DESIGN STUDY OF BOILER FURNACES
Content
Power Technology and Engineering
Vol. 40, No. 3, 2006
DESIGN STUDY OF BOILER FURNACES
A. A. Shatil’1
Translated from Élektricheskie Stantsii, No. 4, April 2006, pp. 5 – 10.
It is shown that the standard method of design of boiler furnaces is not suitable for nonstandard operating conditions. A physicomathematical model of the furnace process is suggested for design studies of promising furnace devices and of conditions of flame extinction. An example of engineering design of a furnace with controlled process is presented.
Keywords: standard method, furnace design, operating conditions, model, furnace process, study, extinction.
The theory and design of heat exchange in furnaces presented in [1, 2] have been developed for the most part by
professors A. M. Gurvich and V. V. Mitor. For many years
these norms have played a determining role in the creation
and maintenance of boiler units in Russia. In 1998 the furnace part of the standard method was amended in the third
edition [3].
Under conditions of unsteady supply of fuel to power
generating units, of the need for changing (or, on the contrary, for preservation of) their operating parameters and for
raising the flexibility of operation, and of increasingly stringent economical and ecological requirements, it has became
necessary to control the processes of combustion and heat
exchange in the furnace chambers of boilers. It is obvious
that the zero-dimensional standard methods of furnace design developed for traditional furnace and burner devices
have limited possibilities as applied to novel technologies of
firing, especially for solid fuel. This circumstance makes it
necessary to develop a zonal engineering method of furnace
design, which will consider simultaneously the processes of
mixing, combustion, and heat exchange in the furnace.
Unfortunately, the procedure for computation of radiant
heat exchange recommended in [3] does not involve the concepts of the degree of flame blackness afl and of the reduced
degree of furnace blackness af widely used in scientific literature. This has led to discrepancies in methodological recommendations on design of furnace devices issued jointly by
the Central Boiler and Turbine Institute (TsKTI) and the
All-Russia Thermal Engineering Institute (VTI) [4].
Considering the computation of heat exchange in furnaces we should not neglect the discussion in [5] concerning
the reliability of the values of the coefficient of thermal efficiency of waterwalls ø recommended in [2, 3], which is determined as the ratio of the radiant heat flows taken up by a
1
waterwall (qt) and incident on it (qi). This inevitably involves
conditional evaluation of the coefficients ø and af and of the
effective radiant temperature of the flame Teff that enter the
equation for computing the specific heat taken up by the
waterwalls [2], i.e.,
4
qt = 4.9 ´ 10–8 ´ øafT eff
.
(1)
Since the value of the heat flow qt is measured quite reliably, for the specified the temperature Teff it is fully determined by the product y a f . This gives some leeway for
choosing the factors ø and af individually. Numerous experimental data show that the gas temperature at the outlet from
boiler furnaces computed by the method of [2] is usually
lower than the actual temperature [5]. This means that the
computed values of the degree of blackness af (or of the coefficient ø) have been overestimated. This drawback is absent in [3]. This could have been done in [2] if the assumption that the areas of the surfaces of the waterwall Fw and of
the flame Ffl in the equation relating af, afl, and ø are the
same were not used; this relation should have the form [6]
af =
a fl
,
a fl + ( F - a fl )y
(2)
where F is a conventional ratio of the areas mentioned,
which depends on the optical density of the flame Bu or on
the proportion of the volume occupied by the luminous
flame.
Thus, in accordance with Eq. (2) the decrease in af at tabulated value of ø [2, 3] causes a decrease in qt and, correspondingly, increases the balance temperature T f at the outlet
from the furnace. It is obvious that in the given case the increase in ø, as suggested in [5], will again disturb the design
heat balance of the furnace and is thus inexpedient.
“NPO TsKTI” Company, St. Petersburg, Russia.
179
1570-145X/06/4003-0179 © 2006 Springer Science + Business Media, Inc.
180
A. A. Shatil’
u
Äw
Äw
u
w3
w2
w1
Fig. 1. Picture of turbulent mass transfer in a flow with velocity
gradient (w3 > w2 > w1): u, w, transverse and longitudinal velocities,
respectively.
In recent years TsKTI specialists developed an approximate zero-one-dimensional zonal model and “TORKA” software for computing the firing process in boilers [7, 8], which
was later generalized to furnace devices with different aerodynamic circuits and methods of fuel firing [9]. This physicomathematical model of the process (stationary combustion, extinction, inflammation, heat exchange) employs the
ideas presented in [10]. Specifically, L. A. Vulis suggests “...
to give up observation of the behavior of individual particles
of fuel or combustible mixture and concentrate on the phenomena occurring in the central part of the furnace volume.”
He also notes that the introduction of volume-averaged values of temperature and concentration better meets the actual
conditions of the occurrence of the process in those parts of
the furnace chamber where the fresh mixture intensely mixes
with the combustion products.
Organization of the firing process is based on mixing of
three flows, i.e., of the oxidizer (air), of the fuel, and of the
combustion products. The applied science of firing considers
two mechanisms of diffusion (mixing), i.e., molecular and
turbulent ones. However, it is obvious that these two mechanisms are accompanied by convective mixing caused by the
aerodynamic structure of the flow, i.e., the structure of the
burning device or of the furnace chamber itself.
In actual furnace devices all the three mixing mechanisms (molecular, turbulent, and convective) act in parallel
and every subsequent mechanism cannot occur without the
preceding ones. In other words division of flows into small
volumes due to the turbulence increases their total surface
for the occurrence of molecular mixing. In its turn, convective mass transfer not only ensures mixing of distant gas volumes but also creates velocity gradients that intensify manyfold the turbulent mass transfer in them. Therefore, it can be
assumed that the latter mechanism plays a determining role
in the three mechanisms of mixing in a furnace. This allows
us to estimate the quality of mixing in actual furnaces by
studying their models.
The mechanism of transverse transfer of momentum in a
turbulent flow with velocity gradient is considered in [11]. It
is shown that according to the theorem of N. E. Zhukovskii a
transverse “lifting force” acting on some small cylinder (volume), which randomly displaces from one “steam line” to
another one that has another velocity, draws it far aside. The
place of the cylinder is taken (by the same mechanism) by
another element moving in the opposite direction. This picture of mixing (mass transfer) is presented in Fig. 1; a simplified computation of parameters characterizing mass transfer
by this mechanism is presented in [8].
