Passive ECC Penetration Duct

GENES4/ANP2003, Sep. 15-19, 2003, Kyoto, JAPAN Paper 1124 Passive ECC Penetration Duct Tae-Soon Kwon *, Chul-Hwa Song, and Won-Pil Baek Korea Atomic Energy Research Institute (KAERI), Thermal Hydraulic Research, Daejeon, Korea The ECC penetration duct of the direct vessel injection () system for the emergency core cooling (ECC) system has been studied to prevent steam-water interaction in the downcomer increasing the ECC bypass fraction during the late reflood phase of a large break loss-of-coolant accident (LBLOCA. The ECC penetration duct is installed from the nozzle at the upper downcomer to the lower downcomer. The performance of the extension duct was evaluated for both LBLOCA and line break of the APR1400 using the MARS code, which is a multi-dimensional best-estimate thermal hydraulic code being developed by KAERI. The shape of an advanced extension duct, which has a two exits at both the top and bottom of the duct, was considered to prevent an inlet-to-outlet switching of the extension duct during the line break. A vent hole is considered at the top of the ECC penetration duct to have the role of the break suction point at the upper downcomer. The break flow is initiated at both holes during the blowdown phase. But, the major break flow is switched from the bottom of the duct to the vent hole at the top caused by the difference in the static head. The results show the new ECC penetration duct having a vent hole at the top of the duct is very useful for both the reflood phase of LBLOCA and the line break. However, the fuel cladding temperature of the simply extended duct into the lower downcomer without vent top holes exceeds the limit value of 1200K. The collapsed water level both at the downcomer and core is also very low compared to those of the typical nozzle. The locations of nozzles both at the lower downcomer below the cold leg about 2 m and at the upper downcomer above the cold leg about 2 m are tested for both the LBLOCA and line break accident. The downcomer boiling and the late reheating occurred when the nozzle was installed at the lower downcomer. KEYWORDS : ECC penetration duct, ECC bypass, line break, LBLOCA I. Introduction APR1400 is a 3983 MWt evolutionary pressurized water reactor with two reactor coolant loops, each of which consists of one steam generator, one hot leg and two cold legs. For the safety injection system, APR1400 has four mechanically separated hydraulic trains with an emergency core cooling system (ECCS) 1) with direct vessel injection (). Each hydraulic train, is composed of a high pressure safety injection pump (SIP) and a safety injection tank (SIT), is connected to the nozzle (21.59 cm I.D.) located approximately 2.1 m above the centerline of the cold legs. Since emergency core cooling (ECC) water is directly injected into the upper part of the reactor vessel downcomer, the thermal-hydraulic phenomena during lossof-coolant accidents (LOCAs) could be quite different from those observed in the previous commercial PWRs adopting cold leg injection (CLI) type. 2) The bypass fraction and liquid entrainment of ECC water in the system is strongly dependent on the steam-water interaction in the downcomer. The ECC water injected into the downcomer is entrained due to high speed steam flow in the circumferential direction during the reflood phase of a LBLOCA. Thus, the in-vessel extension duct of the nozzle is a very useful method to prevent a * Corresponding author, Tel.+82-42-868-8717, Fax. +82-42-868-8362, E-mail: tskwon@kaeri.re.kr steam-water interaction in the downcomer of a reactor vessel. However, if the extension duct has the shape of a simply extended pipe into the lower downcomer, the exit of the extension pipe plays the role of the inlet suction when the line break occurrs. Thus, the collapsed water level of the downcomer may become much lower than those of a typical system without the extension duct. Therefore, the simple pipe inserted into the lower downcomer as an extension duct may not be applicable for the line break only the LBLOCA. The performance of the extension duct should be applicable for both the LBLOCA and line break. If the vent hole is considered at the top of the simplified extension duct like a square chamber or pipe vertically installed in the downcomer, the extension duct has two exits both at the top and bottom. Then the top vent hole at the upper downcomer has a role of inlet suction during the line break. Thus, this design feature has the possibility of preventing the inlet-to-outlet switching of the extension duct while the ECC water may be discharged through both the bottom and top exits during the ECC injection phase of a LBLOCA. The elevation of the nozzle affects the collapsed water level in the downcomer. The peak cladding temperature and the ECC bypass fraction also depend on the injection location of the ECC water. Two types of injection locations, as shown in Fig.1 (a) and (b), are considered at about 2 m above the cold leg and at about 2 m below the cold leg. The collapsed water level of the lower system may be much lower than that of the high system. Because the break suction elevation of the lower system is much lower than the elevation of the cold leg during the line break accident. For LBLOCA,

