chapter

6
Coupling Chemical and Thermal
Analyses of Reacting Systems

OVERVIEW
In Chapter 2, we reviewed the thermodynamics of reacting systems, considering
only the initial and final states. For example, the concept of an adiabatic flame
temperature was derived based on the knowledge of the initial state of the reactants
and the final composition of the products, as determined by equilibrium. Perform-
ing an adiabatic flame temperature calculation required no knowledge of chemical
rate processes. In this chapter, we couple the knowledge gained of chemical kinet-
ics in Chapter 4 with fundamental conservation principles (e.g., mass and energy
conservation) for various archetypal thermodynamic systems. This coupling al-
lows us to describe the detailed evolution of the system from its initial reactant
state to its final product state, which may or may not be in chemical equilibrium.
In other words, we will be able to calculate the system temperature and the various
species concentrations as functions of time as the system proceeds from reactants
to products.
Our analyses in this chapter will be simple, without the complication of
mass diffusion. The systems that are chosen for study in this chapter, shown in
Fig. 6.1, make bold assumptions about the mixedness of the system. Three of the
four systems assume that the systems are perfectly mixed and homogeneous in
composition; the fourth system, the plug-flow reactor, totally ignores mixing and
diffusion in the flow (axial) direction, while assuming perfect mixedness in the
radial direction perpendicular to the flow. Although the concepts developed here
can be used as building blocks for modeling more complex flows, perhaps more
importantly, they are pedagogically useful for developing a very basic understand-
ing of the interrelationships among thermodynamics, chemical kinetics, and fluid
mechanics. In the next chapter, we will extend our simple analysis to include the
effects of mass diffusion.
183

tur80199_ch06_183-219.indd 183 1/4/11 12:18:06 PM

 184 C H A P T E R 6 • Coupling Chemical and Thermal Analyses

Figure 6.1 Simple chemically reacting systems: (a) constant-pressure, fixed mass;
(b) constant-volume, fixed mass; (c) well-stirred reactor; (d) plug-flow reactor.

CONSTANT-PRESSURE, FIXED-MASS REACTOR

Application of Conservation Laws

Consider reactants contained in a piston–cylinder arrangement (Fig. 6.1a) that
react at each and every location within the gas volume at the same rate. Thus,
there are no temperature or composition gradients within the mixture, and a sin-
gle temperature and set of species concentrations suffice to describe the evolution
of this system. For exothermic combustion reactions, both the temperature and
volume will increase with time, and there may be heat transfer through the reaction
vessel walls.
In the following, we will develop a system of first-order ordinary differential
equations whose solution describes the desired temperature and species evolution.

tur80199_ch06_183-219.indd 184 1/4/11 12:18:07 PM

 Constant-Pressure, Fixed-Mass Reactor 185

These equations and their initial conditions define an initial-value problem.

Starting with the rate form of the conservation of energy for a fixed-mass system,
we write
du
Q − W = m . (6.1)
dt
Applying the definition of enthalpy, h ≡ u + Pv, and differentiating gives
du d h dv
= −P . (6.2)
dt dt dt
Assuming the only work is the P−dv work at the piston,
W dv
=P . (6.3)
m dt
Substituting Eqns. 6.2 and 6.3 into Eqn. 6.1, the P dv / d t terms cancel leaving
Q dh (6.4)
= .
m dt
We can express the system enthalpy in terms of the system chemical composition as
N

H
∑ Ni hi (6.5)
i =1
h= = ,
m m
where Ni and hi are the number of moles and molar enthalpy of species i, respectively.
Differentiation of Eqn. 6.5 yields

d h 1 ⎡ ⎛ d Ni ⎞ ⎛ dh ⎞ ⎤
= ⎢ ∑ ⎜ hi ⎟ + ∑ ⎜ Ni i ⎟ ⎥ . (6.6)
dt m ⎣ i ⎝ dt ⎠ i ⎝ dt ⎠ ⎦

Assuming ideal-gas behavior, i.e., hi = hi (T only),

d hi ∂hi d T dT
= = cp, i , (6.7)
dt ∂T d t dt
where cp, i is the molar constant-pressure specific heat of species i. Equation 6.7
provides the desired link to the system temperature, and the definition of the molar
concentration [Xi] and the mass-action expressions, ω i = …, provide the necessary link
to the system composition, Ni , and chemical dynamics, d Ni / d t. These expressions are

Ni = V [ Xi ] (6.8)
d Ni
≡ Vω i , (6.9)
dt
where the ω i values are calculated from the detailed chemical mechanism as dis-
cussed in Chapter 4 (see Eqns. 4.31– 4.33).

tur80199_ch06_183-219.indd 185 1/4/11 12:18:07 PM

 186 C H A P T E R 6 • Coupling Chemical and Thermal Analyses

Substituting Eqns. 6.7–6.9 into Eqn. 6.6, our statement of energy conservation
(Eqn. 6.4) becomes, after rearrangement,

(Q / V ) − ∑ (hiω i )
dT
= i
, (6.10)
dt ∑ ([ X i ] cp, i )
i

where we use the following calorific equation of state to evaluate the enthalpies:
T
hi = h fo, i + ∫ cp, i dT . (6.11)
Tref

To obtain the volume, we apply mass conservation and the definition of [Xi] in Eqn. 6.8:
m
V= . (6.12)
∑([ X i ]MWi )
i

The species molar concentrations, [Xi], change with time as a result of both
chemical reactions and changing volume, i.e.,

d[ X i ] d( Ni / V ) 1 d N i 1 dV
= = − Ni 2 (6.13a)
dt dt V dt V dt
or
d[ X i ] 1 dV
= ω i − [ X i ] , (6.13b)
dt V dt
where the first term on the right-hand side is the chemical production term and the
second term accounts for the changing volume.
The ideal-gas law can be used to eliminate the dV / d t term. Differentiating

PV = ∑ N i Ru T (6.14a)
i

for the case of constant pressure, and rearranging, yields

1 dV 1 d Ni 1 dT
=
V d t ∑ Ni
∑ +
dt T dt
. (6.14b)
i
i

First substituting Eqn. 6.9 into Eqn. 6.14b and then substituting the result into
Eqn. 6.13b, provides, after rearrangement, our final expression for the rate of change
of the species molar concentrations:

d[ X i ] ⎡ ∑ ω i 1 dT ⎤
= ω i − [ X i ] ⎢ ∑ [ X ] + T d t ⎥ . (6.15)
dt ⎢⎣ j j ⎥⎦

tur80199_ch06_183-219.indd 186 1/4/11 12:18:08 PM

 Constant-Volume, Fixed-Mass Reactor 187

Reactor Model Summary

Succinctly stated, our problem is to find the solution to

dT
= f ([ X i ], T ) (6.16a)
dt

d[ X i ]
= f ([ X i ], T ) i = 1, 2, … , N (6.16b)
dt
with initial conditions
T (t = 0) = T0 (6.17a)

and
[ X i ](t = 0) = [ X i ] 0 . (6.17b)

The functional expressions for Eqns. 6.16a and 6.16b are obtained from Eqn. 6.10
and Eqn. 6.15, respectively. Enthalpies are calculated using Eqn. 6.11, and the vol-
ume is obtained from Eqn. 6.12.
To carry out the solution of the above system, an integration routine capable of
handling stiff equations should be employed, as discussed in Chapter 4.

CONSTANT-VOLUME, FIXED-MASS REACTOR

Application of Conservation Laws

The application of energy conservation to the constant-volume reactor follows
closely that of the constant-pressure reactor, with the major difference being the ab-
sence of work in the former. Starting with Eqn. 6.1, with W = 0, the first law takes
the following form:

d u Q (6.18)
= .
dt m
Recognizing that the specific internal energy, u, now plays the same mathematical
role as the specific enthalpy, h, in our previous analysis, expressions equivalent
to Eqns. 6.5–6.7 are developed and substituted into Eqn. 6.18. This yields, after
rearrangement,

(Q / V ) − ∑ (uiω i )
dT
= i
. (6.19)
dt ∑ ([ X i ] cv, i )
i

tur80199_ch06_183-219.indd 187 1/4/11 12:18:08 PM

 188 C H A P T E R 6 • Coupling Chemical and Thermal Analyses

Recognizing that, for ideal gases, ui = hi − Ru T and cv, i = cp, i − Ru , we can express
Eqn. 6.19 using enthalpies and constant-pressure specific heats:

(Q / V ) + Ru T ∑ ω i − ∑ (hiω i )
dT
= i i
. (6.20)
dt ∑ [[ X i ](cp, i − Ru )]
i

In constant-volume explosion problems, the time-rate-of-change of the pressure is

of interest. To calculate dP / dt, we differentiate the ideal-gas law, subject to the con-
stant volume constraint, i.e.,
PV = ∑ N i Ru T (6.21)
i

and

d ∑ Ni
dP dT
V = Ru T i
+ Ru ∑ N i . (6.22)
dt dt i dt
Applying the definitions of [Xi] and ω i (see Eqns. 6.8 and 6.9), Eqns. 6.21 and 6.22
become
P = ∑ [ X i ]Ru T (6.23)
i

and
dP dT
= Ru T ∑ ω i + Ru ∑ [ X i ] , (6.24)
dt i i dt
which completes our simple analysis of homogeneous constant-volume combustion.

