Fate of Nitrogen during Fluidized Incineration of ... - ACS Publications

Article pubs.acs.org/EF

Fate of Nitrogen during Fluidized Incineration of Sewage Sludge. Estimation of NO and N2O Content in the Exhaust Gas Frederic Marias,*,† Abderhamane Benzaoui,† Jean Vaxelaire,† Franck Gelix,‡ and Francois Nicol‡ †

Laboratoire de Thermique Energétique et Procédés, Université de Pau et des Pays de l’Adour, 64000 Pau, France Veolia Environnement Recherche et Innovation, 78520 Limay, France

‡

S Supporting Information *

ABSTRACT: This work aims at studying the chemical mechanisms responsible for the formation and destruction of nitrogen oxides during combustion of sewage sludge in a fluidized bed. To do so, a mathematical model has been built in order to represent the relevant phenomena occurring within the reactor and estimate the nitrous emissions as a function of the operating parameters. An important modeling effort has been done in order to predict what becomes of nitrogen initially present in the sludge. The final model allows computing the local concentration of NO, N2O, NO2, HCN, and NH3 at every location of the bed. Numerical estimations of the model are compared to experimental results obtained at industrial scale in a fluidized bed combustor (for which thermal equilibrium is reached using extra gas). This comparison is very good, allowing one to use the model for design and optimization of such reactors.

1. INTRODUCTION The global generation of sewage sludge, which is a byproduct of wastewater treatment, is increasing continuously. Incineration of this material is a potential alternative for its disposal. It is commonly processed in fluidized beds either as a unique fuel1−7 or in co-combustion.8−11 This incineration is also of great interest since it might contribute a renewable source to our energy mix and lower carbon emission from fossil fuels, since incineration could provide heat for district heating. However, because of the high nitrogen content of the sludge, its combustion into fluidized beds leads to the production of a significant amount of nitrogen oxides, and particularly to the production of nitrous oxide (N2O).12−25 N2O is a toxic compound and a greenhouse gas (with a global warming potential which is 298 times that of carbon dioxide over 100 years); consequently, it is necessary to control and reduce its release into the atmosphere. Such a goal can be achieved using mathematical modeling. Indeed, using this technique, it can be understood how N2O is produced from the nitrogen held with the initial sludge, which is the relevant phenomenon that controls its final production and, hence, how the combustor could be operated in order to reduce N2O generation.26−30 That is the scope of this paper. A mathematical model that was first built to characterize fluidized bed combustion of municipal solid waste31−33 was updated to consider sewage sludge as the fuel. During this first update, processes like drying and pyrolysis were no more considered as instantaneous, and the overall model was set to a transient one.34 Finally, in a recent evolution of this model, a detailed reaction scheme allowing for the estimation of the fate of nitrogen during fluidized bed combustion was added. This new development of the model was achieved in collaboration with Veolia Environnement Company. Indeed, this company is highly invested in “in situ” nitrogen pollutant emission reduction in order to reduce the impact of sludge combustion on the environment. This final model was compared to industrial data obtained in a bubbling fluidized bed combustor with a nominal capacity of 2 t/h of © XXXX American Chemical Society

sludge with an initial moisture content of 75% (on a raw basis). This new development of the model as well as its application to this combustor is further developed in the present paper. The first part of the paper is devoted to the description of the final mathematical model that put into mathematical formalism the different phenomena occurring in such a fluidized bed combustor fueled with sewage sludge. The main assumptions on which the model relies, as well as the main phenomena taken into account by the model, are presented. Because this mathematical model needs to be solved before it might produce numerical results, and because this step can have a strong impact on the time required to obtain such results, the second part of the paper proposes insights on the solving procedure. A comparison between those results and the ones that were obtained using an industrial unit over a week of operation is presented in the third part of the paper. The variation of some of the operating parameters (fuel flow rate, extra-gas flow rate, primary and secondary air flow rates) is also discussed. Then, once the ability of the model to predict the behavior of the combustor has been demonstrated, it is used to show an in-depth analysis of the fate of the “fuel nitrogen” as it is partially converted into a gaseous species during pyrolysis and further reacted within the gaseous phases of the reactor. This analysis will constitute the fourth part of the work. Finally, a conclusion will be drawn.

2. MODELING Before entering the modeling section, some information on the system under consideration has to be given. The first one is relative to the nature of the fluidized bed. As it is depicted in Figure 1, the reactor which is considered here is of dense bed type, operating under bubbling mode. The feeding of the combustion air is split into primary and secondary air, respectively. Received: January 21, 2015 Revised: June 8, 2015

A

DOI: 10.1021/acs.energyfuels.5b01109 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels

to maintain the emulsion at incipient fluidization conditions. This description of the gas held in the dense part of the reactor is completed by the description of the freeboard, which is further decomposed into two plug flow reactors: • A first one representing the disengagement zone where sand might be thrown as the bubbles burst at the top of the dense bed, • A second one, which is supposed to be completely free of sand (post-combustion zone). The scheme of this overall “hydrodynamic” model of the gas phase is depicted in Figure 2. This first description of the gas held within the reactor requires the description of the particles that are present in the system. These are composed of two different categories. The first one represents the bed material which is supposed to be inert (from the chemical point of view). This “sand” might be present in the emulsion zone, in the disengagement zone, and in the buffer zone. In our modeling effort, this sand is characterized by an average temperature, computed according to specific heat transfer coefficient values36 and to a ratio of presence between the dense bed and the disengagement zone. The second category of particles is reactive particles. It represents the particles of sludge that decompose within the reactor. Each particle (which is further assumed to be spherical) is supposed to be composed at any time and any location of • free water • bound water • fresh organic material • char (arising from pyrolysis and that might be composed of carbon, hydrogen, oxygen, nitrogen, and sulfur) • inorganic material (corresponding to the ash remaining after complete combustion of the particle) The transport of these particles within the reactor is supposed to take place as a “diffusion-like” transport within the dense bed, whereas it is mainly described by a “convection-like” law within the freeboard (see Supporting Information). This transport takes into account elutriation of the particles as a function of their size and density (which is a function of the initial characteristics of the sludge to burn and of the thermal and chemical “history” of the particle within the reactor). Indeed, it is assumed that drying of the free water and combustion of char do affect only the external radius of the particle (shrinking core model), whereas drying of bound water and pyrolysis do affect solely its density. The simplified description of the overall model is provided into Figure 3. It has to be noticed that the coupling between the different phases (gas of the five considered regions, sand, and reactive particles of the dense bed or of the freeboard) are mainly taken into account through “source terms” characterizing heat and mass transfer associated with heterogeneous reaction and heat and mass transfer. For detailed information about the mathematical derivation of this “description”, interested readers should refer to refs 32−34 and to the Supporting Information. It has to be noticed also that each “region” of the overall model is not considered as a “steady-state” region but a transient one. The different equations describing the physical and chemical phenomena occurring within each of them includes a transient term representative of the accumulation of mass, species, and energy, respectively. 2.2. Pyrolysis of Sludge. Once the drying step of the particles that enter the reactor is finished, their temperature begins to rise again, and the particles start to decompose under the

Figure 1. Scheme of the combustor under consideration.

