NSE Nuclear Science & Engineering at MIT science : systems : society Progress and Challenges in Predictive Thermal Hydraulic Simulations Massachusetts Institute of Technology Emilio Baglietto
A new approach to Nuclear Reactor Design computational methods drive design PWR Reactor Vessel http://www.neimagazine.com/ Advanced PWR Vessel 2
CFD for Nuclear Applications computational methods drive design Lumped parameter approaches are still the base for reactor design and licensing. 3-Dimensional virtual reactor models are necessary to reduce operating costs. 3-D TH phenomena can cause fatigue cracking, pipe deformations, and additionally lead to anticipated equipment failure. Developing a mitigation strategy requires understanding the mechanisms that lead to the failure: unsteady, 3- dimensional turbulent effects.
Extended range Twin Operations (ETOPS) aka Engines Turning Or Passengers Swimming Extensive use of Predictive Simulation have allowed granting of this ETOPS capability prior to the A350 entrance in service www.youtube.com /user/worltop10
DOE Sponsored Programs Aims to address key challenges of nuclear energy industry, through new M&S technology insights. CASL will deploy a technology step change (VERA) that supports today s nuclear energy industry and accelerates future advances in the development of this cleaner energy source. larger reliance on legacy physics codes early on the program, with selective development of new codes and models NEAMS provides support relevant to both reactor and fuel cycle R&D programs by creating analytic tools, codes and methods for use by scientists and engineers who need to simulate nuclear energy systems. NEAMS is developing a computational ToolKit which is comprised of both reactor and fuel systems analysis capabilities that can be exercised either coupled or independently, depending on the needs of the end user. includes the entire fuel cycle, as well as advanced reactors. Timeline is therefore a longer one, to support a larger, challenging and continuously evolving scope.
A snapshot of the DOE Tools
2009 2012 2014
REACTOR FUEL DESIGN APPLICATIONS
Fuel Applications 1: Press. Drops Extensive validation/application Mature Application Tools have greatly improved Models provide confidence (2006-2014) Trying to collect guidelines to stop re-inventing the wheel (at last) 12 10 8 6 4 2 CASE B = Baglietto, E., 2006, Anisotropic Turbulence Modeling for Accurate Rod Bundle Simulations, ICONE14 K. Ikeda et al. Study Of Spacer Grid Span Pressure Loss Under High Reynolds Number Flow Condition - Proceedings of ICONE17 DP6 (psi) CFD QKE CFD Ke CFD SST QKE = Quadratic k-e Baglietto and Ninokata SST = Menter SST Model 0 20000 30000 40000 50000 60000 70000 80000 R. Sugrue, M. Conner, J. Yan, E. Baglietto, 2013 - Pressure Drop Measurements and CFD Predictions for PWR Structural Grids, LWR Fuel Performance Meeting TopFuel, Sept. 15-19, 2013 - Charlotte, NC.
Fuel Applications 1: Press. Drops Application of ASME V&V20 to Predict Uncertainties in CFD Calc. First-of-kind calculation of uncertainties related to a CFD calculation for nuclear fuel application in the open literature with the ASME V&V20 method CFD modeling to predict pressure losses in rod bundle is optimal E < U val : E is lower than the upper limit of the possible error due to the CFD modeling assumptions and approximations Modeling error within the "noise level" imposed by the numerical, input, and experimental uncertainties Improving the CFD modeling is not possible without an improvement on the numerical, geometric and experimental errors Objective: evaluate the uncertainty due to the modeling, δ s δ D is known from AREVA s large amount of PLC experiments E = validation comparison error is known from AREVA PLC validation between CFD & experiment ASME V&V20 provides a method to evaluate components of δ S C. Lascar, E. Jan, K. Goodheart, T. Keheley, M. Martin, A. Hatman, A. Chatelain, E. Baglietto, 2013 - Example of Application of the ASME V&V20 to Predict Uncertainties in CFD Calculations, Proc. 15 th International Topical Meeting on Nuclear Reactor Thermal Hydraulics (NURETH-15), May 12-17, 2013 Pisa, Italy.
Fuel Applications 2: Velocity predictions Extensive (proprietary) validation/application Mature Application Large validation experience Consistent Industrial Application Accuracy of experimental measurements is critical VALIDATION OF A CFD METHODOLOGY TO PREDICT FLOW FIELDS WITHIN ROD BUNDLES WITH SPACER GRIDS - C. Lascar et al. σ² PIV/CFD =σ² CFD +σ² PIV PIV σ CFD,mean = 1% σ² PIV/CFD =σ² CFD +σ² PIV σ CFD,mean = 1.8% Normalized crossflow velocity [-] Normalized crossflow velocity [-] 0.5 0.4 0.3 0.2 0.1 0.0-0.1-0.2-0.3-0.4-0.5 Position: +2 D h ; Gap #4; Re = Re4 LDA PIV CFD -0.06 0.00 0.06 0.12 0.18 0.24 0.30 0.36 0.42 0.48 0.54 Normalized X [-] Position: 2 D h ; Gap #4; Re = Re1 LDA PIV CFD 1.46 1.10 0.74 0.38 0.02-0.34-0.70-1.06-1.42-1.78-0.06 0.00 0.06 0.12 0.18 0.24 0.30 0.36 0.42 0.48 0.54 Normalized X [-]
Fuel Applications 3: consensus Importance of mesh quality and turbulence modeling [nothing really new] Grid quality and consistency is essential for robust application [experience!, no tets!!] Importance of Anisotropic approach, based on physical representation Demonstrates improved prediction at all locations, including Turbulence Levels EPRI Industrial Benchmark Physically Based Closure Coefficient Quadratic RSM [EdF] RMS errors of the axial fluctuation velocities.
