Desktop version

Home arrow Engineering

  • Increase font
  • Decrease font


<<   CONTENTS   >>

Chemical Looping Reforming for Syngas Production

Syngas, as an important feedstock for the production of many valuable industrial chemicals such as methanol, synthetic liquid fuels, ammonia, and hydrogen, significantly affects the operation and the economics of downstream chemical processes. With the development of the technology, the syngas production process has evolved from simply passing steam through hot coke to using large-scale solids circulating systems which process a variety of hydrocarbon fuels.

CLR for reforming natural gas to syngas represents a potentially efficient means of producing electricity, hydrogen, syngas, and/or liquid fuels with minimal carbon emissions compared to traditional natural gas reforming schemes (Fan et al., 2015). There is a distinct difference in operating strategies between CLR and CLC processes. CLC processes can achieve complete carbon fuel oxidation to CO, and H,0 with a low reactant gas to oxygen earner flow ratio. For CLR processes, partial fuel oxidation requires careful control of the reactant gas to oxygen carrier flow ratio in order to maximize natural gas conversion to syngas.

Multiple CLR processes for syngas production from natural gas have been tested with multiple types of oxygen earner materials and reactor configurations. The present section summarizes several of the CLR processes that have been and are under development to date and is organized based on the type of gas-solid flow regime used for the reducer reactor design.

Fluidized bed chemical looping reforming

CLR processes were typically designed as fluidized bed systems, similar to those of CLC processes where fluidized beds of various forms, i.e., bubbling, turbulent, or fast fluidized beds, are used for the reducer or combustor. The system is configured as a circulating fluidized bed (CFB) system, where metal oxide oxygen earners continuously circulate among the different reactors of the system while undergoing reduction and oxidation reactions in the reducer and the combustor, respectively. Several major fluidized bed CLR process configurations are introduced below.

2.1.1 Welty-process

The first chemical looping reforming approach for syngas generation in fluidized bed reactors from methane was the Welty process, disclosed in a patent by Welty et al. in 1951 (Welty, 1951). The Welty process consisted of two separate fluidized bed reactors, as shown in Figure 3.

In the process, methane was partially oxidized by the oxygen provided by oxygen carriers, such as Fe.O, or CuO, to produce a mixture of H, and CO in the gas converter reactor (reducer). The reduced

Welty process for syngas generation from methane (Welty, 1951)

Figure 3. Welty process for syngas generation from methane (Welty, 1951).

oxygen carrier was re-oxidized with air in the reoxidizer (combustor). In addition to the oxygen earners, nickel-based reforming catalysts with Al,03 or MgO support were also used to selectively convert methane into syngas through steam methane reforming and diy reforming reactions. The addition of a nickel-based reforming catalyst was to inhibit the full oxidation of methane to CO, and H,0 in the reducer. In the initial design, the catalyst was transported to the combustor together with the reduced oxygen carrier where it was regenerated with air. However, this regeneration step also oxidized the reforming catalyst from Ni to NiO, causing the loss of the catalytic activity of the reforming catalyst. A solids-solids centrifugal separator, which is based on the density difference between the oxygen carrier and the nickel-based catalyst, was added to the modified version of the process in order to overcome the drawback. The low-density nickel-based reforming catalyst exited the top of the separator, which was directed back into the reducer, while the high-density oxygen carrier exited at the bottom of the separator and was directed to the combustor for regeneration.

The system had operational issues due to noil-optimal heat integration as there were difficulties in maintaining an operating tempera tine difference between the reducer at 816-927 °C and the combustor at 871-982 °C. In addition to heat integration difficulties, the Welty process was challenging to operate, as carefi.il control of the fluidized gas was required in order to maintain both proper oxygen carrier circulation and separation of the reforming catalyst to prevent its transfer to the combustor. The requirements of both these conditions on the fluidizing gas flow rate substantially reduced the window of operability limits for the Welty process. Between the work of Welty et al. and the early 2000s. little progress was made in further developing the CLR process, until a similar concept was applied by Mattisson et al. for methane reforming, using fluidized beds (Mattisson and Lyugfelt, 2001).