The geometrical and physical similarity of furnace devices is usually disturbed with growth in the capacity of
boiler units. The number of burners and decks increases and
the proportion of the sizes of the burners (nozzles) and of the
furnace changes respectively. The conditions of heat exchange between the flame and the furnace waterwalls change
too and so does the aerodynamic structure of the flows of air
and gases in the volume of the furnace. All these factors
worsen the mixing of the flows of fuel, air, and combustion
products.
The highest intensity of mass transfer is observed in the
initial region of the jets fed into the flow [12]. For straight
round jets its length Li is equal to 4 – 5 diameters of the nozzle d0. In the case of a great number of burners with small diameter (at the same velocities) mass transfer is primarily ensured by the near-wall layer of the ascending (carrying) flow.
This means that it is necessary to choose an optimum relative
hitting range of the jet h, as is recommended, for example, in
[13], for the design of gas burners. Consequently, in order to
improve mixing in the furnace of a more powerful boiler the
burners (nozzles) should have a larger size. This aspect of
mixing should also be matched with the kinetic conditions of
combustion, because the enlargement of the burners (nozzles) increases the sizes of the circulation zones and the time
of residence of the gases ôr in the zone of active combustion
(ZAC).
Many difficulties arising in operation of boilers firing
pulverized coal, for example, slagging of the heating surfaces, deviation from the design values of steam temperature,
narrowing of the working range of loads, and dependence on
the variable characteristics of the fuel, can be eliminated in
furnaces with controlled furnace process. An example of a
furnace device for firing one or several kinds of fuel is a
two-zone furnace that combines an inverted mode of organization of combustion [8, 9] and a system of undergrate blast.
The former design with top blast has been studied under operating conditions in [14, 15]. The latter design is more
widely used in boilers [16] but less studied.
Figure 2 presents a model for aerodynamic “blasting” of
such furnaces with top and undergrate blasts (TB and UB, respectively). The figure also shows the distribution of vertical
velocities of the flow in the longitudinal plane passing
through the axes of the upper nozzles. The results of operation [15] show that the firing process can be controlled effectively using the entire furnace volume and large uniflow noz-
Design Study of Boiler Furnaces
181
X2 = 1.0
á = 1.15
20.3
X2 = 0.66 á = 1.15
580
16.8
X2 = 0.33
á = 1.15
13.3
X1 = 1.0 á = 1.15
X1 = 0.67
a = 0.70
8.5
a
b
X1 = 0.22
a = 0.55
0.50V°
B = 0.9
6.3
Fig. 2. Model of a furnace with top and undergrate blast (a) and distribution of the vertical velocities of flow in it (b).
0.45V°
4.3
0.20V°
zles instead of conventional burners; the control should be
performed not only with the help of the “aerodynamics” but
also by choosing the appropriate place and methods for feeding of the fuel. It is obvious that such organization of the furnace process meets the tendency for improving the aerodynamics of combustion chambers and widening the zone of
active combustion in them [17] at the expense of the zone of
delayed burning (ZDB).
Examples of approximate computation of various devices (a boiler furnace, a combustion chamber, a furnace,
etc.) with the help of the “TORKA” software are presented in
[9]. Here an example of zone-after-zone computation of a
variant of controlled furnace combining TB and UB modes is
presented for a boiler with a steam output of 270 tons/h firing lean Kuznetsk coal with slag-tap removal. The design diagram of the furnace is given in Fig. 3. The diagram presents
the feed places and the flow rates of fuel B (in fractions of
the total flow rate) and of air (in factions of V0), the relative
distances from the initial sections of ZAC (X1) and ZDB
(X2), and the running excess air factors á = Óái/ÓBi. The
main design parameters of the firing process, i.e., the temperature of the flame J, the combustion efficiency â, the mechanical q4 and chemical q3 underfiring, the concentrations
of CO and C NO x (at á = 1.4), and the heat flows qt over the
height of the furnace, are presented in Table 1.
B = 0.9
Fig. 3. Design model of a furnace.
It can be seen from Table 1 that despite the unfavorable
conditions (the content of volatiles in the fuel (Vg = 9%,
ø = 0.45), the stability of combustion of low-reaction coal is
ensured by partial gasification of the fuel in the first two regions of ZAC. In this case the combustion efficiency is high
and the concentration of nitrogen and carbon oxides is low.
In addition, the computed temperature of gases at the outlet
from the furnace is also lower than the design value by
60 – 70°C due to the involvement of the sloping bottom into
the process, which implies the presence of a margin in the
steam rate of the boiler. This example shows that such computational studies make it possible to estimate the expediency of new engineering solutions and the expected parameters of the designed boiler.
The “TORKA” software makes it possible to compute
the parameters of the firing process not only for stationary
combustion modes but also for the modes of extinction and
inflammation. Figure 4 presents a diagram of the thermal regime of combustion of fuel in a flow in the coordinates â – è
(è = RT/E, where R is the gas constant, T is the temperature,
and E is the activation energy) [8 – 10]. The points in the diagram are the points of intersection and tangency of the
TABLE 1. Examples of Combustion Process in a Boiler Firing Lean Coal
X1
X2
B
á
J, °C
â, %
q 4, %
q3, %
CO, %
CNO x , mg/m3
që, Mcal/(m2 · h)
0.22
0.67
1.0
–
–
–
–
–
–
0.33
0.66
1.0
0.9
1.0
1.0
–
–
–
0.55
0.70
1.15
1.15
1.15
1.15
1376
1433
1364
1328
1259
1197
33.49
62.14
83.59
95.84
97.62
98.31
1.08
1.86
11.23
3.44
2.13
1.60
55.53
36.00
5.18
0.72
0.25
0.09
25.34
13.04
1.25
0.17
0.06
0.02
88
245
502
554
562
565
163
181
159
145
115
93
182
A. A. Shatil’
â
1
q3 = 0 is introduced into Table 1 of initial data for computing
the parameters of the limiting mode in which the computation stops upon a change in this or that input parameter. At
the same time, “normal” computation at q3 > 0 is performed
at the same process parameters, which corresponds to the
state of the process before the moment of extinction. These
two modes are “realized” in the algorithm through a kinetic
coefficient m included into the Arrhenius criterion and calculated by the formula
C
Ex
2
4
0.5
m = m0[1 – 0.4(x + q3)],
In
3
0
3
3
0.05
è
Fig. 4. Diagram of the thermal regime of combustion: 1, heat liberation curve; 2, heat removal curves; 3, initial point of the process
(â = 0, è = è0), C, point of stationary regime; Ex, point of extinction;
In, point of inflammation; 4, point of crisis-free regime.
curves of heat liberation and heat removal, the parameters of
which correspond to these modes. Since the stationary modes
have been studied quite well, we will consider in detail the
critical operating conditions of actual furnaces. Note that after the publication of [6], works devoted to determination of
the temperature of extinction Tex in boilers appeared as well.
