A NATURAL CIRCULATION EXPERIMENT OF PASSIVE RESIDUAL HEAT REMOVAL HEAT EXCHANGER FOR AP1000

Transcription

1 A NATURAL CIRCULATION EXPERIMENT OF PASSIVE RESIDUAL HEAT REMOVAL HEAT EXCHANGER FOR AP1000 M. Duan, Y. Chen, Y. Lv, W. L, K. B, W. Wang, K. Du Chna Insttute of Atomc Energy P.O.Box 275(59), Bejng, , Chna H. Wang State Nuclear Power Technology Research &Development Centre ABSTRACT The passve resdual heat removal (PRHR) heat exchanger (HX) s an mportant emergency core coolng system of AP1000. In Amerca, several tests for PRHR HX n AP600 have been performed to study the heat transfer behavor. However, the test secton conssted of three vertcal parallel tubes wthout horzontal sectons. In the present test research, the PRHR heat exchanger was smulated by three C-type tubes, each of whch had a mddle vertcal secton, a top horzontal secton and a bottom horzontal secton wth the same sze and space as n AP1000. The experment purpose was to research the heat transfer behavor nsde and outsde of the HX tubes n dfferent condtons. The test faclty could control the test condtons ncludng the prmary system pressure, the nlet temperature of the HX tubes, the flow rate and the qualty n tubes. The test condtons could vary n a qute wde range whch could cover the operaton condtons n AP1000. Several natural crculaton tests n HX tubes have been performed. The test parameters covered the pressure of MPa, the qualty up to 0.2, the nlet water temperature of , the mass flow rate of kg/s and the heatng power of up to 364 kw. Hundreds of data have been obtaned ncludng the nner-tube water temperatures, the wall surface temperatures along the tubes, the pool water temperatures, the flow rate n tubes and so on. The expermental results were compared wth the calculatons by RELAP5/MOD3.3. The calculaton results agreed well wth the experment. KEYWORDS Natural crculaton, PRHR HX, AP INTRODUCTION AP1000 reactor s the thrd generaton pressurzed water reactor. The passve resdual heat removal (PRHR) heat exchanger (HX) s an mportant emergency core coolng system of AP1000, whch can remove the core decay heat by the coolant natural crculaton as a result of the dfferent water densty. The temperatures between two sdes of the PRHR HX tubes dffer greatly, and change n a wde range. Many knds of sngle-phase and two-phase flow pattern and heat transfer modes are nvolved. Especally n natural crculaton, the flow dynamc characters dffer remarkably wth the forced crculaton because of the low flow rate and the great nfluence by buoyancy. 4450

2 In Amerca, several tests for PRHR HX n AP600 have been performed to study the heat transfer behavor. The HX was smulated by a test secton whch conssted of three vertcal parallel tubes wth no horzontal sectons. The heat transfer has been researched between vertcal tubes and the coolng tank [1]. However, a HX tube n AP1000 s n C shape and conssts of a mddle vertcal secton, a top horzontal secton and a bottom horzontal secton. A majorty of heat s transferred through the top horzontal secton. So t s necessary to smulate the natural crculaton n HX tubes on a PRHR HX faclty wth C-type tubes and research the complcated thermo-hydraulcs phenomenon. In our research, the PRHR heat exchanger was smulated by three C-type tubes wth typcal szes n AP1000. The experment purpose was to research the heat transfer phenomena nsde and outsde of the HX tubes n dfferent condtons and develop proper analyss method. The experment condtons could vary n a wde range ncludng the operaton condtons n AP1000. The test faclty could obtan data ncludng the nner-tube water temperatures, the wall surface temperatures along the tubes, the pool water temperatures, the flow rate n tubes and so on. Several tests have been performed for natural crculaton n tubes. And the expermental results were compared wth the calculatons by RELAP5/MOD TEST FACILITY The PRHR heat exchanger test faclty was desgned to perform heat transfer tests of both natural and forced crculaton n HX tubes. The flow path of the test faclty was desgned as Fgure 1. The man devces of the faclty ncluded a PRHR HX test secton, a canned pump, a heater, a crculaton pump, a heat exchanger n secondary sde, a pressurzer, a hgh water tank, a low water tank, a lft pump, a makeup pump and a spray pump. Fgure 1. Flow Path Chart of the PRHR HX Test Faclty There were three C-type tubes n the faclty wth the same sze as the ones n AP1000. Accordng to AP1000, there are 688 HX tubes (Φ ) and the desgn coolant flow rate s kg/h. so the max desgn flow rate of three tubes should be kg/s. In our faclty, the desgn flow rate was up to 4451

