Oxy-fuel combustion of pulverized fuels: Combustion fundamentals and modeling

Applied Energy 162 (2016) 742–762

Contents lists available at ScienceDirect

Applied Energy journal homepage: www.elsevier.com/locate/apenergy

Review

Oxy-fuel combustion of pulverized fuels: Combustion fundamentals and modeling Chungen Yin a,⇑, Jinyue Yan b,c a

Department of Energy Technology, Aalborg University, 9220 Aalborg East, Denmark School of Chemical Science and Engineering, Royal Institute of Technology, Sweden c School of Sustainable Development of Society and Technology, Mälardalen University, Sweden b

h i g h l i g h t s
The fundamentals underpinning oxy-fuel combustion development thoroughly reviewed.
Oxy-fuel induced changes in combustion physics, chemistry and modeling explained.
Generic modeling strategies for PF oxy-fuel combustion successfully proposed.
Oxy-fuel based power generation and CCS systems and the key issues discussed.
Research needs in oxy-fuel combustion fundamentals and their modeling identified.

a r t i c l e

i n f o

Article history: Received 13 July 2015 Received in revised form 10 October 2015 Accepted 22 October 2015

Keywords: Oxy-fuel combustion Combustion chemistry CFD Radiation System performance Carbon capture and storage

a b s t r a c t Oxy-fuel combustion of pulverized fuels (PF), as a promising technology for CO2 capture from power plants, has gained a lot of concerns and also advanced considerable research, development and demonstration in the past years worldwide. The use of CO2 or the mixture of CO2 and H2O vapor as the diluent in oxy-fuel combustion, instead of N2 in conventional air–fuel combustion, induces significant changes to the combustion fundamentals, because of the great differences in the physical properties and chemical effects of the different diluents. Therefore, some fundamental issues and technological challenges need to be properly addressed to develop oxy-fuel combustion into an enabled technology. Computational Fluid Dynamics (CFD) modeling, which has been proven to be a very useful and cost-effective tool in research and development of conventional air–fuel combustion, is expected to play a similarly vital role in future development of oxy-fuel combustion technology. The paper presents a state-of-the-art review and an in-depth discussion of PF oxy-fuel combustion fundamentals and their modeling, which underpin the development of this promising technology. The focus is placed on the key issues in combustion physics (e.g., turbulent gas–solid flow, heat and mass transfer) and combustion chemistry (e.g., pyrolysis, gas phase combustion and char reactions), mainly on how they are affected in oxy-fuel conditions and how they are modeled and implemented into CFD simulations. The system performance of PF oxy-fuel combustion is also reviewed. Finally, the current status of PF oxy-fuel combustion fundamentals and modeling is concluded and the research needs in these regards are suggested. Ó 2015 Elsevier Ltd. All rights reserved.

Contents 1. 2.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Combustion physics in PF oxy-fuel firing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. Turbulent gas-particle multiphase flow. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.1. Pneumatic transport of PF particles to furnace. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.2. Turbulent gas flow in furnace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.3. PF particles motion in furnace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

⇑ Corresponding author. Tel.: +45 30622577; fax: +45 98151411. E-mail address: [email protected] (C. Yin). http://dx.doi.org/10.1016/j.apenergy.2015.10.149 0306-2619/Ó 2015 Elsevier Ltd. All rights reserved.

744 744 745 745 745 746

C. Yin, J. Yan / Applied Energy 162 (2016) 742–762

743

Nomenclature A

L Lc ma mp mp;0 mv ðtÞ MWN MWNO ni Nu P Patm Ppa Pr qr R Re Rep Ru s Sc Sc Sh SNO t T Tg T p;i u umin

frequency factor in rate coefficient in Arrhenius form (s1) particle BET surface area (m2/kg) particle projected area (m2) projected area of group i particles (m2) particle surface area (m2) emissivity gas temperature polynomial coefficients in WSGGM (–) concentration of PF particles (kg/m3) drag coefficient (–) specific heat (J/(kg K)) particle size (m) conversion rate of particle in different sub-processes (kg/s) pipe diameter (m) mass diffusivity (m2/s) activation energy in rate coefficient in Arrhenius form (J/kmol) scattering factor of group i particles (–) initial moisture fraction (–) gravitational acceleration (m/s2) convective mass transfer coefficients (m/s) convective heat transfer coefficients (W/(m2 K)) radiative intensity at position ~ r in direction ^s (W/ (m2 sr)) kinetic rate (s1) gas thermal conductivity (W/(m K)) absorption coefficient of i-th gray gas in WSGGM (1/ (atm m)) domain-based beam length (m) characteristic length (m) particle ash content (kg) particle mass (kg) initial particle mass at injection (kg) mass of volatile yield up to time t (kg) molecular weight of N (kg/kmol) molecular weight of NO (kg/kmol) number density of group i particles (1/m3) Nusselt number (–) sum of partial pressures of the participating gases (atm) local gas pressure (atm) local gas pressure (Pa) Prandtl number (–) radiative flux (W/m2) conversion rate (s1) Reynolds number (–) particle Reynolds number, Rep ¼ qg ju  v jdp =lg (–) universal gas constant (8315) (J/(kmol K)) path length (m) char burnout rate (kg/s) Schmidt number (–) Sherwood number (–) NO source term (kg/(m3 s)) time (s) temperature (K) local gas temperature (K) temperature of group i particles (K) gas velocity (m/s) the minimum (or saltation) velocity (m/s)

UC

mass of char in particle fraction of unburnt char, U C ¼ mass (–) of char in feed particle

ABET Ap Ap;i Aps be;i;j cs CD Cp dp

dmp dt

D Dg E f p;i f w;0 g; g hM hT Ið~ r; ^sÞ k kg ki

U VM;C

v

V Vp X X NO Y N;char

fraction of unburnt combustibles, mass of volatiles and char in particle U VM;C ¼ mass (–) of volatiles and char in feed particle particle velocity (m/s) cell volume (m3) particle volume (m3) mole fraction (–) mole fraction of NO (–) mass fraction of nitrogen in char (–)

Greek letters a local gas absorption coefficient (m1) a1 ; a2 two yield factors (–) ap equivalent particle absorption coefficient (m1) DH heat effects (J/kg) e total emissivity of local gas mixture (–) ep particle emissivity (–) ep;i emissivity of group i particles (–) g conversion factor (–) radiation temperature (K) hR lg air or gas dynamic viscosity (kg/(m s)) lt turbulent viscosity (kg/(m s)) qg air or gas density (kg/m3) qp particle density (kg/m3) r Stefan–Boltzmann constant (5.67  108) (W/(m2 K4)) rp equivalent particle scattering coefficient (m1) 2 sv particle momentum response time, sv ¼ qp dp =ð18lg Þ (m) / phase function (–) X; X0 solid angle (sr) Abbreviations CCS carbon capture and storage CFD Computational Fluid Dynamics CPD Chemical Percolation Devolatilization CTF combustion test facility DO discrete ordinates (radiation model) DTF drop tube furnace DTR drop tube reactor DTRM discrete transfer radiation model EBU eddy-breakup ED Eddy Dissipation EDC Eddy Dissipation Concept EFR entrained flow reactor EWBM exponential wide band model FG-DVC Functional Group – Depolymerisation Vaporisation Cross-linking FR/ED Finite Rate/Eddy Dissipation FSK full spectrum k-distribution JL 4-step Jones and Lindstedt 4-step LES large eddy simulation PF pulverized fuel RANS Reynolds-Averaged Navier–Stokes RFG recycled flue gas RTE radiative transfer equation TGA thermogravimetric analysis UDF user-defined function VM volatile matters WD 2-step Westbrook and Dryer 2-step WSGGM weighted sum of gray gases model

744

C. Yin, J. Yan / Applied Energy 162 (2016) 742–762

2.2.

3.

4.

5.

6.

Heat and mass transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.1. Convective heat and mass transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.2. Radiative heat transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.3. Heating of PF particles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Combustion chemistry in PF oxy-fuel firing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Ignition characteristics of PF particles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Pyrolysis/devolatilization of PF particles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. Gas-phase combustion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4. Char oxidation and gasification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5. NOx formation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . CFD modeling of PF oxy-fuel combustion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. General modeling routine for PF combustion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2. Modeling of PF oxy-fuel combustion: A handy overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3. Key sub-processes and modeling recommendations for oxy-fuel combustion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . System performance and economic analysis of PF oxy-fuel combustion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1. System integration of PF oxy-fuel combustion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2. Commercialization of oxy-fuel combustion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3. Costs of CO2 avoided by oxy-fuel combustion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions and outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1. Introduction Oxy-fuel combustion is seen as one of the most promising technologies for capturing CO2 from thermal power industry. In oxy-fuel combustion, the fuel is burnt in a mixture of oxygen and recycled flue gas (mainly CO2 and H2O), instead of air, to yield a CO2-rich flue gas stream, which is ready for sequestration after purification and compression. The use of the recycled flue gas together with an overall boiler inlet O2 molar fraction of 27–35% is to moderate the furnace temperatures with attempt to obtain similar adiabatic flame temperature, heat transfer profiles and boiler designs compared to the existing air–fuel PF units. The exact inlet O2 concentration depends on the fuel properties and the boiler design. Oxy-fuel combustion has gained drastically increasing concerns in the past two decades. A lot of research and development have been made and oxy-fuel combustion technology has evolved from pilot-scale to some planned demonstration scale projects, as comprehensively reviewed in [1–7]. This review is focused on the combustion fundamentals and their modeling strategies which underpin oxy-fuel combustion technology development. The combustion triangle for a general combustion process is shown in Fig. 1. Among others (e.g., materials), efforts on the combustion fundamentals (including various issues in combustion physics and combustion chemistry) must be made and integrated in the design, operation and optimization in order to achieve a successful combustion process. For conventional air–fuel combustion, Computational Fluid Dynamics (CFD) modeling, which integrates all the understandings and achievements in combustion fundamentals in the best possible way, has been proven to be a powerful and cost-efficient tool in the design and optimization, as reiterated in [8]. For instance, new combustion systems or design are often conceptually developed using CFD, followed by lab and site testing and adjustments. As a result, CFD modeling is expected to play a similarly important role in the future development of oxy-fuel combustion technology and systems. In this review, the key combustion fundamentals in PF-firing are presented first, somewhat in the order of the sub-processes that take place during PF combustion in an oxy-fuel environment, i.e., from PF particle transport, particle motion and interaction with the turbulent gas flow in a furnace, particle heating subjected to external radiative and convective heat transfer, particle conversion with a series of chemical reactions (i.e., pyrolysis/devolatilization, gas phase combustion, char combustion), to pollutants formation

746 746 747 748 749 749 749 750 751 752 754 754 755 756 757 757 758 758 758 759

and emissions. These key issues and the related sub-models which make the backbone of a CFD analysis of a PF-fired boiler are described in sufficient details so that they can serve as guidelines for CFD application. The differences induced by oxy-fuel combustion are also elaborated. Then, the CFD modeling routine is outlined and a handy overview of the modeling efforts on PF oxy-fuel combustion is given, highlighting how the key issues are addressed in the literature. After that, the system performance of PF oxy-fuel combustion is reviewed and discussed. Finally, the research needs in the combustion fundamentals and their modeling are identified.

2. Combustion physics in PF oxy-fuel firing The different physical properties (as shown in Table 1) and chemical effects of the diluents in oxy-fuel combustion induce substantial changes to the combustion fundamentals and the relevant modeling strategies, compared to conventional air–fuel combustion. They are reviewed and elaborated throughout this paper.

Design, Operation, Optimization of combustors

A successful combustion process: Higher combustion efficiency Maximized energy recovery Lower emissions Better reliability & availability

(1) Development of new physical/process models (2) Integrating models in CFD & System simulation (3) Materials; Manufacturing; and so on

Combustion Physics

• Fluid mechanics (turbulence) • Multiphase flows • Heat and mass transfer • Thermodynamics • Advanced testing, and so on

Combustion Chemistry

• Pyrolysis/devolatilization • Homogeneous reactions • Heterogeneous reactions • Pollutant formation • Ash behavior, and so on

Fig. 1. Combustion triangle in a general combustion process.

745

C. Yin, J. Yan / Applied Energy 162 (2016) 742–762 Table 1 Physical properties of CO2 and N2 at 1 atm pressure and 1273 K [9,10]. Physical property

CO2

N2

Ratio of CO2/N2

Density, q (kg m3) Specific heat capacity, Cp (J kg1 K1) Dynamic viscosity, l (kg m1 s1) Kinematic viscosity, m (m2 s1) Thermal conductivity, k (W m1 K1) Thermal diffusivity, a (m2 s1) Prandtl number, Pr Emissivity and absorptivity (radiation heat transfer) Mass diffusivity of O2 in X, DO2,X (m2 s1)

0.4213 1292

0.2681 1213

1.57 1.07

4.898  105

4.594  105

1.07

1.162  104 0.08491

1.713  104 0.07938

0.68 1.07

1.560  104 0.7455 >0 (participating) 2.133  104

2.440  104 0.7022 0 (nonparticipating) 2.778  104

0.64 1.06 –

mixed in premixed systems to increase the temperature of the reactants above the inner-layer temperature. Oxy-fuel PF combustion further complicates the scenario. The main diluent in oxy-fuel combustion, CO2, has different physical properties from the main diluent in air–fuel combustion, N2, which affect the flow, heat and mass transfer, and thermodynamics. Moreover, CO2 is not as chemically inert in combustion as N2. CO2 can influence homogeneous and heterogeneous reactions via different mechanisms. 2.1.1. Pneumatic transport of PF particles to furnace In pneumatic transport of PF particles, the quantity and temperature of primary air varies, depending on the type of pulverizer, grinding rate, and fuel properties. The minimum primary air velocity allowed in the burner pipes is about 15 m/s, at which the PF particles can be entrained in the primary air system [8]. The minimum (or saltation) velocity, umin , varies with air density and viscosity, particle loading and particle size, as seen from the corre0:75 qffiffiffiffiffiffiffiffi  0:25  D=dp qp =qg [11] and lation umin = gdp ¼ 0:0428Re0:175 p

0.77

2.1. Turbulent gas-particle multiphase flow Fig. 2(a) shows various combustion analysis tools of different complexities, among which stoichiometry, equilibrium and kinetics are applicable to different ideal cases. In real life, industrial PF combustion involves many complex physical phenomena. The PF particles are pneumatically transported by heated primary air via a fuel distribution pipework to the burners and then to the furnace. In the furnace, the PF particles travel through and interact with the gas, get heated up rapidly and undergo a series of conversion processes, as shown in Fig. 2(b), until they reach the outlet. For industrial PF-fired boilers, conflicting requirements are often needed, e.g., rapid mixing and short combustion times versus proper flame stabilization, and high burnout versus low emissions. These are usually achieved by complicated turbulent flow patterns, e.g., swirling flow, recirculation zones, vortex breakdown and precessing vortex core. As a result, turbulent mixing often plays a crucial role in industrial PF furnaces. Combustion occurs only when fuel and oxidizer are well mixed in non-premixed systems, or only when the cold reactants and hot combustion products are well

particle momentum response time sv ¼ qp dp =ð18lg Þ. Please refer to the nomenclature for all the symbols. For the relatively narrow range of density, viscosity and particle size distribution in burner pipes, the minimum velocity limit is adjusted linearly based on particle loading, e.g., from about 15 m/s at 0.5 kg fuel per kg air to about 21.5 m/s at 1.0 kg fuel per kg air [8]. In oxy-fuel PF boilers, recycled flue gas (primarily CO2), instead of primary air, is used as the carrier phase for PF transport. Because of the different density and viscosity of CO2, especially the much higher density (in comparison with N2, as shown in Table 1), the conveying velocity in oxy-fuel combustion is expected to be smaller. More experimental studies are needed to generate useful guidelines for pneumatic transport of PF particles in the fuel distribution pipework in oxy-firing plants. 2

2.1.2. Turbulent gas flow in furnace PF-fired boilers in power plants are often characterized by very complicated turbulent flows, to enhance mixing between fuel and

Flue gas

Fuel & Oxidizer

(b) Conversion processes of a fuel particle Ash

in

out P, T, time

in in

P, T

out

out

Stoichiometry (limited by a reactant)

Equilibrium (output described by equilibrium)

Kinetics (perfectly mixed, but no enough time for reactions)

Turbulent Mixing (fuel/oxidizer not well-mixed or cold reactants/hot products not well-mixed)

(a) Combustion analysis tools of increased complexity Fig. 2. (a) Combustion analysis tools and (b) conversion of a single PF particle.