A computational model of the firing process makes it
possible to analyze the modes of extinction in different kinds
of boilers with allowance for the numerous parameters of the
firing process. Extinction of the flame in furnaces of traditional boilers with burners depends primarily on the kinetics
of combustion and can be caused by growth in the excess air
factor in the burners ág (at ág > 1), in the fraction of the recycling flue gases r, in the coefficient of thermal efficiency of
the waterwalls ø, and in the thermal stress of the furnace volume qv, or by decrease in the temperature of hot air th.a, in the
content of volatiles in the fuel Vg, in the time of residence of
combustion products in ZAC ôr, in the content of oxygen in
the oxidizer O¢2 , etc. On the other hand, the stability of combustion depends on the aerodynamic structure of the flow
near the burners and in ZAC. The sizes of the zones of
recirculation play a substantial role. In operation extinction
of the flame is often caused by an abrupt decrease in the feed
of fuel (disabling of the mills or pulverized coal feeders or
reduction in the combustion value of the fuel), i.e., by step
growth in the excess air factor in the furnace.
Computation with the use of the “TORKA” software
gives the full volume fraction of flue gases in percent and the
value of chemical underfiring q3 in regions of the furnace.
Firing of organic fuel is always accompanied by the presence
of products of incomplete combustion of CO and H2. Therefore, the algorithm stipulates the possibility of determination
of the parameters of the firing process at the moment of
extinction, which is characterized by disappearance of these
gases, i.e., chemical underfiring in ZAC. The condition
(3)
where x is the fraction of “lighting” (highly reactive) fuel in
the total consumption of fuel per boiler (with respect to the
heat). The coefficient m0 for ZAC depends on the type of furnace; for ZDB it is assumed that m0 = 1 and x = 0 [9]. This
simulates a fluctuating combustion process from q3 > 0 to
q3 = 0 (and backward) observed in operation as pulsations of
pressure and temperature in the furnace before extinction of
the flame.
It is presumed in the chosen model of firing (Fig. 4) that
a stationary process with crisis moments of extinction and inflammation under specific conditions can transform into a
crisis-free process without extinction [10]. Transfer from a
crisis form of the â – T dependence to a crisis-free form is
possible, in principle, in one and the same device. Specifically, this can be ensured by increasing the initial temperature of the process T0, using an auxiliary (pilot) flame of
highly reactive fuel (i.e., reducing the coefficient m, which
lowers the “effective” activation energy E), or decreasing the
removal of heat through the walls of the furnace chamber
(reduction of the coefficient ø). A powerful factor of stabilization of combustion is the creation of zones with á < 1 and
zones with combustion of lumped solid fuel in a fluidized (or
another) layer where the chemical underfiring is inconsiderable. It follows from Eq. (3) that this is similar to “lighting”
(x > 0).
The computational dependences of the extinction temperature Tex of the flame of pulverized coal on the content of
volatiles in the fired fuel Vg plotted for several boilers firing
pulverized coal with dry-ash and slag-tap removal, which are
presented in Fig. 5 [8], allow us to write the following equation
Tex = 1660 – 8.3Vg
(4)
Tex = 0.07E/R + 410.
(5)
or
Figure 5 also presents the relation between the computed
values of the temperature Tex and the combustion efficiency
âex at the moment of extinction, which is described by the
equation
Tex = 554/(1 – âex).
(6)
Design Study of Boiler Furnaces
183
Tex, K
This dependence can be written in the form
Tex, K
a
b
9
b ex =
1106 - 8.3V g
1660 - 8.3V g
1500
,
(7)
from which we find that at Vg = 0 (coke, E = 36 ´ 104
kcal/mole) and at Vg = 100% (natural gas, E = 12 ´ 103
kcal/mole) the values of âex amount to 0.67 and 0.33 respectively. Note that these values of âex coincide with the values
of â for extinction and inflammation of a flame of pulverized
coal obtained in [18] on the basis of other assumptions. However, it follows from (7) that the value âex = 2/3 is not constant and decreases to 1/3 upon growth in the reactivity of
the fuel.
On the whole, the thus obtained relationship between the
extinction parameters can be treated as a universal one. However, both these parameters (Tex and âex) are “virtual” and
cannot be used in practice because of the impossibility of
their representative estimation. Therefore, in order to predict
the conditions of stability of combustion we should compute
regime parameters of the process on which it depends and
which should be measurable. The computation should allow
for the following additional circumstances. The first circumstance is the effect of not only the temperature but also the
time ôr of residence of combustion products in the regions of
ZAC on the processes of combustion and extinction. In different flame furnaces it can differ by several times [9].
The effect of the residence time ôr (in seconds) on Tex can
be allowed for by an empirical coefficient determined from
the formula
kô = 1.05 – 0.05ôr.
(8)
Consequently, the temperature Tex at âex > 1 will be
lower and at âex < 1 will be higher than at âex = 1 sec. The
second circumstance is connected with the instability of
feeding of solid fuel by the feeders of pulverized coal [18].
This requires what is known as a margin Äá in the computation of the limiting air excess factor at the moment of extinction [8, 9]. For example, ±3% fluctuations of the rate of fuel
feed (accordingly, Äá = 0.03) can raise the extinction temperature by 80 – 100 K.
Evaluation of the parameters of crisis modes of inflammation does not present great practical interest. It is much
more important to be able to compute the parameters of the
mode of pilot firing of the boiler. In such a computation the
initial data are the air excess factor and the air temperature,
the rates of the main fuel and of the highly reactive fuel
(black oil, gas), and other parameters characterizing the performance of the burners and of the furnace.