3 1.2 kg/s whch could cover the typcal flow rate of the PRHR. By adjustng the motor-drven valves VCO103, VAJ101 and VAJ102 n the prmary loop, the system resstance could be controlled, so that the test flow rate could be changed. A mass flowmeter was nstalled n front of the heater to measure the flow rate n the prmary loop. In AP1000, the PRHR heat exchanger s operated n 15.5MPa, and the desgn pressure s 17 MPa. The nlet temperature of HX tubes s 297.2, whle the outlet temperature s The test faclty preserved the full temperature and pressure condtons, and vared over the expected range for AP1000. The nlet water temperature could be adjusted by controllng the heatng power of the heater. And the prmary pressure was controlled by the pressurzer and the make-up pump. The PRHR HX test secton was the key devce of the faclty. It conssted of three C-type tubes, a coolng tank and nstruments such as thermocouples. The C-type tubes mmerged n the coolng tank whch smulated the IRWST. When the hgh temperature water flowed through the tubes, the HX test secton transferred the heat from the prmary system to the coolng tank by heatng and bolng the water n the tank. The man techncal parameters of the test secton were shown as Table I. Table I. The man techncal parameters of the HX test secton Test Condton Desgn Condton Tank Pressure(MPa) 0.1 Tank Desgn Pressure(MPa) 1 Temperature n Tank() ~100 DesgnTemperature n Tank() 100 Tube Pressure(MPa) ~15 Tube Desgn Pressure(MPa) 17 Inlet Temperature of Tubes() 150~324 DesgnTemperature n Tubes() 350 Flow rate n each tube(kg/m2s) 200~2000 Flud Deonzed Water Tank Materal 06Cr19N10 Tube Materal Inconel C-type Tubes In the test secton, the tube bundle conssted of three C-type tubes. The tubes were n the same vertcal plane. Each tube contaned a top horzontal secton, a bottom horzontal secton and a vertcal secton, whch were connected to each other by cuttng sleeves. The outer dameter of the test tube was mm and the thckness was 1.65 mm as the HX tubes n AP1000 [2-4]. The heat transfer for vertcal tubes has been researched, so ths experment manly focused on the heat transfer n horzontal sectons. Therefore, the space between horzontal tubes n AP1000 was conserved n the test secton. The centerlne-tocenterlne space between adjacent horzontal tubes was 38mm. The tube lengths were shown n Table II. The tubes respectvely smulated the shortest tube, the longest one and the center one of the PRHR HX tubes n AP1000. The materal of the tubes was Inconel 600, whle that n AP1000 was Inconel 690. There s 6 percent dfference n the tube materal thermal conductvty. Table II. The tube length 4452

4 Tube NO. Horzontal Secton Length Vertcal Secton Length Ashortest 2586mm 5284mm Bcenter 3656mm 5360mm Clongest 4724mm 5436mm The tubes were connected to the prmary system ppes by two connectors as shown n Fgure 2. A connector conssted of a vertcal cylnder tank and four nozzles as shown n Fgure 3. A top vertcal nozzle was used for ventng. Three horzontal parallel nozzles were used to connect wth the HX tubes respectvely. A sngle horzontal nozzle was used to connect wth the prmary ppe. The top connector was connected wth the outlet lne of the heater. The heated water flowed out of the heater and nto the test secton through the top connector, then out of the test secton through the bottom connector. When the hgh temperature water flowed through the tubes, the C-type tubes transferred the heat from the prmary system to the coolng tank by heatng and bolng the water n the tank. The nner-tube natural crculaton was developed from the densty dfference between the cold water n the outlet lne and the hot water n the nlet lne. The top connector was about 10.5m above lowest pont of the test loop, so the max crculaton heght was 10.5m. Wndow Overflow Connector Inlet Connector Fgure 2. The Coolng Tank Confguraton 4453

