Thermodynamic analysis of hydrogen production via Sorption Enhanced Chemical-Looping Reforming Pedro da Fonseca Ramos Dissertação para obtenção do Grau de Mestre em Engenharia Mecânica Júri Presidente: Professor Mário Manuel Gonçalves da Costa Orientador: Doutor Rui Pedro da Costa Neto Vogal: Professor João Luís Toste de Azevedo Outubro de 2011
82
Embed
Thermodynamic analysis of hydrogen production via …...reforming (in this case methane reforming) by chemical-looping auto-thermal reforming, with water-gas shift and carbonation
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Thermodynamic analysis of hydrogen production via Sorption
Enhanced Chemical-Looping Reforming
Pedro da Fonseca Ramos
Dissertação para obtenção do Grau de Mestre em
Engenharia Mecânica
Júri
Presidente: Professor Mário Manuel Gonçalves da Costa
Orientador: Doutor Rui Pedro da Costa Neto
Vogal: Professor João Luís Toste de Azevedo
Outubro de 2011
i
Abstract
Hydrogen production via Sorption-Enhanced Chemical-Looping reforming combines hydrocarbons
reforming (in this case methane reforming) by chemical-looping auto-thermal reforming, with water-
gas shift and carbonation reactions. Both H2 production rate and concentration depend on the chemical
equilibrium that is established in the fuel reactor and that is influenced by several parameters
(temperature, pressure, H2O/CH4 content and presence/absence of CaO).
This report focus its attention, in a first part, on understanding, through an Aspen Plus model, the
influence that those parameters have on the equilibrium composition, as well as in the thermal
optimization of the process, so that no heat demand occurs. In a second part, experiments were carried
out, in a bench-scale fluidized bed reactor, in order to demonstrate the feasibility of this process, as
well as the effects created by some of those previously mentioned parameters.
For the thermodynamically balanced system it was possible to produce a high purity H2 (> 95%) at
650°C and 5 atm, using a H2O/CH4 ratio of 2. At these conditions, the process efficiency was 77,8%
and the CO2 capture rate of 95,0%.
In the experimental part it was possible to demonstrate de advantages of mixing oxygen carrier
particles with CO2 sorbents, in order to enhance the H2 production specially at low temperatures (600
A produção de hidrogénio através do mecanismo de reformação por Sorption-Enhanced Chemical-
Looping combina a reformação de hidrocarbonetos por chemical-looping com as reacções de
deslocamento de água e de carbonatação. A taxa de produção de hidrogénio e a concentração
dependem do equilíbrio químico que se estabelece no reactor do combustível e que é influenciado por
diversos parâmetros (temperatura, pressão, razãoH2O/CH4 e presença/ausência de CaO).
Este relatório foca a sua atenção, num primeira parte, através de um modelo desenvolvido no
programa Aspen Plus, no entendimento da influência que esses parâmetros têm na composição de
equilíbrio, assim como na optimização térmica, de modo a que não seja necessário fornecer calor ao
reactor externamente. Num segunda parte, experiências foram levadas a cabo num reactor de leito
fluidizado à escala laboratorial, com o intuito de demonstrar exequibilidade deste processo e os efeitos
criados pelos parâmetros mencionados acima.
Para o caso termodinamicamente integrado, foi possível produzir hidrogénio de elevado nível de
pureza (> 95 %) a 650 °C e 5 atm, usando uma razão H2O/CH4 de 2. Nestas condições, a eficiência do
processo foi 77,8 % (relativamente ao poder calorífico inferior do metano e do hidrogénio) e a taxa de
captura de CO2 de 95,0 %.
Na parte experimental foi possível demonstrar as vantagens de misturar as partículas transportadoras
de oxigénio com os sorventes de CO2, de modo a aumentar a produção a produção de hidrogénio
especialmente a temperaturas baixas (600 °C).
Palavras-Chave: Sorption-Enhanced; Chemical-Looping; reformação de metano; simulação
iii
Agradecimentos
Gostaria de deixar os meus agradecimentos a algumas pessoas que desempenharam um papel
importante ao longo do meu percurso académico que culminou com este trabalho.
Em primeiro lugar quero agradecer muito ao Professor Magnus Rydén pela orientação e o apoio dado
ao longo da realização do trabalho, demonstrando-se extremamente disponível para quaisquer
esclarecimentos ou para qualquer discussão de ideias.
Ao professor Henrik Leion, aos estudantes de doutoramento Mehdi Arjmand e Erik Jerndal e ao
estudande de mestrado Ali Hedayati pelo seu apoio durante o trabalho experimental.
Ao Dr. Rui Neto pela sua contribuição para a revisão do trabalho, tendo estado sempre disponível para
qualquer explicação sobre o melhor modo de apresentação dos diversos pontos do trabalho.
Aos meus Pais, pela educação e apoio prestado ao longo dos anos.
