- Email: [email protected]
Post examination of tungsten monoblocks subjected to high heat flux tests of ITER full-tungsten divertor qualification program
Fusion Engineering and Design 121 (2017) 60–69
Contents lists available at ScienceDirect
Fusion Engineering and Design journal homepage: www.elsevier.com/locate/fusengdes
Post examination of tungsten monoblocks subjected to high heat flux tests of ITER full-tungsten divertor qualification program Zhaoxuan Sun a,b , Qiang Li a,∗ , Wanjing Wang a , Ji-Chao Wang a,b , Xingli Wang a,b , Ran Wei a,b , Chunyi Xie a , Guang-nan Luo a,b,c,d , T. Hirai e , F Escourbiac e , S. Panayotis e a
Institute of Plasma Physics, Chinese Academy of Sciences, Hefei, 230031, China Science Island Branch of Graduate School, University of Science & Technology of China, Hefei, 230031, China c Hefei Center for Physical Science and Technology, Hefei, 230022, China d Hefei Science Center of Chinese Academy of Sciences, Hefei, 230027, China e ITER Organization,Route de Vinon sur Verdon, CS 90 046 13067 Saint Paul lez Durance Cedex, France b
h i g h l i g h t s • • • •
ASIPP manufactured six W monoblock mock-ups that were tested at IDTF for full-tungsten divertor qualification program. Ultrasonic test was performed to investigate the defects of interface. Destructive test was performed on three mock-ups to analysis the damage. FEA was performed to study the temperature and stress distribution of monoblocks.
a r t i c l e
i n f o
Article history: Received 21 December 2016 Received in revised form 30 March 2017 Accepted 6 June 2017 Keywords: ITER Tungsten Divertor High heat flux Monoblock
a b s t r a c t In 2015, as part of ITER Full-Tungsten Divertor Qualification Program, Institute of Plasma Physics Chinese Academy of Sciences (ASIPP) manufactured six small-scale Tungsten (W) monoblock mock-ups that were tested at the electron beam facility, ITER Divertor Test Facility (IDTF, St Petersburg, RF). The high heat flux (HHF) tests consisted of 5000 cycles at 10 MW/m2 , followed by 300 cycles at 20 MW/m2 and additional 700 cycles at 20 MW/m2 . All mock-ups fulfilled the requirements of high heat flux performance successfully. One (WTC-5) of the six mock-ups was then selected to carry out critical heat flux (CHF) test with local heat flux up to 37–39 MW/m2 . After HHF test, ultrasonic test (UT) and destructive tests were performed to characterize the damages of monoblocks. The debonding of the Cu/CuCrZr interface was detected by both nondestructive and destructive tests. Intergranular rupture of W was observed by Scanning Electron Microscope (SEM). The recrystallization was found in the W monoblocks and the recrystallized depth were analyzed by metallography and Vickers hardness measurement. © 2017 Published by Elsevier B.V.
1. Introduction Full tungsten (W) divertor qualification program have been established to develop and validate the performance of W monoblock technology for W divertor in November 2011 [1–3]. Upon the two years of efforts on engineering and physics aspects, the ITER Council endorsed the proposal to start ITER operation with full-W divertor in November 2013. Domestic Agencies in Japan and Europe have succeeded to demonstrate high heat flux (HHF) performance of W monoblocks for a full-tungsten divertor in colFig. 1. Tungsten monoblock mock-ups with swirl tapes before HHF testing. ∗ Corresponding author. E-mail address: [email protected] (Q. Li). http://dx.doi.org/10.1016/j.fusengdes.2017.06.009 0920-3796/© 2017 Published by Elsevier B.V.
Z. Sun et al. / Fusion Engineering and Design 121 (2017) 60–69
61
Fig. 2. IR-pictures of average absorbing heat flux at (a)-24.7 MW/m2 , (b)-25.5 MW/m2 , (c)-26 MW/m2 .
laboration with the ITER Organization (IO) [4,5]. In China, Institute of Plasma Physics Chinese Academy of Sciences (ASIPP) has set up to develop W monoblocks technology since 2010 and applied it to the upper divertor on the EAST Tokamak. In 2015, as a part of qualification of tungsten (W) monoblock technology for ITER divertor, ASIPP and Advanced Technology and Materials Co., Ltd (AT&M) manufactured six small-scale W monoblock mock-ups that were sent to Fig. 3. Possible distribution made by the simulation.
Fig. 4. Average temperature of each monoblock gained by IR-camera of WTC-2 HHF testing: (a)-5000 cycles at 10 MW/m2 , (b)-300 cycles at 20 MW/m2 , (c)-additional 700 cycles at 20 MW/m2 .
62
Z. Sun et al. / Fusion Engineering and Design 121 (2017) 60–69
Fig. 5. Heated surface of 5 mock-ups after 5000 cycles at 10 MW/m2 and 1000 cycles at 20 MW/m2 . The coolant water flow direction is indicated.