2.1.2 Steam methane reforming-chemical looping combustion (SMR-CLC) systems

The conventional SMR process, as briefly described in the introduction to this chapter, can be coupled with CLC for its energy integration, as the external heat generated by the CLC process can be used to compensate for the heat requirement of the endothermic methane reforming reaction while inherently capturing the CO, generated for the combustion of the natural gas. The tubular, packed bed SMR reactors can be embedded in either the reducer or the combustor of the chemical looping combustor system to maintain its operational temperature for continuous operation. This kind of process is referred to as SMR- CLC process. The two different configurations with SMR embedded in the reducer and the combustor, respectively, are shown in Figure 4.

When hydrogen is the desired product, water-gas shift (WGS) reactor and PSA unit are integrated into the SMR-CLC system. The WGS reactor is located after the SMR reactor to maximize the hydrogen yield, with the PSA imit being used to separate the hydrogen from the gas product to obtain purified hydrogen. The off-gas stream from the PSA imit, consisting of CO, produced in the SMR and WGS reactors, unconverted CO and CH4, and a small amount of H,, is sent back to the reducer of the CLC system together with necessary natural gas to reduce the oxygen earner particles and ensure continuous operation of the CLC system. Since the CLC unit generates pure CO, stream from the reducer outlet, the integrated process provides a unique advantage of near 100% CO, capture without any further processing steps.

In the SMR-CLC systems, the SMR reactor can be embedded either in the reducer (Figure 4a) or in the combustor (Figure 4b). The two types of configurations have significantly different process schemes influencing the energy balance and performance of the SMR-CLC process. In the CLC reactor systems, the reaction taking place in the reducer can be either endothermic or slightly exothermic, depending on the fuel and oxygen carrier used, while the reaction in the combustor is strictly exothermic. As a result, more heat in the combustor can be used for the SMR process, and the temperature in the combustor is typically higher than that in the reducer for CLC systems (Fan, 2010; Lyugfelt, 2015).

As in the type a SMR-CLC process, the heat in the CLC reducer needs to support the endothermic SMR reaction and maintain the normally endothermic or slightly exothermic reaction between the metal oxides and fuel, the temperature of the CLC reducer is lower than that of stand-alone CLC systems under equivalent CLC operating conditions. Hie increase in heat requirement of the CLC reducer in

SMR-CLC for H, production

Figure 4. SMR-CLC for H, production: (a) type a SMR embedded in reducer and (b) type b SMR embedded in combustor

(Ryden and Lyngfelt, 2006; Adanez et al., 2012).

the SMR-CLC process can be managed by increasing the recirculating temperature of oxygen earners from the combustor to the reducer or increasing the oxygen earner circulation rate of the CLC system. However, increasing the recirculating temperature of the oxygen carrier particles generally means increasing the operating temperature of the combustor, which puts forward harsher requirements on the physical properties of the oxygen carriers. Wien the oxygen earner circulation rate of the CLC system is increased, the reactor volumes need to be increased, and the attrition rate of the oxygen carrier particles will also increase.

In the type b SMR-CLC process, extensive heat is released in the combustor dining the regeneration process of the reduced oxygen carrier reaction, which can provide enough heat for SMR reaction. Thus, the reducer can be configured the same as that of stand-alone CLC systems. There is no need to increase the operating temperature or the oxygen carrier circulation rate of the CLC system.