Aos meus Amigos, por todos os bons momentos que passamos juntos, diversão e descontracção,
adrenalina, estudo e trabalho.
iv
Contents
List of Figures .................................................................................................................................. vi
List of Tables .................................................................................................................................... ix
Notations .......................................................................................................................................... xi
Figure 6 – Schematic description of Sorption-Enhanced Chemical-Looping ...................................... 11
Figure 7 – Model developed in Aspen Plus ........................................................................................ 22
Figure 8 – H2 dry concentration at 600°C and 1 atm in the reforming reactor for a CH4 flow of 1
kmol/s and for different conditions: ref – without H2O and CaO particles; CaO – with CaO particles
but without H2O; H2O/CH4 – with H2O at different H2O/CH4 ratios but without CaO particles; 2
H2O/CH4 + CaO – with a H2O/CH4 ratio of 2 and CaO particles. ....................................................... 26
Figure 9 – H2 dry concentration evolution with temperature with and without CaO particles at 5 atm in
the fuel reactor and for a CH4 flow of 1 kmol/s, with a NiO/CH4 ratio of 1 and with a H2O/CH4 ratio
of 2 ................................................................................................................................................... 30
Figure 10 – H2 molar flow evolution with temperature with and without CaO particles at 5 atm in the
fuel reactor and for a CH4 flow of 1 kmol/s, with a NiO/CH4 ratio of 1 and with a H2O/CH4 ratio of 2
Figure 11 –Composite Curves considering input streams at 600°C and almost null Heat Duties ......... 33
Figure 12 –Composite Curves considering input streams at 550°C, almost null Heat Duties for the
Calcination and Air reactors and a 650°C temperature in the fuel reactor ........................................... 35
Figure 13 – Reactor location inside the oven ..................................................................................... 36
Figure 14 – Schematic description of experimental setup................................................................... 37
Figure 15 – Laboratory workstation: a - oven with reactor inside; b – electronic equipment for cycle
switch; c – gases flow meters; d – gases piping system; e – cooling unit; f – general perspective; g –
console for flow regulator; h – steam generator; i – analyser; j – valves control. ................................ 38
vii
Figure 16 – Typical evoltion of O2 concentration at the reactor’s exit in a oxidation cycle ................. 39
Figure 17- H2 concentration for different temperatures from experiments carried out with a 5g of
N4MZ1400 particles and CaO particles or sand, for a CH4 flow of 200 ml/min with a H2O/CH4 ratio
of 1,8. ............................................................................................................................................... 42
Figure 18 – Evolution of the H2/(H2+CO2+CO) ratio for different temperatures from experiments
carried out with a 5g of N4MZ1400 particles and CaO particles or sand, for a CH4 flow of 200 ml/min
with a H2O/CH4 ratio of 1,8. ............................................................................................................. 43
Figure 19 – CH4, CO2 and CO concentration for different temperatures from experiments carried out
with a 5g of N4MZ1400 particles and CaO particles or sand, for a CH4 flow of 200 ml/min with a
H2O/CH4 ratio of 1,8. ........................................................................................................................ 44
Figure 20 – CH4, CO2 and CO concentration for different temperatures obtained from Aspen Plus
simulation at 1 atm in the fuel reactor, for a CH4 flow of 1 kmol/s, with a NiO/CH4 ratio of 1, with a
H2O/CH4 ratio of 1.8, with and without CaO particles ....................................................................... 44
Figure 21 - H2 concentration for different temperatures from experiments carried out with a 15g of
N4MZ1400 particles and sand, for a CH4 flow of 200 ml/min with a H2O/CH4 ratio of 1,8. ............... 46
Figure 22 - CH4, CO2 and CO concentration for different temperatures from experiments carried out
with a 15g of N4MZ1400 particles and sand, for a CH4 flow of 200 ml/min with a H2O/CH4 ratio of
In all sets, the inlet flow of methane was 1 kmol/s and the NiO/CH4 ratio of 1, as well. In neither of
them it was taken into account any limitation in the heat duty of the reactors.
When steam was added to the fuel reactor, H2 production rate increased substantially, depending on
the H2O/CH4 ratio. With a ratio of 1, the molar flow of H2 that came out of the reactor increased
85,3%, while with a ratio of 2 it enhanced 139,7%. This large variation was related with the H2O
reaction either via steam reforming or water-gas shift that consequently led to the decrease of
unreformed CH4 equilibrium fraction and the increase of CO2 molar flow. From the combination of
these two reactions, CO decreased.
The presence of CaO particles created a different effect in the equilibrium. In this case, there was
only a small increase of 5,43% in the H2 production rate comparatively with the reference value. In
fact, there was even a small increase in the unconverted methane ratio. Despite all of this, dry H2
purity level of the gas that left the reactor was very similar to the one reached with steam.
Figure 8 allows the comparison of the H2 dry concentration between the several conditions described
above.
Figure 8 – H2 dry concentration at 600°C and 1 atm in the reforming reactor for a CH4 flow of 1 kmol/s and for different
conditions: ref – without H2O and CaO particles; CaO – with CaO particles but without H2O; H2O/CH4 – with H2O at
different H2O/CH4 ratios but without CaO particles; 2 H2O/CH4 + CaO – with a H2O/CH4 ratio of 2 and CaO particles.