Both of them can be heated up and can be composed of air, enriched air, oxygen, or whatever gaseous species. Primary air is introduced at the bottom of the bed of material constituting the bed (sand, alumina, etc.) through a distributor. Sludge can be introduced in the reactor at any location (height above the distributor), either in the dense bed or within the freeboard. Extra gas can be added in the reactor at the freeboard of the reactor or in the dense bed or also at both locations. An extra sand supply ensures that the overall mass of bed material remains constant within the operation. 2.1. Hydrodynamics of the Fluidized Bed. The scope of this paper is not to give a complete description of the overall model that has been developed, but to focus on the main concepts of this model. Dealing with bubbling fluidized bed, the main and first aspect that has to be addressed is the description of the hydrodynamics of the bed. In accordance with previous work,31−34 the description of the dense part of the bed relies on the assumption formulated by Werther35 that a “film” exists between the bubbles conveying the main part of the fluidization gas within the bed and the emulsion part remaining at incipient fluidization conditions. This assumption led us to the description of the gas held within the dense bed of the reactor as a model composed of three parts: • a plug flow reactor, describing the bubbles of the bed • a continuous stirred tank reactor (CSTR) describing the gas held within the emulsion • a two-dimensional (2D) region (hereafter called the “buffer zone”) for which the thickness corresponds to Werther’s film, but where reactions, heat, and mass transfer are described as 2D processes. This region is submitted to ascending gas at incipient fluidization conditions and to a net flux of gas arising from the release of volatiles within the emulsion that must be extracted by the bubbles B


Article

Energy & Fuels

Figure 2. Scheme of the assemblage of models used to describe gaseous processes within the combustor.

Figure 3. Complete description of the different zones of the model and of their connections.

reaction scheme. No information is really given on the composition of the product of the different reactions involved in the two steps scheme. Hence, it is really difficult to introduce such information within a reactor model. It must also be considered that even though such information might be available, it would not allow that the overall balances on carbon, hydrogen, oxygen, nitrogen, sulfur, and chlorine to be satisfied. The procedure that was used to characterize this process in our modeling effort was thus the following:

effect of heat and without any need for an additional reactant. This is the pyrolysis step. This process has been extensively described in the literature.37−45 It has to be noticed that a lot of models were created to describe this phenomenon and the associated kinetics. For our purposes, the choice has been made to consider the pyrolysis step as a single and unique reaction so that the conservation of elements (C, H, O, N, S, and Cl) is ensured during this step. In the literature, it is often shown that this pyrolysis occurs into two successive steps, with the creation of an intermediate species. This description does probably represent with much more reality the exact processes that the one assuming a single reaction. However, the literature does not generally provide the full kinetics information on this two steps

• Assumption of a single and unique reaction leading (from the fresh and dry organic material of the sludge) to the production of two fractions: a solid one and a gaseous one C


Article

Energy & Fuels (a part of which might be condensed at ambient temperature). Such assumption was also used for complete reactor modeling in the literature.45−48 • Use of kinetics data to take into account the rate at which the fresh and dry organic material is decomposed into these two fractions. This information has been extracted from the work of Conesa et al.,49 who assumed a reaction order of 1.31 with respect to the organic material, an activation energy of 63.438 kJ/mol, and a pre-exponential factor of 79.114 (international system of units). • Estimation of the composition of the gaseous and solid fractions of the reaction. A specific pre-computational step was performed. More precisely, we chose to use the results obtained by Gomez Barea et al.39 to determine the composition of the volatiles that would satisfy the experiment performed by this author and that would in turn respect the balances of the elements (C, H, O, N, S, and Cl). Indeed, since the composition of the permanent gases (in terms of CH4, H2, CO, and CO2) was known from the work of Gomez Barea et al., and since specific assumptions were done concerning the element nitrogen, an inverse method was used to determine the composition of the volatiles and of the char of our pyrolysis model. More specifically, the objective function to minimize was that of the composition of the permanent gas, with the constraint of fulfilling the elemental balances, when the particles were subjected to any thermal history. Given the composition of the organic and fresh material of the sludge (as it was provided by the industrial partner, see Table 1),

Table 3. Composition of the Solid Product of the Pyrolysis Step mass content (%) C N

Table 4. Composition of the Gaseous Product of the Pyrolysis Step mass fraction (−) CH4 H2 H2O CO CO2 C6H6

k 2gas 1 O2 ⎯→ ⎯ H 2O 2

(R2)

CO +

k 3gas 1 O2 ⎯→ ⎯ CO2 2

(R3) (R4)

CO2 + H 2 ⎯→ ⎯ CO + H 2O

(R5)

k6gas

C6H6 + 3O2 ⎯→ ⎯ 6CO + 3H 2

(R6)

k 7gas

O + N2 ⎯→ ⎯ NO + N

(R7)

k 8gas

N + NO ⎯→ ⎯ O + N2

(R8)

k 9gas

N + O2 ⎯→ ⎯ O + NO

(R9)

gas k10

O + NO ⎯→ ⎯ N + O2

(R10)

k 3gas 3 O2 ⎯→ ⎯ SO2 + H 2O (R11) 2 The kinetic data associated with the gaseous reaction scheme are further detailed in Table 5. The reactions ranging from R7 to R10 are the ones involved in the “thermal NO” Zeldovich’s mechanism. It is assumed in this one that the oxygen atom concentration is computed according to equilibrium O2↔2O, while the local concentration of nitrogen atom is evaluated according to quasi-steady-state approximation. It has to be noticed that the reactions involving nitrogen have not been presented yet and will be presented below. Only the “thermal NO” route has been added up to this point. 2.4. Heterogeneous Reactions. Once the pyrolysis has occurred, the solid fraction containing mainly (but not only) carbon (the char) needs to be oxidized so that the combustion of sludge can be complete. Depending upon the local concentration of oxygen and the local temperature, this solid fraction can be converted into carbon monoxide and carbon dioxide according to a shrinking core model. The reaction is expected to occur at the external surface of the considered particle.

H 2S +

Provided by the industrial partner.

it was possible, according to the above-mentioned precomputation step, to estimate the composition of the products of the reaction. Such estimation is provided in Tables 2, 3, and 4, respectively. It must also be said that all Table 2. Products of the Pyrolysis Reaction of the Organic Material mass fraction (%) 10.5 89.5

the sulfur and chlorine initially present in the sludge was supposed to be completely devolatilized from the solid particles as H2S and HCl, respectively. 2.3. Reactions in the Gas Phase. Once released from the pyrolysis step and depending upon the local composition of the gas and its temperature, the volatiles might undergo several reactions. This set of reactions is given here: k1gas 1 CH4 + O2 ⎯→ ⎯ CO + 2H 2 2

9.205 × 10−4 1.051 × 10−2 6.303 × 10−3 5.965 × 10−2 0.272

H2 +

k5gas

58.97 8.21 26.77 5.44 0.53 0.08

solid gas

mass fraction (−) HCl HCN H2S NH3 C2H4

CO + H 2O ⎯→ ⎯ CO2 + H 2

mass content (%, dry, ash-free basis)

a

2.996 × 10−2 2.854 × 10−3 0.1288025 0.129 0.151 0.209

k4gas

Table 1. Ultimate Analysis of the Initial Sludgea C H O N S Cl

94.56 5.44

C + ΛO2 → (2 − 2Λ)CO + (2Λ − 1)CO2

Λ = 0.87

The reaction rate associated with this reaction is computed assuming that both the kinetics of the reaction and the external