Pans et al. (Pans et al., 2013) studied the detailed energy balance of both type a and type b SMR-CLC processes with two different n oil-based oxygen carriers, namely pin e non oxide (Fe,03) and alumina supported iron oxide (Fe,03-Al,03). The effects of swinging between different oxidation states, Fe,03/Fe304 scheme for pure non oxide and Fe.O. FeO (Fe,03-Al,03/FeAl,04) for alumina supported iron oxide, were analyzed for the H, yield from the SMR-CLC process. The different swinging schemes of iron oxide oxidation states alter the overall energy balance of the system as the reduction reaction from Fe.O, to Fe3Oa by natural gas is endothermic; while it is exothermic from Fe,0,-Al,03 to FeAl,04. The performance of type a and type b SMR-CLC processes under an autothennal operating condition was compared. The operating temperature of the reactor where the SMR reactor was embedded was set at 900 °C. The type b SMR-CLC system produced a higher hydrogen yield than type a for both oxygen carriers. However, as shown in Figure 5, the difference between the two schemes decreases with the increase of the oxygen carrier circulation rate of the system. For the Fe,0,/Fe,04 scheme, the temperature in the combustor needs to be higher than the reducer for both types of SMR-CLC systems. However, when Fe,0,-Al,03/FeAk04 scheme is used, only in the type a SMR-CLC does the combustor temperature need to be higher than the reducer temperature: while in the type b SMR-CLC system, the reducer temperature is higher than the combustor temperature due to the exothermic characteristics of the reaction in the reducer. This is beneficial for the reactor design as the kinetics of the reaction between the fuel and the oxygen carriers can be enhanced under a higher operating temperature.

Fan et al. conducted an exergy analysis to investigate the benefits of SMR-CLC as compared to conventional SMR process (Fan et al., 2016). The conventional SMR and SMR-CLC processes used in the study were shown in Figure 6. The SMR-CLC system was a type b system where the SMR reactor was embedded in the combustor. Nickel-based oxygen earner particles were used in the CLC system. The temperature of the combustor was set at 1000 °C.

The exergy distributions in the SMR and the SMR-CLC processes were listed in Table 1, generated using Aspen Plus*. The process simulation results showed that the overall exergy efficiency of the SMR- CLC system had an efficiency of 71.4%, an approximately 9.5% advantage over SMR, whose efficiency was 65.2%. The overall exergy destruction of methane in the SMR-CLC process was 217.87 kJ per mole of CH4. while it was 299.66 kJ per mole of CH4 for SMR process. The main exergy destruction

Schematic diagram of conventional SMR and SMR-CLC processes

Figure 6. Schematic diagram of conventional SMR and SMR-CLC processes.

Table 1. Exergy of each component of SMR and SMR-CLC processes.

SMR

SMR-CLC

Exergy (kJ/mol CH4)

°/o of total £лш

Exergy (kJ/mol CH4)

% of total Exm

Exergy m

940.55

100.0

866.17

100,0

Methane

830.19

88.27

830.19

95.85

30.86

3.28

30.86

3.56

0.15

0,01

0.15

0.02

4.97

0.53

4.97

0.57

CO, capture

74.38

7.91

-

-

Exergy out

640.89

68.14

648.30

74.85

Hydrogen

613.59

65.24

618.70

71.43

Exhausted gas

27.30

2.90

29.60

3.42

Exergy destroyed

299.66

31.86

217.87

25.15

Exergy un-used

326.96

34.76

247.47

28.57

Exergy efficiency

65.24

71.43

in the SMR process was the combustor and the CO, capture unit, which contributed 28.5% and 24.9% of the total destroyed exergy, respectively. While in the SMR-CLC process, the exergy destruction for combustion was reduced from 85.3 kJ per mole of CH4in SMR process to 79.1 kJ per mole of CHr The economic feasibility of the SMR-CLC process was also examined by means of financial analysis. The capital cost estimation of both SMR and SMR-CLC processes were listed in Table 2. When compared to the SMR process, the SMR-CLC process can save about 0.02 million euro (M€) total investment costs, mainly due to the reduced total equipment costs which decreased from 15.22 M€ for the conventional SMR process to 15.05 M€ for the SMR-CLC process. This reduction revealed that the SMR-CLC process

Table 2. Capital cost estimation of SMR and SMR-CLC processes (Fan et al., 2016).

