- Email: [email protected]
An experimental study on layer inversion in the corium pool during a severe accident
Nuclear Engineering and Design 278 (2014) 163–170
Contents lists available at ScienceDirect
Nuclear Engineering and Design journal homepage: www.elsevier.com/locate/nucengdes
An experimental study on layer inversion in the corium pool during a severe accident Kyoung-Ho Kang ∗ , Rae-Joon Park, Seong-Ho Hong, Seong-Wan Hong, Kwang Soon Ha Korea Atomic Energy Research Institute, 1045 Daedeok-daero, Yuseong-Gu, Daejeon 305-353, Republic of Korea
h i g h l i g h t s • • • • •
COSMOS tests were performed to investigate the layer inversion of the corium pool using prototypic materials. An induction heating method was implemented for melting of the UO2 –ZrO2 –Zr–Fe mixture. The simulated sequence was the TLFW in APR1400. A metallurgical inspection was performed to investigate the layer inversion. COSMOS test results show the possibility of layer inversion of the heavy metallic material.
a r t i c l e
i n f o
Article history: Received 3 March 2014 Received in revised form 14 July 2014 Accepted 15 July 2014
a b s t r a c t COSMOS (COrium configuration of the molten State in the MOst Severe accidents) tests using prototypic materials have been performed to investigate the layer inversion of the heavy metallic material in the corium pool. An induction heating method using a cold crucible was implemented for melting of the UO2 –ZrO2 –Zr–Fe mixture. Before the main test, three preliminary tests have been performed to enhance the experimental techniques using the real core material. One main test of the COSMOS has been performed to evaluate the corium pool configuration in the lower plenum of the reactor vessel for the TLFW (Total Loss of Feed Water) sequence of the APR (Advanced Power Reactor) 1400 under the IVR-ERVC (InVessel corium Retention through External Reactor Vessel Cooling). A post-test examination of cutting, EPMA, and XRD for the solidified corium pool ingot in the main test has been performed to investigate the layer inversion. From the three preliminary tests, melting techniques of the real core material using a cold crucible were successfully developed. The metallurgical inspection results on the chemical information coincide with the visual observation on the centerline cut ingot in that the upper part is metal and the lower lump is an oxidic mixture with some metal clods in the main test. This means the possibility of layer inversion of the heavy metallic material in the corium pool. © 2014 Elsevier B.V. All rights reserved.
1. Introduction The melted core material relocates to the lower plenum of the reactor vessel during a severe accident and forms the corium pool under the IVR (In-Vessel corium Retention) through the ERVC (External Reactor Vessel Cooling), which is considered an effective means for maintaining the integrity of the reactor vessel of a nuclear power plant (Theofanous et al., 1996, 1997; Rempe et al., 2008). The IVR-ERVC has been adopted in low-power reactors of the AP600 and Loviisa nuclear power plants, and in a medium-power reactor of an AP1000 as a design feature for severe accident mitigation
∗ Corresponding author. Tel.: +82 42 868 2665; fax: +82 42 868 8256. E-mail address: [email protected] (K.-H. Kang). http://dx.doi.org/10.1016/j.nucengdes.2014.07.005 0029-5493/© 2014 Elsevier B.V. All rights reserved.
(Esmaili and Khatib-Rahbar, 2004; Dinh et al., 2003, 2004; Scobel et al., 2002; Kymalainean et al., 1997), and in high-power reactors of the APR (Advanced Power Reactor) 1400 and APR+ as an accident management strategy (KEPCO, 1998; Park et al., 2001, 2006). Fig. 1 shows the typical corium pool formation in the lower plenum of the reactor vessel under the IVR-ERVC. In general, a corium pool including metallic materials, such as stainless steel and Zircaloy, and oxide materials, such as UO2 and ZrO2, may form at the initial stage. The corium pool may be stratified into two layers, due to their density differences, an upper metallic layer without a volumetric heat source and a lower oxidic layer with a volumetric decay heat source, which is an original two-layer formation. The experimental results of the OECD MASCA showed that when a sufficient amount of non-oxidized zirconium (Zr) is available, metallic uranium (U) migrates to the metallic layer (Barrachin
164
K.-H. Kang et al. / Nuclear Engineering and Design 278 (2014) 163–170
Fig. 1. Corium pool formation in the lower plenum of the reactor vessel.