746

C. Yin, J. Yan / Applied Energy 162 (2016) 742–762

oxidizer and/or mixing between the cold reactants and hot products. In turbulent reacting flows, there are a lot of rotating structures (or eddies) of different length, velocity and time scales. For the large eddies whose length scales are comparable to the size of the flow geometry, their motion is very dependent on the flow geometry and highly anisotropic. The large eddies interact with the mean flow and extract energy from the mean flow via vortex stretching. Due to their high Reynolds numbers, the large eddies are unstable and break up, transferring their energy to somewhat smaller eddies. The smaller eddies undergo a similar break-up process and transfer their energy to yet smaller eddies. Such a process continues until the eddy Reynolds number is sufficiently small that the eddy motion becomes stable. On the level of the smallest eddies, molecular viscosity is effective in dissipating the kinetic energy into heat. Compared to the large eddies whose motion is bounded to and dependent on the flow geometry, the statistics of the motion of the small eddies are not dependent on the mean flow or flow geometry and can be assumed to have a universal form. Different eddies are rich in different species, e.g., fuel or oxidizer, reactants or products. The reactions only occur when the fuelrich eddies and oxidizer-rich eddies (or when the reactants-rich eddies and products-rich eddies) are mixed at the smallest eddy level. From the modeling point of view, the Reynolds-Averaged Navier–Stokes (RANS) method has been widely used for industrial PF combustion so far, mainly because it is robust and computationally cheap and still has reasonable accuracy for a wide range of industrial flows. In RANS, the transport equations for the mean flow quantities are solved and all scales of turbulence are modeled. The time step size for transient solution is determined by global unsteadiness. In the last decade, large eddy simulation (LES) of PF combustion has received great interest from academia [12,13]. The interest spreads now to industries, since LES is a good alternative to the two extreme modeling tools, i.e., RANS and the direct numerical simulation. The latter is computationally inhibitive for turbulent reacting flows in industrial PF boilers. Being inherently an unsteady formation, LES computational cost is typically 100 times larger than RANS. In LES, the turbulent fields are separated into large-scale resolved and small-scale unresolved contributions. LES offers two significant advantages over RANS: (1) the large-scale motion of the turbulence, which contains most of the turbulent kinetic energy and controls the dynamics of the turbulence while depends very much on the flow geometry, is computed directly with no need for modeling; (2) the small-scale sub-grid motion of turbulence, which is more or less universal and independent of the flow geometry, can be modeled by dynamic models where the model coefficients are determined as a part of the solution with no need for a priori model constants [12,13]. In LES, the time step size is set by small eddies. Oxy-fuel firing does not make direct difference with air–fuel combustion in terms of turbulence modeling. Though the interest of using LES in PF combustion modeling has spread to industries, RANS is still the main numerical tool in the foreseeing future for industrial PF combustion processes. There are different groups of turbulence models under the RANS framework. The realizable k–e and the SST k–x models are expected to and also have been proven to outperform others in CFD simulations of typical PF-fired furnaces which are characterized by complex turbulent flow patterns (e.g., swirling flow, recirculation zones).

emissions. For instance, an increase in nitrogen conversion to NO is observed when co-firing sawdust with pulverized coal under both air–fuel and oxy-fuel conditions although the sawdust has less fuel-bound nitrogen [14]. It is attributed to that the large biomass particles are not entrained in the near-burner region while break through the flame envelope. Under oxy-fuel combustion conditions, particle breakthrough occurs at smaller diameters, leading to an increased nitrogen conversion to NO when compared to air-firing conditions. For small heavy particles in dilute two-phase flows (which is the case for suspension-firing of pulverized coals), it is sufficient to only retain drag and gravity forces in the equation of motion of particle [15,16].

2.1.3. PF particles motion in furnace The motion pattern of fuel particles in a furnace plays a very important role in air supply and combustion performance, from ignition and flame stability and finally to burnout and pollutant

Nu  hT Lc =kg ¼ 2:0 þ 0:6Re1=2 Pr 1=3

ð2Þ

Sh  hM Lc =Dg ¼ 2:0 þ 0:6Re1=2 Sc1=3

ð3Þ

mp

dv 1 ¼ C D qg Ap ju  v jðu  v Þ þ ðqp  qg ÞV p g dt 2

ð1Þ

All the symbols, including abbreviations, are explained in the nomenclature. For spherical particles, the drag coefficient C D is a function of the particle Reynolds number, Rep ¼ qg ju  v jdp =lg . For non-spherical PF particles, C D also depends on the particle shape factor. When switching from air–fuel to oxy-fuel combustion, the motion pattern of particles will be affected due to the changes in the gas density and viscosity. The model is naturally adjusted to accommodate the changes in the gas properties to update the particle trajectory. The use of gas viscosity lg has to be highlighted here. In turbulent flows of high Reynolds numbers, the turbulent viscosity (or eddy viscosity) largely overwhelms the fluid molecular viscosity. As a result, a constant gas viscosity is commonly used in turbulent combustion modeling, which has negligible direct impact on the gas flow. However, the use of a constant gas viscosity may result in different particle trajectories in a furnace, since the gas viscosity affects the drag coefficient C D via the particle Reynolds number Rep and then affects the particle motion pattern, as seen in Eq. (1). Therefore, the temperature- and compositiondependence needs to be taken into account in the calculation of gas viscosity. In turbulent flows as encountered in industrial PF furnaces, there is interaction between turbulent eddies and immersed small particles. For the same particles released from the same location and at the same initial conditions, they are observed to have different trajectories, i.e., dispersive particle trajectories induced by fluid flow turbulence. From modeling point of view, the mean gas velocity u will be used in the equation of motion of particles to calculate their trajectories. The turbulent dispersion of particles can be taken into account by using stochastic tracking or cloud tracking models. 2.2. Heat and mass transfer The total heat transfer in a boiler includes radiative and convective heat transfer. The former plays the dominant role in the furnace while the latter contributes most to the convective passes. The differences in the physical properties of CO2 and N2, as summarized in Table 1, lead to different heat and mass transfer characteristics between oxy-fuel and air–fuel combustion. 2.2.1. Convective heat and mass transfer The convective heat and mass transfer coefficients, hT and hM , are often evaluated from empirical models. For instance, the widely used Ranz–Marshall model [17] reads,

747

(lines) case 1

(dots) case 2

1

0.1

0.01

(a)

0.001 6

12

18

24

30

36

42

48

Particle/gas emission [w/m3]

Radiation coefficient [m-1]

C. Yin, J. Yan / Applied Energy 162 (2016) 742–762 Particle emission

100000

10000 Gas emission

1000

6

Height along the furnace centerline, Z [m]

(lines) case 1

(b)

100

12

18

24

(dots) case 2

30

36

42

48

Height along the furnace centerline, Z [m]

Fig. 3. (a) Radiation properties and (b) gas/particle emission along the vertical centerline in the PF furnace [38].

When calculating the convective heat and mass transfer between the gas flow and PF particles, the particle Reynolds number ReP and particle diameter will be used in the above equations. Based on such empirical models and the physical properties of gas species (e.g., CO2, N2, O2 and H2O), the impacts of recycle ratio on the convective heat and mass transfer coefficients under oxy-fuel combustion conditions, hT;oxy =hT;air and hM;oxy =hM;air (normalized by the coefficients in air–fuel conditions), can be readily plotted. For boiler retrofit, it is important to match the flame and total heat transfer characteristics as well as the radiativeto-convective heat transfer ratio for oxy-fuel combustion with the counterparts for air–fuel firing [4]. From modeling point of view, the existing empirical models are naturally adjusted to yield the proper convective heat and mass transfer coefficients under oxy-fuel combustion conditions, based on the local gas concentrations, flow conditions and physical properties of the gas mixture. As a result, there is no special effort in the literature on evaluating or modeling convective heat and mass transfer in oxy-fuel combustion. 2.2.2. Radiative heat transfer Comparatively, radiation heat transfer under oxy-fuel combustion environments gains a lot of concerns in the past years. Radiation is the principal mode of heat transfer in combustion furnaces. For a gray participating medium containing absorbing, emitting and scattering particles, the radiative transfer equation (RTE) is,

X rT 4g rT 4p;i dIð~ r; ^sÞ ¼ a þ e  ða þ ap þ rp ÞIð~ r; ^sÞ p;i ni Ap;i i ds p pffl} |fflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl} |fflffl{zfflffl} |fflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflffl gas emission

particle emission

Z

rp þ Ið~ r; ^s0 Þ/ð^s0 ) ^sÞdX0 4p 4p |fflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl}

absorption=scattering losses

ð4Þ

inscattering gain

where the gas absorption coefficient a, and the particle absorption and scattering coefficients, ap and rp , are calculated as follows [18,19],

a ¼ ð1=LÞ  lnð1  eÞ ap ¼

X

ep;i ni Ap;i

ð5Þ ð6Þ

i

rp ¼

X ð1  f p;i Þð1  ep;i Þni Ap;i

e¼

J I X X ae;i ðT g Þð1  eki PL Þ where ae;i ¼ be;i;j T j1 g i¼0

ði ¼ 1; . . . ; IÞ and ae;0

j¼1 I X ¼1 ae;i

ð8Þ

i¼1

Oxy-fuel combustion promotes radiative heat transfer, as a result of the high levels of CO2 and H2O as well as the different CO2/H2O ratio compared to air combustion [2,21–24]. Different operations (e.g., flue gas recycle options) also have clear impacts on gaseous radiative properties and affect the boiler design and boiler efficiency [25]. Various efforts have been made to refine gaseous radiative property models for CFD modeling of oxy-fuel combustion [24,26–30]. Using the exponential wide band model (EWBM) as the reference model, Yin et al. [24] proposed a new four-gray-gas oxy-fuel WSGGM, which is applicable to temperatures in the range of 500–3000 K and pressure path-length of 0.001–60 atm m and also appropriately accounts for variations in H2O vapor and CO2 concentrations in a flame by deriving model parameters for 7 representative molar ratios of H2O vapor to CO2 in the range of 0.125–4. Johansson et al. [28] proposed four-graygas oxy-fuel WSGGM correlations for temperatures in the range of 500–2500 K, pressure path-lengths of 0.01–60 bar m, and molar ratios of H2O vapor to CO2 between 0.125 and 2, by using the statistical narrow band model with the EM2C database as the reference model. Kangwanpongpan et al. [29] proposed new correlations for oxy-fuel WSGGM by fitting the emittance charts calculated from the up-to-date HITEMP 2010 database for temperatures in the range of 400–2500 K, pressure path-length of 0.001–60 bar m and molar ratios of H2O vapor to CO2 between 0.125 and 5. Among the refined oxy-fuel WSGGMs, variations in H2O vapor and CO2 concentrations in a flame are better considered in [28,29], in which the discontinuity in the H2O/CO2 molar ratio in the model correlations is eliminated. The efforts made to refine the gaseous radiative property models are well deserved. Gas phase combustion is always a key process in furnaces burning any kind of fuels (including solid and liquid fuels). The refined models always have the inherent potentials to improve simulation results, particularly for gaseous fuel-fired furnaces under oxy-fuel conditions [24,31,32]. In CFD modeling of oxy-fuel combustion of gaseous fuels, gray and non-gray calculations of a same WSGGM also make remarkable

ð7Þ

i

The total gas emissivity of local gas mixture, e, is often calculated by a weighted sum of gray gases model (WSGGM) in combustion CFD modeling. The WSGGM, originally proposed by Hottel and Sarofim [20] for calculation of the total gas emissivity as a weighted sum of I gray gases and one clear gas, is as follows,

Original population kb Reactive of labile bridges in (heating) bridges the coal lattice

kδ

kc

Side chains

kg

Light gas (g1)

Char bridge + Light gas (g2)

Fig. 4. Coal bridges with different decomposition paths indicated.

748

C. Yin, J. Yan / Applied Energy 162 (2016) 742–762

the literature, e.g., [41–46]. Here the constant values are ep ¼ 0:9 and f p ¼ 0:6. In case 2, the conversion-dependent particle emissivity [47] and scattering factor [38] are used as follows,

ep ¼ 0:4  U C þ 0:6

!

Z rT 4g X rT 4 r  qr ¼ 4p a þ ep;i ni Ap;i p;i þ ða þ ap Þ Ið~ r; ^sÞdX p p X¼4p i ð11Þ 2.2.3. Heating of PF particles When PF particles travel through gas and interact with gas in a furnace, they are rapidly heated up mainly by radiation and undergo a series of conversion processes, as shown in Fig. 2(b). Free moisture in the PF particles is evaporated quickly, followed by pyrolysis or devolatilization in which pyrolysis gases or volatiles are released from the fuel particles. The volatiles are burned out in a turbulent flame, normally stabilized by hot recirculating combustion products. Once the pyrolysis is nearly finished, char combustion dominates the heat release from the particle, normally at a lower heat release rate. During these processes, the particle temperature is updated as follows,

mp C p

100

90

90

N22 N CO CO22

70 60

dT p dmp DH ¼ hT Aps ðT g  T p Þ þ ep Aps rðh4R  T 4p Þ þ dt dt

80

N22 N CO2 CO 2

70 60

Semi-anthracite HVN coal

(a) 50 300

500

700

900

Temperature (K)

1100

ð10Þ

In effect, the particle emissivity ep varies from 1.0 for unburned coal to 0.6 for residual ash. The particle scattering factor changes from 0.9 for unburnt coal (yielding a lower particle scattering coefficient) to 0.6 for residual ash particles (corresponding to a higher particle scattering coefficient), according to Eq. (7). Fig. 3 indicates that particle radiation greatly overwhelms gas radiation in PF combustion. As a result, refining or even using the most accurate gaseous radiative property models is not expected to largely improve the simulation results for PF boilers. On the other hand, neglecting particle radiation in PF combustion modeling is not appropriate. Experimental efforts are needed to develop more accurate and generic models for particle radiative properties as a function of particle conversion degree. The non-gray effects of particle radiation in PF flames as studied in [48] also need to be further investigated. All these hold for both air–fuel and oxy-fuel PFfiring, even though gas radiation in the latter is somehow stronger than that in the former. The heat sources or sinks due to radiation, which can be directly substituted into the energy transport equation, are calculated as follows,

100

80

ð9Þ

f p ¼ 0:9U VM;C þ 0:6ð1  U VM;C Þ

Mass loss (%)

Mass loss (%)

difference in the results [32,33]. Wang et al. [34] conduct a numerical study of oxy-propane combustion, in which detailed hydrocarbon oxidation, soot, and NOx chemistry is employed and two radiation models are implemented for comparison. One accounts for nongray-gas properties and the other does not. It is found that effects of nongray gas-phase radiation are important even in the presence of strong gray soot radiation and must be included to capture the correct distribution of radiative heat loss. Therefore, non-gray implementation of an appropriate oxy-fuel WSGGM is recommended in CFD modeling of oxy-fuel combustion of gaseous fuels. In PF furnaces, it is, however, not appropriate to overemphasize gas radiation while play down or even neglect the role of particle radiation. For instance, the impact of PF particles on the radiation model is improperly disregarded in CFD modeling of biomass/ lignite co-combustion (10% biomass based on thermal input) in a 330 MWe PF furnace by assuming its impact is negligible, while a domain-based WSGGM is used for the absorption coefficient of the gas phase [35]. To gain an impression of the importance of particle radiation in PF flames, Johansson et al. [36] model gas and particle radiation in an axisymmetric cross section of a cylindrical furnace. The particle radiative properties are calculated according to the Mie-theory, accounting for the spectral properties, while the gas radiation is addressed by using a statistical narrow band model as reference. The properties of the combustion gas and the particle load are derived from the measurements in a lignite flame in a 100 kW test rig. The results show that particle radiation dominates the total radiation. The large impact of particles means that differences in the gas composition only have a small effect on the total radiation. Radiative heat transfer in oxy-fuel combustion of pulverized coal is only slightly different from air–fuel combustion, as long as the temperature and particle load are the same. Bäckström et al. [37] perform measurement in the hightemperature zone of a 77 kWth swirling lignite flame and find that the particles are the dominating source of radiation in the crosssection of the flame investigated. Yin [38] evaluates the relative importance of gas radiation and particle radiation via CFD modeling of a 609 MWe conventional PF boiler and demonstrates that particle radiation largely overwhelms gas radiation in PF combustion. Different WSGGMs (i.e., the Smith et al. WSGGM [39] vs. a refined version [40]) for gaseous radiative properties are found to yield almost the same results. On the contrary, slight refinement in particle radiative property models makes much more pronounced difference in the CFD results. Fig. 3 shows the radiation properties and gas and particle emissions along the vertical centerline in the furnace for two computational cases, between which the only difference lies in the particle radiation property. In case 1, constant particle emissivity and scattering factor are employed, as commonly used in PF combustion modeling in

SAB High-volatile bituminous coal

(b) 50 300

500

700

900

1100

Temperature (K)

Fig. 5. Mass loss rate of two coals during devolatilization tests in N2 and CO2 atmospheres (thermogravimetric apparatus, heating rate of 15 K/min) [44].

