BARC BARC PASSIVE SYSTEMS RELIABILITY ANALYSIS USING THE METHODOLOGY APSRA. A.K. Nayak, PhD

Transcription

1 BARC PASSIVE SYSTEMS RELIABILITY ANALYSIS USING THE METHODOLOGY APSRA A.K. Nayak, PhD Reactor Engineering Division Bhabha Atomic Research Centre Trombay, Mumbai , India INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France 1

2 India s Innovative Reactor The AHWR AHWR is a vertical pressure tube type, boiling light water cooled and heavy water moderated reactor using ( 233 U-Th) O 2 and (Pu-Th) O 2 fuel. GRAVITY DRIVEN WATER POOL (GDWP) STEAM DRUM FUELLING MACHINE CALANDRIA INCLINED FUEL TRANSFER MACHINE MAJOR DESIGN OBJECTIVES 1. A LARGE FRACTION OF POWER FROM THORIUM. 2. DEPLOYMENT OF SEVERAL PASSIVE SAFETY SYSTEMS 3 DAYS GRACE PERIOD. 3. NO NEED FOR EMERGENCY PLANNING IN PUBLIC DOMAIN. 4. POWER OUTPUT 300 MWe. REACTOR BUILDING FUEL BUILDING INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France 2

3 General Arrangement of AHWR Core cooling by natural circulation in Main Heat Transport System Direct steam cycle, Moderator heat recovery Decay heat Removal by Isolation Condensers Passive ECCS injection by Accumulators & GDWP; Passive Containment Coolers Moderator cooling, End Shield Cooling Calandria Vault Cooling Systems GDWP Recirculation System INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France 3

4 AHWR Reactor Building VENT SHAFT ISOLATION CONDENSER GRAVITY DRIVEN WATER POOL (GDWP) STEAM DRUM (TYP.) TAIL PIPE TOWER DOWNCOMERS PASSIVE CONCRETE COOLING SYSTEM V1 VOLUME HIGH TEMPERATURE, 285 C PRIMARY CONTAINMENT MAIN HEAT TRANSPORT (MHT) SYSTEM TAIL PIPES GROUND LEVEL REACTOR CAVITY SECONDARY CONTAINMENT Double Containment, Small V1-Volume containing High enthalpy MHT piping Large inventory of water at higher elevation in Gravity Driven Water Pool INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France 4

5 AHWR Reactor Block INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France 5

6 Passive Safety Feature Heat removal from core by natural circulation of coolant in Main Heat Transport System INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France 6

7 Passive Safety Feature Passive core decay heat removal by Isolation Condensers immersed in Gravity Driven Water Pool INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France 7

8 Passive Safety Feature Passive injection of ECC water during LOCA, initially from accumulators and later from the overhead GDWP, directly into fuel cluster. Passive Containment Isolation & Passive Containment Cooling INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France 8

9 Passive Safety Feature Passive Poison Injection in moderator during overpressure transient Passive Poison Injection System actuates during very low probability event of failure of wired shutdown systems (SDS#1 & SDS#2) and non-availability of Main condenser INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France 9

10 Passive Safety Feature Back up heat sink during low probable event of ECCS failure Use of moderator as heat sink Water in calandria vault Flooding of reactor cavity following LOCA INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France 10

11 Main Heat Transport System - AHWR vis-à-vis PHWR AHWR PHWR Light water coolant Coolant boils in core Core flow maintained by natural circulation Vertical flow channel Passive decay heat removal Core coolant transports heat directly to condenser Heavy water coolant No boiling of coolant Core flow caused by pump Horizontal flow channel Active decay heat removal Core coolant transfers heat to secondary coolant which rejects heat in condenser INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France 11