and Defoort, 2004). The transfer of species between the U, O, Zr melt and the steel can result in a significant density increase of the metallic layer. The density increase of the metallic layer can lead to inverse corium pool stratification with an additional heavy metal layer below the oxidic pool, which leads to three-layer formation as shown in Fig. 1. The presence of the metallic layer at the bottom of the lower plenum is likely to decrease the thickness of the top metallic layer and consequently increase the thermal load to the lower head vessel, which is the focusing effect in the top metallic layer. The melt pool configurations inside the lower head of the reactor vessel affect the initial thermal load to the vessel and play
a key role in determining the integrity of the reactor vessel under the IVR-ERVC. Thermodynamic analyses using the GEMINI code were performed to examine the final melt pool configuration at the reactor vessel failure during severe accidents in the APR1400 (Kang et al., 2010). As the representative accident scenarios, a LBLOCA (Large Break Loss of Coolant Accident) without SI (Safety Injection), MBLOCA (Medium Break Loss of Coolant Accident) without SI, SBLOCA (Small Break Loss of Coolant Accident) without SI, SBO (Station Black Out), and TLFW (Total Loss Feed Water) were selected from the level I PSA (Probabilistic Risk Assessment) results. Melt pool configurations were different in the representative accident sequences of the APR1400. In the cases of SBLOCA and LBLOCA where the U/Zr ratio and the initial melt pool temperature were relatively higher, layer-inversion phenomena can be precluded, which results in the original two-layer formation of the corium pool. In other cases of the TLFW, SBO, and MBLOCA, however, a three-layer formation possibility by layer inversion should be considered in evaluating the melt pool configuration and consequent thermal load from the melt pool to the reactor vessel wall (Kang et al., 2010). COSMOS (COrium configuration of the molten State in the MOst Severe accidents) tests have been performed for an investigation on the molten pool configurations considering the layer inversion of the heavy metallic material. In the COSMOS test, an induction heating method using a cold crucible was implemented for a melting of the prototypic UO2 –ZrO2 –Zr–Fe mixture. Three preliminary tests have been performed to enhance the experimental melting techniques for the real core material of the UO2 –ZrO2 –Zr–SS mixture. The main test of the COSMOS to investigate the melt pool configuration in the lower plenum of the reactor vessel has been performed for the TLFW sequence of the APR1400. According to the thermodynamic analysis results, three-layer formation by a layer inversion of heavy metallic material is expected for this melt configuration of the TLFW sequence. The main objectives of this test were not only to provide physical insight into the layer inversion phenomena of the heavy metallic material, but also to produce metallurgical inspection test data to validate the thermodynamic analysis methodology using the GEMINI code. 2. Description of the COSMOS facility and test conditions In the COSMOS test, induction heating method using a cold crucible was implemented for melting of the prototypic UO2 –ZrO2 –Zr–Fe mixture. The maximum electric power and frequency of the generator were 100 kW and 120 kHz, respectively. Fig. 2 shows a schematic diagram of the COSMOS test facility. Photographs of the test facility are shown in Fig. 3. The inner diameter and height of the crucible were 9.45 cm and 25 cm, respectively.
Fig. 2. Schematic diagram of the COSMOS test facility.
K.-H. Kang et al. / Nuclear Engineering and Design 278 (2014) 163–170
Fig. 3. Photograph of the COSMOS test section.
The temperature of the melt pool surface was measured using the two-colored Pyrometer (Model name: IRCON 3R-35C15-0-00-1). The measuring range of the two-colored Pyrometer is from 1500 ◦ C to 3500 ◦ C with a measurement error of 2.0% of the reading.
165
Metallurgical inspection methods such as EPMA (Electron Probe Micro Analyzer) and XRD (X-Ray Diffraction) were implemented to precisely examine the distribution and composition of the melt layer as a post-test examination. The metallurgical homogeneity and chemical compounds of a sampled ingot were analyzed by EPMA and XRD, respectively. Table 1 shows the test condition for the melt compositions in the COSMOS test. Three preliminary tests (CO-PR1, CO-PR2, and COPR3) have been performed to enhance the experimental techniques using a real core material. The main experiment (COSMOS-main) has been performed for the initial melt pool conditions in the lower plenum of the reactor vessel for the TLFW sequence of the APR1400 under the IVR-ERVC. According to the thermodynamic analysis results, a layer inversion of heavy metallic material was expected for this melt configuration of the TLFW sequence. To obtain the ideal melt mass, as shown in Table 1, a UO2 –ZrO2 –Zr–SS (stainless steel) mixture was charged, as shown in Fig. 4. Exact melting of the mixture in terms of the melt mass, however, is a very hard task in the test which implements a cold crucible induction heating method for melting of the prototypic UO2 –ZrO2 –Zr–SS mixture. For this reason, in the COSMOS-main test, actual melt compositions were different from those intended for the TLFW simulation. The actual melt compositions are shown in Table 2. It was found that the U/Zr ratio and Zr oxidation rate in the main test differed from those of the TLFW simulation in which a layer inversion of the heavy metallic material was expected. In Table 2, U/Zr and Cn stand for the molar ratio of Uranium to Zirconium, and the Zr oxidation ratio, respectively. The Zr oxidation rate (Cn) is defined as the oxidation rate of the corium by the molar ratio, ZrO2 /(Zr + ZrO2 ). The experimental results of the OECD MASCA showed that the U/Zr ratio and Zr oxidation rates are key parameters determining the possibility of the layer inversion of the heavy metallic material. The typical charging pattern for the COSMOS-main test is shown in Figs. 4 and 5. The thickness of the ZrO2 bottom liner was 35 mm and the mass was 0.63 kg. The UO2 pellets filled the circumference of the crucible to protect the fingers from an attack of the hot molten metal with a relatively lower melting temperature. In the middle part, the mixture of the UO2 pellets, ZrO2 powder, Zr and stainless steel metal chips were charged. The upper part was filled with the ZrO2 powder. The initiator made of a Zr metal ring, which enabled the charged mixture to be melted at an initial stage, was installed at the upper part. The test vessel protecting the charged cold crucible was purged with argon gas before induction heating, and argon gas was continuously supplied into the vessel at a flow rate of 10 l/min during the heating to prevent a rapid oxidation of the Zr metal chips which could cause a melt eruption. Another reason for the argon gas supply was to vent the aerosols during the melting period in order to measure the melt surface temperature more accurately using the two-colored Pyrometer. Variation of the applied power by the induction heating generator is shown in Fig. 6. The power level was increased step by step. The melting probably started at 770 s because the coupling factor of
Table 1 COSMOS test matrix. Test name
Composition
Ratio (kg/%)
Melt ratio (%)
Comments
CO-PR-01
ZrO2 /Fe/Zr
67.5
Preliminary test
CO-PR-02
ZrO2 /Fe/Zr
86.1
Preliminary test
CO-PR-03
UO2 /ZrO2 /Fe/Zr
40.0
Preliminary test
COSMOS-main
UO2 /ZrO2 /Fe/Zr
4.15/0.85/0.2 (81.3/16/7/2.1) 3.65/1.03/0.12 (76.1/21.5/2.4) 3.82/1.74/1.36/0.1 (54.5/24.8/19.3/1.4) 4.58/0.74/0.47/1.06 (72.8/11.8/7.5/7.9)
84.0
Main test
166
K.-H. Kang et al. / Nuclear Engineering and Design 278 (2014) 163–170
Fig. 4. Schematic diagram of the charging pattern in the main test. Table 2 Test conditions for the melt compositions in the main test.