ð12Þ

749

C. Yin, J. Yan / Applied Energy 162 (2016) 742–762

The convective heat transfer coefficient hT can be evaluated by empirical models, e.g., Eq. (2). The Stefan flow effect can also be taken into account in the calculation of the convective heat transfer coefficient. 3. Combustion chemistry in PF oxy-fuel firing

the Functional Group – Depolymerisation Vaporisation Crosslinking (FG-DVC) model, as briefly introduced below. The first is the single kinetic rate model, assuming the rate of devolatilization is first-order dependent on the amount of volatiles remaining in the particle [76]. The devolatilization kinetic rate, k, is defined by an Arrhenius type pre-exponential factor A and an activation energy E.

3.1. Ignition characteristics of PF particles

k ¼ AeE=ðRu T p Þ

Ignition of PF particles is an important issue in PF combustion and has been widely studied in the literature, e.g., ignition mode, ignition temperature and ignition delay, initially for air–fuel combustion [49–53] and recently for oxy-fuel conditions [54–73]. In general, the primary ignition mode of PF particles can be homogeneous ignition, heterogeneous ignition, or joint hetero-homogeneous ignition. For ‘‘large” fuel particles (typically >100 lm in diameter) under ‘‘slow” heating condition (<100 K/s), the primary ignition mode tends to be homogeneous ignition, characterized by prior pyrolysis and subsequent ignition of the volatiles. The homogeneous ignition is a two-stage process: the first stage is the initial ignition of the volatiles, which forms a circumambient flame and prevents char reactions by effectively screening the solid from access by O2; the second stage is ignition of the char, occurring as pyrolysis terminates. For ‘‘small” particles under ‘‘quick” heating, heterogeneous ignition is more likely the primary ignition mode, by direct O2 attack on the fuel particle. The heterogeneous ignition is generally believed to involve three stages: the first stage is direct attack of the reactant gas on the whole solid fuel particle (not just char) and the heterogeneous reactions remove materials that would otherwise be expelled as volatiles; the second stage is volatile ignition when pyrolysis becomes substantial and the released volatiles quench the initial heterogeneous reactions; the third stage is re-ignition of the char at the end of pyrolysis. At sufficiently high heating rates (>1000 K/s), the joint heterohomogeneous ignition may dominate at all particle sizes [74]. Ignition of PF particles is inherently a resultant phenomenon of various combustion physics issues (e.g., fluid mechanics, heat and mass transfer) and combustion chemistry issues (e.g., pyrolysis, homogeneous and heterogeneous combustion). Ignition of PF particles depends on many factors, e.g., rank and properties of solid fuels, particle size, local gas temperature and composition, heating rate, and even ignition indicators (e.g., light flash, mass loss rate, particle surface temperature, or evolved gases). It is less likely to draw consistent conclusions on the impacts of the combustion atmospheres alone (i.e., oxy-fuel vs. air–fuel) on the ignition characteristics of PF particles. As a result, ignition is not explicitly listed or discussed as a separate issue in this paper, which can also be seen in Fig. 1. Instead, the focus is placed on the basic issues in combustion physics and combustion chemistry.

The second is the two-competing-rate model [77], assuming that the mass loss during devolatilization is described by the two competing rates which control the devolatilization over different temperature ranges,

3.2. Pyrolysis/devolatilization of PF particles Pyrolysis plays a crucial role in a combustion process. It controls the yields of volatiles, tars and char, determines the split of fuel nitrogen into volatiles and char, and affects the porosity and internal surface areas of the resultant char. As a result, ignition, flame stability, char burnout and pollutant formation are all affected. PF pyrolysis characteristics depend on fuel properties and operation conditions (e.g., heating rate, temperature, pyrolysis atmosphere). A proper description of PF pyrolysis is important in PF combustion modeling. The mechanism of coal pyrolysis and product distribution is reviewed in [75]. There are different pyrolysis models, from the simplest single kinetic rate model, to two competing rates model, and to various network pyrolysis models, e.g., the Chemical Percolation Devolatilization (CPD) model, and

R1 ¼ A1 eE1 =ðRu T p Þ

ð13Þ

and R2 ¼ A2 eE2 =ðRu T p Þ

ð14Þ

The two rates are weighted to yield an expression for the devolatilization as

mv ðtÞ ¼ ð1  f w;0 Þmp;0  ma

Z 0

t

ða1 R1 þ a2 R2 Þ dt Rt expð 0 ðR1 þ R2 ÞdtÞ

ð15Þ

The third is the Chemical Percolation Devolatilization (CPD) model, which does not rely on empirically determined Arrhenius rate relationships. Instead, the CPD model characterizes the devolatilization behavior of rapidly heated coal particles based on the physical and chemical transformations of the parent coal structure [78–80]. The coal structure is considered as a simplified lattice or network of chemical bridges that link the aromatic clusters. Upon heating, the original labile bridges in the coal lattice become the set of reactive bridges. Two competing paths are available, as seen in Fig. 4. In one path, the reactive bridges react to form side chains, which may detach from the aromatic clusters to form light gas. As the bridges between neighboring aromatic clusters are cleaved, the clusters become detached from the coal lattice and form heavy-molecular-weight tar precursors, which can vaporize to form tar or reattach to the coal matrix (cross-linking). In the other path, the bridges react and become a char bridge with release of light gases. The necessary input parameters for the CPD model include the initial fraction of bridges in the coal lattice, initial fraction of char bridges, lattice coordination number, cluster molecular weight, side chain molecular weight, and various rate expression constants. The last is the Functional Group – Depolymerisation Vaporisation Cross-linking (FG-DVC) model [81]. Similar to the CPD model, this model uses a simplified coal structure in which functional groups are formed by tightly bound aromatic clusters. These clusters are connected by weaker aliphatic and ether bridges. During decomposition, depolymerization or cleavage of these bridges occurs, accompanied by release of light gases and large fragments (i.e., tar precursors). It is assumed that the bridges are cleaved stepwise, in which tougher bridges require higher temperatures to break. Experimental evidences show that the kinetics of the decompositions is insensitive to the rank of the parent coal, which makes it possible to propose a general model applicable to almost any coal, i.e., using coal-independent rates for decomposition of individual functional groups in coal and char to produce gas species. The above four models are compared and assessed in the CFD prediction of ignition point of pulverized coal flames under oxyfuel conditions [82]. The two network models yield more accurate results than the single rate and two competing rates models and the best performance is achieved by the FG-DVC model. However, it is not stated how the kinetic parameters (A, E) used in the single rate and the two competing rates models are derived, which is actually very important for such an assessment.

750

C. Yin, J. Yan / Applied Energy 162 (2016) 742–762

The most commonly used model is still the single rate model. The values of Arrhenius parameters (A, E) can be obtained from the literature, estimated by network pyrolysis codes, or determined experimentally. The kinetic parameters (A, E) from the literature may have limited applicability since they are dependent on the fuel properties and pyrolysis conditions. Thermogravimetric analyses (TGA) are also often used to experimentally determine the kinetic parameters (A, E). However, the much higher heating rates in typical PF furnaces (on the order of 105 K/s), compared to those in a thermogravimetric apparatus (on the order of 100 K/ min), most likely yield different kinetic parameters (A, E), as shown in the experimental study [83]. Recently, some network pyrolysis codes are used to estimate the kinetic data for a given fuel under a given pyrolysis condition. For example, the FG-DVC code [84] is used to determine the kinetic parameters (A, E) by assuming the heating rate of 105 K/s and using a proper reactor temperature, when the single kinetic rate pyrolysis model is applied in CFD modeling of pulverized coal combustion [43,44,46,85–87]. When switching from air–fuel to oxy-fuel condition, the pyrolysis/devolatilization process is not affected remarkably [88,89]. Experimental evidences show that there is no difference in volatile release and in the properties and reactivity for chars obtained in N2 and CO2. The differences reported in some literature are believed to be induced by CO2 gasification of coal chars [44,89–92]. For instance, Fig. 5 shows the comparison of coal pyrolysis in N2 and CO2 atmospheres [44]. The mass loss curves in the N2 and CO2 atmospheres are found to follow a very similar trend before pyrolysis is finished at around 1150 K. The extra mass loss in the CO2 atmosphere above 1200 K is mainly due to the char-CO2 reaction. As a result, the same pyrolysis kinetic parameters (A, E) can be used in both air–fuel and oxy-fuel combustion modeling, as done in [44,46,86]. Gil et al. [90] also investigate the combustion kinetics of the coal chars obtained from different pyrolysis atmospheres. The coal chars are obtained by devolatilizing the raw coals in an entrained flow reactor at 1000 °C for 2.5 s in N2 or CO2, and the char reactivity tests are conducted in a same oxy-fuel atmosphere (O2/CO2 = 30/70) by isothermal thermogravimetric analysis in a kinetically controlled regime (400–600 °C). The results show that the chars obtained in CO2 have slightly lower reactivity than the chars obtained in N2, which is attributed to that the chars obtained in CO2 have undergone gasification by CO2 during the char preparation process. 3.3. Gas-phase combustion Gas phase combustion is always an important process in any combustor. For typical coal combustion, a great portion of the fuel conversion occurs in gas phase. The volatiles often carry about 50% of the energy of coals. Combustion of the released volatiles plays a key role in ignition, local stoichiometries, flame stability and pollutant emissions. For biomass combustion, volatile combustion is even more important because most of the energy of biomass fuels is in form of volatiles. As a result, gas-phase reaction mechanism is expected to play an important role in CFD modeling of PF combustion [93]. The majority of CFD modeling of gas phase combustion relies on global combustion mechanisms. Two global mechanisms are widely and successfully used in air–fuel combustion modeling. One is the two-step hydrocarbon oxidation mechanism proposed by Westbrook and Dryer (WD) [94,95]. The WD scheme consists of two reactions, where the second step, oxidation of CO to CO2, is reversible. Here CH4 is used as an example to show the scheme.

CH4 þ 1:5O2 ! CO þ 2H2 O

ðR1Þ

CO þ 0:5O2 $ CO2

ðR2Þ

The other is the four-step mechanism developed by Jones and Lindstedt (JL) [96] for alkane hydrocarbons up to butane in mixtures with air in premixed and diffusion flames. The JL 4-step scheme includes two competing fuel breakdown reactions, (R3) and (R4). The two reversible reactions, (R5) and (R6), control the rate of reaction for CO and H2.

CH4 þ 0:5O2 ! CO þ 2H2

ðR3Þ

CH4 þ H2 O ! CO þ 3H2

ðR4Þ

H2 þ 0:5 O2 $ H2 O

ðR5Þ

CO þ H2 O $ CO2 þ H2

ðR6Þ

Neither of the above global mechanisms has been validated against oxy-fuel experimental data. The high-concentration CO2 in oxy-fuel flames has chemical effects, via homogeneous and/or heterogeneous reactions, yielding higher CO concentration [4,6]. However, the detailed pathways or mechanisms are not yet well confirmed. Simulations using detailed reaction mechanisms show that CO2 competes with O2 for atomic hydrogen and enhances CO formation via the reaction CO2 + H M CO + OH [97–100]. Reactions of CO2 with hydrocarbon fragments, e.g., singlet and triplet CH2, and other radicals such as CH3 and CH, may also contribute to CO formation [97,98]. High-concentration CO2 can impact the reaction OH + H2 M H + H2O via OH radical and result in lower H2 and higher H and H2O concentrations in the flame [100]. H2O vapor in oxy-fuel combustion also has chemical effects: it can promote or inhibit CO oxidation depending on the specific conditions. For instance, the experimental and modeling study of CO oxidation at atmospheric pressure and temperatures in the range of 700– 1800 K in O2 environments from significantly rich to lean and variable CO2 (0–75%) and H2O (0.1–10%) concentrations shows that the presence of H2O vapor enhances CO conversion [99]. This is different from other literature results with higher O2 and H2O level [101] in which an inhibiting effect is observed. Andersen et al. [102] refined the two global mechanisms for oxy-fuel combustion by using a detailed chemical kinetic mechanism as the reference model. In the refined schemes, the initiating reactions involving hydrocarbon and oxygen are retained and the H2–CO–CO2 reactions are modified to improve prediction of the major species concentrations. A comparative CFD analysis of a propane oxy-fuel flame is also made. Comparing to the original versions, the refined WD 2-step mechanism improves the prediction of the temperature field and CO in the post flame zone while the refined JL 4-step mechanism slightly better predicts the CO profile in the flame zone. However, the refined JL 4-step scheme involves [H2]0.75 in the reaction rate of (R5), which could cause numerical instability or difficulty in CFD simulations. In the CFD modeling of an 0.8 MW oxy-natural gas flame, three global mechanisms are compared [31], i.e., the original WD 2-step scheme [94,95], the refined WD 2-step scheme [102], and a newly refined JL 4-step scheme in which the only change is to use the H2 oxidation model of Marinov et al. [103] to replace the reversible H2 oxidation reaction (R5) in the original JL 4-step mechanism. The Eddy Dissipation Concept (EDC) is used for turbulence–chemistry interaction in all the simulations. The CFD results are compared against the experimental data (e.g., gas temperature and species). When applied to oxy-fuel combustion, the original WD 2-step scheme is found to over-predict the flame temperature and also largely under-predict the CO level. Both the refined WD 2-step and modified JL 4-step schemes can reasonably well predict the

C. Yin, J. Yan / Applied Energy 162 (2016) 742–762

Volatile-N Light gas-N Tar-N

HCN NH3

N2 Char surface

Fuel-N Char-N

+ O2

NO

Fig. 6. Fuel-N conversion pathways (adapted from [130]).

relatively high CO level in oxy-fuel combustion, in which the latter also reasonably well predicts H2 level and flame temperature [31]. Recently, Chen and Ghoniem [100] perform CFD simulations of a swirling diffusion flame under air–fuel and oxy-fuel conditions, respectively. Three different modeling strategies are compared: (1) the original WD 2-step mechanism [94,95] together with the Eddy Dissipation model for turbulence–chemistry interaction, (2) the WD 2-step mechanism combined with the EDC for turbulence–chemistry interaction, and (3) the WD quasi-global mechanism (12 species and 22 reactions) [94,95] in combination with the EDC. It is found the original WD 2-step global mechanism cannot reasonably predict the CO concentrations, although its joint use with the EDC shows some improvements due to the consideration of CO2 dissociation reaction. The WD quasi-global mechanism is able to capture the chemical effects of CO2 in oxy-fuel combustion and shows improved performance in both air–fuel and oxy-fuel flame CFD simulations. 3.4. Char oxidation and gasification Burnout of PF particles has been widely studied in the literature. Some pilot-scale and lab-scale studies report a higher char burnout in oxy-fuel combustion [4,5], which is interpreted as the results of the longer particle residence times and higher partial pressures of O2 in the vicinity of the burning particles. Gasification by CO2 and/or H2O is also suggested as a contributor to the improved burnout. However, there still is a lot of scatter in the literature. For instance, drop tube furnace experimental results indicate that the char burnout is higher in the oxy-fuel atmosphere than in air–fuel atmosphere for almost the whole range of O2 concentration studied [2]. The authors attribute it to the enhanced char-CO2 gasification reaction under high CO2 concentration; meanwhile they realize that their experimental observations are different from the literature, most of which report lower char reactivity in oxyfuel conditions due to the lower O2 binary diffusivity in CO2. Hecht et al. [104] simulate the combustion of an isolated pulverized coal char particle under oxy-fuel conditions and the results show that the presence of the CO2 gasification reaction increases the char conversion rate for combustion at low O2 concentrations, but decreases char conversion for combustion at high O2 concentrations. The experimental study of Dhaneswar and Pisupati [105] indicates that whether oxy-fuel combustion produces a higher carbon conversion than air combustion or not depends on the rank of coals and combustion temperatures. Chen et al. [6] review the characteristics of char combustion under oxy-fuel conditions and summarize that char oxidation at low temperatures (in zone I) does not show any difference between O2/N2 and O2/CO2 atmospheres while under diffusion controlled conditions (in zone II and III) char oxidation rate becomes lower in CO2 diluent due to the lower O2 diffusivity in CO2. A recent experimental study of char combustion tests under high heating rates in a laboratory-scale drop-tube reactor indicates that the change from O2/N2 to O2/CO2 atmosphere leads to a slight decrease in the mass loss rate of chars while a more drastic decrease in particle temperature [106].

751

These inconsistent findings may be understood, due to the fact that the impacts of CO2 on char conversion are multi-fold. First, the higher heat capacity of CO2, in comparison to N2, tends to lower flame temperatures and reduce char conversion rate [6,107]. Second, the lower O2 binary diffusivity in CO2 (compared to O2 in N2) reduces the O2 flux in the particle boundary layer and in the particle and thus decreases the burning rate of char particles [4,6,107]. Third, gasification reactions become increasingly important in oxy-fuel conditions, which, however, have opposite impacts on char conversion rate [104,108]. On one hand, CO2 and H2O steam gasification reactions tend to increase the overall char consumption rate. On the other hand, both the gasification reactions are strongly endothermic, which tends to cool the char particle and reduces its oxidation rate. The overall impact of oxy-fuel combustion on char conversion rate also depends on char combustion regime, zone I (kinetics-controlled), zone II (diffusion/kinetics-con trolled), or zone III (diffusion-controlled).

