Stress Corrosion Cracking in Repair-Welded 3.5 NiCrMoV Steel in an. Actual Turbine Environment

Transcription

1 FR Stress Corrosion Cracking in Repair-Welded 3.5 NiCrMoV Steel in an Actual Turbine Environment Hitomi Itoh, Takashi Shige (1), Takashi Momoo (2), A. A. (3) (1) Takasago Research & Development Center, Mitsubishi Heavy Industries, Ltd. (2) Takasago Machinery Works, Mitsubishi Heavy Industries, Ltd. (3) The Kansai Electric Power Co., Inc. Key words: PWR, Materials, Repair Abstract Temporary welding repairs are sometimes needed when damage occurs at the teeth of blade grooves in a low-pressure turbine rotor operated at the dry/wet boundary region. When repair welding has been performed for the 3.5 NiCrMoV steel used in low-pressure turbines, the soundness of the weld must be confirmed. For this reason, a laboratory investigation of susceptibility for stress corrosion cracking (SCC) was conducted for test specimens taken from simulated welds, and then an exposure test was conducted in an actual turbine environment for approximately 7,000 hours. As no SCC initiation was detected and also the propagation was extremely small, repair welding is deemed to be applicable. 1. Introduction The 3.5 NiCrMoV steel used in low-pressure turbine rotors is manufactured from gigantic ingots, through forging and subsequent quenching and tempering. As the steel ingots are extremely large, quenching must be effective; in other words, there must be little mass effect. Rotor welding has been discussed in a separate study 1 ', but there appear to be no studies of corrosion resistance using these test materials. Stress corrosion cracking (SCC) damage like that shown in Fig. 1 may occur at the teeth of the blade grooves in low-pressure turbine rotors operated at the dry/wet boundary. In such cases, an enormous amount of time is required to replace the 3.5 NiCrMoV steel rotor, and therefore welding to temporarily repair the rotor is an option. Therefore, in preparation for the time that such repair welding is needed for actual rotors or discs, the authors conducted a welding repair test for a disc made of 3.5 NiCrMoV steel to confirm the corrosion resistance of this area, in particular with respect to SCC and corrosion fatigue. 33/08

2 OGOO 2. Repair welding procedure and results of inspection 2.1 Test materials Fig. 2 shows the shape and tensile properties of the mock-up 3.5 NiCrMoV disc used for the welding test. Gas tungsten arc welding (GTAW) was performed in part of the area near the teeth of the blade grooves. The disc as received had a 0.2% offset strength of 860 MPa; however, after 24 hours at 893 K to simulate stress relieving, the value was 714 MPa. Normally, after welding, the material is checked not by testing the tensile strength but by measuring the hardness; hardness measurements performed at this time are also shown in the figure. The maximum hardness measured as weld was 495 DPH; after stress relieving, this value was 292 DPH. 2.2 Procedure for partial area welding repair Fig. 3 shows the procedure for welding repairs. The test material disc was welded as described above. All of the teeth were removed from the blade grooves of one part of the disc, while only some of the teeth were removed from another part of the disc. These disc sections were then repair-welded and investigated. In both cases, common metal feed wires were used. Post weld heat treatment (PWHT) was performed after welding. Fig. 4 shows the repair welding procedure for the two locations. In one location, only the top tooth was removed from the blade groove and build-up welding was performed to repair that tooth. In the other location, all of the teeth down to the bottom tooth were removed from the blade grooves and the tooth was built up through welding. Different build-up welding methods were used for welding of the first and second layers and welding of the third layer on; the diagram in Fig. 5 illustrates this. As will be explained in detail later, another study was also conducted at the same time in which the entire row was removed and automatic welding was performed. 2.3 Results of test conducted after partial welding repair The welded disc was milled to a shape that would permit the test area to be examined, and then three tests were performed: a fluorescent liquid penetrant test (PT), a magnetic particle test (MT) and an ultrasonic test (UT). In the first two tests, no defects were detected. However, in the latter test, two extremely tiny defects were detected, as shown in Fig. 6. Accordingly, these two defects were eliminated before PWHT was performed and the area was repair-welded. PWHT was conducted for 24 hours at 893 K, and an analysis of the stress distribution both before and after was conducted. Table 1 shows the results of the analysis; the tangential stress at the outer periphery of the disc increased from 84 MPa before heat treatment to 216 MPa after heat treatment. The hanging stress of the disc bore, on the other hand, was almost unchanged at approximately 260 MPa. The increase in tangential stress was within the estimated range.

