LOCA analysis of high temperature reactor cooled and moderated by supercritical light water

Size: px
Start display at page:

Download "LOCA analysis of high temperature reactor cooled and moderated by supercritical light water"

Transcription

1 GENES4/ANP23, Sep , Kyoto, JAPAN Paper 116 LOCA analysis of high temperature reactor cooled and moderated by supercritical light water Yuki Ishiwatari 1*, Yoshiaki Oka 1 and Seiichi Koshizuka 1 1 Nuclear Engineering Research Laboratory, the University of Tokyo, Tokai-mura, Ibaraki, , Japan Tel: , Fax: , ishi@utnl.jp Large break LOCA of a high temperature thermal reactor (SCLWR-H) with descending flow water rods is analyzed. The criterion of the Ni-alloy cladding temperature is set to 126 C. The safety system of the SCLWR-H is similar to that of BWR. The safety principles of the SCLWR-H are (1) feedwater from cold-leg is kept and (2) coolant outlet is opened at hot-leg. To prevent closure of the hot-leg, automatic depressurization valves (ADS) are actuated without delay by the same signal as that of main steam isolation valves (MSIV) such as Flow rate low level 3 or Core pressure low level 2 or, while in ABWR the ADS is actuated by Water level 1 and Containment pressure high with 3 s delay. The low pressure core injection system (LPCI) is actuated by the same signal as that of the ADS with 3 s delay due to the starting of the emergency diesel generators. In the case of 1% cold-leg break LOCA, the core flow rate decreases and then the cladding temperature increases up to about 72 C during blowdown phase. After the ADS are opened, the core flow rate is recovered and the cladding temperature decreases. The core reflooding begins 78 s after the break. The peak cladding temperature in the reflooding phase is about 79 C. In the case of 1% hot-leg break LOCA, the core coolability is much higher than that in cold-leg break. The peak cladding temperature does not exceed that of the normal operation. In the reflooding phase, since the cold-leg loops are isolated by check valves, only upward flow in the core is established. It is a kind of forced cooling by the LPCI. The reflooding phase is also less severe than that of the cold-leg break LOCA. Thus, reflooding phase of hot-leg break LOCA is not calculated in this study. In summary, large break LOCA of the SCLWR-H is mitigated by MSIV, ADS and LPCI. The flow system of the descending flow water rods makes the water inventory large and contributes to the high coolability. This is an advantage. But it should be noticed that the core coolability is small before the ADS are opened at cold-leg break LOCA. KEYWORDS: supercritical water reactor, descending flow water rods, large break LOCA, blowdown, reflooding I. Introduction Supercritical pressure light water cooled reactor (SCR) has been conceptually designed in the University of Tokyo 1)2). It adopts technologies of light water reactors (LWR) and supercritical fossil-fired power plants (FPP). Once through cooling system makes the reactor pressure vessel (RPV) and the containment vessel (CV) small. High specific enthalpy of main steam makes the balance of plant (BOP) compact and thermal efficiency high. The SCR has a potential of a large cost reduction from LWRs. In the past study 3), a LOCA analysis code SCRELA was developed. It consists of the blowdown estimation module and the reflood estimation module. These modules were validated by the REFLA-TRAC code, which was developed in the Japan Atomic Energy Research Institute (JAERI) based on TRAC-PF1. Cold-leg and hot-leg break LOCAs of a low temperature thermal reactor (SCLWR) were analyzed using the SCRELA. After that a high temperature thermal reactor (SCLWR-H) with descending flow water rods has been designed. But the SCRELA does not contain the water rod model. A plant transient analysis code SPRAT-DOWN has been developed containing the descending flow water rod model. But this code can calculate only supercritical pressure region. The purposes of this study are to modify the SPRAT-DOWN for LOCA analysis based on the SCRELA, to analyze large break LOCA of the SCLWR-H, and to clarify its characteristics. II. Reactor core with descending flow water rod The fuel assembly of the SCLWR-H is shown in Fig. 1. It contains many water rods for neutron moderation. The coolant flows downward in the water rods. The flow path at normal operation is shown in Fig. 2. Part of the feedwater is led to the top dome and cools it. Then it flows downward in the control rod guide tubes and the water rods. In the bottom of the fuel assemblies it is mixed with the rest coolant which has descended in the downcomer. This concept has advantages. Since there is no mixing of hot and cold coolant at the upper plenum, the coolant outlet temperature is kept high. Since the difference in the average water density is small between upper and lower parts of the core, the axial distribution of the neutron moderation does not change substantially. * Corresponding author, Tel , Fax , ishi@utnl.jp 1