C þ 0:5O2 ! CO þ 111 kJ=ðmole carbonÞ

ðR7Þ

C þ CO2 ! 2CO  172 kJ=ðmole carbonÞ

ðR8Þ

C þ H2 O ! H2 þ CO  131 kJ=ðmole carbonÞ

ðR9Þ

The activation energy of the char gasification reactions is substantially greater than the 160 kJ/mol (i.e., the activation energy of the char oxidation reaction, R7). For steam gasification (R9), the activation energy is about 222 kJ/mol. For CO2 gasification (R8), the effective activation energy is in the range of 165–283 kJ/mol, where the lower values (165–190 kJ/mol) are expected from the data sets in which diffusional resistance begins to compete with kinetic resistance [104]. The measured data in the literature show that the relative rate of CO2 gasification with respect to char oxidation is between 0.1  104 and 3.0  104 at 1073 K. Based on this, a ‘‘best-guess” value of 251 kJ/mol for the activation energy of the CO2 gasification (R8) is used in char particle combustion modeling and the relative rate of CO2 gasification with respect to oxidation at a typical PF char combustion temperature of 2000 K is estimated to be in the range of (0.1–3.0)  102 [104]. Brix et al. [109] study char conversion in an electrically heated entrained flow reactor at temperatures in the range of 1173– 1673 K and inlet O2 concentrations between 5 and 28 vol% (covering zone I–III). No evidence suggesting an effect of CO2 gasification on char conversion is found in their experimental study. However, their modeling does suggest that CO2 gasification can contribute to char consumption when O2 concentration is low, particle temperature is high and combustion takes place in zone III, even though such conditions may not widely exist in a real boiler. The latest studies recommend or suggest it important to include the steam and CO2 gasification reactions when interpreting experimental data of char combustion or when simulating char combustion, especially in oxy-fuel combustion environments, e.g., [105,108,110,111]. For instance, Geier et al. [110] compare the traditional single-film char reaction model which only accounts for C + O2 ? CO, and the extended single-film model which considers all the char oxidation and gasification reactions, C + O2 ? nCO2 + (1  n)CO, C + CO2 ? CO and C + H2O ? H2 + CO, for predictions of oxy-fuel combustion of pulverized coal char particles. The split parameter n is a function of the particle surface temperature and the oxygen partial pressure at the particle surface: n = 0 at sufficiently high temperatures because the predominant product is CO [112]. The extended single-film model is found to outperform the traditional single-film model for oxy-fuel combustion of coal char. However, the best fit values of the kinetic parameters are somehow outside the range of physical meaningfulness, which is speculated to be induced by the uncertainty with the reaction order of 1 and neglecting CO conversion in the boundary layer.

752

C. Yin, J. Yan / Applied Energy 162 (2016) 742–762

Particle Motion Model Forces on various particles; Turbulent dispersion

Pyrolysis Model

Turbulence Model

Mechanism; Kinetics at proper heating rates/temperatures

RANS vs. LES

CFD Solver: Fluid Flow Transport (mass; momentum; energy; species)

Homogeneous Reaction Model

Radiative Heat Transfer Model

Mechanism and kinetics; Chemistryturbulence interaction; Pollutant

Model to solve RTEs; Gas/particle radiative property models

Char Conversion Model Oxidation/gasification; Mass transfer; Kinetics

Fig. 7. CFD modeling of PF combustion: Implication of challenges with oxy-PF combustion modeling.

Gonzalo-Tirado and Jiménez [113] analyze in detail CO oxidation chemistry in the boundary layer around a coal char particle under oxy-fuel conditions and conclude that the CO flame is less intense in oxy-fuel combustion and can be ignored in char conversion modeling. Kühnemuth et al. [114] investigate CO formation under oxy-fuel conditions and find char-CO2 gasification makes a great contribution to the increased CO formation in oxy-lignite flames (for particle diameter of 50 lm and particle temperature above 1000 °C). Hecht et al. [111] compare the single-film model (no homogeneous reaction in the particle boundary layer), double-film model (assuming an infinitely thin flame sheet that converts any CO to CO2 and any H2 to H2O in the boundary layer) and continuousfilm model for combustion of a single, steady-state particle. Each of the models takes into account the reactions of coal char with O2, CO2 and H2O. All the three reactions are required in order to match particle temperatures and carbon consumption rates over a wide range of gas compositions spanning from air–fuel to oxyfuel conditions. The single-film model which considers all the char oxidation and gasification reactions with O2, CO2 and H2O is concluded as an appropriate sub-model to describe pulverized coal char combustion over a wide range of conditions. Another issue about char conversion in oxy-fuel combustion is the effect of Stefan flow. Stefan flow tends to slow down the char reactions a little bit by somehow screening the solid from the direct attack of O2, CO2 and/or H2O steam. Stefan flow also tends to weaken the heat transfer to the particle surface and accelerate the heat loss from the particle, which lowers the particle surface temperature and delays the char reactions. It is concluded that the effects of Stefan flow on the heat and mass transfer from/to a char particle cannot be neglected in oxy-fuel combustion and the newly derived correction factors for the transfer coefficients of reactive gases (e.g., O2, CO2 and H2O) are found to greatly improve the prediction of the particle temperature, conversion rate and burnout time [115,116].

3.5. NOx formation NOx emissions from air–fuel combustion systems can be formed via thermal NOx (by oxidation of N2 in the combustion air at temperatures above 1500 °C), prompt NOx (by hydrocarbon radicals attacking N2 to form cyanide species and then to NO at the flame front), and fuel NOx (by oxidation of nitrogen contained in the fuel). For conventional air–fuel combustion of pulverized coal, the

majority of the total NOx is from fuel NOx and up to 20% is due to thermal NOx while the prompt NOx is negligible [117]. Switching from air–fuel to oxy-firing significantly lowers NOx emissions. First, the low N2 concentrations in oxy-fuel combustion not only inhibit the thermal and prompt NOx formation but facilitate the reduction of NO into N2 via the reverse Zeldovich mechanism [3,4,118]. Second, flue gas recycling favors NOx reduction via re-burning, e.g., NOx reduced to cyanide and amine intermediates via reactions with the hydrocarbon radicals, reduced to N2 via reactions with other N-volatiles or via heterogeneous reactions over char [4,119]. Flue gas recycling is also supposed to result in a reduced conversion of fuel-N to NOx, probably because water vapor inhibits the oxidation of intermediates to form NOx [120]. Third, the high CO2 and CO concentrations in oxy-fuel combustion tend to suppress fuel NO formation. CO2 inhibits NO formation at both stoichiometric and fuel-lean conditions due to the diminished O/ H radical pool, while promotes NO formation from volatile-N under fuel-rich conditions due to the increased OH-concentration [121– 123]. During char combustion, high CO2 suppresses the formation of NO and its precursors from char-N [124] and high CO promotes NOx reduction on char surface [117,125,126]. The overall fuel-N conversion to NO in O2/CO2 is lower than that in O2/N2, as concluded by recent experimental studies [127–129]. Under oxy-fuel conditions, the NOx emissions per unit of energy generated can be reduced to somewhere between one-third and one-half of that from air–fuel combustion [2,4,118]. The conventional primary and secondary measures can be used in oxy-fuel combustion process for NOx emission control. Primary measures (i.e., combustion control technologies) are believed to be sufficient for oxy-fuel combustion but this will depend on future legislation for CO2 emission and storage [3]. If neglecting soot formation and soot-N conversion, the nitrogen in a parent fuel during PF combustion can be released in the pathways as sketched in Fig. 6 [117,130,131]. First, in the primary pyrolysis the volatile-N is released together with the majority of volatiles. The NOx precursors are mainly HCN and NH3 and may also include other minors (e.g., HNCO). Depending on the O/N ratio in the parent fuel, part of the fuel-N could be directly converted to NO during the primary pyrolysis. Then, in the secondary pyrolysis the thermal cracking of volatiles (mainly tar) provides additional sources of HCN and NH3. The NOx precursors (i.e., HCN and NH3) are competitively oxidized and reduced to form NO and N2, respectively, in which the conversion rates are commonly evaluated by De Soete’s scheme [132]:

753

C. Yin, J. Yan / Applied Energy 162 (2016) 742–762 Table 2 A handy overview of CFD modeling efforts on oxy-fuel combustion of pulverized fuels (PF) in the past decade.a Facility

Fuel

Code

Modeling approaches

Reference

The CANMET 300 kWth down-fired swirl burner furnace

Sub-bituminous coal

CFX [19]

[134,135]

The Chalmers 100 kWth down-fired swirl burner furnace

Lignite

AVL Fire [136]

The RWTH Aachen 100 kWth downfired swirl burner furnace

Lignite

FLUENT [19]

The RWTH Aachen 100 kWth downfired swirl burner furnace

Lignite

FLUENT

The RWTH Aachen 100 kWth downfired swirl burner furnace

Lignite

FLUENT

The IHI 1.2 MWth down-fired swirl burner furnace

Sub-bituminous coal

FLUENT

The INCAR-CSIC near-laminar downfired lab-scale entrained flow reactor (EFR)

Anthracite, semianthracite, bituminous coal

FLUENT

The INCAR-CSIC near-laminar downfired lab-scale EFR

Anthracite, semianthracite, bituminous coal

FLUENT

The INCAR-CSIC near-laminar downfired lab-scale EFR

Olive waste, semianthracite, bituminous coal

FLUENT

The RWEn 0.5 MWth combustion test facility (CTF) equipped with an IFRF burner

Bituminous coal

FLUENT

The RWEn 0.5 MWth CTF equipped with an IFRF burner

Bituminous coal

FLUENT

The RWEn 0.5 MWth CTF equipped with an IFRF burner

Bituminous coal

AVL Fire

The RWEn 0.5 MWth CTF equipped with an IFRF burner

Shea meal, bituminous coal

AVL Fire

The E.ON 1 MWth CTF with single wall-fired swirl staged low-NOx burner

Bituminous coal

FLUENT

A drop tube reactor (DTR), and a 585 kWth research boiler

Bituminous coals

FLUENT

A lab-scale DTR

Bituminous coal

FLUENT

A lab-scale DTR

Bituminous coal

FLUENT

A lab-scale DTR

Victorian brown coal (dried)

FLUENT

Standard k–e; Air–fuel radiation model (assumed to be naturally adjusted to oxy-fuel). Pyrolysis: single kinetic rate model; Volatile combustion: VM + O2 ? CO + H2O, CO + O2 ? CO2, EBU; Char reactions: C + O2, kinetics/ diffusion-limited Standard k–e; DTMR (no particle radiation in RTE), gas absorption coefficient a = 0.31 m1, particle emissivity ep = 0.7. Pyrolysis: single kinetic rate model; Volatile combustion: VM + O2 ? CO2 + H2O, EBU; Char reactions: C + O2, kinetics/diffusion-limited k–e; DO, air–fuel WSGGM, local absorption coefficient – the sum of particle and gas ap + a. Pyrolysis: CPD model; Volatile combustion: VM + O2 ? CO + H2, CO + O2 ? CO2, H2 + O2 ? H2O, FR/ED; Char reactions: C + O2 ? CO, C + CO2 ? CO, C + H2O ? CO + H2, kinetics/diffusion-limited LES and RANS (standard k–e, RNG k–e, SST k–x); DO, oxy-fuel WSGGM. Pyrolysis: single kinetic rate model; Volatile combustion: VM + O2 ? CO + H2O, CO + O2 ? CO2, H2 + O2 ? H2O, ED; Char reactions: C + O2 ? CO, C + CO2 ? CO, C + H2O ? CO + H2, kinetics/diffusion-limited Reynolds stress model; DO, gray air–fuel WSGGM vs. two non-gray oxy-fuel WSGGMs. Pyrolysis: CPD model; Volatile combustion: VM + O2 ? CO + H2, CO + O2 M CO2, H2 + O2 ? H2O, FR/ED; Char reactions: C + O2 ? CO, C + CO2 ? CO, C + H2O ? CO + H2, kinetics-limited Standard k–e; P1, air–fuel WSGGM (cell-based). Pyrolysis: single kinetic rate model; Volatile combustion: single-mixture fraction approach; Char reactions: C + O2, kinetics/diffusion-limited RNG k–e; DO, air–fuel WSGGM, soot included, particle radiation (emissivity ep = 0.9, scattering factor fp = 0.6). Pyrolysis: single kinetic rate model (FG-DVC derived kinetics vs. default kinetics); Volatile combustion: mixture fraction/ PDF; Char reactions: C + O2, intrinsic model; NOx: thermal-NO, fuel-NO (FGDVC determined vs. experimentally determined char-N/volatile-N). RNG k–e; DO, oxy-fuel WSGGM (non-gray, domain-based), particle radiation (ep = 0.9, fp = 0.6). Pyrolysis: single rate model (FG-DVC derived kinetics); Volatile combustion: refined JL 4-step scheme for oxy-fuel, EDC; Char reactions: C + O2 ? CO, C + CO2 ? CO, kinetics/diffusion-limited; NOx: thermal- and fuel-NO (FG-DVC determined VM composition/yields and charN/volatile-N) RNG k–e; DO, oxy-fuel WSGGM (non-gray, domain-based), particle radiation (ep = 0.9 and fp = 0.6 for coal; ep = 0.9 and fp = 0.9 for biomass). Pyrolysis: single rate model (kinetics from FG-DVC for coal and from literature for biomass); Volatile combustion: refined JL 4-step scheme for oxy-fuel, EDC; Char reactions: same as [44]. Fuel-NOx: char-N and volatile-N determined experimentally, NH3 as volatile-N intermediate for biomass and HCN and NH3 for coals Realizable k–e; DO, oxy-fuel WSGGM vs. air–fuel WSGGM vs. constant absorption coefficient for gas, scattering coefficient of 0.2 m1 for particle. Pyrolysis: two-competing-rate model; Volatile combustion: VM + O2 ? CO + H2O, CO + O2 ? CO2, FR/ED; Char reactions: C + O2 ? CO, C + CO2 ? CO, kinetics/diffusion-limited Standard k–e vs. LES; DO, gray air–fuel WSGGM vs. spectral oxy-fuel FSK for gas, particle radiation (ep = 0.9, fp = 0.6). Pyrolysis: single kinetic rate model; Volatile combustion: VM + O2 ? CO + H2O, CO + O2 ? CO2, ED; Char reactions: C + O2 ? CO, intrinsic model Standard k–e; DTRM (no particle effect in RTE), an absorption coefficient of 0.31 m1 used for oxy-fuel. Pyrolysis: single kinetic rate model; Volatile combustion: CH4 + O2 ? CO + H2 + H2O, CO + H2O M CO2 + H2, H2 + O2 M H2O, EBU; Char reactions: C + O2 ? CO2, kinetics/diffusion-limited Standard k–e; DTRM (no particle effect in RTE), air–fuel WSGGM. Pyrolysis: single kinetic rate model; Volatile combustion: CH4 + O2 ? CO + H2 + H2O, CO + H2O M CO2 + H2, H2 + O2 M H2O, CO + O2 M CO2, EBU; Char reactions: C + O2?CO, kinetics/diffusion-limited Standard k–e vs. LES; DO, air–fuel WSGGM, particle radiation (ep = 0.8, fp = 0.9). Pyrolysis: single kinetic rate model (kinetics from FG-DVC); Volatile combustion: VM + O2 ? CO + H2O, CO + O2 ? CO2, ED; Char reactions: C + O2, intrinsic model RNG k–e; P1, air–fuel WSGGM (gray, domain-based). Pyrolysis: single kinetic rate model; Volatile combustion: VM + O2 ? CO + H2O, CO + O2 ? CO2, FR/ED; Char reactions: C + O2 ? CO, C + CO2 ? CO, intrinsic model Standard k–e; DO, air–fuel WSGGM. Pyrolysis: single kinetic rate model; Volatile combustion: VM + O2 ? CO2 + H2O, kinetics-limited; Char reactions: C + O2, kinetics/diffusion-limited Standard k–e; DO, air–fuel WSGGM. Pyrolysis: single rate model vs. twocompeting-rate model vs. CPD model vs. FG-DVC model; Volatile combustion: multistep global schemes, FR/ED; Char reactions: C + O2 ? CO, C + CO2 ? CO, C + H2O ? CO + H2, kinetics/diffusion-limited Realizable k–e; DO, oxy-fuel WSGGM. Pyrolysis: single rate model (kinetics from TGA); Volatile combustion: refined WD 2-step scheme for oxy-fuel, FR/

[42]

[137]

