Influence of Torrefaction on Biomass Gasification ... - ACS Publications

Article pubs.acs.org/EF

Influence of Torrefaction on Biomass Gasification Performance in a High-Temperature Entrained-Flow Reactor Xiaoke Ku,*,† Jianzhong Lin,*,†,‡ and Fangyang Yuan† †

Department of Mechanics, State Key Laboratory of Fluid Power Transmission and Control, Zhejiang University, Hangzhou 310027, China ‡ China Jiliang University, Hangzhou 310018, China ABSTRACT: In the present work, the gasification performances of two pairs of raw and torrefied biomasses, including raw forest residue and torrefied forest residue, and raw spruce and torrefied spruce in a high-temperature entrained-flow reactor, are numerically examined and compared to each other using a Eulerian−Lagrangian CFD model developed based on OpenFOAM. Moreover, the sensitivities of three important operating parameters (excess air ratio, steam/carbon molar ratio, and biomass particle diameter), which vary in the range of practical significance, are also tested. The calculated results are analyzed both qualitatively and quantitatively by five indicators: isothermal profiles, char consumption rate, gas compositions, species yield, and carbon conversion along the reactor length. The obtained results show that torrefied biomass can increase the maximum temperature in the reactor as compared to its raw parent fuel. During gasification, CO, H2, and CO2 are the major species in the product gas and CH4 accounts for only a very small fraction of the syngas at such a high operating temperature of 1400 °C. As expected, the higher the excess air ratio, the lower the H2 yield and higher CO2 production and carbon conversion; a rise in the steam/carbon molar ratio promotes the H2 yield but reduces the CO production; and both the H2 and CO yields and carbon conversion decrease with increasing particle diameter. In addition, three reaction zones can be recognized from the carbon conversion along the reactor length, and such information is useful for optimal reactor design. In all cases, torrefaction consistently reduces both the H2 production and carbon conversion as compared to its raw biomass under the same operating conditions, which suggests that torrefied biomass requires a bigger length entrained-flow gasifier or longer particle residence time to achieve the same level of conversion as that for its raw biomass. temperature range of 200−300 °C. After torrefaction, the fuel properties are improved and the torrefied product has lower moisture content, better hydrophobicity, intensified energy density, and improved grindability and storability in comparison to its raw parent biomass. Moreover, biomass fibers are shortened and the pulverized particles become more spherical by torrefaction, which makes torrefied biomass more easily fluidizable and less prone to agglomeration in the pneumatic dense flow feeding system.18 However, very limited works have been dedicated to the thermochemical conversion behavior of torrefied biomass and there were also some discrepancies in the literature regarding the effect of torrefaction. Chen et al.17 compared the gasification phenomena of raw bamboo, torrefied bamboo, and high-volatile bituminous coal. It was found that the gasification performance of the torrefied bamboo was drastically improved compared to that of raw bamboo. Fisher et al.18 experimentally investigated the effect of torrefaction on the oxidation and gasification reactivity of chars from raw and torrefied willow. Results showed that torrefaction significantly reduced char reactivity and its negative impact was greatest for high-heating-rate char. Kuo et al.19 used Aspen Plus simulator (version 7.3) to thermodynamically evaluate the gasification performances of raw and torrefied bamboo in a fixed-bed reactor, although the simulation conditions were strongly simplified; e.g.,

1. INTRODUCTION Biomass is an abundant, renewable, and environmentally carbonneutral energy resource. Its enhanced utilization will reduce both the dependence on fossil fuels and the net CO2 emissions. One of the most promising technologies for this purpose is biomass gasification which is a thermochemical conversion process that can transform many biomass feedstocks (e.g., straw, wood, forest residues, and so on) into a gas mixture of CO, H2, and CH4 called synthetic gas (syngas). The syngas can be used then in a wide variety of industrial applications, such as liquid transportation fuels (methanol and DME), heat and electricity production, and chemical products (plastics and ammonia). Up to now, the most applied reactors for biomass gasification in industry have still been fixed-bed1−3 and fluidized-bed reactors.4−7 Although entrained-flow reactor for coal gasification is well-researched and widely used because of its high throughput and carbon conversion, low pollutant emissions, and better quality product gas,8−13 the adaption of entrained-flow gasification to biomass is still under research and development due to the inherent limitations of raw biomass properties such as heterogeneous shape and size, high moisture content, and low energy density.14−16 To enable and facilitate the biomass gasification in entrainedflow reactors, a variety of pretreatment methods for improving the properties of raw biomass have been developed,17 among which torrefaction seems to be a very promising approach. Torrefaction is a mild thermal degradation process during which raw biomass is heated under an inert environment in the © 2016 American Chemical Society

Received: January 22, 2016 Revised: April 15, 2016 Published: April 21, 2016 4053

DOI: 10.1021/acs.energyfuels.6b00163 Energy Fuels 2016, 30, 4053−4064

Article

Energy & Fuels

gasification performance. Finally, a short conclusion is presented in Section 5.