12 Safety criteria for advanced reactor systems FREQUENCY (events/year) Risk based approach Accuracy of Current PSA treatment - human reliability? Quantitative Probabilistic Safety Goal unallowable domain Advanced systems - operator action is minimized through passive systems. - reliability of passive systems must be considered. - - allowable domain Residual risk (RR) : no additional public health concerns RADIOLOGICAL CONSEQUENCES - Fig 3. Frequency vs. Consequence Safety Goal INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France 12

13 Why Passive Systems Can Fail? While, Passive systems by definition, should operate only on the basis of fundamental natural physical laws, question arises Can Such Systems Fail? Possibly no for example, gravity does not fail; buoyancy does not fail or in other words mechanism does not fail Possibly Yes for example, mechanism may not fail, but the system may not be able to carry out the required duty or defined objectives whenever called on This is called as Functional Failure of a Passive System, which can happen if the boundary conditions deviate from the specified value on which the performance of the system depends. Mainly because, the driving force of passive systems are small, which can be easily changed even with a small disturbance or change in operating parameters. INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France 13

14 Difficulties in Evaluation of Functional Failure of Passive Systems Lack of Plant Data and Operational Experience Lack of sufficient experimental data from Integral Facilities or even from Separate Effect Tests in order to understand their performance characteristics not only at normal operation but also during transients and accidents. The definition of failure mode of the systems are not well defined. Difficulty in modeling the physical behaviour of such systems; particularly, low flow natural circulation; the flow is not fully developed and can be multi-dimensional in nature flow instabilities which include flashing, geysering, density-wave, flow pattern transition instabilities, etc. critical heat flux under oscillatory condition flow stratification with kettle type of boiling particularly in large diameter vessel thermal stratification in large pools such as in GDWP effect of non-condensable gases on condensation, etc. Capability of so called Best Estimate Codes for such systems - use models applicable for active systems. - applicability for passive systems? Not well known. - Uncertainty of predictions INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France 14

15 Sources of Uncertainties Uncertainties in the best estimate codes can arise due to incapable models built-in the codes to represent a specific phenomena; absence of models to represent a particular phenomena; deviations of the input parameters due to the uncertainties of the instruments and control systems and that of the geometry of the loop; uncertainties in the material properties such as fuel thermal conductivity; fuel-to-clad gap conductance, etc. INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France 15

16 Experimental Programme for Data Generation for Assessment of Code Uncertainties BARC has built many experimental facilities for study of Natural Circulation, Flow Instabilities, CHF Under Oscillatory Condition; Condensation in Presence of Non-condensable; Behaviour of PCCS and PCIS BARC will use its best estimate codes (RELAP5 and others) to compare code prediction with test data and evaluate uncertainties. INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France 16

17 Experimental Facilities for Study of Boiling Two-phase Natural Circulation Objectives To generate date for natural circulation steady state and stability behaviour COOLING WATER IN VENT LINE BLEED LINE CONDENSER COOLING WATER OUT RELIEF VALVE RUPTURE DISC STEAM DRUM BUS BAR Major Design Parameters Design Pressure : 114 kg/cm 2 Design temperature : 315 o C Maximum Power : 80 kw Loop Diameter : 50 mm Elevation : 3000 mm Heated Section : 1000 mm COOLING WATER OUT COOLING WATER IN COOLER FILL LINE DRAIN LINE TEST SECTION BUS BAR INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France 17

18 Experimental Facilities for Study of Boiling Two-phase Natural Circulation (Contd.) OBJECTIVES: APSARA REACTOR CONDENSER STEAM DRUM Test Section Develop flow pattern transition criteria To understand the low power (Type I) and high power (type II) instabilities in natural circulation Measurement of CHF, pressure drop, void fraction and its distribution using NRG Evolution of Start-up procedure NEUTRON BEAM Flow pattern transition studies using neutron radiography Operating Parameters: Pressure : 70 bar Temperature : C Neutron Flux : 10 6 to 10 8 n/cm 2 s INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France 18