From the combination of both effects (steam and CaO particles) it was possible to increase from
0,884 kmol/s of H2 leaving the reactor, with a dry concentration of 46,91%, to a flow of 2,820 kmol/s
with a dry purity level of 97,4%. In this situation, H2 production rate and methane conversion ratio
ref
CaO 1 H2O/CH4
2 H2O/CH4
2 H2O/CH4 +
CaO
0
0,1
0,2
0,3
0,4
0,5
0,6
0,7
0,8
0,9
1
H2
con
cen
tra
tio
n [
% (
V/V
)]
27
were enhanced in a great extent due to the steam, at the same time the CO2 was removed by CaO
particles, which led to the CO reduction as well, allowing a very high purity.
The major advantage existed went these two effects were combined. This allowed the highest H2
production rate as well as the highest H2 purity level for the five tested cases.
5.2. Pressure change effect
The pressure is a very important parameter, since it influences not only the equilibrium to some
extent, but also the whole system dimensioning and efficiency. If the only interest is on the optimal
equilibrium conditions, the value of the pressure in the reactors should be low. However, it should be
taken into consideration that methane needed for this application is available pressurized, and that
the resulting products – H2 and CO2 – have to be at high pressures so that they can be stored. Hence,
by increasing the system pressure conditions, the work needed for the compression stage of the
products would be much lower, increasing process global efficiency.
A pressurized system also allows a proportionally higher production rate (molar flow) for the same
components dimensions comparatively with an atmospheric system, i.e. for the same volumetric
flow.
The values presented in Table 4 were determined in the reforming reactor with a temperature of
600°C and a H2O/CH4 ratio of 2. Like in all cases, the CH4 flow was 1kmol/s and a NiO/CH4 ratio of
1. In this case, the only interest was to understand the pressure influence in the equilibrium
composition.
Table 4 - Molar flows and H2 concentrations at 600°C and different pressure values in the reforming reactor for a CH4 flow
of 1 kmol/s, with a NiO/CH4 ratio of 1 and with a H2O/CH4 ratio of 2 and CaO particles.
Pressure
[atm]
Molar Flow [kmol/s] H2 Dry
Concentration H2 CH4 H2O CO CO2 CaCO3
1 2,820 0,0407 1,098 0,0164 0,0171 0,926 0,974
5 2,548 0,1121 1,227 0,0026 0,0033 0,882 0,956
7 2,471 0,1315 1,266 0,0017 0,0024 0,864 0,948
10 2,383 0,1536 1,309 0,0011 0,0016 0,844 0,938
20 2,198 0,2000 1,401 0,0005 0,0008 0,799 0,916
28
For increasing pressures, CH4 conversion was reduced due to the volumetric increase, inherent to the
reactions in hand, that is hampered at pressurized conditions. Consequently, CO flow and H2
production and concentration were also reduced. Only CO2 sequestration by particles was enhanced,
although the total amount of CO2 produced was also reduced.
5.3. Temperature variation effect
The temperature variation simulation was carried out for a H2O/CH4 ratio of 2, in a pressurized
process at 5 atm a methane flow of 1 kmol/s and a NiO/CH4 ratio of 1. Both cases, with and without
CaO particles, were tested, so that it was possible to understand the influence of these particles, when
the reforming reactor’s temperature changes. The pressure of 5 atm was chosen in order to assure the
pressure effect on these results, but without any specific reason for this particular value.
Table 5 presents the results obtained for the case with CaO particles and at different temperatures.
Table 5 - Molar flows and H2 concentrations at different temperatures and 5 atm in the fuel reactor for a CH4 flow of 1
kmol/s, with a NiO/CH4 ratio of 1 and with a H2O/CH4 ratio of 2 and CaO particles.
Temperature
[°C]
Molar Flow [kmol/s] H2 Dry
Concentration H2 CH4 H2O CO CO2 CaCO3
600 2,548 0,1121 1,227 0,0026 0,0033 0,882 0,956
625 2,558 0,1088 1,224 0,0058 0,0065 0,879 0,955
650 2,567 0,1049 1,223 0,0127 0,0124 0,870 0,952
675 2,574 0,0997 1,226 0,0263 0,0227 0,851 0,945
700 2,577 0,0924 1,238 0,0524 0,0407 0,814 0,933
725 2,572 0,0819 1,264 0,0999 0,0713 0,747 0,910
750 2,550 0,0669 1,316 0,1821 0,1232 0,628 0,872
775 2,493 0,0480 1,411 0,3145 0,2117 0,4258 0,813
800 2,378 0,0284 1,566 0,5088 0,3638 0,0990 0,725
Temperature has a major influence in reactions extension. When this parameter rises, partial
oxidation and steam reforming reactions are intensified, what leads to a higher methane conversion
rate. When the temperature went from 600°C to 750°C, unreformed methane fraction decreased from
29
11,21% to 6,69%, in presence of CaO particles. As a consequence, H2 molar flow tended to increase,
but it only happened until a certain temperature (approximately 700°C) and slowly, because at the
same time, the water-gas shift reaction, that is favoured at low temperatures (typically 200°C –
300°C), had a lower reaction rate. For the same reason, the water content started to increase slowly
(at 675°C).