[138]

[139]

[58,140]

[43,85,86]

[44]

[46]

[141]

[142]

[143]

[144]

[145]

[146]

[147]

[82]

[45]

(continued on next page)

754

C. Yin, J. Yan / Applied Energy 162 (2016) 742–762

Table 2 (continued) Facility

a

Fuel

Code

A 3 MWth pilot-scale tangentially fired furnace

Victorian brown coal

FLUENT

A 330 MWe tangentially-fired PF boiler

Lignite

FLUENT

A 550 MWe tangentially-fired boiler

Brown coal

AVL Fire

A front wall-fired 300 MW utility boiler

Bituminous coal

FLUENT

A 500 MWe coal-fired boiler

Wood, bituminous coal

FLUENT

A 100 MWth front-wall-fired furnace with 4 swirl burners

Coal

FLUENT

A 200 MWe tangentially fired utility boiler

Bituminous coal

FLUENT

A 1000 MWe ultra super critical utility boiler

Powder River Basin coal

CFX

Modeling approaches ED; Char reactions: C + O2 ? CO, C + CO2 ? CO (kinetics both from TGA), C + H2O?CO + H2, kinetics/diffusion-limited Realizable k–e; DO, oxy-fuel WSGGM (domain-based). Pyrolysis: single rate model (kinetics from TGA); Volatile combustion: VM + O2 ? CO + H2, CO + O2 M CO2, H2 + O2 ? H2O, EDC; Char reactions: C + O2 ? CO, C + CO2 ? CO, C + H2O ? CO + H2, kinetics/diffusion-limited Standard k–e; DO, EWBM for gaseous radiative properties. Pyrolysis: single kinetic rate model; Volatile combustion: VM + O2 ? CO + H2O, CO + O2 ? CO2, FR/ED; Char reactions: C + O2, C + CO2, semi global intrinsic kinetic law Standard k–e; DTRM (no particle effect in RTE), air–fuel WSGGM (gray). Pyrolysis: Single rate model; Volatile combustion: VM + O2 ? CO + H2 + H2O, CO + H2O M CO2 + H2, H2 + O2 M H2O, EBU; Char reactions: C + O2 ? CO, C + CO2 ? CO, C + H2O ? CO + H2, kinetics/diffusion-limited Realizable k–e; DO, oxy-fuel WSGGM (domain-based). Pyrolysis: twocompeting-rate model; Volatile combustion: VM + O2 ? CO + H2O, CO + O2 ? CO2, ED; Char reactions: C + O2, kinetics/diffusion-limited Realizable k–e; DO, air–fuel WSGGM (domain-based, soot radiation included). Pyrolysis: single kinetic rate model (kinetics from FG-DVC for coal and from literature for biomass); Volatile combustion: VM + O2 ? CO + H2O, CO + O2 ? CO2, ED; Char reactions: C + O2, intrinsic model Realizable k–e; DO, oxy-fuel WSGGM (gray vs. non-gray, domain-based), particle radiation (ep = 0.9 in pyrolysis, ep = 0.75 after pyrolysis; rp = 0.6 m1). Pyrolysis: single rate model (kinetics from DTF tests); Volatile combustion: VM + O2 ? CO + H2O, CO + O2 ? CO2, FR/ED; Char reactions: C + O2 (kinetics from DTF tests), kinetics/diffusion-limited Standard k–e (with full buoyancy effect); various constant absorption coefficients for combined gas/particle, ag+p = 0.279–0.318 m1. Pyrolysis: CPD model; Volatile combustion: refined JL 4-step mechanism for oxy-fuel, EDC; Char reactions: kinetics/diffusion-limited Standard k–e; DTRM, gas absorption coefficient by a gray gas model, plus 0.4 for particle effect. Pyrolysis: single kinetic rate; Volatile combustion: VM + O2 ? CO + H2O, CO + O2 ? CO2, ED (with modified constants A = 1.5, B = 1); Char reactions: C + O2 ? CO, kinetics/diffusion-limited

Reference

[148]

[149]

[150]

[151]

[87]

[48]

[152]

[153]

All the abbreviations in the table are given and explained in the nomenclature.

k1

HCN þ O2 ! NO þ    k2

HCN þ NO ! N2 þ    k3

NH3 þ O2 ! NO þ    k4

NH3 þ NO ! N2 þ   

R1 ¼ 1  1010 e33728:4=T X HCN X aO2 12 30204:6=T

R2 ¼ 3  10 e

X HCN X NO

ðR10Þ ðR11Þ

R3 ¼ 4  106 e16109:1=T X NH3 X aO2

ðR12Þ

R4 ¼ 1:8  108 e13592:1=T X NH3 X NO

ðR13Þ

The oxygen reaction order, a, is uniquely related to O2 mole fraction in the flame and is in the range of 0–1. Based on the conversion rates R, the source terms of HCN, NH3 and NO can be readily expressed as,

Svolatile;NO ¼ ðR1  R2 þ R3  R4 Þ

MWNO Ppa Ru T

ð16Þ

Finally, most of the char-N is oxidized to NO mainly as a desorption product from the oxidized char nitrogen atoms [133] while the rest is converted to N2, as shown in Fig. 6. This NO source can be calculated as follows, in which the conversion factor, g, is to account for that some of the char-N is converted to N2.

Schar;NO ¼

Sc Y N;char MWNO g MWN V

ð17Þ

SNO ¼ Svolatile;NO þ Schar;NO  SNO;reduction

ð19Þ

As seen from the above analysis, the partitioning of the fuel-N into volatiles and char during pyrolysis, as well as the type of the volatiles, plays an important role in final NOx formation. The fraction of the fuel-N released with volatiles in pyrolysis depends on the fuel type, pyrolysis temperature and residence time. It increases with the oxygen content in the parent fuel. At low temperatures or residence times, fuel-N is preferentially retained in the char; while at high temperatures it is depleted [117]. The NOx precursors (e.g., HCN and NH3) formed during the primary and secondary pyrolysis also play an important role in the final NOx formation. The split ratio of the volatile-N into the different precursors varies with the fuel type, pyrolysis temperature and heating rate. Devolatilization of bituminous coals produces mainly HCN, while more NH3 evolves from biomass and low-rank coals. The HCN/NH3 ratio tends to increase with the heating rates. At low heating rates, NH3 is the dominant precursor for both biomass and coals. At high heating rates, HCN is the main precursor in bituminous coal pyrolysis while pyrolysis of low-rank coals and biomass yields significant amounts of NH3, as summarized in [117]. 4. CFD modeling of PF oxy-fuel combustion 4.1. General modeling routine for PF combustion

The NO formed may be reduced to N2 over the residual char. The NO consumption rate is evaluated as, 17166:3=T

SNO;reduction ¼ cs ABET MWNO  0:23 e

X NO Patm

ð18Þ

If neglecting the minor direct conversion of fuel-N to NO during the primary pyrolysis (as seen in Fig. 6), the source term in the transport equation for the mass fraction of fuel-NO can be evaluated as,

By integrating fluid flow with reacting PF particles, turbulence, heat transfer, and all kinds of chemical reactions (including pollutant formation), as shown in Fig. 7, CFD modeling has been becoming increasingly useful in development of conventional PF-fired burners and boilers, from innovative design to performance evaluation. Besides other factors (e.g., a fine high-quality mesh yielding grid-independent CFD solutions, correct definition of boundary

C. Yin, J. Yan / Applied Energy 162 (2016) 742–762

755

Table 3 Key sub-processes and modeling recommendations for oxy-fuel conditions. Sub-process

Oxy-fuel impact?

Modeling recommendations

Turbulence

No

Particle motion

No

Gas radiation

Yes

Particle radiation

No

PF pyrolysis

No

Gas combustion

Yes

Char reactions NOx emissions

Yes Yes

For industry-scale PF oxy-firing, the realizable k–e is recommended. For lab-scale PF oxy-firing, the SST k–x is a good option since it’s possible to generate a fine enough near-wall mesh to fully exploit its advantages; more accurate but computationally expensive LES can also be used (e.g., in [138,158,159]) For oxy-fuel combustion of pulverized coal, the conventional particle motion model, Eq. (1), is still applicable. For oxy-fuel combustion of biomass fuels, more cares need to be taken in particle motion, as explained in detail below Domain-based oxy-fuel WSGGMs, preferably in non-gray implementation, are recommended, to properly address the impacts of high-concentration CO2 and H2O and to avoid grid-dependent solutions induced by cell-based WSGGMs Conversion-dependent particle radiation property models are recommended, due to the fact that particle radiation overwhelms gas radiation in all PF furnaces The single rate model with kinetic data derived from the similar conditions as in PF flames (e.g., heating rates, temperatures) is a good compromise. The network models may be better if the structural details of PF particles are known Global mechanisms with kinetic parameters refined for oxy-fuel conditions are recommended for large-scale PF oxy-firing, in combination with the EDC for turbulence–chemistry interaction. Alternative options are summarized below The single-film model accounting for all the char oxidation and gasification reactions with O2, CO2 and H2O is recommended Cares need to be taken in the partitioning of fuel-N into volatiles and char during pyrolysis, the type and yield of the volatiles, and the split of the volatile-N into different precursors under oxy-fuel conditions

and/or initial conditions, appropriate numerical methods and convergence), development and use of proper sub-models or schemes for the underlying sub-processes plays a crucial role in a reliable CFD modeling of PF combustion. 4.2. Modeling of PF oxy-fuel combustion: A handy overview CFD modeling studies of PF oxy-fuel combustion have been performed in the past decade. Table 2 gives a handy overview of the representative works, in which the focus is placed on the modeling approaches, i.e., how the key issues in combustion physics and combustion chemistry are addressed. The former mainly includes fuel particle aerodynamics, turbulence and radiation modeling, while the latter addresses pyrolysis, gas phase combustion and char conversion modeling. In the combustion chemistry, the methods to obtain the kinetic parameters for the various chemical reaction models are also highlighted, if they are provided in the source references. The literature review in Table 2 is organized in different orders. Overall, it is in chronological order. Table 2 starts with the modeling studies of real oxy-PF combustion processes which have experimental data available for model comparison. For this section, the various studies which are based on the same or similar test case are presented successively in group in the table. Then, the modeling studies of hypothetical oxy-PF combustion in conventional air–fuel utility boilers are presented in the table. As seen in Table 2, several PF oxy-fuel combustion test cases are often used as the modeling object. The first are tests in different down-fired swirl burner furnaces, e.g., oxy-fuel combustion of a sub-bituminous coal in the CANMET 300 kWth furnace, oxy-lignite combustion in the Chalmers 100 kWth furnace [154], oxy-fuel combustion of a German lignite in the RWTH Aachen 100 kWth furnace, and oxy-firing of an Australian sub-bituminous coal in the IHI 1.2 MWth furnace. These four test cases are numerically investigated in [134,135], [42], [137–139], and [58,140], respectively. The second is the oxy-coal combustion tests and oxy-fuel co-combustion of coal and biomass under various conditions, performed in the near-laminar down-fired lab-scale EFR at INCAR-CSIC (Spain). The different test cases and key topics (e.g., burnout and NOx emissions), as well as impacts of various sub-models, are numerically studied in [43,44,46,85,86]. The third are the oxy-coal combustion tests (or oxy-fuel co-firing of coal/biomass) conducted in the RWEn 0.5 MWth CTF equipped with an IFRF aerodynamically air-staged burner at Didcot power station (UK) [155,156], which are modeled in [141–144]. Table 2 shows

that very different modeling strategies are used in these numerical studies. The other big group is modeling of hypothetical oxy-fuel combustion of PF in industrial boilers, in which the real air–PF combustion is often employed as the baseline case for modeling validation and then oxy-PF combustion is numerically studied either to examine under which oxy-fuel conditions the similar combustion characteristics to air–fuel combustion can be attained or to evaluate the impacts of various sub-models on oxy-fuel combustion, e.g., [48,87,149–153]. Zhang et al. [148] present the first trial in the world for experimental and modeling investigation of oxy-fuel combustion of Victorian brown coal in a 3 MWth pilot-scale tangentially-fired pulverized-coal furnace. Two coal samples with different moisture contents have been tested in air- and oxy-firing modes with oxygen level varying from 27% (vol) to 40% in the furnace. To interpret the experimental data, the test cases are also numerically studied in which a series of validated refined sub-models for oxy-fuel combustion are taken into account. The modeling results show reasonable agreement with the experimental data. Since the pilot-scale test facility system is identical with a real industrial power plant, the testing results and modeling strategies are of great interest for design and optimization of large-scale oxy-coal utility boilers. From the modeling strategy point of view, some of the oxy-fuel CFD studies just stick to the conventional air–fuel combustion modeling method, without sufficiently addressing the main differences between the two combustion conditions (i.e., assuming all the air–fuel combustion sub-models naturally adjusted to oxyfiring). On the contrary, a few CFD modeling studies not only account for the key issues in a general combustion CFD (e.g., mesh and turbulence) but properly address the main differences induced by oxy-fuel combustion (e.g., radiation, gas-phase and char combustion) by refining the relevant sub-models for oxy-fuel conditions and implementing them into CFD, e.g., [44,46,48,148]. Nevertheless, there are still issues that need to be clarified or improved, even in the well-defined CFD studies. For instance, in the CFD modeling of oxy-coal combustion in a 3 MWth tangentially fired furnace [148], particle radiation needs to be explicitly explained and the applicability of the TGA-derived pyrolysis and char reaction kinetics in the pilot-scale PF furnace also needs to be further verified. In PF furnaces, particle radiation plays a more dominant role than gas radiation, as demonstrated in [38]. In addition, the heating rates in a typical PF flame (on the order of 105 K/s) are several orders of magnitude greater than those in a TGA, which is expected to impede the applicability of TGA-derived kinetic

756

C. Yin, J. Yan / Applied Energy 162 (2016) 742–762

N2 Air separation plant

Air

Multi-stage turbines Gas generator

Crude fuel (e.g. gas, oil, coal, or biomass)

HP

IP

Electrical generator

LP

Fuel processing plant Steam/CO2 CO2 to CO2 recovery compression

Heat recovery Condenser Recycle water

Fig. 8. Steam turbine-based oxy-fuel power generation system [165].

Secondary Recycle Option D

Air

ASU

Boiler

Heat Exchanger

Option C Option B Option A

FGD

ESP

SCR

FGC

Primary Recycle Mill

Coal

Fig. 9. Flue gas cleaning and recycling in oxy-fuel combustion [174].

parameters in PF combustion modeling. In the modeling of oxycoal combustion in a 200 MWe tangentially fired utility boiler [48], the dominant role of particle radiation in PF furnaces is properly addressed and the DTF tests are appropriately used to determine the pyrolysis and char oxidation kinetic parameters. However, the turbulence–chemistry interaction can be better addressed using the EDC instead of FR/ED and char gasification reactions may also be better taken into account. Fast chemistry approaches (e.g., ED or EBU) are found to be inadequate in modeling oxy-fuel combustion systems where reverse reactions play an important role, while EDC can produce satisfactory predictions for oxy-fuel combustion [102,157]. In short, efforts still need to be made on oxy-fuel CFD modeling, in order to promote it into a powerful tool in oxy-fuel combustion technology development as it has demonstrated in air–fuel combustion development. 4.3. Key sub-processes and modeling recommendations for oxy-fuel combustion CFD modeling has been used to interpret oxy-fuel combustion test results or study oxy-fuel combustion potentials. Challenges

still exist to evaluate and refine the sub-models or mechanisms which are originally developed for air–fuel combustion and make them applicable to oxy-fuel conditions. Based on the discussions in the above sections, the key sub-processes in PF combustion (corresponding to various aspects of CFD modeling), whether or not they are remarkably or directly affected by oxy-firing, and the modeling recommendations for oxy-fuel combustion are briefly summarized in Table 3. Here, the issue of modeling biomass particle aerodynamics in oxy-fuel co-firing of biomass is elaborated in more detail. Biomass is a non-friable, fibrous material. Biomass particles prepared for suspension-firing often have large and irregular shapes and relatively low density [160], resulting in distinctly different dynamic behavior of the fuel particles and greatly affecting their pneumatic transport, ignition and combustion in the burner quarl, flame characteristics, design and operation of fuel/air staging, and burnout [161]. Irregularly shaped biomass particles under laboratory conditions can spin at up to 12.7 kHz [162], yielding additional forces on the particles (e.g., lift). The instantaneous drag and lift are also dependent on particle orientation. Moreover, biomass often contains a large fraction of moisture and volatiles. The rapid release of moisture and volatiles from biomass particles under PF firing

NaOH / Na2CO3 H2 O

FGD

FGC

30°C, 1atm ~77% CO2 (50-80%)

H2O H2O CaSO4 H2SO4 / Na2CO3 HNO3 / Na2NO3

N2, O2, Ar, <10% total CO2

Compression

1.25 bar

Activated Carbon Beds

H2O, Heavy metals, SOx, HCl, HF

22 bar Molecular 2 Stage Compression Sieves

Heat

H2O

CO2 Liquefaction & Separation 99.7% Liquid CO2 9 t/h, 15-20 bar -15 to -50°C

Fig. 10. Flue gas treatment and compression process of Vattenfall’s 30 MWth oxy-fuel pilot plant [183].