the biomass gasification process was assumed to be isothermal and in steady state. Furthermore, the gasifier was assumed to be running at the thermodynamic equilibrium state, which was only valid in the case of a sufficiently long residence time of reactants. More recently, in our earlier works,20,21 the effect of torrefaction on physical properties and rapid devolatilization behaviors of forest residue and Norwegian spruce in a drop-tube reactor were experimentally investigated. Results showed that torrefaction reduced the particle size after the same milling and sieving procedure and the char yields of the torrefied biomass fuels were higher than those of raw biomasses. But the main focus of our previous works was on the biomass pyrolysis process, and the influence of torrefaction on the progress of biomass gasification process in a high-temperature entrained-flow reactor, from the time that the feedstocks are introduced into the gasifier until they exit the outlet, has not been systematically explored. As the availability of high-performance cluster computing resources has been rapidly increased, CFD simulation has become a powerful tool for enhancing the understanding of the complex turbulent multiphase reactive flow in gasification reactors. In contrast to coal gasification, very few works have been done to study the biomass gasification by CFD analysis. In our earlier paper,22 a multiscale Eulerian−Lagrangian CFD model for biomass pyrolysis and gasification was developed using the open source C++ toolbox OpenFOAM (version 2.1.1)23 and fully validated by experimental data from a laboratory-scale entrained-flow gasifier at the Technical University of Denmark (DTU).15 Furthermore, the sensitivities of different particle shrinkage and devolatilization models were also explored in a more recent work.24 In general, the proposed integrated CFD model can be a very useful tool to evaluate the biomass gasification performance by changing the fuel properties and operating conditions without expensive and laborious experimental measurements. Therefore, the current work will use the developed CFD model to numerically explore the effect of torrefaction on biomass gasification performance in a hightemperature (1400 °C) entrained-flow reactor. Correspondingly, two pairs of raw and torrefied biomasses (raw forest residue, torrefied forest residue, raw spruce, and torrefied spruce) are adopted. Moreover, three important operating parameters (excess air ratio, steam/carbon molar ratio, and mean particle diameter) are also chosen to provide different scenarios in which the effect of torrefaction can be comprehensively tested. Several experimental and numerical works on the impact of these parameters on biomass gasification have been reported.22,25−27 For example, Hernández et al.25,26 found that both a reduction in the biomass particle size and the addition of steam had a positive effect on the gas quality and process performance. However, all of these studies were exclusively performed using raw biomasses as feedstocks, and very few works directly compare raw and torrefied biomasses. Therefore, the objective of the present work is to directly and systematically compare the gasification characteristics between raw and torrefied biomasses when the key operating parameters vary in ranges of practical significance. Such information is very important for identifying reaction zones and the optimal design of torrefied biomass gasifier. This work is structured as follows: Section 2 presents a brief overview of the numerical models used for biomass gasification in an entrained-flow reactor. Section 3 gives the simulation setup and the detailed operating conditions for each test case. Section 4 shows the simulation results which highlight the effects of torrefaction and different operating parameters on the biomass

2. MATHEMATICAL MODELING The integrated CFD model used, which is based on the Eulerian−Lagrangian concept, was developed and implemented using an open source C++ toolbox OpenFOAM (version 2.1.1).23 Details of the governing equations, the numerical schemes, and chemical reaction equations as well as the reaction constants were described in our earlier publication,22 and here only a brief overview is provided. The discrete fuel particles are assumed to be spherical with a prescribed size distribution. Each particle consists of a mixture of moisture, volatile matter, char, and ash. The particle trajectory is tracked in a Lagrangian manner, and influence of the turbulent flow on the particle trajectory is modeled using a stochastic turbulent dispersion model. During biomass gasification, there are basically three sequential mass and heat processes: drying, pyrolysis, and relatively slow char consumption. After the moisture evaporates, the release of volatiles during pyrolysis is modeled according to a single-step first-order Arrhenius reaction equation.8 Furthermore, volatile matter is considered as a mixture of light gases (CH4, H2, CO2, and CO) and tar formation during pyrolysis is disregarded. This is a reasonable hypothesis for high-temperature entrained-flow gasification, where the tar produced during pyrolysis is a nonstable product and very easily decomposes into light gases at high operating temperature.28 This assumption is also consistent with the new experimental findings of Newalkar et al.14 Because of the small particle sizes used in the entrained-flow reactor, the constant volume model with variable density during the particle conversion is adopted.24 After devolatilization, the fuel particle contains only char and ash. It is assumed that char is pure carbon. Several works in the literature reported the catalytic effect of biomass ash on the char reactivity, because of some alkali metal constituents (e.g., K, Na, and Ca) contained in the ash which are believed to be potentially catalytically active mineral substances.25−27,29 However, the catalytic effect is exclusively studied by experimental measurements and K has been observed to form silicates in the presence of Si that eventually deactivates its catalytic effect.30 Thus, to what extent the char consumption rate will be catalytically affected by biomass ash still needs to be further explored. Moreover, there is no simple catalytic model which could be easily combined with the three-dimensional CFD model. Therefore, the ash here is assumed to be transported with the particle and exits the reactor without taking part in any reactions. Char consumption is modeled by three heterogeneous reactions (i.e., partial oxidation reaction and endothermic gasification reactions with H2O and CO2), and the consumption rate takes both the bulk diffusion and kinetic effects into account. At each time step, the particle temperature is calculated considering the convective heat transfer, radiation, the latent heat of vaporization of moisture, and the heat generated by char reactions. Radiation is taken into account using the P-1 radiation model. Note that, due to the lack of specific reaction constants for each biomass species, the same reaction constants, which are suitable for hardwood pyrolysis and char conversion and based on the experimental work of Prakash and Karunanithi,31 are adopted for all of the biomass feedstocks studied in this work. However, new experimental studies need to be carried out in order to obtain the particular reaction constants for both raw and torrefied biomasses, which can be included in the future to further enhance the accuracy of the CFD model. 4054


Article

Energy & Fuels

Table 1. Properties of Raw and Torrefied Biomass Fuels21

For continuous gas phase, the three-dimensional Navier− Stokes equations, together with governing mass, energy, and species conservation equations, are solved using OpenFOAM codes. Turbulence is modeled using the standard k − ε approach. The coupling between the gas phase and the fuel particles is resolved by treating the interphase exchange of mass, momentum, and energy as source terms in the governing equations. In addition, extra source terms considering the homogeneous gas phase reactions and the radiation are also included. Five global gas reactions (i.e., the H2 and CO oxidation reactions, the CH4 oxidation and steam reactions, and the reversible water−gas shift reaction) are chosen to describe the main gas phase chemistry during biomass gasification, and the partially stirred reactor (PaSR) model accounts for the turbulence−chemistry interaction.8