Let is consider as an example the mode of pilot firing of
a TPP-210A twin-furnace boiler firing pulverized coal of
grade ASh. It is assumed that the solid fuel is fed into the furnace when the air is heated by the highly reactive fuel to
1000
500
0
9
4
12 3 11 10
2 1
6
11 4
12
3
10 2
6 1
1500
5
13 7
50
5
7
8
V g, %
1000
500
0
13
8
50
V g, %
Fig. 5. Dependences of computed temperature of extinction on the
content of volatiles in the fuel (a) and on the combustion efficiency
(b) at the moment of extinction for different kinds of furnace, grades
of coal, and types of ash removal: 1, TP-109, G, solid; 2, TP-240, SS,
solid; 3, TP-214, SS, solid; 4, TP-215, SS, solid; 5, BKZ-210, B1,
solid; 6, BKZ-500, B2, solid; 7, PK-10, peat, solid; 8, PK-10, lead,
solid; 9, TP-210A, ASh, liquid; 10, TP-100, G, liquid; 11, TP-87, SS,
liquid; 12, TPB-318, ASh, solid; 13, BKZ-670, B2, solid.
100°C. The rate of feeding of pulverized coal to two burners
out of the six available ones is 10 tons/h (18% of the rated
feeding); that of black oil is about 1 ton/h (x = 0.20). Since
in the pilot firing the flame does not fill the whole of the
cross section of the ZAC, the critical coefficient m0 is taken
to be equal to 0.85 as in low-stress furnaces with bottom (or
elevated) burners [9]. The coefficients of thermal efficiency
of the waterwalls ø are taken to be equal to 0.20 for the ZAC
and 0.45 for the ZDB. The degree of blackness of the furnace
af is determined using formula (2), in which afl = 0.9 and
Fw/Ffl = 4 for the ZAC and afl = 0.5 and Fw/Ffl = 3 for the
ZDB.
The decisive moment in pilot firing is the choice of the
maximum value of the excess air factor. It is assumed in the
initial data for computation with the help of “TORKA” that
in the first region of the ZAC with a height of 5 m á = 1.5,
whereas at the outlet from the second region (H = 8.95 m)
á = 1.70. This increase in the excess air factor is caused by
the arrival of air into the furnace through dead burners. Computations have shown that the flame in the ZAC is stable (JZAC
= 1067 – 1086°C, âZAC = 74.90 – 80.44%) and underfiring at
the outlet from the furnace is about 10%. A check of stability
of combustion (at q3 = 0) has shown that in the first region of
the ZAC combustion is absent and the fuel starts to burn only
in the second region. This means that in the given case pilot
firing at á > 1.70 in the ZAC is impossible with a heat fraction of black oil x = 0.20. Therefore, it is desirable either to
raise the temperature of the hot air, i.e., the boiler should fire
black oil for a longer period, or to increase the proportion of
the highly reactive fuel. This example shows how the personnel can be informed on the conditions of pilot firing of a specific boiler.
184
CONCLUSIONS
1. Development of new technologies of firing organic
fuel and improvement of the operating furnace devices requires creation of a method for their computational study.
2. The computational model of the firing process developed by TsKTI has been tested on a great number of furnaces
of operating boilers firing a wide spectrum of fuel in all firing stages from pilot firing to extinction and on various furnace devices. This allows us to recommend the model and
the “TORKA” software for computational studies not only of
traditional boiler furnaces but also of furnaces employing
novel firing technologies.
3. It is expedient to have at least two modifications of
the software for practical purposes. One of them can be used
for design studies of furnaces and the other (simpler) modification can be used for an operating furnace (or a boiler as a
whole) for estimating the consequences of various possible
working situations.
REFERENCES
1. Thermal Design of Boiler Units (Standard Method) [in Russian],
Gosénergoizdat, Moscow (1957).
2. Thermal Design of Boiler Units (Standard Method) [in Russian],
Énergiya, Moscow (1973).
3. Thermal Design of Boilers (Standard Method) [in Russian], NPO
TsKTI, St. Petersburg (1998).
4. É. Kh. Verbovetskii and N. G. Zhmerik (eds.), Methodological
Recommendations on Design of Furnace Devices [in Russian],
Izd. NPO TsKTI, St. Petersburg (1996).
5. V. I. Antonovskii, “Heat transfer in furnaces of steam boilers.
Retrospective consideration of the development of the standard
computational technique,” Teploénergetika, No. 9 (2004).
A. A. Shatil’
6. A. A. Shatil and E. K. Chavchanidze, “Computational estimation of stability of flame due to firing solid fuels in boiler furnaces,” Teploénergetika, No. 4 (1990).
7. A. A. Shatil and E. Ya. Skripova, “About design of furnaces of
boilers firing pulverized coal,” Teploénergetika, No. 9 (1993).
8. A. A. Shatil, Furnace Processes and Equipment [in Russian],
Izd. TsKTI, St. Petersburg (1997).
9. A. A. Shatil, Computational Study of Furnace Devices [in Russian], Izd. TsKTI, St. Petersburg (2003).
10. L. A. Vulis, Thermal Regime of Combustion [in Russian], Gosénergoizdat, Moscow – Leningrad (1954).
11. N. Ya. Fabrikant, Aerodynamics [in Russian], Nauka, Moscow
(1964).
12. G. N. Abramovich, Turbulent Free Jets of Liquids and Gases [in
Russian], Gosénergoizdat, Moscow – Leningrad (1948).
13. Yu. V. Ivanov, Fundamentals of Computation and Design of
Gas Burners [in Russian], Gosénergoizdat, Moscow (1963).
14. A. A. Shatil and B. N. Barbyshev, “A study of the aerodynamics
of a furnace with inverted flame,” Énerg. Életrifik., No. 2
(1983).
15. A. A. Shatil, V. P. Maistruk, A. E. Surkov et al., “A study of the
method of inverted combustion of coals of grades D and G and
their intermediate product in TP-230-3 boiler,” Élektr. Stantsii,
No. 1 (1986).
16. N. S. Klepikov, L. N. Gusev, A. A. Shatil et al., “Experience of
installation of undergrate blast system designed by Izd. NPO
TsKTI on low-, medium-, and high-capacity boilers,” Trudy
TsKTI, Issue 287.
17. G. G. Olkhovskii and A. G. Tumanovskii, “Problems and prospects of the use of coal in the power industry of Russia,” Énergetik, No. 12 (2004).
18. L. N. Gusev, “A practical method for determining the temperatures of inflammation and extinction of pulverized coal flame in
chamber furnaces,” Tyazh. Mashinostr., No. 6 (2002).