(R1) D


Article

Energy & Fuels

Table 5. Chemical Reaction Rates of the Homogenous Reaction Included in the Model (Molar Concentration in mol·m−3) reaction

reaction rate (mol·m−3·s−1)

rate constant

reference

1

r1 =

100k1gas[CH4]0.7 [O2 ]0.8

⎛ − 202641 ⎞ ⎟ = 1.58 × 1010 exp⎜ ⎝ RT ⎠

Dryer and Glassman50

2

r2 = 0.001k 2gas[H 2][O2 ]

⎛ − 125525 ⎞ ⎟ k 2gas = 1.08 × 1010 exp⎜ ⎝ RT ⎠

Kerinin and Shifrin51

3

r3 = k 3gas[CO][H 2O]0.5 [O2 ]0.25

⎛ − 180032 ⎞ ⎟ k 3gas = 1.78 × 1010 exp⎜ ⎝ RT ⎠

Dryer et Glassman50

4

r4 = 0.1k4gas[CO][H 2O]

⎛ − 12560 ⎞ ⎟ k4gas = 2.778 exp⎜ ⎝ RT ⎠

Biba et al.52

5

r5 = 0.1k5gas[CO2 ][H 2]

⎡ ⎛ 3.473 × 104 ⎞⎤ k5gas = k4gas/⎢0.022 exp⎜ ⎟⎥ RT ⎠⎦ ⎝ ⎣

Biba et al.52

6

r6 = k6gas[C6H6][O2 ]

⎛ − 202641 ⎞ ⎟ k6gas = 1.58 × 1012 exp⎜ ⎝ RT ⎠

Srinivasan et al.53

7

r7 = k 7gas[O][N2]

⎛ − 38370 ⎞ ⎟ k 7gas = 1.8 × 108 exp⎜ ⎝ T ⎠

Glassman54

8

r8 = k 8gas[N][NO]

⎛ − 425 ⎞ ⎟ k 8gas = 3.8 × 107 exp⎜ ⎝ T ⎠

Glassman54

9

r9 = k 9gas[N][O2 ]

⎛ − 4680 ⎞ ⎟ k 9gas = 1.8 × 104T exp⎜ ⎝ T ⎠

Glassman54

gas r10 = k10 [O][NO]

⎛ − 20820 ⎞ gas ⎟ k10 = 3.8 × 103T exp⎜ ⎝ T ⎠

Glassman54

10

k1gas

3. SOLVING The whole model is composed of partial differential equations (PDEs) (space and time dependence for the gas held in the buffer zone, in the bubble zone, in the disengagement zone, or in the post-combustion zone; space and time dependence for the reactive particles held in the dense bed or in the free board), ordinary differential equations (ODEs) (time dependence for the gas held within the emulsion, or for the sand held in the reactor), and algebraic equations (incipient fluidization velocity, heat and mass transfer coefficients, for example). The choice that has been made is to convert the PDEs into ordinary differential ones (method of lines) using the finite volume method59 (with upstream interpolation) for spatial discretization. Given this first “pre-processing” of the PDEs, the resulting system is a mix of ordinary differential and algebraic equations. The following notation is used to explain the mathematical solving: NSG, number of gaseous species taken into account in the model (NSG = 17 in including O2, N2, CH4, H2, H2O, CO, CO2, C6H6, HCl, N2O, SO2, SO3, HCN, H2S, NH3, NO, C2H4) NSS, number of components used to describe the particles (NSS = 8 including free water, bound water, fresh organic material, inorganic material, carbon in the char, nitrogen in the char, sulfur in the char, intermediate) NY, number of cells used to discretize the equations in the “lateral” size of the buffer zone (NY = 4 for this study) NZ, number of cells used to discretize the equations in the “height” size of the buffer zone and of the bubble zone (also for the reactive particles held within the dense bed, NZ = 4 for this study NZD, number of cells used to discretize the disengagement zone (NZD = 4 for this study) NZP, number of cells used to discretize the postcombustion zone (NZP = 4 for this study) The overall system to be solved is of the type

heat transfer of oxygen at the external surface of the particle can be the limiting step. 31−34 The kinetic data arise from the work of Sriramulu et al.,55 while the estimation of the mass transfer coefficient of oxygen to the external surface of the particle is issued from the work of Prins.36 Once again, the fate of the nitrogen held within the solid residue of pyrolysis is not developed here, but it will be considered in the next section. 2.5. Fate of Nitrogen. According to previous work,15 it has been assumed that the initial nitrogen content of the fresh and dry organic matter of the sludge was released during the pyrolysis into two fractions: one in the solid phase, and one in the gaseous phase (volatiles). Because we had no information on the part that was kept within the solid residue of pyrolysis, we assumed that the fraction of nitrogen in the solid residue was equal to the one initially present in the dry organic content of the sludge (5.44%, see Tables 1 and 3). This assumption might be validated by the recent work of Cao et al.56 that showed that the pyrolysis conditions had a very low influence on the molar N/C in the pyrolysis residue. Also, dealing with the fact that nitrogen could be released as volatiles either as HCN or NH3, the work of Liu and Gibbs57 that assumed that the molar ratio of NH3 to HCN released during pyrolysis was equal to 9 was also used in the frame of this study. This is an important assumption that could be confirmed using a pyrolysis experiment in similar conditions with the same raw material. These two data (N/C and NH3/HCN ratios) were added as constraints in the inverse method that was used to estimate the complete composition of the products of pyrolysis (see section 2.2). Dealing with the becoming of both fractions, the work that was done by Sung58 was used to take into account both nitrogen present in the char and released as HCN and NH3, respectively. A previous study has also shown that this mechanism was suitable for our purposes.15 Table 6 sums up the reactions that were taken into account and the associated kinetics.

M (t , y ) E

dy = f (t , y ) dt DOI: 10.1021/acs.energyfuels.5b01109 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels Table 6. Reactions and Associated Rates Included in the Model for Nitrogen-Containing Species58 11 12 13 14 15 16 17 18 19

20

21 22

reaction

rate

rate unit

phase

1 N2O → NO + N2 2

r11 = k11[N2O]

rate constant

⎛ − 29290 ⎞ ⎟ k11 = 8.013 × 108 exp⎜ ⎝ T ⎠

mol·m−3·s−1

homo

N2O → N2 +

1 O2 2

r12 = k12[N2O]

⎛ − 27433 ⎞ ⎟ k12 = 6.886 × 109 exp⎜ ⎝ T ⎠

mol·m−3·s−1

homo

NH3 + O2 →

1 3 N2O + H 2O 2 2

r13 = k13[NH3][O2 ]

⎛ − 50719 ⎞ ⎟ k13 = 2.597 × 1013 exp⎜ ⎝ T ⎠

ppm·s−1

homo

5 3 O2 → NO + H 2O 4 2

r14 = k14[NH3][O2 ]