Reactor Model Summary

Equation 6.20 can be integrated simultaneously with the chemical rate expressions to
determine T(t) and [Xi](t), i.e.,
dT
= f ([ X i ], T ) (6.25a)
dt
d[ X i ]
= ω i = f ([ X i ], T ) i = 1, 2, … , N (6.25b)
dt
with initial conditions
T (t = 0) = T0 (6.26a)
and
[ X i ](t = 0) = [ X i ]0 . (6.26b)

The required enthalpies are evaluated using Eqn. 6.11, and the pressure from
Eqn. 6.23. Again, a stiff equation solver should be used to carry out the integration.

tur80199_ch06_183-219.indd 188 1/4/11 12:18:08 PM

 Constant-Volume, Fixed-Mass Reactor 189

In spark-ignition engines, knock occurs when the unburned fuel–air mixture ahead of the Example 6.1
flame reacts homogeneously, i.e., it autoignites. The rate-of-pressure rise is a key parameter
in determining knock intensity and propensity for mechanical damage to the piston–crank
assembly. Pressure-versus-time traces for normal and knocking combustion in a spark-ignition
engine are illustrated in Fig. 6.2. Note the very rapid pressure rise in the case of heavy knock.
Figure 6.3 shows schlieren (index-of-refraction gradient) photographs of flame propagation
for normal and knocking combustion.
Create a simple constant-volume model of the autoignition process and determine the
temperature and the fuel and product concentration histories. Also determine dP / dt as a func-
tion of time. Assume initial conditions corresponding to compression of a fuel–air mixture
from 300 K and 1 atm to top-dead-center for a compression ratio of 10:1. The initial volume
before compression is 3.68 ⋅ 10−4 m3, which corresponds to an engine with both a bore and a
stroke of 75 mm. Use ethane as fuel.
Solution
We will make some bold and sweeping assumptions about the thermodynamics and the chemi-
cal kinetics to keep the computational complexity to a minimum. Our solution, however, will
still retain the strong coupling between the thermochemistry and chemical kinetics. Our as-
sumptions are as follows:
i. One-step global kinetics using the rate parameters for ethane C2H6 (see Table 5.1).
ii. The fuel, air, and products all have equal molecular weights; i.e., MWF = MWOx = MWPr = 29.
iii. The specific heats of the fuel, air, and products are constants and equal; that is, cp, F =
cp, Ox = cp, Pr = 1200 J / kg-K.
iv. The enthalpy of formation of the air and products are zero; the enthalpy of formation of
the fuel is 4 · 107 J / kg.
v. We assume that the stoichiometric air–fuel ratio is 16.0 and restrict combustion to stoi-
chiometric or lean conditions.
The use of global kinetics is hard to justify for a problem like engine knock where detailed
chemistry is important [3]. Our only justification is that we are trying to illustrate principles,
recognizing that our answers may be inaccurate in detail. Assumptions ii–iv provide values

Figure 6.2 Cylinder pressure-versus-time measurements in a spark-ignition engine for

normal combustion, light knock, and heavy knock cycles. The crank angle interval of 40°
corresponds to 1.67 ms.
SOURCE: Adapted from Refs. [1] and [17] by permission of McGraw-Hill, Inc.

tur80199_ch06_183-219.indd 189 1/4/11 12:18:09 PM

tur80199_ch06_183-219.indd 190
Figure 6.3 Schlieren photographs from high-speed movies of (a) normal combustion and (b) knocking combustion.
Pressure–time traces corresponding to the photographs are also shown.
SOURCE: From Refs. [2] and [17]. Reprinted by permission of McGraw-Hill, Inc.

1/4/11 12:18:09 PM
 Constant-Volume, Fixed-Mass Reactor 191

that give reasonable estimates of flame temperatures, yet trivialize the problem of obtaining
thermodynamic properties [4].
With these assumptions, we can now formulate our model. From Eqn. 5.2 and Table 5.1,
the fuel (ethane) reaction rate is

−15, 098 ⎞
= − 6.19 ⋅10 9 exp ⎛⎜
d[ F ]
⎝ ⎟⎠ [ F ]0.1[O 2]1.65 (I)
dt T
[= ] kmol /m -s,3

where, assuming 21% percent O2 in the air,

[O 2 ] = 0.21[Ox ].

Note the conversion of units from a gmol-cm3 to a kmol-m3 basis for the pre-exponential
factor (1.1 · 1012 · [1000]1−0.1−1.65 = 6.19 · 109).
We can simply relate the oxidizer and product reaction rates to the fuel rate through the
stoichiometry (assumptions ii and v):

d[Ox ] MWF d[ F ] d[ F ]
= ( A /F )s = 16 (II)
dt MWOx d t dt

and
d[ Pr ] MWF d[F ] d[F ]
= −[( A / F )s + 1] = −17 . (III)
dt MWPr d t dt

We complete our model by applying Eqn. 6.20:

d T (Q / V ) + RuT ∑ ω i − ∑(hiω i )

= .
dt ∑ [[ X i ](cp, i − Ru )]

This simplifies, by noting that

Q / V = 0 (adiabatic),
∑ ω i = 0 (assumptions ii and v),
∑ hiω i = ω F h fo, F (assumptions ii–v),
and
P P
∑ [ Xi](cp, i − Ru ) = (cp − Ru )∑[ Xi] = (cp − Ru )∑ χi R T = (cp − Ru ) R T ,
u u

to be

dT −ω F h fo, F
= . (IV)
d t (cp − Ru ) P / ( RuT )

Although our basic model is complete, we can add ancillary relations for the pressure and
pressure-derivative. From Eqn. 6.23 and 6.24,

P = RuT ([ F ] + [Ox ] + [ Pr ]),

tur80199_ch06_183-219.indd 191 1/4/11 12:18:12 PM

 192 C H A P T E R 6 • Coupling Chemical and Thermal Analyses

or
T
P = P0
T0
and
d P P d T P0 d T
= = .
d t T d t T0 d t
Before we can integrate our system of first-order ordinary differential equations
(Eqns. I–IV), we need to determine initial conditions for each of the variables: [F ], [Ox], [Pr],
and T. Assuming isentropic compression from bottom-dead-center to top-dead-center and a
specific heat ratio of 1.4, the initial temperature and pressure can be found:
γ −1 1.4 −1
⎛V ⎞
= 300 ⎛⎜ ⎞⎟
10
T0 = TTDC = TBDC ⎜ BDC ⎟ = 753 K
⎝ VTDC ⎠ ⎝ 1⎠

and
γ
⎛V ⎞ 10 1.4
P0 = PTDC = PBDC ⎜ BDC ⎟ = (1) ⎛⎜ ⎞⎟ = 25.12 atm.
⎝ VTDC ⎠ ⎝ 1⎠

The initial concentrations can be found by employing the given stoichiometry. The oxi-
dizer and fuel mole fractions are
( A /F )s / Φ
χ Ox , 0 = ,
[( A /F )s / Φ] + 1
χ Pr , 0 = 0,
χ F , 0 = 1 − χ Ox , 0 .

The molar concentrations, [Xi] = ci P / (RuT ), are

⎡ ( A /F )s / Φ ⎤ P0
[Ox ]0 = ⎢ ⎥ ,
⎣ [( A /F )s / Φ] + 1 ⎦ RuT0
⎡ ( A /F )s / Φ ⎤ P0
[ F ]0 = ⎢1 − ⎥ ,
⎣ [( A /F )s / Φ] + 1 ⎦ RuT0
[ Pr ]0 = 0.

Equations I–IV were integrated numerically, and the results are shown in Fig. 6.4. From
this figure, we see that the temperature increases only about 200 K in the first 3 ms, while,
thereafter, it rises to the adiabatic flame temperature (c. 3300 K) in less than 0.1 ms. This rapid
temperature rise and concomitant rapid consumption of the fuel is characteristic of a thermal
explosion, where the energy released and the temperature rise from reaction feeds back to
produce ever-increasing reaction rates because of the [−Ea / RuT ] temperature dependence of
the reaction rate. From Fig. 6.4, we also see the huge pressure derivative in the explosive stage,
with a peak value of about 1.9 · 1013 Pa / s.
Comments
Although this model predicted the explosive combustion of the mixture after an initial period of
slow combustion, as is observed in real knocking combustion, the single-step kinetics mechanism

tur80199_ch06_183-219.indd 192 1/4/11 12:18:13 PM

 Constant-Volume, Fixed-Mass Reactor 193

Figure 6.4 Results for the constant-volume reactor model of Example 6.1. Temperature, fuel,
and products concentrations, and rate-of-pressure rise (dP / dt) are shown. Note the expansion of
the time scale at 3 ms and again at 3.09 ms allows the explosion to be resolved.

does not model the true behavior of autoigniting mixtures. In reality, the induction period, or
ignition delay, is controlled by the formation of intermediate species, which subsequently react.
Recall the three basic stages of hydrocarbon oxidation presented in Chapter 5. To accurately
model knock, a more detailed mechanism would be required. Ongoing research efforts aim at
elucidating the details of the “low-temperature” kinetics of the induction period [3].
Control of engine knock has always been important to performance improvements, and,
more recently, has received attention because of legislated requirements to remove lead-based
antiknock compounds from gasoline.

tur80199_ch06_183-219.indd 193 1/4/11 12:18:14 PM

 194 C H A P T E R 6 • Coupling Chemical and Thermal Analyses

WELL-STIRRED REACTOR
The well-stirred, or perfectly-stirred, reactor is an ideal reactor in which perfect
mixing is achieved inside the control volume, as shown in Fig. 6.5. Experimental
reactors employing high-velocity inlet jets approach this ideal and have been used to
study many aspects of combustion, such as flame stabilization [5] and NOx formation
[6–8] (Fig. 6.6). Well-stirred reactors have also been used to obtain values for glo-
bal reaction parameters [9]. The well-stirred reactor is sometimes called a Longwell
reactor in recognition of the early work of Longwell and Weiss [5]. Chomiak [10]
cites that Zeldovich [11] described the operation of the well-stirred reactor nearly a
decade earlier.