2 29.22cm Fig. 1: Fuel assembly Water rod UO 2 Gd 2 O 3 fuel rod Control rod guide tube UO 2 fuel rod Capacity: RCIC(AFS) TD 1 unit: 4%/unit AFS TD 2 units: 4%/unit ADS 8 units: 2%/unit at 25MPa LPCI MD 3 units: 3kg/s/unit at 1.MPa Configuration: TD-AFS LPCI/RHR TD- RCIC LPCI/RHR RCIC: reactor core isolation cooling system RHR: residual heat removal system TD: turbine driven MD motor driven TD-AFS LPCI/RHR Cold leg Top dome Hot leg Upper plenum Active core Bottom dome Control rod drive Control rod guide tube Hot leg Cold leg Down comer Water rod Fig. 2: Flow path at normal operation III. Safety system The plant system of the SCLWR-H is shown in Fig. 3. The safety system is similar to that of BWR. It consists of reactor scram, high-pressure auxiliary feedwater system (AFS), low-pressure core injection system (LPCI), main steam isolation valves (MSIV), safety relief valves (SRV) and automatic depressurization valves (ADS). Capacity and configuration of the safety system are shown in Fig. 4. RPV CR Suppression chamber CV SRV/ADS Turbine control valve Turbine bypass valve AFS Turbine Condenser Fig. 4: Capacity and configuration of safety system IV. Safety principle The SCLWR-H has no water level. To keep the core coolability, flow rate should be watched and maintained while water level is kept in BWR. Decrease in the feedwater flow rate is directly followed by decrease in the core flow rate because the SCLWR-H has no recirculation. Closure of the coolant outlet is also followed by decrease in the core flow rate. The safety principles of the SCLWR-H are to keep feedwater from cold leg and to keep coolant outlet open at hot leg. The system pressure should be also watched. The relation between abnormal levels and actuations of the safety system is shown in Fig. 5. In the LOCA analysis, actuation of the high pressure ECCS (AFS) is neglected as in BWR. Flow rate low Level 1 Reactor scram Level 2 AFS Level 3 MSIV/ADS/LPCI system Pressure low Level 1 Reactor scram Level 2 MSIV/ADS/LPCI system Pressure high Level 1 Reactor scram Level 2 SRV AFS: auxiliary feedwater system MSIV: main steam isolation valve ADS: automatic depressurization system LPCI: low pressure core injection system LPCI LPCI Fig. 5: Abnormal level and actuation of safety system Condensate water storage tank High pressure feedwater heaters Fig. 3: Plant system Main feedwater pumps Low pressure feedwater heaters V. LOCA analysis code 1. Blowdown phase Since water rods are not modeled in the blowdown estimation module of SCRELA 3), a plant transient analysis code SPRAT-DOWN, which includes descending 2