Vol. 40, No. 3, 2006
DESIGN STUDY OF BOILER FURNACES
A. A. Shatil’1
Translated from Élektricheskie Stantsii, No. 4, April 2006, pp. 5 – 10.
It is shown that the standard method of design of boiler furnaces is not suitable for nonstandard operating conditions. A physicomathematical model of the furnace process is suggested for design studies of promising furnace devices and of conditions of flame extinction. An example of engineering design of a furnace with controlled process is presented.
Keywords: standard method, furnace design, operating conditions, model, furnace process, study, extinction.
The theory and design of heat exchange in furnaces presented in [1, 2] have been developed for the most part by
professors A. M. Gurvich and V. V. Mitor. For many years
these norms have played a determining role in the creation
and maintenance of boiler units in Russia. In 1998 the furnace part of the standard method was amended in the third
edition [3].
Under conditions of unsteady supply of fuel to power
generating units, of the need for changing (or, on the contrary, for preservation of) their operating parameters and for
raising the flexibility of operation, and of increasingly stringent economical and ecological requirements, it has became
necessary to control the processes of combustion and heat
exchange in the furnace chambers of boilers. It is obvious
that the zero-dimensional standard methods of furnace design developed for traditional furnace and burner devices
have limited possibilities as applied to novel technologies of
firing, especially for solid fuel. This circumstance makes it
necessary to develop a zonal engineering method of furnace
design, which will consider simultaneously the processes of
mixing, combustion, and heat exchange in the furnace.
Unfortunately, the procedure for computation of radiant
heat exchange recommended in [3] does not involve the concepts of the degree of flame blackness afl and of the reduced
degree of furnace blackness af widely used in scientific literature. This has led to discrepancies in methodological recommendations on design of furnace devices issued jointly by
the Central Boiler and Turbine Institute (TsKTI) and the
All-Russia Thermal Engineering Institute (VTI) [4].
Considering the computation of heat exchange in furnaces we should not neglect the discussion in [5] concerning
the reliability of the values of the coefficient of thermal efficiency of waterwalls ø recommended in [2, 3], which is determined as the ratio of the radiant heat flows taken up by a
1
waterwall (qt) and incident on it (qi). This inevitably involves
conditional evaluation of the coefficients ø and af and of the
effective radiant temperature of the flame Teff that enter the
equation for computing the specific heat taken up by the
waterwalls [2], i.e.,
4
qt = 4.9 ´ 10–8 ´ øafT eff
.
(1)
Since the value of the heat flow qt is measured quite reliably, for the specified the temperature Teff it is fully determined by the product y a f . This gives some leeway for
choosing the factors ø and af individually. Numerous experimental data show that the gas temperature at the outlet from
boiler furnaces computed by the method of [2] is usually
lower than the actual temperature [5]. This means that the
computed values of the degree of blackness af (or of the coefficient ø) have been overestimated. This drawback is absent in [3]. This could have been done in [2] if the assumption that the areas of the surfaces of the waterwall Fw and of
the flame Ffl in the equation relating af, afl, and ø are the
same were not used; this relation should have the form [6]
af =
a fl
,
a fl + ( F - a fl )y
(2)
where F is a conventional ratio of the areas mentioned,
which depends on the optical density of the flame Bu or on
the proportion of the volume occupied by the luminous
flame.
Thus, in accordance with Eq. (2) the decrease in af at tabulated value of ø [2, 3] causes a decrease in qt and, correspondingly, increases the balance temperature T f at the outlet
from the furnace. It is obvious that in the given case the increase in ø, as suggested in [5], will again disturb the design
heat balance of the furnace and is thus inexpedient.
“NPO TsKTI” Company, St. Petersburg, Russia.
179
1570-145X/06/4003-0179 © 2006 Springer Science + Business Media, Inc.
180
A. A. Shatil’
u
Äw
Äw
u
w3
w2
w1
Fig. 1. Picture of turbulent mass transfer in a flow with velocity
gradient (w3 > w2 > w1): u, w, transverse and longitudinal velocities,
respectively.
In recent years TsKTI specialists developed an approximate zero-one-dimensional zonal model and “TORKA” software for computing the firing process in boilers [7, 8], which
was later generalized to furnace devices with different aerodynamic circuits and methods of fuel firing [9]. This physicomathematical model of the process (stationary combustion, extinction, inflammation, heat exchange) employs the
ideas presented in [10]. Specifically, L. A. Vulis suggests “...
to give up observation of the behavior of individual particles
of fuel or combustible mixture and concentrate on the phenomena occurring in the central part of the furnace volume.”
He also notes that the introduction of volume-averaged values of temperature and concentration better meets the actual
conditions of the occurrence of the process in those parts of
the furnace chamber where the fresh mixture intensely mixes
with the combustion products.
Organization of the firing process is based on mixing of
three flows, i.e., of the oxidizer (air), of the fuel, and of the
combustion products. The applied science of firing considers
two mechanisms of diffusion (mixing), i.e., molecular and
turbulent ones. However, it is obvious that these two mechanisms are accompanied by convective mixing caused by the
aerodynamic structure of the flow, i.e., the structure of the
burning device or of the furnace chamber itself.
In actual furnace devices all the three mixing mechanisms (molecular, turbulent, and convective) act in parallel
and every subsequent mechanism cannot occur without the
preceding ones. In other words division of flows into small
volumes due to the turbulence increases their total surface
for the occurrence of molecular mixing. In its turn, convective mass transfer not only ensures mixing of distant gas volumes but also creates velocity gradients that intensify manyfold the turbulent mass transfer in them. Therefore, it can be
assumed that the latter mechanism plays a determining role
in the three mechanisms of mixing in a furnace. This allows
us to estimate the quality of mixing in actual furnaces by
studying their models.
The mechanism of transverse transfer of momentum in a
turbulent flow with velocity gradient is considered in [11]. It
is shown that according to the theorem of N. E. Zhukovskii a
transverse “lifting force” acting on some small cylinder (volume), which randomly displaces from one “steam line” to
another one that has another velocity, draws it far aside. The
place of the cylinder is taken (by the same mechanism) by
another element moving in the opposite direction. This picture of mixing (mass transfer) is presented in Fig. 1; a simplified computation of parameters characterizing mass transfer
by this mechanism is presented in [8].