Unit

SMR

SMR-CLC

Capital cost

Total equipment costs

M€

15.22

15.05

Total install costs (excluding contingency)

M€

29.43

29.1

Total investment costs

M€

36.79

36.37

Fixed O&M costs

Mft'year

4.69

4.5

Variable O&M cost

Fuel cost

Mft'year

8.79

8.79

Non-fuel (including water, power consumption and solvent loss)

Mft'year

1.84

1.82

Total fixed and variable O&M costs

Mft'year

15.32

15.11

Net Hydrogen Production

kg'year

5472000

5472000

Levelized cost of hydrogen

€/kg

3.28

3.24

is economically feasible and attractive, in addition to technically simplifying the overall process. The total fixed and variable O&M costs for SMR-CLC were 15.11 M€ per year, lower than the 15.32 M€ per year for the SMR process. As a result of reduction on both capital cost and O&M cost, the levelized cost of hydrogen for the SMR-CLC process is 3.24 €/kg, lower than the 3.28 €/kg for the SMR process.

Viktor et al. compared simulation results of a type b SMR-CLC process and several other novel SMR integration methods against the conventional SMR process for hydrogen production (Stenberg et al., 2018). Fe-based and Mn-based oxygen earners were used in the simulation, where the temperatures of the reducer and the combustor of the CLC system were assumed to be the same. The simulation results showed that the SMR-CLC process is beneficial as its efficiency is much higher than that of conventional SMR process, almost as high as that of the SMR integrated with oxygen carrier aided combustion process, and it has a unique advantage of near 100% CO, cap true without an additional energy penalty.

Evaluations of SMR-CLC processes were mainly performed using simulations, with minimal experimental tests conducted (Adauez et al., 2012). As CLC was merely used for externally heating the SMR reactor, while the experimental researches on the CLC systems were still in relatively small scales which cannot reach autothennal operating conditions yet. The large surface area to volume ratio of the small scaled CLC reactor makes the heat loss of the system significantly higher than the net energy produced by the system. Therefore, they require external heaters or burners to compensate for the heat loss. In addition, SMR-CLC does not adequately manifest a novel chemical looping technology-based reforming process as CLC is merely applied in order to provide external heat for the endothermic SMR process. This reforming technology is fundamentally based on the already existing SMR technology.

2.1.3 Reactor systems for fluidized bed chemical looping reforming

Unlike the SMR-CLC process, which requires the integration of conventional SMR process with the CLC process, the CLR processes produce syngas directly from partial oxidation of natural gas. while using an oxygen earner to transport oxygen for the reaction. For natural gas, the overall reaction of the CLR process is highly exothermic. This enables the autothennal operation of the CLR processes. A widely applied design for the CLR process is based on the CLC system using a fluidized bed reducer.

The economic analysis of a fluidized bed CLR process was conducted in order to evaluate the natural gas to syngas (NGTS) process of producing syngas for liquid fuel production via Fischer-Tropsch (F-T) gas-to-liquid (GTL) technology. The NGTS process is one of the major CLR processes for syngas production that uses an iron-titanium composite metal oxide (ITCMO) as the oxygen carrier to convert natural gas to syngas (Li et al., 2011).

The liquid fuel production process is based on a reference GTL process with 50.000 banel per day (bpd) liquid fuel production using autothennal refonning (ATR) to convert natural gas to syngas

Overall chemical looping system gas to liquids process for 50,000 bpd plant

Figure 7. Overall chemical looping system gas to liquids process for 50,000 bpd plant.

