Analysis of Accident Scenarios of a Water-Cooled Tokamak DEMO

Transcription

1 1 SEE/P5-1 Analysis of Accident Scenarios of a Water-Cooled Tokamak DEMO M. Nakamura 1, K. Ibano 2, K. Tobita 1, Y. Someya 1, H. Tanigawa 3, W. Gulden 4 and Y. Ogawa 5 1 Japan Atomic Energy Agency, Rokkasho, Japan 2 Graduated School of Engineering, Osaka University 3 Japan Atomic Energy Agency, Naka, Japan 4 F4E: Fusion for Energy, Garching, Germany 5 Graduate School of Frontier Sciences, University of Tokyo, Kashiwa, Japan Corresponding Author: nakamura.makoto@jaea.go.jp Abstract: Of late in Japan, a design study has been undertaken of a tokamak fusion DEMO with pressurized water coolant and solid pebble bed breeding blanket, but safety characteristics of this type of DEMO have not been well examined. In this paper, thermohydraulics analysis of in-vessel and ex-vessel loss-of-coolant accidents of a water-cooled tokamak DEMO is reported. Safety characteristics of water-cooled DEMO, particularly possible loads onto confinement barriers, are discussed based on the thermohydraulics analysis results. Measures to reduce such loads are also proposed. 1 Introduction Safety of a variety of fusion reactor concepts has been analyzed in these two decades. In the ITER project, safety characteristics of a water-cooled fusion device have been clarified extensively and safety systems for ITER have been considerably developed [1, 2]. Aside from the ITER safety case, safety of several fusion reactor concepts, which generate net electricity, was examined [3, 4]. Safety characteristics should be demonstrated also in a fusion DEMO reactor, which will be a fusion power plant of the first generation that can produce net electricity. Of late in Japan, a design study has been undertaken of a tokamak fusion DEMO with pressurized water coolant and solid pebble bed breeding blanket [5]. Safety characteristics of this type of DEMO, however, have not been well examined. The result of a preliminary hazard analysis [6, 7] shows that the decay heat and the coolant enthalpy of a water-cooled DEMO, which can mobilize radioactive materials contained within the reactor and challenge integrity of confinement barriers, are significantly larger than those of ITER. This result suggests that safety characteristics of water-cooled DEMO will be quite different from those of ITER.

2 SEE/P5-1 2 A purpose of this study is to clarify safety characteristics of water-cooled DEMO from the viewpoint of thermohydraulics. In this paper, we report thermohydraulics analysis of in-vessel (in-vv) and ex-vessel (ex-vv) loss-of-coolant accidents (LOCAs) and safety characteristics of water-cooled DEMO is discussed based on the analysis results. 2 DEMO Plant Model Analyzed The DEMO plant model analyzed in this study is based on DT fusion in a steady state tokamak. Basic guidelines for finding out a DEMO design point, which is an input for the safety study, was shown in the earlier papers [6, 7]. Here presented are key aspects of the DEMO design point with respect to the safety study. It is assumed in the DEMO plant model that [6, 7] The first wall (FW), breeding blanket (BLK) modules and divertor (DIV) cassettes are cooled by pressurized water in the pressurized water reactor conditions, i.e MPa, C. The tritium breeding blanket is made up based on know-how of the Japanese ITER Test Blanket Module (TBM) activities [8] and the Broader Approach (BA) DEMO R&D [9]. The structural material is made of reduced activation ferritic martensitic steel, F82H. Mixed solid pebble beds made of lithium-titanate (Li 2 TiO 3 ) and of Be-Ti beryllide (Be 12 Ti) [1] are used as tritium breeding and neutron multiplying material, respectively. The DEMO design parameters selected are summarized in Table I. TABLE I: Specification of the DEMO plant model analyzed in the safety study. Parameter Values Major radius 8.3 m Aspect ratio 3. Fusion power GW Net electricity output 2 3 MW Coolant for in-vessel components Pressurized water ( C, 15.5 MPa) First wall armor material Tungsten Blanket structural material Reduced activation ferritic martensitic steel F82H Tritium breeding material Lithium-titanate Li 2 TiO 3 (pebble beds) Neutron multiplying material Beryllide Be 12 Ti (pebble beds) Divertor armor material Tungsten Divertor structural material Reduced activation ferritic martensitic steel F82H

3 3 SEE/P5-1 Inboard blanket Tokamak building Primary heat transport system Outboard blanket PHTS vault FIG. 1: Schematics of a MELCOR model of ex-vv LOCA. 3 Analysis Method We analyze in-vv and ex-vv LOCAs by using the fully integrated, engineering-level thermohydraulics analysis code MELCOR [11] with modifications for fusion reactor safety applications [12]. The mass and energy of the liquid water, vapor and other non-condensable gases in the in- and out-board blankets, vacuum vessel (VV), first wall/blanket primary cooling loops, heat exchanger and tokamak building, and the environment are nodalized by a small number of control volumes. These volumes are connected with mass and heat flow paths. The heat flow between a volume and a component structure is also modeled by MELCOR. 4 Ex-vessel Loss of Coolant Accident Analysis In order to clarify the maximum load onto the confinement barrier, that is the vault covering the primary cooling loop, we conservatively considered double-ended break of the cooling pipe outside the VV but inside the tokamak building. We assumed that one of four primary cooling loops was broken. It was also assumed in order to identify the maximum load to the confinement barrier containing the primary cooling loop, that there were no pressure ventilation systems in/on the confinement area. Key input parameters for the ex-vv LOCA analysis by MELCOR are summarized in Table II, and the MELCOR model for the ex-vv LOCA analysis is shown in Fig. 1.