Ideal corium Actual corium
UO2 –ZrO2 –Zr mass (kg)
SS mass (kg)
U/Zr
Cn
Comments
4.58/0.74/0.47 3.64/1.23/0.42
1.06 0.89
1.52 0.93
54 69
Target melt compositions Obtained melt compositions
Q decreased slowly from this time. After the applied power reached the maximum level of 63 kW, the power level was maintained for about 14 min and the power was turned off. The maximum absorbed power into the melt was estimated to be about 37.8 kW, which was 59.7% of the total maximum input power of 63 kW. The melting of the surface was confirmed by the visual observation in the COSMOS-main test. After the power was turned off, the melt was solidified in the cold crucible during the cooling process by the
Fig. 5. Inside photograph of the COSMOS test section.
Fig. 6. Variation of the applied power in the main test.
K.-H. Kang et al. / Nuclear Engineering and Design 278 (2014) 163–170
chiller actuation for the purpose of the investigation on the layer inversion phenomena. 3. Discussions on the experimental results The melting techniques of the real core material using a cold crucible were successfully developed from the three preliminary tests. The main test of the COSMOS was performed using these techniques in the TLFW sequence of the APR1400. A schematic diagram and photograph of the solidified corium in the crucible are shown in Figs. 7 and 8, respectively. The upper crust anchored to the cold crucible and was partially cracked. Space between the melted ingot and upper crust was created owing to the porosity of the charged mixture. The mass and thickness of the upper crust were 0.56 kg and 1.0 cm, respectively. The mass and height of the melted ingot were 5.545 kg and 10.9 cm, respectively. There were some un-melted UO2 pellets and ZrO2 powder in the vicinity of the melted ingot. The melted ingot was separated into two layers. The upper part was silver color and the lower lump was black color, which indicates that the upper part may be metal and the lower lump may be an oxidic mixture. Fig. 8 shows a photograph of the solidified ingot and its separated parts. In the main COSMOS test, the total charged mass was 6.86 kg and the total collected mass was 6.80 kg. Therefore, the mass loss
167
was 0.06 kg, which could be attributed to the release by aerosols during the melting process, and also the loss during the disassembling process of the cold crucible after the test. Despite some mass loss, the masses of the charged mixture and the collected products can be considered to be well balanced. In this study, a metallurgical inspection was performed with the aim of a precise investigation on the distribution and composition of the melt layer as the post-test examination. The metallurgical homogeneity and the chemical compounds of a sampled ingot were analyzed by EPMA and XRD, respectively. To perform the metallurgical inspection, the melted ingot was cut along the centerline. Fig. 9 shows the photograph of the melted ingot cut centerline. It can be clearly found that the melted ingot was separated into two layers, and the upper part was estimated to be metal and the lower lump was estimated to be an oxidic mixture. However, there were some white clods in the middle of the lower lump (at location (3) in Fig. 9). These white clods are estimated to be metal which was relocated from the upper metallic part into the lower oxidic lump. The detailed examination to determine the material compositions of these white clods were obtained through a metallurgical inspection. The results of the EPMA analysis are indicated by the relative fractions of the selected target elements, and the sum of those elements is expressed as 100%. Table 3 shows the EPMA analysis
Fig. 7. Schematic diagram of the solidified corium in the main test.
168
K.-H. Kang et al. / Nuclear Engineering and Design 278 (2014) 163–170
Fig. 8. Photographs of the solidified ingot in the main test.
results for the COSMOS-main test specimen. It can be confirmed that the upper part of the melted ingot was a metal mixture and the lower lump was an oxidic mixture, and there were some metal clods inside the lower lump. To obtain the chemical information on the melted ingot, the chemical compounds of the test specimen were analyzed by XRD. Fig. 10 through Fig. 14 show the XRD results for the test specimen in the COSMOS-main test. The lower lump mostly consists of UO2 , as shown in Figs. 11, 13 and 14. In the upper layer of the melted ingot, most of the iron was inferred to form ZrFe2 , as shown in Figs. 10 and 12. The XRD analysis results on the chemical information coincide with the visual observation on the centerline cut ingot in that the upper part is metal and the lower lump is an oxidic mixture with some metal clods, which means the possibility of a layer inversion of the heavy metallic material. The convection of molten corium included Fe during melting period by the cold crucible has not clearly known in experiments. According to the electromagnetic analysis for the only oxide mixture of UO2 and ZrO2 , this molten material is highly possible to be mixed, not separate because the mixture makes a convection flow
Fig. 9. Photograph of the centreline cut ingot and the specimen location for the metallurgical inspection in the main test.
Table 3 EPMA analysis results of the main test (unit: wt%). Specimen location
UO2
ZrO2
Fe2 O3
U
Zr
Fe
1 2 3 4 5
0 66.46 0 69.07 73.07
0 28.79 0 27.03 24.31
0 4.75 0 3.9 2.62
25.12 0 25.42 0 0
15.66 0 16.76 0 0
59.22 0 57.82 0 0 Fig. 10. XRD analysis results for test specimen #1 in the main test.