Application of Conservation Laws

To develop the theory of well-stirred reactors, we review the concept of mass con-
servation of individual species. In Chapter 3, we developed a species conservation
equation for a differential control volume. We now write mass conservation for an
arbitrary species i, for an integral control volume (see Fig. 6.5), as

d mi, cv
= m i′′′ V + m i, in − m i, out . (6.27)
dt
Rate at which mass Rate at which Mass flow Mass flow
of i accumulates mass of i is of i into of i out of
within control generated within control volume control volume
volume control volume

Figure 6.5 Schematic of a well-stirred reactor.

tur80199_ch06_183-219.indd 194 1/4/11 12:18:14 PM

 Well-Stirred Reactor 195

Figure 6.6 Longwell reactor with one hemisphere removed. Fuel–air mixture enters
through small holes in central hollow steel sphere, and products exit through larger holes in
firebrick lining. Scale shown is in inches.
SOURCE: From Ref. [5]. Reprinted by permission, © The American Chemical Society.

What distinguishes Eqn. 6.27 from the overall continuity equation is the presence
of the generation term m i′′′V . This term arises because chemical reactions transform
one species into another; hence, a positive generation rate indicates the formation of
a species, and a negative generation rate signifies that the species is being destroyed
during the reaction. In the combustion literature, this generation term is frequently
referred to as a source or sink. When the appropriate form of Eqn. 6.27 is written for
each of the species in the reactor (i = 1, 2, . . . , N), the sum of these equations yields
the familiar form of the continuity equation,
d mcv
= m in − m out. (6.28)
dt
The mass generation rate of a species, m i′′′, is easily related to the net production
rate, ω i , developed in Chapter 4:
m i′′′ = ω i MWi . (6.29)
Ignoring any diffusional flux, the individual species mass flowrate is simply the
product of the total mass flowrate and that species mass fraction, i.e.,
m i = mY
 i. (6.30)

tur80199_ch06_183-219.indd 195 1/4/11 12:18:14 PM

 196 C H A P T E R 6 • Coupling Chemical and Thermal Analyses

When we apply Eqn. 6.27 to the well-stirred reactor, assuming steady-state opera-
tion, the time derivative of the left-hand side disappears. With this assumption, and
substituting Eqns. 6.29 and 6.30, Eqn. 6.27 becomes

ω i MWi V + m (Yi, in − Yi , out ) = 0 for i = 1, 2, … , N species. (6.31)

Furthermore, we can identify the outlet mass fractions, Yi, out, as being equal to
the mass fractions within the reactor. Since the composition within the reactor is
everywhere the same, the composition at the outlet of the control volume must be
the same as in the interior. With this knowledge, the species production rates are
of the form

ω i = f ([ X i ]cv, T ) = f ([ X i ]out, T ), (6.32)

where the mass fractions and molar concentrations are related by

[ X i ]MWi
Yi = N . (6.33)
∑ [ Xj ]MWj
j =1

Equation 6.31, when written for each species, provides N equations with N + 1 un-
knowns, with the assumed known parameters m and V. An energy balance provides
the additional equation needed for closure.
The steady-state, steady-flow conservation of energy equation (Eqn. 2.28) applied
to the well-stirred reactor is

Q = m (hout − hin ), (6.34)

where we neglect changes in kinetic and potential energies. Rewriting Eqn. 6.34 in
terms of the individual species, we obtain

⎛ N N ⎞
Q = m ⎜ ∑ Yi, out hi (T ) − ∑ Yi, in hi (Tin )⎟ , (6.35)
⎝ i =1 i =1 ⎠

where
T
hi (T ) = h of , i + ∫ cp, i dT . (6.36)
Tref

Solving for the temperature, T, and species mass fractions, Yi, out, is quite similar to our
computation of equilibrium flame temperatures in Chapter 2; however, now the product
composition is constrained by chemical kinetics, rather than by chemical equilibrium.
It is common in the discussion of well-stirred reactors to define a mean resi-
dence time for the gases in the reactor:

t R = ρV / m (6.37)

tur80199_ch06_183-219.indd 196 1/4/11 12:18:20 PM

 Well-Stirred Reactor 197

where the mixture density is calculated from the ideal-gas law,

ρ = PMWmix / Ru T . (6.38)

The mixture molecular weight is readily calculated from a knowledge of the mixture
composition. Appendix 6A provides relationships between MWmix and Yi , ci , and [Xi].

Reactor Model Summary

Because the well-stirred reactor is assumed to be operating at steady state, there is
no time dependence in the mathematical model. The equations describing the reactor
are a set of coupled nonlinear algebraic equations, rather than a system of ordinary
differential equations (ODEs), which was the result for the previous two examples.
Thus, the ω i appearing in Eqn. 6.31 depends only on the Yi (or [Xi]) and temperature,
not time. To solve this system of N + 1 equations, Eqns. 6.31 and 6.35, the generalized
Newton’s method (Appendix E) can be employed. Depending on the chemical sys-
tem under study, it may be difficult to achieve convergence with Newton’s method
and more sophisticated numerical techniques may be necessary [12].

Develop a simplified model of a well-stirred reactor using the same simplified chemistry Example 6.2
and thermodynamics used in Example 6.1 (equal and constant cps and MWs, and one-step
global kinetics). Use the model to determine the blowout characteristics of a spherical
(80-mm-diameter) reactor with premixed reactants (C2H6–air) entering at 298 K. Plot the
equivalence ratio at blowout as a function of mass flowrate for Φ ≤ 1.0. Assume the reactor
is adiabatic.
Solution
Noting that the molar concentrations relate to mass fractions as

PMWmix Yi
[ Xi] = ,
RuT MWi

our global reaction rate, ω F , can be expressed as

m+n m n
d[F ] ⎛ PMWmix ⎞ ⎛ YF ⎞ ⎛ 0.2333YOx ⎞
ω F = = − kG ⎜ ⎜⎝ MW ⎟⎠ ⎜⎝ MW ⎟⎠ ,
dt ⎝ RuT ⎟⎠ F Ox

where m = 0.1 and n = 1.65, the factor 0.233 is the mass fraction of O2 in the oxidizer (air), and
the mixture molecular weight is given by

−1
⎡ Y Y Y ⎤
MWmix = ⎢ F + Ox + Pr ⎥ .
⎣ MWF MW Ox MW Pr ⎦

The global rate coefficient is, as in Example 6.1,

−15, 098 ⎞
kG = 6.19 ⋅10 9 exp ⎛⎜ ⎟⎠ .
⎝ T

tur80199_ch06_183-219.indd 197 1/4/11 12:18:20 PM

 198 C H A P T E R 6 • Coupling Chemical and Thermal Analyses

We can now write species conservation equations for the fuel by applying Eqn. 6.31 to give
0.1 1.65
⎛ YF ⎞ ⎛ 0.233YOx ⎞
⎛ P ⎞
1.75 ⎜⎝ MW ⎟⎠ ⎜⎝ MW ⎟⎠
f1 ≡ m (YF, in − YF ) − kG MWF V ⎜ F Ox
= 0,
⎝ RuT ⎟⎠
1.75
⎡ YF Y Y ⎤
⎢ + Ox + Pr ⎥
⎣ MWF MWOx MWPr ⎦
which further simplifies by applying our assumption of equal molecular weights and noting
that ∑Yi = 1:
1.75
⎛ P ⎞ YF0.1 (0.233YOx )1.65
f1 ≡ m (YF, in − YF ) − kG MW V ⎜ = 0. (I)
⎝ RuT ⎟⎠ 1

For the oxidizer (air),

1.75
⎛ P ⎞ YF0.1 (0.233YOx )1.65
f2 ≡ m (YOx , in − YOx ) − ( A / F )s kG MW V ⎜ = 0. (II)
⎝ RuT ⎟⎠ 1