3 flow water rods model but can calculate only supercritical pressure region, is modified to calculate blowdown phase of the SCLWR-H. The calculation model of the modified SPRAT-DOWN is shown in Fig. 6. The hottest single channel is divided into 2 nodes. A water rod channel is also divided into 2 nodes. Heat transfer between these 2 channels is considered. The main feedwater lines are divided into 1 nodes. The downcomer is divided into 2 nodes including the bottom dome. The upper plenum is divided into 2 nodes including the main steam lines. The top dome is divided into 1 nodes including the CR guide tubes. Since the fuel channel and the water rod is modeled as single channels, the volumes of the top dome, the downcomer, the upper plenum and the main feedwater lines are divided by the total number of the fuel rods. The mass and energy conservations are calculated. where ρ G + = t Z ( ρh ) ( GH ) + = Q' t Z t: time ρ: density G: mass flow rate Z: position H: specific enthalpy Q : heat generation rate peer unit volume In the nodes of tow-phase flow, the average density and the average specific enthalpy are determined as: A = A ( 1 x) A x (3) where l + v '' A: density ρ or specific enthalpy h l: liquid v: vapor x: void fraction The boundaries are the main feedwater pumps, the break point, the MSIV and the ADS. Since the pressure drop in the main feedwater lines and the RPV is much smaller than that at the break, the pressure is assumed as constant in them. The decreasing rate of the pressure in the blowdown phase is governed by the flow rate at the break. The calculation module of the break flow is the same as that in the SCRELA. At subcritical pressure, three correlations of the break flow in superheated vapor, sub-cooled water, and two phase are used. Critical flow in the supercritical pressure is not known. But the pressure at the break is subcritical even if the stagnation pressure is supercritical. The correlations in the subcritical pressure are also used in the supercritical pressure. If the stagnation temperature is below or equal to the pseudo-critical temperature, the critical flow is treated as in the sub-cooled region. If the temperature is higher than the pseudo-critical temperature, the superheated vapor region is assumed. Fig. 7 shows the critical mass fluxes at various pressures. The heat transfer coefficient is evaluated by Dittus-Boelter s correlation in supercritical, sub-cooled and (1) (2) superheated regions. In the film boiling region, Dougal-Rhosenow s correlation of film boiling is used. It is also conservatively used in the nucleate boiling region. The radiation heat transfer is involved. Heat capacities of the RPV and other structures are neglected. The axial power distribution is cosine. The reactor power is calculated by the point-kinetics equation with six delayed neutron groups while the decay heat is calculated using a two-group approximation of 12% of the ANS evaluation 4). Doppler and coolant density feedbacks are considered. Reactor scram is completed 2.8 s after the signal including.55 s delay. The reactivity worth is 1 %dk/k. The reactivity curve shown in Fig. 8 is the same as that of PWR. When LOCA is detected, the main feedwater pumps are assumed to trip. The pump coast-down time is assumed to be 5 s and the flow rate decreases linearly. The flow chart of the calculation is shown in Fig. 9. In the past study 3), the blowdown estimation module of the SCRELA was validated by comparing with the REFLA-TRAC code, which was developed in the Japan Atomic Energy Research Institute (JAERI) based on TRAC-PF1. The calculation started at a core pressure 17 MPa in the REFLA-TRAC code, since this code can not treat supercritical pressure. A low temperature thermal reactor (SCLWR) analyzed in another past study 5) was used for the validation calculation. In this study, the modified SPRAT-DOWN is also compared with these 2 calculations. Fig. 1 shows the pressures of 1% hot-leg LOCA. Main steam line MSIV Check valve Main feedwater line Main feedwater pump ADS line ADS LPCI Bottom dome Top dome Fig. 6: Calculation model of blowdown phase CR guide tube Break (hot leg) Break (cold leg) Upper plenum Fuel channel Water rod Down comer 3