Due also to the increased temperature, particles ability for CO2 capture decreased. Because of this,
and despite the lower amount of CO2 formed, the amount of CO2 in the outlet stream tended to
increase.
Besides CO2, the CO content also increased, because of a greater methane conversion and a
reduction of the reaction between CO and steam by the WGS.
In order to enable the comparison between the previous results, Table 6 is used to demonstrate the
flow values computed for the absence of CaO particles.
Table 6 - Molar flows and H2 concentrations at different temperatures and 5 atm in the fuel reactor for a CH4 flow of 1
kmol/s, with a NiO/CH4 ratio of 1 and with a H2O/CH4 ratio of 2 and without CaO particles.
Temperature
[°C]
Molar Flow [kmol/s] H2 Dry
Concentration H2 CH4 H2O CO CO2 CaCO3
600 1,302 0,3937 1,911 0,1232 0,4831 - 0,566
625 1,484 0,3363 1,843 0,1706 0,4932 - 0,598
650 1,668 0,2762 1,780 0,2274 0,4965 - 0,625
675 1,844 0,2162 1,723 0,2910 0,4928 - 0,648
700 2,003 0,1599 1,677 0,3570 0,4830 - 0,667
725 2,135 0,1112 1,642 0,4200 0,4688 - 0,681
750 2,234 0,0728 1,621 0,4753 0,4519 - 0,691
775 2,298 0,0454 1,611 0,520 0,4341 - 0,697
800 2,335 0,0274 1,611 0,556 0,4166 - 0,700
In the absence of CaO particles, the percentage of unreformed methane decreased more sharply when
the temperature went from 600°C to 750°C, from 39,37% to 7,28%, and therefore, the flow variation
of the other components was different from the previous case.
30
In this situation, the higher variation of methane conversion rate overcame the variation in the WGS
reaction rate. Thus, the flow of H2 and CO was continuously increasing, while the H2O content was
decreasing.
Only the CO2 that is formed just by the WGS reaction showed a reduction of its production rate after
reaching a maximum flow for 650°C, caused by the increase in the steam reforming reaction extent.
Another important aspect was the understanding of the effect that the presence of CO2 sorbent
particles had in the stability of H2 flow, as well as, in the increase of its concentration for lower
temperatures than 800°C, approximately. For higher temperatures, the carbonation reaction is
deprecated comparatively to the calcination reaction, and therefore there was no difference between
the two cases.
These effects can be better perceived in curves from Figures 9 and 10, that were plotted based on
values from Tables 5 and 6.
Figure 9 – H2 dry concentration evolution with temperature with and without CaO particles at 5 atm in the fuel reactor and
for a CH4 flow of 1 kmol/s, with a NiO/CH4 ratio of 1 and with a H2O/CH4 ratio of 2
Regarding H2 dry concentration in presence of CaO particles, the outlet stream increased from 72,5%
at 800°C to 95,6% at 600°C, while without their presence, it decreased from 70,0% to 56,6%, for the
same temperature variation.
0
0,1
0,2
0,3
0,4
0,5
0,6
0,7
0,8
0,9
1
600 650 700 750 800
H2
Vo
lum
etr
ic c
on
cen
tra
tio
n [
%]
Temperature [°C]
With CaO
Without CaO
31
Figure 10 – H2 molar flow evolution with temperature with and without CaO particles at 5 atm in the fuel reactor and for a
CH4 flow of 1 kmol/s, with a NiO/CH4 ratio of 1 and with a H2O/CH4 ratio of 2
Relatively to H2 flow, it had in presence of CaO a slight increase of 7,15%, reaching a maximum of
8,37% at 700°C. Without CaO particles, there was a sharp descent of 44,2% from 2,335 kmol/s to
1,302 kmol/s. At 600°C, it was possible to reach a H2 flow 1,95 times higher with the sorption-
enhanced mechanism.
For a SECLR system operating at 5 atm, with a H2O/CH4 ratio of 2 and without any restrictions in
the reactors’ heat duty, the best operating conditions are between 600°C and 650°C which allow H2
concentrations in the outlet stream of over 95% for similar molar flows reached at a higher
temperature (700°C). Furthermore, if unconverted methane could be easily retrieved from this outlet
stream and reused, it would even result in lower CH4 inlet flow for the same results.
5.4. Thermal optimization
The three reactors have different heat requirements due to the endothermic and exothermic reactions
that take place in each one. In one hand, the air reactor usually releases heat, because the oxidation of
Ni particles into NiO is an exothermic process. On the other hand, calcination reaction is
endothermic and therefore, a heat supply is required for the release of the previously captured CO2 in
the calcination reactor. Furthermore, in the fuel reactor, there is also an heat demand, because of the
partial oxidation and steam reforming, which are both endothermic in this case, despite the
exothermic behaviour of the WGS and carbonation reactions.