The geometrical and physical similarity of furnace devices is usually disturbed with growth in the capacity of
boiler units. The number of burners and decks increases and
the proportion of the sizes of the burners (nozzles) and of the
furnace changes respectively. The conditions of heat exchange between the flame and the furnace waterwalls change
too and so does the aerodynamic structure of the flows of air
and gases in the volume of the furnace. All these factors
worsen the mixing of the flows of fuel, air, and combustion
products.
The highest intensity of mass transfer is observed in the
initial region of the jets fed into the flow [12]. For straight
round jets its length Li is equal to 4 – 5 diameters of the nozzle d0. In the case of a great number of burners with small diameter (at the same velocities) mass transfer is primarily ensured by the near-wall layer of the ascending (carrying) flow.
This means that it is necessary to choose an optimum relative
hitting range of the jet h, as is recommended, for example, in
[13], for the design of gas burners. Consequently, in order to
improve mixing in the furnace of a more powerful boiler the
burners (nozzles) should have a larger size. This aspect of
mixing should also be matched with the kinetic conditions of
combustion, because the enlargement of the burners (nozzles) increases the sizes of the circulation zones and the time
of residence of the gases ôr in the zone of active combustion
(ZAC).
Many difficulties arising in operation of boilers firing
pulverized coal, for example, slagging of the heating surfaces, deviation from the design values of steam temperature,
narrowing of the working range of loads, and dependence on
the variable characteristics of the fuel, can be eliminated in
furnaces with controlled furnace process. An example of a
furnace device for firing one or several kinds of fuel is a
two-zone furnace that combines an inverted mode of organization of combustion [8, 9] and a system of undergrate blast.
The former design with top blast has been studied under operating conditions in [14, 15]. The latter design is more
widely used in boilers [16] but less studied.
Figure 2 presents a model for aerodynamic “blasting” of
such furnaces with top and undergrate blasts (TB and UB, respectively). The figure also shows the distribution of vertical
velocities of the flow in the longitudinal plane passing
through the axes of the upper nozzles. The results of operation [15] show that the firing process can be controlled effectively using the entire furnace volume and large uniflow noz-
Design Study of Boiler Furnaces
181
X2 = 1.0
á = 1.15
20.3
X2 = 0.66 á = 1.15
580
16.8
X2 = 0.33
á = 1.15
13.3
X1 = 1.0 á = 1.15
X1 = 0.67
a = 0.70
8.5
a
b
X1 = 0.22
a = 0.55
0.50V°
B = 0.9
6.3
Fig. 2. Model of a furnace with top and undergrate blast (a) and distribution of the vertical velocities of flow in it (b).
0.45V°
4.3
0.20V°
zles instead of conventional burners; the control should be
performed not only with the help of the “aerodynamics” but
also by choosing the appropriate place and methods for feeding of the fuel. It is obvious that such organization of the furnace process meets the tendency for improving the aerodynamics of combustion chambers and widening the zone of
active combustion in them [17] at the expense of the zone of
delayed burning (ZDB).
Examples of approximate computation of various devices (a boiler furnace, a combustion chamber, a furnace,
etc.) with the help of the “TORKA” software are presented in
[9]. Here an example of zone-after-zone computation of a
variant of controlled furnace combining TB and UB modes is
presented for a boiler with a steam output of 270 tons/h firing lean Kuznetsk coal with slag-tap removal. The design diagram of the furnace is given in Fig. 3. The diagram presents
the feed places and the flow rates of fuel B (in fractions of
the total flow rate) and of air (in factions of V0), the relative
distances from the initial sections of ZAC (X1) and ZDB
(X2), and the running excess air factors á = Óái/ÓBi. The
main design parameters of the firing process, i.e., the temperature of the flame J, the combustion efficiency â, the mechanical q4 and chemical q3 underfiring, the concentrations
of CO and C NO x (at á = 1.4), and the heat flows qt over the
height of the furnace, are presented in Table 1.
B = 0.9
Fig. 3. Design model of a furnace.
It can be seen from Table 1 that despite the unfavorable
conditions (the content of volatiles in the fuel (Vg = 9%,
ø = 0.45), the stability of combustion of low-reaction coal is
ensured by partial gasification of the fuel in the first two regions of ZAC. In this case the combustion efficiency is high
and the concentration of nitrogen and carbon oxides is low.
In addition, the computed temperature of gases at the outlet
from the furnace is also lower than the design value by
60 – 70°C due to the involvement of the sloping bottom into
the process, which implies the presence of a margin in the
steam rate of the boiler. This example shows that such computational studies make it possible to estimate the expediency of new engineering solutions and the expected parameters of the designed boiler.
The “TORKA” software makes it possible to compute
the parameters of the firing process not only for stationary
combustion modes but also for the modes of extinction and
inflammation. Figure 4 presents a diagram of the thermal regime of combustion of fuel in a flow in the coordinates â – è
(è = RT/E, where R is the gas constant, T is the temperature,
and E is the activation energy) [8 – 10]. The points in the diagram are the points of intersection and tangency of the
TABLE 1. Examples of Combustion Process in a Boiler Firing Lean Coal
X1
X2
B
á
J, °C
â, %
q 4, %
q3, %
CO, %
CNO x , mg/m3
që, Mcal/(m2 · h)
0.22
0.67
1.0
–
–
–
–
–
–
0.33
0.66
1.0
0.9
1.0
1.0
–
–
–
0.55
0.70
1.15
1.15
1.15
1.15
1376
1433
1364
1328
1259
1197
33.49
62.14
83.59
95.84
97.62
98.31
1.08
1.86
11.23
3.44
2.13
1.60
55.53
36.00
5.18
0.72
0.25
0.09
25.34
13.04
1.25
0.17
0.06
0.02
88
245
502
554
562
565
163
181
159
145
115
93
182
A. A. Shatil’
â
1
q3 = 0 is introduced into Table 1 of initial data for computing
the parameters of the limiting mode in which the computation stops upon a change in this or that input parameter. At
the same time, “normal” computation at q3 > 0 is performed
at the same process parameters, which corresponds to the
state of the process before the moment of extinction. These
two modes are “realized” in the algorithm through a kinetic
coefficient m included into the Arrhenius criterion and calculated by the formula
C
Ex
2
4
0.5
m = m0[1 – 0.4(x + q3)],
In
3
0
3
3
0.05
è
Fig. 4. Diagram of the thermal regime of combustion: 1, heat liberation curve; 2, heat removal curves; 3, initial point of the process
(â = 0, è = è0), C, point of stationary regime; Ex, point of extinction;
In, point of inflammation; 4, point of crisis-free regime.
curves of heat liberation and heat removal, the parameters of
which correspond to these modes. Since the stationary modes
have been studied quite well, we will consider in detail the
critical operating conditions of actual furnaces. Note that after the publication of [6], works devoted to determination of
the temperature of extinction Tex in boilers appeared as well.
