Storage of Sugar Cane Bagasse

G. Fareo, G. Weldegiyorgis, H. Laurie, M. Kgatle, N. Mhlongo, S.

Mitchell, T. Myers

Mathematics in Industry Study Group (MISG)

January 15, 2016

Introduction

Problem Description

Desired Outcomes

Model

Discussion

Conclusion

References

Introduction
Sugar cane

Sugar cane breaks

Water added

Bagasse

Sugar Extraction

Figure: Moisture levels of 45 - 55%

Problem Description
Stockpile as a resource
Spontaneous combustion
T. F. Dixon (1988)
B. F. Gray et al (2002)

Figure: One Dimensional Model with an insulated bottom

Desired Outcomes

Maximum height of the bagasse heap to avoid spontaneous combustion?

Advantage in adjusting the moisture? (Usable energy per unit area)
Advantage in pelletizing the bagasse?(Usable energy per unit area)

1D-Model formulation: B. F. Gray et. al 2001

Governing equations
U
= Qb ZW exp(E /RU)
t
+ Qw b Zw XW exp(Ew /RU)f (U)

(b cb + mw Xcw )

+ Lv [Zc Y Ze X exp(Lv /RU)] + 2 U, (1)

Y
= Ze X exp(Lv /RU) Zc Y + DY 2 Y ,
t
X
= Ze X exp(Lv /RU) + Zc Y ,
t

(2)
(3)

W
= F b ZW exp(E /RU) F b Zw XW exp(Ew /RU)f (U)
t
+ Dw 2 W . (4)
U is temperature, Y is vapour concentration, X is liquid concentration,
W is oxygen concentration

1D-Model formulation contd

Boundary Conditions
At the bottom, x = 0, we impose the no flow condition (of heat or
material)
U
= 0,
x

Y
= 0,
x

W
= 0,
x

(5)

At the top surface, x = L,

k

U
Y
W
= h(U Ua ), DY
= hY (Y Ya ), DW
= hW (W Wa ), (6)
x
x
x
Initial Conditions
U(x, 0) = U0 (x),

Y (x, 0) = Y0 (x),

(7)

X (x, 0) = X0 (x),

W (x, 0) = W0 (x).

(8)

1D-Model formulation contd

Steady-state equations


2Y
Lv
Zc Y
0 = DY
+ Ze X exp
x 2
RU


Lv
0 = Ze X exp
+ Zc Y
RU


Lv
Zc Ya
Yxx = 0 Ys = Ya , Xs =
exp
Ze
RU

(9)
(10)

(11)





2U
E
Ew
0 = k 2 + Qb ZW exp
+ Qw b Zw Xs W exp
f (U)
x
RU
RU
(12)

1D-Model formulation contd





2W
Ew
E
0 = DW
F b Zw XW exp
f (U)
F b ZW exp
x 2
RU
RU
(13)
If bagasse is hot (everywhere above 58C), then
k 2U
DW 2 W
+
=0
Q x 2
F x 2

(14)

Applying boundary conditions at x = 0

k
DW
U+
W = C0
Q
F
Y = Ya

W = Wa

(15)
Zc Ya
X =
exp
Ze

Lv
RUi


U = Ui Ua

Dimensionless form
Non-dimensional model
t =

t
,
t

x =

X
,
X =
X

x
,
L

U Ua
U =
,
U

Y
Y =
,
Y

= W ,
W
W

Diffusion time scale is

t =

(16)
L2 (b cb + mw cw X )
L2
=
,
k
DU

The liquid equation is



1
Lv
X
exp
= X exp
Ze t
RUi t

Lv (U 1)
Ua + U U

(17)

!
+ Y ,

(18)

where
Lv =

Lv U
.
RUi

(19)

Dimensionless form
Coefficient of LHS is O(105 ), hence
!
1)

(
U
L
X = exp v
Y
Ua + U U

(20)

Lose terms in heat and vapour equations

Vapour equation
Y

2 Y
Y
=
,
x2
t

where

Y =

L2
= O(101 ).
tDY
!
E (U 1)
Ua + U U
!
Ew (U 1)

f (U),
Ua + U U

(21)

U
2 U
exp
(1 + 2 X )
=
+ AE W
x2
t
exp
+ AEw X W

(22)

Dimensionless form

where
mw cw XL2
b cb L2
2 =
kt
 kt
Qb Z WL2
E
AE =
exp
kU
RUi


2
Ew
Qw b Zw X WL
exp
,
AEw =
kU
RUi
E U
Ew U
E =
,
Ew =
RUi
RUi

1 =

(23)
(24)
(25)
(26)

Dimensionless form
The oxygen equation becomes

W
2W
exp
W
=
BE W
x2
t

E (U 1)
Ua + U U

exp
BEw X W

Ew (U 1)
Ua + U U

f (U),

(27)

where
W



L2
F b ZL2
E
=
BE =
exp
tDW
DW
RUi

(28)