⎛ − 34765 ⎞ ⎟ k14 = 5.797 × 108 exp⎜ ⎝ T ⎠

ppm·s−1

homo

r15 = k15[NH3][NO]

⎛ − 29378 ⎞ ⎟ k15 = 2.234 × 109 exp⎜ ⎝ T ⎠

ppm·s−1

homo

NH3 +

2 5 NH3 + NO → N2 + H 2O 3 6

HCN +

3 1 1 O2 → N2O + CO2 + H 2O 2 2 2

r16 = k16[HCN][O2 ]

⎛ − 37639 ⎞ ⎟ k16 = 7.983 × 108 exp⎜ ⎝ T ⎠

ppm·s−1

homo

HCN +

7 1 O2 → NO + CO2 + H 2O 4 2

r17 = k17[HCN][O2 ]

⎛ − 14960 ⎞ ⎟ k17 = 1.610 × 101 exp⎜ ⎝ T ⎠

ppm·s−1

homo

HCN +

5 1 1 O2 → N2 + CO2 + H 2O 4 2 2

r18 = k18[HCN][O2 ]

⎛ − 12632 ⎞ ⎟ k18 = 1.172 × 101 exp⎜ ⎝ T ⎠

ppm·s−1

homo

dmNO dm ⎡ m ⎤ = − char ⎢ N ⎥ dt dt ⎣ mchar ⎦

kinetic and external mass transfer limited

kg·s−1

hetero

⎡ m ⎤ r20 = k 20[NO]⎢ N ⎥ ⎣ mchar ⎦

k 20 = 27

kg·s−1

hetero

[NO]

⎛ − 16900 ⎞ ⎟ k 21 = 275T exp⎜ ⎝ T ⎠

mol·m−3·s−1

hetero

[N2O]

⎛ − 13900 ⎞ ⎟ k 22 = 267T exp⎜ ⎝ T ⎠

mol·m−3·s−1

hetero

char − N +

1 O2 → NO 2

char − N + NO → N2O

NO + C(s) →

1 N2 + CO 2

N2O + C(s) → N2 + CO

r21 = k 21

r22 = k 22

mC 2 4πrpart

mC 2 4πrpart

where y is the vector containing the state variables (which now includes the variables at the nodes of the computational grid), M is the mass matrix, which is singular because of the occurrence of algebraic equations within the overall model. The size of the system is NEQ = 20 + (NSG+4)*(1+NY*NZ+NZ+1+NZD+ NZP)+2+NY*NZ+NZD+NZ+NSS*NZ+6*NZ+NSS*(NZD +NZP)+8*(NZD+NZP)+1+NS*(NZ+NZD+NZP) (NEQ = 1065), with a total number of ordinary differential equations NEQODE = (NSG+1)*(1+NZ+NY*NZ+NZD+NZP)+1+ NZ*(NSS+2)+(NZD+NZP)*(NSS+2) (NEQODE = 643), and a total number of algebraic equations NEQAE = 20 + 3* (1+NZ+NY*NZ+NZD+NZP)+(NSG+4)+(1+NY*NZ+ NZD)+NZ*(5+NSS)+(NZD+NZP)*(NSS+6)+1 (NEQAE = 422). This system has been solved in our case using a DASSL method,60,61 and more specifically the DisCO library62 was used for our integration purpose. To reduce the computational time and improve convergence rate, the Jacobian of f and M mass matrix (involved at the corrector step during the solving) was computed analytically, using the sparse nature of the convergence operator. Since the overall model is of the transient type, a specific scenario was elaborated to obtain steady state information from a given a set of operating conditions. The above-mentioned scenario corresponds to the startup of the unit.

be given. Information on the value of the operating parameters that were modified is provided in a second step. Then, the comparison between industrial data and numerical predictions is shown and discussed. 4.1. Industrial Unit. It is of the bubbling bed type. Figure 1 gives a scheme of such a unit. The complete details of the geometry of the unit are confidential and cannot be given to readers of this paper. Nevertheless, these data have been introduced in the overall model to fit the real case. The main characteristics of the process are as follow: • equivalent diameter of the reactor (considered as a cylinder) = 3.22 m • total height of the equivalent reactor = 11.5 m • total mass of sand held within the reactor during operation = 10 t • diameter of the sand = 800 μm • nature of the extra gas (for dense bed and reactor) = methane • nominal sludge f low rate = 2 t/h • proximate analysis of the sludge being burnt (on a raw basis): ◦ moisture = 75.40% ◦ organic material = 20.44% ◦ inorganic material = 4.16% • lower heating value of the sludge being burnt (raw basis) = 3.8 MJ·kg−1 • ammonia held in the moisture content (mass basis) = 0.309% (it has to be noticed that this ammonia part is released as NH3 in the model during the drying of particles)

4. VALIDATION OF THE MODEL This section of the paper aims at providing the reader information about the industrial unit that was considered to get some information on the influence of the process operating parameters on the nitrogenous emissions at the output of the combustor. In a first step, some insights on the industrial unit will F


Article

Energy & Fuels Table 7. Values of the Operating Parameters during the Seven Runs Considered run no. sludge flow rate (t/h) flow rate of freeboard extra gas (Nm3/h) flow rate of dense bed extra gas (Nm3/h) flow rate of primary air (Nm3/h) flow rate of secondary air (Nm3/h) temperature of primary air (°C) average temperature of the bed (°C)

1

2

3

4

5

6

7

2.38 0.26 0.00 5334 54.9 517 707

2.41 0.52 1.27 5207 231.0 540 705

1.82 0.13 0.00 4932 0.0 503 734

2.11 0.45 5.27 5260 97.5 418 702

2.24 0.91 5.67 5301 156.0 519 738

2.20 0.15 1.40 5298 256.0 532 764

2.18 0.68 2.45 5034 134.0 534 705

The different samplings that were performed on that unit to measure the composition of the gas leaving the furnace was done at the air flume. It must be noticed here that this location for sampling was chosen to analyze the working of the combustor. It must be emphasized here that the composition of the gas that will be presented in the following section, is not the composition of the gas that is released to the atmosphere. Indeed, thanks to the flue gas treatment unit, this composition is adjusted to respect the law in force at the location of the plant. 4.2. Operating Conditions. Series of experimental tests were carried out, over a week, with the above-mentioned industrial unit. For this experimental investigation the following working parameters were modified: • sludge flow rate • flow rate of extra gas to the freeboard • flow rate of extra gas to the bed • flow rate of primary air (fluidizing air) • flow rate of secondary air (freeboard air) • temperature of the primary air Table 7 sums up the evolution of these parameters, which were carried out in seven runs. 4.3. Numerical vs Industrial. To evaluate the ability of the model to properly describe the fate of nitrogen within the combustor, the experimental and simulated compositions of the gas leaving the reactor are studied as a function of the operating conditions. As the full stationary state of the industrial unit could not be completely reached (thermal inertia of the refractory lining of the bed, slight modifications of the operating parameters around the values reported in Table 7 to ensure operation within the expected range), the different thermal levels computed by the model did not exactly match the industrial data. Hence, since the conversion of nitrogen within the reactor strongly depends on these thermal levels, the temperature of the sand was arbitrarily fixed, in the model, to the value of the average temperature of the bed measured in the industrial device. Although this could be considered as an “unsatisfactory” use of the model, it can however show the ability of the model to correctly predict the nitrogen conversion once the temperature of the sand is properly estimated. The composition of the gas leaving the industrial reactor was measured continuously using an FTIR analyzer and a paramagnetic oxygen analyzer, respectively. The gas was sampled within the air flume of the reactor. It must be noticed here that this sample point is located before the flue gas treatment process and that, during the different runs of the campaign, the SNCR used in the industrial unit to reduce the NO emissions was switched off. The duration of this mode operation was kept compatible with the regulation law in force at the industrial plant. Figure 4 shows the evolution of the outlet temperature of the gas as a function of the different test runs. The numerical results were compared with two available industrial data: the