K.-H. Kang et al. / Nuclear Engineering and Design 278 (2014) 163–170
Fig. 11. XRD analysis results for test specimen #2 in the main test.
169
Fig. 14. XRD analysis results for test specimen #5 in the main test.
experimental results show the separated ingot into two layers of metallic and oxidic, which is after slowly solidifying the molten mixture from the required melting temperature. 4. Conclusions
Fig. 12. XRD analysis results for test specimen #3 in the main test.
by electromagnetic force (Song et al., 2005). Also, the composition analysis of the crust formed between molten mixture and figure of the crucible shows that the composition is almost constant with vertical line of the crust. However, if the metal composition of Fe in the mixture of UO2 and ZrO2 is high, the shape of molten material could be different because the Fe with a lower melting point compared to the molten mixture of UO2 and ZrO2 can be first floated on the top of the molten corium and separately makes a convection flow on the top of the molten oxide. For this reason, the current
The COSMOS tests using the prototypic materials have been performed to investigate the layer inversion of the corium pool of the heavy metallic material in the lower plenum of the reactor vessel. An induction heating method using a cold crucible was implemented for melting of the UO2 –ZrO2 –Zr–Fe mixture. Three preliminary tests have been performed to enhance the experimental techniques using real core material. One main test for the initial melt pool condition in the lower plenum of the reactor vessel has been performed in the TLFW sequence under the IVR-ERVC of the APR1400. After the main test, a post-test metallurgical inspection of the cutting and XRD for the solidified corium pool was performed to investigate the layer inversion. In the main test, the U/Zr ratio and Zr oxidation rate were 0.93 and 69%, respectively. The melting techniques of the real core material using a cold crucible were successfully developed from the three preliminary tests. The melted ingot was separated into two layers of metallic and oxidic in the main test. The metallurgical inspection results on the chemical information coincide with the visual observation on the centerline cut ingot in that the upper part is metal and the lower lump is an oxidic mixture with some metal clods, which means the possibility of a three-layer formation. The present test results will be used to produce metallurgical inspection test data to validate the thermodynamic analysis methodology using the GEMINI code. More tests are necessary to validate the corium pool configuration of the two-layer or three layer formations in the lower plenum of the reactor vessel for various sequences of the APR1400, such as LBLOCA, MBLOCA, SBLOCA, SBO. Acknowledgements This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (Ministry of Science, ICT, and Future Planning) (Nos. NRF-2012M2A8A4025893 and NRF-2012M2A8A4025889). References
Fig. 13. XRD analysis results for test specimen #4 in the main test.
Barrachin, M., Defoort, F., 2004. Thermophysical properties of in-vessel corium: MASCA Programme related results. In: Proceedings of MASCA Seminar 2004, Aix-en-Provence, France. Dinh, T.N., Tu, J.P., Salmassi, T., Theofanous, T.G., 2003. Limits of Coolability in the AP1000-Related ULPU-2400 Configuration V Facility, NURETH-10, Seoul, Korea.
170
K.-H. Kang et al. / Nuclear Engineering and Design 278 (2014) 163–170
Dinh, T.N., Tu, J.P., Theofanous, T.G., 2004. Two-phase natural circulation flow in AP1000 in-vessel retention-related ULPU-V facility experiments. In: Proceedings of ICAPP’04, Pittsburgh, PA, USA. Esmaili, H., Khatib-Rahbar, M., 2004. Analysis of In-Vessel Retention and Ex-Vessel Fuel Coolant Interaction, NUREG/CR-6849, ERI/NRC 04-201. Kang, K.H., et al., 2010. Experimental and analytical investigation on the layer inversion of melt pool during the severe accidents in the APR1400. In: Proceedings of the 8th International Topical Meeting on Nuclear Thermal-Hydraulics, Operation and Safety (NUTHOS-8), Shanghai, China. KEPCO, 1998. In-Vessel Retention Workshop on Severe Accident Management of Korea Next Generation Reactor. KEPCO. Kymalainean, O., et al., 1997. In-vessel retention of corium at the Loviisa plant. Nucl. Eng. Des. 169, 109–130. Park, J.W., et al., 2001. Assessment of In-vessel Core Debris Coolability for the APR1400 Design, KHNP Report.
Park, R.J., Kim, S.B., Suh, K.Y., Rempe, J.L., Cheung, F.B., 2006. Detailed analysis of the late-phase melt conditions for the evaluation of an in-vessel corium retention. Nucl. Technol. 156 (3). Rempe, J.L., Suh, K.Y., Cheung, F.B., Kim, S.B., 2008. In-vessel retention of molten corium-lessons learned and outstanding issues. Nucl. Technol. 161 (210). Scobel, J.H., Theofanous, T.G., Conway, L.E., 2002. In-vessel retention of molten core debris in the Westinghouse AP1000 Advanced Passive PWR. In: Proceedings of ICAPP’02, Hollywood, Florida. Song, J.H., Min, B.T., Kim, J.H., Kim, H.W., Hong, S.W., Chung, S.H., 2005. An electromagnetic and thermal analysis of a cold crucible melting. Int. Commun. Heat Mass Transfer 32, 1325–1336. Theofanous, T.G., et al., 1996. The first results from the ACOPO experiment. In: Proceedings of the International Topical Meeting on Probabilistic Safety Assessment (PSA’96), Park City, Utah. Theofanous, T.G., Liu, C., Addition, S., Angelini, S., Kymalainen, O., Salimassi, T., 1997. In-vessel coolability and retention of a core melt. Nucl. Eng. Des. 169 (1.).