RFR

TFR

RS

Proximate Analysis (wt %, As-Received Basis) moisture 6.3 4.2 5.0 volatiles 70.0 61.6 77.1 fixed carbon 21.5 31.5 17.5 ash 2.2 2.7 0.4 Elemental Analysis (wt %, Dry and Ash-Free Basis) C 52.1 59.5 46.7 H 6.1 5.6 6.2 O 41.3 34.3 47.0 others 0.5 0.6 0.1

TS 3.8 72.3 23.4 0.5 52.8 5.8 41.3 0.1

Table 2. Volatile Compositions for the Four Biomass Fuels mass fraction (%)

3. SIMULATION SETUP The laboratory-scale atmospheric pressure entrained-flow reactor at the Technical University of Denmark (DTU) is investigated in the present work. Figure 1 shows the geometry and the main dimensions of the

component

RFR

TFR

RS

TS

H2 CO CO2 CH4

5.0 38.2 44.7 12.1

4.7 37.4 42.5 15.4

6.1 39.3 48.4 6.2

4.8 38.4 45.1 11.7

Figure 2 shows the three different particle size distributions used in the simulations. At reactor walls, the temperature is kept constant and the

Figure 1. (a) Three-dimensional grid system of the DTU entrained-flow reactor and (b) geometry of the top inlets. reactor.22 In short, the reactor features a 202 cm long vertical reactive section with an internal diameter of 8 cm and a detailed design of this reactor can be found in the literature.15 As shown in Figure 1, a grid system consisting of 281280 hexahedral cells is adopted to construct the 3D computational domain and the mesh is refined in the center inlet in order to better predict the devolatilization and combustion of the volatiles. A detailed grid independence study has been carried out in the previous work.22 Pulverized biomass at 300 K is injected into the reactor from the center inlet and a carrier air stream (10 N L/min) at 300 K is employed to aid in the transporting of the fuel particles. A preheated main gas stream, composed of air and steam, is introduced through the concentric annular inlet. Two pairs of raw and torrefied biomasses: raw forest residues (RFR), torrefied forest residues (TFR), raw spruce (RS), and torrefied spruce (TS),21 are selected as the feedstocks to explore the effect of torrefaction on the biomass gasification behavior. Spruce is widely planted in the northern part of China and in Nordic countries (e.g., Norway). It is a fast-growing evergreen coniferous tree and can be used as a pure woody biomass fuel. The torrefied fuels are produced from the torrefaction of raw biomasses at 275 °C for 30 min followed by natural cooling to room temperature. A more detailed description of the torrefaction process has been documented in our previous publication.21 The bulk densities of the four feedstocks (RFR, TFR, RS, and TS) are 248, 227, 227, and 186 kg/m3,21,32 respectively. The properties and volatile compositions of the feedstocks are summarized in Table 1 and Table 2, respectively. Note that the biomass volatile compositions are determined based on the elemental conservation relationships.22 In addition, a Rosin−Rammler distribution is used here to model the particle size distribution and

Figure 2. Particle size Rosin−Rammler distributions adopted in the simulations. emissivity is set to 0.9 which is typically adopted in the CFD modeling of turbulent combustion or gasification in an entrained-flow reactor.13,21,33 Details of the operating conditions for each case are listed in Table 3. Note that, in Table 3, the operating temperature Tr is 1400 °C which denotes the wall temperature of the reactor and is a boundary condition for the CFD simulations. This setup mimics the real experimental runs in which the external walls of the reactor are electrically heated and kept at constant temperature. However, the temperature profile inside the reactor will have an evolution because of the volatiles combustion and char reactions.

4. RESULTS AND DISCUSSION In this section, the gas species production, which is defined as the volume of gas component produced from the gasification of per kilogram of dry and ash-free fuel (expressed in N m3/(daf kg fuel)), is chosen to evaluate the influences of torrefaction and three important operating parameters on biomass gasification behavior. Because the gas production considers both the fuel conversion and gas quality, it will show valuable information on the total gasification process. 4.1. Validation. In contrast to coal, most of the chemical energy in biomass feedstocks is contained in the volatile matters rather than in the char. As shown in Table 1, for both raw and 4055


Article

Energy & Fuels Table 3. Simulation Cases for Biomass Gasification parameter λa

S/Cb

d p̅ c

a

case

Tr (°C)

S/C

λ

biomass fuel

d p̅ (μm)

fuel feeding rate (g/min)

air/steam flow rates through outer annular inlet (g/min)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

1400 1400 1400 1400 1400 1400 1400 1400 1400 1400 1400 1400 1400 1400 1400 1400 1400 1400 1400 1400 1400

0 0 0 0 0 0 0 0 0 0 0 0 0.5 0.5 0.5 0.5 1.0 1.0 1.0 1.0 0

0.25 0.25 0.25 0.25 0.3 0.3 0.3 0.3 0.35 0.35 0.35 0.35 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.3

RFR TFR RS TS RFR TFR RS TS RFR TFR RS TS RFR TFR RS TS RFR TFR RS TS RFR

300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 300 500

15.3 15.3 15.3 15.3 12.8 12.8 12.8 12.8 10.9 10.9 10.9 10.9 12.8 12.8 12.8 12.8 12.8 12.8 12.8 12.8 12.8

9.1/0 13.0/0 6.9/0 10.1/0 9.1/0 13.0/0 6.9/0 10.1/0 9.1/0 13.0/0 6.9/0 10.1/0 9.1/4.6 13.0/5.3 6.9/4.2 10.1/4.8 9.1/9.2 13.0/10.6 6.9/8.4 10.1/9.6 9.1/0

22 23 24 25 26 27 28

1400 1400 1400 1400 1400 1400 1400

0 0 0 0 0 0 0

0.3 0.3 0.3 0.3 0.3 0.3 0.3

TFR RS TS RFR TFR RS TS

500 500 500 700 700 700 700

12.8 12.8 12.8 12.8 12.8 12.8 12.8

13.0/0 6.9/0 10.1/0 9.1/0 13.0/0 6.9/0 10.1/0

λ, excess air ratio. bS/C, steam/carbon molar ratio. cd p̅ , mean particle diameter.