3 oooe 2.4 Procedure for full area welding repair As noted above, a test of welding repairs for the entire blade groove was also performed. Fig. 7 shows how preheating was conducted; the electric heater was placed so as to enclose the disc, and the rotor was preheated as it turned. After preheating, automatic GTAW in which the wire is automatically fed was performed as shown in Fig. 8. Fig. 9 shows the appearance of the mock-up disc after welding. Subsequently, PWHT was conducted under the conditions described earlier. 2.5 Results of test following full area welding repair After PWHT, milling was performed and the blade groove was shaped as shown in Fig. 10. The disc passed the subsequent size inspection, and no defects were detected in the PT, MT and UT tests, confirming that welding can be performed without any significant defects. 3. Test of material properties of welded section Test specimens were sampled from the aforementioned repair-welded sections and the material properties were tested. 3.1 Tensile properties Table 2 shows the results of the tensile tests. The 0.2% offset strength of the weld metal, base metal and weld joint were almost the same at approximately 700 MPa. Ductility was also judged to be satisfactory, with elongation of more than 20% and R.A. of more than 70%. 3.2 Impact Properties Impact test specimens were machined from the heat affected zone (HAZ), the welding bond and the base metal in the radial direction, and impact tests were conducted. In each case, the absorbed energy was 100 J or greater, and the transition temperature was 250 K or less. Fig. 11 shows some typical results; from these results, the impact properties were judged to be satisfactory. 3.3 Hardness Fig. 12 shows the results of hardness measurements. The maximum hardness was found in the HAZ near the weld bond. Compared to the base metal and weld metal, the value was 40 DPH higher in this area, but all of the values were around 300 DPH. From these results, the increased hardness due to welding was judged to be within the allowable range. 3.4 Metallography The macrostructure and microstructure of the welded sections were examined. All areas (weld metal, HAZ and base metal) had a well-tempered martensite structure. 4. Corrosion susceptibility test

4 oooo For the corrosion susceptibility test, test materials prepared from the aforementioned disc as well as several test plates (see Table 3) were prepared; the prepared weld plates are shown in Fig. 13. Stress relieving was conducted for 24 hours every 10 K from 883 K to 913 K to prepare test materials for the corrosion susceptibility test. The aforementioned disc was used as the 893 K material; the tensile properties of three plates made into test materials are shown in Fig. 14, arranged by temper parameters. The 0.2% offset strength of the weld metal (labeled YS in the figure) was approximately 750 MPa to 600 MPa. As shown in Fig. 15, it was confirmed that the hardness of the welded section decreased as the tempering temperature increased. In each case, the maximum hardness for each type of heat treatment was obtained in the HAZ. Three tests were performed for corrosion susceptibility: SCC tests, performed both in a laboratory and in an actual turbine environment using the aforementioned test materials, and corrosion fatigue (CF) tests performed in the laboratory. 4.1 SCC test in the laboratory The test specimens shown in Fig. 16 were used for the SCC test performed in the laboratory. These test specimens were immersed in de-ionized pure water with a DO 2 concentration of 0.1 ppm at a temperature of 403 Kfor 10,374 hours. In the SCC test, crack initiation was investigated first. From the results obtained within the scope of this study, no cracking occurred in any of the test specimens. As shown in Fig. 17, the maximum corrosion pit diameter tended to decrease as the strength increased. This trend is the same as that previously reported by Itoh et al 2). It is thought that this may reflect the increase in the size of the carbides as tempering progressed. Measurements of the crack propagation rate were also investigated. However, during the test, no crack propagation was observed. 4.2 SCC test in an actual turbine environment As there are limits to the number of test specimens that can be used in laboratory tests, an SCC test was performed in an actual turbine environment. Fig. 18 shows the location in the turbine at which the test specimens were installed and the steam conditions at that location. A bypass line leading from the extraction line was installed and the test specimens were placed in the middle of this bypass line. The test time was set at 6,629 hours to match the operation cycle. As an example of test results, Fig. 19 shows the relationship between maximum pit diameter and stress normalized by 0.2% offset strength. There was almost no difference in maximum pit diameter between the weld metal and the HAZ, and stress appeared to have no influence on maximum pit diameter. In the previous test 2 ', neither stress nor the crevices appeared to influence this value for the base metal. In the