Russia for the HHF testing. All mock-ups passed the examinations successfully and met the IO requirements. After HHF test, analyses were performed to evaluate the damages of the mock-ups. The analyses consisted of ultrasonic test to examine the interface of W/Cu and Cu/CuCrZr and destructive test to investigate microstructure of monoblocks. Destructive tests were performed on three mock-ups. Both ultrasonic tests and destructive tests detected the defects of Cu/CuCrZr interface. The depth of cracks and recrystallization were also analyzed to learn more details of the monoblocks. Vickers hardness was also carried out on W monoblocks to validate the relation between recrystallization and Vickers hardness. 2. Mock-up manufacturing and high heat flux testing All small-scale W monoblock mock-ups for high heat flux testing were manufactured by Hot Isostatic Pressing (HIP) technology, which is developed by ASIPP and AT&M. The HIPing conditions were an isostatic pressure of 100 MPa, the heating temperature of 600 ◦ C and the heating time of 2 h. The W monoblocks were cut from rolled W plates and the rolled direction is perpendicular to the loaded surface. Each mock-up consists of 7 W monoblocks separated by a gap of 0.4 mm. between the neighbor monoblocks a molybdenum spacer (15/19 mm (ID/OD)) was placed to maintain the gaps during HIP. The W monoblocks have a width of 28 mm, height of 26 mm and the axial length of 12 mm. The armor thickness (minimum distance between the heat load surface and interlayer (pure copper UNS C11010) is 6 mm. The functional material of pure Cu (about 1 mm thick) is used as interlayer to reduce the stress caused by the differences of W and CuCrZr thermal expansion. The CuCrZr tube manufactured by State Nuclear Bao Ti Zirconium industry, is used as heat sink material for its good thermomechanical behavior. The tube diameter is 12/15 mm (ID/OD). A copper twisted tape, with a
twist ratio of 2 and a thickness of 2 mm, was inserted into the tube to enhance convective heat transfer. The W material, CuCrZr tube and twisted tape were all supplied by AT&M. All mock-ups were examined by ultrasonic test and didn’t show any defects on joints before HHF test. Fig. 1 shows the mock-ups with swirl tapes before HHF testing. The HHF testing was carried out at IDTF, which was designed for HHF testing on the ITER inner and outer vertical targets and domes plasma-facing units [6]. The HHF test consisted of 5000 cycles at 10 MW/m2 , 300 cycles at 20 MW/m2 and additional 700 cycles at 20 MW/m2 . The heat loads duration were set-up at 10 s on and 10 s off. For each mock-up, four out of seven monoblocks were tested. The surface temperature was measured by IR camera and pyrometers, the absorbed heat flux was obtained by global calorimetry from measurement of temperature from thermocouples mounted at the water coolant inlet and the outlet. The hydraulic conditions were set at ITER relevant conditions. The used parameters were a pressure of 3.9 MPa, the water coolant inlet temperature of 70 ◦ C and the mass flow rate of ∼3 kg/s (equivalent the velocity of ∼11 m/s). After 1000 cycles at 20 MW/m2 applied to all the mockups, one mock-up (WTC-5) was selected to carry out the critical heat flux (CHF) testing to evaluate the critical heat flux of the mockups. When the average heat flux increased up to 24.7 MW/m2, overheating of the masks started to occur. The electron-beam raster was modified with the objective to avoid the copper masks’ damage. This change of raster generated non-uniform heating of the mock-up as shown in Fig. 2. At higher heat fluxes, due to the X-ray matrix’s saturation at this stage of experiment verification of heat flux profile was not feasible. Then the finite element analysis was performed according to the temperature gained by IR-camera to simulate the distribution of heat flux when average heat flux was 26 MW/m2 as shown in Fig. 3.
Z. Sun et al. / Fusion Engineering and Design 121 (2017) 60–69
63
Fig. 6. Surface morphology and cut-5-4 of WTC-5 after CHF test. The cutting line for metallographic examination is indicated as 5-4.
Fig. 7. The cutting line and numbering of WTC-2 (a) and WTC-9 (b).
All mock-ups passed the HHF test successfully in the sense that no significant increase of steady state surface temperature along the cycling was observed. Thus met the IO requirements (<30% surface temperature increase in 5000 cycles at 10 MW/m2 and 300 cycles at 20 MW/m2 ). Fig. 4 was the average temperature during HHF testing of four monoblocks (monoblock 1–4) of WTC-2. From this picture it can be seen that the surface temperature increases more at the additional 700 cycles. But it doesn’t exceed 30% of the surface temperature at start of the test at 20 MW/m2 . However, some degradation was found in the mock-ups. After the 5000 cycles at 10 MW/m2 , no visible damage was observed. After 300 cycles at 20 MW/m2 , the surface of the monoblocks became coarse and a longitudinal crack, so-called macro-cracks (self-castellation) appeared at the center of two monoblocks (among 6 × 4 = 24 subjected to the heat flux). After additional 700 cycles of test at 20 MW/m2 , longitudinal cracks can be seen in 9 of 24 monoblocks in all the 6 mock-ups. Fig. 5 shows the surface morphology of 5 mock-ups after 1000 cycles at 20 MW/m2 test. At the loaded surface, roughening and cracks were observed. Crack formation at the middle of monoblocks was also found in the mock-ups made by European industry [1,7,8]. Fig. 6 shows the surface morphology of the WTC-5 after Critical Heat Flux (CHF) test. The melting on loaded surface of the tungsten can be seen clearly. 3. Post test analyses After HHF test, destructive examinations were performed on mock-up WTC-2, WTC-5 and WTC-9. The cutting line and numbering rule of WTC-2 and WTC-9 are shown in Fig. 7. For WTC-5,
Fig. 8. Optical microscope image of (a) cut-2-2 and (b) cut-2-5.
one monoblock (namely cut-5-4,) was cut to find the details of microstructure.
3.1. Crack formation in W monoblocks and fracture interface of the cracks One to two macro-cracks were found on all of the mock-ups. For the WTC-2, two of four tested monoblocks formed a longitudinal crack (cut-2-2 and cut2-5). Fig. 8 show the crack formation of cut2-2 (a) and cut-2-5 (b) by optical microscope. It can be seen from the figure that a single crack propagates from the loaded surface further into W and even reached to the W/Cu interface. The width
64
Z. Sun et al. / Fusion Engineering and Design 121 (2017) 60–69
Fig. 9. SEM image of monoblock’s fracture surfaces. (a) Near loaded side; (b) near Cu side.
of the crack in the W heated surface is larger than that in the W/Cu interface side. Fig. 9 shows the fracture surface of the top side and near Cu interlayer of crack (cut-2-2). The fracture surface structure showed boulder-strewn structure which indicated that the main mechanism of fracture was intergranular rupture. Plastic deformation can be seen in the top side of the monoblock as it was observed on European monoblocks [9]. Recrystallized grain can be found in loaded side of monoblocks (Fig. 9(a)) and the microstructure near the Cu interlayer showed rapture surface reflecting the original grain orientation (Fig. 9(b)). 3.2. Interface Fig. 10 shows the images of the local Cu/CuCrZr interfaces. The destructive analysis showed that the debonding concentrated on the interface of Cu/CuCrZr. Two defects (about 1 mm) were seen
for cut-2-2 while cut-2-5 has three discrete defects, although no defects were detected by ultrasonic inspection before HHF test. The debonding of Cu/CuCrZr interface was located at the loaded side of heat sink. The defects were not found in W/Cu interface. For WTC-9, no defects were found at Cu/CuCrZr as well as W/Cu interfaces in all the examined monoblocks. For WTC-5, melting occurred on both copper interlayer and loaded W surface under CHF test. The joint area was destroyed and voids were found at the W/Cu interface. At the heat flux of 20 MW/m2, there exists a large temperature gradient near the interface of the Cu and the CuCrZr. This large temperature gradient can cause a big thermal stress and plastic strain at the Cu/CuCrZr interface. As the HHF test continue, the Cu/CuCrZr interface would suffer from fatigue damage due to the cyclic thermal stress and plastic strain. Finally, the debonding occurred. The other three mock-ups’ interface was inspected by Ultrasonic Test (UT). Ultrasonic inspection was performed scanning the probe in the tube over the inner surface of tube immersed into water [10].