Chilled NH3

C. Yin, J. Yan / Applied Energy 162 (2016) 742–762

757

Fig. 11. Oxy-fuel combustion integrated with the solid oxide fuel cells (SOFC) and gas turbine [184].

conditions may result in extra intermittent forces. Due to these factors, the applicability of the conventional particle motion model, which solves only translational motion and retains only drag and gravity in the equation of motion, needs to be investigated. An extended particle motion model has been successfully developed to more reliably track large, non-spherical particles in dilute two-phase flows, in which the coupled particle translational and rotational motions are solved simultaneously and other important forces than drag and gravity are also taken into account [163,164]. Such an extended model for particle aerodynamics may need to be adapted and used in CFD study of oxy-fuel combustion of pulverized biomass particles. Methods that can properly address the different gas-phase combustion chemistry induced by oxy-fuel conditions are also summarized in more detail here. The first is the refined WD 2-step or JL 4-step global mechanisms coupled with the EDC for turbulence–chemistry interaction, in which the kinetic parameters are adjusted to make the schemes applicable to oxy-firing modeling. The second is the WD multiple-step quasi-global mechanism (12 species and 22 reactions) [94,95] coupled with the EDC, since the quasi-global mechanism appropriately considers the CO2 chemical effects [100]. The third is the mixture fraction method and the steady flamelet chemistry model with a skeletal mechanism (25 reversible reactions and 17 species) [159]. 5. System performance and economic analysis of PF oxy-fuel combustion The new combustion characteristics and modeling issues in oxy-fuel combustion are summarized and discussed above. Here,

the key issues that influence the performance of oxy-fuel based power generation and carbon capture and storage systems (CCS) are discussed. 5.1. System integration of PF oxy-fuel combustion Oxy-fuel combustion can be applied to all types of fossil fuelbased power generation systems. Fig. 8 shows the most common system configuration under development and demonstration, i.e., a steam turbine-based oxyfuel power generation system [165], in which oxygen is produced from the extra air separation unit (ASU) and used together with the recycled flue gas for fuel combustion to generate high-pressure and high-temperature steam for electricity production. For such a system, the main challenges include the increased costs due to the additional equipment required (e.g., ASU for O2 production and flue gas cleaning unit for CO2 capture), integration of the auxiliary equipment, additional power generation capacity to overcome losses in output, and air separation [166,167]. The system performance significantly depends on how the various equipment or subsystems are designed and integrated as a whole. One of the important new issues in an oxy-fuel power generation system is the energy consumption of the ASU, which is quite intensive in the range of 3–4% [168]. The most common technology used for the ASU is cryogenic distillation, in which compressed air is introduced into the distillation column to separate air into an oxygen-rich stream and a nitrogen-rich stream. Attempts have been made to improve the performance of cryogenic ASU [169]. Other technologies for the ASU include adsorption and membranes [170,171], which have the same level of energy consumption but

758

C. Yin, J. Yan / Applied Energy 162 (2016) 742–762

high costs associated with fabrication, installation, operation and maintenance compared to the cryogenic ASU [168]. Efforts are made to reduce the energy consumption and capital costs associated with oxygen production. For example, the chemical looping based air separation (CLAS) process is proposed as an alternative technology [168], which, similar to chemical looping combustion, uses metal oxide for oxygen production. There are potential to integrate the CLAS with different heat sources including solar energy. Besides the efforts on energy consumption reduction, the use of ASU as energy storage for exploring the cost benefits by peak and off peak operation of oxy-fuel plants is also discussed [172]. Another method is to use the bed materials as thermal energy storage for oxy-fuel fluidized bed combustors [173]. Flue gas cleaning and recycling is another important new issue in an oxy-fuel power generation system. Flue gas recycle in the oxy-fuel combustion system consists of several units, e.g., electrostatic precipitators (ESP), flue gas desulphurization (FGD), selective catalytic reduction (SCR) and flue gas condensation (FGC), as seen in Fig. 9. Depending on the operation requirements and the impurities in the fuels (e.g., S, N, and even Cl), different layouts can be used for removal of the pollutants (e.g., particles, SOx, and NOx), which result in different performances of flue gas recycle [4,25,174,175]. The impacts of the impurities on oxy-fuel combustion based CO2 capture and storage processes have been investigated [176,177]. The impurities in the CO2 stream of oxy-coal combustion include gaseous N2/Ar, O2 and H2O, as well as SOx and NOx, which can change the thermodynamic and transport properties and the phase behavior of CO2 mixtures [177–181] and then affect the system design, operation and optimization. High purity requirements inevitably result in additional costs and high efficiency penalty as well as safety operation issues [6,182]. So far, no generic solution to the impurities and flue gas cleaning is available due to the lack of operation experience of oxy-fuel combustion. The very limited experience in handling the flue gas from oxy-fuel combustion is obtained from Vattenfall 30 MWth oxy-fuel pilot plant, the only plant in operation in the world [183]. Fig. 10 shows the flue gas treatment and compression process in the pilot plant. The removal of SOx, NOx, Hg and other emissions in the flue gas cleaning processes might be useful for the future design of flue gas cleaning process in oxy-fuel combustion systems. Besides the common system configuration as seen in Fig. 8, oxyfuel combustion can also be integrated with more complicated cycles. Fig. 11 shows an integrated power generation system combining solid oxide fuel cell (SOFC) and oxy-fuel combustion technology for high efficiency and CO2 capture [184,185]. The system includes three power generation parts (i.e., SOFC, gas turbine and oxy-combustion cycle) and a power consumption part (i.e., CO2 capture and compression unit). Oxygen is produced from the cathode exit gas using ion transport membrane technology. It should be pointed out that chemical looping combustion [186] also belongs to the category of oxy-fuel combustion, in which the oxygen is carried by solid metal oxides rather than in the form of gaseous oxygen. This is not discussed here since the focus of this review is gaseous oxygen combustion process. 5.2. Commercialization of oxy-fuel combustion The oxy-CFB process has been experimentally studied in about 20 bench scale and pilot scale devices at sizes in the range of 30– 800 kWth [187]. Several demonstration projects of oxy-fuel combustion plants were planned and implemented. Vattenfall’s Schwarze Pumpe Pilot Plant (30 MWth) in Germany which was commissioned in 2008 was successfully completed in testing various components [188]. The operation experiences of Vattenfull’s Schewarze Pumpe plant will be useful for the future implementation of commercial oxy-fuel plants. However, similar as other CCS plants, to achieve

oxy-fuel plants deployment at commercial scale needs policy parity, technology and funding incentives, technology R&D and demonstration, and accumulated installation capacity [189]. 5.3. Costs of CO2 avoided by oxy-fuel combustion The costs of CO2 avoided by CCS are ranged between 48 and 109 US $ per tonne CO2, according to the updates made by Global CCS Institute in 2015 [190]. According to the IEA GHG report of 2005, a plant with 500 MWe net power output, the oxy-fuel case would have a cost of electricity of 72.8 US $/MW h compared with 49 US $/MW h for the base case. The cost of CO2 avoided would be about 41 US $/tonne while the plant thermal efficiency would decrease from 44.2% LHV (net) to 35.4% [191]. Borgert and Rubin [192] estimated the CO2 avoided of oxy-fuel plants at the range of 40–80 US $/tonne. Oboirien et al. [193] made technical and economic viability of oxy-fuel technology for CO2 capture for South African coal-fired power stations of 3600 MW. The total capital costs and cost of electricity for the six plants were different, resulting in the cost of electricity varying from 101 to124 US $/MW h. The cost of CO2 avoided was 75 US $/tonne at the case of Kendal and was 86 US $/tonne at Lethabo. A common method of cost estimation for CO2 capture and storage was developed [194], which can be served as a framework for the CCS cost estimation. 6. Conclusions and outlook The PF oxy-fuel combustion fundamentals and CFD modeling, as well as the system performance of oxy-PF combustion, are thoroughly reviewed, from which the key conclusions are summarized below. (a) Turbulent flow and PF particle aerodynamics and their modeling: Oxy-fuel combustion only has indirect impacts on turbulent flow and particle aerodynamics. The turbulence models and particle motion models for air–fuel combustion can be naturally adjusted to oxy-fuel conditions. (b) Heat transfer and its modeling: Models for gaseous radiative properties and soot formation need to be refined in order to properly address the impacts of the high partial pressures of CO2 and H2O in oxy-fuel combustion. However, in PF oxyfuel furnaces, the high-concentration CO2 and H2O only slightly enhance the total radiative heat transfer because particle radiation always overwhelms gas radiation in PF furnaces. (c) PF pyrolysis and its modeling: Oxy-fuel atmosphere does not remarkably affect PF pyrolysis. So the same pyrolysis kinetic parameters can be used in air–fuel and oxy-fuel combustion modeling. However, it is important to properly determine the pyrolysis kinetic parameters from the similar conditions as in the PF flame under study (e.g., heating rates, temperatures). (d) Gas phase combustion, turbulence–chemistry interaction and their modeling: The high-concentration CO2 and H2O in oxy-fuel flames greatly affect gas phase combustion. Global combustion mechanisms need to be refined for the use in oxy-fuel CFD. Fast chemistry approach (e.g., EBU) becomes inadequate in oxy-fuel modeling and EDC needs to be used, due to the important role of reverse reactions in oxy-fuel systems. (e) Char combustion and its modeling: It is not conclusive if oxy-firing favors a higher char burnout. There is a lot of scatter in the literature, depending on various factors. The extended single-film model with all the char oxidation and gasification reactions can be used for char conversion under oxy-fuel conditions.

C. Yin, J. Yan / Applied Energy 162 (2016) 742–762

(f) NOx formation and its modeling: Oxy-firing largely reduces NOx emissions, up to 70–80% reduction. Fuel-NOx formation mechanisms for air–fuel combustion may be applicable to oxy-fuel conditions, with proper consideration in the partitioning of the fuel-N into volatiles and char during pyrolysis, the type and yield of the volatiles, and the split of the volatile-N into different precursors. The research needs with regard to the combustion fundamentals and modeling are also identified here. (1) The current implementation of WSGGMs should be changed as follows to eliminate the discontinuity problem. If the local gas composition is in-between two composition conditions in a WSGGM, the relevant parameter tables are used to obtain two emissivity values and the emissivity of the gas under consideration is then evaluated from the two values by interpolation. This new implementation is to be demonstrated. More efforts need to be made for particle radiation, since it always overwhelms gas radiation in PF furnaces. (2) Some of the refined global mechanisms for gas phase combustion under oxy-fuel conditions may have numerical instability due to the negative order of reactions with respect to some intermediate species (e.g., H2). More robust and general global mechanisms are needed. The impacts of CO2, CO and H2O on fuel-NOx formation mechanism under oxy-fuel conditions also need to be examined. (3) Ignition and burnout tests under well-defined oxy-fuel conditions are required. Reliable interpretation of experimental observations needs to properly address all the key factors, e.g., the primary ignition mode, fuel particle size, fuel reactivity, and local gas conditions in the vicinity of the particle. (4) The release characteristics of inorganic elements (e.g., K, Na, S, Cl, alkali, and heavy metals such as Hg, As and Se) under oxy-fuel conditions need to be investigated, especially formation mechanisms of the species with corrosive effects (e.g., SO2, SO3, H2S) and their impacts on oxy-fuel systems (e.g., corrosion and CO2 compression). The release of the inorganic elements can also affect the formation of aerosols and fly ash, and aggravate pollutant emissions. Little has been done in this regard [195–200]. (5) CFD modeling needs to be extended to incorporate the latest achievements in sub-model development and to include minor species of high relevance in PF oxy-firing (e.g., SO2, SO3) and deposition and corrosion. (6) Advanced oxy-fuel technologies such as flameless oxycombustion or mild oxy-fuel combustion, as well as oxyfiring of biomass for below-zero CO2 emissions, need to be investigated on a fundamental basis. (7) Design, modeling and optimization of advanced oxy-fuel based power generation and CCS systems for high efficiency and low energy penalty and capital costs by properly integrating individual promising technologies represents the right trend in future and deserves more research and development.

References [1] Buhre BJP, Elliott LK, Sheng CD, Gupta RP, Wall TF. Oxy-fuel combustion technology for coal-fired power generation. Prog Energy Combust Sci 2005;31:283–307. [2] Wall T, Liu Y, Spero C, Elliott L, Khare S, Rathnam R, et al. An overview on oxyfuel coal combustion – state of the art research and technology development. Chem Eng Res Des 2009;87:1003–16. [3] Normann F, Andersson K, Leckner B, Johnsson F. Emission control of nitrogen oxides in the oxy-fuel process. Prog Energy Combust Sci 2009;35:385–97. [4] Toftegaard MB, Brix J, Jensen PA, Glarborg P, Jensen AD. Oxy-fuel combustion of solid fuels. Prog Energy Combust Sci 2010;36:581–625.

759

[5] Scheffknecht G, Al-Makhadmeh L, Schnell U, Maier J. Oxy-fuel coal combustion – a review of the current state-of-the-art. Int J Greenh Gas Control 2011;55:516–35. [6] Chen L, Yong SZ, Ghoniem AF. Oxy-fuel combustion of pulverized coal: characterization, fundamentals, stabilization and CFD modeling. Prog Energy Combust Sci 2012;38:156–214. [7] Yan J. Carbon capture and storage (CCS). Appl Energy 2015;148:A1–6. [8] Kitto JB, Stultz SC. Steam: its generation and use. 41st ed. Ohio: The Babcock & Wilcox Company; 2005. [9] Cengel YA. Heat transfer: a practical approach. 2nd ed. New York: McGrawHill; 2003. [10] Turns SR. An introduction to combustion: concepts and applications. 2nd ed. Singapore: McGraw-Hill; 2000. [11] Cabrejos FJ, Klinzing GE. Pickup and saltation mechanisms of solid particles in horizontal pneumatic transport. Powder Technol 1994;79:173–86. [12] Pitsch H. Large-eddy simulation of turbulent combustion. Annu Rev Fluid Mech 2006;38:453–82. [13] Gicquel LYM, Staffelbach G, Poinsot T. Large eddy simulations of gaseous flames in gas turbine combustion chambers. Prog Energy Combust Sci 2012;38:782–817. [14] Holtmeyer ML, Kumfer BM, Axelbaum RL. Effects of biomass particle size during cofiring under air-fired and oxyfuel conditions. Appl Energy 2012;93:606–13. [15] Maxey MR, Riley JJ. Equation of motion for a small rigid sphere in a nonuniform flow. Phys Fluids 1983;26:883–9. [16] Lazaro BJ, Lasheras JC. Particle dispersion in a turbulent, plane, free shear layer. Phys Fluids A 1989;1:1035–44. [17] Ranz WE, Marshall WR. Evaporation from drops. Part I.. Chem Eng Process 1952;48:141–6. [18] Chui EH, Hughes PMJ, Raithby GD. Implementation of the finite volume method for calculating radiative transfer in a pulverized fuel flame. Combust Sci Technol 1993;92:225–42. [19] Ansys Fluent. Theory guide and users guide. Ansys Inc. USA; 2014. [20] Hottel HC, Sarofim AF. Radiative transfer. New York: McGraw-Hill; 1967. [21] Wang L, Endrud NE, Turns SR, D’Agostini MD, Slavejkov AG. A study of the influence of oxygen index on soot, radiation, and emission characteristics of turbulent jet flames. Combust Sci Technol 2002;174:45–72. [22] Andersson K, Johnsson F. Flame and radiation characteristics of gas-fired O2/ CO2 combustion. Fuel 2007;86:656–68. [23] Andersson K, Johansson R, Johnsson F, Leckner B. Radiation intensity of propane-fired oxy-fuel flames: implications for soot formation. Energy Fuels 2008;22:1535–41. [24] Yin C, Johansen LCR, Rosendahl L, Kær SK. A new weighted sum of gray gases model applicable to CFD modeling of oxy-fuel combustion: derivation, validation and implementation. Energy Fuels 2010;24:6275–82. [25] Hu Y, Yan J, Li H. Effects of flue gas recycle on oxy-coal power generation systems. Appl Energy 2012;97:255–63. [26] Johansson R, Andersson K, Leckner B, Thunman H. Models for gaseous radiative heat transfer applied to oxy-fuel conditions in boilers. Int J Heat Mass Transf 2010;53:220–30. [27] Krishnamoorthy G, Sami M, Orsino S, Perera A, Shahnam M, Huckaby ED. Radiation modeling in oxy-fuel combustion scenarios. Int J Comput Fluid Dyn 2010;24:69–82. [28] Johansson R, Leckner B, Andersson K, Johnsson F. Account for variations in the H2O to CO2 molar ratio when modeling gaseous radiative heat transfer with the weighted-sum-of-grey-gases model. Combust Flame 2011;158:893–901. [29] Kangwanpongpan T, França FHR, da Silva RC, Schneider PS, Krautz HJ. New correlations for the weighted-sum-of-gray-gases model in oxy-fuel conditions based on HITEMP 2010 database. Int J Heat Mass Transf 2012;55:7419–33. [30] Krishnamoorthy G. A new weighted-sum-of-gray-gases model for oxycombustion scenarios. Int J Energy Res 2013;37:1752–63. [31] Yin C, Rosendahl L, Kær SK. Chemistry and radiation in oxy-fuel combustion: a computational fluid dynamics modeling study. Fuel 2011;90:2519–29. [32] Hjärtstam S, Johansson R, Andersson K, Johnsson F. Computational fluid dynamics modeling of oxy-fuel flames: the role of soot and gas radiation. Energy Fuels 2012;26:2786–97. [33] Yin C. Nongragy-gas effects in modeling of large-scale oxy–fuel combustion processes. Energy Fuels 2012;26:3349–56. [34] Wang L, Haworth DC, Turns SR, Modest MF. Interactions among soot, thermal radiation, and NOx emissions in oxygen-enriched turbulent nonpremixed flames: a computational fluid dynamics modeling study. Combust Flame 2005;141:170–9. [35] Nikolopoulos N, Agranoitis M, Violidakis I, Karampinis E, Nikolopoulos A, et al. Parametric investigation of a renewable alternative for utilities adopting the co-firing lignite/biomass concept. Fuel 2013;113:873–97. [36] Johansson R, Leckner B, Andersson K, Johnsson F. Influence of particle and gas radiation in oxy-fuel combustion. Int J Heat Mass Transf 2013;65:143–52. [37] Bäckström D, Johansson R, Andersson K, Johnsson F, Clausen S, Fateev A. Measurement and modeling of particle radiation in coal flames. Energy Fuels 2014;28:2199–210. [38] Yin C. On gas and particle radiation in pulverized fuel combustion furnaces. Appl Energy 2015;157:554–61. [39] Smith TF, Shen ZF, Friedman JN. Evaluation of coefficients for the weighted sum of gray gases model. J Heat Transf-Trans ASME 1982;104:602–8. [40] Yin C. Refined weighted sum of gray gases model for air-fuel combustion and its impacts. Energy Fuels 2013;27:6287–94.