4 25 2 Pressure(bar) 15 1 REFLA-TRAC SCRELA SPRAT-DOWN Time(sec) 4 5 Fig. 7: Critical max flux as a function of stagnation enthalpy and pressure Reactivity ratio Fig. 8: Scram reactivity curve Start Fig. 1: Pressure trend in 1% hot-leg break LOCA 2. Reflooding phase The reflood estimation module of the SCRELA is used. Water rods are not modeled in this code. But a reflooding calculation without considering water rods is conservative because quench front goes up slower and heat transfer from fuel channels to water rods is neglected. It includes System momentum calculation, Thermal equilibrium relative velocity correlation and Quench front velocity correlation. Various heat transfer correlations are prepared according to the flow conditions such as single-phase liquid, saturated two-phase, transient, dispersed and superheated steam flow. The flow chart of the calculation is shown in Fig. 11. This code was also validated by comparing with the REFLA-TRAC code. The detail of this code is explained in Ref. 3). Start Reading calculation condition and blowdown output data Downcomer water level and quench front level Input Momentum conservation in RPV Pressure assumption Intact Loop, ECCS Mass & Energy Conserv. in core nodes Heat transfer Time step change Break flow Mass & Energy Conserv. Flow balance check Yes Heat transfer (Clad - Coolant, Fuel channel - Water rod) Heat conduction in fuel rod Reactivity and Power No P change Flow rate and enthalpy at core top Cladding temperature System momentum equation System (Upper plenum) pressure Pressure at core water level Time step change Finish of Reflood No Yes End No Finish of Blowdown Yes End Fig. 9: Calculation flow chart of blowdown phase Fig. 11: Calculation flow chart of reflood phase VI. Sequence Actuation conditions of the safety system are compared with those of ABWR and PWR in Table 1. The reactor is assumed to be tripped by Flow rate low level 1 or Core 4

5 pressure low level 1 or. The high pressure AFS are assumed to fail. The characteristic of the MSIV, which is the same as that of ABWR, is shown in Fig. 12. If the MSIV are closed and the ADS are not opened at cold-leg break LOCA, the coolant outlet at the hot leg is closed. It means that one of the safety principles to keep coolant outlet open at hot leg is not satisfied and therefore the core flow rate and the core coolability are significantly small. Thus, the ADS are opened without delay by the same signal as that of the MSIV such as Flow rate low level 3 or Core pressure low (level 2) or, while in ABWR the ADS are opened by Water level 1 and Containment pressure high with 3 s delay. It is the most important characteristic of the LOCA sequence of the SCLWR-H. The LPCI are actuated by the same signal as that of the MSIV with 3 s delay due to the starting of the emergency diesel generators. Two out of three LPCI units are assumed to start at a pressure.8 MPa which is conservatively less than the design pressure 1. MPa. Ratio [%] MSIV signal Closed Fig. 12 Characteristic of MSIV Table 1: Actuation conditions of safety system in LOCA analysis (Conditions with underlines are actually used in analysis.) Safety system ABWR PWR SCLWR-H Reactor scram Accumulator High-pressure ECCS Low-pressure ECCS MSIV ADS, or Rapid decrease in core flow rate Fail Water level 1, or (3 s delay) Water level 1.5 (no delay) Water level 1, and (3 s delay) Core pressure low, or ECCS startup Core pressure below 4 MPa, or Core pressure low and pressurizer water level low, or Core pressure abnormally low (32 s delay), or Core pressure low and pressurizer water level low, or Core pressure abnormally low (32 s delay) Flow rate low level 1, or Core pressure low level 1, or Fail Flow rate low level 3, or Pressure low level 2, or (3 s delay) Flow rate low level 3, or Pressure low level 2, or (no delay) VII. LOCA analysis The characteristics of the SCLWR-H analyzed here are shown in Table 2. The plant parameters of the initial condition (normal operation) are as follows: a) core power 1% (23 MWt) b) core pressure 25. MPa c) feedwater flow rate 1% (119 kg/s) d) main steam temperature 5 C e) maximum cladding temperature 643 C f) maximum linear power 39 kw/m The limitation of the cladding temperature of LWR is 126 C for stainless steels, which was obtained considering metal-water reaction. In this study, the limitation of the Ni-alloy cladding temperature is determined to be the same as that of stainless steels. The oxidation characteristics of Ni-alloy should be subject for future study. 5