the lower system is more effective in preventing the direct bypass of the ECC water, especially the liquid entrainment. However, the elevation of the nozzle as a break location is more important in the line break accident because the downcomer water level is strongly dependent on the elevation of the nozzle. To prevent the water level decreasing, the suction elevation of the break flow should be located at least near or above the elevation of the cold legs. In this paper, the performances of the extension duct and the effect of the nozzle elevation are evaluated focused on the peak cladding temperature, collapsed water level, and reheating for both a LBLOCA and line break accidents. High C.L. Hot Leg C.L. Hot Leg Low MARS, the new features of RELAP5/MOD3.2.2 γ-version are also implemented. The MARS code has the capability of analyzing the one-dimensional and/or three-dimensional thermal-hydraulic system and fuel responses of light water reactor transients. In the analysis presented in this paper, a one-dimensional module of the MARS employing the extended 3-D feature of the RELAP5/MOD3 is used. For the approximation of the multi-dimensional effects in the downcomer region, x-, y-, and z- coordinate directional flow geometry input are employed to apply the onedimensional momentum equation to each of the coordinate directions. This feature makes it possible to treat the momentum of each direction being convected by the velocity in the same direction. In this paper, the 8.5 inches line break and cold leg double ended guillotine breaks are taken into account for a reference accident to evaluate the performance of the elevation and the extension duct of the nozzle. The major parameters for comparison include the pressurizer pressure, break flow, cladding temperature, collapsed water level of the core and downcomer, core reheating, and the safety injection flow etc. C.L. (a) High (a) Simple type (b) Low Fig.1 nozzle elevation Hot Leg C.L. Hot Leg (b) Top and bottom hole type Fig. 2 Schematic of break suction II. Analysis Method To investigate the major thermal-hydraulic phenomena of the APR1400, under the anticipated accidental conditions, safety analysis has been performed using the best estimate computer code MARS. 4) The MARS code used for the analysis has been developed at KAERI by unifying and restructuring the RELAP5/MOD3.2.1 and COBRA-TF, the one-dimensional module and the threedimensional module respectively. In the course of the code's development, many improved features have been implemented to extend the code's modeling capability and to enhance user friendliness. In the current version of III. Component Model 1. nozzle and Extension Duct The break point is considered at the nozzle mounted at the upper downcomer in the APR1400 as shown in Fig.1(a). Thus the elevation of the break nozzles both the extended type and typical types are at the same elevation of the mounting point of the typical nozzle in the APR1400. The schematic is shown in Fig.2(a) and (b). However, as shown in Fig.2(a) and (b), the break suction point is at the lowest point of the extension duct in the downcomer not the mounting point of the nozzle because the extended duct into the lower downcomer has a role as a break path during the line break flow. The design features of the extended duct of the system in this test are shown in Fig.3(a), (b) and (c). As shown in Fig.3(c), the shape of the extension duct with a vent hole is somewhat different from the simple pipe particularly it's mounting at the downcomer as shown in Fig.3(b). The flow area in the vertical part of the duct is two times the flow area of the nozzle, while the top vent hole has a flow area same as the nozzle's. In the present analysis, the nozzle is located at 2 m above the cold legs. The duct is extended up to 2 m below the cold legs. The break flow of the line break is discharged through the ECC penetration duct from the lower downcomer region or the core inlet region. In the simplified duct extended into the lower downcomer as shown in Fig.4(a), the collapsed water level of the downcomer may be low because the break suction point is the extended duct exit during the line break. As shown in Fig.5(a) and (b), the extension duct has an internal type of a flow channel so that it may mitigate the phasic interaction between the steam and ECC water in the downcomer region during the late reflood phase of a LBLOCA.

