analyzing the performance of glass furnaces: the role of

1

[email protected]

Analyzing the Performance of Glass Furnaces:The Role of Thermal and Chemical Constraints

Reinhard Conradt

SEMINAR FOR THE GLASS INDUSTRYChulalongkorn UniversityFriday, September 9, 2011

HG

I

KERA

MIK

G

LAS

S

IMUL

ATIO

N

• Introduction & motivation

• Discussion of constraints as seen in a zero-dimensional model

• How to practically distinguish between a thermal anda chemical constraint?(Zero-dimensional analysis of real furnaces)

• Industrial experience with a fast conversion batch

• An extended conclusion: Quo vadis glass melting process?

2






Glass Production World 124MM t/a (Germany 7MM t/a)

49%

32%

8%7% 2% 2%

Container Flat/ Float Fiber Special Art/ CrystalFrit

production capacities:1/3 Asia1/3 America (North & South)1/3 Europe (West & East), Russia

3

0 10 20 30 40 50 60 70 80 90 100 110 120 1301000

1200

1400

1600

1800

2000

2200

2400

oxy-fuel

cross fired

horseshoe flame

ener

gy d

eman

d H

in [k

Wh/

t]; fo

r yC =

0.5

ranking

energy benchmark analysis for > 100 container glass tanks

H. van Limpt, TNO Eindhoven, 2003

normalized to:50 % cullet40 % power station efficiency1 m3 O2 = 1 kWh primary energy

4

1000

abun

danc

e in %

1

510

30

50

70

9095

99

30002000Hin in kWh/t, logartithmic scale

group withH50 = 1349 kWh/tH10 = 1181 kWh/t

group withH50 = 1726 kWh/tH10 = 1587 kWh/t

5

typical measures:

• combustion space- increasing flame emmissivity- new burner concepts (top burners, FLOX burners)- staged combustion- larger combustion space

• heat recovery- regeneratio, recuperation; thermochemical recuperator (CH4 + H2O → CO + 4 H2 at T > 900 °C)- batch pre-heating- pinch analysis of all heat flows

• new furnace concepts- larger furnaces- submerged combustion- segmented melters- all-electric melters

• chemical boosting- low-enthalpy batches- tailored fast conversion batches- in-flight technology

• indirect measures- light weight containers

1600 1800 2000 2200 24006

8

10

12λ = 1.05

HNCV(CH4)

ΔHT(products)

H in

kWh p

er m

3 CH 4

T in °C

adiabatic flame temperature,air: 25 °C

6

1800 2000 2200 2400 2600 2800 30008

10

12

14

ΔHT(products)

HNCV(CH4)

H in

kWh p

er m

3 CH 4

T in °C

1300 °C120011001000temperature Tre ofpre-heated air,λ = 1.05,ΔTre = Tre - 25 °C

read ΔT0.672257 KK

T⋅+=

educts: 1 CH4 + 2.1 O2 (oxyfuel)

7

a. b. c.

e.d.

secondary gas:a. none (rererence case)b. burner sidec. flue gas sided. both sidese. undershooting

100 2150T in °C

staged combustion

a. b. c.

e.d.


0 13000ppm NOx

NOx distribution

8

a. b. c.

e.d.


0 190000ppm NOx

CO distribution

case 1 (reference):pull 455 t/d

case 2:zero port boostpull + 10 % (455 t/d)

case 3:3 top burnerspull + 10 % (455 t/d)


air fuel float glass furnace Ford Nashville; tinted glass 0.6 wt. % Fe2O3; 50 % culllet;boost and top burners: oxy-fuel (by A. Richardson, BOC, Karlovivary 2005)

top burners

9



productionin t/d

L(hot spot)in m

L(batch) kg NOx/per t

case 1 (reference) 455 14.00 13.36 7.34case 2 *) 500 14.00 13.66 6.57case 3 500 14.00 13.36 6.42case 4 550 14.15 13.36 4.95*) glas quality not maintained

case 1 (reference):pull 455 t/d

case 2:zero port boostpull + 10 % (455 t/d)

refractory walloffgasrecuperator

fuel pre-heated air

offgaspre-heated air

reaction zone

flame

FLOX® technology

conventional flame

10

FLOX®

no reaction

instableflame

conventional flame

degree of recirculation

tempe

ratur

e in °

C

CH4 + 2O2 CO2 + 2H2O - 8.99 kWh/m3

CH4 + H2O CO + 3H2 +2.31 kWh/m3

CO+ 1/2O2 CO2 - 3.17 kWh/m3

3H2+ 3/2O2 3H2O - 8.13 kWh/m3

thermochemical recuperator

- 11.30 kWh/m3

H2

CO

H2O

CH4

CO2

°C

m3

CH4+ H2O

flue gas

11

inre

off

sf fireht

ex

wuwo

wxbatch

stack

exch

BATCH PREHEATER

a

a2

a1e

f

hi

b

cd

g

cooling lehrfloat chamber

refiner

melting tank,regenerators

mass flow heated up:

mass flow cooled down:

sketch of a float glass production

12

0 2000 4000 6000 8000 10000 120000

500

1000

1500

2000

2500T

in °C

power in KW

hot stream

cold stream

gas + pre-heated air → offgas

batch → melt + batch gases

Pinch-Analyse

0 2000 4000 6000 8000 10000 120000

500

1000

1500

2000

2500

T in

°C

power in KW

hot stream

cold stream

shift

„pinch“

Pinch-Analyse

13

0 2000 4000 6000 8000 10000 120000

500

1000

1500

2000

2500T

in °C

power in KW

hot stream

cold stream

shift

„pinch“

heating duty

additionalrecoverypotential

Pinch-Analyse

advantages: high pull, high flexibility, melting of high liquidus glassesdisadvantages: relatively high energy demand, no established demonstrators yet

submerged combustion

soure: M. Lindig, Sorg Co.

14

bubblers

electrodeselectrode

flames

feeder

pull

deep refinershallow finer

fossile fuel melter

internalbatchpre-heater

batchmelter

1350 °C1600 °C 1200 °C1300 °C1200 °C

1500 °C

1000 °C

bubblers

batch batch melting quartz primary fining refining conditioningpre-heating dissolution

concept of the „Flex Melter“: segmented melting

Glass Melting Time/ Temperature History of the Fastest Particle

0

200

400

600

800

1.000

1.200

1.400

1.600

1.800

0 50 100 150 200 250 300t/min

T/°C

regular meltingfast melting

all-electric melting

• suitable for all glasse except E, ECR• pull up to 2.7 t/(m²· d)• production up to 220 t.d• 700 to 1100 kWh/t• low investment• short tank lifetime• short dwell times• difficult quartz dissolution and fining

source: M. Lindig, Sorg Co.

15

= radiation

= reflection q2

q4q3q1

q4 - q3

transferred

q2 - q1

melt

crown

batch meltcold steam through tank

hot stream throug combustion space

heat recovery

power station and heat exchanger chemical reactor

fuelair

offgas

wall losses combustion space

wall losses tank exm HmV ;,,τ

batch 25 °C

melt1300 °C tank

1500 °C

radiative heat exchange

batch flame

crown

clear melt

10000

8000

6000

kW

4000

2000

0

1 2 3 4 5 6 7 8

float glass tank,melter 1.2 × 10 × 30 m3,600 - 650 t/d8 burner pairs,air-gas fired,regenerative heat recovery,chambers not separated

6 7 8

1 2 3 4 5 6 7 8

6.875 m

16

burner no. 1:1001 m3/h (9000 kW)+ boosting;batch melting &carry-over

burner no. 2:1040 m3/h (9350 kW)

batch melting &carry-over

burner no. 3:1080 m3/h (9710 kW)

carry-over

burner no. 4:{1080 m3, 9710 kW}

burner no. 5:1103 m3/h (9917 kW)

clear melt surface

burner no. 6:{ 815 m3, 7330 kW}

clear melt surface

burner no. 7:{723m3, 6500 kW}

clear melt surface

burner no. 8:101 m3/h (980 kW)

clear melt surface,perspirating chamber

10000

8000

6000

kW

4000

2000

0

1 2 3 4 5 6

7 8

1 2 3 4 5 6 7 8

6.875 m

power distributionalong the furnace axis

17

10000

8000

6000

kW

4000

2000

0

1 2 3 4 5 6

81 2 3 4 5 6

0

1

2

3

4

5

6isotherms (average between hot and cold cycle):

1000

1100

900

800

1200 °C

6.875

cham

ber h

eight

in m

chamber no.

555045403530252015105

brick

laye

r no.

7

1 2 3 4 5 6 7 8

6.875 m

combustion space

heat flux to tank

flow and temperatures, melt surface

900 1550 1400 1200 1050 °C

130 110 60 10 -5 5 -10 kW/m²

18


• Discussion of constraints as seen in a zero-dimensional model (a)




25

chemical conversion: (1 - yC)· ΔH0chem

made available at pull temperature:Hmelt(Tex) - Hglass(T0)

Carnot type availability (reversibl; ∞ slow)

optimal technical realization; optimal pull

glass quality requireslonger residence times;hence, higher wall losses

finite time heat transfer

theoretical

lower limit Htheorthermal

chemical

Hex

0

Hr

HCARNOT

Hlimit.