Z. Sun et al. / Fusion Engineering and Design 121 (2017) 60–69
65
Fig. 10. SEM image of Cu/CuCrZr interface. (a): cut-2-2; (b): cut-2-5; (c) cut-5-4.
The frequency of the ultrasonic was 15 MHz. The test was focus on the interface of W/Cu and Cu/CuCrZr. Fig. 11(a) shows the UT results of Cu/CuCrZr interfaces. The 0◦ position corresponds to the 6 o’clock position in the monoblock. The amplitudes of the echo signals of the three mock-ups’ Cu/CuCrZr interface at specific zones are much higher (red zone) means that there were defects occurred at these locations. All the defects occurred at the loaded area as indicated
(black box area) in the picture. No defects were detected at W/Cu interface for all 3 mock-ups like WTC-13 in Fig. 11(b). 3.3. Recrystallization of W monoblocks During the 1000 cycles at 20 MW/m2 , the IR camera showed that the surface temperature was more than 1800 ◦ C, which is above the recrystallization temperature of the rolled tungsten. It is known
66
Z. Sun et al. / Fusion Engineering and Design 121 (2017) 60–69
Fig. 11. (a) UT results of Cu/CuCrZr interface of 3 mock-ups; (b) UT result of W/Cu interface of WTC-13. The loaded area was indicated.
Fig. 12. Light microscope image of cross-section through the loaded surfaces of grain microstructure (a) original rolled tungsten; (b) cut-2-2; (c) cut-2-3; (d) cut-5-4.
Z. Sun et al. / Fusion Engineering and Design 121 (2017) 60–69
67
For WTC-5, abnormal huge grain (about 2.6 mm) was found at the surface of the monoblock, which was related to the melting zone of the tungsten at surface. From Fig. 12(d) it can be seen that the recrystallization has extended to the interface of the W/Cu. 3.4. Vickers hardness The Vickers hardness (HV 0.5) measurement was performed to reveal microstructure and recrystallization conditions. Experience showed that the Vickers hardness will decrease when recrystallization occurs on rolled tungsten [1] and the test confirmed it. The result showed that the average hardness of original W is about HV 460 (cut-2-7), while the HHF tested W is changed from HV 380 to HV 460 with depth increasing from the loaded surface to the Cu interface which is related to the recrystallization depth of cut-2-3 (Fig. 13). 3.5. The damage of CuCrZr tube Fig. 13. Vickers’ hardness with the depth of cut-2-3.
For WTC-5, the damage of CuCrZr tube by swirl tape can be seen clearly (Fig. 14). During CHF testing, the temperature of CuCrZr tube could be high than its’ soft temperature which is around 450 ◦ C and the tube became soft. Then there will be the drastic friction between swirl tape and tube under the high water coolant velocity and pressure. Finally the tube got damaged by swirl tape. 4. Discussion Fig. 14. CuCrZr tube’s damage induced by swirl tape.
that the strength of recrystallized W is lower than that of stress released. To understand more detail of the recrystallization condition of W, the monoblocks were cut and polished then etched to observe the grain structure by optical microscope. Fig. 12(a) shows the original grain structure of the rolled W. From the other images of Fig. 12 it can be seen that recrystallization occurred on the loaded side of the W monoblock. In recrystallized regions the isotropic grain replaced the original elongated grain and recrystallized depth of the cut-2-2 and cut-2-3 was 1.7 mm and 2.3 mm respectively.
Recrystallization is one of the important features of W under high heat loads. Indeed the heat affected (recrystallized) zone was observed in deep area (a few mm) in the cross section. The depth of recrystallized zone is related to the distribution of temperature. To learn more details of temperature distribution of the monoblocks, finite element simulation was performed using ANSYS code. The used 3D model of monoblock geometry was set the same as mentioned in Section 2. All the materials were considered as elasto-plastic behavior and the temperature dependent properties were quoted from ITER material database. Hexahedron mesh was used in simulation and the mesh size was 0.5 mm. The properties of stress-relieved W was used in the model. The cooling
Fig. 15. Temperature distribution of monoblocks by ANSYS code.
68
Z. Sun et al. / Fusion Engineering and Design 121 (2017) 60–69
Fig. 16. Thermal stress fields of W at (a): heating stage (10 s) and (b): cooling stage (20 s).