0
0,5
1
1,5
2
2,5
3
600 650 700 750 800
H2
Mo
lar
Flo
w [
km
ol/
s]
Temperature [°C]
With CaO
Without CaO
32
For a self-sufficient process, the heat release from the air reactor should be enough to compensate the
heat demand from the other reactors. Additionally, in order to avoid complex systems for this heat
exchange between reactors, the flow of particles from one reactor to another should be optimized in a
way, so that no extra heat is required.
For this optimization, it was also taken into consideration the possible pre-heating of inlet streams
through a heat exchange with outlet streams. It was assumed that all streams were pre-heated at the
same temperature, and therefore, it was possible with a composite curves analysis to determine the
amount of heat exchanged between these hot and cold streams.
The method of setting the pre-heating temperature was an iterative process. Firstly, inlet streams
were assumed to be pre-heated to certain temperature and based on this assumption, it was
determined the temperature of the reactors, considering a null heat duty on those reactors.
Afterwards, a composite curves analysis was performed, in order to check that the assumed pre-
heating temperature for the inlet streams was acceptable.
The results from Table 7 were achieved with a pre-heating temperature of 600 °C.
The temperature in the calcination reactor (Table 7) was well above the limit temperature for the
calcination reaction to occur, which at 5 atm is approximately 967°C and thus the SE mechanism is
assured.
The temperatures of the three reactors as well as their heat duty are presented in Table 7. The heat
duty is the energy necessary to supply to or to extract from the reactors for assuring the specified
temperature.
Table 7 – Temperatures and Heat Duties of the three reactors considering input streams at 600°C and almost null Heat
Duties
Reactor Temperature [°C] Heat Duty [MW]
Fuel Reactor 702 0,055
Calcination reactor 1071 -0,52
Air Reactor 1190 0,388
The values of the residual heat duty present in Table 7 for those reactors’ temperatures is very low
comparatively to the values involved in this heat exchange process (< 0,5%) and hence it can be
neglected.
33
In Figure 11 the composite curves that correspond to a hot and cold stream was plotted, and from
which was possible to determine the amount of sensible heat that is possible to transfer between
them.
Figure 11 –Composite Curves considering input streams at 600°C and almost null Heat Duties
Based on this analysis, from Figure 11, it was possible to transfer, approximately, 205MW of heat
from hot streams to cold streams, through a heat exchanger network, for each kmol of CH4 that
entered the fuel reactor.
The composition of the stream that leaves the reforming reactor at 702°C is shown in Table 8.
Table 8 – Molar flows and H2 concentration at 702°C and 5 atm in the fuel reactor for a CH4 flow of 1 kmol/s, with a
NiO/CH4 ratio of 1, with a H2O/CH4 ratio of 2 and with CaO particles
Molar flow [kmol/s] H2 Dry
Concentration H2 CH4 H2O CO CO2 CaCO3
2,577 0,0917 1,239 0,0552 0,0426 0,810 0,931
0
200
400
600
800
1000
1200
1400
0 50 100 150 200 250 300
Te
mp
era
ture
[°C
]
Power [MW]
cold stream hot stream
205 MW
34
For the first optimization case, the process efficiency was 77,8%, regarding the conversion of CH4 in
H2 (LHVCH4 = 802MJ/kmol; LHVH2 = 242MJ/kmol), and at the same time, the CO2 capture rate was
85,0%.
With a temperature of nearly 700°C, the H2 produced in the fuel reactor had a purity level lower than
95%. Hence, it was possible to reduce the fuel reactor´s temperature to 650°C, which lead to an
increase in H2 concentration. However, pre-heating temperature had to be reduced due to lower
temperatures in all reactors.
In a similar way, results were recalculated considering a temperature inside the reforming reactor of
650°C and these results are presented in Table 9.
Table 9 – Temperature and Heat Duties of the three reactors considering input streams at 550°C, almost null Heat Duties
for the Calcination and Air reactors and a 650°C temperature in the fuel reactor
Reactor Temperature [°C] Heat Duty [MW]
Fuel Reactor 650 -11,10
Calcination reactor 977 -0,229
Air Reactor 1076 0,177
After adjusting the pre-heating temperature to 550°C, it was possible to optimize the process, with
almost the same amount of H2 produced and with a concentration higher than 95%. For these
conditions, it was necessary to increase the particles flow within the system, so that the temperature
of the calcination reactor was high enough to allow the calcination of CaCO3 particles. As a
consequence, the temperature of the air reactor suffered a decrease.
Like in Table 8, Table 10 presents the composition of the stream that leaves the reforming reactor at
650°C.
