Numerical Modelling of Coal Combustion Processes in the Vortex Furnace with Dual-Port Loading

doi:10.4236/epe.2013.54B060

Paper Menu >>

Journal Menu >>

Energy and Power Engineering, 2013, 5, 306-310

doi:10.4236/epe.2013.54B060 Published Online July 2013 (http://www.scirp.org/journal/epe)

Numerical Modelling of Coal Combustion Processes

in the Vortex Furnace with Dual-Port Loading

I. S. Anufriev1, E. P. Kopyev2, D. V. Krasinsky1, V. V. Salomatov1,2, E. Y. Shadrin2, O. V. Sharypov1,2

1Kutateladze Institute of Thermophysics, Siberian Branch of Russian Academy of Sciences, Novosibirsk, Russia

2Novosibirsk State University, Novosibirsk, Russia

Email: [email protected]

Received March, 2013

ABSTRACT

The work is devoted to numerical simulation of pulverized-coal combustion processes in the vortex furnace which is a

prospective design of a boiler unit for thermal power plants. New modification of this design characterized by addi-

tional tangential-injection nozzle located at the bottom of combustion chamber has been studied. Numerical results for

the case of Siberian brown coal combustion in this vortex furnace with dual-port loading are presented, including 3-D

aerodynamic structure, the fields of temperatures, radiated heat fluxes, species and dispersed phase concentrations, and

NOx emissions.

Keywords: Vortex Furnace, Pulverized-Coal Combustion, Numerical Modelling

1. Introduction

One of prospective technologies to improve the heat

transfer efficiency and ecological performance of boiler

units for thermal power plants utilizing low-grade coals

is the use of furnace devices with vertical jet pulver-

ized-coal combustion. As a kind of these devices, the

so-called vortex furnace has been considered. The vortex

furnace design shown in Figure 1 usually comprises the

vortex combustion chamber in the lower part of the fur-

nace, the diffuser, and the cooling chamber exited with

discharge flue. With this the pulverized coal-air mixture

is injected tangentially (alongside the main vortex flow

in the combustion chamber) via rectangular nozzle lo-

cated in the upper part of combustion chamber, i.e. via

the upper loading port. For the initial vortex furnace de-

sign with only upper nozzle injection, the complex ex-

perimental and numerical study of interior aerodynamics

and turbulent flow features has been carried out at labo-

ratory-scale vortex furnace model in isothermal formula-

tion earlier by authors [1]. In the presented work a new

modification of this design, proposed also by authors [2]

and characterized by additional tangential-injection noz-

zle located horizontally at the bottom of combustion

chamber (bottom loading port), has been studied nu-

merically with comprehensive combustion model for the

full-scale furnace configuration. In this new vortex fur-

nace design with dual-port loading, the stable swirl flow

inside the combustion chamber is fo rmed by two tangen-

tial jets of coal-air mixture released from the upper and

the bottom nozzles. The other distinguishing features of

new vortex furnace design are: the ratio of the diffuser

throat width to effective diameter D0 of vortex combus-

tion chamber is Hx = 0.4; the centers of semi-orifices

(comprising the frontal and back walls in XY-section of

the vortex chamber) are displaced between each other

strictly along vertical line (Figure 1).

The following advantages of new vortex furnace de-

sign can be indicated: – the compact mass-dimensional

parameters, – the distributed (between the upper and the

bottom nozzles) tangential input of air or fuel-air mixture,

which also provides more flexibility to control the fur-

nace working regime, – the horizontal direction of the

axis of rotation which facilitates the completeness of coal

Figure 1. Scheme of the vortex furnace new design (upper

part of the cooling chamber is not shown).

I. S. ANUFRIEV ET AL. 307

burnout process, – the high-temperature combustion of

coals of wide quality range in the regime of liquid slag

removal in the vortex chamber. Obviously the vortex

transfer appears to be the aerodynamic basis of entire

combustion process in the studied vortex furnace. With

this the appropriate aerodynamic structure can enhance

the combustion efficiency, stability of temperature and

heat flux distributions, and reduce toxic emissions.

Therefore the purpose of the presented work is to inves-

tigate the features of two-phase turbulent reacting flow

inside the furnace with aim to estimate and justify the

new design and operating parameters for prospective

boiler units with vortex furnace.