For the product mass fraction, we write

f3 ≡ 1 − YF − YOx − YPr = 0. (III)

Our final equation in the model results from the application of Eqn. 6.35:

f4 ≡ YF ⎡⎣ h of , F + cp, F (T − Tref ) ⎤⎦

+ YOx ⎡⎣ h of , Ox + cp, Ox (T − Tref ) ⎤⎦

+ YPr ⎡⎣ h of , Pr + cp, Pr (T − Tref ) ⎤⎦

− YF, in ⎡⎣ h of , F + cp, F (Tin − Tref ) ⎤⎦

− YOx ,in ⎡⎣ h of , Ox + cp, Ox (Tin − Treef ) ⎤⎦ = 0,

which also further simplifies by the assumptions of equal specific heats and h of , Ox = h of , Pr = 0
to give

f4 ≡ (YF − YF, in )h of , F + cp (T − Tin ) = 0. (IV)

Equations I–IV constitute our reactor model and involve the four unknowns YF , YOx , YPr ,
and T and the parameter m.  To determine the reactor blowout characteristic, we solve the non-
linear algebraic equation set (I–IV) for a sufficiently small value of m that allows combustion
for a given equivalence ratio. We then increase m until we fail to achieve a solution, or until
the solution yields the input values. Figure 6.7 illustrates the results of such a procedure for
Φ = 1. The generalized Newton’s method (Appendix E) was used to solve the equation set.
In Fig. 6.7, we see the decreasing conversion of fuel to products and decreased tempera-
ture as the flowrate increases to the blowout condition (m > 0.193 kg / s). The ratio of the tem-
perature at blowout to the adiabatic flame temperature is given by (1738 K / 2381 K = 0.73),
which is in agreement with the results in Ref. [5].
Repeating the calculations at various equivalence ratios generates the blowout character-
istics shown in Fig. 6.8. Note that the reactor is more easily blown out as the fuel–air mixture

tur80199_ch06_183-219.indd 198 1/4/11 12:18:21 PM

 Well-Stirred Reactor 199

Figure 6.7 Effect of flowrate on conditions inside a model well-stirred reactor. For flowrates
greater than 0.193 kg / s, combustion cannot be sustained within the reactor (blowout).

becomes leaner. The shape of the blowout curve in Fig. 6.8 is similar to those determined for
experimental reactors and turbine combustors.

Comment
Well-stirred-reactor theory and experiments were used in the 1950s as a guide to the development
of high-intensity combustors for gas turbines and ramjets. This example provides a good illustra-
tion of how reactor theory can be applied to the problem of blowout. The blowout condition, plus
some margin of safety, determines the maximum-load condition for continuous-flow combustors.
Although well-stirred-reactor theory captures some of the characteristics of blowout, other theories
also have been proposed to explain flameholding. We explore this topic further in Chapter 12.

Use the following detailed kinetic mechanism for H2 combustion to investigate the behavior Example 6.3
of an adiabatic, well-stirred reactor operating at 1 atm. The reactant stream is a stoichiometric
(Φ = 1) mixture of H2 and air at 298 K, and the reactor volume is 67.4 cm3. Vary the residence
time between the long-time (equilibrium) and the short-time (blowout) limits. Plot the tem-
perature and the H2O, H2, OH, O2, O, and NO mole fractions versus residence time.

tur80199_ch06_183-219.indd 199 1/4/11 12:18:21 PM

 200 C H A P T E R 6 • Coupling Chemical and Thermal Analyses

Figure 6.8 Blowout characteristics of a model well-stirred reactor.

H– O – N mechanism

Forward Rate Coefficienta

No. Reaction A b Ea

lb H + O2 + M ↔ HO2 + M 3.61 ·1017 −0.72 0

2 H + H + M ↔ H2 + M 1.0 · 1018 −1.0 0
3 H + H + H2 ↔ H2 + H2 9.2 · 1016 −0.6 0
4 H + H + H2O ↔ H2 + H2O 6.0 · 1019 −1.25 0
5c H + OH + M ↔ H2O + M 1.6 · 1022 −2.0 0
6c H + O + M ↔ OH + M 6.2 · 1016 −0.6 0
7 O + O + M ↔ O2 + M 1.89 · 1013 0 −1,788
8 H2O2 + M ↔ OH + OH + M 1.3 · 1017 0 45,500
9 H2 + O2 ↔ OH + OH 1.7 · 1013 0 47,780
10 OH + H2 ↔ H2O + H 1.17 · 109 1.3 3,626
11 O + OH ↔ O2 + H 3.61 · 1014 −0.5 0
12 O + H2 ↔ OH + H 5.06 · 104 2.67 6,290
13 OH + HO2 ↔ H2O + O2 7.5 · 1012 0 0
14 H + HO2 ↔ OH + OH 1.4 · 1014 0 1,073
15 O + HO2 ↔ O2 + OH 1.4 · 1013 0 1,073
16 OH + OH ↔ O + H2O 6.0 · 108 1.3 0
17 H + HO2 ↔ H2 + O2 1.25 · 1013 0 0
18 HO2 + HO2 ↔ H2O2 + O2 2.0 · 1012 0 0

tur80199_ch06_183-219.indd 200 1/4/11 12:18:21 PM

 Well-Stirred Reactor 201

H– O – N mechanism (continued )

Forward Rate Coefficienta

No. Reaction A b Ea

19 H2O2 + H ↔ HO2 + H2 1.6 ·1012 0 3,800

20 H2O2 + OH ↔ H2O + HO2 1.0 · 1013 0 1,800
21 O + N2 ↔ NO + N 1.4 · 1014 0 75,800
22 N + O2 ↔ NO + O 6.40 · 109 1.0 6,280
23 OH + N ↔ NO + H 4.0 · 1013 0 0

a
The forward rate coefficient is of the form kf = AT b exp(−Ea / Ru T), where A is expressed in CGS units (cm, s,
K, gmol), b is dimensionless with T in Kelvins, and Ea is given in cal / gmol.
b
When H2O and H2 act as collision partners M, the rate coefficient is multiplied by 18.6 and 2.86,
respectively.
c
When H2O acts as a collision partner M, the rate coefficient is multiplied by 5.

Solution
Although we will use CHEMKIN software to solve this problem, it is instructive to outline the
solution steps as if we were solving the problem from scratch. Examination of the chemical
mechanism above shows that there are 11 species involved: H2, H, O2, O, OH, HO2, H2O2,
H2O, N2, N, and NO.
Thus, we need to write 11 equations of the form of Eqn. 6.31 (or 10 equations of this form
if we choose to invoke ΣYi, out = 1). The Yi, in values are readily calculated from the given stoi-
chiometry, assuming that the only species present in the inlet stream are H2, O2, and N2. For
Φ = 1, a = 0.5 in the combustion reaction, H 2 + a(O 2 + 3.76N 2 ) → H 2 O + 3.76aN 2 ; thus,

χ H2, in = 1/3.38 = 0.2959,

χ O2, in = 0.5 /3.38 = 0.1479,
χ N2, in = 1.88 /3.38 = 0.5562,
χ H,in = χ O2 , in = χ OH, in = . . . = 0.

Using these mole fractions, the reactant mixture molecular weight, Σci MWi , is calculated
to be 20.91. The corresponding mass fractions, ci MWi / MWmix, are

YH = 0.0285,
2 , in

YO = 0.2263,
2 , in

YN = 0.7451,
2 , in

YH, in = YO, in = YOH, in = . . . = 0.

Choosing the equation for O atoms to illustrate one of the equations in the set (Eqn. 6.31), we
write

PMWmixV
ω O MWOV − Y = 0,
RuT t R O, out

tur80199_ch06_183-219.indd 201 1/4/11 12:18:22 PM

 202 C H A P T E R 6 • Coupling Chemical and Thermal Analyses

 , and
where Eqns. 6.37 and 6.38 have been combined to eliminate m

ω O = − k6 f [H][O]( P /(RuT ) + 4[H 2 O])

+ k6r [OH]( P /(RuT ) + 4[H 2 O])
− 2 k7 f [O]2 P /(RuT ) + 2 k7r [O 2 ]P /(RuT )
− k11 f [O][OH] + k11r [O 2 ][H] − k12 f [O][H 2 ]
+ k12r [OH][H] − k15 f [O][HO 2 ]
+ k15r [O 2 ][OH] + k16 f [OH]2 − k16r [O][H 2 O]
− k21 f [O][ N 2 ] + k21r [ NO][ N]
+ k22 f [ N][O 2 ] − k22r [ NO][O].

In the above expression, note that P / (RuT) has been substituted for [M] and how the en-
hanced third-body efficiency for H2O in reaction 6 has been treated. Note also that outlet
quantities are implied for all [Xi]s. Since both YO, out and [O] appear in the O-atom conser-
vation expression, Eqn. 6.33 is needed to express one in terms of the other. As expanding
this expression is straightforward, this step is left to the reader. Again, Eqn. 6.33 would
be employed for each of the 11 species. Closure is obtained to our system of 23 unknowns
(11 [Xi]s, 11 Yi, outs, and Tad) by applying a single conservation-of-energy expression,
Eqn. 6.35:

YH , in hH (298) + YO , in hO (298) + YN , in hN (298)

2 2 2 2 2 2

= YH hH (Tad ) + YH hH (Tad ) + YO hO (Tad )

2 2 2 2

+ YOH hOH (Tad ) +  + YNO hNO (Tad ).