Contents lists available at ScienceDirect
Nuclear Engineering and Design journal homepage: www.elsevier.com/locate/nucengdes
An experimental study on layer inversion in the corium pool during a severe accident Kyoung-Ho Kang ∗ , Rae-Joon Park, Seong-Ho Hong, Seong-Wan Hong, Kwang Soon Ha Korea Atomic Energy Research Institute, 1045 Daedeok-daero, Yuseong-Gu, Daejeon 305-353, Republic of Korea
h i g h l i g h t s • • • • •
COSMOS tests were performed to investigate the layer inversion of the corium pool using prototypic materials. An induction heating method was implemented for melting of the UO2 –ZrO2 –Zr–Fe mixture. The simulated sequence was the TLFW in APR1400. A metallurgical inspection was performed to investigate the layer inversion. COSMOS test results show the possibility of layer inversion of the heavy metallic material.
a r t i c l e
i n f o
Article history: Received 3 March 2014 Received in revised form 14 July 2014 Accepted 15 July 2014
a b s t r a c t COSMOS (COrium configuration of the molten State in the MOst Severe accidents) tests using prototypic materials have been performed to investigate the layer inversion of the heavy metallic material in the corium pool. An induction heating method using a cold crucible was implemented for melting of the UO2 –ZrO2 –Zr–Fe mixture. Before the main test, three preliminary tests have been performed to enhance the experimental techniques using the real core material. One main test of the COSMOS has been performed to evaluate the corium pool configuration in the lower plenum of the reactor vessel for the TLFW (Total Loss of Feed Water) sequence of the APR (Advanced Power Reactor) 1400 under the IVR-ERVC (InVessel corium Retention through External Reactor Vessel Cooling). A post-test examination of cutting, EPMA, and XRD for the solidified corium pool ingot in the main test has been performed to investigate the layer inversion. From the three preliminary tests, melting techniques of the real core material using a cold crucible were successfully developed. The metallurgical inspection results on the chemical information coincide with the visual observation on the centerline cut ingot in that the upper part is metal and the lower lump is an oxidic mixture with some metal clods in the main test. This means the possibility of layer inversion of the heavy metallic material in the corium pool. © 2014 Elsevier B.V. All rights reserved.
1. Introduction The melted core material relocates to the lower plenum of the reactor vessel during a severe accident and forms the corium pool under the IVR (In-Vessel corium Retention) through the ERVC (External Reactor Vessel Cooling), which is considered an effective means for maintaining the integrity of the reactor vessel of a nuclear power plant (Theofanous et al., 1996, 1997; Rempe et al., 2008). The IVR-ERVC has been adopted in low-power reactors of the AP600 and Loviisa nuclear power plants, and in a medium-power reactor of an AP1000 as a design feature for severe accident mitigation
∗ Corresponding author. Tel.: +82 42 868 2665; fax: +82 42 868 8256. E-mail address: [email protected] (K.-H. Kang). http://dx.doi.org/10.1016/j.nucengdes.2014.07.005 0029-5493/© 2014 Elsevier B.V. All rights reserved.
(Esmaili and Khatib-Rahbar, 2004; Dinh et al., 2003, 2004; Scobel et al., 2002; Kymalainean et al., 1997), and in high-power reactors of the APR (Advanced Power Reactor) 1400 and APR+ as an accident management strategy (KEPCO, 1998; Park et al., 2001, 2006). Fig. 1 shows the typical corium pool formation in the lower plenum of the reactor vessel under the IVR-ERVC. In general, a corium pool including metallic materials, such as stainless steel and Zircaloy, and oxide materials, such as UO2 and ZrO2, may form at the initial stage. The corium pool may be stratified into two layers, due to their density differences, an upper metallic layer without a volumetric heat source and a lower oxidic layer with a volumetric decay heat source, which is an original two-layer formation. The experimental results of the OECD MASCA showed that when a sufficient amount of non-oxidized zirconium (Zr) is available, metallic uranium (U) migrates to the metallic layer (Barrachin
164
K.-H. Kang et al. / Nuclear Engineering and Design 278 (2014) 163–170
Fig. 1. Corium pool formation in the lower plenum of the reactor vessel.