Table 10 – Molar flows and H2 concentration at 650°C and 5 atm in the fuel reactor for a CH4 flow of 1 kmol/s, with a
NiO/CH4 ratio of 1, with a H2O/CH4 ratio of 2 and with CaO particles
Molar flow [kmol/s] H2 Dry
Concentration H2 CH4 H2O CO CO2 CaCO3
2,567 0,1049 1,223 0,0127 0,0124 0,870 0,952
At 650°C, the process efficiency had a slight decrease to 77,4%, while the CO2 capture rate increased
to 98,6%.
35
Figure 12 presents a plot of the composite curves that correspond to a hot and cold stream, and from
which was possible to determine the amount of sensible heat that is possible to transfer between
them.
Figure 12 –Composite Curves considering input streams at 550°C, almost null Heat Duties for the Calcination and Air
reactors and a 650°C temperature in the fuel reactor
For this second case, based on Figure 12, it was possible to transfer, approximately, 194MW of heat
from hot streams to cold streams, through a heat exchanger network, for each kmol of CH4 that
entered the fuel reactor.
As there was still a heat release of 11,10 MW in the fuel reactor (Table 8), which is equivalent to
5,7% of the energy involved in the heat exchanged, the H2O/CH4 ratio could be increased.
Nevertheless, this increase could only be from 2 to 2,1, decreasing this heat release to 9,68MW,
since for higher values the temperature in calcination reactor was no longer sufficient for the
calcination reaction.
0
200
400
600
800
1000
1200
0 50 100 150 200 250 300
Te
mp
era
ture
[°C
]
Power [MW]
cold stream hot stream
194 MW
36
6. Experiments description and parameters
The experiments had the main objective to demonstrate, that in fact, this enhancing effect, created by
the addition of Calcium Oxide particles to the standard CLR process, improves the purity of the
hydrogen produced and allows the creation of two separate streams of H2 and CO2. These
experiments were carried out in a bench-scale fluidized bed reactor system at atmospheric pressure
with a bed mass constituted by 1/3 of 40% Ni particles (N4MZ1400) and 2/3 of sand or CaO
particles (limestone, calcined at 1000°C). The quartz reactor used had 820mm long and 22mm wide
with a porous plate located at 370mm from the bottom.
Figure 13 shows the reactor location inside the oven, as well as the two thermocouples introduced
inside.
Figure 13 – Reactor location inside the oven
The experimental procedure is divided in six different stages: setting up the reactor, calibration, start
up, experiments, turn off and cleaning. The procedures for each stage are presented in detailed in
Appendix A.
In set up phase, particles are weighted and introduced inside the reactor. The reactor is then
assembled, attached to the oven and remaining connections and thermocouples are introduced. The
two thermocouples were placed over and above the porous plate inside the reactor where the bed
material is introduced. Before the reactor is insulated, a leakage test is performed to make sure that
all junctions are well tight.
Quartz
Reactor
Thermocouple
Thermocouple
37
During calibration stage, firstly pure nitrogen is used for a zero calibration and hereafter, two
different gases are used for a span calibration, one for CH4, H2, CO and CO2 sensors, and another for
O2 sensor. The zero calibration together with the span calibration were used in order to establish a
linear relation between the sensor output signal and the corresponding concentration value.
In start-up, all other equipment (oven, cooler, heating bands and steam generator) are turned on as
well as the cycle controller window on the computer. Before starting the experiments, a steam
calibration is performed in a reduction cycle.
For the experimental phase, all flows should be set to the correct value, as well as the oven
temperature, and the results ought to be monitored through the control window on the computer.
After all experiments, all equipment and gas flows should be shut down by a specific order and
particles are to be removed to a specific container.
Finally, the reactor must be cleaned properly and before being stored.
Figure 14 shows a schematic description of experimental setup and Figure 15 presents the laboratory
workstation
Figure 14 – Schematic description of experimental setup
38
a b
c d
e f
g
h
i
j
Figure 15 – Laboratory workstation: a - oven with reactor inside; b – electronic equipment for cycle switch; c – gases flow meters; d – gases piping system; e – cooling unit; f – general perspective; g – console for flow regulator; h – steam generator; i – analyser; j – valves control.
39
Experiments consisted in a sequence of oxidation, reduction and inert cycles, so that it was possible
to simulate the conditions for the air reactor, during oxidation, and for the fuel reactor, during
reduction. Inert cycles (N2) were used with a flow of 600ml/min to flush the reactor, in order to avoid
the mixture between the gases from oxidation cycle and the gases from reduction cycle.
During oxidation cycles, a mixture of pure nitrogen and a 20,8% of oxygen in nitrogen flowed
through the reactor until all oxygen carrier particles were fully oxidized with a volumetric flow of
900ml/min. As a criterion it was considered that oxygen carrier particles were fully oxidized when
oxygen concentration at the reactor’s exit was greater than 95% of the inlet value.
Figure 16 presents the typical evolution of O2 concentration in the oxidation cycle of Ni particles.