Nozzle D/C C.L. Hot Leg C.L. Hot Leg (a) Typical Break (b) Simply extended Cold legs Steam Cross Flow (a) Simply extended duct (b) with top vent hole Fig.4 Schematic of line break s ECC Water Fig.3 (c) Top vent hole duct Type of extension duct However, as shown in Fig.4(b), if a vent hole is considered at the top of the ECC penetration duct to have a role of a break suction point at the upper downcomer, the break flow is initiated at both holes during the blowdown phase. The major break flow is switched from the bottom of the duct to the vent hole at the top caused by a difference in the static head of the extended duct. The region where the lines and cold legs are connected is modeled as 6 columns that are split azimuthally every 60 and each column is set to have 5 axial nodes. The lower and upper regions of the downcomer are modeled as a BRANCH component. The core flow goes through two channels: the average core channel and guide tubes. The bypass flow is simplified to have two channel flows: flows through the guide tube and key alignment simulator. For the low elevation of the nozzle, as shown in Fig.1(a), it is considered to be 2 m below the cold legs. (a) Typical (b) Extended duct Fig.5 Schematic of LBLOCA with 2. Reactor Vessel and downcomer As shown in Fig.6, the annular section of the APR1400 downcomer is modeled as 6 columns that are split azimuthally every 60. These columns are also divided by 10 axial nodes and the cross flows between these nodes are modeled to simulate the multi-dimensional flow behavior. The first, third, fourth, and sixth columns replicate the downcomer regions connected to the 4 cold leg nozzles and 4 nozzles, while the second and fifth columns represent the downcomer regions connected in the hot leg nozzle. A broken and three intact cold legs, and 4 lines are connected to the normal direction to the outer surface of the downcomer. The injected vessel flow splits into four channels; average core channel, hottest rod channel, core shroud bypass channel, and core guide tubes. The split vessel flows gather in the upper plenum of the reactor vessel and go out to each hot leg. Fig. 6 MARS 1-D nodalization for APR1400 3. Steam Generator and Steam Lines The APR1400 has two identical steam generators and the generated steam from each steam generator is collected in the common header and delivered to the turbine. In the nodalization of the APR1400, the steam generator is divided into the downcomer, economizer, evaporator and

riser, separator, steam dryer, and steam dome. 4.Safety Injection System As previously described, the APR1400 adopts the direct vessel injection method for the safety system. The safety injection system has four mechanically separated hydraulic trains. Each hydraulic train, is composed of a High Pressure Safety Injection Pump (SIP) and a Safety Injection Tank (SIT), which is connected to the nozzle located at 2 m above the centerline of the cold legs. The nozzle also has the extended duct into the lower downcomer annulus. In the present analysis, two trains of ECCS are considered under the single failure assumption (two SIPs are available) and the nodalizations for the ECCS of the APR1400. 5. Initial and Boundary Condition To simulate the transient phenomena, 102% of the normal power and 1.2 times decay power of the ANS'73 decay heat model are used in the prototype calculation. For the cold leg double-ended guillotine break, the break position is assumed to be at the pump discharge line. To simulate the break flow, two junctions from the pump side and vessel side, respectively, to the atmosphere are configured in the nodalization for the transient calculation. In the transient calculation, the containment pressure is assumed to be atmospheric pressure. IV. Results and Discussion 1. LBLOCA After the LBLOCA is initiated, the primary system pressure of the APR1400 is precipitously dropped, as shown in Fig.7(a), as the blowdown phase, and maintains the low primary pressure condition. As the emergency core cooling (ECC) water is actuated to inject, the core level is recovered by the safety injection tank flow as shown in Fig. 7(b). The fuel cladding temperature reaches the peak values during the reflood phase. The collapsed water level of the core is strongly dependent on the liquid entrainment of the ECC bypass in the downcomer. To continue the high downcomer water level, the ECC extension duct has the role of ECC water protector in the upper downcomer and cold leg nozzles region for ECC penetration to the lower down comer. Fig.7 shows the thermal-hydraulic responses using the ECC extension duct for the reflood phase of the LBLOCA, especially. The results show that the extension duct contributes to a high water level in the downcomer and core cooling. The comparison results are summarized in Table.1. As shown in Fig.7(a), the subcooling degree of coolant at the lower downcomer is much enhanced by the extension duct because the ECC water does not heat up by high temperature steam in the downcomer. The extension duct prevents the steam-water interaction in the downcomer. The core reflood rate is maintained by the increased subcooling degree of the coolant in the lower downcomer without downcomer boiling during the late reflood phase of LBLOCA. But, the additional head of the extension duct in the downcomer does not contribute to the