(Gollener et al., 2013). As shown in Figure 7, the fluidized bed reducer replaces the reference ATR reactor for syngas generation. Tlie NGTS-GTL process is designed to achieve syngas quality and quantity to match the amount of liquid fuel production of the baseline ATR-GTL process. Stoichiometric number (S#) is one of the key parameters to describe the quality of syngas, which is defined as

Here, yA represents the molar concentration of component A in syngas. The Fe,03:C ratio and H:0:C ratio are adjusted to reach specific syngas quality for liquid fuel production and low solids circulation rate to minimize the amount of metal oxide particles in the process. The recycle fuel gas stream is assumed to be completely converted. These assumptions represent the most optimal performance possible for a fluidized bed reactor. In a fluidized bed reactor, solids phase backmixing and heterogeneity of gas-solids fluidization lead to a wide residence time distribution and gas channeling. Experimental results obtained for methane conversion to syngas in a fluidized bed reactor showed that syngas with a H,:CO ratio of 1.8 and a CO:CO, ratio of 12 can be produced from methane in a fluidized bed reactor of CLR process using ITCMO oxygen carrier. The methane conversion is limited to 70%, based on the testing results.

Table 3. Overall performance of the NGTS process in a 50,000 bpd GTL plant (Gollener et al., 2013).

Component

Base case

NGTS (10 atm) fluidized bed

Natural gas flow, kg/hr

354,365

452,992

Natural gas flow, kmol/hr

20,451

26,143

0,68

0.25

2.19

2.18

1.59

-1.90

0.73

0.48

Butane feed flow, kg/hr

18,843

18,843

Diesel fuel, bbl/day

34,543

34,543

Gasoline, bbl/day

15,460

15,460

Total liquids, bbl/day

50,003

50,003

Electrical Load (kWe)

Total Gross Power

303,700

303,700

Net Plant Power

40,800

150,480

These experimental results from fluidized bed CLR were used to develop the models of the fluidized bed NGTS-GTL process for its economic analysis.

Table 3 listed the performance results for the NGTS-GTL process using fluidized bed CLR, as compared to the ATR-GTL process with a 50,000 bpd liquid fuel output. The carbon efficiency of the process is lower than that of the baseline case, even though the net plant power output increases as the fluidized bed CLR eliminated the needs of energy intensive ASU. This is mainly because the fluidized bed CLR consumes 28% more natural gas than the conventional ATR.

300 Wth CLR Unit at Chalmers University of Technology

Chalmers University of Technology (Chalmers) in Sweden investigated a 300 W4 two-compartment fluidized bed reactor system for CLR, as shown in Figure 8, using Ni-based oxygen carriers (Ryden et al., 2006; Ryden et al., 2009; Johansson et al., 2006; Ryden et al., 2008). The unit was initially designed for CLC experiments and was later adopted for CLR for syngas tests (Ryden et al., 2008).

The design of the CLR for syngas system is derived from the shale oil reforming system proposed by Chong et al. in 1986, featured with solids exchange between two adjacent fluidized bed reactors without gas mixing (Chong et al., 1986). The 300 W4 unit consists of two adjacent chambers, the reducer reactor

Chalmers 300 Wfluidized bed CLR system (Kronberger et al., 2004) (Arrows m the figure denotes the dnection

Figure 8. Chalmers 300 Wtfluidized bed CLR system (Kronberger et al., 2004) (Arrows m the figure denotes the dnection

of solids flow).

chamber and the combustor reactor chamber. The chambers are divided by a vertical wall with two slots, as shown in Figure 8 (Kronberger et al., 2004). The reducer chamber had a square cross-sectional area of 25 mm x 25 mm. The combustor had a rectangular cross-section at the bottom with a dimension of 25 mm x 40 mm. The total height of the reactors was 200 mm. with an enlarged section on top for solids disengagement from gas flow. A downcomer with a width of 12 mm was located between the reducer and the combustor for solids transportation from the combustor to the reducer. The two slots on the vertical wall connected the bottom of the two reactors. The slots consisted of two walls, one in each reactor, to minimize the gas leakage between the two reactors. To maintain the constant operating temperature of the reactor system, the 300 Wt unit was placed inside an electrically heated furnace to compensate the heat loss. A water seal with adjustable water column height was added to the exit pipe of the reducer to control the pressure of the reducer so as to minimize the air leakage from the combustor to the reducer. The system used natural gas, with a composition equivalent to C, 14H4,5O001N0005, as the fuel for the CLR experiments.