and Defoort, 2004). The transfer of species between the U, O, Zr melt and the steel can result in a significant density increase of the metallic layer. The density increase of the metallic layer can lead to inverse corium pool stratification with an additional heavy metal layer below the oxidic pool, which leads to three-layer formation as shown in Fig. 1. The presence of the metallic layer at the bottom of the lower plenum is likely to decrease the thickness of the top metallic layer and consequently increase the thermal load to the lower head vessel, which is the focusing effect in the top metallic layer. The melt pool configurations inside the lower head of the reactor vessel affect the initial thermal load to the vessel and play
a key role in determining the integrity of the reactor vessel under the IVR-ERVC. Thermodynamic analyses using the GEMINI code were performed to examine the final melt pool configuration at the reactor vessel failure during severe accidents in the APR1400 (Kang et al., 2010). As the representative accident scenarios, a LBLOCA (Large Break Loss of Coolant Accident) without SI (Safety Injection), MBLOCA (Medium Break Loss of Coolant Accident) without SI, SBLOCA (Small Break Loss of Coolant Accident) without SI, SBO (Station Black Out), and TLFW (Total Loss Feed Water) were selected from the level I PSA (Probabilistic Risk Assessment) results. Melt pool configurations were different in the representative accident sequences of the APR1400. In the cases of SBLOCA and LBLOCA where the U/Zr ratio and the initial melt pool temperature were relatively higher, layer-inversion phenomena can be precluded, which results in the original two-layer formation of the corium pool. In other cases of the TLFW, SBO, and MBLOCA, however, a three-layer formation possibility by layer inversion should be considered in evaluating the melt pool configuration and consequent thermal load from the melt pool to the reactor vessel wall (Kang et al., 2010). COSMOS (COrium configuration of the molten State in the MOst Severe accidents) tests have been performed for an investigation on the molten pool configurations considering the layer inversion of the heavy metallic material. In the COSMOS test, an induction heating method using a cold crucible was implemented for a melting of the prototypic UO2 –ZrO2 –Zr–Fe mixture. Three preliminary tests have been performed to enhance the experimental melting techniques for the real core material of the UO2 –ZrO2 –Zr–SS mixture. The main test of the COSMOS to investigate the melt pool configuration in the lower plenum of the reactor vessel has been performed for the TLFW sequence of the APR1400. According to the thermodynamic analysis results, three-layer formation by a layer inversion of heavy metallic material is expected for this melt configuration of the TLFW sequence. The main objectives of this test were not only to provide physical insight into the layer inversion phenomena of the heavy metallic material, but also to produce metallurgical inspection test data to validate the thermodynamic analysis methodology using the GEMINI code. 2. Description of the COSMOS facility and test conditions In the COSMOS test, induction heating method using a cold crucible was implemented for melting of the prototypic UO2 –ZrO2 –Zr–Fe mixture. The maximum electric power and frequency of the generator were 100 kW and 120 kHz, respectively. Fig. 2 shows a schematic diagram of the COSMOS test facility. Photographs of the test facility are shown in Fig. 3. The inner diameter and height of the crucible were 9.45 cm and 25 cm, respectively.
Fig. 2. Schematic diagram of the COSMOS test facility.
K.-H. Kang et al. / Nuclear Engineering and Design 278 (2014) 163–170
Fig. 3. Photograph of the COSMOS test section.
The temperature of the melt pool surface was measured using the two-colored Pyrometer (Model name: IRCON 3R-35C15-0-00-1). The measuring range of the two-colored Pyrometer is from 1500 ◦ C to 3500 ◦ C with a measurement error of 2.0% of the reading.
165
Metallurgical inspection methods such as EPMA (Electron Probe Micro Analyzer) and XRD (X-Ray Diffraction) were implemented to precisely examine the distribution and composition of the melt layer as a post-test examination. The metallurgical homogeneity and chemical compounds of a sampled ingot were analyzed by EPMA and XRD, respectively. Table 1 shows the test condition for the melt compositions in the COSMOS test. Three preliminary tests (CO-PR1, CO-PR2, and COPR3) have been performed to enhance the experimental techniques using a real core material. The main experiment (COSMOS-main) has been performed for the initial melt pool conditions in the lower plenum of the reactor vessel for the TLFW sequence of the APR1400 under the IVR-ERVC. According to the thermodynamic analysis results, a layer inversion of heavy metallic material was expected for this melt configuration of the TLFW sequence. To obtain the ideal melt mass, as shown in Table 1, a UO2 –ZrO2 –Zr–SS (stainless steel) mixture was charged, as shown in Fig. 4. Exact melting of the mixture in terms of the melt mass, however, is a very hard task in the test which implements a cold crucible induction heating method for melting of the prototypic UO2 –ZrO2 –Zr–SS mixture. For this reason, in the COSMOS-main test, actual melt compositions were different from those intended for the TLFW simulation. The actual melt compositions are shown in Table 2. It was found that the U/Zr ratio and Zr oxidation rate in the main test differed from those of the TLFW simulation in which a layer inversion of the heavy metallic material was expected. In Table 2, U/Zr and Cn stand for the molar ratio of Uranium to Zirconium, and the Zr oxidation ratio, respectively. The Zr oxidation rate (Cn) is defined as the oxidation rate of the corium by the molar ratio, ZrO2 /(Zr + ZrO2 ). The experimental results of the OECD MASCA showed that the U/Zr ratio and Zr oxidation rates are key parameters determining the possibility of the layer inversion of the heavy metallic material. The typical charging pattern for the COSMOS-main test is shown in Figs. 4 and 5. The thickness of the ZrO2 bottom liner was 35 mm and the mass was 0.63 kg. The UO2 pellets filled the circumference of the crucible to protect the fingers from an attack of the hot molten metal with a relatively lower melting temperature. In the middle part, the mixture of the UO2 pellets, ZrO2 powder, Zr and stainless steel metal chips were charged. The upper part was filled with the ZrO2 powder. The initiator made of a Zr metal ring, which enabled the charged mixture to be melted at an initial stage, was installed at the upper part. The test vessel protecting the charged cold crucible was purged with argon gas before induction heating, and argon gas was continuously supplied into the vessel at a flow rate of 10 l/min during the heating to prevent a rapid oxidation of the Zr metal chips which could cause a melt eruption. Another reason for the argon gas supply was to vent the aerosols during the melting period in order to measure the melt surface temperature more accurately using the two-colored Pyrometer. Variation of the applied power by the induction heating generator is shown in Fig. 6. The power level was increased step by step. The melting probably started at 770 s because the coupling factor of
Table 1 COSMOS test matrix. Test name
Composition
Ratio (kg/%)
Melt ratio (%)
Comments
CO-PR-01
ZrO2 /Fe/Zr
67.5
Preliminary test
CO-PR-02
ZrO2 /Fe/Zr
86.1
Preliminary test
CO-PR-03
UO2 /ZrO2 /Fe/Zr
40.0
Preliminary test
COSMOS-main
UO2 /ZrO2 /Fe/Zr
4.15/0.85/0.2 (81.3/16/7/2.1) 3.65/1.03/0.12 (76.1/21.5/2.4) 3.82/1.74/1.36/0.1 (54.5/24.8/19.3/1.4) 4.58/0.74/0.47/1.06 (72.8/11.8/7.5/7.9)
84.0
Main test
166
K.-H. Kang et al. / Nuclear Engineering and Design 278 (2014) 163–170
Fig. 4. Schematic diagram of the charging pattern in the main test. Table 2 Test conditions for the melt compositions in the main test.