Figure 4. Value of the temperature at the outlet of the combustor during the test runs: industrial at the top of the bed and at the air flume, and numerical at the outlet of the equivalent reactor.

temperature at the top of the combustor and the air flume temperature. It can be seen in Figure 4 that the model fits rather well the thermal levels measured on the industrial unit. The small discrepancies that appear are associated with the thermal inertia of the system. Small variations might also occur between the recorded values of the operating parameters and the values actually applied to the process. Figure 5 compares the predicted composition of the gas and of the measured one in terms of volumetric fraction of oxygen (on a dry basis), water, and carbon dioxide (both of them on a raw basis). This comparison is also shown for the HCl concentration on the flue gas. (The reader is reminded that this HCl content was measured before the gas enters the flue gas treatment unit.) Thanks to this figure, it can be concluded that the model accurately predicts the conversion of carbon, oxygen, and hydrogen originating from the sludge (moisture and organic content) and from the combustive air, respectively. The very little discrepancies between simulated and measured data might be associated with fluctuations of the composition of the sludge. Figure 5 also shows the estimation of the outlet concentration in N2O and NO which are compared to the value measured in the industrial device (it is reminded here that these data were collected while the SNCR of the unit was switched off). Dealing with NO estimation, it can be concluded that the computed results match well with industrial data (except for the test runs 4 and 5). Dealing with the estimation of the N2O concentration in the flue gas, the agreement between simulated and industrial G


Article

Energy & Fuels

Figure 5. Value of the composition of the gas during the test runs: industrial at air flume before the flue gas treatment unit and with the SNCR unit switched off, and numerical at the outlet of the equivalent reactor).

with inexact value of the working parameters; see above) may also impact the results. An illustration of this statement is provided in Figure 6, which shows the evolution of the relative error on the value of the concentration of the nitrogen containing

data is less satisfactory. The observed discrepancies (with both species NO and N2O) might be associated with small variation in the composition of the sludge. An inexact prediction of the temperature of the flume temperature (which might be associated H


Article

Energy & Fuels

nitrogen (as an element) entering the combustor is ṁ N = 8.72 × 10−3 kg·s−1 (with 14.72% and 85.28% in the moisture and in the organic fraction, respectively). Some of this incoming rate leaves the combustor with the withdrawal of the ashes at the bottom of the reactor. According to the results of the model, this fraction represents 2.14% of the incoming mass flow rate. Dealing with the fraction elutriated at the top of the bed, the model predicts that there is no nitrogen held within the solid particles leaving the combustor at this point, and hence these particles are solely composed of inorganic material (ashes). This means that 97.86% of the nitrogen entering the reactor is released and converted within the reactor. This is the most important information. The model also allows computing the fate of this nitrogen released in the combustor. Indeed, thanks to the numerical tool, it can be evaluated that, from the nitrogen released in the reactor, 15.00% are released as NH3 during the drying step, while 8.90% and 76.10% are released respectively as char-bound nitrogen and volatiles (with a ratio of NH3 to HCN of 9 on a molar basis, see section 2.5). This also means that 91.10% of the nitrogen released in the bed is first released in the gas phase. Another result which is also brought by the model is the location at which drying and pyrolysis takes place. It can be stated that 99.98% of the nitrogen released during the drying and pyrolysis step is released in the dense bed itself, while 0.02% is released in the disengagement zone, and none is released in the post-combustion zone. 5.2. Secondary Conversion of Nitrogen. Thanks to reactions 11−18 (Table 6), the nitrogen released as gaseous species is converted in a homogeneous way, while some of this fraction might react with the carbon of the char in a heterogeneous way thanks to reactions 20−22 (Table 6). The nitrogen released as char bound nitrogen during the pyrolysis step might also be heterogeneously converted thanks to reactions 19 and 20 (Table 6). The aim of the present subsection is to give insights on these secondary processes. 5.2.1. Homogeneous Reactions. Dealing with the homogeneous conversion, the results are presented in terms of fields of homogeneous net reaction rate (Rk = ∑NREAC vi,kri, where vi,k is i=1 the stoichiometric coefficient of species k in reaction i and mass concentration in the modeled reactor. What is meant by “fields” can be understood thanks to Figure 7, which represents the 2D profiles of temperature within each sub-part of the reactor assemblage. This is a way to represent the evolution of the data computed by the model assuming that all the bubbles of the dense bed cross it in a central core surrounded by the buffer zone (the width of which has been magnified by a factor 25 in the figure), the remaining part of the dense bed being the emulsion. Of course, it does not correspond to what happen in real reactors, this is only a way to represent the results of our model. It can be seen noticed that the lowest levels of temperature are obtained within the emulsion and the buffer part of the dense bed. Because some combustion reaction of the volatiles released during pyrolysis occur within the bubble zone (these volatiles are transferred in the bubble zone to keep the emulsion at incipient fluidization condition) where oxygen is present (the main part of the primary air crosses the bed as bubbles), the thermal level in the bubbles is higher than in the emulsion and buffer zones. This is also the reason why the highest temperatures are reached at the bottom of the disengagement zone where the main part of the combustion of the above-mentioned volatiles takes place. Because of the sand that is held within this region and that returns to the dense bed, some energy from the disengagement region returns to the dense bed, resulting in a slow decrease of

Figure 6. Correlation between the error obtained on the estimation of the nitrogen containing species and on the error on the estimation of the temperature of the air flume.

species as a function of the relative error on the temperature of the air flume. This figure shows that the error on the nitrogen containing species might be correlated to the error on the temperature. Finally, Figure 5 also presents the estimation of the SO2 concentration in the flue gas (before this component is removed from the gas released at the atmosphere in the flue gas treatment unit) compared to the measured value. It shows that the model over predicts systematically the concentration measured on the industrial device. The discrepancy might be associated with the fact that either the initial sulfur content of the sludge is not known. Moreover, all the sulfur is not released as volatiles during the pyrolysis step, and a fraction remains in the solid residue, as it is the case for nitrogen. According to the different comparisons that were provided in this section, we consider that the mathematical model is able to fairly predict the behavior of the industrial unit. Some discrepancies can be highlighted, but it has to be noticed that the experimental data were directly measured on an industrial unit and might suffer from uncertainties.