Hreal

Hmin

real installation; market driven pull

melting behavior=

heat demand⊕

quartz dissolution rate⊕

gas release rate

furnace performance=

heat transfer⊕

dwell time characteristics

the success of optimization strategies essentially depends onthe nature of the constraint which has to be overcome

26

1.5 2.0 2.5 3.0 3.5

1000

1200

1400

1600

1000

1200

1400

1600

spez

ific

heat

dem

and

Hin in

kW

h pe

r t g

lass

specific pull rate r in t/(m2·d)

limit of economic efficiency

rel.

amou

nt o

f gla

ss s

old

rel. fuel consumptionfuel consumption

amou

nt of

glass

to pa

ck

limit determined by turnover rate

furnace A

furnace B

green containers


• Discussion of constraints as seen in a zero-dimensional model (b)




27

cold stream

hot streamTHT0

HH cm ⋅&

LL cm ⋅&

htαA ⋅

LL

HH

cmcm⋅⋅

&

&

1cmAα

LL

ht =⋅⋅

&

Tex

boundary conditions:• can be selected independently• the cold stream has to reach a fixed temperature Tex

problems:• an imbalance of heat capacity flows, given by the ratio

• a thermal constraint: puts an upper limit to

• chemical constraint is an intrinsic limit of the cold stream:

LH m and m &&

Lm&

trL mm && ≤

cold stream

hot streamTH = TadT0

HH cm ⋅&

)ΔH(TΔH)y(1H ex0chemCex +⋅−=

htαA ⋅Tex

Toff

LL cm ⋅&

[ ]kW ht

tkWhp; HΔTcmP exexLLex =⎥⎦

⎤⎢⎣⎡⋅⎥⎦

⎤⎢⎣⎡⋅=⋅⋅= &

in the glass furnace itself:

define the “ideal furnace”:• perfect match between heat capacity flows: • infinitely fast heat transition; αht →∞; • no chemical constraints.

LLHH cm cm ⋅=⋅ &&

( ) 75 %ƒ 1ΔTΔT

HHηPPΔTcm

ad

ex

in

exexreinadHH ≈+⋅==⇒+=⋅⋅& *

Voff / Vfuel, ηre, ...

28

cold stream

hot streamTHT0

HH cm ⋅&

LL cm ⋅&

htαA ⋅

LL

HH

cmcm⋅⋅

&

&

1cmAα

LL

ht =⋅⋅

&

Tex


problems:• an imbalance of heat capacity flows, given by the ratio



LH m and m &&

Lm&

trL mm && ≤

What is an im-balance of heatcapacity flows?

20 min

20 min

1450 °C

500 °C

25 °C

1250 °C

off

re

off

re

offP,

airP,F

F

re

ΔTΔT0.8

ΔTΔT

cc

(off)V(air)Vη

⋅<

⋅⎟⎟⎠

⎞⎜⎜⎝

⎛⋅⎟⎟⎠

⎞⎜⎜⎝

⎛≈

29

cold stream

hot streamTHT0

HH cm ⋅&

LL cm ⋅&

htαA ⋅

LL

HH

cmcm⋅⋅

&

&

1cmAα

LL

ht =⋅⋅

&

Tex


problems:• a mismatch of heat capacity flows, given by the ratio



LH m and m &&

Lm&

trL mm && ≤

cold stream

hot streamTH = TadT0

HH cm ⋅&

)ΔH(TΔH)y(1H ex0chemCex +⋅−=

htαA ⋅Tex

Toff

LL cm ⋅&

[ ]kW ht

tkWhp; HΔTcmP exexLLex =⎥⎦

⎤⎢⎣⎡⋅⎥⎦

⎤⎢⎣⎡⋅=⋅⋅= &

in the glass furnace itself:

define the “ideal furnace”:• perfect match between heat capacity flows: • infinitely fast heat transition; αht →∞; • no chemical constraints.

LLHH cm cm ⋅=⋅ &&

( ) 75 %ƒ 1ΔTΔT

HHηPPΔTcm

ad

ex

in

exexreinadHH ≈+⋅==⇒+=⋅⋅& *

How to derivea realistic valuefor αht ?

Voff / Vfuel, ηre, ...