parameters were as follows: coolant temperature at 70 ◦ C, heat transfer coefficient for pipe with swirl tape 110 kW/K/m2 [11]. Fig. 15 shows distribution of temperature, the highest temperature of the monoblock was 2098 ◦ C located at the edge of the W surface loaded at 20 MW/m2 . The average temperature at the middle of the monoblock’s surface is 1800 ◦ C, which is similar to the value measured from IR camera. It was found that the recrystallization
corresponded to the area above 1300 ◦ C. For copper and CuCrZr, the highest temperature was 572 ◦ C and 495 ◦ C separately. W monoblocks tested at 20 MW/m2 showed non-systematic but frequent appearance of macro-cracks. According to the morphology of the crack and fracture interface, the preliminary conclusion is that the crack initiated at the surface [12]. Fig. 16(a) and (b) shows the thermal stress fields (Cartesian normal component in
Z. Sun et al. / Fusion Engineering and Design 121 (2017) 60–69
the x direction) produced in the W block in the thermal loading stage (10 s) and the cooling stage (20 s) at 20 MW/m2 . A steady state structural analysis was used to simulate the mechanical properties of W and the behavior was set elasto-plastic. The boundary conditions were as follows. Fixed constraint was applied to the end of the tube and a pressure of 3.9 MPa was applied perpendicularly to the CuCrZr tube inner surface.The residual stress caused by HIPing progress was also considered. Thus, the stress free temperature was set to 600 ◦ C which corresponds to HIPing temperature. It is important to emphasize that the joint shall be validated by experiment but not by analysis [8]. So in the article only W surface stress was considered. From the analysis we can induce that in the heating stage, the surface of W suffers compressive stress and occurs plastic deformation. In the end of cooling stage, the surface of W suffers tensile stress, the largest tensile stress occurs at the center of the surface of along toroidal edges and the value is about 980 MPa. This is consistent with the location of macro-cracks as discussed also in [8,12]. The reader may pay attention to the fact that this calculation do not account for change in the material properties due to recrystallization. Indeed, recrystallization induces a decrease of the yield strength. The values of stress described below are thus applicable at the very beginning of the high heat flux test, when the impact of temperature is negligible on the tungsten structure. The material properties as a function of heat flux time exposure due to recrystallization. The accurate description of the thermos-mechanical behavior of the tungsten monoblock may take this phenomenon into account. Additionally, creep in high temperature, recrystallization and thermal fatigue (crack occurs after several hundred cycles at 20 MW/m2 ) can influence the occurrence of macro-crack. 5. Conclusion Accompanying with the application of full tungsten divertor on the EAST Tokamak, ASIPP manufactured small-scale mock-ups as part of full-W divertor qualification program for ITER. All 6 mockups made from ASIPP were tested at IDTF successfully and met the IO requirements. After HHF tests, post analyses were performed to validate the damage of mock-ups and the following characteristics were observed. • Mock-ups withstood HHF tests of 5000 cycles at 10 MW/m2 and 300 cycles at 20 MW/m2 without significant increase of steady state surface temperature along the cycling and withstood additional 700 cycles at 20 MW/m2 . CHF test showed that local critical heat flux of mock-ups is about 37–39 MW/m2 . • Both ultrasonic test and destructive test results showed that Cu/CuCrZr debonding appeared on some of monoblocks after 1000 cycles at 20 MW/m2 . No defects at W/Cu interfaces were detected after the HHF test. • W monoblocks showed non-systematic but frequent appearance of macro-cracks after HHF test. Plastic deformation can be found at the top surface of monoblocks. Thermal fatigue, creep damage and degradation of material are the main reason for the crack initiation. • Recrystallization is one of the important features of tungsten under high heat flux. The depth of recrystallization is up to 1.7–2.3 mm at the center of monoblocks which correspond to the temperature above 1300 ◦ C during heating phase.
69
• Appearance of large thermal stresses at the center of loaded surface was consistent with the location of macro-cracks in monoblocks. Disclaimer The view and opinion expressed herein do not necessarily reflect those of the ITER Organization. Acknowledgments This work is partially funded by the National Natural Science Foundation of China under Contract Nos. 11305213 and 11575242, the National Magnetic Confinement Fusion Science Program under contract Nos. 2013GB105001 and 2013GB105002, Chinese Academy of Sciences maintenance and renovation program of Mega-Project of Science: Maintenance and renovation project of upper divertors in EAST. References [1] T. Hirai, F. Escourbiac, V. Barabash, A. Durocher, A. Fedosov, L. Ferrand, T. Jokinen, V. Komarov, M. Merola, S. Carpentier-Chouchana, N. Arkhipov, V. Kuznetcov, A. Volodin, S. Suzuki, K. Ezato, Y. Seki, B. Riccardi, M. Bednarek, P. Gavila, Status of technology R&D for the ITER tungsten divertor monoblock, J. Nucl. Mater. 463 (2015) 1248–1251. [2] T. Hirai, F. Escourbiac, S. Carpentier-Chouchana, A. Durocher, A. Fedosov, L. Ferrand, T. Jokinen, V. Komarov, M. Merola, R. Mitteau, R.A. Pitts, W. Shu, M. Sugihara, V. Barabash, V. Kuznetsov, B. Riccardi, S. Suzuki, ITER full tungsten divertor qualification program and progress, Phys. Scr. T159 (2014) 014006. [3] T. Hirai, F. Escourbiac, S. Carpentier-Chouchana, A. Fedosov, L. Ferrand, T. Jokinen, V. Komarov, A. Kukushkin, M. Merola, R. Mitteau, R.A. Pitts, W. Shu, M. Sugihara, B. Riccardi, S. Suzuki, R. Villari, ITER tungsten divertor design development and qualification program, Fusion Eng. Des. 88 (2013) 1798–1801. [4] K. Ezato, S. Suzuki, Y. Seki, H. Yamada, T. Hirayama, K. Yokoyama, F. Escourbiac, T. Hirai, Progress of ITER full tungsten divertor technology qualification in Japan: manufacturing full-scale plasma-facing unit prototypes, Fusion Eng. Des. (2016). [5] P. Gavila, B. Riccardi, S. Constans, J.L. Jouvelot, I.B. Vastra, M. Missirlian, M. Richou, High heat flux testing of mock-ups for a full tungsten ITER divertor, Fusion Eng. Des. 86 (2011) 1652–1655. [6] V. Kuznetsov, A. Gorbenko, V. Davydov, A. Kokoulin, A. Komarov, I. Mazul, B. Mudyugin, I. Ovchinnikov, N. Stepanov, R. Rulev, A. Volodin, Status of the IDTF high-heat-flux test facility, Fusion Eng. Des. 89 (2014) 955–959. [7] P. Gavila, B. Riccardi, G. Pintsuk, G. Ritz, V. Kuznetsov, A. Durocher, High heat flux testing of EU tungsten monoblock mock-ups for the ITER divertor, Fusion Eng. Des. 98–99 (2015) 1305–1309. [8] T. Hirai, et al., Use of tungsten material for the ITER divertor, Nucl. Mater. Energy (2017), http://dx.doi.org/10.1016/j.nme.2016.07.003 (in press). [9] G. Pintsuk, M. Bednarek, P. Gavila, S. Gerzoskovitz, J. Linke, P. Lorenzetto, B. Riccardi, F. Escourbiac, Characterization of ITER tungsten qualification mock-ups exposed to high cyclic thermal loads, Fusion Eng. Des. 98–99 (2015) 1384–1388. [10] Q. Li, S. Qin, W. Wang, P. Qi, S. Roccella, E. Visca, G. Liu, G.-N. Luo, Manufacturing and testing of W/Cu mono-block small scale mock-up for EAST by HIP and HRP technologies, Fusion Eng. Des. 88 (2013) 1808–1812. [11] J. Thom, W. Walker, T. Fallon, G. Reising, Boiling in sub-cooled water during flow up heated tubes or annuli, Proc. Inst. Mech. Eng. (Lond.) 180 (Pt. 3C) (1965–1966) 226–246. [12] S. Panayotis, et al., Fracture modes of ITER tungsten divertor monoblock under stationary thermal loads, Fusion Eng. Des. (2017) (in review).