2. Mathematical Model of Aerodynamics

and Coal Combustion Processes

Flow in the vortex furnace is subsonic, turbulent and

essentially three-dimensional due to the constructional

features of the furnace. The mathematical formulation of

the steady-state 3-D turbulent two-phase reacting flow is

based on Euler-Lagrange approach. The Favre-averaged

governing equations of momentum, enthalpy and the

species concentration transport for carrier gas phase are

closed with “realizable” modification of k- turbulence

model [3]. This model, compared to the “standard” k-

model [4], provides more accurate predictions for swirl

flows (which is essential for the vortex furnace) and de-

monstrates good numerical robustness.

For description of the heat-mass transfer and combus-

tion processes, the carrier gas phase is represented by

mixture of the following sp ecies: {O2, CHnOm, CO, CO2,

H2O, N2}. Here CHnOm denotes the generalized formula

for hydrocarbons released during the volatiles’ pyro lysis.

Coefficients m and n in CHnOm formula were calculated

from elementary composition of the coal which is pre-

sumed to consist of the volatiles, ash, wet and free car-

bon. Berezovskij brown coal (rank “2B”) of siberian

Kansk-Achinsk field has been taken in simulations, with

the following data of its proximal analysis: W = 33 %, A0

= 4.7%, Cf0 = 44.2%, OV0 = 14.4%, HV0 = 3.1%, NV0 =

0.4%, S0 = 0.2%. Its volatiles content related to dry

ash-free fraction is Vdaf = 48%, and burnout heat release

Q = 15.66 MJ/kg. Then the coefficients in CHnOm were

estimated as: n = 3.037, m = 0.889, and volatiles’ molar

weight = 29.3 g/mol. The rate of the hydrocarbon

gas-phase burning reaction in the turbulent flow has been

determined according to eddy-dissipation model [5]. The

radiation energy transport equation is based on the

known P-1 approximation of the spherical harmonics

method, with this the gas-phase absorption coefficients

are calculated according to the weighted sum of gray

gases model [6], while the dispersed phase absorption

and scattering coefficients are determined on the basis of

the optically large particles approximatio n [7].

Within Lagrangian particle-tracking framework, ODEs

of mass, momentum and heat transfer for representative

clusters of polydispersed coal particles are solved along

each trajectory calculated via random walk model to ac-

count for turbulent dispersion of particles. The rate of

heterogeneous coke combustion reaction has been mod-

eled by diffusion-kinetic relations [7]. To account for

polydispersity, the clusters of fresh coal particles were

split into ~100 fractions according to Rosin-Rammler

size distribution law, with this the co al milling dispersity

(specified as mass fraction of remainder on 90 m sieve)

was taken as R90 = 0.15 and polydispersity degree  =

1.2.

To predict the nitric oxides emission in the vortex fur-

nace, at the stage of post-processing of main solution

data, the transport equations for the concentrations of

{NO, HCN, NH3} species have been solved. With this

the following NOx formation mechanisms are taken into

account: a) extended “thermal” N2 oxidation mechanism

of Ya.B.Zeldovich; b) “prompt” NO formation in reac-

tion of N2 with hydrocarbon radicals following C. P. Fe-

nimore mechanism; c) “fuel” NO formation mechanism.

Also the factors of NO reduction are accounted: – via

“reburn-mechanism” in reaction with CHi radicals; – via

heterogeneous consumption of NO on the surface of coke

particles.

3. Boundary Conditions and Numerical

Algorithm

Because the studied vortex furnace has a symmetry plane

SXY (in the section zsymm=3 m) between two nozzles (see

Figure 1), the mirror symmetry of the flow is assumed in

relation to this plane, thus to save computational re-

sources only half of the furnace volume was considered

in simulations. The bounding dimensions of computa-

tional domain were taken as: xmax=D0= 6 m, ymax= 36 m,

zmax= zsymm= 3 m. The upper nozzle angle of inclination to

horizon (XZ plane) was =24°. The unstructured com-

putational grid built for finite-volume discrimination of

equations consists of 351784 hexahedral cells. The

“sticking” boundary conditions are set for velocity com-

ponents at walls, and “enhanced wall treatment” method

[8] is applied for near-wall turbulence modeling. A uni-

form profile of mean flowrate velocity is prescribed at

inlet sections of rectangular nozzles, with this the inten-

sity of inlet turbulen t pulsations is assumed equal to 5%.

The known Marsha bound ary conditions [7] f or the ra-

diation energy transport equation have been set at walls,

with emissivity of the furnace walls taken as w = 0.8 and

the temperature of superheated water-vapor inside the

heat-exchange tubes presumed at 390℃. Also to set

boundary conditions for radioactive and convective heat

transfer at the heat-exchange surfaces, the integral heat

transfer coefficient w through the walls (from ash- de-

I. S. ANUFRIEV ET AL.

308

position layer surface to water-vapor inside the tubes)

has been prescribed at different heat-exchange screens

(walls) in the furnace cooling chamber. Its values speci-

fied at these surfaces, as well as the average values of the

screens thermal efficiency av obtained in computations,

are given in Table 1.

