Numerical Simulation of a 10 kW Gas-Fueled Chemical Looping Combustion Unit

Abstract

Chemical looping combustion is one novel technology for controlling CO${}_2$ emission with a low energy cost. Due to a lack of understanding of the detailed and micro behavior of the CLC process, especially for a three dimensional structure, numerical simulations are carried out in this work. The configuration is built according to the experimental unit and gaseous fuel is used in this work. A two-fluid model considering heterogeneous reactions is established, and the flow behaviour and reaction characteristics are obtained. The temperature in the air reactor increases with height owing to the exothermic reaction of the xidation of the oxygen carrier, while the temperature in the fuel reactor decreases with height due to the endothermic reaction. The oxidation level of the oxygen carrier is obtained by simulation, which is hard for measurement, and the difference between the inlet and outlet is 0.065. The influences of the operating temperature and injection rate of fuel are presented to understand the performance of the system. The highest fuel conversion rate reaches 0.92 under high operating temperature. The numerical results are helpful for acquiring insight on the flow and reactive behaviour of CLC reactors.

1. Introduction

Chemical looping combustion is one of the technologies used for carbon dioxide capture, which is beneficial for controlling the impact of greenhouse gas on the climate and on the increase in global temperature [1,2]. This technology provides an alternative method to producing high concentration CO${}_2$ exhaust, with the reactions isolated into two separated reactors. An oxygen carrier transports lattice oxygen and heat between reactors, as a result the air never mixes with the combustion products. The pure CO${}_2$ is obtained after condensing the products of the fuel reactor with less energy for gaseous separation [3]. The net reaction, mass transfer and heat release of CLC are similar to the traditional method of fuel combustion. The difference is that no penalty energy is needed for CO${}_2$ separation. The CLC technology exhibits an obvious advantage for CO${}_2$ capture with low energy cost [4,5].

The typical configuration of a CLC system consists of a fuel reactor, air reactor and loop seals. The diagram for the CLC system is shown in Figure 1. The experiments for the CLC system with dual interconnected fluidized reactors have been successfully carried out at lab scale and pilot scale, with both gaseous fuel and solid fuel [6,7,8], coal [7,9], petroleum [10], biomass [11] and liquid fuel [12], which prove the feasibility of the CLC technology. At the same, researchers have used the numerical method to investigate the operating behaviour of the CLC system, which is difficult for experimental measurement [13,14], and it has played a more and more important role for investigating the CLC system [15,16,17]. Two main methods, two-fluid model and discrete element method (DEM), are utilized for simulating the CLC system [18,19,20]. With DEM, one can track the trajectory of each particle and get the information at particle scale. Kong et al. [21] simulated the CLC system using DEM and the effect of PSD(particle size distribution) on the conversion of fuel conversion. Results showed the positive impact on the performance of the reactor. Lin et al. [22] investigated the CLC system with a dual interconnected reactor using DEM. The residence time can be counted and analysed, and the results provided the particle scale understanding of the operation. The disadvantage of DEM is that the computational cost is huge with the increase in particles in the system. Currently, DEM is not possible for the large scale reactor, especially at a pilot or industrial scale. For the two-fluid model, both the gas phase and solid phase are treated as interpenetrating continua and it obtains considerable calculation speed. Chen et al. [18] utilised a two-fluid model to study the bubble behaviour in a fuel reactor. Francisco et al. [23] developed the model for analysing the 1D fuel reactor. The results show that a higher temperature was desired for high carbon capture, but had a smaller effect on oxygen demand. The carbon capture efficiency increased to 98.6% when raising the temperature to 1100 ${}^{°}$C and with a carbon separation efficiency 98%. Kavitha et al. [24] carried out a 2D simulation of a fuel reactor with coal as fuel. When considering ash during simulations, the performance of CuO and Fe${}_2$O${}_3$ were thesame, if no ash, the CuO performed better. The char gasification was the limited step of the CLC process. Wang et al. [25] pointed out that the formation of large bubbles lead to low fuel conversion. May [26] simulated the 3D CLC system using coal without the cyclone and loop-seal, instead of the changing mass flow rate at the boundary. The utilization of a drag model lead to deviation between the predictions and measurements, especially for the reaction with a different flow regime. Li et al. [27] carried out the 3D simulation of cold flow for hydrodynamic analysis. The full loop structure was constructed. The nonuniformity of flow behavior was observed. Wall roughness showed the indefinite affect and tendency. To sum up, according to the literature review, the fundamental mechanism of the operation of CLC has been investigated through both experiments and theoretical methods, in order to maximize the fuel conversion and carbon capture. However, there are few studies that focus on the full loop CLC system, and the coupling between reactors are missing. Moreover, three dimensional numerical structure is essential to reveal the properties of operation, but currently most works are on the two dimensional structure and only hydrodynamics research. Operating details are needed for which are difficult for experiments.