Ideal corium Actual corium
UO2 –ZrO2 –Zr mass (kg)
SS mass (kg)
U/Zr
Cn
Comments
4.58/0.74/0.47 3.64/1.23/0.42
1.06 0.89
1.52 0.93
54 69
Target melt compositions Obtained melt compositions
Q decreased slowly from this time. After the applied power reached the maximum level of 63 kW, the power level was maintained for about 14 min and the power was turned off. The maximum absorbed power into the melt was estimated to be about 37.8 kW, which was 59.7% of the total maximum input power of 63 kW. The melting of the surface was confirmed by the visual observation in the COSMOS-main test. After the power was turned off, the melt was solidified in the cold crucible during the cooling process by the
Fig. 5. Inside photograph of the COSMOS test section.
Fig. 6. Variation of the applied power in the main test.
K.-H. Kang et al. / Nuclear Engineering and Design 278 (2014) 163–170
chiller actuation for the purpose of the investigation on the layer inversion phenomena. 3. Discussions on the experimental results The melting techniques of the real core material using a cold crucible were successfully developed from the three preliminary tests. The main test of the COSMOS was performed using these techniques in the TLFW sequence of the APR1400. A schematic diagram and photograph of the solidified corium in the crucible are shown in Figs. 7 and 8, respectively. The upper crust anchored to the cold crucible and was partially cracked. Space between the melted ingot and upper crust was created owing to the porosity of the charged mixture. The mass and thickness of the upper crust were 0.56 kg and 1.0 cm, respectively. The mass and height of the melted ingot were 5.545 kg and 10.9 cm, respectively. There were some un-melted UO2 pellets and ZrO2 powder in the vicinity of the melted ingot. The melted ingot was separated into two layers. The upper part was silver color and the lower lump was black color, which indicates that the upper part may be metal and the lower lump may be an oxidic mixture. Fig. 8 shows a photograph of the solidified ingot and its separated parts. In the main COSMOS test, the total charged mass was 6.86 kg and the total collected mass was 6.80 kg. Therefore, the mass loss
167
was 0.06 kg, which could be attributed to the release by aerosols during the melting process, and also the loss during the disassembling process of the cold crucible after the test. Despite some mass loss, the masses of the charged mixture and the collected products can be considered to be well balanced. In this study, a metallurgical inspection was performed with the aim of a precise investigation on the distribution and composition of the melt layer as the post-test examination. The metallurgical homogeneity and the chemical compounds of a sampled ingot were analyzed by EPMA and XRD, respectively. To perform the metallurgical inspection, the melted ingot was cut along the centerline. Fig. 9 shows the photograph of the melted ingot cut centerline. It can be clearly found that the melted ingot was separated into two layers, and the upper part was estimated to be metal and the lower lump was estimated to be an oxidic mixture. However, there were some white clods in the middle of the lower lump (at location (3) in Fig. 9). These white clods are estimated to be metal which was relocated from the upper metallic part into the lower oxidic lump. The detailed examination to determine the material compositions of these white clods were obtained through a metallurgical inspection. The results of the EPMA analysis are indicated by the relative fractions of the selected target elements, and the sum of those elements is expressed as 100%. Table 3 shows the EPMA analysis
Fig. 7. Schematic diagram of the solidified corium in the main test.
168
K.-H. Kang et al. / Nuclear Engineering and Design 278 (2014) 163–170
Fig. 8. Photographs of the solidified ingot in the main test.
results for the COSMOS-main test specimen. It can be confirmed that the upper part of the melted ingot was a metal mixture and the lower lump was an oxidic mixture, and there were some metal clods inside the lower lump. To obtain the chemical information on the melted ingot, the chemical compounds of the test specimen were analyzed by XRD. Fig. 10 through Fig. 14 show the XRD results for the test specimen in the COSMOS-main test. The lower lump mostly consists of UO2 , as shown in Figs. 11, 13 and 14. In the upper layer of the melted ingot, most of the iron was inferred to form ZrFe2 , as shown in Figs. 10 and 12. The XRD analysis results on the chemical information coincide with the visual observation on the centerline cut ingot in that the upper part is metal and the lower lump is an oxidic mixture with some metal clods, which means the possibility of a layer inversion of the heavy metallic material. The convection of molten corium included Fe during melting period by the cold crucible has not clearly known in experiments. According to the electromagnetic analysis for the only oxide mixture of UO2 and ZrO2 , this molten material is highly possible to be mixed, not separate because the mixture makes a convection flow
Fig. 9. Photograph of the centreline cut ingot and the specimen location for the metallurgical inspection in the main test.
Table 3 EPMA analysis results of the main test (unit: wt%). Specimen location
UO2
ZrO2
Fe2 O3
U
Zr
Fe
1 2 3 4 5
0 66.46 0 69.07 73.07
0 28.79 0 27.03 24.31
0 4.75 0 3.9 2.62
25.12 0 25.42 0 0
15.66 0 16.76 0 0
59.22 0 57.82 0 0 Fig. 10. XRD analysis results for test specimen #1 in the main test.