760

C. Yin, J. Yan / Applied Energy 162 (2016) 742–762

[41] Ma L, Gharebaghi M, Porter R, Pourkashanian M, Jones JM, Williams A. Modeling methods for co-fired pulverised fuel furnaces. Fuel 2009;88:2448–54. [42] Al-Abbas AH, Naser J, Dodds D. CFD modeling of air-fired and oxy-fuel combustion of lignite in a 100 KW furnace. Fuel 2011;90:1778–95. [43] Álvarez L, Gharebaghi M, Pourkashanian M, Williams A, Riaza J, Pevida C, et al. CFD modeling of oxy-coal combustion in an entrained flow reactor. Fuel Process Technol 2011;92:1489–97. [44] Álvarez L, Yin C, Riaza J, Pevida C, Pis JJ, Rubiera F. Oxy-coal combustion in an entrained flow reactor: application of specific char and volatile combustion and radiation models for oxy-firing conditions. Energy 2013;62:255–68. [45] Zhang J, Prationo W, Zhang L, Zhang Z. Computational fluid dynamics modeling on the air-firing and oxy-fuel combustion of dried Victorian brown coal. Energy Fuels 2013;27:4258–69. [46] Álvarez L, Yin C, Riaza J, Pevida C, Pis JJ, Rubiera F. Biomass co-firing under oxy-fuel conditions: a computational fluid dynamics modeling study and experimental validation. Fuel Process Technol 2014;120:22–33. [47] Lockwood FC, Rizvi SM, Shah NG. Comparative predictive experience of coal firing. Proc Inst Mech Eng Part C – J Eng Mech Eng Sci 1986;200:79–87. [48] Zhang J, Ito T, Ito S, Riechelmann D, Fujimori T. Numerical investigation of oxy-coal combustion in a large-scale furnace: non-gray effect of gas and role of particle radiation. Fuel 2015;139:87–93. [49] Du X, Annamalai K. The transition of isolated coal particle. Combust Flame 1994;97:339–54. [50] Zhang DK, Wall TF. Ignition of coal particles: the influence of experimental technique. Fuel 1994;73:1114–9. [51] Wendt C, Eigenbrod C, Moriue O, Rath HJ. A model for devolatilization and ignition of an axisymmetric coal particle. Proc Combust Inst 2002;29:449–57. [52] Grotkjær T, Dam-Johansen K, Jensen AD, Glarborg P. An experimental study of biomass ignition. Fuel 2003;82:825–33. [53] Faúndez J, Arenillas A, Rubiera F, García X, Gordon AL, Pis JJ. Ignition behavior of different rank coals in an entrained flow reactor. Fuel 2005;84:2172–7. [54] Molina A, Shaddix CR. Ignition and devolatilization of pulverized bituminous coal particles during oxygen/carbon dioxide coal combustion. Proc Combust Inst 2007;31:1905–12. [55] Shaddix CR, Molina A. Particle imaging of ignition and devolatilization of pulverized coal during oxy-fuel combustion. Proc Combust Inst 2009;32:2091–8. [56] Arias B, Pevida C, Rubiera F, Pis JJ. Effect of biomass blending on coal ignition and burnout during oxy-fuel combustion. Fuel 2008;87:2753–9. [57] Huang X, Jiang X, Han X, Wang H. Combustion characteristics of fine- and micro-pulverized coal in the mixture of O2/CO2. Energy Fuels 2008;22:3756–62. [58] Khare S, Wall TF, Farida AZ, Liu Y, Moghtaderi B, Gupta RP. Factors influencing the ignition of flames from air-fired swirl pf burners retrofitted to oxy-fuel. Fuel 2008;87:1042–9. [59] Ponzio A, Senthoorselvan S, Yang W, Blasiak W, Eriksson O. Ignition of single coal particles in high-temperature oxidizers with various oxygen concentrations. Fuel 2008;87:974–87. [60] Li Q, Zhao C, Chen X, Wu W, Li Y. Comparison of pulverized coal combustion in air and in O2/CO2 mixtures by thermo-gravimetric analysis. J Anal Appl Pyrol 2009;85:521–8. [61] Zhang L, Binner E, Qiao Y, Li CZ. In situ diagnostics of Victorian brown coal combustion in O2/N2 and O2/CO2 mixtures in drop-tube furnace. Fuel 2010;89:2703–12. [62] Riaza J, Álvarez L, Gil MV, Pevida C, Pis JJ, Rubiera F. Effect of oxy-fuel combustion with steam addition on coal ignition and burnout in an entrained flow reactor. Energy 2011;36:5314–9. [63] Riaza J, Gil MV, Álvarez L, Pevida C, Pis JJ, Rubiera F. Oxy-fuel combustion of coal and biomass blends. Energy 2012;41:429–35. [64] Zhang J, Kelly KE, Eddings EG, Wendt JOL. Ignition in 40 kW co-axial turbulent diffusion oxy-coal jet flames. Proc Combust Inst 2011;33:3375–82. [65] Khatami R, Stivers C, Joshi K, Levendis YA, Sarofim AF. Combustion behavior of single particles from three different coal ranks and from sugar cane bagasse in O2/N2 and O2/CO2 atmospheres. Combust Flame 2012;159:1253–71. [66] Khatami R, Stivers C, Levendis YA. Ignition characteristics of single coal particles from three different ranks in O2/N2 and O2/CO2 atmospheres. Combust Flame 2012;159:3554–68. [67] Jovanovic R, Rasuo B, Stefanovic P, Cvetinovic D, Swiatkowski B. Numerical investigation of pulverized coal jet flame characteristics under different oxyfuel conditions. Int J Heat Mass Transf 2013;58:654–62. [68] Momeni M, Yin C, Kær SK, Hansen TB, Jensen PA, Glarborg P. Experimental study on effects of particle shape and operating conditions on combustion characteristics of single biomass particle. Energy Fuels 2013;27:507–14. [69] Momeni M, Yin C, Kær SK, Hvid SL. Comprehensive study of ignition and combustion of single wooden particles. Energy Fuels 2013;27:1061–72. [70] Ahn S, Choi G, Kim D. The effect of wood biomass blending with pulverized coal on combustion characteristics under oxy-fuel condition. Biomass Bioenerg 2014;71:144–54. [71] Riaza J, Khatami R, Levendis YA, Álvarez L, Gil MV, Pevida C, et al. Single particle ignition and combustion of anthracite, semi-anthracite and bituminous coals in air and simulated oxy-fuel conditions. Combust Flame 2014;161:1096–108.

[72] Yi B, Zhang L, Huang F, Mao Z, Zheng C. Effect of H2O on the combustion characteristics of pulverized coal in O2/CO2 atmosphere. Appl Energy 2014;132:349–57. [73] Zou C, Cai L, Wu D, Liu Y, Liu S, Zheng C. Ignition behaviors of pulverized coal particles in O2/N2 and O2/H2O mixtures in a drop tube furnace using flame monitoring techniques. Proc Combust Inst 2015;35:3629–36. [74] Essenhigh RH, Misra MK, Shaw DW. Ignition of coal particles: a review. Combust Flame 1989;77:3–30. [75] Yu J, Lucas JA, Wall TF. Formation of the structure of chars during devolatilization of pf pulverized coal and its thermoproperties: a review. Prog Energy Combust Sci 2007;33:135–70. [76] Badzioch S, Hawksley PGW. Kinetics of thermal decomposition of pulverized coal particles. Ind Eng Chem Process Des Dev 1970;9:521–30. [77] Kobayashi H, Howard JB, Sarofim AF. Coal devolatilization at high temperatures. Proc Combust Inst 1976;16:411–25. [78] Grant DM, Pugmire RJ, Fletcher TH, Kerstein AR. Chemical percolation model of coal devolatilization using percolation lattice statistics. Energy Fuels 1989;3:175–86. [79] Fletcher TH, Kerstein AR, Pugmire RJ, Grant DM. Chemical percolation model for devolatilization. 2. Temperature and heating rate effects on product yields. Energy Fuels 1990;4:54–60. [80] Fletcher TH, Kerstein AR. Chemical percolation model for devolatilization. 3. Direct use of 13C NMR data to predict effects of coal type. Energy Fuels 1992;6:414–31. [81] Solomon PR, Hamblen DG, Carangelo RM, Serio MA, Deshpande GV. General model of coal devolatilization. Energy Fuels 1988;2:405–22. [82] Jovanovic R, Milewska A, Swiatkowski B, Goanta A, Spliethoff H. Sensitivity analysis of different devolatilization models on predicting ignition point position during pulverized coal combustion in O2/N2 and O2/CO2 atmospheres. Fuel 2012;101:23–37. [83] Lemaire R, Menage D, Menanteau S, Harion JL. Experimental study and kinetic modeling of pulverized coal devolatilization under air and oxycombustion conditions at a high heating rate. Fuel Process Technol 2014;128:183–90. [84] Advanced Fuel Research (AFR), Inc. FG-DVC Model; 2009 . [85] Álvarez L, Gharebaghi M, Jones JM, Pourkashanian M, Williams A, Riaza J, et al. Numerical investigation of NO emissions from an entrained flow reactor under oxy-coal conditions. Fuel Process Technol 2012;93:53–64. [86] Álvarez L, Gharebaghi M, Jones JM, Pourkashanian M, Williams A, Riaza J, et al. CFD modeling of oxy-coal combustion: prediction of burnout, volatile and NO precursors release. Appl Energy 2013;104:653–65. [87] Black S, Szuhánszki J, Pranzitelli A, Ma L, Stanger PJ, Ingham DB, et al. Effects of firing coal and biomass under oxy-fuel conditions in a power plant boiler using CFD modeling. Fuel 2013;113:780–6. [88] Rathnam RK, Elliott LK, Wall TF, Liu Y, Moghtaderi B. Differences in reactivity of pulverised coal in air (O2/N2) and oxy-fuel (O2/CO2) conditions. Fuel Process Technol 2009;90:797–802. [89] Brix J, Jensen PA, Jensen AD. Coal devolatilization and char conversion under suspension fired conditions in O2/N2 and O2/CO2 atmospheres. Fuel 2010;89:3373–80. [90] Gil MV, Riaza J, Álvarez L, Pevida C, Pis JJ, Rubiera F. Oxy-fuel combustion kinetics and morphology of coal chars obtained in N2 and CO2 atmospheres in an entrained flow reactor. Appl Energy 2012;91:67–74. [91] Gil MV, Riaza J, Álvarez L, Pevida C, Pis JJ, Rubiera F. Kinetics models for the oxy-fuel combustion of coal and coal/biomass blend chars obtained in N2 and CO2 atmospheres. Energy 2012;48:510–8. [92] Wang C, Zhang X, Liu Y, Che D. Pyrolysis and combustion characteristics of coals in oxyfuel combustion. Appl Energy 2012;97:264–73. [93] Yin C, Kær SK, Rosendahl L, Hvid SL. Co-firing straw with coal in a swirlstabilized dual-feed burner: modeling and experimental validation. Bioresour Technol 2010;101:4169–78. [94] Westbrook CK, Dryer FL. Simplified reaction-mechanisms for the oxidation of hydrocarbon fuels in flames. Combust Sci Technol 1981;27:31–43. [95] Westbrook CK, Dryer FL. Chemical kinetic modeling of hydrocarbon combustion. Prog Energy Combust Sci 1984;10:1–57. [96] Jones WP, Lindstedt RP. Global reaction schemes for hydrocarbon combustion. Combust Flame 1988;73:233–49. [97] Liu F, Guo H, Smallwood GJ, Gülder ÖL. The chemical effects of carbon dioxide as an additive in an ethylene diffusion flame: implications for soot and NOx formation. Combust Flame 2001;2001(125):778–87. [98] Glarborg P, Bentzen LLB. Chemical effects of a high CO2 concentration in oxyfuel combustion of methane. Energy Fuels 2008;22:291–6. [99] Brián M, Giménez-López J, Bilbao R, Alzueta MU. Effect of different concentration levels of CO2 and H2O on the oxidation of CO: experiments and modeling. Proc Combust Inst 2011;33:317–23. [100] Chen L, Ghoniem AF. Modeling CO2 chemical effects on CO formation in oxyfuel diffusion flames using detailed, quasi-global, and global reaction mechanisms. Combust Sci Technol 2014;186:829–48. [101] Le Cong T, Dagaut P. Experimental and detailed modeling study of the effect of water vapor on the kinetics of combustion of hydrogen and natural gas, impact on NOx. Energy Fuels 2009;23:725–34. [102] Andersen J, Rasmussen CL, Giselsson T, Glarborg P. Global combustion mechanisms for use in CFD modeling under oxy-fuel conditions. Energy Fuels 2009;23:1379–89.

