Progress and Challenges in Predictive Thermal Hydraulic Simulations - - PowerPoint PPT Presentation

NSE

Nuclear Science & Engineering at MIT science : systems : society

Massachusetts Institute of Technology

Progress and Challenges in Predictive Thermal Hydraulic Simulations

Emilio Baglietto

Emilio Baglietto - NSE Nuclear Science & Engineering at MIT 2

A new approach to Nuclear Reactor Design

“…computational methods drive design”

PWR Reactor Vessel

http://www.neimagazine.com/

Advanced PWR Vessel

Emilio Baglietto - NSE Nuclear Science & Engineering at MIT

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.

Emilio Baglietto - NSE Nuclear Science & Engineering at MIT

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

Emilio Baglietto - NSE Nuclear Science & Engineering at MIT

DOE Sponsored Programs

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. 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 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.

Emilio Baglietto - NSE Nuclear Science & Engineering at MIT

A snapshot of the DOE Tools

Emilio Baglietto - NSE Nuclear Science & Engineering at MIT

2012 2009 2014

Emilio Baglietto - NSE Nuclear Science & Engineering at MIT

REACTOR FUEL DESIGN APPLICATIONS

Emilio Baglietto - NSE Nuclear Science & Engineering at MIT

Fuel Applications 1: Press. Drops

Extensive validation/application

• K. Ikeda et al. “Study Of Spacer Grid Span Pressure Loss Under

High Reynolds Number Flow Condition” - Proceedings of ICONE17

CASE B = Baglietto, E., 2006,

Anisotropic Turbulence Modeling for Accurate Rod Bundle Simulations, ICONE14

2 4 6 8 10 12 20000 30000 40000 50000 60000 70000 80000

DP6 (psi) CFD QKE CFD Ke CFD SST

QKE = Quadratic k-e Baglietto and Ninokata SST = Menter SST Model

• 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.

Mature Application  Tools have greatly

improved

 Models provide

confidence (2006- 2014)

 Trying to collect

guidelines to stop re-inventing the wheel (at last)

Emilio Baglietto - NSE Nuclear Science & Engineering at MIT

• 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. 15th International Topical Meeting on Nuclear Reactor Thermal Hydraulics (NURETH-15), May 12-17, 2013 – Pisa, Italy.

Fuel Applications 1: Press. Drops

Application of ASME V&V20 to Predict Uncertainties in CFD Calc.

• Objective: evaluate the uncertainty due to

the modeling, δs

• δDis 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 δS

• 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
• ptimal
• E < Uval: 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

Emilio Baglietto - NSE Nuclear Science & Engineering at MIT

Fuel Applications 2: Velocity predictions

Extensive (proprietary) validation/application

Mature Application  Large validation

experience

 Consistent Industrial

Application

 Accuracy of experimental

measurements is critical

PIV

σ²PIV/CFD=σ²CFD+σ²PIV

 σCFD,mean = 1%

Position: +2 Dh; Gap #4; Re = Re4

• 0.5
• 0.4
• 0.3
• 0.2
• 0.1

• 0.06

Normalized X [-] Normalized crossflow velocity [-] LDA PIV CFD Position: 2 Dh; Gap #4; Re = Re1

• 1.78
• 1.42
• 1.06
• 0.70
• 0.34

• 0.06

Normalized X [-] Normalized crossflow velocity [-] LDA PIV CFD

σ²PIV/CFD=σ²CFD+σ²PIV

 σCFD,mean = 1.8%

VALIDATION OF A CFD METHODOLOGY TO PREDICT FLOW FIELDS WITHIN ROD BUNDLES WITH SPACER GRIDS - C. Lascar et al.

Emilio Baglietto - NSE Nuclear Science & Engineering at MIT

• 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

Physica cally Ba Based Cl Closure Co Coeff ffici cient

RMS MS error

• rs

s of the e axial al fluc uctua uation n veloc

• cities

es. Quadratic RSM [EdF]

Fuel Applications 3: consensus

Importance of mesh quality and turbulence modeling [nothing really new]

EPRI Industrial Benchmark

Emilio Baglietto - NSE Nuclear Science & Engineering at MIT

13

Flow structures

Turbulent Jets

• Can you explain the GENX Chevrons ??

http://www.sussex.ac.uk/wcm/assets/media /313/content/9161.250x193.jpg

• A jet nozzle has a sharp edge at which the flow
• separates. The fixed, circular separation line

tends to impose axisymmetry on the initial large- scale eddies.