5. IN-DEPTH ANALYSIS OF THE FATE OF NITROGEN It is now proposed to give insights on the becoming of nitrogen once it is fed to the combustor. The data that were used to achieve this objectives are issued from the test run 2 (Table 7), since in this case the results predicted by the model were in very good agreement with the measurements on the industrial unit. In a first subsection, information on the fate of nitrogen during drying and pyrolysis will be given while in a second subsection, we will focus on its transformation by heterogeneous reactions and homogeneous ones. 5.1. Conversion of Nitrogen during Drying and Pyrolysis. As it has been stated in the previous parts of the paper, nitrogen is contained in the sludge entering the fluidized bed in two fractions: • nitrogen held in the “solid” organic fraction of the sludge • nitrogen held in the moisture content of the sludge (as NH4+) and released as NH3 during drying. So, given the mass flow rate of sludge entering the reactor and its composition, it can be concluded that the mass flow rate of I


Article

Energy & Fuels

Figure 8. Average 1D profiles of temperature along the height of the reactor (gas and solid particles).

Figure 10 shows the net reaction rate for NO (top left) and its concentration within the reactor (top right). Unlike HCN and NH3, the reaction rates for NO are positives, what translates the fact that this species is actually produced in each part of the reactor, although it is involved as a reactant in reaction 15. Most of its production occurs within the bubble zone, which is quite logical considering that both NH3 and HCN are oxidized into NO at this place. To a lower extent, the production of NO is also observed at the bottom of the disengagement zone and remains nearly zero after this point, which results in the concentration of this species being quite constant in the remaining part of the reactor. Dealing with N2O, the net reaction rate is positive within the dense bed and becomes negative at the freeboard of the bed. The highest production is observed at the top of the dense bed, and the highest consumption is observed at the bottom of the disengagement section. This result shows once again the importance of this section in the global conversion of nitrogen within the reactor. From the bottom of the disengagement, the concentration of N2O slowly and continuously decreases mainly under the effect of cracking (reaction 12). Reaction 11 involving NO does not significantly occur since the concentration of NO remains constant in the post-combustion. Figure 11 show the average 1D profiles of nitrogen containing species. The conclusions that were raised in the preceding paragraphs allow justifying the shape of the curves represented on these two figures. 5.2.2. Heterogeneous Reactions. Dealing with heterogeneous reactions 19−22, nitrogen can be either extracted from the char to form NO and N2O (reactions 19 and 20) or reduced into N2 by reaction with carbon of the char (reactions 21 and 22). The first result that comes from the model is that the whole part of the heterogeneous reaction occurs within the dense bed. Indeed, heterogeneous reactions neither occur in the disengagement nor in the post-combustion region. Dealing with the “extraction” of nitrogen from the char by reactions 19 and 22, the total amount of nitrogen which is converted corresponds to the production of nitrogen into char by pyrolysis reaction (which corresponds to 8.90% of the total amount of nitrogen released in the combustor by pyrolysis and drying steps). The information which can be further provided here is relative to the nature of the species that are produced by

Figure 7. Fields of temperature within the reactor.

the temperature in the top of the disengagement zone. The temperature remains almost constant in the post-combustion region which is supposed to be perfectly insulated. To complete this description of the distribution of the temperature, Figure 8 proposes an average 1D profile along the height of the reactor. The average values have been obtained in the dense bed thanks to an averaging procedure that takes into account the volume of each region involved. Figure 9 shows the net reaction rate for HCN (top left) and its concentration (top right) within the reactor. HCN (which is quite wholly released in the dense bed) appears to be mainly converted into the bubble zone, and to a lower extent in the emulsion and buffer zones. The reason for this higher reactivity in the bubbles is associated with the fact that in the bubble zone, both the oxygen content (HCN transformation involves oxygen; see reactions 16−18) and the temperature are higher. It can be also concluded that the consumption of HCN is finished in the disengagement zone and that it no more exists in the postcombustion region. Things are slightly different for NH3. Indeed, due to the kinetics of reactions 13−15 and because NH3 also reacts with NO, the highest reaction rates for NH3 appears in the disengagement region. However, it can be also concluded that this species is wholly consumed and no more exist in the postcombustion region, and hence at the output of the reactor. J


Article

Energy & Fuels

Figure 9. Fields of reaction and mass concentration of HCN and NH3 within the reactor.

indicate an order of magnitude. This part should not be considered in the overall nitrogen balance, since gaseous nitrogen is transformed into gaseous nitrogen by these heterogeneous reactions). To conclude with this reduction of nitrogen, it can be said that in the total amount which is reduced, 94.40% is reduced from NO to N2 and 5.60% from N2O to N2. 5.2.3. Global Conversion. To conclude this in-depth analysis of the fate of nitrogen during fluidized bed combustion of sludge, one can wonder about the final becoming of the nitrogen that

the heterogeneous reaction which might be NO (under the effect of reaction 19) or N2O (thanks to reaction 20). Indeed, according to the results of the model, these species are produced quite equally by these heterogeneous reactions (47.68% of the nitrogen “extracted” from the char is released as NO and 52.32% as N2O). Dealing with the reduction of nitrogen containing species by reduction with carbon held within the char, the total amount of nitrogen consumed is 4.11 × 10−4 kg·s−1. It represents 4.71% of the nitrogen entering the reactor (this value is just given to K


Article

Energy & Fuels

Figure 10. Fields of reaction and mass concentration of NO and N2O within the reactor.

These last results are very interesting since they show that the fluidized bed combustor allows for conversion of 94.15% of the nitrogen held in the sludge and entering the combustor as atmospheric nitrogen.

enters the reactor. As it has been already stated, none of the incoming nitrogen leaves the reactor as HCN or NH3 but as NO, N2O, and N2 or as ashes. It is now quite important to quantify how the incoming nitrogen is distributed over these three gaseous species and ashes: • ratio of the incoming nitrogen released as NO: 2.12% • ratio of the incoming nitrogen released as N2O: 1.59% • ratio of the incoming nitrogen released as N2: 94.15% • ratio of the incoming nitrogen released in the ashes: 2.14%

6. CONCLUSIONS This paper has presented some very recent work that has been done one the modeling of nitrogen conversion within a fluidized bed devoted to sludge combustion. It provides results that were L


Article

Energy & Fuels

Figure 11. Average 1D profiles of mass concentration of NH3 and HCN (left) and NO and N2O (right) along the height of the reactor.

■

obtained through collaboration between an industrial partner (Veolia Environnement Company) and an academic laboratory (LaTEP, Université de Pau et des Pays de l’Adour). The main goal of this collaboration was to build a model allowing description of the relevant phenomena occurring within such a combustor. This model, considers the reactor as an assemblage of several zones, including a 2D one relating the transfer and reaction phenomena occurring within the surrounding of the bubbles crossing the dense bed. It also takes into account the transport of the reactive particles of sludge during ongoing drying, pyrolysis, and heterogeneous reactions. All of these reactions are included in the model with their respective kinetic data, which also ensure conservation of the elements involved in the transformation (carbon, hydrogen, oxygen, nitrogen, sulfur, and chlorine). The existing model has been updated to account for the specificity of nitrogen transformation during the overall process, and more particularly during pyrolysis and further conversion into NO and N2O. The ability of the model to properly describe this transformation has been proven thanks to comparison with data obtained from an industrial operating unit. The model has been also used to give a very accurate description of the fate of nitrogen containing species during the combustion of sludge. Indeed, several analyses have led us to a better estimation of the respective amounts of nitrogen that are converted during each process (drying, pyrolysis, secondary homogeneous, and heterogeneous transformation) at every location of the reactor. Such analyses provide a better understanding on how this “pollutant” is mainly converted to atmospheric nitrogen (94.15% of the nitrogen entering the reactor with sludge leave it as N2). It is thus a promising tool for optimization of the daily working of such units.