To overcome this insufficiency, in this work, a two-fluid model for a full loop three dimensional gas-fueled chemical looping system is developed for exploring the flow and reaction behavior. The CLC geometry is constructed according to the experiment one consisting of two interconnected reactors, cyclone and loop seal. The literature review shows that there is very little research relating to the circulation behaviour and reaction characteristics in full loop structures. The motivation of the current work is to gain the operating behavior of the three dimensional CLC unit at a micro-scale. The data obtained from simulations will be helpful for designing and operating the CLC system.

2. Mathematical Models

A two-fluid model is utilised for modeling the transport equation of mass, momentum, energy and species. The kinetic theory of granular flow is adopted for the closure of the solid phase. The transport equations solved in this work are shown in

Table A1. Transport equation.

	Mass balance equation
T1	${\displaystyle \frac{\partial \left({\alpha }_g{\rho }_g\right)}{\partial t}+\frac{\partial ({\alpha }_g{\rho }_g{U}_{g,j)}}{\partial x_j}={\mathrm{\Gamma }}_g}$
T2	${\displaystyle \frac{\partial \left({\alpha }_s{\rho }_s\right)}{\partial t}+\frac{\partial ({\alpha }_s{\rho }_s{U}_{s,j)}}{\partial x_j}={\mathrm{\Gamma }}_s}$
	Momentum equation
T3	${\displaystyle \frac{\partial {\alpha }_g{\rho }_g{\mathbf{u}}_g}{\partial t}+▽·\left({\alpha }_g{\rho }_g{\mathbf{u}}_g{\mathbf{u}}_g\right)=-{\alpha }_g▽p+▽·\left({\alpha }_g{\tau }_g\right)+{\alpha }_g{\rho }_g\mathbf{g}+\beta ({\mathbf{u}}_s-{\mathbf{u}}_g)}$
T4	${\displaystyle \frac{\partial {\alpha }_s{\rho }_s{\mathbf{u}}_s}{\partial t}+▽·\left({\alpha }_s{\rho }_s{\mathbf{u}}_s{\mathbf{u}}_s\right)=-{\alpha }_s▽p+▽·\left({\alpha }_s{\tau }_s\right)+{\alpha }_s{\rho }_s\mathbf{g}+\beta ({\mathbf{u}}_g-{\mathbf{u}}_s)}$
	k-$\epsilon$ model
T5	${\displaystyle \frac{\partial }{\partial t}\left({\alpha }_g{\rho }_{g}k\right)+▽·\left({\alpha }_g{\rho }_g{u}_{g}k\right)=▽·\left({\alpha }_g\frac{{\mu }_t}{{\sigma }_k}▽k\right)+{\alpha }_g{G}_k-{\alpha }_g{\rho }_g\epsilon }$
T6	${\displaystyle \frac{\partial }{\partial t}\left({\alpha }_g{\rho }_g\epsilon \right)+▽·\left({\alpha }_g{\rho }_g{u}_g\epsilon \right)=▽·\left({\alpha }_g\frac{{\mu }_t}{{\sigma }_{\epsilon }}▽\epsilon \right)+{\alpha }_g\frac{\epsilon }{k}(C_1{G}_K-C_2{\rho }_g\epsilon )-{\alpha }_g{\rho }_g\epsilon }$
	Transport equation of granular temperature
T7	${\displaystyle \frac{3}{2}\left[\frac{\partial }{\partial t}\left({\alpha }_s{\rho }_s\theta \right)+▽·\left({\alpha }_s{\rho }_s\theta \right)u_s\right]=(-▽p_s\mathbf{I}+{\tau }_s):▽u_s+▽·(k_s▽\theta )-{\gamma }_s-3\beta \theta +D_{gs}}$
	Species transport equation
T8	${\displaystyle \frac{\partial {\alpha }_g{\rho }_g{Y}_{g,i}}{\partial t}+▽·\left({\alpha }_g{\rho }_g{\mathbf{u}}_g{Y}_{g,i}\right)=▽·\left({\alpha }_g{J}_{g,i}\right)+S_{g,i}}$
T9	${\displaystyle \frac{\partial {\alpha }_s{\rho }_s{Y}_{s,i}}{\partial t}+▽·\left({\alpha }_s{\rho }_s{\mathbf{u}}_s{Y}_{s,i}\right)=▽·\left({\alpha }_s{J}_{s,i}\right)+S_{S,i}}$
	Transport equation of enthalpy
T10	${\displaystyle \frac{\partial }{\partial t}\left({\alpha }_g{\rho }_g{H}_g\right)+▽·\left({\alpha }_g{\rho }_g{\mathbf{u}}_g{H}_g\right)=▽\left({\kappa }_g▽T_g\right)+h\left(T_g-T_s\right)+{\phi }_g}$
T11	${\displaystyle \frac{\partial }{\partial t}\left({\alpha }_s{\rho }_s{H}_s\right)+▽·\left({\alpha }_s{\rho }_s{\mathbf{u}}_s{H}_s\right)=▽\left({\kappa }_s▽T_s\right)+h\left(T_s-T_g\right)+\phi }$