19 Experimental Facilities for Study of Boiling Two-phase Natural Circulation (Contd.) Objectives: In-phase and out-of-phase instability behaviour of parallel channels in natural circulation mode Effect of void reactivity feed back on thermal hydraulic stability Geometric Details: Number of channels : 4 Elevation : 3000 mm Pipe diameter : 25 mm Heater diameter : 12 mm Length of heater : 1000 mm Operating pressure : 15 bar INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France 19

20 Experimental Facilities for Study of Boiling Two-phase Natural Circulation (Contd.) Generation of database for performance evaluation of following Steady state performance of natural circulation in MHTS - Mass flow rate - Pressure drop - void fraction - CHF - Gravity separation of Steam-water mixture in SD STEAM DRUM GRAVITY DRIVEN WATER POOL TAIL PIPE ISOLATION CONDENSE R N2 CYLINDE R ADVANCED ACCUMULATOR Stability performance of natural circulation in MHTS - Static instability - Dynamic instability ECCS HEADER RUPTURE DISC HEADER Safety systems - Passive decay heat removal system (ICS) - ECCS FUEL CHANNEL SIMULATOR FEEDER INTEGRAL TEST LOOP INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France 20

21 Scaling Philosophy for Design A three level approach is followed (a) GLOBAL SCALING Power to Volume scaling philosophy adopted Pressure, temperature and elevation : 1:1 Volume scaling ratio : 452 (b) BOUNDARY FLOW SCALING Feed water and steam flow simulation Pressure, temperature and enthalpy : 1:1 (c ) LOCAL PHENOMENA SCALED ARE CHF Geysering, flashing, Carry-over and carry-under in steam drum, etc. INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France 21

22 Examples of Uncertainties of RELAP5/MOD3.2 with the in-house natural circulation data Absolute Frequency % Error Uncertainties have been evaluated for - steady state natural circulation, - stability of natural circulation and limited data for CHF Experimental Loop Number of steady state data points Uncertainty Power (kw) Apsara HPNCL ITL Apsara ½ 87 HPNCL 26 ~ 17 % ITL 14 % Error Example of error distribution for the test data of ITL, HPNCL and Apsara natural circulation loops INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France 22

23 Examples of Uncertainties of RELAP5/MOD3.2 with the in-house natural circulation data (Contd.) Uncertainties in code prediction for flow instabilities State or condition of flow - Stable - Unstable - Threshold of Instability Characteristics of Instabilities - Amplitude and frequency of oscillations including flow reversals - Important for simulation of CHF INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France 23

24 Examples of Uncertainties of RELAP5/MOD3.2 with the in-house natural circulation data (Contd.) How to Evaluate Uncertainties for Flow Instabilities Prediction? Current Numerical Codes are formulated based on First-Order-Numerical Discretization. They have inherent numerical problems due to - ill-posedness of basic equations - numerical diffusion - instability whether physical or numerical??? - sensitive to nodalization, etc. Capability of Best-Estimate Codes to flow instabilities are not proven even for the condition or state of instability. Characteristics of Instability???? INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France 24

25 Example of Nodalization Sensitivity of RELAP5 code for Simulation of Flow Instability Mass Flow Rate (kg/s) Steam Tail pipes Steam drums Down Comers Ring Header Number of grids in Riser 4 grids 8 grids 12 grids 36 grids 40 grids 44 grids 48 grids 52 grids 5.1 Fuel bundles Inlet Feeder pipes Time (s) INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France 25

26 Characterization of Uncertainty for Flow Instability Prediction Quantification of Uncertainties in Code Prediction for Instabilities is not possible with the current knowledge. A Qualitative Treatment Can be Given % Error Parameter EXPT. Parameter Parameter EXPT. RELAP5 Error Uncertainty < 10% Low 10%<Error<30% Medium 30%<Error<50% High >50% Severe INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France 26

27 Examples of Uncertainties in RELAP5 code for prediction of CHF induced by flow instability Tube ID (mm) Pressure (bar) Expt. CHF (kw/m2) Predicted CHF %Error Uncertainty LOW LOW MEDIUM MEDIUM LOW HIGH INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France 27