30

Hex

·r

qwu

qht

qwo

qoff

qin

+ qre

glass melt

crown

q2

q4q3

q1

= radiation= reflection

q1 = flux from glassq2 = flux to glassq3 = flux from crownq4 = flux to crown

⇒ q2 – q1 = qht = flux transferred

approach by M. Lindig adopted



( )( )( )( ) 1gas

4gasgasS4

4wo 4wowoS3

3gas 4gasgasS2

2glass4glassglassS1

·qε1 ·T·εCq ·qε1 ·T·εCq

·qε1 ·T·εCq

·qε1·T·εCq

−=

−=

−=

−=

+

+

+

+

melt

*) first term: radiation; second term: reflection; reflectivity = 1 – emissivity

*)

31



( )( )( )( ) 1gas

4gasgasS4

4wo 4wowoS3

3gas 4gasgasS2

2glass4glassglassS1


·qε1 ·T·εCq

·qε1·T·εCq

−=

−=

−=

−=

+

+

+

+

melt

( ) ( )[ ]4glass

4wo

4glass

4gasSht TTB· TTA· C q −+−⋅=

CS = 57.7 kW/m2; T in 1000 K;A, B = dimensionless functions of εgas, εglass, εwo.



( )( )( )( ) 1gas

4gasgasS4

4wo 4wowoS3

3gas 4gasgasS2

2glass4glassglassS1


·qε1 ·T·εCq

·qε1·T·εCq

−=

−=

−=

−=

+

+

+

+

melt

EXAMPLE:εgas = 0.25, εglass = 0.8, εwo = 05;Tgas = 1650 °C, Tglass = 1500 °C, Two = 1580 °C;

Compare to: Hex = 590 kWh/t, r = 3 t / m2· d) ⇒ qex = Hex· r = 74 kW/m2

qht = 106 kW/m2

This may constitute a thermal constraintfor glass melting.

32

cold stream

hot streamTHT0

HH cm ⋅&

LL cm ⋅&

htαA ⋅Tex

What is a ”chemical constraint”?

• the cold stream has to reach a fixed temperature Tex• for a given temperature, the conversion reactions cannot

support any pull p > pR• inspite of many known details, it is not well understood yet

melting behavior=

heat demand⊕


gas release rate


heat transfer⊕


33






time p T01 T03 T05 B1 B2 B3d kg/h °C °C °C Nm3/h Nm3/h Nm3/h

0.00 5855.2 1444 1586 1629 261.6 65.2 225.00.25 5510.8 1445 1586 1624 260.8 65.2 224.40.50 5855.2 1441 1581 1619 260.3 65.3 223.80.75 6199.7 1442 1580 1619 260.5 65.2 224.11.00 5855.2 1443 1580 1619 261.7 65.1 225.01.25 6199.7 1448 1590 1625 261.6 64.9 225.01.50 5510.8 1447 1585 1624 261.2 65.3 224.71.75 5855.2 1442 1584 1621 260.2 65.1 223.82.00 6199.7 1441 1578 1619 260.2 65.1 223.82.25 6199.7 1443 1581 1618 260.7 65.2 224.32.50 5855.2 1443 1581 1618 261.7 65.5 225.12.75 6199.7 1442 1578 1618 261.8 65.4 225.23.00 5855.2 1446 1586 1622 265.5 66.3 228.23.25 5855.2 1450 1596 1629 266.5 66.6 229.23.50 5855.2 1455 1598 1632 263.9 66.1 227.03.75 5855.2 1450 1589 1624 259.7 65.0 223.34.00 5855.2 1446 1585 1621 258.1 64.5 222.04.25 5855.2 1445 1587 1621 256.6 64.1 220.64.50 6199.7 1442 1584 1618 255.4 63.9 219.64.75 5510.8 1442 1581 1617 255.3 63.9 219.6

34

Furnace data, collected over 6 weeks; melting area = 40 m²no. date p r r gas gas Hin 1/r qin Tex Toff Tre Tstack

t/d t/m²·d t/m²·h Nm3/h Nm3/d kwh/kg m²·h/t kW/m21 1 81.0 2.03 0.0844 480 11520 1.209 11.85 0.1022 2 81.0 2.03 0.0844 490 11760 1.234 11.85 0.1043 3 85.2 2.13 0.0888 500 12000 1.197 11.27 0.1064 4 94.2 2.36 0.0981 540 12960 1.169 10.19 0.1155 5 94.2 2.36 0.0981 540 12960 1.169 10.19 0.1156 6 115.7 2.89 0.1205 610 14640 1.076 8.30 0.1307 7 116.4 2.91 0.1213 620 14880 1.087 8.25 0.1328 8 95.8 2.40 0.0998 540 12960 1.150 10.02 0.1159 9 89.2 2.23 0.0929 500 12000 1.143 10.76 0.106