and

Table A2. Conservation equation.

	Stress tensor
C1	${\displaystyle {\tau }_g={\mu }_g[▽{\mathbf{u}}_g+{(▽{\mathbf{u}}_g)}^T]-\frac{2}{3}{\mu }_g(▽·{\mathbf{u}}_g)I}$
C2	${\displaystyle {\tau }_s={\mu }_s[▽{\mathbf{u}}_s+{(▽{\mathbf{u}}_s)}^T]-({\lambda }_s-\frac{2}{3}{\mu }_s)(▽·{\mathbf{u}}_s)I}$
	Interphase momentum transfer coefficient
C3	${\displaystyle \beta =\left\{\begin{array}{cc}\frac{3}{4}\left(\frac{C_D{\alpha }_g{\alpha }_s{\rho }_g|{\mathbf{u}}_g-{\mathbf{u}}_s|}{d_s}\right){\alpha }_s^{-2.65}\hfill & {\alpha }_s<0.2\hfill \\ 150\left(\frac{{\mu }_g{\alpha }_s^2}{{\alpha }_s{d}_s^2}\right)+1.75\left(\frac{{\alpha }_s{\rho }_g}{d_s}\right)|{\mathbf{u}}_g-{\mathbf{u}}_s|\hfill & {\alpha }_s\ge 0.2\hfill \end{array}\right.}$
C4	${\displaystyle C_D=\left\{\begin{array}{cc}\frac{24}{R_e{\alpha }_g}\left[1+0.15{\left(R_e{\alpha }_g\right)}^{0.687}\right]\hfill & R_e<1000\hfill \\ 0.44\hfill & R_e\ge 1000\hfill \end{array}\right.}$
C5	${\displaystyle R_e=\frac{{\rho }_g{d}_s|{\mathbf{u}}_g-{\mathbf{u}}_s|}{{\mu }_g}}$
	shear viscosity
C6	${\displaystyle {\mu }_g=C_{\mu }\frac{k^2}{\epsilon }}$
C7	${\displaystyle {\mu }_s=\frac{4}{5}{\alpha }_s^2{\rho }_s{d}_s{g}_0(1+e)\sqrt{\frac{\theta }{\pi }}+\frac{10{\rho }_s{d}_s\sqrt{\pi \theta }}{96(1+e){\alpha }_s{g}_0}{\left[1+\frac{4}{5}{g}_0{\alpha }_s(1+e)\right]}^2}$
	Solid pressure
C8	${\displaystyle p_s={\alpha }_s{\rho }_s\theta +2{\rho }_s(1+e){\alpha }_s^2{g}_0\theta }$
	Conductivity of fluctuating energy
C9	${\displaystyle k_s=\frac{25{\rho }_s{d}_s\sqrt{\pi \theta }}{64(1+e)g_0}{\left[1+\frac{6}{5}(1+e)g_s{\alpha }_s\right]}^2+2{\alpha }_s^2{\rho }_s{d}_s{g}_0(1+e){\left(\frac{\theta }{\pi }\right)}^{1/2}}$
	Dissipation rate of fluctuating energy
C10	${\displaystyle {\gamma }_s=3(1-e^2){\alpha }_s^2{\rho }_s{g}_0\theta \left(\frac{4}{d_s}\sqrt{\frac{\theta }{\pi }}-▽·u_s\right)}$
	Heat transfer coefficient
C11	${\displaystyle h=\frac{6{\alpha }_s{\alpha }_g{\lambda }_{g}Nu}{d_p^2}}$
C12	${\displaystyle Nu=(7-10{\alpha }_g+5{\alpha }_g^2)(1+0.7R{e}^{0.2}P{r}^{1/3})+(1.33-2.4{\alpha }_g+1.2{\alpha }_g^2)R{e}^{0.7}P{r}^{1/3}}$
C13	${\displaystyle Pr=\frac{{\mu }_g{C}_g}{{\lambda }_g}}$