5 oooe current test, however, although the results for stress were similar, the crevices were found to have some influence and the value for maximum pit diameter was greater. Compared to the base metal, there was more variation in maximum pit diameter in the welded sections. This is thought to reflect the fact that, compared to the uniform base metal, the welded section tended to be less uniform structurally. Fig. 20 shows two parameters needed for stress corrosion cracking initiation: pit growth rate and critical pit size (calculated from K 1SCC P ). Of these two, the rate of growth for the diameter of the corrosion pits is expressed by m in the equation below. where 2c = kt m (1) 2c: Corrosion pit diameter (mm) k: Coefficient t: Time (hours) m: Growth rate power Fig. 21 shows the m values obtained in the field test. For the average value + double the standard deviation ( + 2 ), the m value was This was lower than the value of 0.33 obtained for the base metal alone 2), Compared to the latter, the former value is for a larger number of test specimens installed in a large number of turbines. Therefore, out of consideration for variations, the latter should be used as a conservative value. The results showed that, in the welded sections, the maximum corrosion pit diameter was affected by the crevices. However, from the standpoint of the m value, as shown in Fig. 22, the m value was larger when there were no crevices. Accordingly, it is thought that the use of the aforementioned conservative m value will enable the remaining service life to be evaluated using the data obtained for the base metal. The value for K 1SCC P is derived using the following equation: K.scc" \.\. % (2) ' /Q 1.65 Q (3) where: K P 1SCC : Critical value for SCC initiation from corrosion pits j : Stress a: Pit depth c: 1/2 pit diameter Q: Shape factor P As there was no crack initiation within the scope of this test, the value for K 1SCC

6 oooo could not be obtained directly. However, a provisional value can be calculated from the maximum corrosion pit diameter for which no cracking occurred in the laboratory and actual turbine environments. The maximum pit diameter is expected to be approximately 0.07 mm. Of the test materials, if the one with 0.2% offset strength of 700 MPa is used, the value for K 1SCC P will be approximately 5 MPa m 05, which is a slightly more conservative value than the 5.5 MPa m 05 obtained for the base metal. Therefore, this value should be used as the provisional K 1SCC P value. 4.3 CF test in the laboratory The CF test was performed only in the laboratory. For the weld metal, weld joint and base metal, the Schenk type test specimens shown in Fig. 23 were sampled from the disc material. The CF test was conducted in deaerated pure water at 403 K and the fatigue strength up to 10 8 cycles was obtained. Fig. 24 shows the results of the CF test. The materials can be ranked in the following order according to their fatigue strength up to 10 8 cycles: base metal, weld metal and weld joint. The hardness values for the base metal, weld metal and HAZ were 290, 276 and 258 DPH, respectively; the results of the CF test reflected the hardness of the material. The tensile strength of the base metal was approximately 950 MPa, but the fatigue strength up to 10 8 cycles of CF was approximately 450 MPa, or approximately 0.47 of the tensile strength at room temperature. The fatigue limit is considered to be half the tensile strength, so taking the test temperature into account, there was no significant difference between the results of the fatigue test in the atmosphere and the results of the CF test in deaerated pure water. 5. Conclusion When the 3.5 NiCrMoV steel used in low-pressure turbines is repaired by welding, the soundness of the welds must be confirmed. For this reason, the authors first studied the welding process and then investigated the welds to determine their susceptibility to stress corrosion cracking (SCC) and their corrosion fatigue (CF) characteristics. Laboratory tests were conducted to investigate CF, and tests both in the laboratory and in an actual turbine environment were conducted to investigate SCC. The results were as follows: (1) Using the teeth of the blade grooves in 3.5 NiCrMoV steel discs, gas tungsten arc welding was used for both partial area and full area repair-welding. After the post weld heat treatment, the disc was milled to the shape of a blade groove and a size inspection, fluorescent liquid penetrant test, magnetic particle test and ultrasonic test were performed. Through these tests, it was conformed that welding could be performed with no significant defects. (2) The material properties of the repair-welded sections were tested. The results

7 oooe of tensile property, impact property, hardness, macrostructure and microstructure tests were satisfactory. (3) Repair-welded 3.5 NiCrMoV steel was used as a test material for SCC tests, both in the laboratory and in an actual turbine environment. While no crack initiation occurred, the test found that the corrosion pit growth rate power m used to evaluate remaining service life was approximately 1/3, and a value of approximately 5 MPa m 05 was obtained for the limit value for crack initiation K 1SCC P. This was a slightly more conservative value than the 5.5 MPa m 05 obtained for the base metal. (4) CF tests using 3.5 NiCrMoV steel that had been repair-welded were performed in the laboratory. After 10 8 cycles in deaerated pure water at 403 K, strength values for the base metal, weld metal and heat affected zones that reflected the hardness of these materials were obtained. References 1) E. Kramer, N. Lannefors, B. Scarlin, "Advanced reliable low pressure steam turbine retrofits", PWR-Vol.26, Advances in steam turbine technology for the power generation industry, ASME 1994, p.89 2) H. Itoh, T. Momoo, T. Sakaguchi, "Residual life prediction based on SCC initiation in 3.5NiCrMoV turbine disc steel under field turbine conditions", ICONE-8112, (2000)