Figure 16 – Typical evoltion of O2 concentration at the reactor’s exit in a oxidation cycle
Initially in the oxidation cycle, the O2 concentration at the exit of the reactor was approximately zero
while most of the Ni particles were being re-oxidized, and after a certain stage, the concentration had
a rapid increase tending for the inlet concentration (in this case 5%), revealing that most particles
were already in NiO form.
During reduction cycles, it was introduced in the reactor methane and steam at a specific H2O/CH4
ratio with a CH4 flow of 200 ml/min. The duration of this cycle depended on the amount of Ni
particles introduced inside the reactor and methane flow rate.
The quantity of CaO particles introduced inside the reactor determines the CO2 absorption capacity
of the bed. Thus, after a few number of cycles the reactor had to be heated up at high temperatures
(800°C-900°C) so that sorbent particles were regenerated. This was performed in an inert cycle and
simulated the conditions inside the calcination reactor. The Oxygen carrier used was meant to work
at high temperatures (900°C) and therefore, sintering problems was not expected. As the objective
0
1
2
3
4
5
7200 7400 7600 7800 8000
Co
nce
ntr
açã
o [
% (
V/V
)]
Time reference [s]
Concentração de O2
40
was not the study of the particles performance, no further tests were conducted, but visually there
was not any significant change in the particles dimensions. CO2 sorbent (CaO) was previously
calcined at 1400°C, so it was adequate to work at 900°C. Like the oxygen carriers, there was not any
visual change in its dimensions.
Before starting the actual experiments at lower temperatures (600°C-750°C), several cycles had to be
run at higher temperatures (800°C-900°C) in order to activate the oxygen carrier particles and to
prevent the CO2 sorbent particles from capturing the CO2 released in those cycles.
When establishing comparisons between experimental results and the simulation results, one must be
aware that the gas mixture that left the reactor was not in equilibrium, contrary to what happed in the
simulation results. Neither the residence time of the gases inside the reactor was not enough, nor the
reactor was well stirred, so that equilibrium could be reached.
41
7. Experiments results and discussion
The first two sets analysed consisted in a 15g bed mass, composed by 5g of a 40% Ni particles and
10g of sand or 10g of CaO particles. For these two sets, it was used a flow, during oxidation, with
5% oxygen in nitrogen and a H2O/CH4 ratio of 1,8. There were some difficulties, though, in assuring
a stable steam flow by the steam generator. Nevertheless, reduction cycles were conducted for four
temperature conditions (600°C, 650°C, 700°C and 750°C).
Other difficulties were found in the oscillating values measured by the gas analyser, which only
allowed an estimate of the gases’ concentration that came out of the reactor, as it can be seen in
Appendix C and Appendix D. Typically, errors caused by the accuracy of instruments are smaller
than errors inherent to the whole process (interferences, answering time, gas leakage, pulsing steam
flow, etc.). For a standard measurement of one gas concentration in N2, the analyser could be easily
calibrated with a very high accuracy.
During reduction, the only gases that may have left the reactor in significant concentrations were H2,
CH4, CO2 and CO, but the sum of its measured concentration did not reach 100%. So, despite the
sum did not reach 100%, it was made the assumption that these measured concentrations were in the
right proportion with each other and therefore, the real concentration could be determined. Hence,
each measured value by the analyser was divided by the sum of all concentration values. This
assumption would need a more deep reflection, in order to access what is the reason for this effect in
the measured concentrations.
In Table 11 and Table 12 are presented the concentration results for the several species from the
experiments carried out with 5g of N4MZ1400 particles, with sand (instead of CaO) or with CaO
particles, respectively.
Table 11 – Experiment results from the 5g of N4MZ1400 particles experimental set with sand, for a CH4 flow of 200
Zeman, F. et al. 2004. Frank S. Zeman, Klaus S. Lackner. Capturing carbon dioxide directly from the
atmosphere. World Resource Review. 2004, Vol. 16, pp. 157-172.
52
Appendix A – Experimental and cleaning procedure
53
Preparing the reactor
• Weight sample, (15g, 40g, etc.) introduce it to the reactor and measure height. Attach the
reactor inside the oven, attach thermocouples. Use silicone to avoid leakages.
• Turn on the computer… Start program, click “starta mätning”.
• Open nitrogen tube, open N2 in ceiling, turn up the flow (wheel #8). K2 to analyzer, K3 to
K7, K7 to reactor. Make sure flow is OK. Insulate to oven and above oven with wool (care
not to let it touch your skin). Insulate open surfaces and wool with aluminum foil.
Calibration
• Turn on the CO-measurement device (yellow).
• Use the nitrogen flow you will use during experiment. K2 to K7, K7 to analyzer. All
concentration should be 0. Zero calibrate all lines:
Main F3, Measure, Channel, until CH4, Calib, Zero Calibration
Measure, Channel until CO2, Calib, Zero Calibration
Measure, Channel until CO, Calib, Zero Calibration
Measure, Channel until H2, Calib, Zero Calibration
Measure, Channel until O2, Calib, Zero Calibration
• Turn down N2 flow to 0. Open gas tubes with 4,99% O2 and CH4,CO2,CO and H2. Turn up
calibration gases flow, use the flow you will use during experiment. Turn K1 to ON and
“blå” to CH4, CO2, CO. K2 still to vent. CO2 should be 40%, CO and CH4 20% and H2
19,9%. Calibrate each line by Span calibration:
Measure, Channel until CH4, Clib., Span Calibration
Measure, Channel until CO2, Clib., Span Calibration
Measure, Channel until CO, Clib., Span Calibration
Measure, Channel until H2, Clib., Span Calibration
Reduce calibration gas from the tube.