The two chambers are operated under fluidized bed conditions using different fluidization regimes in the two chambers in order to control the dir ection of oxygen carrier circulation. The combustor reactor chamber is operated at a higher gas velocity than the reducer in order to increase the void fraction in combustor, allowing the oxygen carriers in the reducer to overflow into the combustor. The combustor reactor chamber is tapered at the top, which increases the gas velocity, causing the oxygen carrier to be entrained to the top and pass travel into the downcomer to enter the reducer. The reducer is operated with a lower gas velocity in order to form a dense phase fluidized bed which generated a larger pressure drop than the combustor chamber and in turn causes a pressure difference between the bottoms of the reducer chamber and the combustor chamber. This pressure difference drives the particles to move from the reducer to the combustor through an overflow slot between the two reactors, thus forming a continuous solids circulation loop.

Three types of Ni-based oxygen carriers, whose properties are given in Table 4, were tested in the Chalmers CLR experiments. Experiments were performed with a reducer temperature of between 800 °C and 950 °C, and various air-to-friel ratios for all the oxygen carriers. Steady state of 1-3 horns were maintained for each experiment condition. In the CLR tests, a high fuel flow rate (0.8-1.5 L/miu) and a moderate to high air flow rate (3.8-10 L/min) were used. In addition to the tests with pure CH4, CHj-CO, and CHj-steam, co-feed CLR experiments were also tested, with feeds containing 30% CO, and/or steam in the fuel gas.

Typical results for Chalmers CLR experiments are given in Figures 9-11. Syngas generation by CLR was achieved by adjusting the air-to-fuel ratio. Operating conditions that prer ented carbon deposition, a phenomenon that not only reduces the syngas yield from CH4 and produces CO, emission in the combustor, but also weakens the mechanical strength of oxygen carriers and deactivates the particles, were investigated. As shown in Figure 9, carbon deposition can be eliminated when the operating temperature of the reducer is over 930 °C. Cho et al. reported that solid carbon formation depends strongly on the available oxygen of the Ni-based oxygen carrier (Cho et al.. 2005). Wien 80% of NiO is reduced to Ni, carbon deposition occurs rapidly. However, the Chalmers CLR tests showed that carbon deposition became significant even when there were instances where only 33—44% of NiO was reduced to Ni.

The CLR tests showed that, for a dry natural gas feed case with an operating temperature greater than 930 °C, the syngas quality is improved when the oxygen-to-fuel ratio is reduced, as shown in Figures 10-11. However, there is a lower boundary for the reduction of oxygen-to-fuel ratio, as carbon deposition might be significant. The maximum syngas purity, defined as the molar percentage of CO and H, in the gas product of the reducer, was around 75%, with a H,:CO ratio of 1.7.

Table 4. Ni-based oxygen carrier properties used m 300 W Chalmers’s CLR unit (Ryden et al., 200S).

Oxygen carrier

Chemical composition

Production method

Size (pm)

Porosity (%)

Solids inventory’ (g)

N2AM1400

10% NiO on MgAl,04

Freeze granulation

90-212

35

250

NilS-aAl

18% NiO on a-Al.Oj

Impregnation

90-212

53

180-250

Ni21-yAl

21% NiO on y-Al,Oj

Impregnation

90-250

66

170

Percentage of fuel resulting in carbon deposition (Ryden et al., 2008)

Figure 9. Percentage of fuel resulting in carbon deposition (Ryden et al., 2008).

Syngas purity as a function of air-to-fuel ratio, with and without steam injection for different oxygen earners

Figure 10. Syngas purity as a function of air-to-fuel ratio, with and without steam injection for different oxygen earners

(Ryden et al., 2008).