Figure 3. Particle weight loss along the reactor length for raw and torrefied biomass fuels under pyrolysis condition.

torrefied biomasses studied, the volatiles are always more than 60%. Therefore, the accurate modeling of the pyrolysis process is more important for biomass than coal. Here, we first evaluate our

integrated CFD model by testing the pyrolysis processes for both raw and torrefied biomass feedstocks. The corresponding experimental runs were conducted in a drop-tube reactor and 4056


Article

Energy & Fuels reported in the literature.21 In short, the reactor has a 1.5 m long reactive tube with 5.08 cm inner diameter. The operating temperature is 1200 °C, and the total N2 flow rate is 186 N L/ min. Note that under pyrolysis condition, N2 stream is adopted which is different from the air flow used under gasification condition. During the biomass pyrolysis process, the location where the devolatilization finishes can be determined from the particle weight loss curves. Figure 3 shows the average particle weight loss along the reactor length for the four biomass feedstocks. The horizontal dashed line denotes the total fraction of moisture and volatiles of the fuel based on the proximate analysis shown in Table 1. Through the intersection of the horizontal dashed line and the weight loss curve, a vertical dashed line is drawn downward to meet the z axis where the z coordinate is defined as the location where the devolatilization finishes. In order to see the intersection point clearly, the area within the red dashed circle is replotted in the inset of Figure 3. Table 4 shows the

well with the experimental data. Specifically, the minimum relative errors of H2, CO, CO2, and CH4 productions between the simulations and experimental measurements are 11.8%, 0.8%, 0.5%, and 8.9%, respectively, and the maximum relative errors are 23.5%, 25.5%, 6.1%, and 19.1%, respectively. Moreover, a comprehensive validation against the wide range of experimental data including different operating conditions was reported in our earlier work.22 The preceding analyses have demonstrated the validity of the proposed CFD model by comparing the predicted results for both raw and torrefied biomasses under pyrolysis and gasification conditions with the experimental data reported in the literature. In the following, the CFD model is further applied to numerically investigate the effect of torrefaction on biomass gasification performance. 4.2. Gasification Phenomenon. In this subsection, we first present some qualitative results. Figure 5 shows the predicted

Table 4. Comparison of Predicted Locations along the Reactor Length Where the Devolatilization Finishes with the Experimental Data21 z (m)a RFR

TFR

RS

TS

simulation experiment

0.51 0.53

0.46 0.51

0.53 0.51

0.52 0.57

relative error (%)

3.8

9.8

3.9

8.8

a

z, locations along the reactor length where the devolatilization finishes.

comparison of predicted locations along the reactor length where the devolatilization finishes with the experimental data reported by Li et al.21 It can be observed that good agreement is achieved and the maximum relative error is within 10% for both raw and torrefied feedstocks. In addition to the pyrolysis cases, we also compare the predicted results of the biomass gasification process with the experimental data reported by Qin et al.15 In the experiments, raw beech wood was fed into the entrained-flow reactor operated at five different temperatures (1000, 1100, 1200, 1300, and 1400 °C). Figure 4 shows the comparisons of H2, CO, CO2, and CH4 productions for the five operating temperatures. It can be observed that the calculated results at various temperatures agree Figure 5. Isothermal profiles for the gasification of raw and torrefied biomass fuels (cases 5−8 in Table 3).

isothermal contours in the axial section of the reactor at base conditions (Tr = 1400 °C, S/C = 0, λ = 0.3, and d p̅ = 300 μm) for raw and torrefied biomasses (cases 5−8 in Table 3). It is clearly seen that all fuels exhibit a similar temperature distribution and a high-temperature zone is always located near the fuel and air inlets, where the peak temperatures (around 2300 K) are easily recognized. They are typical of jet diffusion flames.8 The rapid increase in temperature in the vicinity of the inlets is attributed to the fast release of volatiles followed by its mixing with air and the exothermic combustion reactions. Consequently, the substoichiometric amount of O2 introduced is quickly consumed during the combustion of volatiles in the flame and the char combustion with O2 is unlikely to happen in the lower part of the reactor.

Figure 4. Gas production at the exit of the reactor as a function of reactor temperature: λ = 0.3, S/C = 0.5, and raw beech wood. 4057


Article

Energy & Fuels

Figure 6. Mass concentration profiles of (a) H2, (b) CO, (c) CO2, and (d) CH4 at t = 10 s for raw and torrefied biomass fuels (cases 5−8 in Table 3).

Specifically, the maximum temperatures of RFR, TFR, RS, and TS gasification in the reactor are 2307, 2336, 2333, and 2368 K, respectively, revealing that torrefaction can increase the maximum temperature by about 30 K, as compared to its raw parent fuel. The reason is that the torrefied biomass has a lower moisture content (see Table 1), which will lose less heat during the drying process. Moreover, the maximum temperature of RS

gasification is 26 K higher than that of RFR because RS contains relatively more volatiles (77.1%) than that of RFR (70.0%) (see Table 1). Figure 6 presents the mass concentration profiles of both major gas species (H2, CO, and CO2) and minor gas species (CH4) in the axial section of the reactor at base conditions for raw and torrefied biomasses (cases 5−8 in Table 3). Regardless 4058


Article

Energy & Fuels

Figure 7. (a) Mass concentration profiles of char and (b) char consumption rate at t = 10 s for raw and torrefied biomass fuels (cases 5−8 in Table 3).