10 10 95.0 2.38 0.0990 520 12480 1.117 10.11 0.11111 11 95.0 2.38 0.0990 530 12720 1.138 10.11 0.11312 12 95.0 2.38 0.0990 540 12960 1.160 10.11 0.11513 13 116.1 2.90 0.1209 610 14640 1.072 8.27 0.13014 14 116.5 2.91 0.1214 610 14640 1.068 8.24 0.13015 15 116.5 2.91 0.1214 610 14640 1.068 8.24 0.13016 16 110.4 2.76 0.1150 580 13920 1.072 8.70 0.12317 17 110.4 2.76 0.1150 590 14160 1.090 8.70 0.12518 18 110.4 2.76 0.1150 590 14160 1.090 8.70 0.12519 19 98.8 2.47 0.1029 540 12960 1.115 9.72 0.11520 20 98.8 2.47 0.1029 520 12480 1.074 9.72 0.11121 21 95.8 2.40 0.0998 520 12480 1.107 10.02 0.11122 22 95.8 2.40 0.0998 525 12600 1.118 10.02 0.11223 23 94.0 2.35 0.0979 510 12240 1.107 10.21 0.10824 24 94.0 2.35 0.0979 525 12600 1.139 10.21 0.11225 25 94.0 2.35 0.0979 520 12480 1.129 10.21 0.111

80.0

85.0

90.0

95.0

100.0

105.0

110.0

115.0

120.0

0 5 10 15 20 25 30 35 40 45

t in days

prud

uctio

n in

t/d

production p in t/d

35

450

470

490

510

530

550

570

590

610

630

650

0 5 10 15 20 25 30 35 40 45

t in days

gas c

onsu

mptio

n in

m3 p

er h

gas consumption in m3 per h

1.000

1.050

1.100

1.150

1.200

1.250

1.300

0 5 10 15 20 25 30 35 40 45

t in days

H(in)

in 1

000

kWh/

t

energy input H(in) in kWh per t of glass

36

1.000

1.050

1.100

1.150

1.200

1.250

1.300

0 5 10 15 20 25 30 35 40 45

t in days

H(in)

in 1

000

kWh/

t

450

470

490

510

530

550

570

590

610

630

650

0 5 10 15 20 25 30 35 40 45

t in days

gas c

onsu

mptio

n in

m3 p

er h

energy input H(in) in kWh per t of glass

gas consumption in m3 per h

H(in)· r = q(in) ∝ r

0.100

0.105

0.110

0.115

0.120

0.125

0.130

0.135

0.0750 0.0800 0.0850 0.0900 0.0950 0.1000 0.1050 0.1100 0.1150 0.1200 0.1250

pull rate in t/m²·h

q(in)

in 1

000

kW/m

²

q(in) H(in)·r = e + f·re = 0.0427,f = 0.7164,r² = 0.84dq = ±0.0031

37

0.0 0.5 1.0 1.5 2.00

500

1000

1500

2000

2500

P in

kW

p in t/h

180 days,70 data sets

Pin = a + b· pa = 1057 kWb = 1094 kWh/tr² = 0.95

limit

pmax

Even without any modelscenario, the liner relation-ship between power andpull is well substantiated.

A retrospective long-termevaluation of furnace datayields a most valuable basisfor optimization.

0.0 0.5 1.0 1.5 2.00

500

1000

1500

2000

2500

P in

kW

p in t/h


Hex =847±5 kWh/t


limit

Pex = Hex· p

pmax

38

0.0 0.5 1.0 1.5 2.00

500

1000

1500

2000

2500

P in

kW

p in t/h


Hex =847±5 kWh/t


Ploss = Pin - Pex = a + (b – Hex)· p

limit

Pex = Hex· p

pmax

0.0 0.5 1.0 1.5 2.00

500

1000

1500

2000

2500

P in

kW

p in t/h


Hex =847±5 kWh/t



limit

Pex = Hex· p

= 581±5 kWh

pmax

39

0.0 0.5 1.0 1.5 2.00

500

1000

1500

2000

2500

P in

kW

p in t/h


Hex =847±5 kWh/t



limit

Pex = Hex· p

= 581±5 kWh

p1 pmax p2

By statistical evaluation of furnace data over an extended period of time:

By thermodynamic calculation *) from the batch; by evaluation Tex data:

By difference, the cumulative loss is obtained:

independent of p has the same slope as Ploss: slope = ( b - Hex )