in the Appendix A, which has been tested in our previous works [22,28,29]. The fuel used in current work is methane and the reactions of oxygen carrier are following

12Fe₂O₃ + CH₄ → 8Fe₃O₄ + CO₂ + 2 H₂O
4Fe₃O₄ + O₂ → 6Fe₂O₃

The reaction process is described by the changing grain size model [30,31]. Considering the effect of reaction kinetics, the reaction procedure is described as the following:

$$\frac{t}{{\tau }_{chr}}=1-{(1-X_i)}^{1/3}$$

(1)

$${\tau }_{chr}=\frac{{\rho }_m{r}_g}{\overline{b}{k}_s{C}_g^n}$$

(2)

where $X_i$ presents the conversion rate and ${\tau }_{chr}$ presents the corresponding time, which can be calculated based on

Table 1. Kinetic parameters for reduction and oxidation of oxygen carriers.

${\mathit{\rho }}_{\mathit{m}}$ (mol m ${}^{-3}$)	${\mathit{r}}_{\mathit{g}}$	$\overline{\mathit{b}}$	${\mathit{k}}_{\mathit{s}0}$ (mol${}^{1-\mathit{n}}$m${}^{3\mathit{n}-2}$s${}^{-1}$)	${\mathit{E}}_{\mathbf{chr}}$ (kJ mol${}^{-1}$)	n
13,590	$1.25\times {10}^{-6}$	5.78	9.8	$135.2\pm 6$	1

[30].

The conversion rate, $X_i$, is defined as

$$X_r=\frac{m_o-m}{R_{o,ilm}{m}_o}$$

(3)

where $R_{o,ilm}$ represents the oxygen transport capacity. The homogeneous reactions, $C{H}_4+2O_2\to C{O}_2+2H_{2}O$, is considered is this work with the reaction rate ${\displaystyle \frac{d\left[C{H}_4\right]}{dt}=-k{\left[C{H}_4\right]}^{0.7}{\left[O_2\right]}^{0.8}}$, ${\displaystyle k=1.580\times {10}^{23}exp\left(-\frac{24343}{T_g}\right)}$[32].