7b). Further downstream, the slow char gasification reactions involving CO2 and H2O dominate and char consumption rate becomes lower (light blue color). Near the exit, many particles appear as a dark blue color (zero char concentration) and the char consumption rate is close to zero (dark blue color) because the char is completely converted and they contain only ash. In general, raw biomass fuels (RFR and RS) have more blue color particles (low char concentration) at the exit than their torrefied counterparts (TFR and TS) which implies that the raw fuels have a higher char conversion than their torrefied ones. Figures 5−7 overall demonstrate that the raw and torrefied biomasses exhibit qualitative similarities in the gasification phenomenon, although some quantitative differences are still discerned and will be explored in detail in the following sections. 4.3. Impact of Excess Air Ratio. The excess air ratio, λ, which is defined as the ratio between the supplied air mass flow and the air mass flow required for complete combustion of the fuel, is an important index to evaluate the performance of biomass gasification. λ can be changed either by varying the biomass injection rate while keeping the inlet air rate constant or vice versa. Here, the former tuning mechanism is used and λ changes within the range of 0.25−0.35 (cases 1−12 in Table 3). Note that the carrier air stream through the center inlet is also taken into account when calculating the λ. Figure 8 shows the gas species production at the reactor exit as a function of λ for raw and torrefied biomasses. It is seen that the H2 and CH4 productions decrease with increasing λ, regardless of which biomass is used as fuel. This is expected because more oxygen is supplied at bigger λ, leading to oxidation of H2 and CH4. In contrast, the CO curves appear to be rather flat, revealing that an approximately constant CO production is obtained with changing λ. Increasing λ will promote the oxidation of CO,

of which biomass is consumed, higher mass concentrations of H2, CO, CO2, and CH4 in the upper part of the reactor are clearly observed. This is due to the rapid release of volatiles after the fuel injection. In the lower part of the reactor, CO mass concentration becomes higher and CO2 concentration is lower because of the char gasification reaction. In addition, CH4 is quickly consumed and its concentration is very low (dark blue color) in the major part of the reactor (Figure 6d). Torrefaction reduces the mass concentrations of H2, CO, and CO2 in the upper part of the reactor because the torrefied biomass has a lower volatile matter compared to its raw parent fuel (see Table 1), which will release less volatiles during pyrolysis. At the exit, the mass concentrations of H2 and CO from TFR gasification are lower than those of RFR (Figure 6a,b). Similarly, the concentration of H2 of TS gasification is also lower than that of RS but the CO of TNS looks similar to NS (Figure 6a,b). Unlike H2 and CO, the mass concentrations of CO2 looks similar for the four biomasses (Figure 6c). Figure 7 plots the mass concentration of char remaining in the particles and the individual char consumption rate for raw and torrefied biomasses. When a fuel particle is injected into the reactor, it will undergo the following processes: vaporization of moisture, rapid release of volatiles, and slow char reactions. As shown in Figure 7a, just at the top inlets, the particles display light blue color (low char concentration) because they still contain lots of moisture and volatiles and the char mass concentration is low. At this location, the char consumption rate is zero (dark blue color) as shown in Figure 7b. Downstream, most of the particles become a red color (high char concentration; see Figure 7a) because they have only char and ash due to the fast release of moisture and volatiles. Meanwhile, the char consumption rate here is high (red color) due to the existence of O2 (see Figure 4059


Article

Energy & Fuels

Finally, the influence of torrefaction on CO2 production seems to be insignificant. The carbon conversion (CC) is a crucial index used as a measure of the gasification performance.17 Figure 9 plots the CC

Figure 8. Species production at the exit of the reactor as a function of excess air ratio λ: (a) RFR and TFR (cases 1, 2, 5, 6, 9, and 10 in Table 3), and (b) RS and TS (cases 3, 4, 7, 8, 11, and 12 in Table 3).

Figure 9. Carbon conversion along the reactor length for raw and torrefied biomass fuels at different excess air ratios λ: (a) RFR and TFR (cases 1, 2, 5, 6, 9, and 10 in Table 3) and (b) RS and TS (cases 3, 4, 7, 8, 11, and 12 in Table 3). Three different zones are marked by I, II, and III.

whereas it will also strengthen the partial oxidation of char. The former abates the CO production, but the latter intensifies it. When these two factors are considered together, the CO production thus keeps almost constant within the narrow range of λ tested. In addition, the CO2 production increases slightly with λ due to the oxidation of CO. Note that the CH4 production is plotted on a different scale from that of other species and it is 2 orders of magnitude smaller than the productions of the major gas species (H2, CO, and CO2). Generally the CH4 production is very low at such a high operating temperature (1400 °C), and this result is consistent with the experimental findings of Weiland et al.16 A comparison between raw and torrefied biomasses indicates that torrefaction reduces H2 production but increases CH4 yield. The reason is that torrefied biomasses have a lower moisture content (see Table 1) which will retard the water−gas shift reaction, steam gasification of char, and steam re-forming of CH4. This will lead to a smaller H2 production and a bigger CH4 yield. Quantitatively, for forest residue and spruce, torrefaction reduces H2 production by factors of 11.4−14.7% and 3.9−8.2%, respectively, within the λ range. In contrast to H2, the impact of torrefaction on the CO production is not consistent for the two types of biomass. For forest residue, the effect of torrefaction on CO is very slight. However, for spruce, torrefaction amplifies the CO production by factors of 4.5−8.2%. This difference is mainly due to the different natures of the biomass species.