The mass flow rate of pulverized coal loaded per com-

puted half-section of the vortex furnace volume was

Gcoal=3.75 kg/s, the overall air excess coefficient at fur-

nace inlets was set typically for the vortex furnace design

as 1.15. With this the primary stream of coal-air mixture

at inlet temperature of 180℃ was loaded through the

upper nozzle only, while the secondary dry air stream at

320 C was supplied through th e bottom nozzle. Also the

ratio of gas-phase flow rates through the upper and the

bottom nozzles =Gupper/Gbottom was taken as =3 – as it

has been revealed in computations, this parameter influ-

ences the aerodynamic structure formed by two tangen-

tially injected jets insid e the vortex combustion chamber.

For numerical approximation of convective terms of

Favre-averaged momentum equations the second-order

upwind scheme [8] has been applied. The numerical so-

lution at each “global” iteration for the gas phase equa-

tions is obtained according to PISO algorithm [9] for

pressure-velocity decoupling, while the interphone ex-

change of mass, momentum and heat transfer is ac-

counted following the Particle-Source-In-Cell method

[10].

4. Computational Results

Typical flow structure in the studied vortex furnace is

represented in Figure 2 in the form of velocity magni-

tude is contours in XY-section (z = 1.6 m) across the noz-

zle center (zoomed view of only the vortex chamber and

diffuser is shown). The following features of the vortex

furnace aerodynamics can be seen, such as the “glove-

scheme” flow structure when the upper inlet jet (injected

tangentially from the nozzle) evolving alongside the

main vortex flow in the combustion chamber is being

crossed and surrounded by the upstream flow moving out

of this chamber towards the diffuser part of the furnace.

Near the center of vortex combustion chamber (but shifted

in upper-right di rect ion) a region of vortex core is clearly

Table 1. Average heat transfer coefficients (w) and thermal

efficiencies (av) at heat-exhange surfaces.

Heat-exhange

surface location Specified value

of w , W/(m2K) Computed value

of av

Screen “A” in section z=3 m 250 0.590

Two-sided screen “B”

in section z=0 200 0.483

Other screens (front , rear and

ceiling in the cooling chamber) 150 0.415

seen. The upstream flow inside the cooling chamber de-

monstrates noticeable spatial no uniformity (such as

Coanda effect).

This Coanda effect can be also seen from Figure 3

where the temperature field in the new vortex furnace is

demonstrated. High-temperature level inside the vortex

combustion chamber (reaching 1880℃ in the near-wall

reaction zone, see Figure 3) provides th e stable regime of

liquid slag removal. This is also confirmed from Figure

4 where the profile of temperature averaged over hori-

zontal XZ-sections is depicted along furnace height (solid

line in Figure 4). Also the dashed line there indicates the

profile of maximum temperature values along furnace

height. With this the estimated heat release rate per unit

volume inside the combustion chamber is ~800 kW/m3.

The sharp downward trough in averaged temperature

profile (at ~7 m height) in Figure 4 indicates the cold jet

stream evolved from the upper nozzle (see also Figure 3).

Then in the cooling chamber the temperature decreases

rapidly along the furnace height, and the temperature

Figure 2. Contours of velocity magnitude in the new vortex

furnace (section across nozzle center z=1.6 м), m/s.

Figure 3. Temperature field in the new vortex furnace

(section across nozzle center z = 1.6 м), ℃.

I. S. ANUFRIEV ET AL. 309

field becomes more uniform as the difference between

maximum and averaged values becomes smaller. This

improved performance appears due to efficient heat

transfer from the furnace volume to heat-exchange

screens. Indeed the profiles of the screens thermal effi-

ciency  (defined as =Qres/Qinc where Qres is the re-

sulting heat flux at wall surface and Qinc – the incident

heat flux) shown in Figure 5 demonstrate rather good

thermal efficiency values of screens in the cooling

chamber. The dashed line there denotes values at the

middle of screen “A” (mounted in section z=3 m) and the

solid line – at the middle of two-sided screen “B” (in

section z=0). Values av averaged over each screen sur-

face area (shown in Table 1) also indicate the range 0.41

< av < 0.59 which is better than a typical level of ~0.4

for most furnaces.

Figure 4. Averaged (solid line) and maximum (dashed line)

temperature distribution along vortex furnace height, C.