Contents lists available at ScienceDirect
Fusion Engineering and Design journal homepage: www.elsevier.com/locate/fusengdes
Post examination of tungsten monoblocks subjected to high heat flux tests of ITER full-tungsten divertor qualification program Zhaoxuan Sun a,b , Qiang Li a,∗ , Wanjing Wang a , Ji-Chao Wang a,b , Xingli Wang a,b , Ran Wei a,b , Chunyi Xie a , Guang-nan Luo a,b,c,d , T. Hirai e , F Escourbiac e , S. Panayotis e a
Institute of Plasma Physics, Chinese Academy of Sciences, Hefei, 230031, China Science Island Branch of Graduate School, University of Science & Technology of China, Hefei, 230031, China c Hefei Center for Physical Science and Technology, Hefei, 230022, China d Hefei Science Center of Chinese Academy of Sciences, Hefei, 230027, China e ITER Organization,Route de Vinon sur Verdon, CS 90 046 13067 Saint Paul lez Durance Cedex, France b
h i g h l i g h t s • • • •
ASIPP manufactured six W monoblock mock-ups that were tested at IDTF for full-tungsten divertor qualification program. Ultrasonic test was performed to investigate the defects of interface. Destructive test was performed on three mock-ups to analysis the damage. FEA was performed to study the temperature and stress distribution of monoblocks.
a r t i c l e
i n f o
Article history: Received 21 December 2016 Received in revised form 30 March 2017 Accepted 6 June 2017 Keywords: ITER Tungsten Divertor High heat flux Monoblock
a b s t r a c t In 2015, as part of ITER Full-Tungsten Divertor Qualification Program, Institute of Plasma Physics Chinese Academy of Sciences (ASIPP) manufactured six small-scale Tungsten (W) monoblock mock-ups that were tested at the electron beam facility, ITER Divertor Test Facility (IDTF, St Petersburg, RF). The high heat flux (HHF) tests consisted of 5000 cycles at 10 MW/m2 , followed by 300 cycles at 20 MW/m2 and additional 700 cycles at 20 MW/m2 . All mock-ups fulfilled the requirements of high heat flux performance successfully. One (WTC-5) of the six mock-ups was then selected to carry out critical heat flux (CHF) test with local heat flux up to 37–39 MW/m2 . After HHF test, ultrasonic test (UT) and destructive tests were performed to characterize the damages of monoblocks. The debonding of the Cu/CuCrZr interface was detected by both nondestructive and destructive tests. Intergranular rupture of W was observed by Scanning Electron Microscope (SEM). The recrystallization was found in the W monoblocks and the recrystallized depth were analyzed by metallography and Vickers hardness measurement. © 2017 Published by Elsevier B.V.
1. Introduction Full tungsten (W) divertor qualification program have been established to develop and validate the performance of W monoblock technology for W divertor in November 2011 [1–3]. Upon the two years of efforts on engineering and physics aspects, the ITER Council endorsed the proposal to start ITER operation with full-W divertor in November 2013. Domestic Agencies in Japan and Europe have succeeded to demonstrate high heat flux (HHF) performance of W monoblocks for a full-tungsten divertor in colFig. 1. Tungsten monoblock mock-ups with swirl tapes before HHF testing. ∗ Corresponding author. E-mail address: [email protected] (Q. Li). http://dx.doi.org/10.1016/j.fusengdes.2017.06.009 0920-3796/© 2017 Published by Elsevier B.V.
Z. Sun et al. / Fusion Engineering and Design 121 (2017) 60–69
61
Fig. 2. IR-pictures of average absorbing heat flux at (a)-24.7 MW/m2 , (b)-25.5 MW/m2 , (c)-26 MW/m2 .
laboration with the ITER Organization (IO) [4,5]. In China, Institute of Plasma Physics Chinese Academy of Sciences (ASIPP) has set up to develop W monoblocks technology since 2010 and applied it to the upper divertor on the EAST Tokamak. In 2015, as a part of qualification of tungsten (W) monoblock technology for ITER divertor, ASIPP and Advanced Technology and Materials Co., Ltd (AT&M) manufactured six small-scale W monoblock mock-ups that were sent to Fig. 3. Possible distribution made by the simulation.
Fig. 4. Average temperature of each monoblock gained by IR-camera of WTC-2 HHF testing: (a)-5000 cycles at 10 MW/m2 , (b)-300 cycles at 20 MW/m2 , (c)-additional 700 cycles at 20 MW/m2 .
62
Z. Sun et al. / Fusion Engineering and Design 121 (2017) 60–69
Fig. 5. Heated surface of 5 mock-ups after 5000 cycles at 10 MW/m2 and 1000 cycles at 20 MW/m2 . The coolant water flow direction is indicated.
