analyzing the performance of glass furnaces: the role of
TRANSCRIPT
1
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
• 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?
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.
secondary gas:a. none (rererence case)b. burner sidec. flue gas sided. both sidese. undershooting
0 13000ppm NOx
NOx distribution
8
a. b. c.
e.d.
secondary gas:a. none (rererence case)b. burner sidec. flue gas sided. both sidese. undershooting
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)
case 4:5 top burnerspull + 20 % (550 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
case 3:3 top burnerspull + 10 % (455 t/d)
case 4:5 top burnerspull + 20 % (550 t/d)
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
• Introduction & motivation
• Discussion of constraints as seen in a zero-dimensional model (a)
• 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?
19
20
21
22
23
24
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
• Introduction & motivation
• Discussion of constraints as seen in a zero-dimensional model (b)
• 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?
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
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 && ≤
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
boundary conditions:• can be selected independently• the cold stream has to reach a fixed temperature Tex
problems:• a mismatch 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 ≈+⋅==⇒+=⋅⋅& *
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
q1 = flux from glassq2 = flux to glassq3 = flux from crownq4 = flux to crown
⇒ q2 – q1 = qht = flux transferred
( )( )( )( ) 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
q1 = flux from glassq2 = flux to glassq3 = flux from crownq4 = flux to crown
⇒ q2 – q1 = qht = flux transferred
( )( )( )( ) 1gas
4gasgasS4
4wo 4wowoS3
3gas 4gasgasS2
2glass4glassglassS1
·qε1 ·T·εCq ·qε1 ·T·εCq
·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.
q1 = flux from glassq2 = flux to glassq3 = flux from crownq4 = flux to crown
⇒ q2 – q1 = qht = flux transferred
( )( )( )( ) 1gas
4gasgasS4
4wo 4wowoS3
3gas 4gasgasS2
2glass4glassglassS1
·qε1 ·T·εCq ·qε1 ·T·εCq
·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⊕
quartz dissolution rate⊕
gas release rate
furnace performance=
heat transfer⊕
dwell time characteristics
33
• 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?
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
180 days,70 data sets
Hex =847±5 kWh/t
Pin = a + b· pa = 1057 kWb = 1094 kWh/tr² = 0.95
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
180 days,70 data sets
Hex =847±5 kWh/t
Pin = a + b· pa = 1057 kWb = 1094 kWh/tr² = 0.95
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
180 days,70 data sets
Hex =847±5 kWh/t
Pin = a + b· pa = 1057 kWb = 1094 kWh/tr² = 0.95
Ploss = Pin - Pex = a + (b – Hex)· p
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
180 days,70 data sets
Hex =847±5 kWh/t
Pin = a + b· pa = 1057 kWb = 1094 kWh/tr² = 0.95
Ploss = Pin - Pex = a + (b – Hex)· p
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
180 days,70 data sets
Hex =847±5 kWh/t
Pin = a + b· pa = 1057 kWb = 1094 kWh/tr² = 0.95
Ploss = Pin - Pex = a + (b – Hex)· p
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
• 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?
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
• 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?
= 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⊕
quartz dissolution rate⊕
gas release rate
furnace performance=
heat transfer⊕
dwell time characteristics
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
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
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/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
860 °C Tm of Na2CO3
eutectic Na2-Ca-CO3 785 °C
eutectic S-NC3S6-NS2 760 °C< 0.4 %
< 2 %
60 %
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/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
examples ofclever fast con-version batches ...
in food industry:instant coffee,baking mixtures
image sources: steaming.files.com; backen mit essknete.deKröger 1957
860 °C Tm of Na2CO3
eutectic Na2-Ca-CO3 785 °C
eutectic S-NC3S6-NS2 760 °C< 0.4 %
< 2 %
60 %
53
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
• 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
liqsolid,thermalbranch
solid,impuritybranch
KClaqueous
sat.
1 mol/l
distance
rea contact aturnoverreaction ×= atomic mobility
Kröger 1957
860 °C Tm of Na2CO3
eutectic Na2-Ca-CO3 785 °C
eutectic S-NC3S6-NS2 760 °C< 0.4 %
< 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
sand limestone soda ash sand limestone
sand limestone soda ash sand limestone
56
sand limestone soda ash sand limestone
sand limestone soda ash sand limestone
57
sand limestone soda ash sand limestone
sand limestone soda ash sand limestone
58
sand limestone soda ash sand limestone
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!