Page 1
INGAS 18 months meeting, Paris, INGAS 18 months meeting, Paris, 20./21.5.201020./21.5.2010
Institut für Chemische VerfahrenstechnikD-70199 Stuttgart, Böblingerstr. 72
InGas 18 months meetingMay, 20th/21st 2010
Paris, France
WP B2.3: Exhaust heating/Catalyst concepts
Institute for Chemical Process Engineering of Stuttgart University, Germany
-USTUTT-
Page 2
INGAS 18 months meeting, Paris, INGAS 18 months meeting, Paris, 20./21.5.201020./21.5.2010
INGAS INtegrated GAS PowertrainINGAS INtegrated GAS Powertrain
2
• Heat exchanger experiments:– Setup for stationary heat exchanger experiments (ICVT)– Stationary results of ICVT prototype– Basic adaption of simulation model– Setup for stationary heat exchanger experiments (Delphi)– Stationary results of Delphi prototype
Activity outline
• Comparison experimental results ICVT/Delphi– Amplification factors– Pressure drop– Conversion behavior
• Conclusions/Outlook
• Appendix
Page 3
INGAS 18 months meeting, Paris, INGAS 18 months meeting, Paris, 20./21.5.201020./21.5.2010
INGAS INtegrated GAS PowertrainINGAS INtegrated GAS Powertrain
3
Setup for stationary hex experiments
FIC
Air
H2
CH4
FIC
FIC
TIR TIR
TIR
TC
Hood
TIR TIR TIR TIR
TIRTIRTIR
FID
General conditions:• Air flows up to 30 m3/h• CH4 conc. up to 5000
ppm • Inflow temperatures:
20 – 400 °C• Fuel lean operation• Hydrogen-assisted
heat up
PIR
PIR
Sensors:• 10 Thermocouples
(Type K)
• 2 pressure sensors
• THC analytics
Flowchart of test rig
Page 4
INGAS 18 months meeting, Paris, INGAS 18 months meeting, Paris, 20./21.5.201020./21.5.2010
INGAS INtegrated GAS PowertrainINGAS INtegrated GAS Powertrain
Positions of axially aligned Thermocouples
4
Δ-Pressure measurement
Additional insulation was applied around the heat exchanger
Due to severe heat losses, the burner interface (tube) was heated during the second set of stationary experiments
32,5 cm
Setup for stationary hex experiments Sensor equipment of hex prototype (ICVT)
Page 5
INGAS 18 months meeting, Paris, INGAS 18 months meeting, Paris, 20./21.5.201020./21.5.2010
INGAS INtegrated GAS PowertrainINGAS INtegrated GAS Powertrain
5
Setup for stationary hex experiments Control loop for setpoint experiments
Tmax
FIC
PID
-+
CH4
Tset
Main advantages:
Stationary temperature profile is reached much faster
Approach:
Heuristic design of
controller parameters
i.e. analysis of step
response (of yCH4,in)
Resulting amplification factor is equal to system without control:
Without control Controlled system
= f(yCH4,in,Vair)
= const.
= const.
= f(yCH4,in,Vair)adT
TA
max
adT
TA
max
.
.
Page 6
INGAS 18 months meeting, Paris, INGAS 18 months meeting, Paris, 20./21.5.201020./21.5.2010
INGAS INtegrated GAS PowertrainINGAS INtegrated GAS Powertrain
1. Startup:• Inflow temperature @ 200 °C, Volume flow @ 12 m3/h
• H2 in air for fast ignition
• H2 + CH4 in air to further heat up the system
• CH4 in air until stationary point is reached
6
# Constant parameters Varied parameters Target
I Tin, Tmax (setpoint) Air flow, yCH4 (control) Amplification factor
II Tin, yCH4 Air flow Amplification factor
III Air flow, Tmax (setpoint) Tin, yCH4 (control) Amplification factor
2. Stationary experiments performed:
3. Stationary experiments, modified (electrically heated burner interface):
• Repetition of 2.I and 2.III (II omitted due to inertia of system) Setpoint Ttube = Setpoint Tmax
Setup for stationary hex experiments Experimental procedure (ICVT)
Page 7
INGAS 18 months meeting, Paris, INGAS 18 months meeting, Paris, 20./21.5.201020./21.5.2010
INGAS INtegrated GAS PowertrainINGAS INtegrated GAS Powertrain
7
Stationary results of ICVT prototype Axial temperature: Experiments with active control
6 m3/h (GHSV: 24 000 1/h), Tin = 300 °C, Tset= 630 °C, yCH4,in= 3174 ppm (control)
Heat sink due to non-insulated connection tube
Tmax
TH,out
TC,in
TC,out
TH,in
inCinH
outHinHhex TT
TT
,,
,,
εhex = 91 %
Page 8
INGAS 18 months meeting, Paris, INGAS 18 months meeting, Paris, 20./21.5.201020./21.5.2010
INGAS INtegrated GAS PowertrainINGAS INtegrated GAS Powertrain
8
Stationary results of ICVT prototype Axial temperature: Experiments with active control
11 m3/h (GHSV: 44 000 1/h), Tin = 300 °C, Tset= 630 °C, yCH4,in= 3155 ppm (control)
Tmax
Tmax shifted to outflow channels
Smaller influence of heat sink
εhex = 84 %
Page 9
INGAS 18 months meeting, Paris, INGAS 18 months meeting, Paris, 20./21.5.201020./21.5.2010
INGAS INtegrated GAS PowertrainINGAS INtegrated GAS Powertrain
9
Stationary results of ICVT prototype Axial temperature: Experiments with active control
11 m3/h (GHSV: 44 000 1/h), Tin = 300 °C, Tset= 630 °C, yCH4,in= 3155 ppm (control)
Tmax
εhex = 81 %
Page 10
INGAS 18 months meeting, Paris, INGAS 18 months meeting, Paris, 20./21.5.201020./21.5.2010
INGAS INtegrated GAS PowertrainINGAS INtegrated GAS Powertrain
Stationary results of ICVT prototype
10
• Catalyst activity is low and steadily decreasing (see later)
• Conversion is distributed over inflow- and outflow channels
• Mass transfer limitation Active surface needs to be
increased!