Figure 5. Heat-exhange screens thermal efficiency  distri-

bution along furnace height in the cooling chamber; dashed

line – screen “A”, solid line – screen “B”.

The contours of resulting heat fluxes (at heat-exchange

screens) are shown in Figure 6. The profiles of O2, CO,

NO concentrations averaged over horizontal XZ-sections

are plotted in Figure 7 along furnace height.

The following integral parameters have been obtained

in the vortex furnace outflow section (before exhaust

flue): mean temperature Tmean= 980℃, maximum tem-

perature Tmax= 1055℃, coke burnout incompleteness

coefficient q4~1.3%, averaged volumetric concentrations

[O2]= 2.47%, [CO]= 14 ppm, [NO]= 307 ppm. It is seen

that NOx emission level remains contained within ac-

ceptable limits – notwithstanding the high temperature

level inside the vortex combustion chamber – mainly due

to the effect of recirculation of combustion products

Figure 6. Heat fluxes absorbed at the surface of heat-ex-

change screens, kW/m2.

Figure 7. Distributions of O2, CO, NO concentrations (by

volume), averaged over horizontal XZ-sections, along vortex

furnace height.

I. S. ANUFRIEV ET AL.

310

towards reaction zones inside the vortex chamber. Also

the improved thermal efficiency of heat-exchange screens

in the cooling chamber has been demonstrated.

5. Conclusions

The numerical simulation of 3-D aerodynamics and pul-

verized brown coal combustion processes in the new

vortex furnace configuration with dual-port loading has

been performed, and detailed flow field information in-

cluding the flow structure, the fields of temperatures,

radiated heat fluxes, species and dispersed phase concen-

trations have been obtained. With this it has been shown

that the integral heat engineering and ecological parame-

ters of new vortex furnace meet the ranges corresponding

to engineering practice.

6. Acknowledgements

Computations have been performed with the use of CFD

package FLUENT at the supercomputer clusters

NKS-160 and NKS-30T (SSCC SB RAS, Novosibirsk,

Russia). The work has been supported by Russian Minis-

try of Education and Science (Agreement No. 8187).

REFERENCES

[1] V. V. Salomatov, D. V. Krasinsky, Yu. A. Anikin, I. S.

Anufriev, O. V. Sharypov and Kh. Enkhjargal, “Experi-

mental and Numerical Investigation of Aerodynamic

Characteristics of Swirling Flows in a Model of the Swirl-

ing-type Furnace of a Steam Generator,” J. of Engineer-

ing Physics and Thermophysics, Vol. 85, No. 2, 2012, pp.

282-293.doi:10.1007/s10891-012-0651-8

[2] N. V. Golovanov, V. E. Nakoryakov, A. P. Burdukov, V.

V. Salomatov and A. A. Dorozhkov, “Vortex Furnace,”

Russian Federation patent No.2042084, 1995 (in Rus-

sian).

[3] T.-H. Shih, W. W. Liou, A. Shabbir, Z. Yang and J. Zhu

“A New k- Eddy-Viscosity Model for High Reynolds

Number Turbulent Flows – Model Development and Va-

lidation,” Computers&Fluids –Vol. 24, No. 3, 1995, pp.

227-238. doi:10.1016/0045-7930(94)00032-T

[4] B. E. Launder and D. B. Spalding, “Lectures in Mathe-

matical Models of Turbulence,” London (England): Aca-

demic Press, 1972.

[5] B. F. Magnussen, “The Eddy Dissipation Concept,” IEA,

1989.

[6] T. F. Smith, Z. F. Shen and J. N. Friedman “Evaluation of

Coefficients for the Weighted Sum of Gray Gases Mod-

el,” Proceedings of XX-th National ASME-AIChE Heat

Transfer Conference, Milwakee, USA, August 2-5, 1981.

[7] E. P. Volkov, L. I. Zaichik and V. A. Pershukov, “Model-

ling the Solid Fuel Combustion,” Moscow: Nauka, 1994

(in Russian).

[8] FLUENT 6.3 User’s Guide, Fluent Inc., 2006.

[9] R. I. Issa, “Solution of Implicitly Discretized Fluid Flow

Equations by Operator Splitting,” Journal of Computa-

tional Physics, Vol. 62, No. 1, 1986, pp. 40-65.

doi:10.1016/0021-9991(86)90099-9

[10] C. T. Crowe, M. P. Sharma and D. E. Stock, “The

Particle-Source-In-Cell (PSI-CELL) model for

gas-droplet flows,” ASME Journal of Fluids Engi-

neering, Vol. 99, No. 2, 1977, pp. 325-332.

doi:10.1115/1.3448756