BEw



F b Zw XL2
Ew
=
exp
.
Dw
RUi

(29)

Dimensionless form
Boundary conditions
U
= 0,
At x = 0 :
x

Y
= 0,
x

W
= 0,
x

at x = 0,
(30)

At

x = 1 :
U

= U,
x

where
hL
=
,
k

Y =

Y
= Y (Y 1),
x

hY L
,
DY

W =

W
1),
= W (W
x
(31)

hW L
.
DW

(32)

Note = O(10), Y = W = O(105 ) so we may simplify the boundary

= 1 at x = 1. The initial conditions are
conditions Y = W
U = U0 ,

Y = Y0 ,

W = W0 ,

at t = 0

(33)

Discussion
Simplest model

Steady-state temperature

2 U
0=
+ A exp
x2

(U 1)
Ua + U U

This is standard form, but leads to very small piles

(34)

Discussion
What happens when the density is not assumed constant?

Discussion
Pseudo steady-state
W , Y small
L (U 1)
X = exp v
Ua + U U

!
Y

Y = 1

!
E (U 1)
Ua + U U
!
1)

(
U
E
w
exp

+ AEw X W
f (U),
Ua + U U
!

2W

(
U

1)
E
exp
0=
BE W
x2
Ua + U U
!

(
U

1)
Ew
exp

BEw X W
f (U),
Ua + U U

(35)

U
2 U
exp
(1 + 2 X )
+ AE W
=
x2
t

(36)

(37)

Discussion
Almost full problem
L (U 1)
X = exp v
Ua + U U

!
Y

Y
2 Y
=
,
x2
t
2 U
U
exp
+ AE W
(1 + 2 X )
=
x2
t

(38)

(39)
!

E (U 1)
Ua + U U
!
1)

(
U
E
w
exp

+ AEw X W
f (U),
Ua + U U
!

2W
W

(
U

1)
E
exp
=
W
BE W
x2
t
Ua + U U
!

(
U

1)
Ew

exp
BEw X W
f (U),
Ua + U U

(40)

(41)

Note, insulated bottom and 100% humidity. Pile height increases with
lower humidity

Typical evolution of temperature

Often appears piles can be very large without ignition but ...

What if it rains?

Ignition model words

Puzzle: why do apparently stable heaps ignite after getting soaked?

wet reaction is fast, but turns off for temperatures above 58 C,
dry reaction is slower
So, near centre the bagasse dries out and starts to heat above the 58 C
limit. We imagine two steady states:
inner: hot and dry, insulated inner end, at 58 C at interface
outer: warm and wet, 58 C at interface, cooling condition at surface
Stefan problem with moving boundary

Ignition model diagram

Please use conformal mapping to

imagine this as a
square with a hot
yellow and a warm
red band . . .

We have a good handle on equations for the steady states, but havent got
a formulation for the velocity of the moving interface

Conclusion
We have a model for temperature evolution in bagasse piles - can be
made simpler

Conclusion
We have a model for temperature evolution in bagasse piles - can be
made simpler
Steady-state models should be sufficient - to provide bifurcation
diagram

Conclusion
We have a model for temperature evolution in bagasse piles - can be
made simpler
Steady-state models should be sufficient - to provide bifurcation
diagram
For any ambient conditions we can cause ignition, by making the pile
sufficiently large

Conclusion
We have a model for temperature evolution in bagasse piles - can be
made simpler
Steady-state models should be sufficient - to provide bifurcation
diagram
For any ambient conditions we can cause ignition, by making the pile
sufficiently large
Under normal conditions pile does not burn, but adding water can
then cause ignition

Conclusion
We have a model for temperature evolution in bagasse piles - can be
made simpler
Steady-state models should be sufficient - to provide bifurcation
diagram
For any ambient conditions we can cause ignition, by making the pile
sufficiently large
Under normal conditions pile does not burn, but adding water can
then cause ignition
We have looked at a worst case scenario - insulated bottom, no heat
loss at sides. Model can be improved.
Future work will constitute consideration of a more realistic boundary
condition at the bottom, 2D model with heat loss at the sides;
compare full system to simplified models.

References

B. F. Gray, M. J. Sexton, B. Halliburton, C. Macaskill.

Wetting-induced ignition in cellulosic materials. Fire Safety J 37
(2002) 465 - 479.

T. F. Dixon. Spontaneous combustion in bagasse stockpiles.

Proceedingsof the Australian Sugar Cane Technology, Mackay,
Queensland, Australian, 1988, p. 53 - 61.

C. Macaskill, M. J. Sexton, B. F. Gray. A reaction diffusion model of

stored bagasse. Anziam J 43 (2001) 13 - 35.