pressure of the downcomer. If the liquid entrainment in the downcomer is mitigated sufficiently by the extension duct, the increased downcomer water level affects the void fraction or quality of the upstream condition of the break. As shown in Fig.7(b), the break flow is a little distorted. As shown in Fig.7(c), the water level of the downcomer shows that the extension duct is a very useful tool to maintain the DC water level by mitigating the steam-water interaction in the downcomer. For the first stage of the refill phase, the water level of the downcomer is refilled very fast by the SIT high flow, and maintains the high liquid level until the SIT is empty. The core reflood rate is governed by the collapsed water level of the downcomer during the reflood phase of the LBLOCA, and dependent on the ECC penetration rate and ECC bypass rate. If the ECC bypass in the downcomer is mitigated well by adapting the extension duct of the nozzle, the liquid level of the downcomer may increase up to the onset-of-sweep out liquid level during the period of SIT activation. The liquid level of the downcomer is governed by the sweep out near the broken cold leg. The excessive SIT flow above the sweep out level goes out through the break. However, the degree of subcooling and liquid level is governed by ECC bypass during the reflood phase. As shown in Fig.7(c) and (d), the extension duct increases the collapsed water level of the downcomer and core to about 2 m higher than that of the typical nozzle without any extension part. As shown in Fig.7(e), the peak cladding temperature in case of the typical nozzle shows the late reheating during the reflood phase after about 1600 seconds because the downcomer water level is much decreased due to the liquid entrainment in the downcomer. However, for the case of the installed extension duct, the reflood peak cladding temperature has much lower values compared to those of the typical case. Furthermore, late reheating during the reflood phase does not appear when the extension duct is used. The extension duct of the nozzle may increase the subcooling degree of the coolant in the lower downcomer because the ECC water does not heat up by the steam when the ECC water flows down. Thus downcomer boiling during the late reflood phase of the LBLOCA does not occur when the extension duct is used as a part of ECC system with. As shown in Fig.7(f), the safety injection flows are not different from the two injection types. The fluidic device model is applied in the injection flow rate, thus the high safety injection flow continues to about 200 sec. The downcomer water level is maintained at high level during the SIT activation. 2. Line Break 2.1 Elevation Effect : High and Low Downcomer The effect of the elevation of the nozzle is investigated under the same conditions except the elevation of the nozzle in the downcomer. After the line break is initiated, the primary system pressure dropped to a saturation condition of about 8 MPa, and maintains twophase blowdown in the saturation condition. As the primary coolant is further leaked out, the pressure curves show the

gradual decrease and secondary plateau at about 2MPa as shown in Fig.8(a). The pressure trends with respect to time agree well with each other before SIT initiation at 200 seconds. However, the pressure of the lower nozzle is lower than that of the high nozzle during the reflood phase after 300~400 seconds. : Typical (APR1400) : Extension Duct (e) Cladding temperature (a) Subcooling degree at lower downcomer 20000 15000 10000 5000 0 Rx Side : No Ext. Rx Side : Ext. Pump Side : No Ext. Pump Side : Ext. 0 200 400 600 (b) Break flow rate (c) Collapsed D/C water level (d) Collapsed core water level Fig. 7 (f) SI flow rate LBLOCA transient for a typical nozzle (No. Ext.) vs. a simply extended nozzle. The break flowrate has a similar trend except for some perturbation around 400 seconds. This fluctuation means that the upstream of the break located at the lower downcomer is changed to a single phase by the SIT high injection flow. As shown in Fig.8(c), the water level of the downcomer shows that the coolant flows out the break nozzle in the lower downcomer. The DC water level is decreased with a level difference between the two nozzles in high and low elevation. For the first stage of the SIT discharge phase, the water level of the downcomer is refilled for a very short period by the SIT high flow. However, the liquid level of the downcomer is decreased by the sweep out phenomena near the broken nozzle. Thus the liquid level of the downcomer does not go up above the cold leg under the condition of the same HPSI flow. As shown in Fig.8(c) and (d), the water level of the downcomer and core in case of the lower nozzle is about lower 2~3 m than that of the high level nozzle without an extension part during the late reflood phase of the line break. However, the typical high nozzle maintained a high water level during the reflood phase of the line break. As shown in Fig.8(e), the peak cladding temperature of the low nozzle shows late reheating during the late reflood phase after about 1600 seconds because the downcomer water level is decreased by the low water level of the downcomer. However, for the case of the high, the reflood peak cladding temperature has wetted values compared to those of the low case.