8 oooo YS MPa Tensile properties PS MPa El % RA % Max. hardness ol weld IOPH) As received S After heal treatment 893K-24H TU { As weld ^ i 495 j Keyway \ / Sampling Fig.1 Areas in which SCC was detected at disc rim attachment a Results of mock-up test for welds on..., disc steel Disk Rotor Welding wire Certification test before welding Tensile test, hardness measurement PWHT [ Removal of the groove area Shrink mung ** Pre-heating 1 Welding Full area Local area Size measurement Residual stress measurement 1 UT, MT, PT I Finished performance \ MT, PT i Size measurement \ Residual stress measurement I Removing of test specimens \ Certification test Fig.3 Procedure for mock-up test applied to full scale disc First and 2nd layer weld build-up welding After 3rd layer Chilt Build-up welding (a) Local tooth area weld (b) Full tooth area weld Fig.4 Shape of local welds at rim attachment s58 Automatic welding Fig.5 Shape of full area welds at rim attachment

9 oooo Vertical and UT build-up welding build-up welding 320 Local area weld (a) Local tooth area weld (b) Full tooth area weld Fig.4 Shape of local welds at rim attachment Fig.6 Results of UT test for mock-up test weld Electric heater Table 1 Analytical residual stress values in mock-up test Q rotor r> Direction of rotation Before repair After repair Radial stress at rim oe(mpa) Hanging stress at disc bore o, (MPa) Fig.7 Illustration of pre-heating treatment torch,n~7welding wire "?L- dlsc direction of rotation ---0 full area repair, AX) \ local area repair Fig.8 Rim attachment welding Fig.9 General view of mock-up

10 oooo Table 2 Tensile properties of mock-up welds YS (MPa) i i TS (MPa) El (%) RA (%) weld metal radial base metal radial tangential Fig.10 View of welded rim attachment af ter completion weld joint tangential 705 " Reference valub I >a o weld, HAZ base metal Test temperature K ) ( distance from weld bond (mm) Fig.11 Results of Charpy impact test at HAZ of weld joint Fig.12 Hardness of weld joint after stress relieving Table 3 Chemical composition and mechanical properties of test plate c Chemical composition trims % S. Mn! P j 8 Nr Cr Mo V YS MPa rs MPa >3(W*jjr properties % SCVA" K J [ \ Tensile test pieces DIGS scataca E3t39 \ ^SCC tesl pidces ! i :! 1 1.S S ? 21? Fig. 13 Test weld plate and test pieces removed from test weld plate

11 - oooo 913K ttfths ? a 0 o S e i KAZ. base metal 190 us temper parameter T.P«Tx(20»IOflt)x10' T: tefnpefature K I: time h Fig.14 Tensile strength of weld metal after heat treatment o 1 J- Fig.15 Distance from weid bond mm Hardness of weld plate 10? sxposufa ^ot 1O3?4h R2 / R2 Observation area Fig.16 SCC test specimen Fig.17 LMP (T(20»log t) x 10 l ) Maximum pit diameter after exposure in water at 403K (Water : de-ionized water with DO ppm) LP turbine SR :893K Field test conditions temp.: 393K time: 6629h weld metal _ Extraction line I Heater -\zzm\- Specimens Orifice base metbl o Fig.18 Condenser Location at which specimens were installed o i 0 (! aid, Fig.19 Maximum pit diameter obtained in field test

12 oooo Corrosion pit size (log) Crack t f Pjt Critical pit size calculated from Kwcc maximum pit size at inspection Extension life before SCC initiation crack propagation crack initiation Fig.20 Time (log) Values needed to evaluate service life Fig ! m value of weld obtained in field test 0.7 0$ SR:893K Weld test conditions temp.. 393K lime : 6629h wold metal MA2 i O "" To""" o o in 94 o o i 0.5 s ei I j f tola t 4 L total The heat affected zone is the center y of the specimen when the weld joint was used. SR temp. 893K ' Weld metal 0.1 Fig.22 weld melal HA2 m value obtained in field test Fig.23 Test specimen for corrosion fatigue test * * a o 0 A k- aerated pure water {OOA < 5ppb) at 403K o Base niaiertai A Weld joint a WeH metal SR lemp 893* * 10' 10' 10- N cycle Fig.24 Result of corrosion fatigue toct