Of course, a thermodynamic database is needed to relate the hi values and Tad .

The above framework for our solution is embodied in the Fortran computer code
PSR [12]. This code, together with the subroutines contained in the CHEMKIN library [16],
is used to generate numerical results for our particular problem. Input quantities include
the chemical mechanism, reactant-stream constitutents, equivalence ratio, inlet tempera-
ture, and reactor volume and pressure. The residence time is also an input variable. With
some experimentation, a residence time of 1 s was found to yield essentially equilibrium
conditions within the reactor, while blowout occurs near tR = 1.7 · 10−5 s. Figures 6.9
and 6.10 show predicted mole fractions and gas temperatures within this range of resi-
dence times. As expected, the adiabatic temperature and H2O product concentration drop
as residence times become shorter; conversely, the H2 and O2 concentrations rise. The
behavior of the O and OH radicals is more complicated, with maxima exhibited between
the two residence-time limits. The NO concentrations fall rapidly as residence times fall
below 10−2 s.
Comment
This example provides a concrete illustration of how relatively complex chemistry couples
with a simple thermodynamic process. A similar analysis was used to generate the reaction
pathway diagrams for CH4 combustion shown in Chapter 5.

tur80199_ch06_183-219.indd 202 1/4/11 12:18:22 PM

 Well-Stirred Reactor 203

Figure 6.9 Predicted mole fractions of H2O, H2, O2, and OH in a well-stirred reactor for
conditions between blowout (tR ≈ 1.75 ⋅ 10–5 s) and near equilibrium (tR = 1 s).

Figure 6.10 Predicted temperatures and O-atom and NO mole fractions in a well-stirred
reactor for conditions between blowout (tR ≈ 1.75 ⋅ 10–5 s) and near equilibrium (tR = 1 s).
Also shown are O-atom concentrations associated with equilibrium at the kinetically derived
temperature together with NO concentrations based on these equilibrium O atoms.

tur80199_ch06_183-219.indd 203 1/4/11 12:18:23 PM

 204 C H A P T E R 6 • Coupling Chemical and Thermal Analyses

Example 6.4 Explore the degree to which nonequilibrium O atoms affect NO formation in the well-stirred re-
actor of Example 6.3. Assume O atoms, O2, and N2 are at their equilibrium values for the reactor
temperatures predicted by the full kinetics. Use the following corresponding temperatures and
residence times from Example 6.3: Tad = 2378 K for tR = 1 s, Tad = 2366.3 K for tR = 0.1 s, and
Tad = 2298.5 K for tR = 0.01 s. Also assume that N atoms are in steady state and that NO forma-
tion is controlled by the simple Zeldovich chain-reaction pair (Chapter 5, Eqns. N.l and N.2):
k1 f
N 2 + O ⇔ NO + N
k1r
k2 f
N + O 2 ⇔ NO + O.
k 2r

Solution
We first formulate the simplified reactor model. Since the temperature and equilibrium values
for cO, χ O , and χ N will be inputs to the model, we need only write the NO species conserva-
2 2
tion equation (Eqn. 6.31):

ω NO MWNOV − mY
 NO = 0,

where YNO is the NO mass fraction within the reactor. Rearranging the above and converting
YNO to a molar concentration basis (Eqn. 6A.3) yields
m
ω NO − [ NO] = 0,
ρV
or, more simply,

ω NO − [ NO] /t R = 0. (I)
We now need only to apply the simple Zeldovich mechanism to express ω NO in Eqn. I in
terms of presumably known quantities (T, [O2]e, [O]e) and the unknown [NO]. The result is a
complex transcendental equation with [NO] as the only unknown variable. Accordingly,
ω NO = k1 f [O]e [ N 2 ]e − k2r [ NO][O]e + [ N]ss ( k2 f [O 2 ]e − k1r [ NO]). (II)
Applying the steady-state approximation for the N atom, we obtain
k1 f [O]e [ N 2 ]e + k2r [ NO][O]e
[ N]ss = . (III)
k1r [ NO] + k2 f [O 2 ]e

Substituting this result for [N]ss back into our expression for ω NO yields

ω NO = k1 f [O]e [ N 2 ]e ( Z + 1) + k2r [ NO][O]e ( Z − 1), (IV)

where
k2 f [O 2 ]e − k1r [ NO]
Z= .
k1r [ NO] + k2 f [O 2 ]e

Before we can solve Eqn. I for [NO], values for [O2]e , [O]e , and [N2]e are needed. Since
several equilibrium expressions are involved, finding these values is nontrivial; however, we
can make this job simple by employing the code TPEQUIL (Appendix F). Although the code

tur80199_ch06_183-219.indd 204 1/4/11 12:18:23 PM

 Well-Stirred Reactor 205

is set up to deal with fuels of the type Cx HyOz, neither x nor y may be identically zero, so deal-
ing with pure H2 does not appear possible. We can still utilize the code, however, by setting x
equal to unity and y to some very large integer, say 106. The small amount of carbon allows
the code to avoid division by zero, without introducing any significant error in the evaluation
of the species mole fractions. For example, CO and CO2 concentrations are less than 1 ppm
for all calculations performed for this example. The following table shows the equilibrium O,
O2, and N2 mole fractions calculated using TPEQUIL :

tR (s) Tad (K) χO, e χO2, e χN2, e

1.0 2378 5.28 · 10−4 4.78 · 10−3 0.6445

0.1 2366.3 4.86 · 10−4 4.60 · 10−3 0.6449
0.01 2298.5 2.95 · 10−4 3.65 · 10−3 0.6468

Equation I was solved using a Newton–Raphson algorithm, i.e., [NO]new = [NO]old − f ([NO]old) /
f ′([ NO]old ), where f ([ NO]) ( = 0) represents Eqn. I, implemented by spreadsheet software.
Within the spreadsheet, the input mole fractions were converted to molar concentrations, for ex-
ample, [O]e = c O, e P / (RuT), and rate coefficients were evaluated from the following expressions
presented in Chapter 5:

k1 f = 1.8 ⋅ 1011 exp[−38, 370 / T (K)],

k1r = 3.8 ⋅ 1010 exp[− 425 / T (K)],
k2 f = 1.8 ⋅ 10 7 T (K) exp [− 4680 / T (K)],
k2r = 3.8 ⋅ 10 6 T (K) exp [− 20, 820 / T (K)].

The results of this iterative solution for the three residence times are shown below, along with
the full-chemistry results generated in Example 6.3 for comparison. These results also are
plotted in Fig. 6.10.

Equilibrium O-atom Assumption Full Chemistry

tR (s) χO, e (ppm) χNO (ppm) χO (ppm) χNO (ppm)

1.0 528 2,473 549 2,459

0.1 486 1,403 665 2,044
0.01 295 162 1,419 744

Comments
Figure 6.10 shows the dramatic departure of the kinetically derived O-atom concentration
from the corresponding equilibrium values as residence times are reduced below 1 s. This is
a result of there being insufficient time for the recombination reactions to form stable species
from the radicals. In turn, the superequilibrium O-atom concentration results in greater NO
production; for example, at tR = 0.01 s, the NO concentration predicted by the full kinetics is
nearly five times that based on the assumption of equilibrium O atoms.

tur80199_ch06_183-219.indd 205 1/4/11 12:18:24 PM

 206 C H A P T E R 6 • Coupling Chemical and Thermal Analyses

PLUG-FLOW REACTOR

Assumptions
A plug-flow reactor represents an ideal reactor that has the following attributes:
1. Steady-state, steady flow.
2. No mixing in the axial direction. This implies that molecular and / or turbulent
mass diffusion is negligible in the flow direction.
3. Uniform properties in the direction perpendicular to the flow, i.e., one-dimensional
flow. This means that at any cross section, a single velocity, temperature, composi-
tion, etc., completely characterize the flow.
4. Ideal frictionless flow. This assumption allows the use of the simple Euler equa-
tion to relate pressure and velocity.
5. Ideal-gas behavior. This assumption allows simple state relations to be employed
to relate T, P, r, Yi, and h.

