0 (micro-grain size number), that is, an average grain diameter of 0.36 mm. High magnification examinations of these sectiors did not reveal any sensitization.
Examinarfon of the microstructures observed in the 16-inch leg of the samplas from unused cross No. 2 showed that in general, the grain size for these samples was uniform through the wall thickness and was measured as ASTM 3.0-3.5. This grain size is very much smaller than
I PAGE 5 0F 16 that of the 24-inch legs.- Examination of the metallographic sections estabitshed that the material was not sensitized.
3.0 METALLURGICAL EVALUATION Meta 11ographic examinations were performed on specimens from both groups of sample material (Unused cross No. I and cross No. 2). In the case of unused cross No. 1, the objectives of these examinations were to characterize the microstructure of the material, establish the nature of the defects observed in previous visual and LP inspections of the outside surface, and identify the microstructural features of observed defects. The objectives of the examinations of the samples from unused cross No. 2 were to compare the microstructure with that of unused cross No. 1, determine whether any inherent cracks or flaws were present and compare any observed defects with those of unused cross No. 1.
Sections through the locations of externally visible indications of unused cross No. I revealed a considerable network of crack like defects or grain boundary abnormalities at or near the outer surface of the cross. The abnormalities (defects) were completely intergranular and relatively wide open. Some grain boundary segments and junctions of grain boundary segments exhibited an abnormal grain boundary constituent. A detailed examination of the defects in the as polished condition revealed a deposit material within the defect itself and a unique microstructure adjacent to the defect.
Examples of these features are shown in Figure 7 (b) and (c) of Interim Report 85-M-017 (Attachment to ACCM-85/0372). Typically, a nearly continuous layer of a uniformly grey deposit material was present at or near the edge of the defect itself and a two phase structure was present adjacent to the
r PAGE 6 0F 16 defect. The two phase region consists of small, rounded, grey colored particles dispersed in the unetched matrix. Scaning electron microscope (SEM) and Auger spectroscopy reveal that this is a grain boundary matrix of chromium-rich iron / chromium oxide with islands of chromium depleted base metal. These microstructural features are characteristic of progressive oxidation outward from the defect into the base metal.
Although unused cross No. 2 did not exhibit LP indications on the outside surface, metallographic examination revealed a network of grain boundary defects immediately below the surface. This condition is the same as that observed in unuced cross No. 1, with the exception that the grain boundaries are generally tighter (not as open) and do not open to the outer surface.
The two-phase constituent is identical and quantitative Energy Dispersive X-ray Spectrographics (EDS) analysis performed at locations where the tight
. defects appear contain a single phase constituent identified the as a chromium-rich oxide.
4.0 DEPOSIT ANALYSIS A k-inch thick specimen containing one of the visible surface defect indicationa was cut from an outside surface specimen taken from unused cross No. 1 (11tSI Treated) . The specimen was broken open in bending to expose the cracked surface and examined in the SEM. Fractographs of the exposed surface are shown in Figure 17 of Interim Report 85-M-017 (attachment to AECM-85/0372). The intergranular nature of the fracture surface is apparent in Figure (17a) and the inherent granular oxido deposit is apparent in Figure (17b). An x-ray energy spectrum obtained from the crack surface is shown in Figure (17c) and identifies the surface deposit as a chromium rich oxide.
PAGE 7 0F 16 Qualitative EDS analysis of surface deposits detected the three principal elements of the base metal; iron, chromium, and nickel. There were no other extraneous elements detected.
A near surface deposit filled defect was analyzed by EDS and Auger spectroscopy confirming that the deposits within the grain boundary defects and the dispersed phase adjacent to the defects consists of chromium iron oxides in a chromium depleted matrix with the oxide layers concentrated near the outside surface.
The nature of the deposit material and the adjacent two phase region consisting of layered oxides at the surface with dispersed Fe-Cr oxides in a chromium depleted matrix is characteristic of high temperature oxidation of metal alloys. These results establish that the grain boundary defects in both crosses were present during one or more of the hot working and heat treating processes employed in fabrication of the crosses.
5.0 TENSILE TESTS Liquid penetrant examinations failed to show any surface defect under no load conditions for the unused Cross No. 2 which had not been subject to IHSI treatment. However, metallographic studies demonstrated the existence of grain boundary defects within the outer 0.3 inch of the 24 inch cross pipe.
