Flow Regimes and Mechanistic Predictions of Critical Heat Flux under Subcooled Flow Boiling Conditions

1
Flow Regimes and Mechanistic Predictions of
Critical Heat Flux under Subcooled Flow Boiling
Conditions

• Jean-Marie Le Corre
• Westinghouse Electric Sweden AB
• Carnegie Mellon University, Pittsburgh , USA

2
Outline

• Introduction
• Visual experiments, Flow regime types and Flow
  regime map at DNB
• Postulated mechanistic modeling at DNB
• Selected DNB model(s)
• Model validation and applications (1D and 3D)
• Conclusions and on-going work

3
Introduction

• Boiling crisis is an important limiting parameter
  in boiling system
• DNB boiling crisis under subcooled boiling
  conditions
• Physical modeling of DNB is not well established
• Mechanistic DNB prediction is useful in
  development of new fuel design
• Both 1D and 3D applications are desirable

4
Two-phase Flow Regimes at DNB
5
DNB visual experiments flow regime map

• Review of visual experiments available in the
  literature
• Various flow regimes were reported
• Dimensional analysis reveals relevant parameters
• Consistent calculation of relevant local
  parameters
• Preliminary map of flow two-phase flow regimes at
  DNB is established
• Low pressure
• Limited geometric range
• X and We as relevant parameters
• More systematic experimental work is needed

6
DNB visual experiments flow regime map

• Three main types of flow regime at DNB
• Type 1 Bubbly flow
• Type 2 Near-wall vapor clots
• Type 3 Slug flow

7
DNB visual experiments
8
DNB flow regime map
Type 1
Type 2
Type 3
9
DNB Theoretical Modeling
10
Postulated mechanistic modeling at DNB

• Various DNB physical modeling can be found in
  literature
• Most used models relied on near-wall two-phase
  flow hydrodynamics only
• Experimental evidence show that
• Various flow pattern can exist at DNB (from
  non-packed bubbly flow to slug flow)
• No near-wall macroscopic change at DNB
• Wall effect (e.g. thickness) is important
• Goal Select model in agreement with experimental
  observations

11
DNB modeling in the literature

• Theoretical studies
• Near-wall bubble crowding model (Weisman and Pei,
  1983)
• Liquid sublayer dryout model (Lee and Mudawar,
  1988)
• Many others
• Experimental studies
• Three main types of flow regime at DNB
• Dry patch formation quenching prevention has
  been mentioned for Type 1 and 3 (few theoretical
  studies)
• Bubbly layer lift-off model (Mudawar)
  hypothesized for Type 2

12
Selected DNB model

• Basic DNB modeling is based on a dry spot created
  under a nucleating bubble (1) or a vapor clots
  (2) or a vapor slug (3)
• Temperature locally increases under dry area then
  decreases due to quenching
• Quenching may be prevented in the limiting case
  (Leidenfrost)
• Resulting dry patch may spread through radial
  conduction

13
DNB physical modeling (Type 1)

• 2D transient wall thermal response to nucleation
  cycle is calculated (ADI scheme transient heat
  flux boundary conditions)
• Model consistent with wall boiling model is
  desirable
• Most needed parameters are in use in wall
  partitioning model (e.g. RPI model)

14
DNB physical modeling (Type 1)

• A limiting nucleation site is considered
  (stochastic nature)
• Domain extend to as many averaged nucleation
  sites as necessary

15
DNB physical modeling (Type 1)

• Needed (optional) constitutive relations
• Bubble departure diameter ( bubble growth rate)
• Time of evaporation
• Bubble departure frequency
• Nucleation site density
• Evaporation heat flux (transient form)
• Quenching heat flux (transient form)
• Limiting conditions

16
DNB Model Validation
17
Model validation and applications

• Model validation
• Limiting nucleation site is the key to the model
• Use detailed boiling data at DNB (bypass most
  constitutive relations) to show Leidenfrost
  effect can happen
• Model various CHF points (1D) to study the
  limiting nucleation site
• Model applications
• 1D
• 3D

18
Model validation (Del Valle Kenning data)

• Detailed boiling information were reported form
  70-95 of DNB (bubble departure diameter, bubble
  departure frequency, nucleation site density,)
• Parameters from most limiting nucleation site
  calculated from statistical distribution
• Used for DNB model validations (Type 1)
• Wall superheat around 100 C can be reached
  allowing for Leidenfrost effect (150 50 C)

19
Model validation (Del Valle data)

• 2D transient wall thermal response to Nucleation
  cycle
• DNB occurrence
• Wall thermal response immediately after DNB (dry
  patch spreading)

20
Model validation (Del Valle data)

• 2D transient wall thermal response to Nucleation
  cycle
• DNB occurrence
• Wall thermal response immediately after DNB (dry
  patch spreading)

L
1
2
3
21
Model validation (Del Valle data)
Hot spot superheat
Dry patch radius
Time
22
Model applications (on going)

• 1D applications in progress
• Look-up CHF database
• Study of limiting nucleation site
• 3D applications
• Limited by current advances in the field
• Prototype CFX-5.7.1 was used in simple geometry
• Validated at high pressure
• Validation at low pressure performed in this work
• Applied to CFD experiments at high pressure
  (DeBortoli, 1958)
• Approach to complex geometry and fuel assembly
  design

23
3-D CFX-5.7.1 validation (Bartel data)
Volumetric interfacial area
Void fraction
24
3-D DNB Model Applications (on-going)

• DeBortoli data (1958)
• Local We 2426 at DNB, local x -0.086
• Type 2 region but probably Type 1 at high
  pressure
• Modified Unals model was used
• Limiting nucleation site 2 averaged bubble
  diameter
• DNB model application
• Peak wall superheat 105 C for 0.5 mm SS heater
• Peak wall superheat 95 C for 1.0 mm SS heater

25
3-D application in complex geometries

• Not in the scope of the current research program
• No accurate prediction of CHF is expected
• Correct parametric trends and correct treatment
  of 3D effects are expected
• Approach
• Compute peak wall superheat in each near-wall
  computational cell (CFD post-processing)
• Show local weak spot relative to DNB
• Design goal is a low ( uniform) peak wall
  superheat

26
Conclusions

• Different model of DNB can apply (compete)
  depending on conditions (pressure, We, x)
• A most likely mechanism is identified for Type
  1 (and Type 3)
• Model is validated using detailed boiling data
• Definition of limiting nucleation site is the key
• Additional validations, 1D and 3D applications
  are on going
• Accurate prediction of CHF is not expected
• Help in increasing CHF performance of complex
  systems