54
• Turn up the 4,99% O2 flow to a flow you will be using during the experiment. Turn “blå” to
O2. Calibrate O2 line:
Measure, Channel until O2, Calib., Span Calibration
Reduce O2 from the tube, K1 to off, K2 to analyzer. K7 to reactor.
Start up
• Name files (2, pressure file too)
• Turn on Air tube, set the flow you are going to use for Air and N2(ox).
• Open the “ventil tider”, put a long time (99999) to Oxidation and start.
• Start heating by turning on the power for the oven and heating bands, the cooling will also
start. Choose temperature on oven.
• K3 to Inert, “Starta logging” in program (2, pressure as well).
• Control files are logging when red “5min” is shown.
• Turn on remaining tubes you are going to use, make sure yellow CO measurement device in
on.
• When desired temperature is reached, rejoice! When you start, set the oxidation time to some
time soon (like10 seconds forward in time) and press enter.
• Turn up reduction flow.
Turn off
• Turn off during an oxidation (or inert) phase, reduce fuel flow.
• Turn off yellow CO measurement device.
• Turn off oven and carefully open it a little using gloves.
• “Stoppa logging”, (2 files) save data on USB.
• Turn off cooling and heat bands.
• When cooled down (to at least 300degrees), reduce flow.
• Turn off program and computer.
• When cooled, take out reactor, measure the height, weight and save sample.
• Clean reactor
55
Cleaning the reactor
• Make sure that you have removed the sample from the reactor main body and weight it;
• Start by uncoupling the reactor upper and lower parts from the main part;
• Remove the silicone used to avoid leakages in the connections. Frist clean it with some
paper and then use a solvent for the remaining’s;
• Clean with water the glass stick of the lower and upper parts and dry them with a paper;
• For the main part start by washing it just water a few times. Then, use water with soap and a
brush to reach the interior walls. Be careful, not to damage the reactor’s bed. Wash it again
with water to remove the soap from the interior;
• Take the reactor to the hotte and make sure the reactor bottom end is closed with a glass
stopper;
• Pour a bit of hydrochloric acid (≈50 ml) inside the reactor, cover the upper opening with a
cover connected to a tube and introduce the other end of the tube inside a recipient (beaker)
with water. Turn on the heater to boil the acid.
• When the acid is boiling, turn off the heater and remove the tube from the water, to avoid
suction of water to the interior of the reactor as the acid cools down;
• Pour the acid into a recipient and wash it with water inside the hotte;
• Wash the reactor with water from the water the water company and then with distilled water.
• Dry it out with compressed air before store the reactor.
56
Appendix B – Experimental equipment specifications
57
Equipment Brand Model Accuracy Obs.
Quartz reactor Custom Made -
820 mm long
22 mm wide
porous plate at 370 mm
from the bottom
Oven ElectroHeat Custom Made - -
Cooler M&C Peltier coolers
ECP1000 - -
Heating bands
system Termonic Serie 16150 -
1,5 kW
-15°C…+150°C
58
Mass Flow
regulation
console
unknown unknown unknown -
Mass Flow
controller Brooks 5800S Series ± 0,7% Several working ranges
Cycle regulation Parker
131V5406-
2995-
481865C1
-
Direct operated (3 way
corrosion resistant)
solenoid valve
Steam Generator Cellkraft E-1000 -
100% steam
0…18 g/min
≈150°C
1 atm
59
Valves Swagelok Ball - 2 or 3 way valves
Thermocouples -
K Type unknown unknown ± 2,2°C or 0,75% -
Analyser Rosemount
Analytical NGA 2000
H2: ≤ 2% Thermal conductivity
CH4: ≤ 1% Infrared spectra
CO2: ≤ 1% Infrared spectra
CO: ≤ 1% Infrared spectra
O2: ≤ 1% Para-magnetism
Typically, errors caused by the accuracy of instruments are smaller than errors inherent to the whole process (interferences, answering time, gas leakage,
pulsing steam flow, etc.). For a standard measurement of one gas concentration in N2, the analyser could be easily calibrated with a very high accuracy.
60
Appendix C – Experimental results for 5g N4MZ1400 with 10g sand
61
Figure 23 – Experimental results for 5g N4MZ1400 with 10g sand at 600°C and 1 atm in the fuel reactor, with a H2O/CH4
ratio of 2
Figure 24 - Experimental results for 5g N4MZ1400 with 10g sand at 650°C and 1 atm in the fuel reactor, with a H2O/CH4