To overcome the limitation on oxygeu-to-fuel ratio, steam or CO, co-feeding with natural gas were introduced in order to reduce or eliminate carbon deposition in the reducer. When 30% steam or CO, was со-fed with natural gas into the reducer with an operating temperature of 950 °C, a high syngas purity was obtained with a low air-to-fuel ratio. The co-feeding of steam with natural gas feed assisted the reactions in two aspects. The CH4-steam co-feeding could yield a higher H, concentration compared to CH4 only feedstock in the gas product due to the SMR reaction with Ni as a SMR catalyst. Furthermore, carbon deposition was reduced as the solid carbon was gasified by steam. As shown in Figure 10, the CFl4-steam co-feeding with an operating temperature of over 930 °C showed promising results for producing syngas with a high purity as the reducer was operated nearly under a condition of a steam methane reformer. The performance for 30% CO, co-feeding with CFI4 was given in Figure 11. The CFl4-CO, co-feeding also decreased carbon deposition, as the solid carbon was gasified by the CO,. The maximum syngas concentration achieved was around 90%, as shown in Figure 11(a). However, the maximum H,:CO ratio was only 1.3, as shown in Figure 11(b).

140 kW(h DCFB CLR Unit at Vienna University of Technology

Vienna University of Technology (VUT) in Austria developed a dual circulating fluidized bed (DCFB) reactor configuration for the CLR process to generate syngas from natural gas (Pro et al., 2010; Proll et al., 2010). The system, as shown in Figure 12(a), uses Ni-based oxygen carriers and has a thermal capacity of 140 kW (Proll et al., 2005; Proll et al., 2009). The reducer of the VUT DCFB system was operated under a turbulent fluidized bed regime, and the combustor was under a fast fluidized bed regime. The reactor system consisted of two interconnected circulation loops with two reactors, three loop seals, and two cyclones. The primary loop of the system consisted of the combustor, the reducer, the primary cyclone and two loop seals, one connecting between the bottoms of the combustor and the reducer, and the other connecting between the bottom of the primary cyclone and the middle of the reducer. The lower loop seal that cormects the bottoms of the reducer and the combustor maintains a stable solids distribution between the two reactors. The internal loop, consisting of the reducer, an internal loop seal and an internal cyclone, enabled local solids circulation within the reducer. This design ensured a long average residence time of

VUT 140 kW dual circulatuig fluidized bed reactor for CLR (Proll et al, 2009)

Figure 12. VUT 140 kW4 dual circulatuig fluidized bed reactor for CLR (Proll et al, 2009).

the oxygen carrier particles in the reducer, allowing for the reactions in the reducer to reach equilibrium state. This internal loop also enabled a local solids circulation rate of the reducer, independent of the global solids circulation rate of the system. Steam was used as the sealing gas in the loop seals.

Pressure profiles through the DCFB system, with a sample as shown in Figure 12(b), were obtained for several operating conditions in order to analyze the solids distribution in the system under different conditions. The sample pressure profile was obtained when the system was operated at a full capacity of 140 kWth with a total solids inventory of 65 kg, with around 30 kg in the DCFB reactors and the balance being distributed in the cyclones and loop seals (Proll et al., 2010). The global air-to-fuel ratio of the system was 1.1, with the reducer temperature maintained at 900 °C (Proll et ah, 2009). The pressure profile along the height of the combustor represented a typical fast fluidized bed operation. In the reducer, the pressure drop in the lower section was sharper, representing a typical dense-phase fluidized bed, while the top section had a much smaller pressure drop, reflecting the solids disengagement zone of the fluidized bed. Due to the different fluidization regimes of the reducer and the combustor, the overall pressure drop in the reducer was larger than the combustor. Correspondingly, more solids were present in the reducer compared to the combustor. For all the loop seals, the pressures at the gas inlets were higher than the both gas outlets, showing that the inert gas split and flew to the both ends, and good gas sealing between the connected reactors was achieved.

The oxygen carrier particles used in the system, whose production methods were described in detail by Linderholm et al. and Jemdal et al. (Linderholm et al.. 2009; Jemdal et al., 2009), were made by mixing two different types of oxygen earner materials with a 50:50 weight ratio. One oxygen carrier material is a sintered mixture of NiO/NiAl,04 (N-YTTO) synthesized from NiO and AfO,. The other was a sintered mixture of Ni0/MgAl,04-NiAl.,04 (N-VITOMg) synthesized from NiO, Al.O,, and MgO. In total, there was about 40 wt% of active NiO material in the oxygen carrier particles. The mean diameter of the oxygen carrier particles was 135 pm.