K.-H. Kang et al. / Nuclear Engineering and Design 278 (2014) 163–170
Fig. 11. XRD analysis results for test specimen #2 in the main test.
169
Fig. 14. XRD analysis results for test specimen #5 in the main test.
experimental results show the separated ingot into two layers of metallic and oxidic, which is after slowly solidifying the molten mixture from the required melting temperature. 4. Conclusions
Fig. 12. XRD analysis results for test specimen #3 in the main test.
by electromagnetic force (Song et al., 2005). Also, the composition analysis of the crust formed between molten mixture and figure of the crucible shows that the composition is almost constant with vertical line of the crust. However, if the metal composition of Fe in the mixture of UO2 and ZrO2 is high, the shape of molten material could be different because the Fe with a lower melting point compared to the molten mixture of UO2 and ZrO2 can be first floated on the top of the molten corium and separately makes a convection flow on the top of the molten oxide. For this reason, the current
The COSMOS tests using the prototypic materials have been performed to investigate the layer inversion of the corium pool of the heavy metallic material in the lower plenum of the reactor vessel. An induction heating method using a cold crucible was implemented for melting of the UO2 –ZrO2 –Zr–Fe mixture. Three preliminary tests have been performed to enhance the experimental techniques using real core material. One main test for the initial melt pool condition in the lower plenum of the reactor vessel has been performed in the TLFW sequence under the IVR-ERVC of the APR1400. After the main test, a post-test metallurgical inspection of the cutting and XRD for the solidified corium pool was performed to investigate the layer inversion. In the main test, the U/Zr ratio and Zr oxidation rate were 0.93 and 69%, respectively. The melting techniques of the real core material using a cold crucible were successfully developed from the three preliminary tests. The melted ingot was separated into two layers of metallic and oxidic in the main test. The metallurgical inspection results on the chemical information coincide with the visual observation on the centerline cut ingot in that the upper part is metal and the lower lump is an oxidic mixture with some metal clods, which means the possibility of a three-layer formation. The present test results will be used to produce metallurgical inspection test data to validate the thermodynamic analysis methodology using the GEMINI code. More tests are necessary to validate the corium pool configuration of the two-layer or three layer formations in the lower plenum of the reactor vessel for various sequences of the APR1400, such as LBLOCA, MBLOCA, SBLOCA, SBO. Acknowledgements This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (Ministry of Science, ICT, and Future Planning) (Nos. NRF-2012M2A8A4025893 and NRF-2012M2A8A4025889). References
Fig. 13. XRD analysis results for test specimen #4 in the main test.
Barrachin, M., Defoort, F., 2004. Thermophysical properties of in-vessel corium: MASCA Programme related results. In: Proceedings of MASCA Seminar 2004, Aix-en-Provence, France. Dinh, T.N., Tu, J.P., Salmassi, T., Theofanous, T.G., 2003. Limits of Coolability in the AP1000-Related ULPU-2400 Configuration V Facility, NURETH-10, Seoul, Korea.
170
K.-H. Kang et al. / Nuclear Engineering and Design 278 (2014) 163–170
Dinh, T.N., Tu, J.P., Theofanous, T.G., 2004. Two-phase natural circulation flow in AP1000 in-vessel retention-related ULPU-V facility experiments. In: Proceedings of ICAPP’04, Pittsburgh, PA, USA. Esmaili, H., Khatib-Rahbar, M., 2004. Analysis of In-Vessel Retention and Ex-Vessel Fuel Coolant Interaction, NUREG/CR-6849, ERI/NRC 04-201. Kang, K.H., et al., 2010. Experimental and analytical investigation on the layer inversion of melt pool during the severe accidents in the APR1400. In: Proceedings of the 8th International Topical Meeting on Nuclear Thermal-Hydraulics, Operation and Safety (NUTHOS-8), Shanghai, China. KEPCO, 1998. In-Vessel Retention Workshop on Severe Accident Management of Korea Next Generation Reactor. KEPCO. Kymalainean, O., et al., 1997. In-vessel retention of corium at the Loviisa plant. Nucl. Eng. Des. 169, 109–130. Park, J.W., et al., 2001. Assessment of In-vessel Core Debris Coolability for the APR1400 Design, KHNP Report.
Park, R.J., Kim, S.B., Suh, K.Y., Rempe, J.L., Cheung, F.B., 2006. Detailed analysis of the late-phase melt conditions for the evaluation of an in-vessel corium retention. Nucl. Technol. 156 (3). Rempe, J.L., Suh, K.Y., Cheung, F.B., Kim, S.B., 2008. In-vessel retention of molten corium-lessons learned and outstanding issues. Nucl. Technol. 161 (210). Scobel, J.H., Theofanous, T.G., Conway, L.E., 2002. In-vessel retention of molten core debris in the Westinghouse AP1000 Advanced Passive PWR. In: Proceedings of ICAPP’02, Hollywood, Florida. Song, J.H., Min, B.T., Kim, J.H., Kim, H.W., Hong, S.W., Chung, S.H., 2005. An electromagnetic and thermal analysis of a cold crucible melting. Int. Commun. Heat Mass Transfer 32, 1325–1336. Theofanous, T.G., et al., 1996. The first results from the ACOPO experiment. In: Proceedings of the International Topical Meeting on Probabilistic Safety Assessment (PSA’96), Park City, Utah. Theofanous, T.G., Liu, C., Addition, S., Angelini, S., Kymalainen, O., Salimassi, T., 1997. In-vessel coolability and retention of a core melt. Nucl. Eng. Des. 169 (1.).