6 Table 2: Characteristics of SCLWR-H Core Core diameter / height [m] 3.6 / 4.2 Number of fuel assemblies 96 Coolant inlet / outlet temperature [ C] 28 / 5 Coolant density coefficient [dk/k/(g/cm 3 )].2 Doppler coefficient [dk/k/ C] Maximum linear power [kw/m] 39 RPV and Main loop Inner diameter / wall thickness / total height [m] 4.34 /.35 / 15 Volume of top dome / upper plenum / bottom dome / down comer [m] 55 / 24 / 21 / 26 Inner diameter of main feedwater line / main steam line [m].27 /.46 Length of main feedwater line / main steam line [m] (1 loop) 2 / 2 Number of main loops 2 Fuel assembly Fuel rod diameter / pitch [mm] 1.2 / 11.2 Cladding material / thickness [mm] Ni-alloy /.63 Water rod wall material / thickness [mm] Ni-alloy /.2 Number of fuel rods / water rods 3 / 36 Mass flux in fuel channel / water rod [kg/s/m 2 ] 1161 / 45 System Core pressure [MPa] 25. Thermal / electric power [MW] 23 / 1 Thermal efficiency [%] 43.5 Feedwater flow rate [kg/s] Cold leg break LOCA (a) Blowdown phase The time sequence is shown in Table 3. The pressure and the peak cladding temperature are shown in Fig. 13. The flow rates and the reactor power are shown in Fig. 14. Before the ADS are opened, the core flow rate is small because a large quantity of high-density water in the top dome and the water rods flows to the break without passing through the core. After the ADS are opened, the core flow rate is recovered and the cladding temperature decreases. The reactor power is promptly decreased by density feedback and scram. When the LPCI flow from the suppression chamber reaches the core bottom at 78 s, the blowdown calculation is finished. The highest cladding temperature is 721 C. The influences of various parameters are shown in Table 4. The peak cladding temperature is sensitive to the ADS parameters such as the time delay and the number of valves opened. But it still has a good margin compared with 126 C even if the delay is a little longer and some of the valves are not opened. (b) Reflooding phase The reflooding phase starts from 78 s. The quench front level, the downcomer level, the peak cladding temperature and its axial position are shown in Fig. 15. The downcomer is filled with the water from the LPCI at 112 s. The axial position of the peak cladding temperature goes up with the quench front level. The highest cladding temperature is 792 C at 255 s. The quench front reaches the core at about 5 s. The effect of the LPCI capacity is shown in Table 5. If it is smaller, the reflooding phase begins later. But the peak cladding temperature is not so sensitive to the LPCI capacity because the highest water level in the downcomer is constant. Table 3: Time sequence of 1% cold-leg break LOCA 1 % Cold leg break Scram signal by Pressure low level 1 (24.MPa) MSIV/ADS/LPCI signal by Pressure low level 2 (23.5MPa) ADS opened Scram completed MSIV closed Pressure.8 MPa LPCI actuated Start of reflooding phase Highest cladding temperature 792 C (Reflooding) Complete of reflooding phase s Pressure [MPa] pressure peak cladding temperature Temperature [ o C] Fig. 13: Pressure and peak cladding temperature at 1% cold leg break LOCA (blowdown phase) 6