A computational model of the firing process makes it
possible to analyze the modes of extinction in different kinds
of boilers with allowance for the numerous parameters of the
firing process. Extinction of the flame in furnaces of traditional boilers with burners depends primarily on the kinetics
of combustion and can be caused by growth in the excess air
factor in the burners ág (at ág > 1), in the fraction of the recycling flue gases r, in the coefficient of thermal efficiency of
the waterwalls ø, and in the thermal stress of the furnace volume qv, or by decrease in the temperature of hot air th.a, in the
content of volatiles in the fuel Vg, in the time of residence of
combustion products in ZAC ôr, in the content of oxygen in
the oxidizer O¢2 , etc. On the other hand, the stability of combustion depends on the aerodynamic structure of the flow
near the burners and in ZAC. The sizes of the zones of
recirculation play a substantial role. In operation extinction
of the flame is often caused by an abrupt decrease in the feed
of fuel (disabling of the mills or pulverized coal feeders or
reduction in the combustion value of the fuel), i.e., by step
growth in the excess air factor in the furnace.
Computation with the use of the “TORKA” software
gives the full volume fraction of flue gases in percent and the
value of chemical underfiring q3 in regions of the furnace.
Firing of organic fuel is always accompanied by the presence
of products of incomplete combustion of CO and H2. Therefore, the algorithm stipulates the possibility of determination
of the parameters of the firing process at the moment of
extinction, which is characterized by disappearance of these
gases, i.e., chemical underfiring in ZAC. The condition
(3)
where x is the fraction of “lighting” (highly reactive) fuel in
the total consumption of fuel per boiler (with respect to the
heat). The coefficient m0 for ZAC depends on the type of furnace; for ZDB it is assumed that m0 = 1 and x = 0 [9]. This
simulates a fluctuating combustion process from q3 > 0 to
q3 = 0 (and backward) observed in operation as pulsations of
pressure and temperature in the furnace before extinction of
the flame.
It is presumed in the chosen model of firing (Fig. 4) that
a stationary process with crisis moments of extinction and inflammation under specific conditions can transform into a
crisis-free process without extinction [10]. Transfer from a
crisis form of the â – T dependence to a crisis-free form is
possible, in principle, in one and the same device. Specifically, this can be ensured by increasing the initial temperature of the process T0, using an auxiliary (pilot) flame of
highly reactive fuel (i.e., reducing the coefficient m, which
lowers the “effective” activation energy E), or decreasing the
removal of heat through the walls of the furnace chamber
(reduction of the coefficient ø). A powerful factor of stabilization of combustion is the creation of zones with á < 1 and
zones with combustion of lumped solid fuel in a fluidized (or
another) layer where the chemical underfiring is inconsiderable. It follows from Eq. (3) that this is similar to “lighting”
(x > 0).
The computational dependences of the extinction temperature Tex of the flame of pulverized coal on the content of
volatiles in the fired fuel Vg plotted for several boilers firing
pulverized coal with dry-ash and slag-tap removal, which are
presented in Fig. 5 [8], allow us to write the following equation
Tex = 1660 – 8.3Vg
(4)
Tex = 0.07E/R + 410.
(5)
or
Figure 5 also presents the relation between the computed
values of the temperature Tex and the combustion efficiency
âex at the moment of extinction, which is described by the
equation
Tex = 554/(1 – âex).
(6)
Design Study of Boiler Furnaces
183
Tex, K
This dependence can be written in the form
Tex, K
a
b
9
b ex =
1106 - 8.3V g
1660 - 8.3V g
1500
,
(7)
from which we find that at Vg = 0 (coke, E = 36 ´ 104
kcal/mole) and at Vg = 100% (natural gas, E = 12 ´ 103
kcal/mole) the values of âex amount to 0.67 and 0.33 respectively. Note that these values of âex coincide with the values
of â for extinction and inflammation of a flame of pulverized
coal obtained in [18] on the basis of other assumptions. However, it follows from (7) that the value âex = 2/3 is not constant and decreases to 1/3 upon growth in the reactivity of
the fuel.
On the whole, the thus obtained relationship between the
extinction parameters can be treated as a universal one. However, both these parameters (Tex and âex) are “virtual” and
cannot be used in practice because of the impossibility of
their representative estimation. Therefore, in order to predict
the conditions of stability of combustion we should compute
regime parameters of the process on which it depends and
which should be measurable. The computation should allow
for the following additional circumstances. The first circumstance is the effect of not only the temperature but also the
time ôr of residence of combustion products in the regions of
ZAC on the processes of combustion and extinction. In different flame furnaces it can differ by several times [9].
The effect of the residence time ôr (in seconds) on Tex can
be allowed for by an empirical coefficient determined from
the formula
kô = 1.05 – 0.05ôr.
(8)
Consequently, the temperature Tex at âex > 1 will be
lower and at âex < 1 will be higher than at âex = 1 sec. The
second circumstance is connected with the instability of
feeding of solid fuel by the feeders of pulverized coal [18].
This requires what is known as a margin Äá in the computation of the limiting air excess factor at the moment of extinction [8, 9]. For example, ±3% fluctuations of the rate of fuel
feed (accordingly, Äá = 0.03) can raise the extinction temperature by 80 – 100 K.
Evaluation of the parameters of crisis modes of inflammation does not present great practical interest. It is much
more important to be able to compute the parameters of the
mode of pilot firing of the boiler. In such a computation the
initial data are the air excess factor and the air temperature,
the rates of the main fuel and of the highly reactive fuel
(black oil, gas), and other parameters characterizing the performance of the burners and of the furnace.