Application of Conservation Laws

Our goal here is to develop a system of first-order ODEs whose solution describes
the reactor flow properties, including composition, as functions of distance, x. The
geometry and coordinate definition are schematically illustrated at the top of Fig. 6.11.
Table 6.1 provides an overview of the analysis listing the physical and chemical prin-
ciples that generate 6 + 2N equations and a like number of unknown variables and
functions. The number of unknowns could be easily reduced by N, by recognizing that
the species production rates, ω i, can be immediately expressed in terms of the mass
fractions (see Appendix 6A) without the need to explicitly involve the ω 1 . Explicitly

Table 6.1 Overview of relationships and variables for plug-flow reactor with N species

Number of
Source of Equations Equations Variables or Derivatives Involved
d ρ d v x d P d h d Yi
Fundamental conservation 3+N , , , , (i = 1, 2, … , N ), ω i (i = 1, 2, … , N )
principles: mass, dx dx dx dx dx
x-momentum, energy,
species
Mass action laws N ω i (i = 1, 2, … , N )
d ρ d P dT d MWmix
Equation of state 1 , , ,
dx dx dx dx
dh dT dYi
Calorific equation of state 1 , , (i = 1, 2, … , N )
dx dx dx
d MWmix dYi
Definition of mixture 1 , (i = 1, 2, … , N )
molecular weight dx dx

tur80199_ch06_183-219.indd 206 1/4/11 12:18:24 PM

 Plug-Flow Reactor 207

Figure 6.11 Control volumes showing fluxes of mass, x-momentum, energy, and species
for a plug-flow reactor.

retaining them, however, clearly reminds us of the importance of chemical reactions in

our analysis. Although not shown in Table 6.1, the following parameters are treated as
known quantities, or functions, and are necessary to obtain a solution: m,  ki (T ), A( x ),
and Q ′′( x ). The area function A(x) defines the cross-sectional area of the reactor as a
function of x; thus, our model reactor could represent a nozzle, or a diffuser, or any
particular one-dimensional geometry, and not just a constant cross-sectional device
as suggested by the top sketch in Fig. 6.11. The heat flux function Q ′′( x ), although
explicitly indicating that the wall heat flux is known, is also intended to indicate that
the heat flux may be calculated from a given wall-temperature distribution.

tur80199_ch06_183-219.indd 207 1/4/11 12:18:24 PM

 208 C H A P T E R 6 • Coupling Chemical and Thermal Analyses

With reference to the fluxes and control volumes illustrated in Fig. 6.11, we can
easily derive the following conservation relationships:

Mass Conservation
d( ρ vx A)
= 0. (6.39)
dx
x-Momentum Conservation
dP dv
+ ρ vx x = 0. (6.40)
dx dx
Energy Conservation

d ( h + v x2 / 2 ) Q ′′ P (6.41)
+ = 0.
dx m
Species Conservation

dYi ω i MWi
− = 0. (6.42)
dx ρvx
The symbols vx and P represent the axial velocity and local perimeter of the reactor,
respectively. All of the other quantities have been defined previously. The derivation
of these equations is left as an exercise for the reader (see problem 6.1).
To obtain a useful form of the equations where the individual variable deriva-
tives can be isolated, Eqns. 6.39 and 6.41 can be expanded and rearranged to yield
the following:
1 dρ 1 d vx 1 d A
+ + =0 (6.43)
ρ d x vx d x A d x
dh d v Q ′′ P
+ vx x + = 0. (6.44)
dx dx m
The ω i s appearing in Eqn. 6.42 can be expressed using Eqn. 4.31, with the [Xi]s
transformed to Yi s.
The functional relationship of the ideal-gas calorific equation of state,
h = h(T , Yi ), (6.45)

can be exploited using the chain rule to relate dh / dx and dT / dx, yielding

dh d T N dYi
= cp + ∑h . (6.46)
dx d x i =1 i d x
To complete our mathematical description of the plug-flow reactor, we differen-
tiate the ideal-gas equation of state,
P = ρ Ru T / MWmix, (6.47)

tur80199_ch06_183-219.indd 208 1/4/11 12:18:25 PM

 Plug-Flow Reactor 209

to yield

1 d P 1 d ρ 1 dT 1 d MWmix (6.48)
= + − ,
P d x ρ d x T d x MWmix d x

where the mixture molecular weight derivative follows simply from its definition
expressed in terms of species mass fractions, i.e.,
−1
⎡N ⎤
MWmix = ⎢ ∑ Yi / MWi ⎥ (6.49)
⎣ i =1 ⎦

and
N
d MWmix 1 dYi
dx
= − MWmix
2
∑ . (6.50)
i =1 MWi d x

Equations 6.40, 6.42, 6.43, 6.44, 6.46, 6.48, and 6.49 contain in a linear fashion the
derivatives dr / d x, dvx / d x, dP / d x, dh / d x, dYi / d x (i = 1, 2, . . . , N), dT / d x, and
d MWmix / d x. The number of equations can be reduced by eliminating some of the
derivatives by substitution. One logical choice is to retain the derivatives dT / d x,
dr / d x, and dYi / d x (i = 1, 2, . . . , N ). With this choice, the following equations con-
stitute the system of ODEs that must be integrated starting from an appropriate set
of initial conditions:

⎛ ⎞ ρ Ru ⎛ ⎞
2 2 ⎛ 1 d A⎞
N
Ru MWmix
⎜ 1 − c MW ⎟ ρ v x ⎜⎝ A d x ⎟⎠ + v c MW ∑ MWiω i ⎜ hi − MW cp T ⎟
dρ ⎝ p mix ⎠ x p mix i =1 ⎝ i ⎠
= ,
dx ⎛ vx ⎞
2
P ⎜1 + ⎟ − ρ vx
2
⎝ cp T ⎠
(6.51)

dT v 2 dρ v x2 ⎛ 1 d A ⎞ 1 N
= x + ⎜ ⎟ −
d x ρcp d x cp ⎝ A d x ⎠ v x ρcp
∑ hiω i MWi , (6.52)
i =1

d Yi ω i MWi
= (6.53)
dx ρvx

Note that in Eqns. 6.41 and 6.52, Q ′′ has been set to zero for simplicity.
A residence time, tR, can also be defined, and one more equation added to
the set:

d tR 1
= . (6.54)
d x vx

tur80199_ch06_183-219.indd 209 1/4/11 12:18:25 PM

 210 C H A P T E R 6 • Coupling Chemical and Thermal Analyses

Initial conditions necessary to solve Eqns. 6.51–6.54 are

T (0) = T0 , (6.55a)

ρ ( 0 ) = ρ0 , (6.55b)

Yi (0) = Yi 0 i = 1, 2, . . . , N , (6.55c)

t R ( 0 ) = 0. (6.55d)

In summary, we see that the mathematical description of the plug-flow reac-

tor is similar to the constant-pressure and constant-volume reactor models in that
all three result in a coupled set of ordinary differential equations; the plug-flow
reactor variables, however, are expressed as functions of a spatial coordinate
rather than time.

APPLICATIONS TO COMBUSTION SYSTEM MODELING

Various combinations of well-stirred reactors and plug-flow reactors are fre-
quently used to approximate more complex combustion systems. A simple il-
lustration of this approach is shown in Fig. 6.12. Here we see a gas-turbine com-
bustor modeled as two well-stirred reactors and a plug-flow reactor, all in series,
with provisions for some recycle (recirculation) of combustion products in the
first reactor, which represents the primary zone (see Fig. 10.4a in Chapter 10).
The secondary zone and dilution zones are modeled by the second well-stirred
reactor and the plug-flow reactor, respectively. To accurately model a real com-
bustion device, many reactors may be required, with judicious selection of the
proportioning of the various flows into each reactor. This approach relies much
on the art and craft of an experienced designer to achieve useful results. Reactor
modeling approaches are often used to complement more sophisticated finite-
difference or finite-element numerical models of turbine combustors, furnaces,
and boilers, etc.

Figure 6.12 Conceptual model of gas-turbine combustor using a combination

of well-stirred and plug-flow reactors.
SOURCE: After Ref. [13].

tur80199_ch06_183-219.indd 210 1/4/11 12:18:25 PM

 Nomenclature 211

SUMMARY
In this chapter, four model reactors were explored: a constant-pressure reactor, a
constant-volume reactor, a well-stirred reactor, and a plug-flow reactor. A description
of each of these systems was developed from fundamental conservation principles and
linked to chemical kinetics. You should be familiar with these principles and be able to
apply them to the model reactors. A numerical example of a constant-volume reactor
was developed employing three species (fuel, oxidizer, and products) with one-step
global kinetics and simplified thermochemistry. With the model, some characteristics
of thermal explosions were elucidated and related to autoignition (knock) in reciprocat-
ing engines. As a second example, an equally simple numerical model of a well-stirred
reactor was developed. This model was exercised to demonstrate the concept of blow-
out and the dependence of the blowout mass flowrate on equivalence ratio. With a firm
grasp of these simple models, you should be in a good position to understand more
complex and more rigorous analyses of combustion systems. Moreover, these simple
models frequently are useful as a first step in analyzing many real devices.