Evaluating Pin according to the heat balance yields:

Pht Ploss-fire same slope as Ploss

What is left to do is to derive a threshold of Pht-MAX, to determine the slope of Pin - Pht-MAX,and to compare it to the slope of Ploss..

firelosshtinstackwxwowuexin

stackwxwowu

exexinloss

exmeltPchemCexexex

in

PPPPPPPPP

PPPPpHbaPPP

TcHyHpHP

pbaP

−=−⇒++++=

+++=>⋅−+=−=

Δ⋅+Δ⋅−=⋅=

⋅+=

0)(

)1( ; ,0

40

In order to assess Pht, let us introduce the following abbreviations:

the imbalance ratio between hot and cold stream,

a reciprocal ref. temperature for j = off, oxy-off, and air NCV

fueljjj

LL

HHHL

HVVc

z

cmcm

/

ƒ '

'

⋅=

⋅⋅

=

j = cj in

Wh m3·K

Vj/Vfuel in

m3

m3

zj in

1 103 K

off 0.49 10.6 0.52off-oxy 0.65 9.6 0.20air 0.43 3.1 0.41

500 1000 1500 2000 25000.00.20.40.60.81.01.21.41.61.82.0

m'L·cL = Hex/ΔTex

m'H·cH, oxy-fuel

m'H·cH, air-fuel

m·c/p

in kW

h/(t·K

)

Hin in kWh/t

In order to assess Pht, let us introduce the following abbreviations:

the imbalance ratio between hot and cold stream,

a reciprocal ref. temperature for j = off, oxy-off, and air

According to the theory of heat exchangers (co-current mode), the maximum transferredpower at infinitely high αht is given by

≤ ½ ≈ 1 ⇒ factor x ≤ ½

A thermal constraint is reached if the slope of the losses Ploss, (b - Hex) exceeds (1-x)· b = bimb.

As long as 0 < (b - Hex) < bimb, the only constraint is due to the imbalance of cold andhot stream heat capacity flows.

pbxPTzP

HVVc

z

cmcm

inadoffHL

MAXht

NCV

fueljjj

LL

HHHL

⋅⋅∝⋅Δ⋅⋅+

=

⋅=

⋅⋅

=

− ƒ11

/

ƒ '

'

41

case threshold of slope∂Ploss/∂p

principal constraint involved

A b – Hex = 0 Pstack independent of pull p (does notincrease with increasing power input Pin);heat transferred to the tank just followsthe pull; no thermal constraint involved;chemical constraint

B 0 < b – Hex ≤ bimb Pstack increases with Pin; an increasingportion of Pin cannot be transferred to thetank due to flow imbalance only;αht not constrained; at prevailing flowimbalance: chemical constraint

C b – Hex > bimb Pstack increases with Pin; an increasingportion of Pin cannot be transferred to thetank due to a constrained αht:thermal constraint

0.0 0.5 1.0 1.5 2.00

500

1000

1500

2000

2500

P in

kW

p in t/h


Hex =847±5 kWh/t



Pex = Hex· p

P loss, min

42

0 1 2 3 4 5 60

1000

2000

3000

4000

5000

6000

0 1 2 3 4 5 6

P in

kW

p in t/h p in t/h

two “practically identical” horseshoe flame container glassfurnaces40 m2, large regenerator 40.5 m2, medium size regenerator

Pin Pin

PexPex

Ploss =Pin - Pex

Ploss =Pin - Pex

P loss, min

P loss, min

2.4 2.6 2.8 3.01.000

1.050

1.100

1.150

2.4 2.6 2.8 3.01.100

1.150

1.200

1.250

dolomite

half-burnt dolomite

dolomite

half-burnt dolomite

ener

gy co

nsum

ption

in kW

h per

kg gl

ass

pull rate r in t/(m²·d) pull rate r in t/(m²·d)

two “practically identical” horseshoe flame container glassfurnaces40 m2, large regenerator 40.5 m2, medium size regenerator