Russia for the HHF testing. All mock-ups passed the examinations successfully and met the IO requirements. After HHF test, analyses were performed to evaluate the damages of the mock-ups. The analyses consisted of ultrasonic test to examine the interface of W/Cu and Cu/CuCrZr and destructive test to investigate microstructure of monoblocks. Destructive tests were performed on three mock-ups. Both ultrasonic tests and destructive tests detected the defects of Cu/CuCrZr interface. The depth of cracks and recrystallization were also analyzed to learn more details of the monoblocks. Vickers hardness was also carried out on W monoblocks to validate the relation between recrystallization and Vickers hardness. 2. Mock-up manufacturing and high heat flux testing All small-scale W monoblock mock-ups for high heat flux testing were manufactured by Hot Isostatic Pressing (HIP) technology, which is developed by ASIPP and AT&M. The HIPing conditions were an isostatic pressure of 100 MPa, the heating temperature of 600 ◦ C and the heating time of 2 h. The W monoblocks were cut from rolled W plates and the rolled direction is perpendicular to the loaded surface. Each mock-up consists of 7 W monoblocks separated by a gap of 0.4 mm. between the neighbor monoblocks a molybdenum spacer (15/19 mm (ID/OD)) was placed to maintain the gaps during HIP. The W monoblocks have a width of 28 mm, height of 26 mm and the axial length of 12 mm. The armor thickness (minimum distance between the heat load surface and interlayer (pure copper UNS C11010) is 6 mm. The functional material of pure Cu (about 1 mm thick) is used as interlayer to reduce the stress caused by the differences of W and CuCrZr thermal expansion. The CuCrZr tube manufactured by State Nuclear Bao Ti Zirconium industry, is used as heat sink material for its good thermomechanical behavior. The tube diameter is 12/15 mm (ID/OD). A copper twisted tape, with a
twist ratio of 2 and a thickness of 2 mm, was inserted into the tube to enhance convective heat transfer. The W material, CuCrZr tube and twisted tape were all supplied by AT&M. All mock-ups were examined by ultrasonic test and didn’t show any defects on joints before HHF test. Fig. 1 shows the mock-ups with swirl tapes before HHF testing. The HHF testing was carried out at IDTF, which was designed for HHF testing on the ITER inner and outer vertical targets and domes plasma-facing units [6]. The HHF test consisted of 5000 cycles at 10 MW/m2 , 300 cycles at 20 MW/m2 and additional 700 cycles at 20 MW/m2 . The heat loads duration were set-up at 10 s on and 10 s off. For each mock-up, four out of seven monoblocks were tested. The surface temperature was measured by IR camera and pyrometers, the absorbed heat flux was obtained by global calorimetry from measurement of temperature from thermocouples mounted at the water coolant inlet and the outlet. The hydraulic conditions were set at ITER relevant conditions. The used parameters were a pressure of 3.9 MPa, the water coolant inlet temperature of 70 ◦ C and the mass flow rate of ∼3 kg/s (equivalent the velocity of ∼11 m/s). After 1000 cycles at 20 MW/m2 applied to all the mockups, one mock-up (WTC-5) was selected to carry out the critical heat flux (CHF) testing to evaluate the critical heat flux of the mockups. When the average heat flux increased up to 24.7 MW/m2, overheating of the masks started to occur. The electron-beam raster was modified with the objective to avoid the copper masks’ damage. This change of raster generated non-uniform heating of the mock-up as shown in Fig. 2. At higher heat fluxes, due to the X-ray matrix’s saturation at this stage of experiment verification of heat flux profile was not feasible. Then the finite element analysis was performed according to the temperature gained by IR-camera to simulate the distribution of heat flux when average heat flux was 26 MW/m2 as shown in Fig. 3.
Z. Sun et al. / Fusion Engineering and Design 121 (2017) 60–69
63
Fig. 6. Surface morphology and cut-5-4 of WTC-5 after CHF test. The cutting line for metallographic examination is indicated as 5-4.
Fig. 7. The cutting line and numbering of WTC-2 (a) and WTC-9 (b).
All mock-ups passed the HHF test successfully in the sense that no significant increase of steady state surface temperature along the cycling was observed. Thus met the IO requirements (<30% surface temperature increase in 5000 cycles at 10 MW/m2 and 300 cycles at 20 MW/m2 ). Fig. 4 was the average temperature during HHF testing of four monoblocks (monoblock 1–4) of WTC-2. From this picture it can be seen that the surface temperature increases more at the additional 700 cycles. But it doesn’t exceed 30% of the surface temperature at start of the test at 20 MW/m2 . However, some degradation was found in the mock-ups. After the 5000 cycles at 10 MW/m2 , no visible damage was observed. After 300 cycles at 20 MW/m2 , the surface of the monoblocks became coarse and a longitudinal crack, so-called macro-cracks (self-castellation) appeared at the center of two monoblocks (among 6 × 4 = 24 subjected to the heat flux). After additional 700 cycles of test at 20 MW/m2 , longitudinal cracks can be seen in 9 of 24 monoblocks in all the 6 mock-ups. Fig. 5 shows the surface morphology of 5 mock-ups after 1000 cycles at 20 MW/m2 test. At the loaded surface, roughening and cracks were observed. Crack formation at the middle of monoblocks was also found in the mock-ups made by European industry [1,7,8]. Fig. 6 shows the surface morphology of the WTC-5 after Critical Heat Flux (CHF) test. The melting on loaded surface of the tungsten can be seen clearly. 3. Post test analyses After HHF test, destructive examinations were performed on mock-up WTC-2, WTC-5 and WTC-9. The cutting line and numbering rule of WTC-2 and WTC-9 are shown in Fig. 7. For WTC-5,
Fig. 8. Optical microscope image of (a) cut-2-2 and (b) cut-2-5.
one monoblock (namely cut-5-4,) was cut to find the details of microstructure.
3.1. Crack formation in W monoblocks and fracture interface of the cracks One to two macro-cracks were found on all of the mock-ups. For the WTC-2, two of four tested monoblocks formed a longitudinal crack (cut-2-2 and cut2-5). Fig. 8 show the crack formation of cut2-2 (a) and cut-2-5 (b) by optical microscope. It can be seen from the figure that a single crack propagates from the loaded surface further into W and even reached to the W/Cu interface. The width
64
Z. Sun et al. / Fusion Engineering and Design 121 (2017) 60–69
Fig. 9. SEM image of monoblock’s fracture surfaces. (a) Near loaded side; (b) near Cu side.
of the crack in the W heated surface is larger than that in the W/Cu interface side. Fig. 9 shows the fracture surface of the top side and near Cu interlayer of crack (cut-2-2). The fracture surface structure showed boulder-strewn structure which indicated that the main mechanism of fracture was intergranular rupture. Plastic deformation can be seen in the top side of the monoblock as it was observed on European monoblocks [9]. Recrystallized grain can be found in loaded side of monoblocks (Fig. 9(a)) and the microstructure near the Cu interlayer showed rapture surface reflecting the original grain orientation (Fig. 9(b)). 3.2. Interface Fig. 10 shows the images of the local Cu/CuCrZr interfaces. The destructive analysis showed that the debonding concentrated on the interface of Cu/CuCrZr. Two defects (about 1 mm) were seen
for cut-2-2 while cut-2-5 has three discrete defects, although no defects were detected by ultrasonic inspection before HHF test. The debonding of Cu/CuCrZr interface was located at the loaded side of heat sink. The defects were not found in W/Cu interface. For WTC-9, no defects were found at Cu/CuCrZr as well as W/Cu interfaces in all the examined monoblocks. For WTC-5, melting occurred on both copper interlayer and loaded W surface under CHF test. The joint area was destroyed and voids were found at the W/Cu interface. At the heat flux of 20 MW/m2, there exists a large temperature gradient near the interface of the Cu and the CuCrZr. This large temperature gradient can cause a big thermal stress and plastic strain at the Cu/CuCrZr interface. As the HHF test continue, the Cu/CuCrZr interface would suffer from fatigue damage due to the cyclic thermal stress and plastic strain. Finally, the debonding occurred. The other three mock-ups’ interface was inspected by Ultrasonic Test (UT). Ultrasonic inspection was performed scanning the probe in the tube over the inner surface of tube immersed into water [10].