It has been shown that these defccts existed before the application of IHSI to the Unit I recirculation system. Therefore it was decided to conduct tensile tests on the unused cross No. 2 base materini to satisfy two major i
objectives.
- 1. To evaluate the stress levels required to open up the surface and subsurface defects in a SA403 WP304 cross with no IHS1 treatment, and
( ..
PAGE 8 0F 16
- 2. To assess the significance of the defects in terms of the integrity of the piping crosses.
5.1 Specimen Description Four tensile test specimens were cut out. Specimen Nos. 1, 2 and 3 were nominally 5/8 inch thick, 1-1/4 inches wide and 12 inches long. These specimens were taken from the outer diameter of the pipe. Specimen No.
4 was taken at the ID of the pipe, with dimensions 1 inch thick,'1-1/4 inches wide and 12 inches long. A fifth specimen (Specimen No. 5) was taken to establish the metallurgical condition of the material before the tenaile tests were conducted. The tensile test specimens were machined at the ends for a length less than 3 inches to provide for good gripping action in the tensile test machine. In all cases a minimum length of 6 inches of the original cross pipe surface was preserved.
5.2 Test Description and Results Test Specimens 2, 3, 4 and 5 were used in the test program. Specimen No. I was not tested. The surfaces of the test specimens were visually or LP examined. There were no detectable indications on the original cross pipe surfaces. There was evidence of some indications in the machined areas of the specimens.
The following is a description of the various tests conducted and the results obtained for the tests.
- 1. Test Specimen 5.
This specimen was used to establish the baseline metallurgical characteristics for the tensile test program. Metallurgical examinations revealed material defects on the outer diameter surface. The defects were tight and not detectable by penetrant examination.
/
PAGE 9 0F 16
- 2. Test Specimen.3 A LP exemination of the surface material of Test Specimen No. 3 revealed no detectable indications. The specimen, with a cross-sectional area of 0.78125 square inches (1.25 inches x 0.625 inches), was placed in a tensile machine and loaded initially in-increments of 2,500 pounds to a load of 15,000 pounds. At this load level (net section stress o = 19,200 psi) a LP examination was
. performed under load. There were no LP indications noticed. The specimen was further loaded to 20,000 pounds (o - 25,600 psi) and finally to 25,000 pounds (o = 32,000 psi). Additional LP examinations were performed at the respective loads. Several indications were noticed at the 20,000 pound load level, with approximately fifteen indications appearing at the 25,000 pound load level.
For evaluation purposes the material yield strength was assumed to t
be 32,000 psi. Initial noninearity of the load displacement curve was noticed around the 25,000 pound load level. After unloading, the specimen was reexamined under a 50X magnification. The defects that may have opened under the 20,000-pound load appear to have reclosed and were not visible.
- 3. Test Specimen 2 The purpose of this test was to study the effect of loads beyond net section yield stress on the material with the internal defects. An I initial LP examination was conducted on the specimen surface. The l specimen, with a 1.247 x 0.67-inch cross section was loaded in steps i
l to load levels 20,000 pounds, 25.000 pounds, and finally to 35,200 l
l pounds. At each load level the specimen was unloaded and examined i
I L_
PAGE 10 0F 16 f
under a 50X magnification. No indications were found for loads of f
20,000 pounds and 25,000 pounds. However, significant nonlinearity developed in the load-displacement trace for leads exceeding 25,000 pounds. Surface indications developed and were visible at the 35,200-pound load level. On unloading, the surface flaws remained open and resembled the surface of the cross pipe treated by IHSI.
The 35,200-pound load level corresponds to approximately 130% of yield load.
- 4. Test Specimen 4 This specimen was obtained from the ID of the cross pipe. The purpose of this test was to demonstrate that no defects opened in the ID area under load. Specimen No. 4 was larger in cross-section that the others tested. The computed cross-sectional area of the specimen was 1.332 square inches. Pretest examination (LP) indicated no flaws or defects. The specimen was loaded in three stages. The load levels included 32,000 pounds, 39,900 pounds and 48,400 pottnds. After each load test, the specimen was examined under a 50X magnification. There was no evidence of any defect formation even beyond net section yield load conditions.