5 Fgure 3. The Connector Confguraton 2.2. Coolng Tank In Amerca, several tests for PRHR HX n AP600 have been performed. The results ndcated that when the space between adjacent tubes was large enough, the thermal effect was ndependent to each other. Typcal ptch-to-dameter raton was 1.3 to 1.5, and n AP600 P/D was 2 for adjacent tubes. The tests also ndcated that when the dstance between the tube and the tank wall was large enough, the thermal effect by the tank wall could be gnored [1]. Therefore, n our test faclty, the IRWST was predgested to a cylnder tubular tank, as shown n Fgure 2. The coolng tank conssted of three parallel vertcal tubes connected to a top and bottom horzontal tube. Each vertcal tube of the tank respectvely contaned one vertcal secton of the HX tubes, whle the horzontal tube of the tank contaned three horzontal sectons of the HX tube bundle. The sze of the horzontal tubes was Φ mm, and that of the vertcal tubes was Φ mm. The dstance between the tank wall and the HX tubes was much larger than 38mm, so the tank sze wouldn t affect the heat transfer from the HX tubes to the coolng water, whch meant that the predgested coolng tank was feasble. The secondary coolng water flowed nto the coolng tank from the nlet at the bottom, and left t from the outlet on the top flange cover. There was another outlet for the coolng water overflow. In order to observe the heat transfer phenomena n the tank, there were two wndows on the tank Thermocouples In the test secton, many temperature data needed to be acqured, ncludng the nner-tube water temperatures, the wall surface temperatures along the tubes and the pool water temperatures. In order to measure the water temperature n tubes, several thermocouple traps were mounted nto the tubes. A thermocouple trap was made of a Φ30.5mm seamless stanless steel tube wth one end closed and the other open. Several sheathed thermocouples of 0.5mm dameter were nserted nto the trap from the open end. The thermocouple traps were nserted nto and welded to the HX tubes, and supported n the centre of the tubes. By ths way, t was assured that the thermocouples measured the centre temperature n the HX tubes. In order to mnsh the trap nfluence on the flow n HX tubes, the trap sze should be as small as possble. However, there were 10 thermocouples to be nstalled n A tube, 12 n B tube, 13 n C tube. In order to contan so many thermocouples, the trap sze was desgned to be Φ30.5mm. Because 4454

6 of the thermocouple traps, the flow cross secton n the HX tubes was reduced by 3.6%, whch could be accepted. Several thermocouples were welded on the outer wall surface of the HX tubes to measure the wall surface temperatures. There were 10 thermocouples on A tube, 14 on B tube, 13 on C tube. Each thermocouple on the wall surface was located between two thermocouples nsde the HX tubes. Besdes, 18 thermocouple traps were nserted nto and welded to the tank. One thermocouple was nstalled n each trap to measure the pool water temperatures. 3. TEST PROCEDURES Before the natural crculaton test started, close valve VAJ101 and VAJ102 to solate the canned pump from the test loop. Make sure valve VCO103 open. Open valve VHO202 and VHO203, and start the lft pump to pump the water n the low water tank to fll the hgh water tank. When the hgh water tank overflowed, shut down the lft pump, and turn off the valve VHO203. Turn on valve VHO105, VHO104 and VHO 204. Because of the gravty, water n the hgh water tank flled the test loop ncludng the pressurzer, prmary ppes and the coolng tank. When the test loop was full of water, turn off valve VHO105. Feed ntrogen gas nto the pressurzer, and start the make-up pump to ncrease the prmary system pressure to the test condton. Then start the heater, whch was a vertcal ppe connected wth the postve pole of DC power at the mdpont and wth the negatve pole at two ends. When the prmary water flowed though the heater, ts temperature rose and the densty decreased, so the water flowed upwards nto the test secton. Adjust the heatng power to make sure the nlet water temperature of the test secton reached the test condton. Adjust the motor-drven valve VCO103 to change the system resstance and control the flow rate n the prmary loop. When the hot water flowed through the test secton, t was cooled by the coolng tank. The water densty ncreased, so the water flowed downwards and back nto the heater. Ths course formed the natural crculaton n the prmary loop. Meanwhle, start the crculaton pump to crculate the water n the coolng tank through the secondary heat exchanger, and turn on valve VHO303 and VHO304 to cool the tank water. The natural crculaton test n HX tubes was a steady test. When the prmary water temperature was under the saturaton temperature, there was sngle-phase flow n the prmary loop. In ths case, the natural crculaton n HX tubes was stable. If the water out of the heater was supersaturated, there was two-phase flow n the HX tubes. The prmary system pressure fluctuated strongly. In ths case, turn down valve VCO103 to ncrease the system resstance. When the system resstance was large enough, the prmary pressure stopped fluctuatng and the system became stable. When the system was stable, the computer began to record the test data every second ncludng the nnertube water temperatures, the wall surface temperatures along the tubes, the pool water temperatures, the flow rate n the prmary loop, the heatng power and so on. 4. TEST RESULTS AND ANALYSIS The purpose of the PRHR tests was to provde a database to research the heat transfer behavor nsde and outsde of the PRHR tubes. In our tests, the parameters covered the prmary pressure of MPa, the qualty n tubes up to 0.2, the nlet water temperature of , the mass flow rate of kg/s and the heatng power of up to 364 kw. The test condtons could cover the operaton range for PRHR HX n AP1000. In all tests, the system could acheve a steady state. Even f there was two-phase flow nsde tubes, the system pressure and the natural crculaton flow rate n tubes were stable. 4455