along the reactor length as a function of λ for raw and torrefied biomasses under the same operating conditions. As shown in Figure 9, three different zones can be recognized: (I) drying, (II) fast pyrolysis and char oxidation, and (III) slow char gasification and re-forming zones. In order to see the drying stage clearly, the region near the inlets is replotted in the insets of Figure 9. In zone I, the CC does not start and the cold biomass particles entering the reactor are heated and the moisture evaporates. This zone locates very close to the inlets and only covers a reactor length of about 0.03 m which can be seen in the insets of Figure 9. Note that the raw biomass (RFR or RS) has a longer drying process than its torrefied product (TFR or TS) due to the higher moisture content in the raw fuel (see Table 1). In zone II, where the drying has already finished, the fuel particles are further heated and rapidly consumed. The CC quickly increases mostly because of the fast devolatilization and partly because of the volatiles combustion and char oxidation which leads to a hightemperate flame region as shown in Figure 5. In zone III, where O2 is completely consumed and the pyrolysis has already finished, the CC becomes more slow which is dominated by the slow char gasification reactions with CO2 and H2O. Also obviously, CC rises monotonically with increasing λ, regardless of which fuel is gasified. There are several reasons for this. First, an increase in λ will promote the carbon oxidation 4060


Article

Energy & Fuels reactions. Second, an elevated λ will result in a higher gasification temperature which is advantageous for char gasification. Finally, in the constant volume reactor, increasing λ by decreasing the fuel load will increase the particle residence time inside the reactor.22 At λ = 0.35, a CC of over 90% can be achieved at the rector exit for all the fuels except for TFR which is featured by the lowest volatiles matter and highest fixed carbon (see Table 1). As also shown in Figure 9, although the torrefied biomasses generally have a similar carbon conversion behavior along the reactor length to raw fuels, torrefaction has a pronounced negative influence on the CC. Quantitatively, for forest residue, torrefaction reduces CC at the reactor exit by factors of 11.9%, 10.3%, and 8.6%, respectively, when λ = 0.25, 0.3, and 0.35. For spruce, the negative effect lessens and torrefaction lowers the CC at the reactor exit by factors of 6.4%, 5.4%, and 4.9%, respectively, when λ = 0.25, 0.3, and 0.35. Two major conversion mechanisms are responsible for carbon conversion: (i) the fast devolatilization in zone II and (ii) the slow char heterogeneous reactions in zone III. Compared to raw fuels, the torrefied fuels have a lower volatile matter and higher fixed carbon content (see Table 1), both of which are disadvantageous for carbon conversion. Thus, torrefied biomasses need a longer carbon conversion process than the corresponding raw fuels; i.e., a longer particle residence time is required to ensure complete conversion of torrefied fuels. 4.4. Impact of Steam/Carbon Molar Ratio. In this subsection, attention is paid to the effects of torrefaction and steam on the quantitative gasification results. We use the steam/ carbon molar ratio (denoted by S/C) as an indicator of steam supply. Accordingly, 12 different cases (cases 5−8 and 13−20 in Table 3), where the S/C ratio ranges from 0 to 1.0 for raw and torrefied biomasses with otherwise the same operating conditions, are taken into account. Note that the air and steam flow rates through the outer annular inlet are listed in Table 3. Figure 10 depicts the gas species production at the reactor exit as a function of S/C ratio for raw and torrefied biomasses. It is easily observed that increasing S/C ratio facilitates the H2 production, no matter which fuel is injected. This is a result of more hydrogen produced from both the water−gas reaction (C + H2O → CO + H2) and the water−gas shift reaction. In addition, the CO yield decreases and the CO2 production increases with an increase in S/C ratio, reflecting the role of water−gas shift reaction which can convert CO and steam to CO2 and H2. Again, the CH4 production is fairly low at Tr = 1400 °C and reduces with S/C ratio due to the promotion of the steam re-forming of CH4 at bigger S/C ratio. Regarding torrefaction, although RFR gets higher H2 and CO productions compared to TFR and RS obtains a little higher H2 and lower CO productions than TS, the differences between raw and torrefied biomasses are not significant and the effect of torrefaction on the productions of the major gas components (H2, CO, and CO2) is slight within the range of S/C ratio tested. For minor species CH4, again, torrefaction enhances the CH4 yield because of the lower moisture content in torrefied fuels. Figure 11 shows the CC along the reactor length as a function of S/C ratio for raw and torrefied biomasses. Three different zones are also marked by I, II, and III, and the drying stage is highlighted in the insets of the figure. Instead of the monotonic evolution with S/C ratio, the maximum CC occurs at S/C = 0.5 within the range of S/C ratio studied. This result is in consistency with the experimental findings of Hernández et al.26 At S/C = 0.5, the values of CC at the reactor exit from the gasification of RFR, TFR, RS, and TS are 89.1%, 78.9%, 93.7%, and 88.3%, respectively. Again, torrefaction has a negative effect on the

Figure 10. Species production at the exit of the reactor as a function of steam/carbon molar ratio S/C: (a) RFR and TFR (cases 5, 6, 13, 14, 17, and 18 in Table 3) and (b) RS and TS (cases 7, 8, 15, 16, 19, and 20 in Table 3).