4 SEE/P5-1 4 TABLE II: Key input parameters for MELCOR analysis of an ex-vessel loss-of coolant of a first wall/blanket cooling loop Parameter Value Coolant water inventory per a primary cooling loop 24 m 3 /loop Number of the primary cooling loops 4 Inner diameter of the broken cooling pipe.727 m Break area.83 m 2 Volume of the confinement area covering a primary cooling loop 38, m 3 The analysis results of the thermohydraulic transients to the ex-vv LOCA indicate that the following event sequences will happen following the ex- VV LOCA, as shown in Fig. 2. The water pressure in the primary cooling loop decreases from 15.5 MPa to 11 MPa in.2 s after the pipe break because of the loss of the coolant (Fig. 2(a)). The pressure decrease cause flashing and then two-phase flow of the coolant. The elevation of the water in the pressurizer begins to decrease at.2 s after the pipe break, and is almost completely depleted in 3 s (Fig. 2(b)). Discharge of the liquid coolant water at the pipe break area lasts for 5.6 s after the pipe break, after which discharge of the vapor coolant continue for several tens seconds (Fig. 2(c)). The pressure in the confinement vault containing the broken primary cooling loop, reaches the maximum of 144 kpa at 97.6 after the pipe break (Fig. 2(d)). Pressure (MPa) Elevation (m) Mass flow rate (ton/s) Pressure (kpa) (a) Coolant pressure Outboard inboard Inboard (b) Water elevation in the pressurizer (c) Mass flow rate from the break area Steam Liquid water Total (d) Pressure in the vault Elapsed time (s) FIG. 2: Thermohydraulic transient behavior to the ex- VV LOCA. The analysis results indicate that the maximum pressure is about 44 kpa larger than the atmospheric pressure. In general, it is difficult to provide the large vault, such as the

5 5 SEE/P5-1 Inboard blanket Primary heat transport system Divertor heat transport system Outboard blanket FIG. 3: Schematics of a MELCOR model of in-vv LOCA. tokamak building, of pressure tightness. In order to reduce such overpressure of the vault, we propose the following strategies: (i) to enlarge the volume of the vault and (ii) to implement an intermediate small vault of pressure tightness, covering the primary cooling system inside the confinement area. TABLE III: Key input parameters for MELCOR analysis of an in-vessel loss-of coolant of a first wall/blanket cooling loop Parameter Value Volume of the vacuum vessel (VV) 3,8 m 3 Design pressure of the VV.5 MPa Volume of the VV pressure suppression system (PSS) 5,6 m 3 Volume of the water in the VVPSS 2,8 m 3 Disk rupture set point between the VV and PSS.2 MPa Area of the disk rupture 4. m 2 Cross section of a FW cooling pipe 64 mm

6 SEE/P In-vessel Loss of Coolant Accident Analysis We analyzed multiple double-ended break of the cooling pipes of the FW in the VV in order to clarify the load to the primary confinement barrier, i.e. the VV and its pressure suppression system (PSS). The VV is connected with the PSS to relieve overpressure by the in-vv LOCA. The VV and PSS are segregated by a rupture disk, of which set point is.2 MPa, in the relief pipe between the VV and PSS. The VV design pressure is assumed to be.5 MPa, which is similar to that of ITER-FDR. Key input parameters for the in-vv LOCA analysis by MELCOR are summarized in Table III, and the MELCOR model for the in-vv LOCA analysis is shown in Fig. 3. The total break area of the FW cooling pipes, which is equivalent to the number of the pipe break, is a scan parameter in this analysis. We analyzed an in-vv LOCA caused by break of all the outboard cooling pipes, which are in the poloidal direction, of the whole toroidal perimeter. The total break area of the FW cooling pipes are.82 m 2. The analysis results of the thermohydraulic transients to the in-vv LOCA indicate that the following event sequences will happen following the in-vv LOCA, as shown in Fig. 4. The in-vv LOCA leads sudden decrease in the coolant pressure in the outboard FW cooling channel, and at.1 s after the pipe break the coolant pressure in the inboard side begins to decrease (Fig. 4(a)). At about 1 s after the pipe break, the coolant in the pressurizer is completely exhausted (Fig. 4(b)). Figure 4(c) indicates that the ingress of water to the vacuum vessel lasts for about 1 s after the in-vv pipe break. The oscillation of the mass flow rate, which begins at 2 s after the pipe break, is attributed to the generation Pressure (MPa) Elevation (m) Mass flow rate (ton/s) Pressure (MPa) Vacuum vessel Relief pipe Distributor Suppression tank (a) FW Coolant pressure Outboard inboard (b) Water elevation in the pressurizer (c) Mass flow rate from the break area Liquid water Steam Total (d) Pressure Elapsed time (s) FIG. 4: Thermohydraulic transient behavior to the in- VV LOCA. of the two-phase flow of the coolant. At.61 s after the pipe break, the pressure in the VV reaches the disk rupture set point of.2 MPa (Fig. 4(d)). After that, nevertheless,