■

(1) Werther, J.; Ogada, T. Sewage Sludge Combustion. Prog. Energy Combust. Sci. 1999, 25, 55−116. (2) Urciuolo, M.; Solimene, R.; Chirone, R.; Salatino, P. Fluidized bed combustion and fragmentation of wet sewage sludge. Exp. Thermal Fluid Sci. 2012, 43, 97−104. (3) Murakami, T.; Suzuki, Y.; Nagasawa, H.; Yamamoto, T.; Koseki, T.; Hirose, H.; Okamoto, S. Combustion characteristics of sewage sludge in an incineration plant for energy recovery. Fuel Process. Technol. 2009, 90, 778−783. (4) Werle, S.; Wilk, R. K. A review of methods for the thermal utilization of sewage sludge: The Polish perspective. Renewable Energy 2010, 35, 1914−1919. (5) Khiari, B.; Marias, F.; Vaxelaire, J.; Zagrouba, F. Incineration of a small particle of wet sewage sludge: A numerical comparison between two states of the surrounding atmosphere. J. Hazard. Mater. 2007, 147, 871−882. (6) Khiari, B.; Marias, F.; Zagrouba, F.; Vaxelaire, J. Use of a transient model to simulate fluidized bed incineration of sewage sludge. J. Hazard. Mater. 2006, 135, 200−209. (7) Fytili, D.; Zabaniotou, A. Utilization of sewage sludge in EU application of old and new methodsA review. Renew. Sustain. Energy Rev. 2008, 12, 116−140. (8) Åmand, L.-E.; Leckner, B. Metal emissions from co-combustion of sewage sludge and coal/wood in fluidized bed. Fuel 2004, 83, 1803− 1821. (9) Shimizu, T.; Toyono, M.; Ohsawa, H. Emissions of NOx and N2O during co-combustion of dried sewage sludge with coal in a bubbling fluidized bed combustor. Fuel 2007, 86, 957−964. (10) García, G.; Arauzo, J.; Gonzalo, A.; Sánchez, J. L.; Á brego, J. Influence of feedstock composition in fluidised bed co-gasification of mixtures of lignite, bituminous coal and sewage sludge. Chem. Eng. J. 2013, 222, 345−352. (11) Hein, K. R. G.; Bemtgen, J. M. EU clean coal technologycocombustion of coal and biomass. Fuel Process. Technol. 1998, 54, 159− 169. (12) Sänger, M.; Werther, J.; Ogada, T. NOx and N2O emission characteristics from fluidised bed combustion of semi-dried municipal sewage sludge. Fuel 2001, 80, 167−177. (13) Mura, E.; Debono, O.; Villot, A.; Paviet, F. Pyrolysis of biomass in a semi-industrial scale reactor: Study of the fuel-nitrogen oxidation during combustion of volatiles. Biomass Bioenergy 2013, 59, 187−194. (14) Li, S.; Cadet, C.; Thivel, P. X.; Delpech, F. Towards the modelling and control of NOx emission in a fluidized bed sludge combustor. Comput. Chem. Eng. 2011, 35, 1281−1294. (15) Benzaoui, A.; Vaxelaire, J.; Marias, F.; Gelix, F.; Pillet, S.; Cretenot, D. Emission of Nitrogen Compounds by Incineration of a Wastewater

ASSOCIATED CONTENT

S Supporting Information *

The main equations that make the overall model. The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.energyfuels.5b01109.

■

REFERENCES

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Telephone: 33 559 407 809. Notes

The authors declare no competing financial interest. M


Article

Energy & Fuels Sludge Particle Under Fluidized Bed Conditions: A Numerical Approach. Waste Biomass Valor. 2013, 4, 181−195. (16) Shao, L.-M.; Fan, S.-S.; Zhang, H.; Yao, Q.-S.; He, P.-J. SO2 and NOx emissions from sludge combustion in a CO2/O2 atm. Fuel 2013, 109, 178−183. (17) Cammarota, A.; Chirone, R.; Salatino, P.; Solimene, R.; Urciuolo, M. Particulate and gaseous emissions during fluidized bed combustion of semi-dried sewage sludge: Effect of bed ash accumulation on NOx formation. Waste Manage. 2013, 33, 1397−1402. (18) Shimizu, T.; Toyono, M. Emissions of NOx and N2O during cocombustion of dried sewage sludge with coal in a circulating fluidized bed combustor. Fuel 2007, 86, 2308−2315. (19) Gavin, D. G.; Dorrington, M. A. Factors in the conversion of fuel nitrogen to nitric and nitrous oxides during fluidized bed combustion. Fuel 1993, 72, 381−388. (20) Johnsson, J. E. Formation and reduction of nitrogen oxides in fluidized-bed combustion. Fuel 1994, 73, 1398−1415. (21) de las Obras-Loscertales, M.; Mendiara, T.; Rufas, A.; de Diego, L. F.; García-Labiano, F.; Gayán, P.; Abad, A.; Adánez, J. NO and N2O emissions in oxy-fuel combustion of coal in a bubbling fluidized bed combustor. Fuel 2015, 150, 146−153. (22) Feng Duan, C.-S. C.; Zhang, L.-h.; Chi, Y.-T. Effect of the molecular structure of nitrogen compounds on the pollutant formation in a bubbling fluidized-bed combustor. Energy 2015, 83, 383−402. (23) Qian, F. P.; Chyang, C. S.; Huang, K. S.; Tso, Jim Combustion and NO emission of high nitrogen content biomass in a pilot-scale vortexing fluidized bed combustor. Bioresour. Technol. 2011, 102, 1892−1898. (24) Zhou, W.; Zhao, C. S.; Duan, L. B.; Chen, X. P.; Liang, C. Twodimensional computational fluid dynamics simulation of nitrogen and sulfur oxides emissions in a circulating fluidized bed combustor. Chem. Eng. J. 2011, 173, 564−573. (25) Konttinen, J.; Kallio, S.; Hupa, M.; Winter, F. NO formation tendency characterization for solid fuels in fluidized beds. Fuel 2013, 108, 238−246. (26) Bin Yang, Y.; Sliwinski, L.; Sharifi, V.; Swithenbank, J. Dynamic behavior of sewage sludge incineration in a large-scale bubbling fluidised bed in relation to feeding-rate variations. Fuel 2008, 87, 1552−1563. (27) Bin Yang, Y.; Sliwinski, L.; Sharifi, V.; Swithenbank, J. Mathematical modelling of sewage sludge incineration in a bubbling fluidised bed with special consideration for thermally-thick fuel particles. Waste Manage. 2008, 28, 2245−2258. (28) Gómez-Barea, A.; Leckner, B. Estimation of gas composition and char conversion in a fluidized bed biomass gasifier. Fuel 2013, 107, 419− 431. (29) Nilsson, S.; Gómez-Barea, A.; Fuentes-Cano, D.; Ollero, P. Gasification of biomass and waste in a staged fluidized bed gasifier: Modeling and comparison with one-stage units. Fuel 2012, 97, 730− 740. (30) Gómez-Barea, A.; Leckner, B. Modeling of biomass gasification in fluidized bed. Prog. Energy .Combust. Sci. 2010, 36, 444−509. (31) Marias, F. Incinération d’un déchet modèle: de l’approche physique aux simulations en lit fluidisé. Ph.D. Thesis, Université de Pau et des Pays de l’Adour, France, 1999. (32) Marias, F.; Puiggali, J. R.; Flamant, G. Modelling for Simulation of Fluidized-Bed Incineration Process. AIChE J. 2001, 47, 1438−1460. (33) Marias, F.; Puiggali, J. R.; Flamant, G. Effects of freeboard volatile release during fluidized bed incineration of a model waste. Trans. Inst. Chem. Eng. Part B: Process Safety Environ. Protect. 2001, 79, 244−252. (34) Khiari, B. Modélisation et simulation de l’incinération de boues en lit fluidisé. Ph.D. Thesis, Université de Pau et des Pays de l’Adour, France, 2006. (35) Werther, J. Modelling and scale-up of industrial fluidized bed reactors. Chem. Eng. Sci. 1980, 35, 372−379. (36) Prins, W. Fluidized bed combustion of a single carbon particle. Ph.D. Thesis, University of Twente, The Netherlands, 1987. (37) Scott, S. A.; Davidson, J. F.; Dennis, J. S.; Hayhurst, A. N. The devolatilisation of particles of a complex fuel (dried sewage sludge) in a fluidised bed. Chem. Eng. Sci. 2007, 62, 584−598.