NOMENCLATURE
A Area (m2)
A /F Mass air–fuel ratio (kg / kg)
c p , cp Constant-pressure specific heat (J / kg-K or J / kmol-K)
cv , cv Constant-volume specific heat (J / kg-K or J / kmol-K)
h of , h fo Enthalpy of formation (J / kg or J / kmol)
h, h, H Enthalpy (J / kmol or J / kg or J)
k Chemical kinetic rate coefficient (various units)
m Mass (kg) or reaction order with respect to fuel
m Mass flowrate (kg / s)
m ′′′ Volumetric mass production rate (kg / s-m3)
MW Molecular weight (kg / kmol)
n Reaction order with respect to oxygen
N Number of moles
P Pressure (Pa)
P Perimeter (m)
Q Heat transfer rate (W)
Q ′′ Heat flux (W / m2)
Ru Universal gas constant (J / kmol-K)
t Time (s)
T Temperature (K)
u, u , U Internal energy (J / kmol or J / kg or J)
v Velocity (m / s)
v Specific volume (m3/ kg)
V Volume (m3)
V Velocity vector (m / s)
W Power (W)

tur80199_ch06_183-219.indd 211 1/4/11 12:18:26 PM

 212 C H A P T E R 6 • Coupling Chemical and Thermal Analyses

x Distance (m)
Y Mass fraction (kg / kg)
Greek Symbols
g Specific heat ratio, cp / cv
r Density (kg / m3)
Φ Equivalence ratio
c Mole fraction (kmol / kmol)
ω Species production rate (kmol / s-m3)
Subscripts
ad Adiabatic
BDC Bottom-dead-center
cv Control volume
e Equilibrium
f Forward
F Fuel
G Global
i ith species
in Inlet condition
mix Mixture
out Outlet condition
Ox Oxidizer
Pr Product
r Reverse
ref Reference state
R Residence
s Stoichiometric
ss Steady state
TDC Top-dead-center
x x-Direction
0 Initial
Other
[X ] Molar concentration of species X (kmol / m3)

REFERENCES
1. Douaud, A., and Eyzat, P., “DIGITAP—An On-Line Acquisition and Processing System
for Instantaneous Engine Data—Applications,” SAE Paper 770218, 1977.
2. Nakajima, Y., et al., “Analysis of Combustion Patterns Effective in Improving Anti-
Knock Performance of a Spark-Ignition Engine,” Japan Society of Automotive Engineers
Review, 13: 9–17 (1984).
3. Litzinger, T. A., “A Review of Experimental Studies of Knock Chemistry in Engines,”
Progress in Energy and Combustion Science, 16: 155–167 (1990).

tur80199_ch06_183-219.indd 212 1/4/11 12:18:26 PM

 Problems and Projects 213

4. Spalding, D. B., Combustion and Mass Transfer , Pergamon, New York, 1979.
5. Longwell, J. P., and Weiss, M. A., “High Temperature Reaction Rates in Hydrocarbon
Combustion,” Industrial & Engineering Chemistry, 47: 1634–1643 (1955).
6. Glarborg, P., Miller, J. A., and Kee, R. J., “Kinetic Modeling and Sensitivity Analysis of
Nitrogen Oxide Formation in Well-Stirred Reactors,” Combustion and Flame, 65: 177–
202 (1986).
7. Duterque, J., Avezard, N., and Borghi, R., “Further Results on Nitrogen Oxides Produc-
tion in Combustion Zones,” Combustion Science and Technology, 25: 85–95 (1981).
8. Malte, P. C., Schmidt, S. C., and Pratt, D. T., “Hydroxyl Radical and Atomic Oxygen
Concentrations in High-Intensity Turbulent Combustion,” Sixteenth Symposium (Inter-
national) on Combustion, The Combustion Institute, Pittsburgh, PA, p. 145, 1977.
9. Bradley, D., Chin, S. B., and Hankinson, G., “Aerodynamic and Flame Structure within a
Jet-Stirred Reactor,” Sixteenth Symposium (International) on Combustion, The Combus-
tion Institute, Pittsburgh, PA, p. 1571, 1977.
10. Chomiak, J., Combustion: A Study in Theory, Fact and Application, Gordon & Breach,
New York, p. 334, 1990.
11. Zeldovich, Y. B., and Voyevodzkii, V. V., Thermal Explosion and Flame Propagation in
Gases, Izd. MMI, Moscow, 1947.
12. Glarborg, P., Kee, R. J., Grcar, J. F., and Miller, J. A., “PSR: A Fortran Program for Mod-
eling Well-Stirred Reactors,” Sandia National Laboratories Report SAND86-8209, 1986.
13. Swithenbank, J., Poll, I., Vincent, M. W., and Wright, D. D., “Combustion Design Fun-
damentals,” Fourteenth Symposium (International) on Combustion, The Combustion
Institute, Pittsburgh, PA, p. 627, 1973.
14. Dryer, F. L., and Glassman, I., “High-Temperature Oxidation of CO and H2,” Fourteenth
Symposium (International) on Combustion, The Combustion Institute, Pittsburgh, PA,
p. 987, 1972.
15. Westbrook, C. K., and Dryer, F. L., “Simplified Reaction Mechanisms for the Oxidation of
Hydrocarbon Fuels in Flames,” Combustion Science and Technology, 27: 31–43 (1981).
16. Kee, R. J., Rupley, F. M., and Miller, J. A., “Chemkin-II: A Fortran Chemical Kinetics
Package for the Analysis of Gas-Phase Chemical Kinetics,” Sandia National Laboratories
Report SAND89-8009, March 1991.
17. Heywood, J. B., Internal Combustion Engine Fundamentals, McGraw-Hill, New York,
1988.
18. Incropera, F. P., and DeWitt, D. P., Fundamentals of Heat and Mass Transfer, 3rd Ed.,
John Wiley & Sons, New York, p. 496, 1990.

PROBLEMS AND PROJECTS

6.1 Derive the basic differential conservation equations for the plug-flow re-
actor (Eqns. 6.39–6.42) using Fig. 6.11 as a guide. Hint: This is relatively
straightforward and does not involve much manipulation.

tur80199_ch06_183-219.indd 213 1/4/11 12:18:26 PM

 214 C H A P T E R 6 • Coupling Chemical and Thermal Analyses

6.2 Show that

d( ρ vx A) 1 dρ
=0= +  , etc. (see Eqn. 6.43).
dx ρ dx
6.3 Show that
d ⎛ RT ⎞ 1 dP 1 dρ
⎜ P=ρ u ⎟⇒ = +  , etc. (see Eqn. 6.48).
dx ⎝ MWmix ⎠ P dx ρ dx

6.4 Show that

d MWmix
dx
= − MWmix
2
∑ (dYi /d x ) MWi−1 .
i

6.5* A. Use MATHEMATICA or other symbolic manipulation software to

verify Eqns. 6.51–6.53.
B. Add the heat flux distribution, Q ′′( x ), to the problem defined by
Eqns. 6.51–6.53.
6.6 In the well-stirred-reactor literature, a “reactor loading parameter” is frequently
encountered. This single parameter lumps together the effect of pressure,
mass flowrate, and reactor volume. Can you identify (create) such a param-
eter for the well-stirred-reactor model developed in Example 6.2? Hint: The
parameter is expressed as P a m bV c . Find the exponents a, b, and c.
6.7 Consider a nonadiabatic well-stirred reactor with simplified chemistry, i.e.,
fuel, oxidizer, and a single product species. The reactants, consisting of fuel
(YF = 0.2) and oxidizer (YOx = 0.8) at 298 K, flow into the 0.003-m3 reactor
at 0.5 kg / s. The reactor operates at 1 atm and has a heat loss of 2000 W.
Assume the following simplified thermodynamic properties: cp = 1100
J / kg-K (all species), MW = 29 kg / kmol (all species), h of , F = −2000 kJ/ kg ,
h of , Ox = 0 , and h of , Pr = −4000 kJ/ kg. The fuel and oxidizer mass fractions
in the outlet stream are 0.001 and 0.003, respectively. Determine the tem-
perature in the reactor and the residence time.
6.8 Consider the combustion of a fuel and oxidizer in an adiabatic plug-flow
reactor of constant cross-sectional area. Assume that the reaction is a single-
step reaction with the following stoichiometry and kinetics: 1 kgF + ␯kgOx →
(1 + ␯)kgPr and ω F = − A exp(− Ea / Ru T )[ F ][Ox ]. Assume also the follow-
ing simplified thermodynamic properties: MWF = MWOx = MWPr , cp, F =
cp, Ox = cp, Pr = constant, h of , Ox = h of , Pr = 0, and h of , F = Δhc .
Develop a conservation-of-energy relationship in which the temper-
ature, T, is the principal dependent variable. Express all concentration
variables, or parameters that depend on concentrations, in terms of mass
fractions, Yi . The only unknown functions that should appear in your final

*
Indicates required use of computer.