Experiments with active control
• Tmax = Tset = 660 °C
Page 11
INGAS 18 months meeting, Paris, INGAS 18 months meeting, Paris, 20./21.5.201020./21.5.2010
INGAS INtegrated GAS PowertrainINGAS INtegrated GAS Powertrain
Stationary results of ICVT prototype
11
1. As can be seen in the previous results, the burner interface acts as a heat sink Coupling equation between inflow and outflow channel at U-turn:
ambwall
p
tube
outinTT
cmAk
TT
1,2,
Fitting parameter
2. The catalyst is less active than initially assumed: Mass transfer surface as well as kinetic parameters (act. energy, preexp. factor)
were modified
3. Global heat loss was increased by 3 %
4. Laminar pressure drop is accounted for by:
2
hdm
Cdzdp
Fitting parameter
Model adaptions I
Page 12
INGAS 18 months meeting, Paris, INGAS 18 months meeting, Paris, 20./21.5.201020./21.5.2010
INGAS INtegrated GAS PowertrainINGAS INtegrated GAS Powertrain
12
7 m3/h (GHSV: 28 000 1/h), Tin = 300 °C, yCH4,in= 3500 ppm
Heat sink caused by connection tube for burner
Stationary results of ICVT prototype Results with fitted model: no control, no heating
Page 13
INGAS 18 months meeting, Paris, INGAS 18 months meeting, Paris, 20./21.5.201020./21.5.2010
INGAS INtegrated GAS PowertrainINGAS INtegrated GAS Powertrain
Stationary results of ICVT prototype
13
8 m3/h (GHSV: 33 000 1/h), Tin = 300 °C, Theat=Tmax=630 °C
Results with fitted model: control, heating
Heat sink relation set to 0 in model
Page 14
INGAS 18 months meeting, Paris, INGAS 18 months meeting, Paris, 20./21.5.201020./21.5.2010
INGAS INtegrated GAS PowertrainINGAS INtegrated GAS Powertrain
Stationary results of ICVT prototype
14
Amplification factor is calculated as:
A is corrected by CH4 conversion (ΔTad referred to converted CH4)
adT
TA
max
Amplification factors
Page 15
INGAS 18 months meeting, Paris, INGAS 18 months meeting, Paris, 20./21.5.201020./21.5.2010
INGAS INtegrated GAS PowertrainINGAS INtegrated GAS Powertrain
Stationary results of ICVT prototype
15
No heating, conversion
No heating, yCH4,in
Heating, conversion
Heating, yCH4,in
• Tmax is kept constant in both experimental runs (630 °C)
yCH4,in has to change with volume flow
Inverse shape as of A vs. volume flow
• Severe decay of catalyst activity between two runs
What happend to the catalyst?
Deactivation!
CH4 conversion and yCH4,in
Page 16
INGAS 18 months meeting, Paris, INGAS 18 months meeting, Paris, 20./21.5.201020./21.5.2010
INGAS INtegrated GAS PowertrainINGAS INtegrated GAS Powertrain
Stationary results of ICVT prototype
16
Linear decay of conversion, leading to linear decay of Tmax
Thermal ageing is not likely. Spalling of washcoat?
Catalyst deterioration
Page 17
INGAS 18 months meeting, Paris, INGAS 18 months meeting, Paris, 20./21.5.201020./21.5.2010
INGAS INtegrated GAS PowertrainINGAS INtegrated GAS Powertrain
Stationary results of ICVT prototype
17
Washcoat crumbs!