43

6400 6500 6600 6700 6800 6900 70003000

3100

3200

3300

3400

3500

3600

3700

3800

powe

r in kW

power Pin in kW

Poff

kW

Pex

2100

2000

1900

1800

1700

1600

1500

1400

1100

the combustion spaceresponds very well; allheat required can beactually made availableto the basin (lines converge)

case study: tank 5

6300 6400 6500 6600 6700 6800 6900 7000 7100

1300

1400

1500

1600

1700

1800

1900

2000

P in

kW

Pex in kW

Poff

Pwalltoo much of the heatconveyed to the basinis lost through the peri-phery

case study: tank 5

44

0 5 100 5 10 150

2000400060008000

100001200014000160001800020000

0 5 10p in t/h

P in

kW

p in t/h p in t/h

furnace A furnace B furnace C

beforecold repairafter

cold repair

0 5 10 150

2000400060008000

100001200014000160001800020000

P in

kW

p in t/h

in

loss

ex

furnace A

furnace C

furnace B

45

4 6 8 10 12 14 16 18

1000

1500

2000sp

ecific

ener

gy de

mand

Hin i

n kW

h/t

p in t/h

furnace A

furnace C

furnace B






46

0 100 200 300 400 5002000

2200

2400

2600

2800

3000

3200sp

ecifi

c he

at c

onsu

mpt

ion

Hin in

kW

h/t

time in documentation units

normal batch pre-conditioned batch normal batch

Due to the manager‘s impatience, no significant difference was found ...

0.6 0.8 1.0 1.2 1.41800

2000

2200

2400

2600

2800

3000

3200

3400

spec

ific

heat

con

sum

ptio

n H

in in

kW

h/t

pull rate r in t/(m²·d)

normal batch: r² = 0.899±δ = 57

fast conversion batch: r² = 0.926±δ = 44

... although the new batch quickly makes the furnace more stable.

47

0 30 60 90 120

780

800

820

840

860

880

900H

in in

kW

h/t

time in days

Same type of batches in a different factory.

normal batch fast conversion batchThis time, an energy savings effect was found,however, statistically blurred. The furnace seems to run in a less stable way.

320 330 340 350 360 370 380 390 400 410

125

130

135

140

145

150

155

160

165

170

pow

er o

utpu

t QT i

n kW

production p in kg/h

pTHQ exT ⋅Δ= )(

The new batch yielded a yetunreached production ...

Evaluating the mass and power pulled from the furnace speaks a different language!

48

320 330 340 350 360 370 380 390 400 4101245

1250

1255

1260

1265

1270

1275

1280

1285

1290

1295te

mpe

ratu

re T

ex in

°C

production p in kg/h

... but nobody noticed thatTex was unneccessarily high.So the energy savings effectwas much smaller than itcould have been.

Intermediate conclusion:

Enhancing glass melting by overcomingchemical constraints (i.e., by speeding upthe conversion processes) seems to havean especially high and widely unexploitedpotential for optimization.

49






= radiation

= reflection q2

q4q3q1

q4 - q3

transferred

q2 - q1

melt

crown

batch meltcold steam through tank

hot stream throug combustion space

heat recovery

power station and heat exchanger chemical reactor

fuelair

offgas

wall losses combustion space

wall losses tank exm HmV ;,,τ

batch 25 °C

melt1300 °C tank

1500 °C

radiative heat exchange

batch flame

crown

clear melt

50

Start 880 °C 1090 °C 1375 °C

1 mm

soda qtz. lime

0

20

40

60

80

100

limestone

dolomite

sand

% ab

unda

nce

0

20

40

60

80

100

different qualitiesdense soda ash,

light soda ash

40 63 125 180 250 355 500 700 1000 2000 40 63 125 180 250 355 500 700 1000 2000µm µm

melting behavior=

heat demand⊕


gas release rate


heat transfer⊕


0 200 400 600 800 1000 1200 14000.0

0.2

0.4

0.6

0.8

1.0

degr

ee o

f tur

nove

r

T in °C

release ofchemical water

dolomitedecomposition

limestonedecomposition

silicate formation

release ofphysical water

0 200 400 600 800 1000 1200 14000.0

0.2

0.4

0.6

0.8

1.0

degr

ee o

f tur

nove

r

T in °C

release ofchemical water

dolomitedecomposition

limestonedecomposition

silicate formation

release ofphysical water

51

Na2CO3 + SiO2 → Na2SiO3 + CO2

Teq = 287 °C; ΔHchem = 185 kWh/tNa2CO3 + 2SiO2 → Na2Si2O5 + CO2

Teq = 256 °C, ΔHchem = 127 kWh/t2Na2CO3 + CaCO3 + 3SiO2 → N2CS3 + 3CO2

Teq = 281 °C, ΔHchem = 191 kWh/tNa2CO3 + 3CaCO3 + 6SiO2 → NC3S6 + 4CO2

Teq = 152 °C, ΔHchem = 128 kWh/t






Kröger 1957

860 °C Tm of Na2CO3

eutectic Na2-Ca-CO3 785 °C

eutectic S-NC3S6-NS2 760 °C< 0.4 %

< 2 %

60 %

52





Teq = 152 °C, ΔHchem = 128 kWh/t1 2 3

-8

-6

-4

-2

0

log κ

, κ in

(ohm

·cm)-1

1/T in 1000/K

atomi

c mob

ility

liqsolid,thermalbranch

solid,impuritybranch

KClaqueous

sat.