7 Ratio [%] Flow rate at core top -5 Flow rate at core bottom Flow rate at water rod top -1 Reactor power Fig. 14: Flow rate and reactor power at 1% cold-leg break LOCA (blowdown phase) Table 4: Sensitivity analysis of 1% cold-leg break LOCA (blowdown phase) Break size [%] PCT [ o C] ADS delay [s] PCT [ o C] hot-leg break LOCA is expected to be much less severe than that of cold-leg break LOCA. Thus, only blowdown phase is analyzed in this study. The time sequence of 1% hot-leg break LOCA is shown in Table 6. The pressure and the peak cladding temperature are shown in Fig. 16. The flow rates and the reactor power are shown in Fig. 17. The core flow rate is significantly increased because the high-density water in the RPV and the main feedwater lines flows through the core to the break and the ADS. At the beginning the core power is temporally increased by density feedback. But the core flow rate is much larger. Thus, the cladding temperature does not exceed that of normal operation. The blowdown calculation is finished at 66 s. s Table 6: Time sequence of 1% hot-leg break LOCA 1 % Hot leg break Scram signal by Pressure low level 1 (24.MPa) MSIV/ADS/LPCI signal by Pressure low level 2 (23.5MPa) ADS opened Scram completed MSIV closed LPCI actuated (Pressure.8 MPa) LPCI flow reaches core bottom. Axial position [m] Number of ADS opened PCT [ o C] PCT: Peck Cladding Temperature Quench front level Downcomer level Position of PCT PCT PCT: Peak cladding temperature Temperature [ o C] Pressure [MPa] pressure peak cladding temperature Temperature [ o C] Fig. 16: Pressure and peak cladding temperature at 1% hot-leg break LOCA (blowdown phase) 4 Fig. 15: Reflooding phase of 1% break cold-leg LOCA Table 5: Influence of LPCI capacity LPCI capacity [kg/s/unit] When reflooding starts [s] PCT [ C] PCT: Peak cladding temperature 2. Hot leg break In the case of hot-leg break LOCA, the reflooding phase is different from that of cold-leg break LOCA. Since the cold leg pipes are isolated by the check valves, the coolant outlet is only at the break and the ADS lines. It is a kind of forced cooling by the LPCI. Thus, the reflooding phase of Ratio [%] Flow rate at core top Flow rate at core bottom Flow rate at water rod top Reactor power Fig. 17: Flow rates and reactor power at 1% hot-leg break LOCA (blowdown phase) 7

8 VIII. Discussion At normal operation the coolant density in the top dome, which has a large volume fraction in the RPV, is the same as that of the feedwater. The density in the water rods is also high. Thus, the water inventory in the RPV of the SCLWR-H is much larger than that without water rods. It contributes to the high coolability at large break LOCA. Fig. 18 and Fig. 19 respectively show the cladding temperatures in blowdown and reflooding phase of 1% cold-leg break LOCA analyzed in the past study 3) without considering water rods. The peak cladding temperatures of the blowdown and the reflooding phase were respectively 8 C and 98 C with a LPCI capacity 85 kg/s/unit. In this study they are about 72 C and 79 C with a LPCI capacity only 3 kg/s/unit. But the core coolability is small if the ADS are not opened. That s why the ADS are opened by or logic of signals without delay while in ABWR they are opened by and logic with 3 s delay. References 1) Y. Oka and S. Koshizuka, Design Concept of One-Through Cycle Supercritical-Pressure Light Water Cooled Reactors, Proc. 1st Int. Symposium on Supercritical Water-cooled Reactors, Design and Technology, Tokyo, Japan, Nov. 6-9, 2, 1-22 (2) 2) Y. Oka, S. Koshizuka, Y. Ishiwatari and A. Yamaji, Elements of design Consideration of Once-Through Cycle, Supercritical-Pressure Light Water Cooled Reactor, Proc. Int. Conf. on Advanced Nuclear Power Plants (ICAPP), Hollywood, Florida, June 9-13, 22, Sec. 3.4 (22) 3) J. H. Lee, Y. Oka and S. Koshizuka, Development of a LOCA Analysis Code for the Supercritical Pressure Light Water Cooled Reactors, Ann. Nucl. Energy, Vol. 25, No. 16, (1998) 4) Decay energy release rates following shutdown of uranium-fueled thermal reactors, proposed standard ANS , American Nuclear Society (1971) 5) S. Koshizuka et al., Large Break Loss of Coolant Accident Analysis of a Direct-Cycle Supercritical Pressure Light Water Reactor, Ann. Nucl. Energy, Vol. 21, No. 3, (1994) Fig. 18: Cladding temperatures in blowdown phase of 1% cold-leg break LOCA analyzed in past study Fig. 19: Cladding temperatures in reflooding phase of 1% cold-leg break LOCA analyzed in past study IX. Conclusion Large break LOCA of the SCLWR-H is mitigated by MSIV, ADS and LPCI. The RPV structure with descending flow water rods gives a large water inventory and makes the core coolability high at LOCA. The highest cladding temperature is only 792 C even though the LPCI capacity is 3 kg/s/unit (37% of that in the past study). But it should be noticed that ADS actuation is important at cold-leg break LOCA. 8