Available online at www.sciencedirect.

com

ScienceDirect
Procedia Engineering 71 (2014) 446 – 453

The Introduction of Criteria Parameter in Spontaneous Combustion

Problem
Quan-bing Luoa,b, Li Wanga,b, Dong Lianga,b*
a
School of Engineering, Sun Yat-sen University, Guangzhou 510006, China
b
Guangdong Provincial Key Laboratory of Fire Science and Technology, Guangzhou 510006, China

Abstract

The governing equation of spontaneous combustion is a partial differential equation (PDE) of heat conduction. With the help of numerical
simulation, we can get the solution of this problem. But in order to identify the risk of spontaneous combustion in engineering problem,
the method of numerical simulation seems too complex and is hard to be used by engineers. So we wish that we can get a criteria
parameter to identify the risk of spontaneous combustion more easily. There are two theories about the spontaneous combustion-the
theory of Semenov and the theory of Frank-Kamenetskii. In this thesis, we will introduce their theories first and then we’ll introduce the
theory of criteria parameter in the problem of spontaneous combustion through analyzing the mathematical model and simplifying the
problem. At last, we’ll give our criteria to identify the risk of spontaneous combustion problem.

© 2014
© 2014 The
The Authors.
Authors. Published
Published by
by Elsevier
Elsevier Ltd. Open access
Ltd. Selection under
and peerCCreview
BY-NC-ND
underlicense.
responsibility of the Academic Committee
Peer-review under
of ICPFFPE 2013. responsibility of School of Engineering of Sun Yat-Sun University

Keywords: spontaneous combustion, criteria parameter, Frank-Kamenetskii

1. Mathematical Model

The mathematical model of spontaneous combustion is actually a partial differential equation (PDE) of heat
conduction.
Governing equation:
From the theory of the heat transfer, we get the governing equation[1].
wT § w 2T w 2T w 2T · Q (T )
a¨ 2  2 + 2 ¸ (1)
wt © wx wy wz ¹ U c
Q (T ) is the heat source, W / m3 .
O is thermal conductivity, W / (m ˜ K )
a O / U c is thermal diffusivity.
And from the theory of chemical reaction kinetics, we get the approximate expression of heat source[2].That is
Q c1 exp   E / RT  ,  c1 ! 0  (2)
E is the activation energy of chemical reaction, J / mol .
R is universal gas constant,which is 8.314 J /  mol ˜ K 

*
Corresponding author. Tel.: +86-20-39332230.
E-mail address: gzliangd@163.com

1877-7058 © 2014 The Authors. Published by Elsevier Ltd. Open access under CC BY-NC-ND license.
Peer-review under responsibility of School of Engineering of Sun Yat-Sun University
doi:10.1016/j.proeng.2014.04.064
 Quan-bing Luo et al. / Procedia Engineering 71 (2014) 446 – 453 447

We can get the solution with the initial condition and the boundary condition. But that is still a very complex problem,
we can only get the numerical solution with numerical method. In the thesis[3] , we have introduced some numerical
method with the help of MATLAB. And through contrast with the analytical solution, we have proved its rationality. And
in that thesis, we have discussed the analytical solution of a one-dimensional problem.
As some parameters cannot be got easily and accurately, there would be a great gap between the result of numerical
simulation and the result of experiment. In the thesis[4] , we can estimate these parameters with the MCMC (Markov Chain
Monte Carlo) method.
However, we find that it's hard to get the solution of PDE function. So we should need other good methods to solve the
spontaneous combustion in the engineering problem.

2. Theory of Spontaneous Combustion

2.1. The Theory of Semenov

Semenov theory tells us that the happening of spontaneous combustion is the result of combined action of heat
generation and heat dissipation[5]. If heat generation dominate in the process, spontaneous combustion would happen in the
end. On the other hand, if heat dissipation dominate in the process, spontaneous combustion would not happen in the end.
So, for the aim of simplifying the problem, he made the hypotheses below:
ķ The temperature of outside environment keeps steady.
ĸ The temperature of the spontaneous combustion system is uniformly distributed.
Ĺ The convective heat transfer coefficient of the system is a constant.
Based on these hypotheses, we will explain his theory through the figure below.