As shown in Fig.8(f), the HPSI flows are not so different from two injection types. The downcomer water level is maintained by HPSI injection during the reflood phase. But, the HPSI flows are not so different for the two cases. Thus, the low downcomer water level is caused by the elevation of the nozzle in the low system. (d) Collapsed core water level (a) Pressurizer pressure (b) Break flow rate 3000 2500 2000 1500 1000 500 0 0 400 800 1200 1600 2000 (c) Collapsed D/C water level Fig. 8 (e) Cladding temperature (f) SI flow rate line break transient for high and low level elevation of nozzle 2.2 Effect Of The Top Vent Hole of The Extension Duct To investigate the effect of the top vent hole of the extension duct in preventing the inlet-to-outlet switching of the extension duct, three types of the nozzles are considered. The first case is for a typical nozzle, which is installed at the upper downcomer in the APR1400 plant. The index of the first case is "No Ext." in Fig. 9. The elevation of the nozzle is the upper downcomer. The second case is for the extension duct of the nozzle into the lower downcomer. The extension duct has two exits at both the top and bottom. The bottom exit is for the ECC injection phase, and the top exit is for the steam suction for the line break. The index of the second case is "exit. with top hole" in Fig.9. The elevation of the exit for the ECC injection is the lower downcomer. Lastly, the third case is for the simple extension pipe without a top hole. The index of the third case is "Simple Ext." in Fig.9. The elevation of the nozzle is the lower downcomer. Thus, the static head in the extension duct may exist for the second case and the third case if the extension duct is filled with a liquid or two phase during the discharge phase including the break phase of the line break accident. In the simple extension duct as a pipe, the discharge outlet has the role of a break suction inlet when the line break occurs. Thus the exit in the downcomer during the discharge phase switches into the break suction inlet during the line break. The inlet-to-outlet switching of

the extension duct is the major phenomena in this test. Only in the simplified extension pipe, the inlet-to-outlet switching phenomenon occurs for the line break, not in the extension duct with a top vent hole. After the line break is initiated, the primary system pressure dropped to a saturation condition of about 8 MPa, and gradually decreases at about 2MPa as shown in Fig.9(a). The pressure trends with respect to time agree well with each other before the SIT initiated at around 200 seconds. However, the pressure of the simple extension nozzle is lower than that of the no extension nozzle during the reflood phase after 300~400 seconds. The ECC injection flowrate of the simple extension nozzle will becomes larger than that of the no extension nozzle because the system pressure is low as shown in Fig. 9(a). The break flowrate has a similar trend except some perturbation around 300~600 seconds. This fluctuation indicates that the upstream of the break node, located at the lower downcomer, is under the condition of a two-phase discharge condition. The safety injection water fills the downcomer from the bottom. Thus, the flow condition of the break suction in the lower downcomer is changed more easily to a two-phase or single-phase liquid from the steam condition after the refill phase than the other cases. Therefore, the break flow of the third case of the simple extension is higher than those of the other two cases. In the case of the top hole, the steam in the upper downcomer is mainly discharged because of the difference in the static head of the extension duct. Thus, the break flow rate in case of the no extension duct shows a trend nearly similar to the case of the extension duct with the top hole. As shown in Fig.9(c), the water level of the downcomer shows that the third case using a simple extension pipe without a top hole is much lower than those of the other cases. Fig.9(c) also shows that the effect of the