7 The PRHR heat exchanger had several dfferent modes of heat transfer. When the power of the heater was not large enough, water out of the heater couldn t acheve the saturaton temperature. So the heat transfer nsde the HX tubes was natural convecton. When the power was large enough, the prmary water out of the heater was supersaturated. So the heat transfer was nucleate bolng nsde tubes. Outsde the tubes, the heat transfer was bolng and natural convecton n pool. RELAP5/MOD3.3 has been used to model the test faclty, smulate the heat transfer nsde and outsde of the HX tubes and assess the calculaton model. There were 3 nner-tube natural crculaton test condtons calculated n Table III. Table III. Test condtons NO. Pressure(MPa) Heatng Power(kW) Inlet Temperature() Mass Flow Rate(kg/s) Qualty(x) FN FN FN Heat Transfer n the Tube The heat transfer was natural convecton or nucleate bolng n the HX tubes dependng on the heatng power. In RELAP5, the heat transfer coeffcent could be calculated by Dttus-Boelter correlaton for sngle-phase lqud turbulent forced convecton [5-7], whch s gven as: Nu DB hdb D k Re Pr (1) Where D means the nsde tube dameter, k means the flud thermal conductvty, Re means the flud Reynolds number, Pr means the flud Prandtl number. Dttus-Boelter correlaton could be appled wthn the range of parameters:0.7 Pr 160, Re 6000, L 60 (L s the tube length ). D L In our tests, Prandtl number was between 0.8 and 3.2, Reynolds number was above 18000, and 300 D. Though t was natural crculaton n tubes, the Reynolds number was very large. Therefore, t was turbulence n tubes, and Dttus-Boelter correlaton was chosen n RELAP5 for heat transfer coeffcent calculaton n these test condtons. In RELAP5, the heat transfer coeffcent could be calculated by Chen correlaton for nucleate bolng [5, 7], whch s gven as: h h h TP NB FC (2) 4456

8 Where h NB means the nucleate bolng heat transfer coeffcent, h FC means the convecton heat transfer coeffcent whch could be calculated by Dttus-Boelter correlaton. In RELAP5, nput the pressure, nlet temperature, the heatng power, the mass flow rate and the pool temperature dstrbuton as the boundary condtons, the nner-tube temperature dstrbuton along the tube length could be calculated. The calculaton results of three tubes have been shown n Fgure 4-6. From the results, t can be ndcated that the major heat transfer took place at the horzontal secton of the tubes. Along the tube length, the nner-tube temperature decreased because of the secondary coolng water. Near the nlet, the flud temperature was hgh, the heat transfer coeffcent and the heat flux were large, so the prmary temperature decreased quckly. Along the tube length, the heat flux decreased and the nnertube temperature dropped gradually slowly. Also, n the horzontal secton, the temperature of the upper water tank was hgher than the bottom, so the heat flux of the upper tube was smaller than the bottom tube. The nner-tube temperature of the upper tube was hgher than the bottom one at the same dstance from the nlet. The comparson between the test data and the calculaton results ndcated that n both sngle-phase and two-phase heat transfer condtons, the calculated nner-tube temperatures of three HX tubes agreed well wth the test data. Therefore, t can be concluded that Dttus-Boelter and Chen correlatons could be appled n RELAP5 to calculate the heat transfer nsde the tubes wthn the test condtons. Fgure 4. The Inner-tube Temperature Comparson of FN01Test 4457