(38) Soria-Verdugo, A.; Garcia-Hernando, N.; Garcia-Gutierrez, L. M.; Ruiz-Rivas, U. Analysis of biomass and sewage sludge devolatilization using the distributed activation energy model. Energy Convers. Manage. 2013, 65, 239−244. (39) Gomez-Barea, A.; Nilsson, S.; Vidal Barrero, F.; Campoy, M. Devolatilization of wood and wastes in fluidized bed. Fuel Process. Technol. 2010, 91, 1624−1633. (40) Neves, D.; Thunman, H.; Matos, A.; Tarelho, L.; Gómez-Barea, A. Characterization and prediction of biomass pyrolysis products. Prog. Energy Combust. Sci. 2011, 37, 611−630. (41) Urban, D. L.; Antal, M. J., Jr. Study of the kinetics of sewage sludge pyrolysis using DSC and TGA. Fuel 1982, 61, 799−806. (42) Chao, C.; Chiang, H.; Chen, C. Pyrolytic kinetics of sludge from a petrochemical factory wastewater treatment planta transition state theory approach. Chemosphere 2002, 49, 431−437. (43) Calvo, L. F.; Otero, M.; Jenkins, B. M.; Garcia, A. I.; Moran, A. Heating process characteristics and kinetics of sewage sludge in different atmospheres. Thermochim. Acta 2004, 409, 127−135. (44) Font, R.; Fullana, A.; Conesa, J. Kinetic models for the pyrolysis and combustion of two types of sewage sludge. J. Anal. Appl. Pyrol. 2005, 74, 429−438. (45) Loha, C.; Chattopadhyay, H.; Chatterjee, P. Three dimensional kinetic modeling of fluidized bed biomass gasification. Chem. Eng. Sci. 2014, 109, 53−64. (46) Klimanek, A.; Adamczyk, W.; Katelbach-Wozniak, A.; Wecel, G.; Szlek, A. Towards a hybrid Eulerian−Lagrangian CFD modeling of coal gasification in a circulating fluidized bed reactor. Fuel 2015, 152, 131− 137. (47) Petersen, I.; Werther, J. Experimental investigation and modeling of gasification of sewage sludge in the circulating fluidized bed. Chem. Eng. Processing 2005, 44, 717−736. (48) Guedea, I.; Díez, L. I.; Pallarés, J.; Romeo, L. M. On the modeling of oxy-coal combustion in a fluidized bed. Chem. Eng. Sci. 2013, 228, 179−191. (49) Conesa, J. A.; Marcilla, A.; Prats, D.; Rodriguez-Pastor, M. Kinetic Study of the Pyrolysis of Sewage Sludge. Waste Manage. Res. 1997, 15, 293−305. (50) Dryer, F. L.; Glassman, I. High temperature oxidation of CO and CH4. Proceedings of the 14th International Symposium on Combustion, Pittsburgh, 1973; pp 987−1003. (51) Kerinin, E. V.; Shifrin, E. I. Mathematical Model of Coal Combustion and Gasification in a Passage of an Underground Gas Generator, Combustion, Explosion, and Shock Waves. Acad. Sci., Russia 1993, 148−154. (52) Biba, V.; Macak, J.; Klose, E.; Malecha, J. Mathematical Model for the Gasification of Coal under Pressure. Ind. Eng. Chem. Process Design Dev. 1978, 17, 92−98. (53) Srinivasan, S.; Siramulu, S.; Kulasekaran, S.; Agarwal, P. K. Mathematical modeling of fluidized bed combustion-2 combustion of gases. Fuel 1998, 77, 1033−1049. (54) Glassman, I. Combustion, 3rd ed.; Academic Press: San Diego, 1996. (55) Sriramulu, S.; Sane, S.; Agarwal, P. K.; Mathews, T. Mathematical modelling of fluidized bed combustion: 1. Combustion of carbon in bubbling beds. Fuel 1996, 75, 1351−1362. (56) Cao, J. P.; Li, L. Y.; Morishita, K.; Xiao, X. B.; Zhao, X. Y.; Wei, X. Y.; Takarada, T. Nitrogen transformations during fast pyrolysis of sewage sludge. Fuel 2013, 104, 1−6. (57) Liu, H.; Gibbs, B. M. Modelling of NO and N2O emissions from biomass-fired circulating fluidized bed combustor. Fuel 2002, 81, 271− 280. (58) Sung, L. Y. J. NOx and N2O emissions from circulating fluidized bed combustion. Master’s thesis, University of British Columbia, 1995. (59) Patankar, S. V. Numerical Heat Transfer and Fluid Flow; Hemisphere Publishing Corp.: Washington, DC, 1980. (60) Brenan, K. E.; Campbell, S. L.; Petzold, L. R. Numerical solution of initial-value problems in Differential algebraic equations; Classics in Applied Mathematics; Society for Industrial and Applied Mathematics: Philadelphia, PA, 1996. N


Article

Energy & Fuels (61) Ascher, U. M.; Petzold, L.R. Computer methods for ordinary differential equations and Differential Algebraic equations; Society for Industrial and Applied Mathematics: Philadelphia, PA, 1998. (62) Sargousse, A.; Le Lann, J. M.; Joulia, X.; Jourda, L. Disco, un nouvel environnement de simulation orientée objet. Presented at MOSIM’99, Modélisation et simulation des flux physiques et informationnels, 2ème Conférence Francophone de Modélisation et Simulation, Annecy, 1999.

O


Fate of Nitrogen during Fluidized Incineration of ... - ACS Publications

Recommend Documents