CC under the same operating conditions. Compared to raw biomasses (RFR and RS), the values of CC at the reactor exit from the gasification of TFR and TS at S/C = 0.5 are reduced by factors of 11.4% and 5.8%, respectively. 4.5. Impact of Particle Diameter. Three mean particle diameters (d p̅ = 300, 500, and 700 μm), but otherwise the same operating conditions, are adopted to investigate the effects of torrefaction and particle size on the gasification results. Figure 12 shows the species production at the reactor exit as a function of d p̅ for raw and torrefied biomasses. Overall, the H2, CO, and CH4 productions decrease and the CO2 production slightly increases with an increase in d p̅ , regardless of whether the biomass is torrefied. Quantitatively, for forest residue, torrefaction reduces H2 and CO productions by factors of 14.6−22.6% and 0.5− 10.4%, respectively, within the d p̅ range tested. For spruce, torrefaction reduces H2 production by factors of 8.1−9.3% and increases CO production by factors of 2.7−4.5%. Again, torrefaction promotes the CH4 production and has very little effect on the CO2 production. Figure 13 depicts the CC along the reactor length as a function of d p̅ for raw and torrefied biomasses under the same operating conditions. In addition, three different zones are denoted by I, II, and III and the drying stage is redrawn in the insets of the figure. Clearly, the particle size has a distinct and consistent effect on CC 4061


Article

Energy & Fuels

Figure 11. Carbon conversion along the reactor length for raw and torrefied biomass fuels at different steam/carbon molar ratios S/C: (a) RFR and TFR (cases 5, 6, 13, 14, 17, and 18 in Table 3) and (b) RS and TS (cases 7, 8, 15, 16, 19, and 20 in Table 3). Three different zones are marked by I, II, and III.

Figure 12. Species production at the exit of the reactor as a function of mean particle diameter d p̅ : (a) RFR and TFR (cases 5, 6, 21, 22, 25, and 26 in Table 3) and (b) RS and TS (cases 7, 8, 23, 24, 27, and 28 in Table 3).

for all the biomasses; i.e., CC decreases with increasing d p̅ . Furthermore, the drying process becomes shorter with a decrease in d p̅ . This is reasonable because smaller fuel particles have a bigger ratio of external surface to volume and a relatively longer residence time compared to larger fuel particles, both of which are conducive to the particle heat-up and carbon conversion. Again, torrefaction reduces CC as compared to the raw parent biomass. Quantitatively, for forest residue, torrefaction reduces CC at the reactor exit by factors of 11.9%, 13.9%, and 11.8%, respectively, when d p̅ = 300, 500, 700 μm. For spruce, the decrease trend is less marked and torrefaction decreases CC at the reactor exit by factors of 6.4%, 7.3%, and 7.3%, respectively, when d p̅ = 300, 500, 700 μm. Therefore, it is expected that, under the same operating conditions, raw biomass particles can be relatively larger than torrefied biomass particles to achieve the same conversion. Generally, sufficient carbon conversions can be achieved for raw biomasses when d p̅ < 500 μm, whereas torrefied

the same CC as raw fuels. This conclusion is consistent with the experimental works reported in the literature.18,21 Even though torrefaction has a negative effect on the CC, it does not mean that the torrefied biomass is inferior to raw biomass. In contrast, it indirectly proves that the torrefied feedstock has a higher energy density than its raw fuel, which thus leads to a lower carbon conversion under the same operating conditions. Moreover, raw biomasses have many shortcomings, such as high moisture content, low energy density, high oxygen content, hygroscopic behavior, and very heterogeneous physical and chemical properties, which result in long-distance transport challenges, storage problems, and utilization limitations. On the other hand, torrefaction can release most of the moisture and some of the volatiles and hemicellulose, enhance the energy density, reduce the O/C and H/C ratios (see Table 1), and improve the hydrophobic nature and grindability property.19 Furthermore, torrefaction can also homogenize the physical and chemical properties of biomass fuels (e.g., shortened fibers and more spherical particle shape). All these characteristics make the torrefied feedstock superior to raw biomass for transportation, storage, grinding, and feeding. Recently, by plotting the increased O/C and H/C ratios on the Van Krevelen diagram, Xue et al.29 experimentally found that the fuel properties of torrefied biomass were similar to peat in terms of the fuel classification. This might imply that the torrefied biomass is more suitable than raw biomass for cofiring with pulverized coal in existing large industrial facilities.

biomasses need to be ground to a smaller particle size (e.g., d p̅ < 300 μm) in order to obtain a same high carbon conversion. This is feasible because torrefaction can disintegrate the biomass structure and significantly improve the grindability. 4.6. Discussion. In the preceding subsections, we find that the torrefied feedstock always has a lower carbon conversion than its raw counterpart under the same operating conditions (e.g., Tr, λ, S/C, and d p̅ ). This finding is useful for the reactor design, and it implies that a higher operating temperature or a longer length gasifier will be needed for torrefied feedstocks in order to achieve 4062


Article

Energy & Fuels

Very few studies directly compare the gasification performances between raw and torrefied biomasses in a high-temperature entrained-flow reactor. This work generally shows how the process temperature, product gas compositions, syngas yield, char consumption rate, and carbon conversion are affected by torrefaction as well as systematic variation of three important process parameters. In all cases, torrefaction consistently reduces both the H2 production and carbon conversion as compared to its raw parent biomass under the same operating conditions. The carbon conversion differences reveal that process control or reactor design may need to be adjusted when changing from a raw biomass to a torrefied product. From a process viewpoint, the results imply that torrefied biomass requires a longer burn-out time and a gasifier processing torrefied fuel will need to be longer to obtain the same level of conversion as will be achieved for raw biomass.

■

AUTHOR INFORMATION

Corresponding Authors

*(X. Ku) Tel.: +86 57187952221. E-mail: [email protected]. *(J. Lin) Tel.: +86 57186836009. E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank the National Natural Science Foundation of China for Grant No 11132008 and the startup fund from the ‘“hundred talents program”’ of Zhejiang University for financial support.