inlet-tooutlet switching of the extension duct occurs during the line break. The collapsed water level of the downcomer is strongly dependent on the location of the break suction in the downcomer during the line break. As shown in Fig.9(e), the collapsed water level of the core has the similar trends with the downcomer water levels. The water level of the core for the third case with a simple extension pipe is lower than 3 m. The HPSI flows are not so different according to the injection types as shown in Fig.9(f). Therefore, the low water level is caused by inletto-outlet switching phenomena of the simplified extension pipe in the lower downcomer. As shown in Fig.9(e), the peak cladding temperature for the third case with a simple extension nozzle is heated up from the early reflood phase because the downcomer water level is decreased by the break suction in the lower downcomer for the break flow. However, for the case of the second one with the extension duct with the top hole, the cladding temperature shows continuously the wetted values during the reflood phase. The comparison results are summarized in Table 1. (a) Pressurizer pressure 4000 3000 2000 1000 0 0 200 400 600 800 1000 (b) Break flow rate (c) Collapsed D/C water level (d) Collapsed core water level

(e) Cladding temperature (f) SI flow rate No Ext.(High ) Ext. with top hole Simple Ext. Fig. 9 Line break ; No Ext. for Typical Type; Ext. for extension duct with vent hole; Simple Ext. for a simplified extension nozzle. Table 1 Comparison results for Types during the line break. The ECC penetration duct is installed from the nozzle at the upper downcomer to the lower downcomer. The shape of an advanced extension duct, which has two exits at both the top and bottom of the duct, was considered to prevent inlet-to-outlet switching of the extension duct during the line break. A vent hole is considered at the top of the ECC penetration duct for a role of break suction at the upper downcomer. The break flow is initiated at both holes during the blowdown phase. But, the major break flow is switched from the bottom exit of the duct to the vent hole at the top of duct caused by the difference in the static head. The results show that the new ECC penetration duct having a vent hole at the top of the duct is a very useful design for both the reflood phase of LBLOCA and the line break. The late reheating during the reflood phase of a LBLOCA was successfully prevented by extension duct having a top vent hole. In the extended duct without vent holes, however, the reflood PCT exceeds the limit value of 1200K for the low downcomer water level caused by the outlet-to-suction inlet switching of the simplified extension pipe into the lower downcomer during the line break even though its a useful tool in preventing the steam-water interaction during the refoold phase of the LBLOCA. The collapsed water level both at the downcomer and core is also very low compared to those of the typical type of the nozzle. Acknowledgment This research has been performed as a part of the nuclear R&D program supported by the Ministry of Commerce, Industry and Energy (MCIE) of the Korean government. V. Conclusion References 1) KEPCO, Standard Safety Analysis Report for Korean Next Generation Reactor, KEPCO report (1999). 2) Y. J. Chung et al. Test Requirements for the KNGR ECCS Performance Test, KAERI report, SA- -99008 (1999). 3) C. H. Song et al., Scaling Analysis of the Integral Test Loop to Simulate Korean PWR Plants, KAERI/TR- 1783/2001, KAERI (2001). 4) J. J. Jeong et al., Development of a multi-dimensional Thermal-hydraulic System Code, MARS 1.3.1, Annals of Nuclear Energy, 26(18), 1611-1642 (1999). The ECC penetration duct of the direct vessel injection () system for the emergency core cooling (ECC) system has been studied to prevent steam-water interaction in the downcomer which leads to increasing the ECC bypass fraction during the late reflood phase of a large break loss-of-coolant accident (LBLOCA), and to prevent the collapsed water level which is usually decreasing