Flow structures Turbulent Jets Can you explain the GENX Chevrons?? A jet nozzle has a sharp edge at which the flow separates. The fixed, circular separation line tends to impose axisymmetry on the initial largescale eddies. Axisymmetry can be broken by corrugating the lip of the nozzle, which breaks up axisymmetric vortices into smaller, irregular eddies. http://www.sussex.ac.uk/wcm/assets/media /313/content/9161.250x193.jpg 13
Mass flow measurement by means of orifice plates q m = p 4 C 1 1- b d 2 2(p 4 1 - p 2 )r
Mass flow measurement by means of orifice plates: LES Results 100% power level 80% power level Extruded 3D d 3D Base size 2D d 3D 50% power level
CFD Activities in Support of Thermalhydraulic Modeling of SFR Fuel Bundles Emilio Baglietto, Joseph William Fricano, Eugeny Sosnovsky
Model Geometry Modeling inlet region of the test section shown to be important
2.5% 7.5% 12.5% 17.5% 22.5% 27.5% In-Bundle Comparison (2014) Compare to 36 different thermocouples for each case Plot below shows the experimental measurement for each thermocouple matches the at least one of the CFD probes Analyzed the complete data set CDF of all the error of the measurement and nearest probe for all data points for all 7 cases 2 1.5 1 0.5 0-0.5 100% exp a 80% 60% b c 40% 0 5 10 15 20 25 30 35 18
Distorted fuel analysis 19
Crossflow Plane Section Index Mass flow rate (kg/s) -0.05 0 0.05 0.1 0.15 0.2 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Nominal Fully Deformed (Left: Nominal geometry; Right: Deformed geometry) 20
Irradiation-caused Deformation Consequences (coolant) Parameter Value Pressure drop -2.04% Hot channel outlet temperature +6.99K Average mass crossflux -11.4% Sodium temperature penalty factor 1.058 The sodium temperature penalty factor is: The ratio of the hottest subchannel s outlet temperature increase to the nominal difference between this subchannel s inlet and outlet temperatures. 21
NSE Nuclear Science & Engineering at MIT science : systems : society boiling heat transfer void fraction DNB Multiphase CFD the grand challenge Massachusetts Institute of Technology
A CASL-centric view With contributions from: Mark Christon (LANL) Area Lead Igor Bolotnov (NCSU) Gretar Tryggvason (ND) Jacopo Buongiorno (MIT) Yassin Hassan (TAMU) Nam Dinh (NCSU) Mike Podowski (RPI) Annalisa Manera (UM) 23
Good news: mature baseline CASL Validation has Demonstrated Maturity of Closures Demonstrated Portability of Closures (STAR-CCM+) The DEBORA Test Case Results are shown below L2:THM.P7.01 Demonstration & Assessment of Advanced Modeling Capabilities for Multiphase Flow with Sub-cooled Boiling
Demonstration of GEN-I M-CFD Closure for onset of DNB Industrial Application has demonstrated: Usability of GEN-1 Closure up to onset of DNB Good trend predictions Good generality Ongoing work is looking at: Extended generality via more realistic mechanistic representation Extension to oxidized/crudded surface Incorporating realistic DNB Mechanism Synthetic CRUD (MIT) DNB inception - Nam Dinh (NCSU) Large HQ database Jin Yan -ISACC-2013, Xian, China 25
GEN-II Heat Partitioning: Improved Physical Understanding GEN-I (J. Buongiorno, MIT) GEN-II q fc q q q e Mechanistic model proposed by Judd and Hwang (1976) Adapted by Kurul and Podowski (1990) for wall heat flux partitioning during pool nucleate boiling. While limited it is de-facto the only model in M-CFD. Erroneous representation of physical boiling. Subgrid Representation of Surface (flow boiling) Key challenges/approach: Tremendously complex surface interactions, cannot be resolved by first principle: Selection of local characteristic in the CFD solution to drive the SGS Model representation Fully Mechanistic representation to extend generality and allow leveraging experimental microscale measurements Tracking of subgrid surface characteristics to: Include influence of surface evolution (oxidation, crud, etc.) Extension to CHF description as surface hydrodynamic phenomenon
GEN-II Heat Partitioning: Quick Overview 1. Mechanistic Representation of Bubble Lift off and Departure Diameters 2. Accurate evaluation of evaporation heat flux by modeling effective microlayer 5. Account for bubble interaction on surface 3. Account for sliding bubble effect on heat transfer and nucleation sites Flow 4. Account surface quenching after bubble departure
GEN-II Heat Partitioning: assessment Validation performed against MIT boiling curves Allows validating separate model components Calibration-free demonstrated generality deriving from improved physical representation Pressure = 1.0 bar and 10 C Subcooling SLIDING: Dominant effect on heat transfer and nucleation sites Pressure = 2.0 bar and 15 C Subcooling Bucci, Su, 2015 Evaporation term is not dominant contribution Effect of bubble sliding dominates Flow Boiling Heat Transfer (previously postulated by Basu) The new model demonstrates improved predictions at all conditions Enhanced robustness at higher heat fluxes
Tackling the grand-challenge: CHF Bubbles merge on heater surface prior to departure Indicates size of dry surface patches Fraction of nucleation sites ACTIVE at a point in time N b = ft g N P = 1 e N b πd d 2 complete spatial randomness methods (CSR) I can track the wet and dry surface in a cell This allows me to split the heat transfer into 2 components where q" tot = A dry q" vapor_film + (1 A dry )q" Nucleate.. as the heat flux increases, heat removed by the wetted area can t keep up, leading to larger coalescence between bubbles, and further decreases in wetted area, resulting in surface dryout. 29