Fig. 1. The Theory of Semenov

In the figure, q1 , q1c stand for the rate of heat generation (the heat generated by heat source), q2 , q2c , q2cc stand for the
rate of heat dissipation (through the method of heat convection).

q1 V Q (T ) c1V exp   E / RT  (3)

q2 hS (T  Tf ) (4)
V is the volume . S is surface area.
So, from the theory of Semenov, only when the curve of q1 is totally above the curve q2 , spontaneous combustion can
happen in the end. The critical situation is that there is only one point of tangency between the curve of q1 and q2 .
But we should know that the theory of Semenov is based on three hypotheses. In these hypotheses, the uniformly
distributed temperature of the spontaneous combustion system is very hard to be satisfied in the actual situation. So the
theory of Semenov can not forecast the risk of spontaneous combustion accurately.
448 Quan-bing Luo et al. / Procedia Engineering 71 (2014) 446 – 453

2.2. The Theory of Frank-Kamenetskii

The theory of Frank-Kamenetskii tells us that if the temperature of the system can reach a steady state in the end, the
phenomenon of spontaneous combustion cannot be gotten. Otherwise the spontaneous combustion would happen in the end.
So the problem is whether we can reach a steady state. We can explain his theory through the figure below.

Fig. 2. The Theory of Frank-Kamenetskii

That means whether there is a solution of the steady state conduction function below.
w 2T w 2T w 2T Q
   0 (5)
wx 2 wy 2 wz 2 O
The advantage of theory of Frank-Kamenetskii is that it gave us a more reasonable theory to solve the spontaneous
combustion problem.
But the disadvantage is that it is still too hard to identify whether there is a solution of the steady state conduction
function under boundary condition. So we should think other ways to simplify and solve this problem.

3. Criteria Parameter of Spontaneous Combustion

From the theory of Frank-Kamenetskii, we want to know whether the solution of the steady state conduction function
existed. It is very hard for the three-dimensional problem, so we should simplify it to one-dimensional problem. We should
use a characteristic dimension l to describe the size of the problem. Then we will give the governing equation and
boundary conditions of the problem. That is like the problem of an infinite plane.

Fig. 3. The One-dimensional Problem

Governing equation:
d 2T Q
 0 (6)
dx 2 O
 Quan-bing Luo et al. / Procedia Engineering 71 (2014) 446 – 453 449

Boundary conditions:
­ dT
° 0,
° dx x 0
® (7)
°O dT h( T x l  Tf ),
°̄ dx x l
Heat source:
Q c1 exp   E / RT  ,  c1 ! 0  (8)
But it is too hard to get analytical solution for heat source in this form. We should think about a method to simplify it
to an accepted form. Now we’ll start it.
1
When x is a very small number, we can get (1  x) | 1  x . So we can get the approximation below.
E E
 
RT R (T Tf Tf )
e e
E

RTf [1 (T Tf )/ Tf ]
e
(9)
E [1 (T Tf )/ Tf ]

|e RTf

E E (T Tf )

RTf2
e RTf
˜e
E E (T Tf )

Q c1e RTf
˜e RTf2
,  c1 ! 0  (10)
E (T  Tf )
To simplify and nondimensionalize the form of our function, we define two new variables T
RTf2
x
and x1 . So we can get T RTf2T / E  Tf and x lx1 . Then we can get
l
d 2T
 G eT 0 (11)
dx12
E
El 2 
G c e RTf (12)
O RTf 2 1

Now we define dimensionless number G as the criteria parameter of spontaneous combustion. It decides whether there
is analytical solution of our problem. Of course, it means whether the spontaneous combustion can happen or not. There is a
critical value G cr of the criteria parameter of spontaneous combustion.

4. Critical Value of Criteria Parameter

Now we’ll identify the range of criteria parameter.

dT d 2T dp dp dT dp
To the governing equation, suppose p , so p , then the governing equation can
dx1 dx12 dx1 dT dx1 dT
be turned to
dp
p  G eT 0 (13)
dT
Then
450 Quan-bing Luo et al. / Procedia Engineering 71 (2014) 446 – 453

pdp G eT dT
1 2
Ÿ p G eT  aG
2
dT ­° 2aG ˜ 1  eT / a , ( x  0)
Ÿ p ®
1
(14)
T
dx1 °̄ 2aG ˜ 1  e / a , ( x1 t 0)
­ 1
° dT 2aG dx1 ˜, ( x1  0)
° 1  eT / a
Ÿ®
° 1
dT  2aG ˜ dx1 , ( x1 t 0)
°¯ 1  eT / a
­ 1 1
° x1
° 2aG
³ 1  eT / a
dT  b, ( x1  0)
Ÿ®
°x  1 1
°¯ 1
2aG
³ 1  eT / a
dT  b, ( x1 t 0)

Where a, b are undetermined coefficient. And to the integration, we get

1
³ 1  eT / a
dT

1
³ eT / a T
1 e / a
d (eT / a )

1
1  eT / a o t ³ d (1  t 2 )
(1  t )t
2

(15)
1
2³ dt
t 2 1
t 1
ln C
t 1
§ 1  1  eT / a ·
t o 1  eT / a ln ¨ ¸C
¨ 1  1  eT / a ¸
© ¹
So we can get