28 Assessment of Passive Systems ReliAbility (APSRA) BARC has developed a methodology for Assessment of Passive Systems ReliAbility known as APSRA. It mainly considers the functional failure of the system to carry out the desired function as the basis of the failure of the passive systems. The functional failure due to deviation of parameters are correlated with the failure of actual components through root diagnosis. The methodology relies on in-house experimental data from simulated facilities in addition to best estimate codes for evaluation of reliability. The method has been evaluated to evaluate the reliability of various Passive Systems of the AHWR. INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France 28

29 APSRA - How it works? INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France 29

30 Reliability Evaluation of Natural Circulation Using APSRA Step I Passive System For example, Natural Circulation in the MHT System of the AHWR INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France 30

31 Reliability Evaluation of Natural Circulation Using APSRA (Contd.) Step II Identification of its operational mechanisms: Natural circulation operates by difference in density in hot and cold legs (known as buoyancy force) balanced by flow resistances. Identification of its failure: Natural circulation failure in AHWR can be identified by - rise in clad surface temperature above a critical value (400 o C) or/and - occurrence of CHF by flow induced instability INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France 31

32 Reliability Evaluation of Natural Circulation Using APSRA (Contd.) Flow rate - kg/sec Step III Parameters affecting the operation Natural Circulation Performance depends on - operating pressure - fission heat - level in the steam drum - feed water temperature/ core inlet subcooling - presence of non-condensable gases - flow resistances in the system Pressure - Mpa Fig.6 Effect of pressure on primary system flow rate INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France 32

33 Reliability Evaluation of Natural Circulation Using APSRA (Contd.) Mass flow rate (kg/s) CHFR Step IV Key parameters causing the failure T sub =25 K Pressure = 70 bar - fission heat generation rate high - level in steam drum low - pressure in the system too low - feed water temperature too low or too high - concentration of non-condensables gases high MW (100% FP) MW (136% FP) MW (139% FP) Time (s) Failure can happen if these parameters exceed their limits to cause the failure as discussed in Step II 100% FP (2.6 MW) 0 136% FP (3.536 MW) 139% FP (3.614 MW) Flow oscillation induced CHF at high power INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France Time (s)

34 Reliability Evaluation of Natural Circulation Using APSRA (Contd.) How to determine the limits of the parameters - Through use of best estimate codes supplemented by experiments in order to reduce the uncertainties in the best estimate codes. - BARC has a full scaled facility of the AHWR, known as the Integral Test Loop (ITL). This facility operates at the same pressure and temperature conditions of the AHWR. - BARC also has number of experimental facilities for study of boiling two-phase natural circulation. - Experiments will be conducted in these facilities in order to confirm the limits of the parameters at which failure of natural circulation occurs. FUEL CHANNEL SIMULATOR STEAM DRUM GRAVITY DRIVEN WATER POOL ECCS HEADER TAIL PIPE FEEDER RUPTURE DISC HEADER ISOLATION CONDENS ER N2 CYLINDE R ADVANCED ACCUMULATOR INTEGRAL TEST LOOP INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France 34

35 Reliability Evaluation of Natural Circulation Using APSRA (Contd.) % Full power Step V : Generation of failure surface Success 70 Pressure (bar) Failure Subcooling (K) Failure Surface generated by taking into account 3 parameters INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France 35

36 Programme for Validation of Failure Surface with Test Data Power (kw) Range of Key Parameters to cause failure to be determined by Best Estimate Codes Set the Key Parameters To the Desired Value as the input for the experiments Experimental Facilities ITL HPNCL PCL Stable Experimental Data Unstable data Stable data Monitor the Failure Variables Benchmarking 20 Compare code prediction with test data Subcooling (K) Unstable Pressure (bar) Input to step V Failure Surface of Passive System Determine the Uncertainty and modify the failure data points Failure data point as input to Mathematical Model to generate failure surface INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France 36