Z. Sun et al. / Fusion Engineering and Design 121 (2017) 60–69
65
Fig. 10. SEM image of Cu/CuCrZr interface. (a): cut-2-2; (b): cut-2-5; (c) cut-5-4.
The frequency of the ultrasonic was 15 MHz. The test was focus on the interface of W/Cu and Cu/CuCrZr. Fig. 11(a) shows the UT results of Cu/CuCrZr interfaces. The 0◦ position corresponds to the 6 o’clock position in the monoblock. The amplitudes of the echo signals of the three mock-ups’ Cu/CuCrZr interface at specific zones are much higher (red zone) means that there were defects occurred at these locations. All the defects occurred at the loaded area as indicated
(black box area) in the picture. No defects were detected at W/Cu interface for all 3 mock-ups like WTC-13 in Fig. 11(b). 3.3. Recrystallization of W monoblocks During the 1000 cycles at 20 MW/m2 , the IR camera showed that the surface temperature was more than 1800 ◦ C, which is above the recrystallization temperature of the rolled tungsten. It is known
66
Z. Sun et al. / Fusion Engineering and Design 121 (2017) 60–69
Fig. 11. (a) UT results of Cu/CuCrZr interface of 3 mock-ups; (b) UT result of W/Cu interface of WTC-13. The loaded area was indicated.
Fig. 12. Light microscope image of cross-section through the loaded surfaces of grain microstructure (a) original rolled tungsten; (b) cut-2-2; (c) cut-2-3; (d) cut-5-4.
Z. Sun et al. / Fusion Engineering and Design 121 (2017) 60–69
67
For WTC-5, abnormal huge grain (about 2.6 mm) was found at the surface of the monoblock, which was related to the melting zone of the tungsten at surface. From Fig. 12(d) it can be seen that the recrystallization has extended to the interface of the W/Cu. 3.4. Vickers hardness The Vickers hardness (HV 0.5) measurement was performed to reveal microstructure and recrystallization conditions. Experience showed that the Vickers hardness will decrease when recrystallization occurs on rolled tungsten [1] and the test confirmed it. The result showed that the average hardness of original W is about HV 460 (cut-2-7), while the HHF tested W is changed from HV 380 to HV 460 with depth increasing from the loaded surface to the Cu interface which is related to the recrystallization depth of cut-2-3 (Fig. 13). 3.5. The damage of CuCrZr tube Fig. 13. Vickers’ hardness with the depth of cut-2-3.
For WTC-5, the damage of CuCrZr tube by swirl tape can be seen clearly (Fig. 14). During CHF testing, the temperature of CuCrZr tube could be high than its’ soft temperature which is around 450 ◦ C and the tube became soft. Then there will be the drastic friction between swirl tape and tube under the high water coolant velocity and pressure. Finally the tube got damaged by swirl tape. 4. Discussion Fig. 14. CuCrZr tube’s damage induced by swirl tape.
that the strength of recrystallized W is lower than that of stress released. To understand more detail of the recrystallization condition of W, the monoblocks were cut and polished then etched to observe the grain structure by optical microscope. Fig. 12(a) shows the original grain structure of the rolled W. From the other images of Fig. 12 it can be seen that recrystallization occurred on the loaded side of the W monoblock. In recrystallized regions the isotropic grain replaced the original elongated grain and recrystallized depth of the cut-2-2 and cut-2-3 was 1.7 mm and 2.3 mm respectively.
Recrystallization is one of the important features of W under high heat loads. Indeed the heat affected (recrystallized) zone was observed in deep area (a few mm) in the cross section. The depth of recrystallized zone is related to the distribution of temperature. To learn more details of temperature distribution of the monoblocks, finite element simulation was performed using ANSYS code. The used 3D model of monoblock geometry was set the same as mentioned in Section 2. All the materials were considered as elasto-plastic behavior and the temperature dependent properties were quoted from ITER material database. Hexahedron mesh was used in simulation and the mesh size was 0.5 mm. The properties of stress-relieved W was used in the model. The cooling
Fig. 15. Temperature distribution of monoblocks by ANSYS code.
68
Z. Sun et al. / Fusion Engineering and Design 121 (2017) 60–69
Fig. 16. Thermal stress fields of W at (a): heating stage (10 s) and (b): cooling stage (20 s).