■

REFERENCES

(1) Couto, N.; Rouboa, A.; Silva, V.; Monteiro, E.; Bouziane, K. Energy Procedia 2013, 36, 596−606. (2) Razmjoo, N.; Sefidari, H.; Strand, M. Fuel Process. Technol. 2014, 124, 21−27. (3) Umeki, K.; Yamamoto, K.; Namioka, T.; Yoshikawa, K. Appl. Energy 2010, 87, 791−798. (4) Ku, X.; Li, T.; Løvås, T. Chem. Eng. Sci. 2015, 122, 270−283. (5) Meng, X.; de Jong, W.; Fu, N.; Verkooijen, A. H. M. Biomass Bioenergy 2011, 35, 2910−2924. (6) Song, T.; Wu, J.; Shen, L.; Xiao, J. Biomass Bioenergy 2012, 36, 258− 267. (7) Tremel, A.; Becherer, D.; Fendt, S.; Gaderer, M.; Spliethoff, H. Energy Convers. Manage. 2013, 69, 95−106. (8) Abani, N.; Ghoniem, A. F. Fuel 2013, 104, 664−680. (9) Chen, C.; Hung, C.; Chen, W. Appl. Energy 2012, 100, 218−228. (10) Kumar, M.; Ghoniem, A. F. Energy Fuels 2012, 26, 464−479. (11) Richter, A.; Vascellari, M.; Nikrityuk, P. A.; Hasse, C. Fuel Process. Technol. 2016, 144, 95−108. (12) Watanabe, H.; Otaka, M. Fuel 2006, 85, 1935−1943. (13) Á lvarez, L.; Gharebaghi, M.; Pourkashanian, M.; Williams, A.; Riaza, J.; Pevida, C.; et al. Fuel Process. Technol. 2011, 92, 1489−1497. (14) Newalkar, G.; Iisa, K.; D’Amico, A. D.; Sievers, C.; Agrawal, P. Energy Fuels 2014, 28, 5144−5157. (15) Qin, K.; Jensen, P. A.; Lin, W.; Jensen, A. D. Energy Fuels 2012, 26, 5992−6002. (16) Weiland, F.; Wiinikka, H.; Hedman, H.; Wennebro, J.; Pettersson, E.; Gebart, R. Fuel 2015, 153, 510−519. (17) Chen, W.; Chen, C.; Hung, C.; Shen, C.; Hsu, H. Appl. Energy 2013, 112, 421−430. (18) Fisher, E. M.; Dupont, C.; Darvell, L. I.; et al. Bioresour. Technol. 2012, 119, 157−165. (19) Kuo, P.; Wu, W.; Chen, W. Fuel 2014, 117, 1231−1241. (20) Li, T.; Geier, M.; Wang, L.; Ku, X.; Güell, B. M.; Løvås, T.; Shaddix, C. R. Energy Fuels 2015, 29, 177−184. (21) Li, T.; Wang, L.; Ku, X.; Güell, B. M.; Løvås, T.; Shaddix, C. R. Energy Fuels 2015, 29, 4328−4338.

Figure 13. Carbon conversion along the reactor length for raw and torrefied biomass fuels at different mean particle diameters d p̅ : (a) RFR and TFR (cases 5, 6, 21, 22, 25, and 26 in Table 3) and (b) RS and TS (cases 7, 8, 23, 24, 27, and 28 in Table 3). Three different zones are marked by I, II, and III.

5. CONCLUSION Gasification performances of two pairs of raw and torrefied biomasses (RFR and TFR; RS and TS) in a high-temperature entrained-flow reactor are numerically investigated by a multiscale Eulerian−Lagrangian CFD model developed in our earlier work.22 Moreover, three important operating parameters (λ, S/C, and d p̅ ) are chosen to evaluate the effect of torrefaction on the gasification behavior. The calculated results show that CO, H2, and CO2 are the major species in the product gas and CH4 constitutes a very small fraction of the syngas at such a high operating temperature of 1400 °C. As expected, the higher the λ, the lower the H2 production and higher the CO2 production; a rise in S/C promotes the H2 and CO2 yields but reduces the CO production; and increasing d p̅ decreases the H2 and CO productions but increases CO2 yield. In addition, carbon conversion (CC) along the reactor length for both raw and torrefied biomass fuels at different operating conditions are also presented, from which three reaction zones can be recognized. Moreover, CC increases with an increase in λ or a decrease in d p̅ regardless of whether the biomass is torrefied. 4063


Article

Energy & Fuels (22) Ku, X.; Li, T.; Løvås, T. Energy Fuels 2014, 28, 5184−5196. (23) OpenFOAM Documentation; OpenFOAM Foundation, London, U.K., 2012; http://www.openfoam.org/docs/. (24) Ku, X.; Li, T.; Løvås, T. Energy Fuels 2015, 29, 5127−5135. (25) Hernández, J. J.; Aranda-Almansa, G.; Bula, A. Fuel Process. Technol. 2010, 91, 681−692. (26) Hernández, J. J.; Aranda, G.; Barba, J.; Mendoza, J. M. Fuel Process. Technol. 2012, 99, 43−55. (27) Hernández, J. J.; Aranda-Almansa, G.; Serrano, C. Energy Fuels 2010, 24, 2479−2488. (28) Gerber, S.; Behrendt, F.; Oevermann, M. Fuel 2010, 89, 2903− 2917. (29) Xue, G.; Kwapinska, M.; Kwapinski, W.; Czajka, K. M.; Kennedy, J.; Leahy, J. J. Fuel 2014, 121, 189−197. (30) Lundberg, L.; Tchoffor, P. A.; Pallarès, D.; Johansson, R.; Thunman, H.; Davidsson, K. Fuel Process. Technol. 2016, 144, 323−333. (31) Prakash, N.; Karunanithi, T. J. Appl. Sci. Res. 2008, 4, 1627−1636. (32) Ragland, K. W.; Aerts, D. J.; Baker, A. J. Bioresour. Technol. 1991, 37, 161−168. (33) Velusamy, K.; Sundararajan, T.; Seetharamu, K. N. J. Heat Transfer 2001, 123, 1062−1070.

4064