37 Reliability Evaluation of Natural Circulation Using APSRA (Contd.) Step VI : Root Diagnosis After establishing the domain of failure surface, Next task is to Identify the causes for the deviation of key parameters This must be done carefully through experts judgments. The key parameters deviations are either caused by failure of some active components such as - valves, pumps, instruments, control systems, etc. Or, due to failure of some passive components such as - rupture disc, check valves, passive valves, etc. INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France 37

38 Reliability Evaluation of Natural Circulation Using APSRA (Contd.) Step VII Once the causes of failure of key parameters (either due to active components or passive devices) are known in Step V, the failure probability of the components can be evaluated in the conventional way. To evaluate the failure probability of certain components such as a globe valve at partial open positions, a new methodology is being developed. An example of event tree/fault tree for high feed water temperature or low inlet subcooling INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France 38

39 Reliability Evaluation of Natural Circulation Using APSRA (Contd.) HIGH FEED WATER TEMPERATURE FEEDTEMP w=1.150e-1 LOW FEED WAT ER FLOW VALVES FEED&STEAM SIDE LOWFEEDFLOW w=1.150e-1 FEEDVALVESMALFUNCTI ONI NG w=2.064e-10 Malfuctioning of level control Valves Check Valves in the feed water line Malfuction Condensate extration pump malfuction Feed water Pump malfuction Steam Drum Level controller malfuctioning Inadvertant opening of VALVES Isolation valves feed water side 2 LEVELCNTRL VAL CHECK VALVE w=1.240e-2 CEP-MKV FWP-MKV SD-LEVEL-CNTRL- FAIL1 w=2.018e-7 VAL-ST EAM w=2.694e-2 VAL-FEED w=2.656e-7 w= e-005* w= * w= * before level control valves - Check Valve 1 stuck close After level control valves - Check Valve2 stuck close Steam Drum Level controller-1 malfuctioning Steam Drum Level controller -2 malfuctioning Steam Drum Level controller-3 malfuctioning Inadvertant opening of C/V in the steam side of temp control heater Parallel MANUAL valve fails to remain closed Isolation valve-1 in the temp control heater feed water side fails to remain closed Isolation valve-2 in the temp control heater feed water side fails to remain closed LVL-CHV1 LVL-CHV2 SD-LVL-CNTRL1 SD-LVL-CNTRL2 SD-LVL-CNTRL3 CV-STEAM MANUAL VALVE-STEAM ISOVAL1-FEED ISOVAL2-FEED r= r= r= r= r= r= r= r= r= INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France 39

40 Reliability Evaluation of Natural Circulation Using APSRA (Contd.) Subcooling (K) Step VIII Evaluation of Reliability Of NC System Constant % full power lines Failure frequency E-9 3E-9 2.5E-9 2E-9 1.5E-9 1E Pressure (bar) 5E E-10 INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France 40

41 APSRA applications: other examples Water level in tank (% of design value) Failure region Isolation Condenser Passive Containment Isolation System (PCIS) Failure region Failure due to insufficient V1-V2 pressure differential to raise water to spill into duct % of Non-condensables Success region Failure due to insufficient inventory in the tank to form liquid seal Success region GDWP water temperature ( o C) % Break size Failure region % Height Exposure of IC Tubes Failure probability for IC to maintain Hot-SD ~ 8x1e-7/ yr INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France 41

42 RMPS vs. APSRA There are certain points which are common in both the methodologies; for example, treatment of the functional failure as the failure of the system identification of functional failure criteria evaluation of uncertainties in code prediction Consideration of uncertainties in prediction of functional failure of system. However, there are differences; for example, treatment of deviation of key parameters causing the failure generation of failure data/surface consideration of test data/code-to-code differences for calculation of uncertainties. INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France 42

43 INPRO Consultancy Meeting JUNE 16-17, 2008, Cadarache, France 43