• Axisymmetry can be broken by corrugating the

lip of the nozzle, which breaks up axisymmetric vortices into smaller, irregular eddies.

Emilio Baglietto - NSE Nuclear Science & Engineering at MIT

Mass flow measurement by means

• f orifice plates

qm = p 4 C 1 1-b

4 d 2 2(p1 - p2)r

Emilio Baglietto - NSE Nuclear Science & Engineering at MIT

80% power level 50% power level 100% power level

Mass flow measurement by means

• f orifice plates: LES Results

Extruded 3D d 3D Base size 2D d 3D

Emilio Baglietto, Joseph William Fricano, Eugeny Sosnovsky

CFD Activities in Support of Thermal- hydraulic Modeling of SFR Fuel Bundles

Emilio Baglietto - NSE Nuclear Science & Engineering at MIT

Model Geometry

 Modeling inlet region of the test section

shown to be important

Emilio Baglietto - NSE Nuclear Science & Engineering at MIT 18

• 0.5

0.5 1 1.5 2 5 10 15 20 25 30 35 exp a b c 40% 60% 80% 100% 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

Emilio Baglietto - NSE Nuclear Science & Engineering at MIT

Distorted fuel analysis

19

(Left: Nominal geometry; Right: Deformed geometry)

• 0.05

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 Mass flow rate (kg/s)

Crossflow Plane Section Index

Nominal Fully Deformed 20

Emilio Baglietto - NSE Nuclear Science & Engineering at MIT

Irradiation-caused Deformation Consequences (coolant)

21

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.”

Emilio Baglietto - NSE Nuclear Science & Engineering at MIT

NSE

Nuclear Science & Engineering at MIT science : systems : society

Massachusetts Institute of Technology

Multiphase CFD … the grand challenge

boiling heat transfer DNB void fraction

Emilio Baglietto - NSE Nuclear Science & Engineering at MIT 23

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)

A CASL-centric view

Emilio Baglietto - NSE Nuclear Science & Engineering at MIT

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

Emilio Baglietto - NSE Nuclear Science & Engineering at MIT 25

Demonstration of GEN-I M-CFD Closure for onset of DNB

Large HQ database

Jin Yan -ISACC-2013, Xian, China

• 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

DNB inception - Nam Dinh (NCSU) Synthetic CRUD (MIT)

Emilio Baglietto - NSE Nuclear Science & Engineering at MIT

GEN-II Heat Partitioning: Improved Physical Understanding

• 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
• nly model in M-CFD.
• Erroneous representation of

physical boiling.

𝒓𝒈𝒅

′′

𝒓𝒇

′′

𝒓𝒓

′′

GEN-I

Subgrid Representation of Surface (flow boiling)

(J. Buongiorno, MIT)

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

Emilio Baglietto - NSE Nuclear Science & Engineering at MIT

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

• 3. Account for sliding bubble

effect on heat transfer and nucleation sites

Flow

• 4. Account surface quenching

after bubble departure

• 5. Account for bubble

interaction on surface

Emilio Baglietto - NSE Nuclear Science & Engineering at MIT

Pressure = 1.0 bar and 10°C Subcooling Pressure = 2.0 bar and 15°C Subcooling

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

• 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

SLIDING: Dominant effect on heat transfer and nucleation sites

Bucci, Su, 2015

Emilio Baglietto - NSE Nuclear Science & Engineering at MIT 29 Fraction of nucleation sites ACTIVE at a point in time

Tackling the grand-challenge: CHF

• Bubbles merge on heater surface prior to departure
• Indicates size of dry surface patches

𝑂𝑐

′′ = 𝑔𝑢𝑕𝑂′′

𝑄 = 1 − 𝑓−𝑂𝑐

′′𝜌𝐸𝑒 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

𝒓"𝒖𝒑𝒖 = 𝑩𝒆𝒔𝒛 𝒓"𝒘𝒃𝒒𝒑𝒔_𝒈𝒋𝒎𝒏 + (𝟐 − 𝑩𝒆𝒔𝒛)𝒓"𝑶𝒗𝒅𝒎𝒇𝒃𝒖𝒇 .. 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.