9 Fgure 5. The Inner-tube Temperature Comparson of FN02Test Calculaton Measurement Calculaton Measurement Temperature o C P = 6.49 MPa Power = 203 KW M = kg/s x = C Tube Length (m) Temperature o C B Tube Length (m) P = 6.49 MPa Power = 203 KW M = kg/s x = 0.17 Temperature o C Calculaton Measurement P = 6.49 MPa Power = 203 KW M = kg/s x = A Tube Length (M) Fgure 6. The Inner-tube Temperature Comparson of FN03Test 4458

10 4.2. Heat Transfer n the Coolng Tank There were two heat transfer modes outsde of the HX tubes: bolng and natural convecton, whch depended on the tube-wall temperature. Bolng occurred at the horzontal sectons of the HX tubes. Near the HX tube nlet, the tubes had suffcent wall superheat ( T wo T s ), so the local nucleate bolng occurred. Along the tube length, the wall temperature dropped because of the secondary coolng water. When the wall superheat wasn t large enough, bolng ddn t occur. Consderng the large volume of the tank, the flow rate n the secondary crculaton was very small, so the heat transfer mode could be consdered to be natural convecton n large room. The local wall heat flux could be calculated by the steady-state energy balance equaton as below: q w mh ( H 1) DL (3) Where q w means the local wall heat flux, m means the nner-tube mass flow rate, H means the local flud enthalpy at poston whch was determned by the nner-tube temperature, means the dstance between two nner-tube temperature measure postons and +1. The bolng heat transfer coeffcent could be calculated by Chen correlaton as mentoned above. The bolng heat transfer curve could be correlated by the local wall heat flux and the wall superheat Two Ts as Fgure 7. The calculaton results by RELAP5 were also shown n t. Generally speakng, the calculaton trend agreed well wth the test data. At low superheat regon, the calculated heat flux was close to the test data. At hgh superheat regon, the calculated heat flux was hgher than the test data. When the local wall heat flux ncreased largely, the wall superheat ncrease slowed down. The wall superheat was less than 30. Compared wth the large temperature dfference between the nner-tube water and the coolng tank water, the temperature dfference n the tank water was relatvely small. Therefore, Chen correlaton was enough to calculate the bolng heat transfer n the test condtons. L Fgure 7. The Bolng Heat Transfer Curve n the Coolng Tank 4459

11 The natural convecton heat transfer coeffcent could be calculated by Churchll-Chu and McAdams correlaton n RELAP5 [5]. The natural convecton heat transfer curve could be correlated by the local wall heat flux and the dfference between the wall temperature and the secondary temperature ( Two Tf ) as Fgure 8. The calculaton results were also shown n t. The calculated heat flux was lower than the test data, and the calculaton trend agreed well wth the test data. Fgure 8. The Natural Convecton Heat Transfer Curve n the Coolng Tank 5. CONCLUSIONS Accordng to the nner-tube natural crculaton experment research of PRHR HX for AP1000, t can be concluded that: a) A PRHR heat exchanger test faclty wth 3 typcal C-type HX tubes of AP1000 has been bult. It can smulate the natural crculaton n HX tubes under the pressure of MPa, the qualty up to 0.2, the nlet water temperature of and the flow rate of kg/s, whch could cover the operaton range n AP1000. b) The experment provded hundreds of data to research the heat transfer behavor nsde and outsde of the C-type tubes. c) RELAP5 could be appled wth proper correlatons to analyss the thermal transfer n PRHR heat exchanger under the test condton. ACKNOWLEDGMENTS Ths research was supported by Large Advanced PWR Plant Natonal Scence and Technology Major Project (2011ZX ). REFERENCES 1. L.E.Hochreter, F.E.Peters and D.L.Paulsen, AP600 Passve Resdual Heat Removal Heat Exchanger Test Fnal Report, Ln Chengge and Yu Zusheng, An Advanced Passve Plant AP1000, RRYES J. and WOODS B., Testng of Passve Safety System Performance for Hgher Power Advanced Reactors. US: Oregon State Unversty, SCHULZ T L., Westnghouse AP1000 advanced passve plant, Nuclear Engneerng and Desgn, Informaton Systems Laboratores, Inc., RELAP5/MOD3.3 Code Manual. Volume IV: Models and Correlaton, December

12 6. Corlett, M. M., Hochreter, L. E. and W. A. Stewart, AP600 Passve Resdual Heat Removal Heat Exchanger Test, Fnal Report, WCAP Rev. 0, December Hao Laom. Bolng Heat Transfer and Gas (Vapor)-Lqud Two-phase Flow, August