Aspects of Cold Cap Melting12th ESG 2014

Pavel Hrma

Pacific Northwest National Laboratory

Richland, Washington, USA

23 September 2014

Hanford Tank Waste Treatment

and Immobilization Plant

200,000 m3 of radioactive

waste from plutonium

production: 1943 − 1987

Photos provided by handfordvitplant.com from Feb. 2013 (top) and Aug. 2013 (bottom).

Pre-Treatment

Facility

HLW Vitrification

Facility

LAW Vitrification

Facility

Outline• Glass-melting furnace (melter) and cold cap (batch

blanket)

• Melter feed conversion to molten glass

• Mathematical modelling

• Data for modelling:– Gas generation kinetics: thermal and evolved gas analyses

– Effective heat capacity: differential scanning calorimetry

– Density and porosity

– Heat conductivity

– Viscosity

– Crystalline phases: XRD and SEM-EDS

• Model results: Cold cap temperature profile, glass production rate

• Cold cap produced in laboratory

• Bubbling

Glass-discharge

port

Electrode

s

Glass melting• Nuclear waste will be vitrified in all-electric (Joule-heated) melters.

• Glass-formers are mixed with the waste and charged into melters operating at 1150°C.

• Melter feed (the slurry batch) forms a cold cap floating on molten glass.

Picture courtesy of Jarrett Rice

Melter feed conversion to molten glass

T ≈ 100°C

T ≈ 800°C

T ≈ 960°C

T ≈ 1070°C

T ≈ 1100°C

Reacting

Feed

Primary Foam

Secondary Foam

Molten Glass Melt

Boiling Slurry

Picture adapted from R. Pokorny and P. Hrma, J. Nucl. Materials 429, 245-256 (2012).

Upper Heat Flux

Top Heat Flux

Bottom Heat Flux

Feed Mass Flux

Gas Mass Flux

Glass Mass Flux

Mathematical model: Mass and

energy balance

bbbb r

dx

vd

dt

d=+

)(ρρ 0=+ bg rr

( )2

2

dx

Td

dx

dTcjcj

dt

dTc

Eff

gg

Eff

bbbb λρ −−=

ρ is the spatial density

v is the velocity

r is the mass change rate (via chemical reactions)

c is the heat capacity

cbEff is the effective heat capacity (includes reaction heat)

x is the spatial coordinate (vertical position)

t is the time

j is the mass flux

subscriptsb and g denote the condensed phase and the gas phase

1. Cold cap gas phase and

condensed phases

(solids, molten salts,

glass-forming melt) move

vertically (1D model).

2. Condensed phases

(solids, molten salts,

glass-forming melt) move

with the same velocity.

3. Finite volume method

is simple, efficient and

adequate to problem

Coupling cold cap model with melter model

Reaction kinetics– TGA, EGA, and DSC

Reaction rates from thermogravimetric analysis (TGA), evolved gas analysis (EGA, and differential scanning calorimetry (DSC)

• The nth-order reaction kinetics satisfactorily describes most of the melting reactions

• EGA can be calibrated for quantitative analysis

7

( )

−−=∑

T

Bk

t

in

i

N

i

ii exp1

d

dα

α

effp p Tc c H α= + ∆ ∂

is the effective heat capacity

cp is the true heat capacity

∆H is the specific reaction enthalpy

Feed density and foaming

8

Triple foam layer occurs under the cold cap:

• primary foam (from trapped batch gases)

• gas cavities (moving sideways)

• secondary foam (from rising bubbles)

“Foaming curves” are obtained from

feed expansion experiments.

Feed pellets are photographed and

their profile area is measured. Pellet

volume, density, and void fraction

(porosity, p) is then computed.

m

bpρ

ρ=

ρb bulk density

ρm material density

Below: Effect of heat-treatment

temperature, T, on ρm: sample quenched

from T to room-temperature and sample

heated to T at 10 K min-1 (computed)

The rate of the feed-to-glass

conversion (the rate of

melting) is controlled by the

heat delivered to the cold

cap across the foam layer

and from the plenum space.

Heat conductivity

9

• Heat conductivity, λEff, estimated from crucible experiments

The glass-forming

melt became

connected at TP.

Foam evolved

between TP

and

TC

and collapsed

at T > TC.