7 7 SEE/P5-1 the VV pressure continues to increase and reaches the maximum of 2.94 MPa at 6.5 s after the pipe break (Fig. 4(d)). It is noted that the maximum pressure exceeds the VV design pressure of.5 MPa. The ingress of water to the VV lasts for 7 s after the pipe break, and the VV and VVPSS pressures reach the equilibrium of 12 kpa. This result suggests that the VVPSS of the present design cannot cope with the in-vv LOCA caused by the break of all the outboard FW cooling pipes of the whole toroidal perimeter. The total break area of the FW cooling pipes has to be decreased in order to reduce the overpressure of the VV for the in-vv LOCA. Figure 5 shows that the total break area has to be below about 6 mm 2 to reduce the maximum VV pressure below the VV design pressure of.5 MPa. A way to reduce the FW break area could be to change the direction of the cooling pipes: for example, to arrange the pipes along with the toroidal direction rather than the poloidal direction. Pressure (MPa) Total break area of the FW cooling pipes (m 2 ) 6 Conclusions FIG. 5: Dependence of the maximum pressure of the VV on the total break area of the FW cooling pipes. We have analyzed the in-vv and ex-vv large LOCAs of the water-cooled tokamak DEMO. The analyses have identified the event sequences following the in-vv and ex-vv LOCAs. We have assessed the load onto the confinement area covering the broken primary cooling loop caused by the large ex-vv LOCA. It was found that the maximum load is so large that it is difficult make a large volume, such as the tokamak building, pressure-tight. The analysis result suggests that the primary cooling pipes should be covered by a small vault of pressure tightness or with a pressure suppression system. We have assessed the load onto the VV caused by the in-vv multiple break of the outboard FW cooling pipes of the whole toroidal perimeter. It was found that the maximum load, i.e. the maximum VV pressure, is about 6 times larger than the VV design pressure. The analysis result suggests that a measure to reduce the total FW break area will be required. A possible way is to arrange the pipes along with the toroidal direction rather than the poloidal direction. Acknowledgment This work is partially supported by the Broader Approach (BA) and by a Grant-in-Aid for Scientific Research from Japan Society for the Promotion of Science (JSPS). M.N.

8 SEE/P5-1 8 and K.T. thank Dr. Brad Merrill (Idaho National Laboratory, INL) for kindly providing us with MELCOR with the fusion modifications and his hospitality during their visit to INL. They also thank Mrs. Takao Araki (Toshiba Inc.) and Kazuhito Watanabe (Toshiba Inc.) for supporting MELCOR modeling. References [1] ITER Generic Site Safety Report (GSSR) as summarized in ITER Technical Basis, ITER EDA Documentation Series No. 24, IAEA, Vienna (22). [2] TAYLOR, N., et al., Updated safety analysis of ITER, Fusion Eng. Des. 86 (211) 619. [3] RAEDER, J., et al., Safety and environmental assessment of fusion power (SEAFP), EURFUBRU XII-217/95 (1995). [4] GULDEN, W., et al., Main safety issues at the transition from ITER to fusion power plants, Nucl. Fusion, 47 (27) S415. [5] TOBITA, K., et al., Research and development status on fusion DEMO reactor design under the Broader Approach, Fusion Eng. Des. 89 (214) 187. [6] NAKAMURA, M., et al., Study of safety features and accident scenarios in a fusion DEMO reactor, Fusion Eng. Des., 89 (214) 228. [7] NAKAMURA, M., et al., Key aspects of the safety study of a water-cooled fusion DEMO reactor, Plasma Fusion Res., in press (214). [8] ENOEDA, M., et al., Development of the Water Cooled Ceramic Breeder Test Blanket Module in Japan, Fusion Eng. Des., 87 (212) [9] NISHITANI, T., et al., Progress of fusion nuclear technologies in the broader approach framework, Fusion Eng. Des., 87 (212) 535. [1] TSUCHIYA, K., et al., Development of advanced tritium breeders and neutron multipliers for DEMO solid breeder blankets, Nucl. Fusion, 47 (27) 13. [11] GAUNTT, R.O., et al., MELCOR Computer Code Manuals vol. 1: Primer and Users Guide Version 1.8.5, NUREG/CR-6119, Vol. 1, Rev. 2, SAND2-2417/1, USNRC Report, Sandia National Laboratory (2). [12] MERRILL, B.J., et al., Modifications to the MELCOR code for application in fusion accident analyses, Fusion Eng. Des (2) 555.