parameters were as follows: coolant temperature at 70 ◦ C, heat transfer coefficient for pipe with swirl tape 110 kW/K/m2 [11]. Fig. 15 shows distribution of temperature, the highest temperature of the monoblock was 2098 ◦ C located at the edge of the W surface loaded at 20 MW/m2 . The average temperature at the middle of the monoblock’s surface is 1800 ◦ C, which is similar to the value measured from IR camera. It was found that the recrystallization
corresponded to the area above 1300 ◦ C. For copper and CuCrZr, the highest temperature was 572 ◦ C and 495 ◦ C separately. W monoblocks tested at 20 MW/m2 showed non-systematic but frequent appearance of macro-cracks. According to the morphology of the crack and fracture interface, the preliminary conclusion is that the crack initiated at the surface [12]. Fig. 16(a) and (b) shows the thermal stress fields (Cartesian normal component in
Z. Sun et al. / Fusion Engineering and Design 121 (2017) 60–69
the x direction) produced in the W block in the thermal loading stage (10 s) and the cooling stage (20 s) at 20 MW/m2 . A steady state structural analysis was used to simulate the mechanical properties of W and the behavior was set elasto-plastic. The boundary conditions were as follows. Fixed constraint was applied to the end of the tube and a pressure of 3.9 MPa was applied perpendicularly to the CuCrZr tube inner surface.The residual stress caused by HIPing progress was also considered. Thus, the stress free temperature was set to 600 ◦ C which corresponds to HIPing temperature. It is important to emphasize that the joint shall be validated by experiment but not by analysis [8]. So in the article only W surface stress was considered. From the analysis we can induce that in the heating stage, the surface of W suffers compressive stress and occurs plastic deformation. In the end of cooling stage, the surface of W suffers tensile stress, the largest tensile stress occurs at the center of the surface of along toroidal edges and the value is about 980 MPa. This is consistent with the location of macro-cracks as discussed also in [8,12]. The reader may pay attention to the fact that this calculation do not account for change in the material properties due to recrystallization. Indeed, recrystallization induces a decrease of the yield strength. The values of stress described below are thus applicable at the very beginning of the high heat flux test, when the impact of temperature is negligible on the tungsten structure. The material properties as a function of heat flux time exposure due to recrystallization. The accurate description of the thermos-mechanical behavior of the tungsten monoblock may take this phenomenon into account. Additionally, creep in high temperature, recrystallization and thermal fatigue (crack occurs after several hundred cycles at 20 MW/m2 ) can influence the occurrence of macro-crack. 5. Conclusion Accompanying with the application of full tungsten divertor on the EAST Tokamak, ASIPP manufactured small-scale mock-ups as part of full-W divertor qualification program for ITER. All 6 mockups made from ASIPP were tested at IDTF successfully and met the IO requirements. After HHF tests, post analyses were performed to validate the damage of mock-ups and the following characteristics were observed. • Mock-ups withstood HHF tests of 5000 cycles at 10 MW/m2 and 300 cycles at 20 MW/m2 without significant increase of steady state surface temperature along the cycling and withstood additional 700 cycles at 20 MW/m2 . CHF test showed that local critical heat flux of mock-ups is about 37–39 MW/m2 . • Both ultrasonic test and destructive test results showed that Cu/CuCrZr debonding appeared on some of monoblocks after 1000 cycles at 20 MW/m2 . No defects at W/Cu interfaces were detected after the HHF test. • W monoblocks showed non-systematic but frequent appearance of macro-cracks after HHF test. Plastic deformation can be found at the top surface of monoblocks. Thermal fatigue, creep damage and degradation of material are the main reason for the crack initiation. • Recrystallization is one of the important features of tungsten under high heat flux. The depth of recrystallization is up to 1.7–2.3 mm at the center of monoblocks which correspond to the temperature above 1300 ◦ C during heating phase.
69
• Appearance of large thermal stresses at the center of loaded surface was consistent with the location of macro-cracks in monoblocks. Disclaimer The view and opinion expressed herein do not necessarily reflect those of the ITER Organization. Acknowledgments This work is partially funded by the National Natural Science Foundation of China under Contract Nos. 11305213 and 11575242, the National Magnetic Confinement Fusion Science Program under contract Nos. 2013GB105001 and 2013GB105002, Chinese Academy of Sciences maintenance and renovation program of Mega-Project of Science: Maintenance and renovation project of upper divertors in EAST. References [1] T. Hirai, F. Escourbiac, V. Barabash, A. Durocher, A. Fedosov, L. Ferrand, T. Jokinen, V. Komarov, M. Merola, S. Carpentier-Chouchana, N. Arkhipov, V. Kuznetcov, A. Volodin, S. Suzuki, K. Ezato, Y. Seki, B. Riccardi, M. Bednarek, P. Gavila, Status of technology R&D for the ITER tungsten divertor monoblock, J. Nucl. Mater. 463 (2015) 1248–1251. [2] T. Hirai, F. Escourbiac, S. Carpentier-Chouchana, A. Durocher, A. Fedosov, L. Ferrand, T. Jokinen, V. Komarov, M. Merola, R. Mitteau, R.A. Pitts, W. Shu, M. Sugihara, V. Barabash, V. Kuznetsov, B. Riccardi, S. Suzuki, ITER full tungsten divertor qualification program and progress, Phys. Scr. T159 (2014) 014006. [3] T. Hirai, F. Escourbiac, S. Carpentier-Chouchana, A. Fedosov, L. Ferrand, T. Jokinen, V. Komarov, A. Kukushkin, M. Merola, R. Mitteau, R.A. Pitts, W. Shu, M. Sugihara, B. Riccardi, S. Suzuki, R. Villari, ITER tungsten divertor design development and qualification program, Fusion Eng. Des. 88 (2013) 1798–1801. [4] K. Ezato, S. Suzuki, Y. Seki, H. Yamada, T. Hirayama, K. Yokoyama, F. Escourbiac, T. Hirai, Progress of ITER full tungsten divertor technology qualification in Japan: manufacturing full-scale plasma-facing unit prototypes, Fusion Eng. Des. (2016). [5] P. Gavila, B. Riccardi, S. Constans, J.L. Jouvelot, I.B. Vastra, M. Missirlian, M. Richou, High heat flux testing of mock-ups for a full tungsten ITER divertor, Fusion Eng. Des. 86 (2011) 1652–1655. [6] V. Kuznetsov, A. Gorbenko, V. Davydov, A. Kokoulin, A. Komarov, I. Mazul, B. Mudyugin, I. Ovchinnikov, N. Stepanov, R. Rulev, A. Volodin, Status of the IDTF high-heat-flux test facility, Fusion Eng. Des. 89 (2014) 955–959. [7] P. Gavila, B. Riccardi, G. Pintsuk, G. Ritz, V. Kuznetsov, A. Durocher, High heat flux testing of EU tungsten monoblock mock-ups for the ITER divertor, Fusion Eng. Des. 98–99 (2015) 1305–1309. [8] T. Hirai, et al., Use of tungsten material for the ITER divertor, Nucl. Mater. Energy (2017), http://dx.doi.org/10.1016/j.nme.2016.07.003 (in press). [9] G. Pintsuk, M. Bednarek, P. Gavila, S. Gerzoskovitz, J. Linke, P. Lorenzetto, B. Riccardi, F. Escourbiac, Characterization of ITER tungsten qualification mock-ups exposed to high cyclic thermal loads, Fusion Eng. Des. 98–99 (2015) 1384–1388. [10] Q. Li, S. Qin, W. Wang, P. Qi, S. Roccella, E. Visca, G. Liu, G.-N. Luo, Manufacturing and testing of W/Cu mono-block small scale mock-up for EAST by HIP and HRP technologies, Fusion Eng. Des. 88 (2013) 1808–1812. [11] J. Thom, W. Walker, T. Fallon, G. Reising, Boiling in sub-cooled water during flow up heated tubes or annuli, Proc. Inst. Mech. Eng. (Lond.) 180 (Pt. 3C) (1965–1966) 226–246. [12] S. Panayotis, et al., Fracture modes of ITER tungsten divertor monoblock under stationary thermal loads, Fusion Eng. Des. (2017) (in review).