tur80199_ch06_183-219.indd 214 1/4/11 12:18:27 PM

 Problems and Projects 215

result are T, Yi , and the axial velocity, vx , each being a function of the axial
coordinate x. For simplicity, neglect kinetic energy changes and assume
that the pressure is essentially constant. Hint: You may find species con-
servation useful.
6.9* Create a computer code embodying the simple constant-volume reactor
developed in Example 6.1. Verify that it reproduces the results shown in
Fig. 6.4 and then use the model to explore the effects of P0, T0, and Φ
on combustion times and maximum rates-of-pressure rise. Discuss your
results. Hint: You will need to decrease the time interval between output
printings when combustion rates are rapid.
6.10* Develop a constant-pressure-reactor model using the same chemistry and
thermodynamics as in Example 6.1. Using an initial volume of 0.008 m3,
explore the effects of P and T0 on combustion durations. Use Φ = 1 and
assume the reactor is adiabatic.
6.11* Develop a plug-flow-reactor model using the same chemistry and thermody-
namics as in Example 6.1. Assume the reactor is adiabatic. Use the model to
A. Determine the mass flowrate such that the reaction is 99 percent com-
plete in a flow length of 10 cm for Tin = 1000 K, Pin = 0.2 atm, and
Φ = 0.2. The circular duct has a diameter of 3 cm.
B. Explore the effects of Pin, Tin, and Φ on the flow length required for
99 percent complete combustion using the flowrate determined in
Part A.
6.12* Develop a model of the combustion of carbon monoxide with moist air in
a constant-volume adiabatic reactor. Assume the following global mecha-
nism of Dryer and Glassman [14] applies:

kf
CO + 12 O 2 ⇔ CO 2 ,
kr

where the forward and reverse reaction rates are expressed as

d[CO]
= − k f [CO][H 2 O]0.5 [O 2 ]0.25
dt
d[CO 2 ]
= − kr [CO 2 ],
dt
where
−0.75
⎡ kmol 1⎤ ⎡ −1.674 ⋅ 108 ( J/ kmol) ⎤
k f = 2.24 ⋅ 1012 ⎢⎛⎜ 3 ⎞⎟ ⎥ exp ⎢ ⎥
⎣⎝ m ⎠ s⎦ ⎣ Ru T (K ) ⎦
⎡ −1.674 ⋅ 108 ( J/ kmol) ⎤
kr = 5.0 ⋅ 108 ⎛⎜ ⎞⎟ exp ⎢
1
⎝ s⎠ ⎥.
⎣ Ru T (K ) ⎦

tur80199_ch06_183-219.indd 215 1/4/11 12:18:27 PM

 216 C H A P T E R 6 • Coupling Chemical and Thermal Analyses

In your model, assume constant (but not equal) specific heats evaluated
at 2000 K.
A. Write out all the necessary equations to describe your model, ex-
plicitly expressing them in terms of the molar concentrations of CO,
CO2, H2O, O2, and N2; individual cp values, cp, CO , cp, CO2, etc.; and
individual enthalpies of formation. Note that the H2O is a catalyst so
its mass fraction is preserved.
B. Exercise your model to determine the influence of the initial H2O
mole fraction (0.1–3.0 percent) on the combustion process. Use maxi-
mum rates-of-pressure rise and combustion durations to characterize
the process. Use the following initial conditions: T0 = 1000 K, P =
1 atm, and Φ = 0.25.
6.13* Incorporate the global CO oxidation kinetics given in problem 6.12 in a
model of a well-stirred reactor. Assume constant (but not equal) specific
heats evaluated at 2000 K.
A. Write out all of the necessary equations to describe your model, ex-
plicitly expressing them in terms of the molar concentrations of CO,
CO2, H2O, O2, and N2; individual c p values, cp, CO , cp, CO , etc.; and
2
individual enthalpies of formation.
B. Exercise your model to determine the influence of the initial H2O
mole fraction (0.1–3.0 percent) on the blowout-limit mass flow-
rate. The incoming gases are a stoichiometric mixture of CO and
air (plus moisture) at 298 K. The reactor operates at atmospheric
pressure.
6.14* Incorporate Zeldovich NO formation kinetics in the well-stirred-reactor
model presented in Example 6.2. Assume that the NO formation kinetics
are uncoupled from the combustion process, i.e., the heat lost or evolved
from the NO reactions can be neglected, as can the small amount of mass.
Assume equilibrium O-atom concentrations.
A. Write out explicitly all of the equations required by your model.
B. Determine the NO mass fraction as a function of Φ (0.8–1.1) for
m = 0.1 kg / s, Tin = 298 K, and P = 1 atm. The equilibrium constant for

Kp
1
2 O2 ⇔ O

is given by

Kp = 3030 exp(−30, 790 / T ).

6.15* Develop a model of a gas-turbine combustor as two well-stirred reactors in

series, where the first reactor represents the primary zone and the second

tur80199_ch06_183-219.indd 216 1/4/11 12:18:27 PM

 Problems and Projects 217

reactor represents the secondary zone. Assume the fuel is decane. Use the
following two-step hydrocarbon oxidation mechanism [15]:
kF
C x H y + ⎛⎜ + ⎞⎟ O 2 → x CO + H 2 O
x y y
⎝ 2 4⎠ 2
kCO, f
CO + 12 O 2 ⇔ CO 2 .
kCO, r

The rate expressions for the CO oxidation step are given in problem 6.12,
and the rate expression for decane conversion to CO is given by

d[C10 H 22 ]
= − kF [C10 H 22 ]0.25 [O 2 ]1.5 ,
dt
where
−15, 098 ⎤
kF = 2.64 ⋅ 10 9 exp ⎡⎢ ⎥⎦ (SI units).
⎣ T
A. Write out all of the governing equations treating the equivalence ra-
tios, Φ1 and Φ2, of the two zones as known parameters. Assume con-
stant (but not equal) specific heats.
B. Write a computer code embodying your model from part A. Perform
a design exercise with objectives and constraints provided by your
instructor.
6.16* Use CHEMKIN [16] subroutines to model H2–air combustion and thermal
NO formation, including superequilibrium O-atom contributions, for the
following systems:
A. A constant-volume reactor.
B. A plug-flow reactor.
C. Exercise your models using initial and / or flow conditions provided
by your instructor.
6.17* Consider a simple tube furnace as shown on next page in the sketch. A
natural gas (methane)–air mixture (Φ = 0.9) is burned in a rapid mix-
ing burner at the inlet of the furnace, and heat is transferred from the hot
products of combustion to the constant-temperature wall (Tw = 600 K). As-
sume that the nitric oxide produced by the burner is negligible compared
with that produced in the postflame gases in the furnace. The natural gas
flowrate is 0.0147 kg / s. Assuming that the hot products enter the furnace
at 2350 K and exit the furnace at 1700 K, determine the following, taking
into account the constraints given below:
A. For a furnace diameter of 0.30 m, what is the length of the furnace?
B. What is the nitric oxide (NO) mole fraction at the exit?

tur80199_ch06_183-219.indd 217 1/4/11 12:18:28 PM

 218 C H A P T E R 6 • Coupling Chemical and Thermal Analyses

C. For the same fuel flowrate and stoichiometry, can the NO emissions
be reduced by changing the diameter (D) of the furnace? (Note that
the length (L) of the furnace will have to be changed accordingly to
maintain the same outlet temperature of 1700 K). What are the values
of D and L that result in lower NO if the mean velocity at the entrance
to the furnace is constrained to be within a factor of two (up or down)
from the original design point? On a single graph, plot cNO versus
pipe length for each case.

Additional constraints and assumptions are the following:

1. The pressure is constant at 1 atm.
2. The flow is fully developed, both hydraulically and thermally, within
the tube furnace.
3. Use the Dittus–Boelter equation [18] NuD = 0.023 ReD0.8 Pr 0.3 for your
heat-transfer analysis, where the properties are evaluated at (Tin + Tout) / 2.
Use air properties to simplify your calculations.
4. Assume Zeldovich NO kinetics apply. Do not neglect reverse reactions.
Use the steady-state approximation for N atoms and assume that O2
and O are in simple equilibrium. Furthermore, assume that the O2 mole
fraction is constant at 0.02, even though it changes slightly with temper-
ature. Use the equilibrium constant Kp = 3.6 · 103 exp[−31,090 / T (K)]
for the 12 O 2 ⇔ O equilibrium.

APPENDIX 6A
SOME USEFUL RELATIONSHIPS AMONG MASS
FRACTIONS, MOLE FRACTIONS, MOLAR
CONCENTRATIONS, AND MIXTURE
MOLECULAR WEIGHTS
Mole fraction / mass fractions:
χ i = Yi MWmix / MWi (6A.1)

Yi = χ i MWi /MWmix . (6A.2)

tur80199_ch06_183-219.indd 218 1/4/11 12:18:28 PM

 Appendix 6A 219

Mass fraction / molar concentration:

[ X i ] = PMWmixYi /(Ru TMWi ) = Yi ρ / MWi (6A.3)

[ X i ]MWi
Yi = . (6A.4)
∑[ X j ] MW j
j

Mole fraction / molar concentration:

[ X i ] = χ i P / Ru T = χ i ρ / MWmix (6A.5)

χ i = [ X i ] / ∑ [ X j ]. (6A.6)
j

Mass concentration:

ρi = ρYi = [ X i ]MWi . (6A.7)

MWmix defined in terms of mass fractions:

1
MWmix = . (6A.8)
∑ Yi / MWi
i

MWmix defined in terms of mole fractions:

MWmix = ∑ χ i MWi . (6A.9)

i

MWmix defined in terms of molar concentrations:

∑[ X i ]MWi
MWmix = i
. (6A.10)
∑[ X i ]
i

tur80199_ch06_183-219.indd 219 1/4/11 12:18:28 PM