Quartz dissolution and spinel formation

0.0

0.2

0.4

0.6

0.8

1.0

400 500 600 700 800 900 1000 1100 1200

Un

dis

solv

ed

qu

artz

frac

tio

n

Temperature, °C

A0-AN1 5

A0-AN1 45

A0-AN1 75

A0-AN1 150

A0-AN1 195

The legend indicates that the average

particle sizes

Quartz dissolution can be represented as

nth-order process:

0

0.01

0.02

0.03

0.04

0.05

0.06

0 200 400 600 800 1000 1200

Cry

sta

l ma

ss f

ract

ion

Temperature (°C)

Hematite, 5 K/min Hematite, 15 K/min

Spinel, 5 K/min Spinel, 15 K/min

Spinel under cold cap

11

ggss

M

F fff ϕϕη

η++= 0log

• ϕs – volume fraction

of dissolved quartz

• ϕg – volume fraction

of gas phase

• fg, fs, f0 – constants

Transient glass-forming phase viscosity

]/)(exp[ TBA SM ϕη =

Model resultsMelting rate versus cold cap

bottom temperature

No fitting coefficients were used.

Blue points represent melter data

reported by the Vitreous State

Laboratory (VSL).

0

200

400

600

800

1000

1200

0 0.01 0.02 0.03 0.04

Tem

pe

ratu

re (

°C)

Distance from cold cap bottom (m)

1100

1050

Cold cap temperature distribution

for two bottom temperatures

250

350

450

550

650

750

850

950

1050

1150

1000 1050 1100 1150

Me

ltin

g r

ate

(k

g m

-2d

ay

-1)

Bottom temperature (°C)

Updated CC

model

VSL at 1150°C

VSL at 1200°C

• Reacting feed with open pores

• Transition between reacting layer and

foaming layer

Laboratory scale melter (LSM)

Fracture through cold cap

LSM cold cap temperature profile

14

Cold cap temperature profile was obtained by

comparing optical and SEM images of

designated cold cap areas with samples heat

treated to various temperatures.

Additional check was performed by

comparing micro-XRD data of the cold cap

with XRD of heat-treated samples.

Temperature, crystallinity, and amorphous

phase distribution in LSM cold cap

15

0.4

0.5

0.6

0.7

0.8

0.9

1.0

300

500

700

900

1100

0 5 10 15 20

Am

orp

ho

us

ph

ase

fra

ctio

n

Tem

pe

ratu

re (

°C)

Distance from cold cap bottom (mm)

LSM cold cap temperature

Amorphous phase fraction

Left: micro-XRD of the LSM cold

cap compared with XRD of

heat-treated samples.

Below: Temperature and

amorphous phase distribution

in the LSM cold cap.

Effect of bubbling on melting rate

Bubbling

outlets

Production

rate

kg/m2/day

4 1060

5 1290

6 1400

1. Bubbling generates forced convection in

the molten glass • velocity gradients become steeper

• thermal boundary layer is suppressed

• cold cap bottom temperature rises

2. Bubbles from bubblers sweep the

secondary foam formed under the cold

cap into vent holes.

3. Feed is stirred into the melt at the edges

of vent holes, exposing a fraction of the

feed to high temperature.

Energy Solution

melter test data

No bubbling:

300-500 kg/m2/day

0

200

400

600

800

1,000

1,200

1,400

1,600

1,800

0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35

Calculated specific bubbled area, m2 / m

2

Sp

ec

ific

pro

du

cti

on

ra

te,

kg

/m2-d

ay ♦♦♦♦ DM1200 y = 3703x + 245

���� DM100 y = 3256x + 400

LSM-model comparison of

temperature profiles

Red points: LSM data

shifted by secondary

foam and cavity

thickness; the slope

was adjusted assuming

that the unquenched

sample had 57%

porosity (primary

foam).

Important difference: The top surface of the LSM cold cap was dry

(400°C), whereas the model surface was wet (100°C).

y = -78.385x + 1056.2

y = -33.565x + 1396.7

200

400

600

800

1000

1200

0 5 10 15 20 25

Tem

pe

ratu

re (

°C)

Distance from cold cap bottom (mm)

Cold cap model

LSM cold cap

LSM-porosity correction