Due to non-fixed spacer structures (fins), the washcoat was mechanically not stable!
Catalyst deterioration
Page 18
INGAS 18 months meeting, Paris, INGAS 18 months meeting, Paris, 20./21.5.201020./21.5.2010
INGAS INtegrated GAS PowertrainINGAS INtegrated GAS Powertrain
Setup for stationary hex experiments
Fitting of additional Thermocouples (sliding through center channels)
18
Δ-Pressure
measurement
Delphi hex at test rig
Additional insulation
Sensor equipment of hex prototype (Delphi)
Page 19
INGAS 18 months meeting, Paris, INGAS 18 months meeting, Paris, 20./21.5.201020./21.5.2010
INGAS INtegrated GAS PowertrainINGAS INtegrated GAS Powertrain
Setup for stationary hex experiments
19
Positions of additional TCs
Inflow U-turn
outflow
Positions of TCs placed by Katcon
1* 2* 3*4 5
67
# Position [cm]
Type Diameter [mm]
1 7.5 K 1
2 15 K 1
3 22.5 K 1
*: 2 TCs of same type (redundance)
# Position [cm]
Type Diameter [mm]
Inflow 0 K 1
4 20 K 0.5
5 24 K 0.5
U-turn 30 K 1
6 28 K 0.5
7 26 K 0.5
outflow 0 K 1
Axial temperature measurement (Delphi)
(All positions measured from inflow end)
Page 20
INGAS 18 months meeting, Paris, INGAS 18 months meeting, Paris, 20./21.5.201020./21.5.2010
INGAS INtegrated GAS PowertrainINGAS INtegrated GAS Powertrain
1. Startup:• Inflow temperature @ 200 °C, Volume flow @ 11 m3/h
• H2 in air for fast ignition
• H2 + CH4 in air to further heat up the system
• CH4 in air until stationary point is reached
20
# Constant parameters Varied parameters Target
I Tin, Tmax (setpoint) Air flow, yCH4 (control) Amplification factor
2. Stationary experiments performed:
Setup for stationary hex experiments Experimental procedure (Delphi)
Due to insulated burner interface, additional heating was not required!
Page 21
INGAS 18 months meeting, Paris, INGAS 18 months meeting, Paris, 20./21.5.201020./21.5.2010
INGAS INtegrated GAS PowertrainINGAS INtegrated GAS Powertrain
21
Stationary results of Delphi prototype Axial temperature: Experiments with active control
ICVT: 6 m3/h (GHSV: 24 000 1/h), Tin = 300 °C, Tset= 630 °C, yCH4,in= 3174 ppm (control)
Delphi: 7 m3/h (GHSV: 24 000 1/h), Tin = 278 °C, Tset= 630 °C, yCH4,in= 4041 ppm (control)
Page 22
INGAS 18 months meeting, Paris, INGAS 18 months meeting, Paris, 20./21.5.201020./21.5.2010
INGAS INtegrated GAS PowertrainINGAS INtegrated GAS Powertrain
22
Stationary results of Delphi prototype Axial temperature: Experiments with active control
ICVT: 11 m3/h (GHSV: 44 000 1/h), Tin = 300 °C, Tset= 630 °C, yCH4,in= 3155 ppm (control)
Delphi: 13 m3/h (GHSV: 44 000 1/h), Tin = 293 °C, Tset= 630 °C, yCH4,in= 3484 ppm (control)
Kink due to increased cell density!
Simulation result taken from DB2.5 report (m=15kg/h)..
Page 23
INGAS 18 months meeting, Paris, INGAS 18 months meeting, Paris, 20./21.5.201020./21.5.2010
INGAS INtegrated GAS PowertrainINGAS INtegrated GAS Powertrain
23
Stationary results of Delphi prototype Axial temperature: Experiments with active control
ICVT: 15 m3/h (GHSV: 60 000 1/h), Tin = 300 °C, Tset= 630 °C, yCH4,in= 3472 ppm (control)
Delphi: 18 m3/h (GHSV: 60 000 1/h), Tin = 297 °C, Tset= 630 °C, yCH4,in= 4379 ppm (control)
Page 24
INGAS 18 months meeting, Paris, INGAS 18 months meeting, Paris, 20./21.5.201020./21.5.2010
INGAS INtegrated GAS PowertrainINGAS INtegrated GAS Powertrain
24
Comparison exp. results ICVT/Delphi Amplification factors
Strong influence of axial heat conduction in wall material!