3. Simulation Setup

A chemical looping combustion system with dual interconnected reactors, as shown in Figure 2, is adopted to investigate the reactive characteristics and hydrodynamic behaviour. It consists of an air reactor, fuel reactor, cyclone and loop seal. The configuration used here is built according to the work of Johansson et al. [33] and Guann et al. [34]. The total height is 1.9 m. The diameter is 0.19 m for both air reactor and fuel reactor. $k-\epsilon$ model is adopted for simulating gas turbulence. A non slip boundary condition is set for the wall. The methane is similarly adopted to the experiment and ilmenite is utilized as the oxygen carrier. More information about the numerical method can be found in our previously published works [28,35].

4. Results and Discussion

4.1. Flow Behavior of CLC System

The profile of instantaneous distribution of solid volume fraction is displayed in Figure 3. The particles circulate between reactors to exchange the lattice oxygen. Initially, the particles are located at the bottom of air reactor, fuel reactorand the loop seal. When the fluidizing gas is introduced to the reactor, the particles switch to the fluidizing state and different flow regime can be found in the CLC system. The particles in AR are entrained by gas phase and flow out of the riser, then they are separated by cyclone and enter FR. With the loop seal, particles will be sent back to AR and realize the circulation. Here, the loop seal works with a low velocity for preventing the exchange between AR and FR, which is prohibited for CLC system. The corresponding flow rate of the oxygen carrier is shown in Figure 4. The measured position is near the outlet of riser. The flow rate fluctuates with the time which is caused by the formation and breakup of bubbles and the interaction between gas and solid. After 35 s, the reactors tend to be steady state, so the solid flow rate of AR is equal to FR, and to the loop seal. Additionally, the solid flow rate is hard to measure during the experiment, while we can get it from the simulation. Under current conditions, the solid circulating rate is about 0.4 kg/s.

4.2. Reaction Behavior of CLC System

Figure 5 shows the profile of the gas species mass fraction at the outlets of AR and FR with the inlet rate of methane 0.5 g/s with temperature 200 ${}^{°}$C. The total solid inventory of the oxygen carrier is 34.6 kg. At the initial time, some methane crosses the bed materials due to the insufficient oxygen carrier. After 40s, the mass fraction of gaseous species reaches a steady value and the particles have shifted from air reactor to the fuel reactor for the balance of mass. The methane is not fully converted under current conditions. The main products detected at the outlet of the fuel reactor are CO${}_2$ and H${}_2$O. In the air reactor, the oxygen is consumed by the oxidation of oxygen carrier, and this value is equal to the consumption of oxygen in the fuel reactor. Much less CO${}_2$ is found in the air reactor due to leakage of the gas from loop seal.

The distribution of temperature in both AR and FR is shown in Figure 6. In AR, the oxygen carrier reacts with oxygen, which is an exothermic process. In FR, the reduction of the oxygen carrier is an endothermic process. As shown in Figure 6, the temperature in AR is low at the bottom due to the injection of air with a low temperature. Then, the temperature increases because of the heat transfer between oxygen carrier and gas phase. The decrease in temperature at a height of 0.3 m is caused by the inlet of the loop seal, where the cold gas and particles are injected into the air reactor. Then, temperature keeps increasing due to the reaction heat releasing during the oxidation process of oxygen carrier. At the bottom of FR, the temperature is high because oxygen carriers from the air reactor with high temperature enter FR. The temperature decreases with height and it is nearly constant (957K) at the end of the reactor. The distribution of temperature in the system is consistent with the theoretical analysis. We get a similar trend of temperature as in the work of Want et al. [17] on a CLC simulation.

To assess the state of the oxygen carrier, the oxidation level is calculated at the inlet and outlet of AR and shown in Figure 7. The difference between the two curves reflects the oxygen transfer between AR and FR, also the oxygen consumption in the fuel reactor. These values are the same with the outlet and inlet value of the fuel reactor, since the reactors are connected. Initially, fully reduced particles fill the loop seal for maintaining the temperature of the system. So, the oxidation level is low at the inlet of the air reactor at the beginning. After 30 s, the system becomes steady and the oxidation level is around 0.935 at the inlet of the AR. In fact, the oxidation level is decided by the reaction of the oxygen carrier in FR. For the outlet of the air reactor, the oxidation level is about 1, and the oxygen carrier has been fully oxidized. Based on the solid circulating rate and the conversion rate of methane, we can calculate the theoretical oxidation level, as shown in Figure 7 in blue. The simulated result is close to the theoretical one. Moreover, we find that the oxidation of the oxygen carrier is fast and the reduction process is the limitation of system.