C. Yin, J. Yan / Applied Energy 162 (2016) 742–762 [103] Marinov NM, Westbrook CK, Pitz WJ. Detailed and global chemical kinetics model for hydrogen. In: Chan SH, editor. Transport phenomena in combustion. Washington (DC): Taylor and Francis; 1996. [104] Hecht ES, Shaddix CR, Molina A, Haynes BS. Effect of CO2 gasification reaction on oxy-combustion of pulverized coal char. Proc Combust Inst 2011;33:1699–706. [105] Dhaneswar SR, Pisupati SV. Oxy-fuel combustion: the effect of coal rank and the role of char-CO2 reaction. Fuel Process Technol 2012;102:156–65. [106] Tolvanen H, Raiko R. An experimental study and numerical modeling of combusting two coal chars in a drop-tube reactor: a comparison between N2/ O2, CO2/O2, and N2/CO2/O2 atmospheres. Fuel 2014;124:190–201. [107] Maffei T, Khatami R, Pierucci S, Faravelli T, Ranzi E, Levendis YA. Experimental and modeling study of single coal particle combustion in O2/N2 and oxy-fuel (O2/CO2) atmospheres. Combust Flame 2013;160:2559–72. [108] Hecht ES, Shaddix CR, Geier M, Molina A, Haynes BS. Effect of CO2 and steam gasification reactions on the oxy-combustion of pulverized coal char. Combust Flame 2012;159:3437–47. [109] Brix J, Jensen PA, Jensen AD. Modeling char conversion under suspension fired conditions in O2/N2 and O2/CO2 atmospheres. Fuel 2011;90:2224–39. [110] Geier M, Shaddix CR, Davis KA, Shim HS. On the use of single-film models to describe the oxy-fuel combustion of pulverized coal char. Appl Energy 2012;93:675–9. [111] Hecht ES, Shaddix CR, Lighty JS. Analysis of the errors associated with typical pulverized coal char combustion modeling assumptions for oxy-fuel combustion. Combust Flame 2013;160:1499–509. [112] Tognotti L, Longwell JP, Sarofim AF. The products of the high temperature oxidation of a single char particle in an electrodynamic balance. Proc Combust Inst 1990;23:1207–13. [113] Gonzalo-Tirado C, Jiménez S. Detailed analysis of the CO oxidation chemistry around a coal char particle under conventional and oxy-fuel combustion conditions. Combust Flame 2015;162:478–85. [114] Kühnemuth D, Normann F, Andersson K, Johnsson F. On the carbon monoxide formation in oxy-fuel combustion – contribution by homogenous and heterogeneous reactions. Int J Greenh Gas Control 2014;25:31–41. [115] Yu J, Zhou K, Ou W. Mass transfer coefficients considering effects of steam in oxy-fuel combustion of coal char. Fuel 2013;111:48–56. [116] Zhou K, Yu J, Ou W. Effects of Stefan flow on ignition and heat transfer of a char particle in O2/CO2 atmospheres. Energy Fuels 2013;27:3454–9. [117] Glarborg P, Jensen AD, Johnsson JE. Fuel nitrogen conversion in solid fuel fired systems. Prog Energy Combust Sci 2003;29:89–113. [118] Andersson K, Normann F, Johnsson F, Leckner B. NO emission during oxy-fuel combustion of lignite. Ind Eng Chem Res 2008;47:1835–45. [119] Okazaki K, Ando T. NOx reduction mechanism in coal combustion with recycled CO2. Energy 1997;22:207–15. [120] Stadler H, Christ D, Habermehl M, Heil P, Kellermann A, Ohliger A, et al. Experimental investigation of NOx emissions in oxycoal combustion. Fuel 2011;90:1604–11. [121] Mendiara T, Glarborg P. Reburn chemistry in oxy-fuel combustion of methane. Energy Fuels 2009;23:3565–72. [122] Mendiara T, Glarborg P. Ammonia chemistry in oxy-fuel combustion of methane. Combust Flame 2009;156:1937–49. [123] Watanabe H, Yamamoto J, Okazaki K. NOx formation and reduction mechanisms in staged O2/CO2 combustion. Combust Flame 2011;158:1255–63. [124] Park DC, Day SJ, Nelson PF. Nitrogen release during reaction of coal char with O2, CO2, and H2O. Proc Combust Inst 2005;30:2169–75. [125] Aarna I, Suuberg EM. The role of carbon monoxide in the NO-carbon reaction. Energy Fuels 1999;13:1145–53. [126] López D, Calo J. The NO-carbon reaction: the influence of potassium and CO on reactivity and populations of oxygen surface complexes. Energy Fuels 2007;21:1872–7. [127] Sun S, Cao H, Chen H, Wang X, Qian J, Wall T. Experimental study of influence of temperature on fuel-N conversion and recycle NOx reduction in oxyfuel combustion. Proc Combust Inst 2011;33:1731–8. [128] Shaddix CR, Molina A. Fundamental investigation of NOx formation during oxy-fuel combustion of pulverized coal. Proc Combust Inst 2011;33:1723–30. [129] Wang G, Zander R, Costa M. Oxy-fuel combustion characteristics of pulverized-coal in a drop tube furnace. Fuel 2014;115:452–60. [130] Yin C, Rosendahl L, Kær SK. Grate-firing of biomass for heat and power production. Prog Energy Combust Sci 2008;34:725–54. [131] Hill SC, Smoot LD. Modeling of nitrogen oxides formation and destruction in combustion systems. Prog Energy Combust Sci 2000;26:417–58. [132] De Soete GG. Overall reaction rates of NO and N2 formation from fuel nitrogen. Proc Combust Inst 1975;15:1093–102. [133] Lockwood FC, Romo-Millares CA. Mathematical modeling of fuel–NO emissions from PF burners. J Inst Energy 1992;65:144–52. [134] Chui EH, Douglas MA, Tan Y. Modeling of oxy-fuel combustion for a western Canadian sub-bituminous coal. Fuel 2003;82:1201–10. [135] Chui EH, Majeski AJ, Douglas MA, Tan Y, Thambimuthu KV. Numerical investigation of oxy-coal combustion to evaluate burner and combustor design concepts. Energy 2004;29:1285–96. [136] AVL Fire CFD Solver v 8.5 manual. A-8020 Gras, Austria; 2008 . [137] Toporov D, Bocian P, Heil P, Kellermann A, Stadler H, Tschunko S, et al. Detailed investigation of a pulverized fuel swirl flame in CO2/O2 atmosphere. Combust Flame 2008;155:605–18.

761

[138] Chen L, Ghoniem AF. Simulation of oxy-coal combustion in a 100 kWth test facility using RANS and LES: a validation study. Energy Fuels 2012;26:4783–98. [139] Kangwanpongpan T, da Silva RC, Krautz HJ. Prediction of oxy-coal combustion through an optimized weighted sum of gray gases model. Energy 2012;41:244–51. [140] Cao H, Sun S, Lui Y, Wall TF. Computational fluid dynamics modeling of NOx reduction mechanism in oxy-fuel combustion. Energy Fuels 2010;24: 131–5. [141] Hu Y, Yan J. Numerical simulation of radiation intensity of oxy-coal combustion with flue gas recirculation. Int J Greenh Gas Control 2013;17:473–80. [142] Edge P, Gubba SR, Ma L, Porter P, Pourkashanian M, Williams A. LES modeling of air and oxy-fuel pulverised coal combustion – impact on flame properties. Proc Combust Inst 2011;33:2709–16. [143] Bhuiyan AA, Naser J. Numerical modeling of oxy fuel combustion, the effect of radiative and convective heat transfer and burnout. Fuel 2015;139:268–84. [144] Bhuiyan AA, Naser J. Computational modeling of co-firing of biomass with coal under oxy-fuel condition in a small scale furnace. Fuel 2015;143:455–66. [145] Gharebaghi M, Irons RMA, Ma L, Pourkashaniana M, Pranzitelli A. Large eddy simulation of oxy-coal combustion in an industrial combustion test facility. Int J Greenh Gas Control 2011;5S:S100–10. [146] Naredi P. Coal combustion in O2/CO2 medium for power generation: Numerical modeling of char burnout, NOx and CO emissions. PhD Thesis. The Pennsylvania State University; 2009. [147] Jovanovic R, Milewska A, Swiatkowski B, Goanta A, Spliethoff H. Numerical investigation of influence of homogeneous/heterogeneous ignition/combustion mechanisms on ignition point position during pulverized coal combustion in oxygen enriched and recycled flue gases atmosphere. Int J Heat Mass Transf 2011;54:921–31. [148] Zhang J, Dai B, Meng Y, Wu X, Zhang J, Zhang X, et al. Pilot-scale experimental and CFD modeling investigations of oxy-fuel combustion of Victorian brown coal. Fuel 2015;144:111–20. [149] Nikolopoulos N, Nikolopoulos A, Karampinis E, Grammelis P, Kakaras E. Numerical investigation of the oxy-fuel combustion in large scale boilers adopting the ECO-scrub technology. Fuel 2011;90:198–214. [150] Al-Abbas AH, Naser J, Dodds D. CFD modeling of air-fired and oxy-fuel combustion in a large-scale furnace at Loy Yang A brown coal power station. Fuel 2012;102:646–65. [151] Hu Y, Li H, Yan J. Numerical investigation of heat transfer characteristics in utility boilers of oxy-coal combustion. Appl Energy 2014;130:543–51. [152] Guo J, Liu Z, Wang P, Huang X, Li J, Xu P, et al. Numerical investigation on oxycombustion characteristics of a 200 MWe tangentially fired boiler. Fuel 2015;140:660–8. [153] Gao H, Runstedtler A, Majeski A, Yandon R, Zanganeh K, Shafeen A. Reducing the recycle flue gas rate of an oxy-fuel utility power boiler. Fuel 2015;140:578–89. [154] Andersson K. Characterization of oxy-fuel flames – their composition, temperature and radiation. Chalmers University of Technology. PhD Thesis, Göteborg; 2007. [155] Smart JP, Patel R, Riley GS. Oxy-fuel combustion of coal and biomass, the effect on radiative and convective heat transfer and burnout. Combust Flame 2010;157:2230–40. [156] Smart JP, O’Nions P, Riley GS. Radiation and convective heat transfer, and burnout in oxy-coal combustion. Fuel 2010;89:2468–76. [157] Galletti C, Coraggio G, Tognotti L. Numerical investigation of oxy-natural-gas combustion in a semi-industrial furnace: validation of CFD sub-models. Fuel 2013;109:445–60. [158] Muto M, Watanabe H, Kurose R, Komori S, Balusamy S, Hochgreb S. Largeeddy simulation of pulverized coal jet flame – effect of oxygen concentration on NOx formation. Fuel 2015;142:152–63. [159] Prieler R, Demuth M, Spoljaric D, Hochenauer C. Numerical investigation of the steady flamelet approach under different combustion environments. Fuel 2015;140:731–43. [160] Rosendahl LA, Yin C, Kær SK, Friborg K, Overgaard P. Physical characterization of biomass fuels prepared for suspension firing in utility boilers for CFD modeling. Biomass Bioenerg 2007;31:318–25. [161] Baxter L. Biomass-coal co-combustion: opportunity for affordable renewable energy. Fuel 2005;84:1295–302. [162] Roberts K, Yan Y, Carter RM. On-line sizing and velocity measurement of particles in pneumatic pipelines through digital imaging. J Phys: Conf Ser 2007;85:012019. [163] Yin C, Rosendahl L, Kær SK, Sørensen H. Modeling the motion of cylindrical particles in a nonuniform flow. Chem Eng Sci 2003;58:3489–98. [164] Yin C, Rosendahl L, Kær SK, Condra T. Use of numerical modeling in design for co-firing biomass in wall-fired burners. Chem Eng Sci 2004;59:3281–92. [165] Hu YK, Yan J. Oxy-fuel combustion for CO2 capture. In: Yan J, editor. Mitigation technologies. Handbook of clean energy systems, vol. 2. Wiley; 2015. [166] IPCC. Special report on carbon dioxide capture and storage. Cambridge: Cambridge University Press; 2005. [167] IEA. Technology roadmap – carbon capture and storage; 2009. [168] Shah K, Moghtaderi B, Zanganeh J, Wall T. Integration options for novel chemical looping air separation (ICLAS) process for oxygen production in oxy-fuel coal fired power plants. Fuel 2013;107:356–70.

762

C. Yin, J. Yan / Applied Energy 162 (2016) 742–762

[169] Aneke M, Wang M. Potential for improving the energy efficiency of cryogenic air separation unit (ASU) using binary heat recovery cycles. Appl Therm Eng 2015;81:223–31. [170] Castillo R. Thermodynamic analysis of a hard coal oxyfuel power plant with high temperature three-end membrane for air separation. Appl Energy 2011;88:1480–93. [171] Smith AR, Klosek J. A review of air separation technologies and their integration with energy conversion processes. Fuel Process Technol 2001;70:115–34. [172] Hu Y, Li X, Li H, Yan J. Peak and off-peak operations of the air separation unit in oxy-coal combustion power generation systems. Appl Energy 2013;112:747–54. [173] Arias B, Criado YA, Sanchez-Biezma A, Abanades JC. Oxy-fired fluidized bed combustors with a flexible power output using circulating solids for thermal energy storage. Appl Energy 2014;132:127–36. [174] Hu Y, Yan J. Characterization of flue gas in oxy-coal combustion processes for CO2 capture. Appl Energy 2012;90:113–21. [175] Nakayama S, Noguchi Y. Pulverized coal combustion in O2/CO2 mixtures on a power plant for CO2 recovery. Energ Convers Manage 1992;33:379–86. [176] Li H, Yan J, Yan J, Anheden M. Impurity impacts on the purification process in oxy-fuel combustion based CO2 capture and storage system. Appl Energy 2009;86:202–13. [177] Li H, Yan J. Evaluating cubic equations of state for calculation of vapour– liquid equilibrium of CO2 and CO2 mixtures for CO2 capture and storage processes. Appl Energy 2009;86:826–36. [178] Li H, Yan J. Impacts of equations of state (EOS) and impurities on the volume calculation of CO2 mixtures in the applications of CO2 capture and storage (CCS) processes. Appl Energy 2009;86:2760–70. [179] Li H, Ji X, Yan J. A new modification on RK EOS for gaseous CO2 and gaseous mixtures of CO2 and H2O. Int J Energ Res 2006;30:135–48. [180] Li H, Jakobsen JP, Wilhelmsen Ø, Yan J. PVTxy properties of CO2 mixtures relevant to CO2 conditioning and transport: review of available experimental data and theoretical models. Appl Energy 2011;88:3567–79. [181] Li H, Wilhelmsen Ø, Lv Y, Wang W, Yan J. Viscosity, thermal conductivity and diffusion coefficients of CO2 mixtures: review of experimental data and theoretical models. Int J Greenh Gas Control 2011;5:1119–39. [182] Yan J, Anheden M, Bernstone C, Liljemark S, Pettersson H, Li H, et al. Impacts of non-condensable components on CO2 compression/purification, pipeline transport and geological storage. In: 1st IEA oxyfuel combustion conference, Cottbus, Germany, September 8–11, 2009. [183] Wall T, Stanger R, Liu Y. Gas cleaning challenges for coal-fired oxy-fuel technology with carbon capture, and storage. Fuel 2013;108:85–90. [184] Park SK, Kim TS, Sohn JL, Lee YD. An integrated power generation system combining solid oxide fuel cell and oxy-fuel combustion for high performance and CO2 capture. Appl Energy 2011;88:1187–96.

[185] Park SK, Ahn JH, Kim TS. Performance evaluation of integrated gasification solid oxide fuel cell/gas turbine systems including carbon dioxide capture. Appl Energy 2011;88:2976–87. [186] Lyngfelt A. Chemical-looping combustion of solid fuels – status of development. Appl Energy 2014;113:1869–73. [187] Stanger R, Wall T, Spörl R, Paneru M, Grathwohl S, Weidmann M, et al. Oxyfuel combustion for CO2 capture in power plants. Int J Greenh Gas Control 2015;40:55–125. [188] Anheden M, Burchhardt U, Ecke H, Faber R, Jidinger O, Giering R, et al. Overview of operational experience and results from test activities in Vattenfall’s 30 MWth oxyfuel pilot plant in Schwarze Pumpe. Energy Procedia 2011;4:941–50. [189] National Coal Council. Fossil forward: revitalizing CCS bringing scale and speed to CCS deployment; 2015. [190] Global CCS Institute. The costs of CCS and other low-carbon technologies in the United States: 2015 update; 27 July 2015. [191] Davidson RM, Santos SO, Oxyfuel combustion of pulverised coal. IEA Greenhouse Gas R&D Programme (IEAGHG); 24 August 2010. [192] Borgert KJ, Rubin ES. Oxyfuel combustion: technical and economic considerations for the development of carbon capture from pulverized coal power plants. Energy Procedia 2013;37:1291–300. [193] Oboirien BO, North BC, Kleyn T. Techno-economic assessments of oxy-fuel technology for South African coal-fired power stations. Energy 2014;66:550–5. [194] Global CCS Institute. Toward a common method of cost estimation for CO2 capture and storage at fossil fuel power plants; 2013. [195] Zhang L, Jiao F, Binner E, Chen L, Bhattacharya S, Ninomiya Y, et al. Evolution of organically bound metals during coal combustion in air and O2/CO2 mixtures: a case study of Victorian brown coal. Proc Combust Inst 2011;33:2795–802. [196] Fryda L, Sobrino C, Glazer M, Bertrand C, Cieplik M. Study of ash deposition during coal combustion under oxyfuel conditions. Fuel 2012;92:308–17. [197] Li G, Li S, Dong M, Yao Q, Guo CY, Axelbaum RL. Comparison of particulate formation and ash deposition under oxy-fuel and conventional pulverized coal combustions. Fuel 2013;106:544–51. [198] Wen C, Xu M, Yu D, Sheng C, Wu H, Zhang P, et al. PM10 formation during the combustion of N2-char and CO2-char of Chinese coals. Proc Combust Inst 2013;34:2383–92. [199] Li W, Wang L, Qiao Y, Lin JY, Wang M, Chang L. Effect of atmosphere on the release behavior of alkali and alkaline earth metals during coal oxy-fuel combustion. Fuel 2015;139:164–70. [200] Morris WJ, Yu D, Wendt JOL. Soot, unburnt carbon and ultrafine particle emissions from air- and oxy-coal flames. Proc Combust Inst 2011; 33:3415–21.