1 mol/l

distance

rea contact aturnoverreaction ×= atomic mobility

Kröger 1957




< 2 %

60 %





Teq = 152 °C, ΔHchem = 128 kWh/t1 2 3

-8

-6

-4

-2

0

log κ

, κ in

(ohm

·cm)-1

1/T in 1000/K

atomi

c mob

ility



KClaqueous

sat.

1 mol/l

distance


examples ofclever fast con-version batches ...

in food industry:instant coffee,baking mixtures

image sources: steaming.files.com; backen mit essknete.deKröger 1957




< 2 %

60 %

53






• the optimization potential of the steps takingplace prior to the high-T melting process iswidely unexploited

• option 1: developing „instant coffee“ batches• option 2: developing a prototype of a combined mixer & grider & pre-heater reactor unitworking under hydrothermal conditions and exploiting heat at T levels < 300 °C

1 2 3-8

-6

-4

-2

0

log κ

, κ in

(ohm

·cm)-1

1/T in 1000/K

atomi

c mob

ility



KClaqueous

sat.

1 mol/l

distance


Kröger 1957




< 2 %

60 %

thumb movie:limestone + soda ash + sand

(not activated)

54

25 - 800 °C

834 °C

905 °C

1085 °C

1350 °C

sand lime- soda sand lime-stone ash stone

25 - 650 °C

812 °C

844 °C

1025 °C

1350 °C

910 °C

sodaash sand cullet

sand limestone soda ash sand limestone

55



56



57



58


0

90

DTA s

ignal

in µ

V

0.0

6.0

TG

sig

nal

in m

g

sample 1 sample 2

59

1400 °C 1400 °C

200 gglass melt

at 1400 °C

50 g batch

experimental parameters:Tmelt = 1400 °Cthold = 10 und 15 minTanneal = 550 °Cqcool = -2 K/min

modified Batch-Free-Time test

batch-free timeexperiments

ref 1. 2.

4.3.

6.

5.

7.

60

black-white contrast;black = clear melt

no. 1 no. 2 no. 3

no. 4 no. 5 no. 6

no. 7 no. 8

batch-free time tests

re-evaluation of free surface by black-white contrast image analysis;

calibration area = 100 %

61

1. 2.

3. 4.

∅ = 757 mm

h= 20

00 m

m

T = const. orL = const. heating mode

batch heap (4 kg)

glass melt (7 kg)

1

34

2 5 cm

sensors

1 cm

1

2

3

4

1± 0,2 cm

62

⎟⎟⎟⎟⎟⎟

⎠

⎞

⎜⎜⎜⎜⎜⎜

⎝

⎛

002

01

,2,1,

21

21

...::::

...

...

...

n

nPPP

n

n

HHH

cccmmmTTT&&&

modeldata acquisition

step 1: retrospective analysis:heat & power balance;casting into a model

63

⎟⎟⎟⎟⎟⎟

⎠

⎞

⎜⎜⎜⎜⎜⎜

⎝

⎛

002

01

,2,1,

21

21

...::::

...

...

...

n

nPPP

n

n

HHH

cccmmmTTT&&&

data bank

model

actuali-zation

step 2: retrospective analysis:reactor (dwell time) behavior;on-line actualization of model

lab

batchhouse

tankhotstream

tank,coldstream

step 3: in-factory analysis of heat and mass flows;data acquisition and formatting

64

⎟⎟⎟⎟⎟⎟

⎠

⎞

⎜⎜⎜⎜⎜⎜

⎝

⎛

002

01

,2,1,

21

21

...::::

...

...

...

n

nPPP

n

n

HHH

cccmmmTTT&&&

data bank

model

actuali-zation

lab

batchhouse

tankhotstream

tank,coldstream

step 4: comprehension and linking of data

⎟⎟⎟⎟⎟⎟

⎠

⎞

⎜⎜⎜⎜⎜⎜

⎝

⎛

002

01

,2,1,

21

21

...::::

...

...

...

n

nPPP

n

n

HHH

cccmmmTTT&&&

data bank

model

actuali-zation

lab

batchhouse

tankhotstream

tank,coldstream

display

data & software architecture

step 5: realization: structuring of the processes

step 6: factory test and evaluation

65

Thank youfor your kind attention!

analyzing the performance of glass furnaces: the role of

Documents