Page 25
INGAS 18 months meeting, Paris, INGAS 18 months meeting, Paris, 20./21.5.201020./21.5.2010
INGAS INtegrated GAS PowertrainINGAS INtegrated GAS Powertrain
25
Comparison exp. results ICVT/Delphi Pressure drop
Page 26
INGAS 18 months meeting, Paris, INGAS 18 months meeting, Paris, 20./21.5.201020./21.5.2010
INGAS INtegrated GAS PowertrainINGAS INtegrated GAS Powertrain
26
Comparison exp. results ICVT/Delphi Conversion behavior
Wrong measurement due to non-uniform mixing in U-turn section
Page 27
INGAS 18 months meeting, Paris, INGAS 18 months meeting, Paris, 20./21.5.201020./21.5.2010
INGAS INtegrated GAS PowertrainINGAS INtegrated GAS Powertrain
27
• Experiments with ICVT heat exchanger: Stationary profiles require long time experiments
(very slow system response) Experimental procedure with controlled Tmax is significantly faster
Insulation/Heating of top section is critical for hex performance Severe catalyst deterioration
• Experiments with Delphi heat exchanger: Results are comparable with ICVT heat exchanger Worse conversion behavior due to lower cell density (i.e. lower active surface) Much lower pressure drop than ICVT prototype Up to now no information regarding catalyst deterioration
Conclusions
• Fitting: Temperature profiles fit nicely Pressure profiles fit as well Fitting is not reliable due to severe catalyst deterioration
Page 28
INGAS 18 months meeting, Paris, INGAS 18 months meeting, Paris, 20./21.5.201020./21.5.2010
INGAS INtegrated GAS PowertrainINGAS INtegrated GAS Powertrain
28
• Dynamic experiments: ICVT prototype was damaged during first experimental attempt (see
Appendix) Heat up tests with fixed ICVT prototype Heat up tests with Delphi prototype
• Stationary experiments with Delphi heat exchanger: More experimental data to evaluate ageing/deactivation
Outlook
• Fitting: Fitting of stationary results obtained with Delphi hex Improvement of pressure relation Fitting of dynamic experiments as soon as data is available
Page 29
INGAS 18 months meeting, Paris, INGAS 18 months meeting, Paris, 20./21.5.201020./21.5.2010
INGAS INtegrated GAS PowertrainINGAS INtegrated GAS Powertrain
29
-Appendix-
Page 30
INGAS 18 months meeting, Paris, INGAS 18 months meeting, Paris, 20./21.5.201020./21.5.2010
INGAS INtegrated GAS PowertrainINGAS INtegrated GAS Powertrain
30
Characterization of burner system Thermal power output / Lambda of exhaust
Page 31
INGAS 18 months meeting, Paris, INGAS 18 months meeting, Paris, 20./21.5.201020./21.5.2010
INGAS INtegrated GAS PowertrainINGAS INtegrated GAS Powertrain
31
Characterization of burner system Mean temperature / Volume flow
Page 32
INGAS 18 months meeting, Paris, INGAS 18 months meeting, Paris, 20./21.5.201020./21.5.2010
INGAS INtegrated GAS PowertrainINGAS INtegrated GAS Powertrain
32
Damaged ICVT prototype
• Temperature in outflow channels was too high. Meltdown of channel ends due to compression of hex core Rapid heat accumulation / pressure increase
• Damage only at the very end of outflow channels Hex core was shortened and will be used again for heat up experiments
Page 33
INGAS 18 months meeting, Paris, INGAS 18 months meeting, Paris, 20./21.5.201020./21.5.2010
INGAS INtegrated GAS PowertrainINGAS INtegrated GAS Powertrain
33
PHASE I: Stationary tests without burner
• Heating up the system with high throughput of hot exhaust (~ 500°C) with low THC concentration (low amplification factor)
• Subsequently, different characteristic operating points in engine map are tested, i.e. low/middle/high load at low/middle/high rpm?
Testing different constant exhaust compositions @ λ = 1: Variation of H2 ,CO/CH4 ratio in exhaust
Can CH4 conversion be boosted by H2 / CO (similar tests at ICVT test bench?) ?
PHASE II: Dynamic tests without burner
• Can CH4 light-off be sped-up by increasing CO concentration in exhaust (+ flap) ?
• Simulating fuel shut-off under overrun conditions @ constant rpm (possible at test bench?) How long can system be kept above CH4 light-off ?
• λ spikes @ constant rpm+load: monitoring CH4 slip
Test program Delphi bench scale prototype
Page 34
INGAS 18 months meeting, Paris, INGAS 18 months meeting, Paris, 20./21.5.201020./21.5.2010
INGAS INtegrated GAS PowertrainINGAS INtegrated GAS Powertrain
34
PHASE III: Dynamic tests with burner
• Cold start @ constant rpm and burner mass flow
• Cold start under NEDC conditions
• Burner operation after cold start: Preventing extinction during fuel shut-off (overrun) @ high rpm Is addition of CH4 more effective ?
More details need to be defined after dynamic laboratory experiments!
For bench scale, a specifically designed burner system should be ordered (experience will
be gained during laboratory tests)
Test program Delphi bench scale prototype