­ 1 § 1 1  eT / a ·
° x1 ln ¨ ¸  b, ( x1  0)
°° 2aG ¨© 1  1  eT / a ¹¸
® (16)
° 1 § 1 1 e / a ·
T

° x1  ln ¨ ¸  b, ( x1 t 0)
°¯ 2aG ¨© 1  1  eT / a ¸¹

hl
Now we consider the boundary conditions. Here we suppose the Boit number ( Bi ) of spontaneous combustion
O
problem is very big. So the temperature of boundary can be equal to the temperature of environment. And considering
symmetry of boundary conditions, we change the boundary conditions as follows.
 Quan-bing Luo et al. / Procedia Engineering 71 (2014) 446 – 453 451

T x1 1
Tx
0ˈ 1
0 (17)
1

With boundary conditions, we get

­ 1 § 1  1  1/ a ·
°1 ln ¨¨ ¸b
° 2aG © 1  1  1/ a ¸¹
® (18)
° 1 § 1  1  1/ a ·
°1  2aG ln ¨¨ 1  1  1/ a ¸¸  b
¯ © ¹
So
2
1 § 1  1  1/ a ·
G ¨ ln ¸ ,a t1 (19)
2a ¨© 1  1  1/ a ¸¹
b 0 (20)
With the help of MATLAB, we can get the range of G . It was showed below.

Fig. 4. The Range of G

From the figure, we can get the range of G . So we can get the approximate value of G cr , that is G cr | 0.88 .

5. Summary

In the analysis above, we introduced the criteria parameter of spontaneous combustion. And we also gave the critical
value of criteria parameter. That is G cr | 0.88 . If G ! 0.88 , from the theory of Frank-Kamenetskii, spontaneous
combustion would happen in the end. Otherwise it would not happen. So it will be an easy way to identify the risk of
spontaneous combustion. The process of using criteria parameter can be showed in the picture below.
452 Quan-bing Luo et al. / Procedia Engineering 71 (2014) 446 – 453

Fig. 5. The Process of Using Criteria Parameter

But we should pay attention to some important points of this theory.

hl
ķ To the boundary conditions, we suppose the Boit number ( Bi ) of spontaneous combustion problem is very
O
big. This would make the critical value of criteria parameter G cr a little bigger. But this is acceptable for our problem.
ĸ When we simplify our problem, we give a characteristic dimension l to indicate the influence of size and
geometry. For some regular shape like cube or sphere, we can simplify it through mathematical method. But for our real
problem, it may not be that easy to get the characteristic dimension. That still wait for us to research in the later.

Acknowledgements

This work was supported by funds of Guangdong Provincial Scientific and Technological Project (No.
2011B090400518) and Guangdong Provincial Key Laboratory of Fire Science and Technology (No. 2010A060801010).

References
[1] YANG Shi-ming,TAO Wen-quan, 2006. Heat Transfer (4th Edition)(Chinese), Higher Education Press, Beijng ,China. pp.123-133.
[2] HUANG Yong, 2009. Combustion And Combustor (Chinese), Beihang University Press, Beijng ,China. p.48-58.
[3] LUO Quan-bing et al, 2013. Numerical Methods in Spontaneous Combustion with the Help of MATLAB, Procedia Engineering 52 ( 2013 ) 245 ̢ 253
[4] LUO Quan-bing et al, 2013. MCMC Parameters Estimation in the Inverse Problem of Spontaneous Combustion, Applied Mechanics and Materials
Vols. 444-445 (2013) pp 1440-1445.
[5] ZHANG Ying-hua, HUANG Zhi-an, 2010. Combustion and Explosion(Chinese), Metallurgical Industry Press, Beijng ,China .p.48-58.
 Quan-bing Luo et al. / Procedia Engineering 71 (2014) 446 – 453 453

[1] YANG Shi-ming,TAO Wen-quan, 2006. Heat Transfer (4th Edition)(Chinese), Higher Education Press, Beijng ,China. pp.123-133.
[2] HUANG Yong, 2009. Combustion And Combustor (Chinese), Beihang University Press, Beijng ,China. p.48-58.
[3] LUO Quan-bing et al, 2013. Numerical Methods in Spontaneous Combustion with the Help of MATLAB, Procedia Engineering 52 ( 2013 ) 245 ̢ 253
[4] LUO Quan-bing et al, 2013. MCMC Parameters Estimation in the Inverse Problem of Spontaneous Combustion, Applied Mechanics and Materials
Vols. 444-445 (2013) pp 1440-1445.
[5] ZHANG Ying-hua, HUANG Zhi-an, 2010. Combustion and Explosion(Chinese), Metallurgical Industry Press, Beijng ,China .p.48-58.