5.3 Summary The tensile tests reveal the following:
- a. The OD material defects were tight and were not easily detectable.
- b. Loads in excess of 80% of net section yield load were required before surface defects opened up.
- c. The surface defects experienced crack tip blunting at 130% net section yield load. The defects did not grow in any through
PAGE 11 0F 16 thickness direction'. They were stable even at net section yield stress conditions.
- d. The inside surface of the cross was free of any defects. No defects were detected even at loads exceeding not section yield loads.
6.0
SUMMARY
AND CONCLUSIONS i
A comprehensive metallurgical and structural test program was conducted to evaluate the post IHSI surface cracking of SA403 WP304 pipe crosses in the recirculation piping of Unit 1 at Grand Gulf Nuclear Station. The principal observations from this study are summarized as follows:
- 1. The chemical composition of the material used in the fabrication of the pipe crosses conformed to the ASME Specifications and the material passed the check for sensitization.
i
- 2. The material of the 24-inch legs of both crosses exhibited an abnormally large grain size.
- 3. Numerous grain boundary defects were present in all samples from the 24-inch legs which were examined (Unused cross No. I and Cregs No. 2).
The samples included bc,th material that had been subjected to IHSI treatment and virgin material.
- 4. The grain boundary defectn were confined to the outer 0.3-inch region of the crosses. The maximum linear cracking observed was 0.22 inch.
- 5. The grain boundary defects included oxidation attack and grain boundary separation. The deposits within the defects and the dispersed phase adjacent to the defects consist of Cr-Fe oxides in a Cr depleted matrix. t i
The oxide layers tend to concentrate nearer the outside surface.
i
/ PAGE 12 OF 16
- 6. No defects were found in the material at the inside surface of either leg of either cross..
l
- 7. Tensile load' tests indicate the following-i
- s. Up to a. load level of 100% of not section yield stress there was no evidence of open defects on the original surface of the test specimens taken from the outside surface of the cross.
- b. At a load equivalent to 130% of not section yield stress-(OD material) s.urf ace cracks were detectable by both visual (50X) and liquid penntrant examination. Metallurgical examinations after testing verified crack opening under load. Plastic deformation was observed at the tips of the opened defects but no crack propagation was evident. i
- c. Material from the inside of the cross loaded to over 120% of net section yield stress displayed no defects on the inside surface.
The metallographic evaluations of the sample material from Unused cross No. I and Unused cross No. 2 revealed numerous defects on the 0.D. of the 24-inch legs of both crosst s. In both cases the defect zone extended approximately 0.S inch below the outside surface. All of the defects in both sets of samples were intergranular in nature (except for a few short transgranular extensions of predominately intergranular defects in samples from unused cross No. 2). Auger spectrographic analysis, together with the microstructural features, provided positive evidence of oxidation attack along defects as deep as 0.17 inch below
'the outside surface. All of the microstructural data serves as positive l
evidence that the defects in both crossen are essentially identical in l nature. Since unused cross No. 2 did not receive any IHSI treatment at l any time in its history, the similarity of the defects in the two l
4
I l
.i.
PAGE 13 0F 16 crosses is cor.clusive evidence that the defects were inherent in the crosses. Therefore, it is evident that the defects in unused cross No.
I were not caused by the IHSI treatments.
The metallographic examinations established that the grain size of the 16-inch arms in Cross No. 3 was considerably smaller than that of the 24-inch legs. The only mechanism by which such a situation could occur is by partial hot working of an originally large-grain-size materisi.
In such a case, recrystallization and grain refinement would occur in any ntwly hotworked material producing significantly different grain structures at different locations within the same part. It was reported that the final hot forming operation for the crosses involves drawing out the 16-inch legs from the 24-inch OD cylindrical blank. This operation results in hot working only that material which is drawn into the' arms (the major portion of the 24-inch diameter legs is unaffected) and accounts for the difference in grain size. This factor establishes that the large grain size developed in some operation prior to the forming operation for the 16-ineb legs.
The intergranular nature of the defects in the 24-inch legs and the absence of defects in the 16-inch arms establishes that the defects developed at some stage of manufacturing during o after the development of the abnormally large grain size.