4.3. Influence of Fuel Inlet Rate

The influence of the inlet rate of methane is tested, which corresponds to the CO${}_2$ capture capacity of the system. The mass fraction of gaseous species at the outlet of the FR are shown in Figure 8 for three cases with an inlet rate of 0.25 g/s, 0.5 g/s and 1.0 g/s. The CO${}_2$ fraction decreases with the methane inlet rate and the same tendency for the water vapour. Additionally, more methane is found at the FR outlet, because of the short contact time between gas and oxygen carrier. The corresponding distribution of gas species is illustrated in Figure 9. It shows intuitively the variety of the product and reactant with inlet rate. We can find CO${}_2$ in the loop seal and air reactor, which is caused by the leakage and is quite low in this work. The results are consistent with the work of Johansson [33].

The performance of CLC can be assessed by fuel conversion rate, which is defended as following [36],

$$X_{C{H}_4}=\frac{F_{in,C{H}_4}-F_{out,C{H}_4}}{F_{in,C{H}_4}}$$

(4)

The carbon capture efficiency used in this work is defined as following [37],

$${\eta }_{cc}=\frac{F_{FR,C{O}_2}-F_{AR,C{O}_2}}{F_{in,C{H}_4}}$$

(5)

From Figure 10, the conversion rate decreases with the methane injection rate, also less methane is converted into CO${}_2$ under current conditions. The resistance time decreases when increasing the FR inlet rate. Under current conditions, the operating temperature of FR is about 940 K and the total mass inventory is 34.6 kg, it is hard for the system to convert more methane and most of it escapes the reactor. The capture efficiency is quite low with a high inlet rate of methane. One possible solution is to increase the inventory or the operating temperature.

4.4. Influence of Operating Temperature

The performance of the CLC system under different temperatures is analysed in this section. As shown in Figure 11, the mass fraction of CH${}_4$ at the FR outlet decreases with temperature, which is due to the kinetic reaction increasing with temperature. More products, CO${}_2$ and H${}_2$O, are generated simultaneously. The nitrogen is also detected at FR outlet, which is from the loop seal. The conversion of fuel is an important parameter to assess the CLC performance. As shown on the right of Figure 11, the conversion rate of methane increases from 0.33 to 0.92 when raising the operating temperature from 930 K to 1290 K. The relationship is approximately linear, with a slope of 0.0024 per Kelvin. The oxygen demanded is the fraction of O${}_2$ required to fully burn the fuel and O${}_2$ needed by the unconverted gas exiting FR [38]. The similar effect of temperature has been presented in the work of Liu et al. [39]. From the figure, the oxygen demanded decreases with temperature, because more methane has been converted in FR with a higher temperature. Under current conditions, high operating temperature is beneficial for improving the performance of system.

5. Conclusions

A 3D model for a full loop CLC system is built in the frame of the Eulerican-Eurlerian approach for investigating the operation of a 10 kW gas-fueled CLC system. The temperature increases with height in the air reactor, due to the exothermic oxidation process of the oxygen carrier, and the temperature in the fuel reactor decreases slightly due to the reduction of oxygen carrier being a endothermic process. The temperature difference between reactors is 24 K The oxidation level is calculated during the simulation, which is difficult in the experiment. The difference in oxidation levels between the inlet and outlet of the air reactor is about 0.065, which is consistent with the theoretical value. The higher injection rate of methane will lead to a lower fuel conversion rate, due to the short resistance time and interaction time between the gas and oxygen carrier. The carbon capture efficiency decreases from 0.54 to 0.2 when the gas flow rate increases from 0.5 g/s to 1.0 g/s The operating temperature is an essential parameter for the performance of the CLC system. Under certain cases, the conversion rate of methane reaches 0.92 when raising the operating temperature to 1290 K. The industrial simulations will be carried out in following research based on the current work.