Three different operation temperatures, 750 °C, 800 °C and 900 °C, were tested. Natural gas from the Viennese grid (98.7% CH4) was introduced into the reducer and maintained at 140 kVh capacity during stable operation. The global stoichiometric air-to-fiiel ratio was decreased stepwise from 1.1 to 0.5, with an interval of 0.1, while the cooling duty of the reactor was adjusted accordingly to maintain the constant temperature.

The operational results of the YUT CLR system as a function of different air-to-fuel ratios under different operating temperatures of the systems were shown in Figure 13 (Proll et al., 2010). As shown in Figure 13(a), under a given operating temperature of the system, the syngas purity increased with the decrease of the air-to-fuel ratio. Increasing the operating temperature of the reducer decreased the syngas purity under a given air-to-fuel ratio.

The combustor operating temperature of the VUT CLR system was designed to be higher than the reducer, such that the temperature difference is sufficient to obtain autotliennal operation. The operating

temperatures of the combustor under different reducer operating temperature conditions as a function of different air-to-fiiel ratios were given in Figure 13(b) (Proll et al., 2010). Under a given reducer temperature, a higher air-to-fuel ratio translated to a lower operational temperature of the combustor. When the air-to-fuel ratio was above 1.1, the system was essentially operated in the CLC mode. The heat generated from combustion was sufficient to provide the heat required for frill oxidization of fuel in the reducer. On the other hand, decreasing the air-to-fuel ratio resulted in a larger temperature difference requirement between the two reactors in order to achieve heat balance. The minimum air-to-fuel ratio was around 0.5.

Duiing the operation of the VUT CLR system, gas leakage between the reducer and the combustor was observed. This may lead to a hazardous condition as the reducing gas in the reducer was mixed with the oxidizing gas from the combustor at high temperatures. Dining the operation, this gas leakage rate was controlled to be no more than 0.5% volume of the combustor gas, as deemed acceptable for safety and process performance. However, due to the gas leakage issue, a large amount of sealing gas in the loop seals flowed to the reactors. The majority of the gas entered the combustor and exited in the combustor exhaust gas. As much as 5% volume of steam was detected in the gas at the combustor outlet. The sealing gas leakage to the reducer could have a slight effect on decreasing the syngas concentration in the reducer product gas. During the experiments, there was no CO, or CO detected in the combustor, indicating minimal carbon deposition in the reducer and no gas leakage from reducer to combustor. The study is one of a few cases of fluidized bed CLR processes where no obvious carbon deposition was observed in the reducer without steam or CO, co-feeding with natural gas. However, the syngas purity under these conditions was relatively low, at less than 60%.

In summary of experimental studies on fluidized bed CLRs, the fluidized bed CLR processes are capable of syngas production from methane reforming. The syngas purity from these reducers can achieve from 0% (CLC case) to ~ 70% by adjusting the air-to-fuel ratio from above 1 (for combustion) to about 0.4 (for reforming) without steam or CO, co-feeding. Further lowering the air-to-fuel ratio will result in significant carbon deposition, which lowers the syngas purity, affects the chemical and physical properties of the oxygen earner particles and increases the CO, footprint of the process. Avoiding carbon deposition by limiting the oxygen carrier reduction to a low level leads to a high oxygen carrier circulation rate for the CLR system. However, steam and or CO, co-feeding with natural gas can alleviate the carbon deposition issue. With the addition of steam and/or CO„ the air-to-fuel ratio can be further lowered with a higher syngas purity achieved. Feedstock conversion may not be complete due to the inherent solids backmixing and channeling of gaseous hydrocarbons produced in a fluidized bed reducer.

 
<<   CONTENTS   >>

Related topics