In summary, the known factors concerning the development of the defects are as follows:
- 1. The defects developed in some stage fabricatien prior to the forming of the 16-inch arms.
l l
L
PAGE 14 0F 16
- 2. The defects developed or were open during elevated temperature processing (hot working or heat treating) as evidenced by the oxidation within the defects.
r
- 3. The defects developed during or after the operation which resulted in the large grain size of the original 24-inch diameter cylinder.
The grain size observed in the 24-inch legs of both crosses is much larger than is usual for this type of wrought product. Th*.s factor, together with relatively severe oxidation attack within ard adjacent to the defects suggests that the parts were subjected to abnormally high temperatures and/or heated for abnormally long periods of time during the fabrication of the 24-inch cylindrical preform.
On the basis of the results of this investigation the following
. conclusions may be drawn.
- 1. The grain boundary defects were inherent in the crosses and preexisted the piping installation. They were present or were formed during one or more of the high-temperature processen employed in the original fabrication of the crosses.
- 2. IHSI was not the cause of the defect formation.
- 3. IHSI probably contributed to the opening up of the preexisting defects such that they became detectable by LP inspectior.
- 4. Tensile tests demonstrated that the preexisting, near-surface defects will not open up under operational or design loads. The required design factor of safety is expected to be maintained in the particular cronses examined. r l
l l
l 1
(
I PAGE 15 0F 16 7.0 DETERMINATION OF GENERIC APPLICABILITY AT GCNS The recirculation system 24" x 16" crosses are SA-403 WP304 with a base material of SA-182 and are supplied by Taylor Forge. SA-403 is a specification for the forming of fittings from other preexisting materials (e.g. SA-240, SA-182, SA-376), and not a material specification in itself.
Tha size and thickness of the crosses makes them unusual for the forging process, and would require a mill to establish unique processes and tooling for their manufacture. Normal forging operations are performed by a hammer, press, forging machine, or forging rolls. The attached figure indicated that a normal forging process was only used in the early stages of the component and that the actual forming was performed by a back extrusion and expansion process. In other words, the crosses are not a catalog fitting or shelf item, and assembly line processes were not applicable.
SA-403 is a common material as described in the design documents (NON-NSSS) of CGNS. For ASME Class 1 fittings greater than 8", and ASME Class 2 and 3 fittings greater than 10", SA-403 WP304W is specified. The "W" suffix of 304W indicates that it is a welded fitting (fabricated); and is normally fabricated / welded from formed SA-240 plate, thus the forging process is not used. The use of a forged base product in the smaller diameters is not precluded, but with the thickness associated with these smaller diameters and the BWR design, normal forging operations would he used.
ASME Class 1, 2, and 3 stainless steel flanges as described by CGNS design documents are SA-182 and are manufactured by the simplest forging process.
Large diameter flanges are of the weld neck design and both the I.D. and 0.D.
surfaces are machined.
e.w
- *~
PAGE 16 0F 16 A review of the recirculation system has identified the crosses, 16" and caps and the 16" x 12" header to riser sweapolets as the only significant sized forgings. The sweepolets and end caps, though large, are still simple forgings and would not require special tooling or process development for their manufacture. Other major fitting (e.g. elbows, tees) are manufactured from formed plate (SA-240).
Based on the above it is concluded that no like components are installed in GGNS Unit I.
In the unlikely event that grain boundary anomalies did exist in other components, they are not discernible until the tensile stresses induced by IHSI are employed to aggravate the preexisting condition. The Non-IHSI treated anomalies exist on a micro scale and require 0.D. surface tensile stresses substantially higher that what is typically seen during plant operation to aggravate the anomalies into becoming discontinuities.
Therefore, similar conditions as identified on the cwo (2) 24" x 16" recirculation system crosses are not suspected on other stainless steel fittings at GGNS-Unit I.
8.0 REFERENCES
8.1 Southwest Research Institute Project 06-8915-001 Final Report,
" Recirculation Piping Center Cross Surface Cracking Evaluation-GGNS Unit gn 8.2 Futech Engineers Final Report MTL-03-226 Rev. O. " Failure Mode Evaluation of Liquid Penetrant Indications" 8.3 Electric Power Research Institute Lettern W. J. Childs, to J. M. Fell, t MP&L dated November 12, 1985 and December 19, 1985.
_w