Let is consider as an example the mode of pilot firing of
a TPP-210A twin-furnace boiler firing pulverized coal of
grade ASh. It is assumed that the solid fuel is fed into the furnace when the air is heated by the highly reactive fuel to
1000
500
0
9
4
12 3 11 10
2 1
6
11 4
12
3
10 2
6 1
1500
5
13 7
50
5
7
8
V g, %
1000
500
0
13
8
50
V g, %
Fig. 5. Dependences of computed temperature of extinction on the
content of volatiles in the fuel (a) and on the combustion efficiency
(b) at the moment of extinction for different kinds of furnace, grades
of coal, and types of ash removal: 1, TP-109, G, solid; 2, TP-240, SS,
solid; 3, TP-214, SS, solid; 4, TP-215, SS, solid; 5, BKZ-210, B1,
solid; 6, BKZ-500, B2, solid; 7, PK-10, peat, solid; 8, PK-10, lead,
solid; 9, TP-210A, ASh, liquid; 10, TP-100, G, liquid; 11, TP-87, SS,
liquid; 12, TPB-318, ASh, solid; 13, BKZ-670, B2, solid.
100°C. The rate of feeding of pulverized coal to two burners
out of the six available ones is 10 tons/h (18% of the rated
feeding); that of black oil is about 1 ton/h (x = 0.20). Since
in the pilot firing the flame does not fill the whole of the
cross section of the ZAC, the critical coefficient m0 is taken
to be equal to 0.85 as in low-stress furnaces with bottom (or
elevated) burners [9]. The coefficients of thermal efficiency
of the waterwalls ø are taken to be equal to 0.20 for the ZAC
and 0.45 for the ZDB. The degree of blackness of the furnace
af is determined using formula (2), in which afl = 0.9 and
Fw/Ffl = 4 for the ZAC and afl = 0.5 and Fw/Ffl = 3 for the
ZDB.
The decisive moment in pilot firing is the choice of the
maximum value of the excess air factor. It is assumed in the
initial data for computation with the help of “TORKA” that
in the first region of the ZAC with a height of 5 m á = 1.5,
whereas at the outlet from the second region (H = 8.95 m)
á = 1.70. This increase in the excess air factor is caused by
the arrival of air into the furnace through dead burners. Computations have shown that the flame in the ZAC is stable (JZAC
= 1067 – 1086°C, âZAC = 74.90 – 80.44%) and underfiring at
the outlet from the furnace is about 10%. A check of stability
of combustion (at q3 = 0) has shown that in the first region of
the ZAC combustion is absent and the fuel starts to burn only
in the second region. This means that in the given case pilot
firing at á > 1.70 in the ZAC is impossible with a heat fraction of black oil x = 0.20. Therefore, it is desirable either to
raise the temperature of the hot air, i.e., the boiler should fire
black oil for a longer period, or to increase the proportion of
the highly reactive fuel. This example shows how the personnel can be informed on the conditions of pilot firing of a specific boiler.
184
CONCLUSIONS
1. Development of new technologies of firing organic
fuel and improvement of the operating furnace devices requires creation of a method for their computational study.
2. The computational model of the firing process developed by TsKTI has been tested on a great number of furnaces
of operating boilers firing a wide spectrum of fuel in all firing stages from pilot firing to extinction and on various furnace devices. This allows us to recommend the model and
the “TORKA” software for computational studies not only of
traditional boiler furnaces but also of furnaces employing
novel firing technologies.
3. It is expedient to have at least two modifications of
the software for practical purposes. One of them can be used
for design studies of furnaces and the other (simpler) modification can be used for an operating furnace (or a boiler as a
whole) for estimating the consequences of various possible
working situations.
REFERENCES
1. Thermal Design of Boiler Units (Standard Method) [in Russian],
Gosénergoizdat, Moscow (1957).
2. Thermal Design of Boiler Units (Standard Method) [in Russian],
Énergiya, Moscow (1973).
3. Thermal Design of Boilers (Standard Method) [in Russian], NPO
TsKTI, St. Petersburg (1998).
4. É. Kh. Verbovetskii and N. G. Zhmerik (eds.), Methodological
Recommendations on Design of Furnace Devices [in Russian],
Izd. NPO TsKTI, St. Petersburg (1996).
5. V. I. Antonovskii, “Heat transfer in furnaces of steam boilers.
Retrospective consideration of the development of the standard
computational technique,” Teploénergetika, No. 9 (2004).
A. A. Shatil’
6. A. A. Shatil and E. K. Chavchanidze, “Computational estimation of stability of flame due to firing solid fuels in boiler furnaces,” Teploénergetika, No. 4 (1990).
7. A. A. Shatil and E. Ya. Skripova, “About design of furnaces of
boilers firing pulverized coal,” Teploénergetika, No. 9 (1993).
8. A. A. Shatil, Furnace Processes and Equipment [in Russian],
Izd. TsKTI, St. Petersburg (1997).
9. A. A. Shatil, Computational Study of Furnace Devices [in Russian], Izd. TsKTI, St. Petersburg (2003).
10. L. A. Vulis, Thermal Regime of Combustion [in Russian], Gosénergoizdat, Moscow – Leningrad (1954).
11. N. Ya. Fabrikant, Aerodynamics [in Russian], Nauka, Moscow
(1964).
12. G. N. Abramovich, Turbulent Free Jets of Liquids and Gases [in
Russian], Gosénergoizdat, Moscow – Leningrad (1948).
13. Yu. V. Ivanov, Fundamentals of Computation and Design of
Gas Burners [in Russian], Gosénergoizdat, Moscow (1963).
14. A. A. Shatil and B. N. Barbyshev, “A study of the aerodynamics
of a furnace with inverted flame,” Énerg. Életrifik., No. 2
(1983).
15. A. A. Shatil, V. P. Maistruk, A. E. Surkov et al., “A study of the
method of inverted combustion of coals of grades D and G and
their intermediate product in TP-230-3 boiler,” Élektr. Stantsii,
No. 1 (1986).
16. N. S. Klepikov, L. N. Gusev, A. A. Shatil et al., “Experience of
installation of undergrate blast system designed by Izd. NPO
TsKTI on low-, medium-, and high-capacity boilers,” Trudy
TsKTI, Issue 287.
17. G. G. Olkhovskii and A. G. Tumanovskii, “Problems and prospects of the use of coal in the power industry of Russia,” Énergetik, No. 12 (2004).
18. L. N. Gusev, “A practical method for determining the temperatures of inflammation and extinction of pulverized coal flame in
chamber furnaces,” Tyazh. Mashinostr., No. 6 (2002).
